Barry Kauler

Blog archived

ANNOUNCEMENT, OCTOBER 30, 2012

This blog is archived, read-only.

The gzipped database is almost 8MB, so I have decided that is big enough.

My new blog:

http://bkhome.org/blog2/

Blog fixed

The Category and number of comments of each post was not displaying. I spent all day yesterday trying to find out what was wrong with the database -- PPLOG has it's own flat-file plain-text database.

I got somewhat depressed as I couldn't fix it.

Then last night I installed a few other blogs and played with them for awhile.

Then this morning I took a fresh look at my PPLOG blog. It occurred to me that the problem might be with my host, hostgator.com, some subtle change in how Perl works.

That's what it turned out to be. I have just about zero knowledge of Perl, so it was a painful process, but I discovered a bug in some "print" statements. My guess is that Hostgator have updated Perl, which has a subtle change in syntax requirement. Hmmm, I would have expected this of Python, not Perl.

Anyway, I was able to fix the syntax.

Blog broken

Something has just gone wrong with my blog, it is not displaying properly.

I will disable posting until I can fix it.

Pfontview, fc-query

Thunor has developed Pfontview, a TTF viewer. It requires the 'fc-query' utility from the 'fontconfig' package.

However, Woof has a template for fontconfig that excludes fc-query. Remedied. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/390bf2138d

Forum thread for Pfontview:
http://murga-linux.com/puppy/viewtopic.php?t=81534

I tried Pfontview 0.1.1 in Precise Puppy, but get this error:

# pfontview /usr/share/fonts/default/TTF/DejaVuSans-BoldOblique.ttf 



** (gtkdialog:4773): ERROR **: gtkdialog: Error in line 7, near token 'string': syntax error



/usr/bin/pfontview: line 275:  4773 Trace/breakpoint trap   $GTKDIALOG --space-expand=true --space-fill=true --file=$TEMP_DIR/winMain

#

Anyway, in future builds, fc-query will be there, so Pfontview can potentially be a nice addition to Puppy.

Planning SP1 for Precise 5.4

Bugs are being discovered in Precise Puppy 5.4, which is to be expected. Of course, those who monitor my blog and the appropriate Forum thread can apply fixes as they are found.

However, there is definitely going to be a need for a Service Pack 1, for those who already have 5.4, and a corresponding bug-fix release 5.4.1.

I intend to work on this, and read the Precise 5.4 Forum feedback/bugs thread, with a view to creating SP1 in about a week from now. Hopefully, anyway.

Personals
I noticed on the Forum, ttuuxxx is back -- welcome back! For those who don't know, ttuuxxx is a long-time valued Puppy developer/tester, who took time off for a holiday in the USA and Canada.

Precise: Aqualung 20121029

I have compiled Aqualung music player in Precise Puppy 5.4. I got the source from SVN, 2012-10-29.

Aqualung project page:
http://aqualung.factorial.hu/

Configure options:
# ./autogen.sh
# ./configure --build=i486-pc-linux-gnu --prefix=/usr --sysconfdir=/etc --localstatedir=/var --with-jack=no --with-mod=no --with-mpc=no --with-mac=no --with-lua=no --disable-systray

I disabled the systray because a new instance starts in the tray every time a sound file is clicked-on in a ROX-Filer window.

PETs (343K, 333KB):
http://distro.ibiblio.org/quirky/pet_packages-precise/aqualung-20121029-up.pet
http://distro.ibiblio.org/quirky/pet_packages-precise/aqualung_NLS-20121029-i486.pet

Note, Aqualung is in the Ubuntu repo, but requires GTK3 and a lot more dependencies.

Precise: Pmusic bug

I am running Precise Puppy 5.4, which has Pmusic 2.6.6.

Yesterday, I clicked on an .mp3 audio file, and Pmusic failed to start, though it is set as the default music player.
This morning, I have investigated this further, clicking any audio file, Pmusic does not start.

Yet, run 'pmusic' without any file on the commandline, and it does start, and then any music file can be opened and played.

This is a serious bug. How come no one else has reported this in any of the alphas, betas and RCs?

If you are using Precise, do you get this bug?
But then, Pmusic was upgraded to 2.6.6 on 21st September, so maybe the bug came in then?

Slacko: Woof fixes

01micko advised of a fix for the 'gcc_dev' template, and an update for support/slacko/woof_base_hack.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/64d64908fd

Slacko: 'strings' utility

01micko advised me that /usr/bin/strings is missing from Slacko builds. 01micko advised that it is in the 'util-linux' package.

This is a bit weird. The 'strings' utility is not in any util-linux official sources. I looked here:
http://www.kernel.org/pub/linux/utils/util-linux/
Refer also here:
http://en.wikipedia.org/wiki/Util-linux

The 'strings' utility is in the official 'binutils' package source, and this is what all distros (except for Slackware) use. Slackware renames it to 'strings-GNU'.

I was tempted to renamed 'strings-GNU' as 'strings', as this is the standard utility that all other distros use.

However, I went with 01micko's suggestion, and modified the template of 'util-linux' to grab /usr/bin/strings if it exists.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/969643fd37

Extract devx-only packages

The PPM references /root/.packages/woof-installed-packages to determine what packages are built-in to Puppy, and thus already installed.

Until now, packages that are in the devx SFS only, are also listed in woof-installed-packages.

However, this does cause problems if the user does not have the devx SFS loaded. The PPM will think a certain dependency is present, when in fact it isn't.

I had this problem with Python, and I got around it for Precise Puppy 5.4 final by building Python into the main Puppy SFS, despite it's size.

Yesterday, another dependency, 'libpq5', was mentioned in the Forum with the same problem. The PPM thought it is already installed, when it is only in the devx.

I have modified the script 'findwoofinstalledpkgs' in Woof, to exclude devx-only packages from woof-installed-packages, and instead log them in a new file, 'devx-only-installed-packages'.

This new file is also copied into /root/.packages during the build, alongside woof-installed-packages.

Woof commits:
http://bkhome.org/fossil/woof2.cgi/info/88855e95ea
http://bkhome.org/fossil/woof2.cgi/info/3c9e97accd

This does mean that I could move Python back to the devx for future releases of Precise Puppy.
I think that some PETs that 666philb has created recently assume the presence of Python, so that would need to be addressed.

Precise: Exaile 3.2.2

Forum member 666philb has created a PET of Exaile music player for Precise:
http://murga-linux.com/puppy/viewtopic.php?t=81743

I have uploaded the PET to the 'precise' repo (25.4MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/exaile-3.2.2_precice.pet

Icon Finder 1.0

That's unfortunate, I just discovered that Icon Finder is not in the menu. It is in Precise Puppy 5.4, but user's won't know!

You can run it from a terminal window, just run "iconfinder".

I looked at the Forum, SFR has released version 1.0. Precise has version 1.0b3. Forum:
http://murga-linux.com/puppy/viewtopic.php?t=81271

Version 1.0 doesn't fix the menu though. I modified the .desktop file so that Icon Finder will appear in the Graphic category, and have uploaded the fixed PET (9KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/iconfinder-1.0-1.pet

Precise: Openshot 1.4.3

Great, Forum member 666philb has created a PET (and SFS) of Openshot 1.4.3, a video editor that enables you to create your own video DVD:
http://murga-linux.com/puppy/viewtopic.php?t=81680

I have added the PET to the 'precise' repo (57MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/openshot-1.4.3_precise.pet

Good, there's not much in the 'precise' PET repo so far.

Precise Puppy kiosk

Forum member shadower_sc has created this nice tutorial on how to lock-down your pup for a kiosk-type of usage:

http://murga-linux.com/puppy/viewtopic.php?t=81701

This is good, makes it easier for others wanting to do something similar and having to try and discover all the things needed to lock-down their pup.

Precise: executables now shared libraries

Forum member FeodorF posted to me about this. An example is /usr/bin/Xorg, which 'file' reports:

# file /usr/bin/Xorg
/usr/bin/Xorg: ELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.24, BuildID[sha1]=0x85311bab3ebf5578bfe33ac5b43e6910010c83b8, stripped

Whereas a normal binary executable looks like this:

# file /usr/bin/xmodmap
/usr/bin/xmodmap: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.15, BuildID[sha1]=0x65ff654e3e14b27ca4ec06b3e4969d7b333c1b29, stripped

Precise is built with a 'file' PET, not with the Ubuntu 'file' (and 'libmagic1) DEBs. I installed the DEBs, same problem.

It seems that some projects are now creating binary executables with a different format that 'file' does not properly recognise.
The Ubuntu DEB is version 5.11, which is the latest.

Raspberry Pi GPU driver open-sourced

Forum members pakt and Dougal have been very excited to inform us about this:
http://www.raspberrypi.org/archives/2221

Yes, this looks like a game-changer.

Heh, heh, I see from the comments in the above link, there is some criticism regarding just how much is open source and what is still hidden as a firmware/binary blob.

It will be very exciting when a hardware-accelerated X.org driver emerges for that GPU.

I will dust off my RP, but for now have a few things to do with Woof -- work is needed on MoManager and the PPM.

"retro" Precise Puppy 5.4

Precise Puppy 5.4 was announced yesterday:
http://www.bkhome.org/blog/?viewDetailed=03063

I promised that I would upload another build which will work on older non-PAE-capable CPUs. These are some early Pentium-M 400MHz-bus CPUs from around 1995 vintage, and some pre-Pentium-Pro CPUs.

The Wikipedia has further explanation on PAE:
http://en.wikipedia.org/wiki/Physical_Address_Extension

Ok, Here it is:
http://distro.ibiblio.org/quirky/precise-5.4-retro/

It also has a mini-readme:
http://distro.ibiblio.org/quirky/precise-5.4-retro/precise-5.4-retro-README.htm

This is somewhat less tested than the "main" Precise Puppy, so take it as being an unofficial release. Well, it is the same in all respects except for kernel and video drivers.

Important: This retro-Precise has a 3.2.32 kernel, the "main" precise has a 3.2.29 kernel. You will find various kernel drivers in the repositories for the 3.2.29 kernel, and they won't work in the retro-Precise.

It is probably not a good idea for developers to start compiling drivers for the 3.2.32 kernel, as I consider it to be a work-in-progress, optimising it for older hardware, so will probably recompile the kernel soon.

Anyway, the source for the 3.2.32 kernel was posted this morning:
http://www.bkhome.org/blog/?viewDetailed=03065

3.2.32 kernel

I have configured this for i586, smp, no-pae.

The intention is to build Precise Puppy with this kernel, for computers that have a pae-incapable CPU.
However, I was uncertain whether any of these, such as the Pentium-M, are recognised by the kernel as having more than one processor.
If not, then it would have been better to have configured without smp support.

Anyway, as it is, this kernel can be used for all CPUs, from i586 upward.

This is the PET, for use by Woof developers:
http://distro.ibiblio.org/quirky/pet_packages-precise/linux_kernel-3.2.32-i586-smp-nopae-up.pet

The sources:
http://bkhome.org/sources/kernel-3.2.32-nopae/

Woof for Precise 5.4

This is the version of Woof used to build Precise Puppy 5.4:
http://bkhome.org/fossil/woof2.cgi/info/da49a4dae4

Precise Puppy 5.4

This is it, the very first official release of Precise Puppy! Brief announcement:

Precise Puppy is built from Ubuntu Precise Pangolin 12.04.1+ binary DEB packages, hence has binary compatibility with Ubuntu and access to the vast Ubuntu package repository. Couple that with Puppy's tiny size, speed and ease-of-use, and this is one incredible pup!

This is the very first release of Precise Puppy. It is assigned version 5.4 to indicate it's position relative to the other puppies, such as Wary 5.3 and Slacko 5.3.3 (5.4 coming soon).

A lot of work has happened at the "Woof-level" since the release of Wary 5.3 in April 2012 -- of particular importance to Precise are the many enhancements to the Puppy Package Manager (PPM). At the "Precise-level" there has been a very long period of testing and refinement, over several months.

As Ubuntu Precise Pangolin is a 5-year Long Term Supported release, we expect that Precise Puppy will be also.

The detailed announcement and Release Notes:
http://distro.ibiblio.org/quirky/precise-5.4/release-Precise-5.4.htm
...please read!

Download live-CD image (157.5MB):
http://distro.ibiblio.org/quirky/precise-5.4/precise-5.4.iso
...note, files can be extracted from this for other modes of installation.

The "devx" SFS, to turn Puppy into a complete compiling environment (126.3MB):
http://distro.ibiblio.org/quirky/precise-5.4/devx_precise_5.4.sfs

Some extra notes:

Nouveau
This is the open-source nVidia video driver. It is still immature on some hardware -- for example, you just get a black screen. Fallbacks are the 'nv' or 'vesa' X.org drivers, or the commercial nVidia driver -- see Forum if you can't figure out how to fallback to these.

PAE kernel
See the Release Notes above. "Pentium M" CPUs are the problem -- many of these are i686 but lacking PAE capability.
I have decided to do a build of Precise Puppy with non-PAE kernel. Please wait a couple of days for this. It will be announced on my blog.

Martian modem firmware

Rerwin has extracted the martian modem firmware from Woof, to a PET:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=855

I have implemented this in Woof, PET will be uploaded to 'common' repo soon. Ditto, Woof will be uploaded soon.

remasterpup2 fix

Rerwin posted a fix for script 'remasterpup2':
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=840

Precise: libicu48, Python

Getting the PPM to work flawlessly with the Ubuntu repositories is a work-in-progress.

I tested installing the 'scribus' package, and after installation ran "scribus" in a terminal window. It reported 'libicui18n.so.48' missing, and a quick check at packages.ubuntu.com showed that package 'libicu48' is required.

The problem is, the Ubuntu package database for 'scribus' does not show 'libicu48' as a needed dependency. Unfortunately, the Debian and Ubuntu package databases do not always show correct dependencies -- sometimes they show a dependency that actually does not exist, or they miss out a needed dependency.

In the case of libicu48, the scribus developers probably never realised the oversight as libicu48 is likely in all builds of Ubuntu. It is only in a cut-down distro such as Puppy that the oversight shows up.

So, I decided to add 'libicu48' inbuilt to Precise Puppy. I am grumbling about this, as the package is about 8MB. It is needed by webkit, midori, etc, so I suppose has some usefulness being pre-installed.

Then there is Python. Scribus needs Python, but we build Puppy with Python in the 'devx' SFS.
Scribus will not run unless the 'devx' is loaded.
I need an elegant solution to this, but will do that after Precise 5.4-final is released -- for now, will advise users to load the 'devx' for any apps that need Python.

defaults-chooser 0.8.6

Precise RC1 tester bigpup reported that Default Applications Chooser (Setup menu) is not working quite right:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=795

This was sc0ttman's baby, but shinobar has been recently upgrading it. Shinobar has reponded to bigpup's reports, and created a new version, 0.8.6.

PET (6KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/defaults-chooser-0.8.6.pet

Grub4dos menu bug

Bigpup reported that grub4dos appears in three different places in the menu, in Precise RC1.

Hmmm, yes, the latest .desktop file has this:

Categories=X-SystemMemory;X-SetupUtility;System

I have restored it to just this:

Categories=X-SystemMemory

New PET (183KB):
http://distro.ibiblio.org/quirky/pet_packages-common/grub4dos-0.4.4.v1.8.0-1.pet

Precise Puppy RC2

This is version 5.3.97. Download from here:

http://puppylinux.com/test/precise-5.3.97-rc2/

I was unable to upload to ibiblio.org, so have used my own site. Go easy on the downloads!

I have written the Release Notes for 5.4-final, that you will find in the 'Help' menu. As usual, very interesting to see just how much has been done since the last official Puppy release!

Precise: libcurl3-gnutls

Forum member James C reported that dependency 'libcurl3-gnutls' was not installed when 'conky' was installed via the PPM. Fixed.

langpack_ru 20121012

I have updated this PET in the 'noarch' repo, created by rodin.s.

Can't login to ibiblio.org right now, will (hopefully) upload it soon.

Precise: libdbus, pupSysInfo

Forum member 666philb reported that the official libdbus is required for the Qt4 DEB to work. OK, done.

Forum member bigpup has reported that pupSysInfo is broken. This is 'Sys-Info System Information' in the 'System' menu.
This is Forum member TaZoc's baby, and he will have to be notified about the problem.

I have removed it ('sys-info' PET) from the Precise inbuilt package list.

pAVrecord, Precord, VLC-GTK, cursor_themes

I have updated these PETs:

pAVrecord 0.9.1
Precord 8.1.4
VLC-GTK 2.5.9

Forum member bigpup reported that Pcur, the mouse cursor selector, does not properly identify where to obtain the cursor theme PET.
I fixed that, and also have placed the 'cursor_themes' PET into the 'noarch' repo.

Monitor Gamma Calibration fixed

Forum member bigpup reported that this utility, in the X GUI version of the Xorg Wizard, which is script /usr/sbin/xgamma-gui, is broken in Precise Puppy. Bigpup reported that it does work in Lucid Puppy.

This script is actually in Woof, it is not a separate PET.

Various people have worked on this script. I found that both the latest in Woof, and the one in Lucid, are broken.
The one in Lucid is the best. It was worked on by Forum member EW in 2010. So, I used this as the basis.

I fixed a couple of bugs, and internationalized it. The script can be found in the Woof commit.

Woof commits:
http://bkhome.org/fossil/woof2.cgi/info/128b3464d3
http://bkhome.org/fossil/woof2.cgi/info/c238a7cbeb

Neon icons, PPM missing pkg

Forum member vicmz composed the nice simple Neon desk icons that I have used in Precise Puppy RC1.

One of them is a yellow that is difficult to see, and they can all be difficult to see on a non-white background.
I ran them all through mtPaint, changed the yellow to something more visible, and bumped the saturation of all of them to 100% -- perhaps some subtly is lost on some of them, but they now stand out better on non-white backgrounds.

I cannot login to ibiblio right now, so have not yet uploaded the updated Neon PET.

I was discussing recently that the local package database may get out of date. PPM now puts up a message if a package does not exist in a repository, that the local database may need updating (with instructions how to do that).
Updated Woof will be uploaded soon.

Precise: ffmpeg-extra

Forum member shinobar has advised that ffmpeg can be enhanced by installing more DEBs:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=750

This is them:

libavcodec-extra-53_0.8.3ubuntu0.12.04.1_i386.deb
libavdevice-extra-53_0.8.3ubuntu0.12.04.1_i386.deb
libavfilter-extra-2_0.8.3ubuntu0.12.04.1_i386.deb
libavformat-extra-53_0.8.3ubuntu0.12.04.1_i386.deb
libavutil-extra-51_0.8.3ubuntu0.12.04.1_i386.deb
libdirac-encoder0_1.0.2-4build1_i386.deb
libopenjpeg2_1.3+dfsg-4_i386.deb
libpostproc-extra-52_0.8.3ubuntu0.12.04.1_i386.deb
libswscale-extra-2_0.8.3ubuntu0.12.04.1_i386.deb
libvo-aacenc0_0.1.1-2_i386.deb
libvo-amrwbenc0_0.1.1-2_i386.deb

Forum member mcewanw has indicated that these may introduce a problem with libxvid?

Maybe RC1 testers can try shinobar's PET, see if there are any side-effects?

I am on the home-run for RC2/final. It is about 7pm here, I plan to do a few more things tonight, then upload RC2/final tomorrow.
If anyone wants to report on shinobar's PET, and mcewanw's pAVrecord prior to that, please do!

I want to get this pup "out the door". I will give RC2 version number 5.4 and if it looks OK will promote it to the 5.4-final release.

Precise: more wallpaper

Forum member futwerk creates incredible backgrounds for all of our puppies. Here are four more for Precise Puppy:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=765

They are png images and very big. I have put them into a PET, and I converted them to jpg, however I used quality-factor-100, so they are still big (but about one-quarter that of the pngs).

PET (3.2MB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/desk_background_precise2-1.pet

PMirrorget 0.2.1

Forum member gposil wrote this website downloader back in 2009:
http://www.murga-linux.com/puppy/viewtopic.php?p=354753

This is in Precise Puppy RC1, however there is a bug report that the menu icon is missing. There is also an icon missing in a button.

I have fixed it, and assigned version number 0.2.1 (2KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/pmirrorget-0.2.1.pet

It is a very simple script that uses gtkdialog for the GUI and wget to download. I have not internationalized it -- we will target doing that after Precise-final is out.

Icon-Finder, pAVrecord, Precord

Two new PETs, one upgrade!

Icon Finder 1.0b3
This is very handy. Created by Forum member SFR:
http://murga-linux.com/puppy/viewtopic.php?t=81271

pAVrecord 0.9.0
Forum member mcewanw created this audio/video recorder:
http://murga-linux.com/puppy/viewtopic.php?t=81332
Ah, our keen developers! I just now added version 0.8.8 to the PPM database, then checked online now and found 0.9.0 is out, as well as a special version for Precise Puppy!

Precord 8.1.3
Forum member mcewanw has updated this nice application:
http://murga-linux.com/puppy/viewtopic.php?t=49907

I have uploaded all of these to the 'noarch' repo at ibiblio.org, and all three are going to be in Precise Puppy final.

Precise: rerwin: Intel 537* drivers

Rerwin has compiled drivers for the Intel 537, 537aa, 537ea and 537sp modems:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=720

I have added these to the 'common' repo and uploaded them to ibiblio.

Peasy* apps

rcrsn51 has created a range of very useful applications, all named with the prefix "Peasy". rcrsn51 has advised of updates, plus new ones...

PeasyPrint 2.0
http://www.murga-linux.com/puppy/viewtopic.php?p=546198#546198

PeasyPDF 2.2.1
http://murga-linux.com/puppy/viewtopic.php?p=551954#551954

PeasyScale 1.4
http://murga-linux.com/puppy/viewtopic.php?p=617584#617584

PeasyPort 1.7
rcrsn51 reports problems with Superscan, so I am using this instead for next Precise build.
http://murga-linux.com/puppy/viewtopic.php?p=642902#642902

...in fact, all of the above will be in the next Precise Puppy.

PETs (4KB, 12KB, 3KB, 3KB):
http://distro.ibiblio.org/quirky/pet_packages-common/peasyscale-1.4.pet
http://distro.ibiblio.org/quirky/pet_packages-common/peasyport-1.7.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/peasypdf-2.2.1.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/peasyprint-2.0-1.pet

Precise: SeaMonkey 2.13.1

Precise Puppy is at RC1 and there are now supposed to be essential bug fixes only. However, I have compiled the latest SeaMonkey, version 2.13.1, up from 2.12.1.

PET (23.3MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/seamonkey-2.13.1-i686-up.pet

Yep, bigger and bigger, in leaps and bounds, with each new release. The 2.12.1 PET is 22.8MB.

Precise: Firefox 16.0.1

I have compiled Firefox 16.0.1 in Precise Puppy.

PET uploaded (20MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/firefox-16.0.1-i686-up.pet

Ubuntu: Firefox DEB removed

Hmmm. Rodney sent me an email about being unable to download files. At first I didn't understand and asked for more information. Rodney sent a snapshot that showed the problem.

This is something that I did not anticipate. The Ubuntu package repo has, for example, Firefox 11.0 DEB which I think was the version in the first release of Ubuntu Precise Pangolin 12.04.

However, Ubuntu later released an upgrade, 12.04.1, which had Firefox 14.0.

However, Ubuntu has now removed that DEB and replaced it with version 16.0. The original 11.0 is still there:
http://archive.ubuntu.com/ubuntu/pool/main/f/firefox/

This is a nuisance. It makes be think that I should have kept Precise Puppy at the original v12.04. If packages are going to disappear from the Ubuntu repo, then PPM is going to fail to download sometimes.

The PPM does have a database update feature in the Configuration window.

This is a very awkward situation. I am going to have to sync with the latest repo when I build Precise Puppy 5.4-final, but in time some packages will be dropped from the repo, and a database update will be required.
The user will only know something is wrong when a package fails to download.

I had expected, reasonably I think, that when they announced 12.04.1, that all the DEBs for that release would stay put in the repo.

rediATM failure

Most, perhaps all, credit unions in Australia, and at least one bank, use the rediATM network of automatic teller machines.

I am a member of a credit union and have used rediATMs without any problem for years. Until about a month ago...

I tried to draw $900 from a rediATM, but got a "network timeout" message, followed by a "transaction cancelled" message.
I then tried another nearby rediATM and tried to draw $300, same problem.
I then used a Commonwealth Bank ATM to draw from my credit union account, successfully.

About a week later, the periodic statement arrived in the mail from the credit union, and it showed both transactions as having succeeded!
In other words, the amount of $1200 was deducted from my account, although the rediATMs had failed to give me the money.

I filled in a complaint form, which was investigated promptly and my account corrected.

However, this has left me with a deep worry about rediATMs. The communication protocol between ATM and server should be such that the server only recognises the cash has been dispensed when the ATM advises successful dispensing. This is such basic common sense, which leads me to think that the people who have programmed the rediATM communications protocol are incompetent.

I have suffered a loss of confidence, and from now on will use Commonwealth Bank ATMs to draw from my credit union account, despite the $2 fee.

Woof commit for Precise RC1

This is the version of Woof used to build Precise Puppy RC1. That is, it is the latest Woof.

http://bkhome.org/fossil/woof2.cgi/info/5a3bff2e07

Precise Puppy RC1

Nearly there! This is the 5.4 Release Candidate 1. Here is the announcement for the previous release, beta7:
http://bkhome.org/blog/?viewDetailed=03000

RC1 has been given version number 5.3.96, and can be downloaded from here:
http://distro.ibiblio.org/quirky/test/precise-5.3.96-rc1/

The theme that I chose is interesting? Good? Bad? Anyway, this is it:
JWM: stark-blueish
GTK: stark-blueish
Desk icons: Neon
Wallpaper: precise-futwerk-1b

At this stage of course, we only want to make essential bug fixes. I will probably remove hcfpcimodem and hsfmodem drivers from the final -- probably we need to provide documentation to let people know that PETs are available, for those who have those (and other) modems.

Another thing, I have yet to write the Release Notes.

Whether there is an RC2 depends on what we discover with testing RC1. It may be possible to go straight to the final release.

Note, there is no non-PAE build for RC1. I have been thinking of simplifying things. A non-PAE build of Precise Puppy can be done by the Puppy community, along with other modifications.

There are not that many computers (as a percentage) that have i686-class CPUs that are not PAE-capable, so I might not bother with them. People could of course use Wary.

Regarding Wary, I am also thinking of dropping Racy. But, again, someone in the community can build a Racy Puppy if they want.

grub4dos 0.4.4.v1.8.0

Shinobar has two versions of grub4dos:
http://www.murga-linux.com/puppy/viewtopic.php?t=51697

The 'common' repo had v1.7.2, and I have now added v1.8.0 (183KB):
http://distro.ibiblio.org/quirky/pet_packages-common/grub4dos-0.4.4.v1.8.0.pet

I added "+linux_kernel&ge3.0" in the dependency field of v1.8.0, so that when building in Woof the right 'grub4dos' PET will get chosen.

Screeny screen capture

01micko created a very nice screen and window capture application, named Screeny:
http://www.murga-linux.com/puppy/viewtopic.php?t=76204

I have been stuck on the very simple but basic mtPaint capture capability for a long time, I even resisted some nicer-looking GUIs for it.

But, Screeny is very nice, particularly as it easily captures windows, something that I need frequently.

Screeny needs a small utility, 'xwd', and I have modified the xorg_base* templates in Woof to include 'xwd', so it will be in all future puppies.

I also added 'screeny' to the package-lists of most of the build flavours, including Precise Puppy.

PET uploaded (3KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/screeny-0.7-noarch.pet

Racy/Wary (and anyone else without 'xwd') will need this PET (11KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/xwd-1.0.2-racy-i486.pet

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/b53071effd

Precise: Chromium browser

Forum member peebee has created PETs and SFSs for Chromium, for both Slacko and Precise:
http://www.murga-linux.com/puppy/viewtopic.php?t=81351

I added 'libpn14' dependency into the 'pet.specs' file, and applied the modified 'chromium' script to allow older Flash player, and have uploaded the PET (42MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/chromium-22.0.1210.0-i686.pet

The dependency (75KB):
http://distro.ibiblio.org/quirky/pet_packages-precise/libpng14-14-slackware.pet

Precise: rerwin: 536/537 modem drivers

Rerwin has compiled kernel modules for the 3.2.29 kernel in Precise Puppy:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=720

I have uploaded these to the 'common' repo. These are for people to experiment with, I don't intend to put them into the next release of Precise Puppy (417KB, 1.4MB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_module_intel536ep-20120710-k3.2.29-smp.pet
http://distro.ibiblio.org/quirky/pet_packages-common/linux_module_intel537ep-20120710-k3.2.29-smp.pet

They require (6KB, 7KB):
http://distro.ibiblio.org/quirky/pet_packages-common/firmware_linux_module_intel536-20121007.pet
http://distro.ibiblio.org/quirky/pet_packages-common/firmware_linux_module_intel537-20121007.pet
Notes for rerwin:

1.
In 'pinstall.sh':
[ "`pwd`" = "/" ] && depmod
...so that depmod will not execute if the PET is builtin when building a puppy in Woof.

2.
I have fleshed out the pet.specs files:

linux_module_intel536ep-20120710-k3.2.29-smp|linux_module_intel536ep|20120710-k3.2.29-smp||BuildingBlock|1087K||linux_module_intel536ep-20120710-k3.2.29-smp.pet|+linux_kernel&eq3.2.29,+firmware_linux_module_intel536|Intel 536 modem SMP driver|ubuntu|precise||
linux_module_intel537ep-20120710-k3.2.29-smp|linux_module_intel537ep|20120710-k3.2.29-smp||BuildingBlock|3975K||linux_module_intel537ep-20120710-k3.2.29-smp.pet|+linux_kernel&eq3.2.29,+firmware_linux_module_intel537|Intel 537EP modem SMP driver|ubuntu|precise||

Note, I have also done this to the other modem driver PETs.

rerwin: modem firmware PETs

Rerwin has moved more modem firmware out of Woof into PETs. He has also updated some firmware PETs that I uploaded previously. See 'modem_firmware_packages-complete set-20121007.tar.gz' posted to Forum:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=720

The PETs have been uploaded to the 'noarch' and 'common' repos.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/c489357b11

01micko: delete params on Exec line

The script /usr/local/petget/installpkg.sh has code to delete extra parameters prefixed with "%" on the "Exec=" line in .desktop files.
For example "Exec=gimp %U" will have the "%U" deleted.

It seems that Billtoo has reported a problem with this, and 01micko has proposed alternate code:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=720

...which is a bit of a problem for me, as I can't see anything wrong with my code, and I don't have the explicit example of when it "goes wrong".
However, I have implemented 01micko's change. It will be in the next upload of Woof.

01micko posted a useful reference on these extra parameters:
http://standards.freedesktop.org/desktop-entry-spec/latest/ar01s06.html

rerwin: modems: module loading fix

Rerwin has posted improvements to scripts /sbin/pup_event_modprobe_backend_protect and /etc/rc.d/functions4puppy4:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=720

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/20181a4a05

Modem firmware: ltmodem, pctel, slmodem

Rerwin has extracted this firmware from Woof to separate PETs:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=705

I have uploaded the PETs (599KB, 2KB, 2KB):
http://distro.ibiblio.org/quirky/pet_packages-common/firmware_linux_module_slmodem-20120930.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/firmware_linux_module_ltmodem-20120927.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/firmware_linux_module_pctel-20120930.pet

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/321c9a8cce

ESS modem driver

Rerwin has compiled the ESS modem driver for the 3.2.29 kernel used in Precise Puppy, and has extracted the ESS firmware from Woof to a separate PET:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=705

I have uploaded the PETs (2KB, 187KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/firmware_linux_module_ess-20120930.pet
http://distro.ibiblio.org/quirky/pet_packages-common/linux_module_ess-2.6-0.7-k3.2.29.pet

Woof ESS changes (plus a few other things):
http://bkhome.org/fossil/woof2.cgi/info/563611cb4c

gtk-youtube-viewer 3.0.2

666philb has updated gtk-youtube-viewer:
http://www.murga-linux.com/puppy/viewtopic.php?t=76835

I have uploaded the PET for Precise Puppy:
http://distro.ibiblio.org/quirky/pet_packages-precise/gtk_youtube_viewer-3.0.2_precise.pet

rerwin: dgc hsf modem firmware

Rerwin has posted update PETs for the dgcmodem and hsfmodem firmware:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=705

I have uploaded the PETs (30KB, 97KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/firmware_linux_module_dgcmodem-20120927.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/firmware_linux_module_hsfmodem-20120927.pet

Precise: wallpaper files

Forum member futwerk has posted some wallpaper images for Precise Puppy:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=690

Thanks for that!

I have added two of these to my collection (the PET previously only had two images) for Precise (315KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/desk_background_precise-1.1.pet

Precise: dgc, hcf, hsf modem drivers

Rerwin has compiled dgc, hcf and hsf modem kernel drivers for the 3.2.29 kernel used in Precise Puppy:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=645

I have uploaded these to the 'common' repo (8KB, 472KB, 1MB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_module_dgcmodem-1.13-20120901-k3.2.29.pet
http://distro.ibiblio.org/quirky/pet_packages-common/linux_module_hcfpcimodem-1.21full-20120927-k3.2.29.pet
http://distro.ibiblio.org/quirky/pet_packages-common/linux_module_hsfmodem-7.80.02.06full-dellhybrid-20120901-k3.2.29.pet

The 3.2.29 Linux kernel PET is also in the 'common' repo, with the intention that other (non-Precise) Puppy builds can also use it. Although, note that it was compiled in Precise Puppy, so any other 3rd party drivers must be also.

I will include these three modules in the next release of precise, even though the hsfmodem PET is big (1MB).

nobus, dbus library stubs

technosaurus has created a PET of nobus, modified as shared libraries, which are do-nothing stubs for the 'dbus' and 'dbus-glib' libraries. This is for apps linked against those libs but which will still work if they are only stubs.

Forum post:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=645

Uploaded PET in 'common' repo (5KB):
http://distro.ibiblio.org/quirky/pet_packages-common/nobus-0.2-i486.pet

Precise: NVIDIA and AMD drivers

Forum member 666philb has created NVIDIA and AMD video driver PETs for Precise Puppy and the 3.2.29 kernel:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=645

I have added these to the repo and uploaded them (50.7MB, 27.4MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/amd_fglrx-12.8-precise-k3.2.29.pet
http://distro.ibiblio.org/quirky/pet_packages-precise/nvidia-glx-304.43-precise-k3.2.29.pet

PPM: category translations

L18L reported a problem with category names not translated in PPM post-install window:
http://bkhome.org/blog/?viewDetailed=03020

I have fixed it. Maybe a clumsy way to do it, but it is a fix. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/84478fc183

Langpack post-install script

I recently modified the 'pinstall.sh' script in langpack PETs, reported here:
http://bkhome.org/blog/?viewDetailed=03007

See also a slightly earlier report:
http://bkhome.org/blog/?viewDetailed=03006

I have responded to queries from rodin.s and L18L in the top link.

Multiarch: builtin files-list

Reiterating the history, Puppy contains lists of builtin files in /root/.packages/builtin_files, as explained here:
http://bkhome.org/blog/?viewDetailed=01620

We ran into a problem with Ubuntu/Debian multiarch. Puppy introduced symlinks to force files back into "normal" locations:
http://bkhome.org/blog/?viewDetailed=02752

However, those symlinks broke the code that composes the builtin files-lists, as reported by mavrothal:
http://bkhome.org/blog/?viewDetailed=03000

...as the above thread explained, I applied a fix, but it was not 100% successful. So, I have now applied a fix for the fix. The appropriate script is '3builddistro' and is in this Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/98952ea684

defaults-chooser internationalized

Default Applications Chooser version 0.8 is in all recent puppies. This was created by sc0ttman:
http://www.murga-linux.com/puppy/viewtopic.php?t=56282

Vicmz advised that this has been internationalized by shinobar:
http://murga-linux.com/puppy/viewtopic.php?mode=attach&id=57099

Thanks guys, I have updated the PET in the 'noarch' repo (6KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/defaults-chooser-0.8-20120522.pet

...my PET does not have the 'ja' translation (or any others), as that belongs in a langpack.

VLC-GTK 2.5.4

Sc0ttman has updated VLC-GTK, a GUI frontend for VLC commandline media player. See Forum:
http://www.murga-linux.com/puppy/viewtopic.php?t=54753

I have uploaded the PET to the 'noarch' repo (28KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/vlc-gtk-2.5.4.pet

There are two dependencies, 'xdotool' and 'rtmpdump', that I have uploaded to the 'common' repo (78KB, 47KB):
http://distro.ibiblio.org/quirky/pet_packages-common/rtmpdump-2.4-i486.pet
http://distro.ibiblio.org/quirky/pet_packages-common/xdotool-2.201012.3049.pet

Well, also 'vlc_nogui' package is also required.

Planner menu double-entry

I reported that Planner project manager is in the menu under both 'Business' and 'Utility' categories. Foo- replied that the "ProjectManagement" category is listed in both /etc/xdg/menus/puppy-business.menu and puppy-utility.menu, hence appears in both menus.

Yes, we often have this conflict. Project management can apply to both business and code-development projects. I am placing code-development related apps under Utility.

So, I decided to leave the double-entry as-is.

The Planner PET compiled in Wary also works fine in Precise Puppy, so I have added the PET to the Precise build. It's a very good app to have in the base suite builtin to Puppy.

GetFlash 1.2

Shinobar advised that version 1.2 of GetFlash is available. Forum:
http://www.murga-linux.com/puppy/viewtopic.php?t=74491

I have added the PET to the 'noarch' repository on ibiblio (8KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/getflash-1.2.pet

ARM defeated, Haiku OS

Every now and again, I look around to see what is happening with alternative operating systems.

Yesterday, I took a good look at Haiku, which is an open-source re-implementation of beOS (a defunct commercial OS). Actually, what lead me there, indirectly, was I checked out progress on Lima, the open-source project for a GPU driver for the Mali GPUs used in many ARM chips -- only to be very disappointed when I found the last git commit was in February 2012 -- in other words, a lot of activity and publicity late last year, and early in 2012, then the development has faded.

Which makes me think that perhaps the ARM chips are going to miss their window-of-opportunity to supplant x86 in small netbooks and even tablets, and running a good OS. Intel will bring out their low-power 22nm low-power SoC x86 chips, and maybe it is game over? Even if the Intel SoC's consume more power, relative to other components such as screen, overall it won't make much difference.

This lead me onto thinking about x86 OS's, and I also reflected on the Unix security model -- which I don't like.

I was looking at various alternatives to Linux, many of which are extremely immature or too specialised, or not very active. Then I looked at Haiku and was very impressed with the level of active development. This is a very nice OS, and they are looking at a release-candidate in 2013. This has been a long time coming, I think the project started in 2001.

Haiku home:
http://www.haiku-os.org/
Haiku news:
http://dailyhaiku.net/
Introduction to Haiku:
http://spectrum.ieee.org/computing/software/the-dawn-of-haiku-os/0

However, I am on the lookout for an OS that has a builtin security model that is superior to Unix, but Haiku has no security at all. It is a single-user system, which is fine by me, but I still consider there must be security mechanisms -- the Haiku forum has discussion on this, but nothing has been implemented.

I really don't think it is something that can be tacked on later. It needs to be designed into the OS.

Looking at their apps, there isn't even a firewall.

Haiku is designed for traditional mouse/keyboard and there does not seem to be any consideration for running it with a touchscreen. I think that is another serious oversight, as Haiku could have found a niche for itself with touchscreen computers.

...these are just some of my rambling thoughts about ARM, security and Haiku!

back on-board

My eight-day Royal Show job finished yesterday, so now I can catch-up with Puppy stuff.

I have a backlog of other non-Puppy stuff to do also. A lot of paperwork -- one thing, one of my elderly relatives went into a nursing home a little over a week ago, and I am managing all his finances and personal affairs.

Thought-provoking that, seeing someone leaving behind a house with lifetime collection of possessions, just one suitcase packed for the nursing home. Then of course, the "final journey", nothing is taken. Anyway, this kind of personal reflection is a bit off-track for what I intend this blog to be for -- I was going to start "Barry's Other Blog" for that kind of post!

For Puppy development, I will attempt to read all posts and catch up. I will document on this blog as I do things, and if you see after awhile that I have missed something important, let me know.

The objective will be a Precise Puppy RC1 released soon, maybe in a week, also beta2 of Wary and Racy.

Electric vehicles

I have been waiting for this:
http://www.extremetech.com/extreme/136894-will-high-mileage-nissan-leafs-need-costly-battery-replacements-soon

I should really have started "Barry's Other Blog" to write about these things. One of my pet peeves is "knowledge filtering", which means we get incomplete, misleading or wrong information.

Electric vehicles are an example. What is the real cost of these vehicles? This needs to cost based on the total energy required for the vehicle over it's operating life.

What is the environmental impact of manufacturing (and disposal) of these battery packs?

Royal Perth Show job

My Royal Show job starts tomorrow, running for eight days. All day, from dawn until after dark. I will get home about 8.30pm, so don't expect much puppy development from me for the next eight days!

Royal Show page:
http://www.perthroyalshow.com.au/

I'll try to do a little bit each day though, a hour or so.

Woof: built Wary/Racy 5.3.90

This is the Woof version used to build Wary and Racy 5.3.90 (5.4beta1):
http://bkhome.org/fossil/woof2.cgi/info/d574f24630

Wary and Racy 5.3.90 (5.4beta1)

Ok, here they are! These are the first betas for the next release of Wary and Racy, which will be version 5.4. The current official release is 5.3.

Wary 5.3.90
Get it from here:
http://distro.ibiblio.org/quirky/test/wary-5.3.90/

There are some issues with this one. The 'welcome1stboot' window does not appear. It was built without an Adobe Flash Player, also without Getflash. But, you can grab either, either via PPM, or direct (5.1MB, 8KB):
http://distro.ibiblio.org/quirky/pet_packages-common/flashplayer10-10.3.183.23.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/getflash-1.1.pet

Racy 5.3.90
Get it from here:
http://distro.ibiblio.org/quirky/test/racy-5.3.90/

Racy does have both Adobe Flash Player and Getflash. Also 'welcome1stboot' works. Only one issue I have noticed so far: Planner project management appears in both Business and Utility sub-menus.

The kernels are the same as for 5.3. Lots of infrastructure upgrades with Woof, plus many apps updated, including cups now 1.4.8 and seamonkey now 2.12.1.

rerwin: modem fixes

Rerwin posted about fixes to scripts /usr/sbin/modemprobe_erase and pupdial:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=690

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/fdcf56890f

Wary: older Geeqie

Forum member tuxtoo reported that the Geeqie image viewer (from wary5 repo) crashes in Wary:
http://murga-linux.com/puppy/viewtopic.php?t=77303&start=165

Tuxtoo found an older version, compiled back in 2009 by puppyluvr, that works:
http://www.murga-linux.com/puppy/viewtopic.php?p=300260

I have added an icon to puppyluvr's PET and rolled back to this older version in the ibiblio 'wary5' repo. I will upload the PET soon.

Wakepup2, Precord

Forum member Crash has updated Wakepup2, a tool for booting puppy from floppy disk:
http://www.murga-linux.com/puppy/viewtopic.php?t=80291

I have updated the PET in the ibiblio 'common' repo.

Forum member mcewanw is working at a furious pace on Precord, a audio recorder/player app:
http://www.murga-linux.com/puppy/viewtopic.php?t=49907

I have upgraded the PET at ibiblio 'noarch' repo to version 8.0.8.

I will upload the PETs soon.

Improved .desktop translation

MoManager manages translations for the .desktop files in /usr/share/applications.

Previously, only the 'Name' field was translated, I have now added the 'Comment' field.

A langpack PET created by MoManager has /usr/share/applications.in, which is a collection of .desktop files with the appropriate translated 'Name' and 'Comment' fields.
Refer to /usr/sbin/momanager to see how this directory is created, and /usr/share/doc/langpack-template/pinstall.sh for the langpack post-install script (which will perform the translations on /usr/share/applications).

I have extended this so that the Puppy Package Manager will recognise if /usr/share/applications.in exists (meaning a langpack installed) and will apply a translation if a matching .desktop file exists in /usr/share/applications.in when a package is installed.
Refer to script /usr/local/petget/installpkg.sh.

I think that L18L and others were modifying 'pinstall.sh' in their langpacks, but now need to consider return to the official pinstall.sh script.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/1e322e20c7

Fix DejaVu font problem for non-Latin

The guys who want to localise Puppy for Chinese and Japanese (and other non-Latin) languages, have a particular problem as many configuration files specify the 'DejaVu Sans' font.

The DejaVu home page gives an overview of the languages supported:
http://dejavu-fonts.org/wiki/Main_Page

To localise to Chinese or Japanese, a special font package is required to be installed. To get applications such as JWM (window manager) to use it, their config files must replace "DejaVu Sans" with just "Sans".
If the system recognises "Sans" as mapping to the new Chinese/Japanese font set, then it will work.

Actually, in theory the fontconfig mapping could remap "DejaVu Sans" to the new Chinese/Japanese font set, but it seems that no one has figured out how to do that. It would be the simplest solution though.

Anyway, I have modified two scripts to replace all occurrences of "DejaVu Sans" with just "Sans".

/usr/share/doc/langpack-template/pinstall.sh
This file is read by MoManager and becomes the post-install script for a langpack.
There is no Chinese or Japanese langpack yet (there is one for some older 4.x and early 5.x puppies but that is not compatible with recent pups). When someone uses MoManager to create a Chinese/Japanese langpack, it will have this new pinstall.sh in it.

/usr/local/petget/hacks-postinstall.sh
This takes care of replacing "DejaVu Sans" with "Sans" when certain packages are installed by the user.

Note, it is not just JWM that has this "DejaVu problem". There are various JWM and GTK theme PETs, and other PETs such as openbox, seamonkey, firefox.

You can find the updated scripts here:
http://bkhome.org/fossil/woof2.cgi/info/02cc3006a0

Planning for next Wary/Racy

To keep everyone informed about what I am currently doing...

I am reading through the Wary/Racy 5.3 feedback thread:
http://murga-linux.com/puppy/viewtopic.php?t=77303

I plan to build a beta for the next release, which will probably be 5.4.

I don't know about the kernels though, might leave them alone. The thing is, we have some drivers, for example nVidia, compiled for those kernels, and if I upgrade the kernels then those PETs will also require upgrading.

If there is no urgent need to upgrade the kernels, I will leave them alone.

So, the next Wary/Racy will have all the new features of Woof, plus various application upgrades.

Wary: Sakura 2.4.2

Karl Godt compiled Sakura terminal emulator in Racy 5.3:
http://www.murga-linux.com/puppy/viewtopic.php?t=77387

He also compiled the dependency 'vte' and provided a PET, however the Wary5 repo already has the 'vte' package.

I tested KG's Sakura with the current 'vte' PET, works. I tweaked the Sakura PET a bit and have added it to the wary5 repo.

PETs (30KB, 5KB, 14KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/sakura-2.4.2-i686.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/sakura_DOC-2.4.2-i686.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/sakura_NLS-2.4.2-i686.pet

You will also need to install the 'vte' PET, but of course once 'sakura' is in the PPM, this dependency will be taken care of.

Wary: SeaMonkey 2.12.1

I have started to look at getting Wary and Racy organised for the next release.

I was unable to compile SeaMonkey 2.12.1. It required upgrading of libvpx (from 0.9.7 to 1.1.0) and sqlite (from 3.7.10 to 3.7.14), however it seems that the version of gcc in Wary is too old. I did some research on this, fiddled around for sometime, but was unable to compile SM.

So, I used the official binary from the SM website, made it into a PET. This requires 'dbus-glib', so I enabled that (and 'dbus') in the Wary (and Racy) package-list.
Note, it does not use the system libvpx and sqlite, has it's own inbuilt. Ditto for cairo.

As this new PET includes the IRC-chat client module, I added a menu entry under the 'Internet' category.

Sigh... I don't like to see this. Just looking at the SM PETs, version 1.1.19 is 10.1MB, 2.0.11 is 11.2MB, and 2.12.1 is 21.5MB!

The PETs (252KB, 307KB, 21.5MB, 380KB, 443KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/libvpx-1.1.0-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/libvpx_DEV-1.1.0-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/seamonkey-2.12.1-x86.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/sqlite-3.7.14-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/sqlite_DEV-3.7.14-w5c.pet

Gpart 0.1h-11

The docs state:

Gpart is a tool which tries to guess the primary partition table of a PC-type disk in case the primary partition table in sector 0 is damaged, incorrect or deleted.

This utility is in Precise Puppy. Apparently Gparted will use it if it exists.

I have now compiled it in Wary and placed it in the 'common' repo for all builds to use.

PET (20KB):
http://distro.ibiblio.org/quirky/pet_packages-common/gpart-0.1h-patched_11debian-i486.pet

Gpart Debian page:
http://packages.debian.org/squeeze/gpart

I have uploaded the patched source to my sources repo:
http://bkhome.org/sources/alphabetical/g/

Note, Gpart will be in the next build of Wary and Racy.

Woof: built Precise beta7

This is the version of Woof used to build Precise Puppy 5.3.95 5.3.95.1 beta7.

Woof commits:
http://bkhome.org/fossil/woof2.cgi/info/daca77b5c2
http://bkhome.org/fossil/woof2.cgi/info/81b3499837

Precise Puppy beta7

Here is what I hope will be the last beta of Precise Puppy. This is built with 3.2.29 PAE kernel (meaning it supports more than 4GB RAM).

Download:
http://distro.ibiblio.org/quirky/test/precise-5.3.95-beta7-pae/

Kernel sources:
http://bkhome.org/sources/kernel-3.2.29/
Us#er#na#me#: p#up#py P#as#sw#or#d: l#in#ux

I am also building beta7 with non-PAE kernel, targeting older hardware, including older video hardware, which will be uploaded soon.

Langpacks: es, de

I have updated these langpack PETs:

langpack_es 20120905 (vicmz)
langpack_de 20120830 (L18L)

Heh heh, with all of these upgrades, you may have guessed that I am getting ready to build the next beta of Precise Puppy!

Yes, it will be beta7, but we can't just drag on releasing betas. So, I propose that this be the last beta, followed by an RC, hopefully followed by the final release of Precise Puppy.

As beta7 will have the 3.2.29 kernel, which I intend that we will stick with for the final and well into the future, I request peebee and others who have been creating extra kernel modules, to please do so for this kernel, so it will all be available for the RC.

As I want to move onto the RC fairly promptly, I want to minimise changes. In Woof, bug-fixing only, no new features.

Updated PETs from zigbert and mcewanw

From zigbert:

Pburn 3.7.5
Pclock 0.7.1
Pmirror 0.6.1
Pmusic 2.6.6

From mcewanw:

Pfetch 0.9.2
Precord 7.0.2

I have uploaded these as-is to the 'noarch' repo, except for Precord I added some deps to the pet.specs file:

+ffmpeg,+gtkdialog3,+vorbis-tools,+lame,+alsa-utils

...although not *really* needed, as those are likely to be in all pup builds. We might sometimes do a build without ffmpeg though.

Pfetch Forum thread:
http://murga-linux.com/puppy/viewtopic.php?t=80887

Precord Forum thread:
http://murga-linux.com/puppy/viewtopic.php?t=49907

...notice, Precord already later version available. It is currently under rapid development.

Flash Player 10.3.183.23

Precise Puppy is built with the 10.3.x series of the Adobe Flash Player. Beta6 has version 10.3.183.18.

There have been bugfix releases, now up to 10.3.183.23, as advised by Forum member xman:
http://murga-linux.com/puppy/viewtopic.php?t=77256

I have updated the PET in the 'common' repo in ibiblio (5.1MB):
http://distro.ibiblio.org/quirky/pet_packages-common/flashplayer10-10.3.183.23.pet

Desk icons LagaBlueNight, Neon

Vicmz has put together the "Neon" desktop icon theme, see forum:
http://murga-linux.com/puppy/viewtopic.php?t=31059&start=330

I have added this to the 'noarch' repo on ibiblio:
http://distro.ibiblio.org/quirky/pet_packages-noarch/desk_icon_theme_neon-1.pet

In the same forum link, Forum member starhawk has made a nice little modification to the LagaBlueNight desk icon theme. The hard-drive and usb-drive icons were previously the same, starhawk has now created a special icon for usb-drives.

I have updated the PET in the 'noarch' repo to version 1.1:
http://distro.ibiblio.org/quirky/pet_packages-noarch/desk_icon_theme_lagabluenight-1.1.pet

Woof: rerwin: fixes '3f'

Rerwin posted some fixes here, related to Broadcom drivers (see down page):
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=570

I have implemented these changes in Woof.

However, these "firmware directories" have been taken out of Woof, now separate PETs:
b43, b43legacy, brcm, dgcmodem, hsfmodem, wl

Here are the PETs:
http://distro.ibiblio.org/quirky/pet_packages-noarch/firmware_linux_module_b43-120920.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/firmware_linux_module_b43legacy-20120920.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/firmware_linux_module_brcm-20120920.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/firmware_linux_module_dgcmodem-1.13-20120920.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/firmware_linux_module_hsfmodem-7.80.02.06-20120920.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/firmware_linux_module_wl-20120920.pet

Woof commits:
http://bkhome.org/fossil/woof2.cgi/info/44d63b25ab
http://bkhome.org/fossil/woof2.cgi/info/dd9818dd95
http://bkhome.org/fossil/woof2.cgi/info/40f0491f21

Recovered, etc...

The "flu" has gone, so I'm getting up to speed with Puppy again.

Today have to visit an elderly relative in hospital, the other side of the city. Just about to head off now, back onto Puppy this evening.

dgcmodem, hsfmodem firmware

In Puppy, we have a directory /lib/modules/all-firmware, which contains what we used to call the "firmware tarballs" for modems.

Except that I stopped making them into tarballs, left them as directories, as the SFS file compressed everything anyway.

These "firmware tarballs" are more than just firmware. They may contain data files, scripts, binary executables, anything needed to get the modem working. Take the case of the 'dgcmodem' -- the 'dgcmodem' directory would get installed if Puppy detected the presence of such a modem. Otherwise, the files are kept out of the way in the 'all-firmware' directory, so as not to interfere with the system.

These firmware tarballs used to be kept in kernel PETs, but I had a problem with maintaining all the different versions. So I moved them into Woof, even though some of these "firmware" directories also contain binary executables.

Moving to the present, rerwin is taking a new direction, starting with 'dgcmodem' and 'hsfmodem'. These are deleted from Woof, although I have left empty /lib/modules/all-firmware/dgcmodem and hsfmodem directories in Woof. These two are now separate PETs.

It is good I think that these are out of Woof. But, any puppy build that has a kernel with dgcmodem and/or hsfmodem modules, will have to also build the pup with these corresponding "firmware" PETs.

Here is where rerwin has posted these PETS:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=480
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=570

I have uploaded them (28KB, 88KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/dgcmodem-1.13-20120910-1.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/hsfmodem-7.80.02.06-20120911-1.pet

guess_fstype: 120911: btrfs support

A little while ago (120705), I fixed ext4 detection in guess_fstype. Now, technosaurus has added btrfs detection, bugfixed adfs, and replaced printf() with write() for smaller size:
http://murga-linux.com/puppy/viewtopic.php?p=651646#651646

I have uploaded the x86 PET (7KB):
http://distro.ibiblio.org/quirky/pet_packages-common/guess_fs-20120911.pet

The source can be found here:
http://bkhome.org/sources/alphabetical/g/

This utility is compiled statically with dietlibc, and is also in Woof, in the initrd. This update will be in the next commit of Woof.

Note, I need to compile for ARM boards, will do that soon.

3.2.29 kernel for x86 Precise

I had hoped to stay with the 3.2.28 kernel as used in Precise Puppy beta6, up until final release and beyond.

However, the filesystem-check at startup for a full-hd installation, fails. The reason is that the ramdisk size is 16MB, which is no longer big enough -- a small Linux is created on this on-the-fly at startup, and the main hd f.s. is unmounted and a f.s. check performed.

I have now set the ramdisk size at 32MB.

3.2.29 PET (23.4MB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_kernel-3.2.29-pae-i686-up.pet

3.2.29 sources:
http://bkhome.org/sources/kernel-3.2.29/

SeaMonkey 2.12.1 for Precise

Here are the PETS (22.8MB, 20.5MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/seamonkey-2.12.1-i686-up.pet
http://distro.ibiblio.org/quirky/pet_packages-precise/seamonkey_DEV-2.12.1-i686-up.pet

Compiled in Precise Puppy beta6.

Woof: for Precise beta6

This is the version of Woof used to build Precise Puppy 5.3.93 (beta6).

Commit:
http://bkhome.org/fossil/woof2.cgi/info/baff2d1b29

Precise Puppy beta6

My Puppy work has slowed right down, due to myself having "the flu". There is a lot more that I could have done before releasing beta6, but I think that there is enough new stuff to release it now.

Download:
http://distro.ibiblio.org/quirky/test/precise-5.3.93/

Regarding the PPM, rodin.s reported that a search for deps for 'kajongg' application found 165 deps and took a few minutes. The latest PPM is faster, on my laptop it took 33 seconds and found 173 missing deps.

VLC is a good one to test. PPM deps search found 68 missing deps and the search took about 18 seconds. The automatic fix using bbe to allow running as root works, I just tried it.

Enjoy!

Back home

I flew back from Melbourne on Thursday afternoon. At that stage, I had a tickle in my throat, urge to cough but controllable. Now it is a raging flue.

This is a "present" from Melbourne! Almost every day I was riding around in trains and trams. Difficult situation when people cough in an enclosed and crowded environment.

Anyway, I feel like sh*

I am doing the usual things, tea with lemon juice, lozenges, today I bought Irish Moss cough syrup. I read somewhere that a soak in a hot tub is beneficial, so I'll try that tonight.

Staying home as much as possible, and managing to soldier-on with Puppy development.

My Optus wireless broadband is choosing now to behave very poorly. Lots of connection timeouts, to the extent that I sometimes just give up and disconnect from the Internet.

VLC: hack to run as root

This was discussed in the previous post:
http://bkhome.org/blog/?viewDetailed=02985

I have put hacks into Woof to fix this. The FIXUPHACK in vlc template applies the fix if a pup is built with vlc. The script /usr/local/petget/hacks-postinstall.sh applies the fix if vlc is installed in a running pup.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/617697370d

PPM: dependency search deeper, faster

I have been puzzling over this for the last three days, finally nailed it.

When you click on a package in the PPM, then on "Examine dependencies" button, the script 'dependencies.sh' searches for all missing dependent packages. I have been using 'vlc' (audio/video player) for most of my testing, as that has a lot of missing deps -- 72 in fact.

Previously, not all missing deps were being found, also the search was slow. Also, the "hierarchy view" was not what I intended it should be.

I have fixed all three of those. Now, dependencies.sh goes down up to 11 levels searching for missing deps, and does it fast. The "hierarchy view" now looks like what it should be.

There are two scripts that I have modified, dependencies.sh and installpreview.sh. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/3343d0f880

Woof: for Precise beta5

This is the version of Woof used to build Precise Puppy 5.3.91 (beta5):

http://bkhome.org/fossil/woof2.cgi/info/559a668e3e

Precise Puppy beta5

Beta4 (5.2.73) was released recently, see blog post:
http://bkhome.org/blog/?viewDetailed=02976

Read my blog for reports on changes since then. Beta5 is now out, and I have assigned it as version 5.3.91, as I intend the final release will be 5.4. Summary of the recent versions:
http://distro.ibiblio.org/quirky/test/precise-5.3.91/precise-5.3.91-readme.htm

Download:
http://distro.ibiblio.org/quirky/test/precise-5.3.91/

I have been working on the Puppy Package Manager (PPM) the last few days, fixing some things. I fixed a chronic speed problem. There is another issue, the final "success" window when install a lot of packages at once, can be too high for the screen -- needs a vertical scrollbar -- I might do that tonight.

This pup is looking good!

Note, built with 3.2.28 PAE kernel. I plan to build a non-PAE version later.

sys-info 1.3

TaZoc has fixed a little bug in sys-info:
http://murga-linux.com/puppy/viewtopic.php?t=60462

I have uploaded version 1.3 to ibiblio (46KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/sys-info-1.3.pet

PPM: deps versioning re-fixed

The previous announcement about fixing versioned dependencies:
http://bkhome.org/blog/?viewDetailed=02978

However, I found that was broken. One significant problem is that it did not find any dependencies below "level 1", that is, nested dependencies.

So, I reverted that fix, then re-implemented it. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/b2e2a54947

Note, the code that npierce created was pretty good. The logic of this part of the PPM is mind-bending, even for me, so getting it just-right is far from easy. Npierce's work was a good starting point for me to look at where the fixes are required.

Gimp, nVidia SFS's

Shinobar has prepared two SFS files for Precise Puppy.

Gimp SFS is based on one created by Forum member scabz, the nVidia SFS created with getnvidia-0.7:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=420

I have uploaded the SFS files to ibiblio (20.8MB, 28MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/gimp-2.8.2-1-precise.sfs
http://distro.ibiblio.org/quirky/pet_packages-precise/nvidia-glx-304.43-k3.2.28.sfs

PPM: repo radiobuttons fixed

This is an ancient bug! Users have reported that after selecting repositories in the Configuration window, then (after restarting ppm and) back in the main window, the repo radiobuttons were incorrect.

I had some very messy code in script /usr/local/petget/pkg_chooser.sh, now cleaned up, and that bug fixed.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/a9cddfa2f6

PPM: deps versioning fixed

A Puppy-format package database entry has a dependencies field that may have optional versioning, examples:

|+ncurses,+readline&ge2.3.5,+glibc|
|+ncurses,+readline,+linux_kernel&ge2.6.39<3.0|

Where the operators may be:
&ge, >, &eq, &le, <

...this was documented awhile back in my blog.

At the time, I didn't quite implement it fully in the Puppy Package Manager, or rather I did but didn't test it. Forum member npierce has been investigating this, and has posted a solution to the Forum:
http://www.murga-linux.com/puppy/viewtopic.php?p=648934#648934

This fix looks good, I have put it into Woof:
http://bkhome.org/fossil/woof2.cgi/info/82b60b8633

Precise buglets fixed

01micko reported that at first bootup, when QuickSetup runs, making a change that requires a restart of X, then at restart there is no "welcome1stboot" nor the two barks.

Yes, this is a logical bug in /usr/sbin/delayedrun, and is a Woof-level bug applying to all pups. Fixed, will be in next Woof upload.

Clicking on the clock, bottom-right of tray, is supposed to bring up a calendar, however Precise Puppy is missing the 'minixcal' PET. Fixed. This is not a Woof bug, as such, or rather it is, as the package-list used to build Precise is in Woof.
But it is really classified as a Precise bug.

Precise Puppy 5.2.73 (beta4)

Here is the third variant. The brothers were released today and yesterday:
5.2.72: http://bkhome.org/blog/?viewDetailed=02975
5.2.71: http://bkhome.org/blog/?viewDetailed=02974

5.2.73 is the same, except that I used Ubuntu 'mesa' DEBs (not all of them) instead of my 'mesa' PET. Some notes:
http://distro.ibiblio.org/quirky/test/precise-5.2.73/precise-5.2.73-readme.htm

Download:
http://distro.ibiblio.org/quirky/test/precise-5.2.73/

This one is the most "Ubuntu standard", so preferred from the compatibility viewpoint. Ubuntu splits the original 'mesa' package into about 26 DEBs (although I am not sure if all of them are taken out of the original mesa pkg). These are some that I did not include in the build:

libegl1-mesa-drivers
libegl1-mesa
libegl1-mesa-dev
libgbm1
libgbm1-dev
libgl1-mesa-swx11
libgl1-mesa-swx11-dev
libgles1-mesa
libgles1-mesa-dev
libgles2-mesa
libgles2-mesa-dev
libopenvg1-mesa
libopenvg1-mesa-dev
libosmesa6
libosmesa6-dev
libxatracker1
libxatracker-dev

This is what I did include:

libglapi-mesa
mesa-common-dev
libgl1-mesa-dri
libgl1-mesa-glx
libgl1-mesa-dev
libglu1-mesa
libglu1-mesa-dev
libglapi-mesa
mesa-utils

I don't know if I included enough mesa DEBs, but it is running ok with my Intel video -- though I note neither the i915_dri.so nor swrast_dri.so modules have loaded.
'libGL.so*' is being provided by 'libgl1-mesa-glx' DEB, but note that alternative libGL files are in 'libgl1-mesa-swx11' DEB. I don't know what the difference is.

Precise Puppy 5.2.72 (beta4) PAE-enabled

The non-PAE brother of this build was announced yesterday:
http://bkhome.org/blog/?viewDetailed=02974

Some notes for 5.2.72:
http://distro.ibiblio.org/quirky/test/precise-5.2.72-pae/precise-5.2.72-readme.htm

Download:
http://distro.ibiblio.org/quirky/test/precise-5.2.72-pae/

...lots of delta files this time, if you have already downloaded an earlier Precise beta.

Precise Puppy 5.2.71 no-PAE (beta4)

I have created "beta4", with these features:
http://distro.ibiblio.org/quirky/test/precise-5.2.71-nopae/precise-5.2.71-readme.htm

Download:
http://distro.ibiblio.org/quirky/test/precise-5.2.71-nopae/

There are no delta-files, due to this build using gzip. As earlier builds use xz compression, the delta files are actually bigger than the original files!

I will now do another build, with the PAE-enabled kernel and Radeon-KMS enabled. Also the SFS xz'ed for smaller size.

Basically, the "no-pae" build will suit older hardware, also (probably) older Radeon video cards. But, it will be interesting to try both, if you have ATI/Radeon video (and CPU can handle PAE, which is most of them, even older i686).

Note, do not re-use the same save-file when testing the next pae-enabled build, they are incompatible.

Another note: beta3 reports, some issues with printing. I did upgrade cups from 1.3.11 to 1.4.8, relative to beta2, so maybe some problems got introduced? I don't have access to a printer right now, can't test.

Woof commit of version used for this build:
http://bkhome.org/fossil/woof2.cgi/info/5103776349c

...however, /etc/DISTRO_SPECS needs to have the appropriate kernel PET selected:

DISTRO_KERNEL_PET='linux_kernel-3.2.28-nopae-i686-aufs-up.pet'

Radeon firmware

01micko originally created this PET. I have added more firmware to it.

PET (66KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/radeon_firmware-120829.pet

Timezone fix

Forum members shinobar and ramachandran reported that timezone files are missing out of /usr/share/zoneinfo. This problem is in Precise Puppy beta3, not in beta2.

Beta3 is built from upgraded 12.04.1 DEBs, and the 'tzdata' package has been updated, which is the culprit.

The layout of files in /usr/share/zoneinfo has changed dramatically. If this change has been done by the 'tzdata' developers, not Debian or Ubuntu developers, then this problem is going to affect all Puppy builds as they move to latest tzdata.

I have fixed the 'glibc' template in Woof, so now all files are in /usr/share/zoneinfo.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/1487d9a91e

/dev/sda16+ not recognized

Forum member eps reported this. It applies to all puppies.

The nodes in /dev in the initrd are static, and only go up to sda15. Thus, if a hard drive has more partitions, they are not recognized.

After bootup, when the main filesystem is running, the kernel will auto-create the extra nodes if new hardware is plugged in, but not in the case of an internal drive present at bootup.

However, eps found that running Gparted does cause the extra device nodes to be created.

In Woof, I have created extra nodes for the initrd, up to sda23. I also did the same thing for the main filesystem.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/1d56dff622

Precise puppy beta4 coming

Probably tomorrow.

I think that maybe the Radeon problems are sorted out, judging from feedback so far. I intend to build two beta4's:

1. non-PAE kernel, Radeon without KMS.
2. PAE kernel, Radeon with KMS.

Number 1 will suit older hardware.

From feedback, it seems that the Ubuntu Radeon package doesn't need llvm. It seems that my 'z_mesa' PET has removed that requirement.
So, #2 will get built with the Ubuntu ATI/Radeon DEBs, #1 with my 'z_xf86-video-ati' PET.

There does seem to be a problem with timezones, reported by shinobar. This seems to be due to the upgraded DEBs, from 12.04 to 12.04.1. I will await shinobar's feedback (on the Precise Puppy Forum thread) and fix this, before building beta4.

rerwin: 3+ fixes

There were a couple of oversights in my previous application of rerwin's '3' and '3a' fixes for Woof, as rerwin has posted about here:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=330

I have applied those fixes. I also found a "-E" is needed for grep on one line, otherwise the search will always return nothing, in script functions4puppy4.

Explanation: a default grep expression treats "|" as just another string character, whereas extended-regular-expression treats it as a OR operation (latter is what we want).

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/4ca1741e19

PET uploads

Pfind 5.04 (zigbert)
radeon_firmware 120813 (01micko)
gimp 2.8.2 (scabz)

PETS (18KB, 14KB, 9.6MB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/pfind-5.04.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/radeon_firmware-120813.pet
http://distro.ibiblio.org/quirky/pet_packages-precise/gimp-2.8.2-precise.pet

The Gimp PET includes 'babl' and 'gegl' dependencies, so nothing else needed to install.

01micko has been investigating problems with Radeon video on Precise Puppy beta3, and found firmware needed, see Forum thread:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=315

Note, for further testing of Radeon video, I plan to upload a "beta4" tomorrow, with a non-PAE kernel, and KMS disabled.

PPM: mark already-installed pkgs

01micko discovered a bug in the Puppy Package Manager, running Precise Puppy beta3. A search for "xserver-xorg-video-radeon" has a hit, however that package is already installed. There was no indication of that fact, furthermore, clicking the 'Examine dependencies' button did not work.

Now, the icon of an already-installed package displays the /usr/local/lib/X11/mini-icons/mini-tick.xpm icon, also the string "(ALREADY INSTALLED)" is prefixed to the description field.

Also, the 'Examine dependencies' button is removed for an already-installed package. The package can still be installed however, in case there is a need to re-install it.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/e93ae86ff9

Woof commit 2012-08-25

This is Woof used to build Precise Puppy 5.2.69 (beta3).

Last three commits:
http://bkhome.org/fossil/woof2.cgi/info/7e6b7ee930
http://bkhome.org/fossil/woof2.cgi/info/158e3df752
http://bkhome.org/fossil/woof2.cgi/info/cd7d41afb1

Precise Puppy 5.2.69 (beta3)

Precise Puppy beta2 was released on June 6, 2012, see blog announcement:
http://bkhome.org/blog/?viewDetailed=02860

I was involved in various other things for awhile, now back on developing the x86 Precise Puppy. Version 5.2.69 (beta3) is now available.

Download from here:
http://distro.ibiblio.org/quirky/test/precise-5.2.69/

Beta3 has the 3.2.28 kernel, PAE-enabled. See blog post:
http://bkhome.org/blog/?viewDetailed=02961

Note, I am thinking of releasing Precise Puppy final with this kernel, plus another build with a non-PAE kernel. The latter will suit some older CPUs, and I will also set the default for Radeon video to non-KMS, and build the SFS with gzip compression (rather than xz), again to suit older hardware.

Beta3 has been built from Ubuntu Precise Pangolin 12.04.1 DEBs, plus many PETs that we have compiled ourselves. That is, I have used the latest DEBs from the recently-released Ubuntu Precise Pangolin upgrade (or Service Pack 1 if you want to call it that).

I am using beta3 right now, it looks good. We need a desktop background and a new theme for the RC...

Various updates

I am getting ready to build Precise Puppy beta3, so doing an update of various packages:

GetFlash 1.1 (shinobar)
Pburn 3.7.4 (zigbert)
pfilesearch 1.32 (zigbert)
NO Pfind 5.04 -- did not upgrade, zigbert's pet corrupted.
Pfind 5.01 (zigbert) -- note, removed trailing comma from pet.specs deps field.
Pmusic 2.6.5 (zigbert) -- note, slight modif. to pinstall.sh
Pprocess 2.3.2 9zigbert) -- note, removed trailing comma from pet.specs deps field.

I also added this DEB package to Precise Puppy:

gpart
http://murga-linux.com/puppy/viewtopic.php?t=80251

usb-modeswitch 1.2.4

Rerwin has upgraded usb-modeswitch to version 1.2.4 (my latest PET was 1.2.1). He has also enhanced the 'pinstall.sh' script to make some customizations that I had previously done manually. I just made one tiny change to pinstall.sh.

Forum post:
http://www.murga-linux.com/puppy/viewtopic.php?p=647379#647379

PETs uploaded (30KB, 15KB):
http://distro.ibiblio.org/quirky/pet_packages-common/usb-modeswitch-1.2.4-up.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/usb-modeswitch-data-20120815.pet

gtk-youtube-viewer

Forum member 666philb has created this:
http://www.murga-linux.com/puppy/viewtopic.php?t=76835

Great, he has created a PET for Precise Puppy, so I have uploaded it. So, it will be in the PPM for beta3.

PET (2.8MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/gtk_youtube_viewer-2.05_precise.pet

There's not much in that 'precise puppy' repo yet, but it will grow. Especially as it is likely Precise Puppy is going to be a Long Term Supported release.

There are some other PETs compiled for Precise Puppy, announced on the Forum, I will see if I can add some of these over the next couple of days.

Linux kernel 3.2.28

I compiled the 3.2.27 kernel only a couple of days ago:
http://bkhome.org/blog/?viewDetailed=02952

I have configured the 3.2.28 kernel the same, except enabled some more digital radio/tv drivers, plus enabled KMS by-default for Radeon and Intel i915 video. I also enabled the Intel i810 driver -- never noticed that one before.

PETs, for Woof developers (834KB, 23.3MB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_headers-3.2.28.pet
http://distro.ibiblio.org/quirky/pet_packages-common/linux_kernel-3.2.28-pae-i686-aufs-up.pet

Sources:
http://bkhome.org/sources/kernel-3.2.28/

Woof: rerwin: fixes '3'

I have applied rerwin's fixes from 'precise_rerwin_fixes-3.pet', see Forum post:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=255

Note, I earlier applied '3a' fixes, see previous blog post:
http://www.bkhome.org/blog/?viewDetailed=02959

My notes:


/etc/modules/firmware.dep

...fixes in this file in Woof.

Note that a couple of entries rerwin recommended to remove but left in, as they are still relevant for older kernels.

That is, these left in:

ar9170usb:ar9170usb.ko

rt2860sta-fw:rt2860sta.ko

rt2870sta-fw:rt2870sta.ko

Also, rerwin recommended adding "bcma.ko":

brcm:brcmfmac.ko,brcmsmac.ko,bcma.ko

...however, the 3.2.x kernel does not have that module, and I don't know what does, so I did not make that change.

...rethinking, I have appended "bcma.ko". don't know anything about it, but doesn't do any harm being there!



/etc/rc.d/functions4puppy

#120823 rerwin: multisession: screen out /.XLOADED when save.

...applied.



/etc/rc.d/MODULESCONFIG

...added the "bcma:wl" preference.



/usr/sbin/puppyinstaller

#120823 rerwin: preserve user-modified files in full-install upgrade.

...applied.

Note that Woof-built puppies now also have /root/.profile



/usr/sbin/remasterpup2

#120605 rerwin: omit /dev/snd content, /dev/.udev subdirectory and modem daemons from new master (because daemons reinstalled from firmware if needed).

#120606 rerwin: support users' replacement of stripped /root & /etc with complete directories (for "boot disk"). 

#120607 rerwin: Remove indicators/files for integrated user-installed packages.

#120628 shinobar: no change id-string for reuse

#120721 revert 'wildcards' option (from 120605); improve /dev/snd file exclusion; exclude /usr/share/icons/hicolor/icon-theme.cache (shinobar).

...all applied.



/usr/sbin/shutdownconfig

#120823 rerwin: avoid saving personal data if flag set in /etc (by remasterpup2, file /etc/personal_data_save_disabled).

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/58fc508fde

Woof: rerwin: fixes '3a'

I have applied some fixes from 'rerwin_woof_fixes-delta-3a.pet', see Forum:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=255

Not fully though, as I didn't understand some things. My notes:

/etc/rc.d/rc.shutdown:

#120603 rerwin: Retain deletions in first pupsave file by copying whiteout files.

...have not applied this. it seems only applies to the top-level directories, why would they be deleted?



/sbin/pup_event_backend_modprobe

#120823 rerwin: $FIRMPKG always written. 

#120823 rerwin: --use-blacklist to apply the blacklist commands in the configuration files (if any) to module names as well.

...applied.

#120608 rerwin: for rule-loaded modules either blacklisted or supported by modaliases, load no module

...have not applied this, don't understand why MODULES has to be cleared.



/usr/sbin/connectwizard

#120823 rerwin: add frisbee.

#120823 rerwin: get correct pid in wait loop.

...applied.



/usr/sbin/delayedrun

#120823 rerwin: skip quicksetup if /etc/personal_settings_popup_disabled file created by remasterpup2 (modified by rerwin).

#120823 rerwin: use distro-specific delayedrun if available.

...applied.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/ad1e860d8c

Precise Puppy schedule

Vicmz expressed confusion over what is the "official" Precise Puppy.

Ok, I just looked through the Puppy Forum. I see the point of confusion, oldyeller's "Precise final" thread here:
http://murga-linux.com/puppy/viewtopic.php?t=80175
And "Precise 5.3.0" thread here:
http://murga-linux.com/puppy/viewtopic.php?t=80431

That's fine, but it doesn't mean they are "official", just puplets that someone is building.

There is a lot of interest in Precise Puppy. My "absence" while working on the ARM boards does not mean that Precise Puppy got neglected -- not really, as I made many improvements to Woof that apply to both ARM and x86 builds.

Anyway, as I am back working on the x86 Precise Puppy, I guess that I should make a definite statement about where I am going with this, and make a commitment.

The thread for my beta releases is here:
http://murga-linux.com/puppy/viewtopic.php?p=632032#632032

I am reading through this thread, currently at about page 17. When I release beta3, feedback will continue on the same thread.

So yeah, it is probably best to keep it going to RC and final release, so that will become the official Precise Puppy. Others can then bring out puplet variants as they desire.

At my current pace, not sure when beta3 will be out, maybe Sunday or Monday. Any suggestions, please post to my "beta" thread.

PPM: dependencies fix

This bug has been there forever, good to finally fix it!

Precise Puppy is built with a 'cups' PET rather than the Ubuntu DEBs. As per normal, Ubuntu split a package into many DEBs, and cups becomes several, including 'libcups2'.

I noticed 'libcups2' showing up as a missing dependency when I was going to install 'gnome-control-center'. That is wrong, as the 'cups' PET is already installed.

If the 'libcups2' DEB was installed, it might cause problems as might not be compatible with the rest of the 'cups' PET.

I have fixed this problem. These "false dependencies" no longer show up. I modified scripts pkg_chooser.sh and findmissingpkgs.sh in /usr/local/petget.

This fix will apply to Puppy builds from Debian, Ubuntu and Raspbian DEBs.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/8ec06c1cb7

Melbourne again!

On the weekend I will once again be flying to Melbourne. Yay!

I created a little "report" on an earlier trip in 2010:
http://bkhome.org/bkauler/melbourne-dec-2010.htm

This time I will be there for almost two weeks. Taking the laptop of course, but have also packed the ODROID-X -- just a tiny board, power adaptor, keyboard (very small), easily fits in my carry-bag. I will plug the ODROID-X into the TV at my daughter's place.

My airline ticket is one of those cheapies, carry-on luggage only.

I might upload Precise Puppy beta3 after I have arrived there, take advantage of their super-fast Internet connection.

Sylpheed 3.2

Sylpheed in Ubuntu Precise Pangolin is compiled for GTK3, which unfortunately is the case for many applications. As we only have GTK2 in Precise Puppy, we have to compile it ourselves.

However, we found this quite difficult to do. This is because Debian and Ubuntu remove many of the .la files from /lib and /usr/lib. These files are in the original installation when compiled from source, however they don't put them into their DEBs.

I think that the reason for this is their "multi-architecture", where they put libraries into special directories.
Personally, I am inclined to the view that this multi-architecture stuff is bad news. Create a distro for one architecture, and be done with it. Trying to support different architectures, say i486 and i686, in the same distro causes so many complications.

Such as the .la files not working, hence they have to be left out. Some source packages, such as Sylpheed, need some .la files, and as they don't exist, compile fails.

I have created a PET '0hack1_precise_dev' in which I created the .la files needed for Sylpheed to compile, and this will be in the 'devx' of the next Precise Puppy (beta3).

Here are the Sylpheed PETs (671KB, 1.3MB, 167KB, 1.3MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/sylpheed-3.2.0-i686-up.pet
http://distro.ibiblio.org/quirky/pet_packages-precise/sylpheed_DEV-3.2.0-i686-up.pet
http://distro.ibiblio.org/quirky/pet_packages-precise/sylpheed_DOC-3.2.0-i686-up.pet
http://distro.ibiblio.org/quirky/pet_packages-precise/sylpheed_NLS-3.2.0-i686-up.pet

GtkHash 0.6.0

Forum member Billtoo reported that GtkHash in Precise Puppy beta2 does not work. This is from the official Ubuntu repository.

Billtoo has recompiled it, and it now works. Billtoo's post:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=210

I have uploaded the PETs (40KB, 8KB):
http://distro.ibiblio.org/quirky/pet_packages-precise/gtkhash-0.6.0-i686-up.pet
http://distro.ibiblio.org/quirky/pet_packages-precise/gtkhash_NLS-0.6.0-i686-up.pet

GtkHash home page:
http://gtkhash.sourceforge.net/

puppyClock timing tools

Forum member don570 has created puppyClock:
http://murga-linux.com/puppy/viewtopic.php?t=73168

I have uploaded the PET to the 'noarch' repo on ibiblio.org (18KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/puppy_clock-0.5.2-1.pet

Linux kernel 3.2.27

I have compiled the Linux kernel 3.2.27 in Precise Puppy, to be used in the next release (beta3).

PETs (for Woof usage only) (23.2MB, 834KB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_kernel-3.2.27-pae-i686-aufs-up.pet
http://distro.ibiblio.org/quirky/pet_packages-common/linux_headers-3.2.27.pet

The sources, including a kernel SFS file, are here:
http://bkhome.org/sources/kernel-3.2.27/
Us#er#na#me#: pu#pp#y Pa#ss#wo#rd#: li#nu#x

"Windows upgrade"

Terry sent me a pm with this very amusing paragraph:

I have two friends that I booted their Windows computer with Puppy to fix a problem who called and told me how much they LOVE THE WINDOWS UPGRADE!!! I went back to their house, and found that I had not pulled out the Puppy disk and re-booted after we started drinking beer, chatting, and I went on home a couple hours later!!!

Macpup 529

I have just seen this on distrowatch.com, Macpup 529:
http://distrowatch.com/?newsid=07396

I knew about Macpup, but I didn't know the guys (runtt21 and prit) have based the latest on the Woof-built Precise Puppy 5.2.60 (beta2), which I announced awhile back:
http://bkhome.org/blog/?viewDetailed=02860

Hmmm, that was June 6. After that I was absorbed by the ARM boards. Well, now I'm "back" so there is going to be a beta3 soon.

It is good to see the Macpup initiative. if they stay synced with latest Woof, they will be able to incorporate the significant improvements that I have recently done for package categories and icons -- though I do need to tidy it up over the next few days, then will (or rather plan to) build beta3 and upload latest Woof.

Improved categories and xdg menu

I have been working on improving the categories management and icons for packages, see earlier blog posts:
http://bkhome.org/blog/?viewDetailed=02948
http://bkhome.org/blog/?viewDetailed=02947

Now I have mostly completed the upgrade, with further refinements to the categories and icons, plus improvements to the Puppy Package Manager. Woof commits (most recent listed first):

some fixes for pkg categories, small ppm fix:
http://bkhome.org/fossil/woof2.cgi/info/c79220de2a

updated db lists to use xdg_puppy pet with binary executables only (other components are now in Woof):
http://bkhome.org/fossil/woof2.cgi/info/2d457e4d22

improved Categories= and Icon= assignments in .desktop files, refer scripts 2createpackages and /usr/local/petget/installpkg.sh:
http://bkhome.org/fossil/woof2.cgi/info/f2dec708da

improved xdg menu categories, /etc/xdg, /usr/share/desktop-directories created in Woof:
http://bkhome.org/fossil/woof2.cgi/info/83fce4bb09

improved the naming of sub-categories in pkg db entries, improved icon naming:
http://bkhome.org/fossil/woof2.cgi/info/2a178f9b89

I need to refine the support/find_cat utility, which assigns the category[;subcategory] for the packages. I am currently focused on improving this for Ubuntu packages -- find_cat is run when the Puppy db files are created for the Ubuntu DEBs.

So, it looks like I will be bringing out another release of x86 Precise Puppy. Probably before the end of the week.

Debian, Ubuntu package categories

I have been making improvements to the 'category' field in package database entries, see recent blog posts.

I am continuing this, according to a strategy that I have had in mind since I first designed the package db format for the Puppy 2.x series.

Today there is just a little improvement, but significant. For puppies built from Debian, Ubuntu (and Raspbian) DEBs (binary packages), I have improved the assignment of an appropriate 'category[;subcategory]' field.

The problem has been that Debian/Ubuntu sometimes uses odd names for packages, which confuses the support/find_cat utility in Woof, that is mostly looking for the original package names.
For example, if I recall rightly, Debian renames 'pidgin' as 'finch'. Some other packages have odd numbers and letters appended (apart from the version number).
Also, packages are often broken up into several (or more) separate packages, each with a different name.

I have found a simple way around this problem.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/c5f76a9101

PPM: application icons, take-2

A little while ago I implemented a mechanism for having application icons in the Puppy Package Manager:
http://bkhome.org/blog/?viewDetailed=02851

I have now re-implemented the mechanism, a complete redesign. I have mentioned this in the Forum:
http://murga-linux.com/puppy/viewtopic.php?p=646355#646355

The package database format now extends the 'category' field to have an optional sub-category, like this:

pkgname|nameonly|version|pkgrelease|category[;subcategory]|size|path|fullfilename|dependencies|description|compileddistro|compiledrelease|repo|

Example:
homebank-4.3-w5|homebank|4.3-w5||Business;finance|604K||homebank-4.3-w5.pet|+gtk+|HomeBank finance management|puppy|wary5||

There are generic icons in /usr/local/lib/X11/mini-icons for all 'category' and 'subcategory' names, and they now display in the PPM.

In Woof, the support/find_cat tool, called from '0setup', now generates category;subcategory, the latter being optional, if it can be determined.
Other scripts have been updated to handle the change in category field.

Woof commits, latest first:
http://bkhome.org/fossil/woof2.cgi/info/5d6118ef94
http://bkhome.org/fossil/woof2.cgi/info/fba51bbf16
http://bkhome.org/fossil/woof2.cgi/info/e2620791bb
http://bkhome.org/fossil/woof2.cgi/info/0eca8ff7a2

Slow going

I have run into a problem with the latest builds of Wary and Racy. ROX-Filer refuses to display thumbnails. It outputs an error message that I cannot understand.

I have been trying to figure out what is wrong, no success.

So, I am now going through a very painfully slow process of elimination. I have checked out the version of Woof used to build Wary/Racy 5.3 -- commit 9d81d0919b, April 3, 2012 -- build Racy, test ROX, then update Woof by say one month, build Racy and test ROX, etc.

gtk 2.24.11 with paste patch

I have compiled GTK+ version 2.24.11 in Wary.

Forum member disciple reported on a patch for a keypress to replace the mouse middle-button, to paste the primary selection:
http://murga-linux.com/puppy/viewtopic.php?t=80150

The patches for GTK+ 2.24.x and 3.x are here:
https://bugzilla.gnome.org/show_bug.cgi?id=643391

Here are the PETs (2.3MB, 266KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/gtk+-2.24.11-patched_paste-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/gtk+_DEV-2.24.11-patched_paste-w5c.pet

Here is how I configured it:

# patch -p1 < ../gtk2-0001-Add-paste-selection-keybind-signal-2.24.1.patch

# ./configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var --build=i486-t2-linux-gnu --with-xinput=yes --enable-debug=minimum

ShareInternet and dnsmasq

Forum member gyro has created this very interesting application to turn a Puppy-PC into a router:
http://www.murga-linux.com/puppy/viewtopic.php?t=37520

Gyro's PET has 'dnsmasq', a binary executable, in it. I have taken that out, so there are now two PETs (106KB, 14KB):
http://distro.ibiblio.org/quirky/pet_packages-common/dnsmasq-2.62-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/shareinternet-2.2.4-1.pet

CUPS 1.4.8

I have compiled CUPS 1.4.8 in Wary, for use in Wary and Racy. This is to fix a problem with landscape printing, mentioned here:
http://bkhome.org/blog/?viewDetailed=02940

The PETs (2.2MB, 35KB, 603KB, 589KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/cups-1.4.8-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/cups_DEV-1.4.8-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/cups_DOC-1.4.8-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/cups_NLS-1.4.8-w5c.pet

Note that the PET package for 1.3.11, as used in Wary/Racy up until now, is 1.3MB, this new PET is 2.2MB.

Here is how I configured it:

# LDFLAGS='-L/usr/X11R7' DSOFLAGS='-L/usr/X11R7' CPPFLAGS='-I/usr/X11R7/include' CFLAGS='-I/usr/X11R7/include'  ./configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var --disable-dbus --disable-pam --enable-libpaper --disable-ldap --disable-dnssd --disable-launchd --disable-slp  --disable-gssapi --disable-use-network-default --disable-pap --enable-openssl --disable-libusb --with-pdftops=/usr/bin/pdftops

...this took me awhile to get right. The '--disable-libusb' is required so that CUPS will use the old "usblp kernel module" method -- otherwise, the 'usblp' module has to be blacklisted. Besides, I'm not sure if the CUPS developers have got the new "libusb" method fully sorted out.

'cnxsoft' very busy guy

I only know him (or her) as 'cnxsoft'. I often look at his site:

http://www.cnx-software.com/

If your addiction to ARM boards requires you to get a regular hit on the latest developments, this is the place to go

There is something new posted everyday. 'cnxsoft' is keeping up a very intense schedule, plus he finds out about all kinds of things, garnered from the remotest corners of the Internet.

In particular, there is a lot of info about the Allwinner A10 boards, including tutorials and download links.

Woof: 0setup fix

For doing a pet-based build, for example Wary and Racy, I found some bugs, one in script '0setup' and one in support/fix-puppy-dbs. Fixed.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/afe97bc120

Wary/Racy plans

My ODROID-X quad-core Cortex-A9 ARM board has arrived, and I ran Android 4.0.4 on it. Nice and fast!

However, I plan to put the ARM boards on the back-burner for awhile and do some work on Wary and Racy, look toward releasing version 5.4.

The last release was 5.3, on April 5, 2012, announced here:
http://bkhome.org/blog/?viewDetailed=02784

I intend to start reading the Wary/Racy feedback thread on the Forum, from page 7:
http://murga-linux.com/puppy/viewtopic.php?t=77303&start=90

I think that I will upgrade CUPS from 1.3.11 to 1.4.8. There is a problem with printing in landscape mode, reported by vtpup:
http://murga-linux.com/puppy/viewtopic.php?t=77303&start=135

rcrsn51 posted something that may be helpful:
http://murga-linux.com/puppy/viewtopic.php?p=624541#624541

RP: JWM problem?

I upgraded JWM from 574 to 579, see earlier blog post. Since then, there has been some peculiar behaviour, and I am now reverting to 574.

I have made a bug report to Joe, the JWM developer:
http://sourceforge.net/tracker/?func=detail&aid=3553850&group_id=157836&atid=805416

I have recompiled 'pngoverlay' in sap6, as 01micko reported a bug. I think that previously I had compiled it in Gentoo on RP. This will be in the next Woof commit.

Fossil: 'getwoof' script

I need to automate the download of Woof from the online repository, so I wrote a script, /usr/sbin/getwoof. It also takes an optional parameter "update" to update an existing downloaded Woof.

Previously, there were manual steps involved, as explained here:
http://bkhome.org/fossil/woof2.cgi/wiki?name=Getting+Started+with+Woof

I had a great deal of difficulty writing this script, as I could not get some of the fossil CLI commands to work as the docs seemed to indicate how they should work. I ended up writing what I think is a clumsy script, but it does work.

Fossil commit:
http://bkhome.org/fossil/woof2.cgi/info/996378481d

MoManager: updated howto

With MoManager becoming better and better, I realised that the "HOWTO Internationalization" page (see "Help" menu) is becoming a bit outdated.

So, I have updated it. This will be in the "Help" of future Puppy builds, however you can also view it online:
http://bkhome.org/fossil/woof2.cgi/artifact/ee29820a5d24df4f84cc2c449152df4fefdcbb79

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/26c45f7e8b

MoManager: multiple scripts one TEXTDOMAIN

This is great news, I have unleashed MoManager. Now, you can have one TEXTDOMAIN for multiple script. Also, a pre-existing .pot file will be automatically used.

One TEXTDOMAIN
If multiple scripts have the same TEXTDOMAIN, for example:
export TEXTDOMAIN=myapp
then just one myapp.pot file and hence myapp.mo file will be created for all of those scripts.
Some people prefer this, as it is more efficient than having one .pot file for each script.

Note that these multiple scripts with same TEXTDOMAIN can be located anywhere. They do not have to be in the same directory.

Pre-existing .pot
A pre-existing .pot, for example /usr/share/doc/nls/myapp/myapp.pot, will automatically be used, instead of generating it from the script(s).
I could not see any reason why this should not be automatic.

Check out the changes in /usr/sbin/momanager, Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/6486b73f7e

A note about entering the TEXTDOMAIN into scripts:
export TEXTDOMAIN=myapp
...do not put quotes, do not append any comment, do not indent! Any of those might break MoManager.

Wary: ffmpeg improved

K Godt, npierce and shinobar have been investigating improvements to ffmpeg in Wary and Racy:
http://murga-linux.com/puppy/viewtopic.php?t=76182&start=270

I have uploaded the ffmpeg PETs compiled by shinobar, with x11grab and network support (2.3MB, 2.8MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/ffmpeg-0.8.12-1-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/ffmpeg_DEV-0.8.12-1-w5c.pet

This works with shinobar's latest ffconvert 1.3 (61KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/ffconvert-1.3.pet

This improvement will of course be in the next release of Wary/Racy.

MoManager: one .pot for many scripts

A problem with MoManager was reported to this blog:
http://bkhome.org/blog/?viewDetailed=02928

And solutions proposed on the Puppy Forum:
http://www.murga-linux.com/puppy/viewtopic.php?p=643538

The solution proposed by L18L looks good. I presume for the example of 'pprocess', that the scripts 'box_ok' and 'func' will not have the "export TEXTDOMAIN=..." declaration?

I have implemented L18L's solution, Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/4eebc559f0

Host Woof on ARM board

So far, the host system on which I have used Woof is an x86 computer. So when building a Puppy for an ARM board (such as the Raspberry Pi), I have specified host as "x86" and target as "arm" in Woof -- this is asked for in script 'merge2out'.

There are various problems with this "cross build", one of which is that certain operations (such as 'depmod') cannot be performed on the target environment, and they have to be deferred until first bootup of Puppy on the target board.
This is one of the factors that causes first-boot to be slow.

I have now added some armv6 executables needed to run Woof in an ARM environment, that is, host and target can both be "arm". Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/985145716c

I haven't tried it yet. I will probably have a go at that tomorrow. It needs to be done with a USB hard drive, and probably the Raspberry Pi can handle it.

RP: recompiling 3.1.9 kernel

Today I have downloaded the latest patched 3.1.9 kernel source for the Raspberry Pi, from here:
https://github.com/raspberrypi/linux

I had previously been through this exercise, and that resulted in the kernel used in Squeezed Arm Puppy alpha4. That was about one month ago.

Now I am doing it again, with a few changes to the .config:

bcm2835 sound driver built-in
deadline scheduler (instead of cfq)
disabled automatic devtmpfs mounting on /dev

Also, this time I did not apply any of my own patches, that is, left out unionfs and usb-printk patches. Only applied the loglevel patch.

I have just started the compile on my RP.

I have also downloaded the latest firmware from here:
https://github.com/raspberrypi/firmware

Fixed f.s. check at bootup

A full installation of Puppy will perform a filesystem check at bootup if there was a previous abnormal shutdown. I did some work on this for the Raspberry Pi Puppy alpha4, see earlier blog posts:
http://bkhome.org/blog/?viewDetailed=02907
http://bkhome.org/blog/?viewDetailed=02908

You might be interested in one of the original blogs post about this mechanism:
http://bkhome.org/blog/?viewDetailed=01434

However, the mechanism is still broken in RP Puppy alpha4. I have now fixed it, see Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/72799a5530

...see script /sbin/initNEW, that becomes /sbin/init in a running Puppy.

RP: faster booting

I have been working on streamlining the bootup on the ARM boards. I am aiming for 20 seconds on the Raspberry Pi -- that is a target, not there yet! It is an optimistic target, not sure if I can get there.

I haven't had much time the last few days, mostly due to doing things for elderly relatives.

Today I setup wireless broadband Internet on a laptop for one of my elderly relatives, a person who has never used a computer before. She was recently given an old laptop, which has Windows 2000 on it. I didn't put Puppy on it, left Windows 2000 as-is, just installed the latest Firefox and Adobe Flash players that work with Windows 2000.

I took her to the local shopping centre and bought an Optus Prepaid Mobile Broadband usb stick, for AU$69.95. Even though it specified at least Windows XP required, I ignored that -- it installed ok and runs fine on Windows 2000.

We were quickly online and surfing the web. It was interesting watching someone using a mouse for the very first time. She is in her early-80s and her hands are not so nimble.

I took her through all the steps, but I am wondering if she will be too scared to turn it on on her own. I'll pop back in a day or too and see what has happened.

ODROID-X production underway

Here are some photos showing the board being manufactured:
http://odroid.foros-phpbb.com/t1144-odroid-x-odroid-q-shipment-ready-almost

I placed my order on July 11, so hopefully I will be in the first batch.

'de', 'ru' langpacks

L18L is the creator and maintainer of the German (de) language pack, see Puppy Forum:
http://murga-linux.com/puppy/viewtopic.php?t=76368

rodin.s is the creator and maintainer of the Russian (ru) language pack.

These guys have recently updated their langpacks. L18L has it on the Forum, rodin.s pm'ed it to me.

I have uploaded them (2.4MB, 2.1MB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/langpack_de-20120718.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/langpack_ru-20120720.pet

arm: gtkdialog reverse-order fixed

I posted awhile back about the widgets rendering in reverse order in a window on ARM architectures:
http://bkhome.org/blog/?viewDetailed=02809

Puppy Forum member jamesbond fixed it:
http://murga-linux.com/puppy/viewtopic.php?t=79359

Now, the maintainer of Gtkdialog, thunor, has applied the fix to the official source, revision 321:
http://code.google.com/p/gtkdialog/source/detail?r=321

RP: gutenprint updated

01micko has compiled the latest Gutenprint package on Squeezed Arm Puppy (sap6) on the Raspberry Pi, see Forum post:
http://murga-linux.com/puppy/viewtopic.php?t=79359&start=165

I have uploaded the PETs to ibiblio and they will be in the next release of sap6 (2.6MB, 621KB, 1.5MB):
http://distro.ibiblio.org/quirky/arm/pet_packages-squeeze/gutenprint-5.2.8-sap6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-squeeze/gutenprint_DEV-5.2.8-sap6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-squeeze/gutenprint_NLS-5.2.8-sap6.pet

Simple touch-screen driver

taca0 posted about the WM8650 ARM-based netbooks:
http://murga-linux.com/puppy/viewtopic.php?t=79896

This link is interesting:
http://pond-weed.com/wmt8650/index.html

...it has the source for a simple touch-screen driver to replace a mouse.

I thought that I would post the link here for future reference.
I am gradually coming round to the thinking that a touch-screen with proper keyboard is a good combo. It solves the problem of using a mouse when you are on-the-go.
It would certainly be much better than a touchpad, and I don't see why we cannot have our traditional UI, no need for a dumbed-down UI as in tablets these days.

Except of course, we have three buttons on a mouse, and would ideally need to emulate that.
The driver in the above link seems to only be supporting the left-mouse-button.

Desktop icon themes

Jemimah and vicmz have created some nice icons for the desktop and menu:
http://murga-linux.com/puppy/viewtopic.php?t=31059&start=315

I have uploaded these to the 'noarch' repo on ibiblio:

Jemimah (106KB, 104KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/desk_icon_theme_fs-1.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/desk_icon_theme_neu-1.pet

vicmz (131KB, 125KB, 151KB, 135KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/desk_icon_theme_jungle-1.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/desk_icon_theme_malays_uniblack-1.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/desk_icon_theme_malays_uniblue-1.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/desk_icon_theme_shining_silver-1.pet

...note, these are smaller than vicmz's original PETs, as some redundant images removed.

RP: heatsink

I reported recently that the ODROID-X will ship with a heatsink:
http://bkhome.org/blog/?viewDetailed=02913

Heh heh, you can also put a heatsink on your Raspberry Pi:
http://www.chrisnewland.com/raspberry-pi-heatsink-mod-251

Sys-Info, xdriinfo

TaZoc has updated Sys-Info. I separated out the xdriinfo binary executable, as sys-info PET is in the 'noarch' repo.

sys-info (41KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/sys-info-0.9.pet

xdriinfo (4KB, 3KB):
http://distro.ibiblio.org/quirky/pet_packages-common/xdriinfo-1.0.4-x86.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/xdriinfo-1.0.4-sap6.pet

RP: recompiled tray applets

01micko reported that the slowness of the network_tray applet in sap6 was not due to JWM, but due to the applet itself.

I had compiled the tray applets in Gentoo on the RP. I have now recompiled them in sap6.

PETs (7KB, 9KB, 6KB):
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/freememapplet_tray-2.4-sap6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/network_tray-2.5-sap6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/powerapplet_tray-2.5-sap6.pet

RP: alsaconf broken

The 'alsaconf' script does not work on the ARM boards, due to absence of modules.pcimap and modules.isapnpmap. Perhaps the script could be modified to work without those?

Anyway, I have disabled 'alsaconf' for the ARM board. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/8473af633d

Well, well, I actually got sound out of my RP! Just one song, /usr/share/audio/ivy.mp3. None of the other audio files in /usr/share/audio work.

I played it like this:

# playmusic /usr/share/audio/ivy.mp3

'playmusic' is thunor's frontend to various players, including 'mplayer'. It wouldn't play the other files.

The 'aplay' utility doesn't work at all. I also tried 'wavplay' to play /usr/share/audio/2barks.wav, no go, it just hung.

...in fact, after playing one file, all the players hang and nothing else will play. Even that 'ivy.mp3' only played once.

This is testing on Squeezed Arm Puppy alpha4.

I am not going to mess around with this anymore. The kernel sound driver needs to be fixed.

RP: JWM 579

01micko reported that JWM in Squeezed Arm Puppy on the RP has a problem with the tray -- applets slow to update. He compiled the latest, version 579, and that fixed the problem.

The problem though, is probably not the version. I compiled version 574 in the Gentoo build for the RP. I have found that some apps compiled in Gentoo misbehave in sap6.

I have compiled JWM 579 in sap6, and uploaded the PET (79KB):
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/jwm2-579-sap6.pet

Disable text-mode xorgwizard on ARM boards

The text-mode Xorg Wizard is not appropriate for the ARM boards.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/a913fe8a97

Mixmos tray applet

01mick created a tray applet for the Mixmos audio mixer. This is a very simple applet for use in the Raspberry Pi. 01micko's Forum post:
http://murga-linux.com/puppy/viewtopic.php?t=79359&start=150

I have placed 'mixmos_tray' into the 'mixmos' PET (75KB):
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/mixmos-0.2.0-1-sap6.pet

RP: Gmeasures, Superscan

Two little apps missing from Squeezed Arm Puppy are Gmeasures and Superscan. The first is a units converter and the second a port scanner, both GUI apps. These have traditionally been included in our puppies, to round out the app-suite.

I have compiled them in Squeezed Arm Puppy on the Raspberry Pi, and they will be in the next release.

However, you can get them now (25KB, 22KB):
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/gmeasures-0.7-sap6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/superscan-0.8-patched2-sap6.pet

Woof: builds pup from Raspbian pkgs

It is pre-alpha as although there is a desktop, some things are broken. But yes, Woof can now build a Puppy from Raspbian Wheezy DEBs.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/49f23f3d31

There is a mystery. Some shared libraries, that were compiled in Squeezed Arm Puppy (built from Debian Squeeze DEBs) do not work.
For example, 'ldd' reports that /usr/lib/libnetpbm.so.10 is "not a dynamic executable".

Huh? Those libraries work fine in Squeeze. They are "armel", compiled for armv4 architecture, but that is not a problem.

Note, this Woof can now also be used to build from straight Debian Wheezy DEBs. It is the same situation as for Raspbian, just different repositories.

dpkg-deb with xz support

Building Puppy from Raspbian DEBs, I thought that some were missing from the repository, however I discovered that '1download' was downloading then rejecting some DEBs.

The reason is that some DEBs have xz compression, which the 'dpkg-deb' utility in Wary and Racy (and most other pups) does not understand.

Support for xz was introduced with version 1.15.6 of the 'dpkg' package, of which 'dpkg-deb' is a utility therein.

I have compiled a recent version and uploaded to the 'common' repo (44KB):
http://distro.ibiblio.org/quirky/pet_packages-common/dpkg-deb-1.16.4.3-w5c.pet

ODROID-X now with heatsink

Hardkernel are the developers of the ODROID-X quad-core A9 ARM board that I have ordered, see my recent blog post:
http://bkhome.org/blog/?viewDetailed=02901

Hardkernel have tested the board with all four cores running "flat out" and the chip does get hot. This is not the situation for normal use, however they decided to include a passive heatsink on the production board:

Their Forum post explaining this decision:
http://odroid.foros-phpbb.com/t1127-the-first-sample-of-pre-production

Apart from whether it is really needed or not, the heatsink gives it a certain look, like it really means business. A mean machine!

Woof: Raspbian support, new DB format

01micko posted about a change in the package database format in Debian Wheezy:
http://bkhome.org/blog/?viewDetailed=02900

Actually, I was looking through the DB files for Raspbian Wheezy, and it has a mix of old and new formats.
Therefore, I have modified script 0setup to automatically detect which format.

I have also added preliminary support to build Puppy from Raspbian packages.

I want to build a pup from Raspbian debs and see if it is sane, then I will upload Woof.

Woof to build sap6 alpha4

This is the version of Woof used to build Squeezed Arm Puppy alpha4 (for the Raspberry Pi).

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/904fb8d4a3

Puppy for Raspberry Pi alpha4

Announcement of previous alpha3:
http://bkhome.org/blog/?viewDetailed=02899

Announcement of alpha4:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.105-alpha4/sap6-5.105-alpha4-readme.htm

Download SD card image (87.1MB):
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.105-alpha4/raspi-sd-4gb-sap6-5.105.img.xz

"devx" (119.9MB):
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.105-alpha4/devx_sap6_5.105.sfs

Howto copy image to SD-card:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.105-alpha4/howto-copy-to-sd-card.htm

md5sums:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.105-alpha4/md5sums.txt

RP: X server refuses to die

Sometimes when I shutdown, the Raspberry Pi screen goes blank and it just hangs. I examined the log, in /tmp/xerrs.log, and it has this:

waiting for X server to shut down XIO:  fatal IO error 11 (Resource temporarily unavailable) on X server ":0.0"

      after 61 requests (59 known processed) with 0 events remaining.

..........



/usr/bin/xinit:  X server slow to shut down, sending KILL signal.



waiting for server to die ...



/usr/bin/xinit:  Can't kill server

Curious Google finds nothing for the string "/usr/bin/xinit: Can't kill server".

Anyone got any thoughts on how to force the X server to die? Anyway, why won't it die?

Full install: f.s. check maximal count

K Godt gave me the idea for this, see previous blog post:
http://bkhome.org/blog/?viewDetailed=02907

A f.s. check is performed if an abnormal shutdown, now also if maximal mount count reached.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/ac97a7c17b

Full install: f.s. check at bootup

For a full installation of Puppy to hard drive, including the SD-card installations for the ARM boards, also known as PUPMODE=2, if there is an abnormal shutdown then there is supposed to be an automatic f.s. check at next bootup. Except that it was broken.

There is no initrd, and the first thing that executes at bootup is /sbin/init in the main f.s. This is a script, that normally passes execution to Busybox init and bootup happens normally.

However, if there is an abnormal shutdown, then the file /fsckme.flg exists at bootup and the /sbin/init script goes into a f.s.-check mode. This consists of creating a ramdisk with a mini-Linux distro on it, "pivot_root" into it, then do a f.s. check of the main partition.

This mechanism was broken, and on the Raspberry Pi Puppy alpha3 it was not happening at all.

I have fixed it, Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/6dccd3e818

However, avoid deliberate abnormal shutdowns. Really, they should be avoided completely. If there is one, having a journaled f.s. will not save you from possible file corruption. A f.s. repair at bootup will not save you from potential file corruption either. I wrote about this on my blog recently.

Simple Wallpaper Setter

Pwallpaper, used in alpha3 in the Raspberry Pi, segfaults.

01micko created a very simple wallpaper setter using gtkdialog, named appropriately Simple Wallpaper Setter, or just SWS. It had a bug with changing the pinboard, so I hacked it a bit, now it is working for me.

Uploaded (2KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/simple_wall_set-0.6-noarch.pet

Open source graphics driver for ARM

The lack of accelerated graphics for X applications has been a very big issue. Linux on the Raspberry Pi and the Mele A1000 uses the kernel framebuffer driver for X, which is dumb and does not make any use of the GPU (Graphics Processing Unit). Except, in the Raspberry Pi there are some apps that use the GPU for video playing.

Anyway, something, apparently, is about to change:
http://www.phoronix.com/scan.php?page=news_item&px=MTE0MDA

'cnxsoft' at http://www.cnx-software.com/ thinks that it will be the Mali drivers from ARM, that are currently proprietary. Those drivers are partly GPLed and it makes good sense for them to make the drivers fully GPL.

SoftMaker Office UK English

I created a PET of the free version of Softmaker Office 2008, back in August 2011:
http://bkhome.org/blog/?viewDetailed=02435

The base PET package has US and UK English spell-checking, however it was reported that UK could not be selected.

Tuxtoo discovered that a file is missing from the base PET, installation of which then enabled UK English to be chosen. Tuxtoo posted a PET with the missing file:
http://www.murga-linux.com/puppy/viewtopic.php?p=640462#640462

...thanks very much for that!

I have added that missing file and uploaded the SoftMaker Office PET (12.1MB):
http://distro.ibiblio.org/quirky/pet_packages-common/softmakeroffice-2008-1-x86.pet

The docs and other languages are also available (13.2MB, 24.2MB):
http://distro.ibiblio.org/quirky/pet_packages-common/softmakeroffice_DOC-2008-1-x86.pet
http://distro.ibiblio.org/quirky/pet_packages-common/softmakeroffice_NLS-2008-1-x86.pet

RP: Audio output control in QuickSetup

QuickSetup is now customised for the Raspberry Pi, the start of it anyway.

There are now radiobuttons to choose whether the routing of audio output is automatic, to the analog socket, or to HDMI.

QuickSetup could in future have widgets to control entries in 'config.txt' in partition #1.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/41197d3a04

RP: Screensaver fixed

On the top-right of screen is the "lock" icon, which is for when you go away from your computer and don't want anyone else to use it, or see what was on the screen. This system is known as 'Xlock' or 'xlockmore'.

This is different from the inactivity time-out screensaver, which is just a black screen. That is setup by the 'pupX' application in the 'Desktop' menu.

Anyway, Xlock is broken. The 'xlock' application was compiled in RP on a Gentoo distro, which has much later packages than those in Debian squeeze, and there is some kind of mismatch that causes the utility to crash in the Debian Squeeze build of Puppy.

So, I have recompiled it, and now it works. The PET (132KB):
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/xlockmore-5.31-patched-sap6.pet

ODROID-X development platform

I have become addicted to these ARM boards I was drooling over this latest one, that will be available late July, price only US$129:
http://www.hardkernel.com/renewal_2011/products/prdt_info.php?g_code=G133999328931&tab_idx=1

Here is the CPU module:
http://odroid.foros-phpbb.com/t1062-exynos-4412-cpu-module

...oh joy, notice that it has a little battery! That means a battery-backed hardware clock!

This is just too nice, I had to order it, have decimated my PayPal balance! I ordered those extra goodies:

WiFi module
USB-UART module
eMMC module
5V/2A adaptor
Bluetooth module
HDMI cable

The board comes with Android ICS, however also Ubuntu 12.04 runs on it:
http://odroid.foros-phpbb.com/t1075-ubuntu-1204-on-odroid-x-exynos-4412

...but not yet any GPU acceleration for Ubuntu. Though, it has the Mali 400 GPU, so we are optimistic that there will be an accelerated X11 driver "sometime".

Explanation of "eMMC"?
http://www.samsung.com/global/business/semiconductor/product/flash-emmc/overview

Woof commit 19ad701efd

This is the Woof used to build Squeezed Arm Puppy alpha3 for the Raspberry Pi.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/19ad701efd

Note that sap6 alpha3 was built with an SD-card skeleton image dated 09-July-2012, I will upload that now.

Puppy for Raspberry Pi alpha3

Here it is, our alpha3 release, and it is looking good!

Release notes:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.102-alpha3/sap6-5.102-alpha3-readme.htm

...please read these notes before running your pup on the Raspberry Pi.

Download the SD-card image, only 86.7MB:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.102-alpha3/raspi-sd-4gb-sap6-5.102.img.xz

How to copy the SD-image file to an SD memory card:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.102-alpha3/howto-copy-to-sd-card.htm

If you want to turn your pup into a complete compiler environment, do it with just one file (read Release Notes for details), file size 119.9MB:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.102-alpha3/devx_sap6_5.102.sfs

If you want to read about the previous alpha2 release, see here:
http://bkhome.org/blog/?viewDetailed=02879

Additional usage notes
It is very important to be aware that the first bootup is very slow, so be patient, but subsequent bootups are very fast. I will be improving the first-bootup speed soon.

Additional issues
I have just noticed that the screen-lock does not work (icon at top-right of screen). I will look into that asap.
There is no volume control in the tray, but there is a audio mixer in the Multimedia menu.

Pnethood 0.7, RP: udev

Shinobar has updated Pnethood, and it is also internationalised, now at version 0.7:
http://murga-linux.com/puppy/viewtopic.php?t=70532

I have uploaded the PET (18KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/pnethood-0.7.i18n.pet

I was trying to figure out why my Optus broadband wireless USB modem wasn't getting properly detected when I plugged it into my Raspberry Pi. Then I made the discovery that the 'udevd' daemon was not running. I tested it and it aborted at startup with the message:

Root privileges required

Hmmm, udevd is confused. I compiled it statically sometime ago, before acquiring my RP, cross-compiled statically with uClibc I think. I don't recall what toolchain was used.

Anyway, I recompiled it in the Squeezed Arm Puppy on the RP, and all is well. Now I plug in my USB modem and I get the familiar "3G USB modem now ready for use" popup.

That's great, fixing udevd will have fixed a whole swag of problems. It will probably slow down the bootup though -- I am getting bootup time of 24 seconds, which seemed to be too good to be true -- and as it turns out, it is too good to be true.

The new udev PET (144KB):
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/udev_167-167-patched_t2-sap6.pet

ARM: wvdial, wvstreams

Squeezed Arm Puppy alpha2 has wvdial 1.41, which is an old version that might not work properly with broadband usb devices.

I compiled wvdial 1.61 and wvstreams 4.61 (library needed by wvdial). PETs uploaded (76KB, 5KB, 743KB, 1.5MB, 3KB):
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/wvdial-1.61-sap6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/wvdial_DOC-1.61-sap6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/wvstreams-4.6.1-patched_t2-sap6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/wvstreams_DEV-4.6.1-patched_t2-sap6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/wvstreams_DOC-4.6.1-patched_t2-sap6.pet

This will be in Squeezed Arm Puppy alpha3.

Playmusic music player

We have to use zigbert's older Pmusic 1.x in our RP Puppy, due to older ffmpeg. But, it doesn't work.

Sound is a work-in-progress on our Squeezed Arm Puppy!

Thunor has a neat little music player named 'Playmusic', and 01micko has hacked it to use mplayer, and had created a PET:
http://murga-linux.com/puppy/viewtopic.php?t=79359&start=60

This will be in the next Squeezed Arm Pup, alpha3. I have uploaded the PET to ibiblio (19KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/playmusic_mplayer-0.1.7-noarch.pet

01micko,
You mentioned that you are looking for a mixer that works in the RP. I compiled 'mixmos' 0.2.0, it looks good. But of course on my board I can't get sound to work so I can't verify that it is a good mixer.

Woof: getting close to RP pup alpha3

This is the latest Woof upload, with many fixes and tweaks for RP Puppy, the precursor to alpha3.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/a0cca3798b

I have got a bit more to do, so probably will upload alpha3 in about 2 days from now. It is Monday evening here, so perhaps sometime on Wednesday.

However, that's if I can get enough space on ibiblio.org. That is in great doubt though, anyone got somewhere I can upload to?

In the past, I have received offers to host my files, and I have always replied "she's right mate". I have treated ibiblio.org like a bottomless pit, which it isn't.
Then there's my own site with puppylinux.com, bkhome.org and goosee.com -- my host is hostgator.com -- those guys have warned me sometimes, my site gets too busy. Some of the guys at Hostgator are fans of Puppy, so I think they are extremely tolerant, but there are limits. As my site is now bulging at the seams, I have to think about uploading stuff elsewhere.

Pcurlftp, Planner ARM PETs

I compiled Planner project manager, that technosaurus had hacked awhile back to not need gconf, runs nice on the RP (390KB. 2.6MB, 737KB):
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/planner-0.14.5-patched1-sap6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/planner_DOC-0.14.5-patched1-sap6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/planner_NLS-0.14.5-patched1-sap6.pet

RP Puppy alpha2 does not have curlftpfs file-sharing PET, however 01micko has separated out the binary executables, compiled them on the RP and created PETs. I have uploaded them (97KB, 17KB, 5KB, 21KB, 5KB, 30KB):
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/pure-ftpd-1.0.36-armv6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/pure-ftpd_DOC-1.0.36-armv6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/mpscan-0.1.0-armv6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/curlftpfs-0.9.2-armv6.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/curlftpfs_DOC-0.9.2-armv6.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/pcurlftp_file_sharing-0.2.pet

Pcurlftp_file_sharing is a great little app created by kirk for sharing files over a network, see Forum post:
http://murga-linux.com/puppy/viewtopic.php?t=25537

...I dubbed it with the name "Pcurlftp".

Kirk reported yesterday that ibiblio.org has run out of space. Yes, today I have found that also, was unable to upload my PETs. So, I deleted a big un-needed file, then upload succeeded.

I then realised that some other distro will grab the space that I have released! So I did a ssh login and copied some big files, until it stopped with "no space left on device". Heh heh, next time that I want to upload something, I will delete one or more of those un-needed files, upload, then fill up the remaining space again.
Hopefully though, ibiblio will add some more space.

RP: desktop drive icons fixed

Testing Puppy alpha2 on the Raspberry Pi, Lobster reported that when a USB drive is unplugged, the partition icon remains on the desktop.

The culprit turns out to be our old friend Busybox again.

I built Puppy with the 'diff' DEB, which has become a dummy package as the proper package is now 'diffutils'. Consequently Puppy alpha2 is using the Busybox 'diff'.

This comes as a surprise to me, but the output format of the Busybox 'diff' is totally different to that of the diffutils 'diff'. Which is weird, because I seem to recall using diff in older versions of Busybox, and the format was the same as the full version of diff.

It seems that the Busybox diff is defaulting to the "unified" output format. This is supposed to be only done if the "-u" option is used.

Anyway, the problem is solved by installing the proper diff:
http://packages.debian.org/squeeze/diffutils

Alpha3 will have this fix.

Configuring Chromium

One thing that struck me about Chromium web browser is the limited configuration options in the "Preferences" window.

I wish to configure Chromium to minimise writes to Flash memory. The cache is the big one, and I ended up killing it this way:

# ln -s /dev/null /root/.cache

Chromium does complain at startup, but this method does work.

I wanted to know what commandline options there are, but the 'man' page has hardly anything. Instead, you have to look in the source code:
http://src.chromium.org/svn/trunk/src/chrome/common/chrome_switches.cc
http://src.chromium.org/svn/trunk/src/base/base_switches.cc

Apparently, these switches are not stable and liable to change at the whim of the developers.

One guy made ~/.config/chromium/Default/History and ~/.config/chromium/Default/Thumbnails into symlinks to /dev/null, as this files get written to a lot.
However, this causes an error message popup window at startup, although the browser still works. I was not able to suppress that error window.

Hardware setup of Raspberry Pi

I have written a page describing my hardware setup for the Raspberry Pi, with some technical notes and useful links:

http://bkhome.org/arm/rp-hardware-setup.htm

Latest SD image file for RP

I have uploaded my latest SD skeleton image file for the Raspberry Pi. Please not, this is for Woof developers, it does not contain Puppy!

Image file, compressed (2.2MB):
http://distro.ibiblio.org/quirky/arm/sd-skeleton-images/raspi-sd-4gb-skeleton-ext4_nojournal-07jul2012.img.xz

md5sum:
http://distro.ibiblio.org/quirky/arm/sd-skeleton-images/raspi-sd-4gb-skeleton-ext4_nojournal-07jul2012.img.xz.md5.txt

I have taken the time to write a very nice documentation page:
http://distro.ibiblio.org/quirky/arm/sd-skeleton-images/raspi-sd-4gb-skeleton-ext4_nojournal-07jul2012-README.html

RP: sound is broken

I cannot get sound to work on the RP. I tried everything described here:
http://elinux.org/R-Pi_Troubleshooting#Sound

I don't get sound from HDMI or the speaker socket.

Here are some RP Forum posts:
http://www.raspberrypi.org/phpBB3/viewtopic.php?f=28&t=10269
http://www.raspberrypi.org/phpBB3/viewtopic.php?f=28&t=10392
http://www.raspberrypi.org/phpBB3/viewtopic.php?f=28&t=9680
http://www.raspberrypi.org/phpBB3/viewtopic.php?f=28&t=8192
http://www.raspberrypi.org/phpBB3/viewtopic.php?f=38&t=8783

So, I'm not the only one. Reading through the RP Troubleshooting section, I guess that I should be thankful that I have got Puppy mostly working, sound only is broken.

RP: Full-screen video

Puppy alpha1 and alpha2 do not fill a HDMI screen. It is supposed to be 720p, however somewhere there is "overscan" imposed, which creates a black border.

The way to get the video to fill the entire screen is by appending this line to file 'config.txt' in the first partition:

disable_overscan=1

You can alternatively specify the overscan (black border) on each side, which you might want to do if the video is rendering too large on your screen and overlapping the edges.

The overscan settings are described here:
http://elinux.org/RPi_config.txt

Windows RT ARM laptops and secure boot

There are very interesting developments afoot. I was most intrigued when I first learned that MS was porting Windows to the ARM architecture. Now we are about to see the outcome.

MS plan to bring out a tablet with plastic keyboard, called "Surface tablet":
http://wp-life.com/microsoft-unveils-the-windows-rtpro-surface-tablet/

Toshiba also have a ARM RT laptop with touch screen and keyboard:
http://www.netbooknews.com/50789/toshiba-arm-with-windows-8-rt-ti-omap-4470-dual-core-in-video/

Of course, as we would expect, MS will do whatever they legally can to restrict competition. So what about this "secure boot" thing?
http://en.wikipedia.org/wiki/Windows_8

It seems that Ubuntu, Red Hat and Fedora are going to be able to get around the secure boot problem:
http://www.neowin.net/news/windows-8-secure-boot-plans-for-ubuntu-linux-revealed

...so, what do we do? It is a great shame that there is no open reference design ARM computer, like there was back in the early days of the IBM PC. An opportunity missed perhaps. If the Windows RT laptops/tablets become the "reference designs", we will want to be able to boot Puppy on them too.

They look like they might be very nice, the kind of laptop that I would like to own.

Plan to write an app

I might release one more official Wary/Racy, and probably get the Raspberry Pi and Mele pups to release quality (or close to). After that I might not release any more puppies, but will keep working on Woof at a slow pace.

The reason is, I want a change, and want to work an a big application. I have got some ideas.

This means that someone in our community can take on the next Dpup, Upup, (or Arch-pup, Mageia-pup) as an "official" Puppy. I think that 01micko will be developing another Slacko, as time allows.

ext4 reported as ext2

I have created Puppy on an SD card with a partition 4MB-aligned and ext4 f.s. without journal. For the record, it was created like this:

mke2fs -t ext4 -O ^has_journal -L puppy -m 0 -b 4096 -E stride=2,stripe-width=1024 ${SDDEVICE}2

I added that -E option after reading this link provided by mavrothal:
http://blogofterje.wordpress.com/2012/01/14/optimizing-fs-on-sd-card/

Puppy boots ok, but the problem is if I plug the SD card into my PC, the partition is identified as 'ext2'.

'disktype', 'guess_fstype' and 'blkid' all report it as ext2:

# disktype /dev/sdc2

--- /dev/sdc2

Block device, size 3.262 GiB (3502243840 bytes)

Ext2 file system

  Volume name "puppy"

  UUID F13CEC24-BB3B-4BEE-9C2C-BF4ACD4E1BF0 (DCE, v4)

  Last mounted at "/"

  Volume size 3.262 GiB (3502243840 bytes, 855040 blocks of 4 KiB)



# guess_fstype /dev/sdc2

ext2



# blkid /dev/sdc2

/dev/sdc2: LABEL="puppy" UUID="f13cec24-bb3b-4bee-9c2c-bf4acd4e1bf0" TYPE="ext2"

It won't mount as ext2 though, gives this error:

EXT2-fs (sdc2): error: couldn't mount because of unsupported optional features (240)

However, it does mount as ext4, with the system log:

EXT4-fs (sdc2): mounted filesystem without journal. Opts: (null)

Anyone think of a solution for this?
Perhaps I should try to contact Jesse, the guy who wrote guess_fstype -- he is a super clever hardware programming guy. He moved on from Puppy involvement a long time ago.
Maybe...

Yes! I have already solved the problem. The full 'blkid' from the util-linux package (version 2.18) reports correctly:

# ./blkid /dev/sdc2
/dev/sdc2: LABEL="puppy" UUID="f13cec24-bb3b-4bee-9c2c-bf4acd4e1bf0" TYPE="ext4"

So, now I have to implement this correct detection into the 'probepart' utility.

Anyone want to volunteer to make a bug report to the the Busybox developers? The version of Busybox that I am using is 1.19.3.

A problem with 'disktype' is, last time I checked, the ext4 support is broken. I did report this, and informed that there is another ext4 patch available (created by the Pardus distribution, for disktype v9) that works, which is what we use. However, I don't think the official disktype has been corrected. So, if anyone feels like reporting the ext4-reported-as-ext2 bug, you would also have to remind them that their ext4 patch is broken and they should use the Pardus patch.

Enhancing running on SD

Thanks to feedback from mavrothal and others, there have been excellent links to read.

I found these two also:
http://tombuntu.com/index.php/2008/09/04/four-tweaks-for-using-linux-with-solid-state-drives/
http://www.lesswatts.org/tips/disks.php

I am implementing many of the ideas into the next alpha. I have created a new file, /etc/rc.d/BOOTCONSTRAINED, which has various variables to set how the system can be constrained at bootup.

For example, the "VM writeback time" defaults to 5 seconds, I have changed that to 15 (see second link above).

I have changed the default I/O scheduler to 'deadline' (see first link above).

This is in addition to my SD image now having partitions with 4MB alignment. Also, the main filesystem mounts with 'relatime' (only updates access-time of files sometimes).

The next alpha will also use ext2 and swap will be off (even though the SD card has a swap partition).

ext2 with xip

As a followup to the previous thread:
http://bkhome.org/blog/?viewDetailed=02882

I have done a test build with ext2. I also put this into rc.sysinit:

mount -o remount,rw,xip /

I then opened 8 tabs in Chromium (each showing a different web page), then started Abiword and opened a document, then Gnumeric and opened a document, then Geany and again opened a document. These are all open at once, and no sign of breakage yet.

'free' shows usage of the swap partition steadily climbing as I started up more apps. Note, it only climbed a little bit as I opened up each tab in Chromium. Now using 6MB of swap.

This is very good news.

The swap partition is 256MB, and as more use of it is made, I would expect a slow down.

Hmmm, I have got to see when this breaks... opened Gnome-mplayer and played a tiny example video. Swap usage has climbed to 10MB.
Opened Osmo, Ayttm, Notecase... this pup is still going strong!

I am running out of things to open. But of course, this test is only going to be meaningful if I do the same thing without xip...

ext2 without xip
Ahem, I loaded all the same apps, got the same result.

Even going further, opened a second Chromium window, no problem.

So, xip doesn't seem to do anything. This has got me very curious, as Lobster reporting Chromium crashing after opening 4 tabs.
Let's see, trying one of my SD cards with ext4 f.s...

ext4
No problem, opens all the same apps. Interesting though, it is using much less swap.

This is fascinating, only 256MB and I it doesn't break. But, I wonder if there was no swap partition...

ext2, xip, no swap
Everything opens, no problem.

ext2, no xip, no swap
Ditto

I was expecting noticeable slowdown as I opened more apps, as pages get dropped from RAM, but responsiveness seems to remain the same. Very interesting.

These tests are telling me that the swap partition isn't needed, not for normal usage anyway. Maybe for heavier stuff like compiling. Nor xip, it doesn't seem to make any difference.

There is a good reason for not having the swap partition, is reduces writes to the Flash card. It seems that just allowing pages to be dropped from RAM does not lower reponsiveness noticeably.
...this reminds me, back in my uni lecturing days, when I thought that swap partitions were illogical, a cop-out for a badly designed OS.

ext2 vs ext3 vs ext4

All of the distros for the RP are using ext4 on the SD card, all the ones I have looked at anyway.

They are doing this because of the supposed advantages of journaling, however I had a reminder a couple of days ago not to rely too much that journaling will save the day.

The power went off my RP unexpectedly, and even after I had done a filesystem check, I still had some corruption. A couple of files were truncated, due to the filesystem buffers in RAM being unable to flush completely.

Also, journaling is bad for Flash memory as there are more writes to memory. It is also slower.

Then there is Execute In Place (xip) suggested by technosaurus. This will be very good for a system with not much RAM. Xip is only available for ext2 I think.

Puppy will do a f.s. check at bootup if there was an abnormal shutdown. This is done by creating a mini-Linux in a ramdisk, chrooting into that, unmounting the main partition then doing a f.s. check, then reboot. This all happens automatically. Fortunately, as I have compiled my own kernel, I set the ramdisk big enough (I think).

All of the above indicates to me that ext2 is the best choice for the RP. My kernel also has enabled ext2 support, with xip enabled. So, I reckon this will be the next thing to try.

Chromium home page

Here is a little tweak for Chromium, for those interested in configuring this web browser.

In Puppy Linux on the RP, alpha2, the first time that Chromium is run, it opens debian.org web site. Even though I have specified file:///usr/share/doc/home.htm as the home-page in file /root/.config/chromium/Default/Preferences.

The Mozilla browsers are also like this, they ignore the home-page setting on the first startup.

In the case of Chromium, you just have to create a file "/root/.config/chromium/First Run", 0-bytes in size, and at first startup Chromium will go to our Puppy home page. Also, a popup window asking if we want to set the Chromium browser as the default browser is avoided.

I have implemented this in the "chromium" template in Woof:
http://bkhome.org/fossil/woof2.cgi/info/872a712118

Woof version used to build Puppy for RP Alpha2

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/298047627a

Puppy for Raspberry Pi, alpha2

This pup runs well! Significant improvements relative to alpha1, which was uploaded only a few days ago:
http://bkhome.org/blog/?viewDetailed=02874

This pup is feature-complete, that is, to use that expression, it includes the "kitchen sink". But of course, as we expect with Puppy, still tiny -- the SD-card image download file is only 85.6MB.

Release Notes
Here is the main announcement and release notes:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.96-alpha2/sap-5.96-alpha2-readme.htm

...as you can see, there are still some bugs that we need to fix, plus I am sure more will be discovered. Sound... that will probably work after the correct kernel driver is loaded, and I will see about automating that for alpha3/beta1.
In general though, it is looking good, and I am personally quite surprised at how nippy the performance of the big apps like Chromium are on the RP.

Note that first-time bootup is much slower than subsequent bootups. So, if it seems stuck at first bootup, just wait. Note, there are things that I can do to significantly speedup the first-boot, which I will attend to soon.

Download
Here is the SD-card image (85.6MB):
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.96-alpha2/raspi-sd-4gb-sap6-5.96.img.xz

And instructions for writing it to an SDHC card:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.96-alpha2/howto-copy-to-sd-card.htm

If you want to do any C/C++/Vala/Genie/BaCon compiling, grab the "devx" file (see Release Notes for more details) (117.8MB):
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.96-alpha2/devx_sap6_5.96.sfs

Sha1sums and md5sums:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.96-alpha2/sha1sums.txt
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.96-alpha2/md5sums.txt

Personal comment
I have got to say, and of course I am probably biased, but I think that this pup is shaping up as the simplest and most complete out-of-the-box experience for newcomers to the RP.
Just about every app they might want ready-to-go, the Puppy Package Manager can install any app from the huge Debian repository, our simple "devx" one-file-install to transform our pup into a complete C/C++/Vala/Genie/BaCon compiler environment -- I recommend BaCon as a very easy language for programming-newbies.

Or, you can create utility and GUI apps without the the "devx" file, using just Ash/Bash, Perl or jimTcl scripting and gtkdialog (see Pburn, Pmusic and dozens of other apps created this way).
...yes, our tiny 85.6MB pup also has Perl and Tcl, albeit cutdown versions. If you really *must* have Python (see here why *not*: http://bkhome.org/genie/) then it is in the "devx". Heh, heh, I am just stirring some Python fans up, of course Python is ok if you want, but not much good for writing apps in Puppy as the "devx" is needed.

noarch PET upgrades

I have upgraded these in the 'noarch' repo, thanks to shinobar and zigbert for their ongoing commitment to improving these apps!

ffconvert 1.3 (shinobar)
pburn 3.6.1 (zigbert)
pfilesearch 1.30 (zigbert)
pfind 5.0 (zigbert)
pmusic 2.6.3 (zigbert)

Improve "welcome1stboot" window

The first time that Puppy is booted, after the QuickSetup window comes the "welcome1stboot" window -- with the title "Welcome, woof, woof" -- also, Puppy barks twice.

the SD cards we are booting off now on the ARM boards, are full installations, no layered filesystem, and there is no need for selecting where to save the session at first shutdown.

So, I have modified the welcome window to omit the save-session section when not wanted.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/449d8e6e1d

Raspberry Pi success

I have got to report this before signing off for tonight...

I managed to successfully compile my own kernel on the RP, and it boots. And, magically, the keyboard now works properly -- that is, arrow keys and del key now work.

Well, there were some things in the official RP kernel .config file that were turned off, that I thought should be enabled, so I did.

Furthermore, Retrovol now runs, except shows sound as muted. The Retrovol mixer window is completely broken. I tested playing a audio file with 'aplay', it seems to play but there is no actual sound -- but I don't know if I need to do anything special to get sound on my hdmi tv.

I will be busy with family stuff most of tomorrow, but in the evening I will do some more testing. It looks like 'alpha2' isn't far off!

Set date if no hardware clock

Squeezed Arm Puppy for the Raspberry Pi is built with 'Psync', an application created by tasmod to set the time/date from the Internet.

The first time that you bootup RP, if there is an Internet connection then QuickSetup will offer to run Psync.

However, if you don't have an Internet connection, then the date will be January 1, 1970. You would have to run 'Set date and time' in the 'Desktop' menu. Running with such an old date can cause problems.

So, I thought that Puppy should detect if there is no hardware clock, then set the date to something reasonable. I modified /etc/rc.d/rc.country and rc.shutdown.
rc.country will set the date/time to that when the Puppy was built (my local time here in Western Australia). rc.shutdown will save the time/date at shutdown, rc.country will read that at next bootup.

So, even though the time/date will be wrong, it won't be wildly wrong. It will also keep updating every time you bootup, but will fall behind, until of course you get onto the Internet.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/6de4ee9075

Squeezed Arm Puppy alpha1

Only 10 days ago I uploaded Squeezed Arm Puppy (sap) pre-alpha for the Raspberry Pi, announced here:
http://bkhome.org/blog/?viewDetailed=02865

A lot of bugs have been fixed, and I have added many applications, to bring this pup up to what we expect of our puppies. That is, to use that expression, it includes the "kitchen sink". But of course, as we expect with Puppy, still tiny -- the SD-card image download file is only 85.4MB.

Here is the main announcement and release notes:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap-5.95-alpha1/sap-5.95-alpha1-readme.htm

...as you can see, there are still some significant bugs that we need to fix, plus I am sure more will be discovered. In general though, it is looking good, and I am personally quite surprised at how nippy the performance of the big apps like Chromium are on the RP. Note that first-time bootup is slower than subsequent bootups.

Here is the SD-card image (85.4MB):
http://distro.ibiblio.org/quirky/arm/test/raspi-sap-5.95-alpha1/raspi-sd-4gb-sap6-5.95.img.xz

And instructions for writing it to an SDHC card:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap-5.95-alpha1/howto-copy-to-sd-card.htm

If you want to do any C/C++/Vala/Genie/BaCon compiling, grab the "devx" file (see Release Notes for more details) (117.8MB):
http://distro.ibiblio.org/quirky/arm/test/raspi-sap-5.95-alpha1/devx_sap6_5.95.sfs

Sha1sums and md5sums are online:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap-5.95-alpha1/sha1sums.txt
http://distro.ibiblio.org/quirky/arm/test/raspi-sap-5.95-alpha1/md5sums.txt

Improved menu assignment

When a Puppy is built in Woof, or a package installed with the Puppy Package Manager (not a PET), placement in the menu has up until now been somewhat unsatisfactory.

For example, the Sylpheed .deb package from Debian Squeeze creates two menu entries, in Network and Internet categories -- the former is not appropriate. Sometimes there is no menu entry created at all when there should be.

I have given this problem some attention and made improvements to scripts 2createpackages and installpkg.sh. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/6bc0629099

Another A10-based media player

Here is yet another media player thingy based on the Allwinner A10:
http://dx.com/p/mini-android-2-3-hd-1080p-network-media-player-w-wi-fi-hdmi-usb-av-tf-134686

sap6 progress

There has been steady progress with Squeezed Arm Puppy, and it is about time to do another upload. I might designate this as "alpha 1".

Yesterday I compiled the kernel for the RP, using the official patched 3.1.9 source from git. I followed the instructions carefully, and only made small changes to the official .config file -- for example, Puppy needs the ramdisk size to be 32MB. After doing it all so carefully -- I even compiled the kernel on the RP -- I was unhappy when I tested it and got a "kernel panic".

I can't be bothered to go through that again. Next time that I upgrade the kernel, I will just use the latest official compiled kernel and modules.

Anyway, "alpha 1" should be out in a couple of days.

Support kbd package

Sometime ago I removed support for setting the keyboard keys layout from the layout files provided by the 'kbd' package. The change was made back in late 2009:
http://bkhome.org/blog/?viewDetailed=01218
http://bkhome.org/blog/?viewDetailed=01219

I have now modified the scripts in Woof to recognise if the 'kbd' package is installed, and to use those keyboard layout files. Note, I have not yet tested it.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/fdec6256c2

BootManager: SFS handling improved

The BootManager offers to load SFS files at bootup, that is one of it's functions.

However, this cannot be done in the case of a full installation where there is no initrd, and hence no layered filesystem (unionfs or aufs).

I have modified BootManager to recognise this situation and explain how a SFS can be permanently installed by copying the files out of the SFS.

The BootManager also now recognises if shinobar's SFS-Load application is installed, and offers that as an alternative if load/unload on-the-fly is desired.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/1a28927a14

QuickSetup enhanced

I have fixed a couple of things in QuickSetup that GCMartin reported recently.

One of those problems was a confusing button to change the hostname. Another problem was the network-frame sometimes not displaying.

The network-frame has two new checkboxes, to enable the firewall and to enable setting time from NTP server -- the latter will only be displayed if tasmod's Psync is installed.

The NTP checkbox is to make it simpler to setup on small boards like the Raspberry Pi that do not have a hardware clock.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/ce27ec3fbe

Psync 2.8: get time/date from Internet

The Raspberry Pi does not have an internal hardware clock, so the date and time have to be set manually or via synchronisation from an NTP server on the Internet.

Forum member tasmod created Psync, that uses the 'ntpdate' utility (part of the 'ntp' package) to set time and date from the Internet:
http://murga-linux.com/puppy/viewtopic.php?t=42978

I already had Psync version 2.7 in the 'noarch' repo at ibiblio, however that was a mistake, as the package has the 'ntpdate' binary executable. I have now extracted the binary executable.

Latest Psync (6KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/psync-2.8.pet

'ntpdate' for x86 and arm (36KB, 40KB):
http://distro.ibiblio.org/quirky/pet_packages-common/ntpdate-4.2.6p5-i486.pet
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/ntpdate-4.2.6p5-armv6.pet

Now, on the RP, how do we set that up so that the date/time gets automatically set at bootup?
I haven't actually tried Psync yet, does it install something to do that?

Woof used to build sap6 pre-alpha

See previous blog post:
http://bkhome.org/blog/?viewDetailed=02865

I have uploaded the Woof used to build this pup, commit d3694f3cd8 (June 19, 2012):
http://bkhome.org/fossil/woof2.cgi/timeline

Squeezed Arm Puppy pre-alpha

Squeezed Arm Puppy for armv6, or just "sap6", for the Raspberry Pi pre-alpha build is available. This is for debugging, not yet release quality.

Find out more:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.91-pre-alpha/sap6-5.91-readme.htm

How to copy image to SD card:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.91-pre-alpha/howto-copy-to-sd-card.htm

Download files:
http://distro.ibiblio.org/quirky/arm/test/raspi-sap6-5.91-pre-alpha/

...yes, that is fully uploaded, the SD image file is just 64.3MB.

Debugging help wanted!

RasPi progress

A quick progress report, for those Raspberry Pi fans wondering what is happening...

I have Puppy booting on the RasPi. But, still some issues to resolve.

What has consumed much time is my decision to build from Gentoo binary packages. I have had to learn from scratch how Gentoo split up some packages and some oddities in the directory layout of some packages.

This morning I got a desktop, with mouse and keyboard working (that took awhile!). I am finding some problems, for example jesse's 'guess_fstype' does not detect ext4 or swap partition types, only fat -- so I have put a workaround in the 'probepart' utility to fallback to using 'blkid'.

I am now going to experiment with a build from Debian Squeeze packages.

armv6 pet packages uploaded

I have created about 200 PET packages for an armv6 CPU, specifically for the Raspberry Pi. I did this manually, and I have lost count of how many days it has taken.

Here they are:
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/

These are the 'common' PETs, that are likely to be needed for all Puppy builds.

Raspberry Pi, second impression

I deleted my post from yesterday, and all comments. Well, this is my personal blog, it is not like a public forum.

Anyway, I decided that I had got off on the wrong foot with the RP. I have sorted out the initial problems, now running Gentoo, and intend to get stuck into compiling apps so that I can build a Puppy.

Gotta run, but will post more details soon.

Woof used to build Precise Puppy 5.2.60

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/f155b46d8b

Precise Puppy 5.2.60 (beta2)

Precise Puppy beta1 was uploaded on June 3, see blog announcement:
http://bkhome.org/blog/?viewDetailed=02856

Beta2 is now out, download from here:
http://distro.ibiblio.org/quirky/test/precise-5.2.60/

Forum thread for feedback, discussion of beta2 starts on page 15:
http://murga-linux.com/puppy/viewtopic.php?p=632032#632032

My Raspberry Pi has arrived, so I will switch over to playing with that for awhile.

Linux + X on one floppy

I recall back in 2003, there were a couple of Linux distributions that were compiled from source and ran on one or two floppies -- and they had X windows and a desktop.

1diskXwin
This afternoon I tried to find them. Took a lot of searching, but I found one, known as '1diskXwin' or '2diskXwin':
http://soundproofingforum.co.uk/soundproofing/soundproofing.htm
http://linux.wikia.com/wiki/Mungkie
http://freecode.com/projects/natld

Thinking hard about this, I seem to recall there was one that had gtk 1.2 and Netscape/Mozilla and it still fitted on 2 floppies ...or, maybe my memory is faulty.
Or, it may have been 1diskXwin, as that has build scripts to download and compile GTK 1.2 and Mozilla.

Xwoaf
Xwoaf is another one. Puppy Forum member goingnuts is our specialist in small Linuxes, and has resurrected Xwoaf here (the original site is dead):
http://www.goingnuts.dk/linux.asp
Here is a Xwoaf download site. Ah, xwoaf is based upon 1diskXwin and needs it's sources:
http://www.mmnt.net/db/0/0/ftp.uni-duisburg.de/Linux/Xwoaf
...but that is only 0.1.0, and I think the latest is 0.2.0.

Blueflops
As far as I can determine, there is no source, only binaries.
http://sourceforge.net/projects/blueflops/

pUPnGO
This is created by goingnuts. Does not fit on 2 floppies but still tiny:
http://www.goingnuts.dk/linux.asp

REQUEST
Does anybody have Vmware installed? If so, I was wondering if you could do something for me. Get hold of this (27MB):
http://soundproofingforum.co.uk/soundproofing/mungdev20100301.rar
which is described here:
http://soundproofingforum.co.uk/soundproofing/soundproofing.htm

...and extract the files from the Vmware image, make it into a tarball.

The reason is, the other 1diskXwin source tarballs do not have the sources for some extra apps, such as 'mdesk' -- I would like to examine those sources and compile them.

RasPi coming soon

I received the paperwork from Element14, dated 29 May, which probably means that my Raspberry Pi was posted also. Tomorrow (Monday) is a public holiday here, so I am expecting the parcel to arrive on Tuesday.

When it does arrive, I will most likely switch over to working intensely on porting Puppy to it.

Which means that the x86 Precise Puppy will languish. I uploaded beta1 early this morning, and already have fixed many things. I will continue that until Tuesday morning I think, then upload beta2.

After that, Precise Puppy development will slow down for awhile.

Woof commit 2012-06-03

This is the Woof used to build Precise Puppy 5.2.58 (beta1). Commit:
http://bkhome.org/fossil/woof2.cgi/info/1343570ea5

For an introduction to Woof and how to download Woof, go here:
http://bkhome.org/fossil/woof2.cgi/home

Precise Puppy 5.2.58 (beta1)

I am being bold, and announcing this latest build of Precise Puppy as the first "beta quality"!

There are lots of improvements relative to the alpha3 (5.2.55) release on May 24, which was announced here:
http://bkhome.org/blog/?viewDetailed=02842

Hopefully the xorg video tests will be better this time. The 'llvm' package has been included, needed by the radeon kms xorg driver, which has bumped the .iso file by about 5MB, which is unfortunate, but we have to live with it -- just for the sake of supporting the 'r300' series of Radeon (ATI) video chips.

This pup can be downloaded from here:
http://distro.ibiblio.org/quirky/test/precise-5.2.58/

I have compiled the 3.2.18 kernel in Precise Puppy, and propose that this be the one used for the release version of Precise Puppy -- and well into the future. So, anyone who feels inspired to compile a 3rd-party kernel driver, go for it! nVidia and ATI PETs are welcome!

The kernel .sfs can be got from here (though I intend to also upload it to ibiblio, later today, so will be installable via the 'install' icon on desktop):
http://bkhome.org/blog/?viewDetailed=02854

Forum thread for feedback, beta1 feedback starts on page 12 of this thread:
http://murga-linux.com/puppy/viewtopic.php?p=631371#631371

Known issues
A couple of things with the PPM:

1. No app icons, need to restart PPM.
2. search all repos, click an app, cannot determine size and deps.

For the first one, you have to exit then restart PPM to see the application icons. I have already fixed this in Woof.

For the second one, not so good, if you search all repos, then click to install an app, PPM cannot determine size and deps. I will look into that very soon.

Getting ready for Precise beta1

I think that I might be able to build Precise Puppy beta1 tonight.

Over the last day or so, I have fixed a few things, updated some PETs, but haven't documented it here.

But, some notes on stuff done this morning...

Pemasu sent me a 'probepart' script that was modified by Jemimah. While I was playing with that, I realised that probepart outputs duplicate /dev/sr0 lines. This seems to be a problem introduced with the 3.2.x kernel. I fixed it, see next Woof commit for details.

Vicmz sent me updated Spanish langpack, and another langpack for LibreOffice. These are uploaded to the 'noarch' PET repo.

Um, thinking back to yesterday: I updated FFConvert, Pburn and sfs_load, added brcm driver firmware to Woof.

Big family outing today, so I'll get back to Puppy stuff this evening.

Linux kernel 3.2.18

I have compiled this kernel in Precise Puppy alpha3. Only change from before is that I have enabled the brcm network drivers.

PET packages (808KB, 23.2MB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_headers-3.2.18.pet
http://distro.ibiblio.org/quirky/pet_packages-common/linux_kernel-3.2.18-pae-i686-aufs-up.pet

The sources and build scripts are here:
http://bkhome.org/sources/kernel-3.2.18/

ROX-Filer squashed log-window

Iguleder reported about a problem with a squashed log-window and a patch that he had created:
http://bkhome.org/blog/?viewDetailed=02849

The patch is here:
http://repo.or.cz/w/rox-filer.git/patch/37a0c90d8a7e69cedf94cee6e84eafd945ffd02f

I have applied the patch and created a new PET (377KB):
http://distro.ibiblio.org/quirky/pet_packages-common/rox_filer-20111022-patched3-w5c.pet

The new source package is 'rox-filer-20111022-patched3.tar.gz', it is patched with 'rox-filer-iguleder-20120526.patch', both uploaded to my sources repo:
http://bkhome.org/sources/alphabetical/r/

Installwatch 0.7

My 'common' PET repo has had an old version of Installwatch, 0.6.3, for a very long time. This is broken with recent util-linux (or was that coreutils) package, which causes 'new2dir' to be broken.

Some other builds, I think that Slacko is one example, have created their own updated Installwatch PET.

Anyway, I have finally updated the official PET in the 'common' repo, now version 0.7 (20KB):
http://distro.ibiblio.org/quirky/pet_packages-common/installwatch-0.7-i486.pet

The sources, as usual, are in my 'sources' repo. Note that Installwatch is part of another package, Checkinstall, but I have extracted it to it's own package 'installwatch-0.7.tar.gz', see here:
http://bkhome.org/sources/alphabetical/i/

I think that there were reports that 'new2dir' does not work in Precise Puppy alpha3, this should hopefully fix it.

PPM: Application icons

I discussed the possibility recently of displaying an icon alongside every package entry in the Puppy Package Manager, and zigbert advised how that can be done with gtkdialog.

So, I have implemented it, for both the Classic and Ziggy user interfaces in PPM. It is optional, set by a checkbox in the Configure panel.

There is a rather heavy price to pay for this bit of extra glamour in the PPM. I created a PET with a large collection of 16x16 xpm application icons. Actually, there are "only" 532 icons, but 3788 symlinks. The PET is 211KB.

It does need some work though, as some of the symlinks are ridiculous. I used a script to create them. It now needs a lot of manual tweaking. Maybe even a 24x24 set would be nice?

PET (211KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/icons_puppy_app16-20120529.pet

Woof commits:
http://bkhome.org/fossil/woof2.cgi/info/530dac8798
http://bkhome.org/fossil/woof2.cgi/info/5e592eeebf

I will put this into the next release of Precise Puppy.

xpm versus png load time

I am creating a set of 16x16 icons for the PPM. Following zigbert's advice, I have got icons working in the PPM main window, and can have a unique icon for each package. That's another topic, I'll announce that soon.

The icons though... if we are going to have an icon for each package, installed or not, that is a lot of icons. Hundreds of them. But, I thought, if I can make the icons small enough, it won't increase the size of puppy much (and, I need to make their usage optional in the PPM).

I found a nice collection of SVG application icons in the Open Icon Library:
http://openiconlibrary.sourceforge.net/downloads.html

I wrote scripts to convert them all to tiny xpm 16x16 images. I will bundle those scripts together here for the record:

for ONESVG in `find open_icon_library-full/icons/svg/apps -type f -name '*.svg' | tr '\n' ' '`

do

 ONEBASE="`basename $ONESVG .svg`"

 echo "$ONEBASE"

 rsvg-convert -w 16 -h 16 -o png-16x16/${ONEBASE}.png $ONESVG

 sync

done



for ONEPNG in `find png-16x16 -type f -name '*.png' | tr '\n' ' '`

do

 ONEBASE="`basename $ONEPNG .png`"

 echo "$ONEBASE"

 pngtopnm $ONEPNG | pnmquant 16 | ppmtoxpm > xpm-16x16/${ONEBASE}.xpm

 sync

done



for ONEXPM in `find xpm-16x16 -type f -name '*.xpm' | tr '\n' ' '`

do

 ONEBASE="`basename $ONEXPM .xpm`"

 echo "$ONEBASE"

 sPTN='s%"  c black"%"  c None"%'

 sed -e "$sPTN" $ONEXPM > xpm-16x16-alpha/${ONEBASE}.xpm

 sync

done

I have a fondness for xpm images, although I realise they are "yesterday's technology'. I like being able to open the file in a text editor!

I was wondering how these tiny images compare, xpm versus png. taking abiword image as an example, sizes in bytes (using default compression factors):

abiword.xpm 663
abiword.xpm.gz 382
abiword.xpm.xz 412

abiword.png 258
abiword.png.gz 277
abiword.png.xz 316

Xz does not do well with such small files, but in a .sfs file would probably get much smaller size.

Well, comparing abiword.xpm.gz and abiword.png, the former is slightly bigger. I wonder if abiword.xpm would get closer to abiword.png size with a bigger data-set as in a squashfs? As it stands, png is the winner here.

But, what about loading time, that is, to read the file and render it on-screen? Given that png images will have to be fed through a decompression routine, I always assumed them to be inherently slower to load than xpm. Xpm does of course need translation of the text, so it also has an overhead. I hunted around on the web couldn't find any comparison data. Perhaps the difference is negligible, and I should just choose the smaller png files.

Regarding size, so far I have created 294 xpm images. If I convert those to png, each about 250 bytes, the total size is about 75000 bytes, about 73KB. That's not too bad. Note, in a squashfs .sfs file there is no such thing as a minimum file size (as when stored separately on a drive), tiny files are packed together and there is no wasted space.

Note, I am not going to put these images into Woof! They will go into a PET. Woof already has too many images. The PET can then be optionally used in a Puppy build.

Gconv-modules cache

This is amazing. Pemasu and Radky have been puzzling over why gtkdialog-based applications are slow to startup. The amazing thing is, they have found out why, and found the fix. This can shave noticeable startup time off of some apps! And we thought that Puppy was already super-fast!

Here is pemasu's report:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=105

The problem is that Puppy lacks the cache file /usr/lib/gconv/gconv-modules.cache, which is a compiled data file. The file is created in the Glibc source package when it is compiled, but we never put it into Puppy.

The matter is made worse, because Puppy has always had a cutdown selection of files in /usr/lib/gconv. Therefore, using the cache file as-is is not correct.

I wrote a little script that creates a cutdown /usr/lib/gconv/gconv-modules, and then a correct /usr/lib/gconv/gconv-modules.cache is generated using the 'iconvconfig' executable.

The correct files, including the script, are now in Woof, in the 'glibc' template, in /usr/lib/gconv-cut, and FIXUPHACK updates /usr/lib/gconv. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/36a75da77e

I also put the corrected files into the 'glibc' PET package for Wary. See:
http://distro.ibiblio.org/quirky/pet_packages-wary5/glibc-2.10.1-1-w5c.pet

Note: I posted a 'gconv-modules.cache' to the Forum, however that one is not quite correct. The files in the Woof commit are the latest.

Why not sfs load-on-the-fly

Peebee asked why I don't include shinobar's SFS load on-the-fly PET:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=90

Over the years I have learned to be very wary of modifying the Aufs or Unionfs layers "on-the-fly". It often just doesn't work right.

Anyway, I decided to give it a go. Running Precise Puppy alpha3. I got 'sfs_load' out of shinobar's PET 1.9.4, and placed it at /usr/sbin.

I modified /usr/sbin/filement to recognise if sfs_load exists, and to offer the choice of mounting a file for viewing (prior behaviour), or loading it with sfs_load.

Choosing the latter, it loads ok, as bottom layer, and reports success. However, the files on the bottom layer are not visible "on top". the remount worked to append the layer, but did not update the filesystem.

Was I amiss in only installing /usr/sbin/sfs_load?
Note, my /etc/rc.d/BOOTCONFIG already had the EXTRASFSLIST variable with a "devx" SFS file in it, and sfs_load inserted the new SFS file correctly.

The files on the bottom layer will become visible after a reboot, but that defeats the purpose.

mtPaint 3.40+ 20120525

Precise Puppy testers have reported that mtPaint is very old. I have tried as much as possible to use the official Ubuntu DEBs, including mtPaint, however that is version 3.34.

So, I have compiled mtPaint in Wary 5.3, downloaded from wjaguar's git, which is officially 3.40 but has some things done since the 3.40 release. I have put the PETs into the 'common' repo for the convenience of all puppies to include in their package-list.

PETs (311KB, 3KB, 360KB):
http://distro.ibiblio.org/quirky/pet_packages-common/mtpaint-3.40-20120525-i486.pet
http://distro.ibiblio.org/quirky/pet_packages-common/mtpaint_DOC-3.40-20120525-i486.pet
http://distro.ibiblio.org/quirky/pet_packages-common/mtpaint_NLS-3.40-20120525-i486.pet

this mtPaint will be in the next release of Precise Puppy, but testers are of course welcome to try it now.

PPM: bug fix, help update

I reported a bug in PPM here (in Precise alpha3 release):
http://bkhome.org/blog/?viewDetailed=02843

I have fixed that bug, also have updated the PPM help page. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/85ec359858

Translation team
The PPM help page has been updated, so translations also have to be updated.

You can download the Help page:
http://bkhome.org/fossil/woof2.cgi/artifact/fff94b0e6b88516166c955fba1cb8202c92b4dc9
(if you are logged in, you get a download link)

Precise testers
Precise Puppy alpha3 testers can also grab the latest /usr/local/petget/postfilterpkgs.sh, which has the bug fix:
http://bkhome.org/fossil/woof2.cgi/artifact/3534dde2db4d76b67de06bdf9d6c467940862582

Precise: SeaMonkey 2.9.1

I have compiled SeaMonkey 2.9.1 in Precise Puppy alpha3.

PETs (19.7MB, 19.6MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/seamonkey-2.9.1-up.pet
http://distro.ibiblio.org/quirky/pet_packages-precise/seamonkey_DEV-2.9.1-up.pet

PPM: simplifying DEB management

There is something that has been in the back of my mind for awhile. Given that Upups and Dpups are popular, despite the rampant dependencies, increasing complexity and considerable rearranging of file layout (all of which, as a KISS proponent, I find quite depressing), we need to make the best of the situation.

In Woof-built puppies, I undo a lot of the mangling wraught by the Debian and Ubuntu developers. One thing that I do is put many packages back together again. For example, in Precise Pangolin, the Gutenprint package is split up into these:

cups-driver-gutenprint_5.2.8~pre1
gutenprint-doc_5.2.8~pre1
gutenprint-locales_5.2.8~pre1
libgutenprint2_5.2.8~pre1
libgutenprint-dev_5.2.8~pre1
libgutenprint-doc_5.2.8~pre1
libgutenprintui2-1_5.2.8~pre1
libgutenprintui2-dev_5.2.8~pre1
printer-driver-gutenprint_5.2.8~pre1

It is a good thing to have -doc_, -locales_ and -dev_ packages split out, however they often go a lot further than that. The above is a mild example, showing separate library packages.

Now, the difficulty that arises. "Zillions" of little packages displayed in the PPM causes confusion. really, all that you want is to have the correct name of the package displayed, in this example "ghostscript", and PPM has checkboxes for you to filter out the exe, doc, dev, and nls components.

If only the generic names are displayed, this would oh so much simplify and greatly reduce the number of packages displayed in PPM. There could be a checkbox to make this switchable, that is, show all packages or only generic names.

So, if generic names checkbox is ticked, only "ghostscipt" would display instead of all those above. Clicking on the ghostscript entry would then install all of the above, subject to the exe,dev,doc,nls filtering of those checkboxes.

This would, I think, greatly improve the PPM for Debian and Ubuntu repos. I have been thinking about how to code it, and it is do-able. Any thoughts on this are welcome.

Just an extra thought, for those who know a lot more about gtkdialog than me: is it possible to have a different icon alongside each package name in the main window? Right now, there is just a list of packages, and you click on the one that you want to install. No icons. But, I think, back when I first coded it, it was possible to have an icon on each row, but only the same icon on every row -- I want to be able to have a unique icon on a row.

Woof: build Precise Puppy alpha3

This is the version of Woof used to build Precise Puppy alpha3 (5.2.55). Last two commits:

Latest: http://bkhome.org/fossil/woof2.cgi/info/a0f79d8bd7
http://bkhome.org/fossil/woof2.cgi/info/0364d17677

Precise Puppy alpha3

Precise Puppy, built from Ubuntu Precise Pangolin binary packages, has reached 'alpha3' (version 5.2.55).

Alpha 2 was released on March 26:
http://bkhome.org/blog/?viewDetailed=02766

Download version 5.2.55 from here:
http://distro.ibiblio.org/quirky/test/precise-5.2.55/

A lot of bugs have been fixed since alpha2, and this build is working very nicely, I am running it right now.

There are some issues with the Puppy Package Manager:

1. I clicked on 'desk_icon_theme_darkfire' and it installed that, but also wanted to install 'desk_icon_theme_original'. Weird.
2. The EXE, DEV, DOC, NLS checkboxes seem to be working for PET packages, but not for DEBs.

So go easy on using PPM, I will check it out tonight or tomorrow morning.

Note, I also intend to compile the latest SeaMonkey tonight.

The Forum thread for feedback, discussion on alpha3 starts on page 7:
http://murga-linux.com/puppy/viewtopic.php?t=76961&start=90

Pmusic, FFConvert, Peasyscan

PET updates:

Pmusic 2.6.1 (zigbert)
FFConvert 1.2.2 (shinobar) -- also FFConvert NLS pkg.

I move this from 'noarch' to 'common' repo, as it has a binary executable...

Peasyscan 2.0 (rcrsn51)

VIA neo-itx $49 ARM board

Ha ha, here they come, the Raspberry Pi look-alikes!

This one is called 'APC' and is fairly similar to the RasPi -- same arm11 (armv6) class of CPU. It has been announced all over the place, here is one link:
http://www.reghardware.com/2012/05/22/chip_maker_via_outs_49_dollar_raspberry_pi_alike/

The official website is down right now, probably overloaded:
http://apc.io/

This site has a video:
http://liliputing.com/2012/05/via-apc-a-49-android-computer-with-an-arm11-cpu.html

It only has Android 2.3, but for sure there will be plenty of people endeavoring to get Linux running on it.

Precise Puppy progress

Just a short note to let anyone interested, that I am back on developing Precise Puppy. Using it now. I intend to bring out "alpha 3" soon.

Note, I discovered a bug in rc.shutdown, that was introduced recently, like in the last week. I will do another Woof commit tonight with the fix.

ROX-Filer: Open With menu icons

My ROX-Filer PET packages have never had icons in the "Open With..." menu. That is, you get a list of applications, abiword, geany, gnumeric, etc., but with generic icons alongside.

It is only a small thing perhaps, but it is nice to have the appropriate icon for each application. So, I have implemented that and updated the ROX-Filer PET package.

PET here (376KB):
http://distro.ibiblio.org/quirky/pet_packages-common/rox_filer-20111022-patched2-1-w5c.pet

Tray applets internationalised

Rodin.s has internationalised our little tray applets: freememapplet_tray, network_tray and powerapplet-tray, refer to earlier blog thread about Xarchive internationalisation:
http://bkhome.org/blog/?viewDetailed=02836

I have made these into PETs (4KB, 6KB, 5KB):
http://distro.ibiblio.org/quirky/pet_packages-common/freememapplet_tray-2.4.pet
http://distro.ibiblio.org/quirky/pet_packages-common/network_tray-2.5.pet
http://distro.ibiblio.org/quirky/pet_packages-common/powerapplet_tray-2.5.pet

The sources are uploaded to my sources repo:
http://bkhome.org/sources/alphabetical/

The .pot files have been added to Woof, for the convenience of use with MoManager:
http://bkhome.org/fossil/woof2.cgi/info/e5a7542639

Xarchive internationalised

I announced support for .rar archive files in Puppy, and we were discussing Xarchive. L18L had reported that the C source is not internationalised, and rodin.s took up the challenge and did it:
http://bkhome.org/blog/?viewDetailed=02828

Thanks very much! I have compiled the new internationalised Xarchive in Wary and uploaded the PET to the 'common' repo (38KB):
http://distro.ibiblio.org/quirky/pet_packages-common/xarchive-0.2.8.6.i18n-w5c.pet

Note, I have not yet updated the ARM PET.

I uploaded the new source to my sources repo (xarchive-0.2.8-6-i18n.tar.gz):
http://bkhome.org/sources/alphabetical/x/

For the convenience of creating translations with MoManager, I have placed the 'xarchive.pot' file into Woof, in rootfs-skeleton/usr/share/doc/nls/xarchive -- MoManager will automatically read this and offer it to be translated.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/0d0f0f3d8b

I just remembered something. L18L and rodin.s, if you read this, kindly let me know: I think that you guys i18n'ed a couple of the tray apps?
I still have the non-translated versions in the 'common' PET repo. Would you mind posting the links to where you did this work? I will update those also.

Automounting fixed

Forum member bala4dj asked how to automount drives when they are plugged in:
http://murga-linux.com/puppy/viewtopic.php?t=78321

I replied that the Puppy Event Manager already supports that. Bala4dj replied that it only launches Pmount, doesn't automount.

The Event Manager is in the System menu, and the 'Icon Handler' tab has checkboxes labeled:

Tick box for auto-detection of target application
Tick box for auto-launch 'handler' when media plugged in

The first is ticked by default in Puppy. if the second one is ticked, if you plug in a DVD video then the default DVD player will automatically run. Similarly, plug in a music CD and the default music player will run.

The original intention was that this auto-launching of the 'handler' would also automount a drive if it is plugged in. However, at the time I constrained it to only launch Pmount.

I have now fixed it. If both of those boxes are ticked (and X restarted), if a USB drive is plugged in, all of it's partitions will automount and ROX-Filer windows will open.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/c2d6e84dff

A note to developers about /usr/local/bin/drive_all. When a drive is plugged in, icons get created on the desktop, and corresponding directories get created in /root/.pup_event.
For example, for /dev/sda3, directory /root/.pup_event/drive_sda3 is created, with script 'AppRun' inside it -- that script is a copy of 'drive_all'.
When someone clicks on the 'sda3' icon, 'AppRun' will execute.
The script responsible for setting all this up is /sbin/pup_event_frontend_d, which is launched as a daemon when X starts.

PPM: Mageia pkgs, bug fixes

I have enhanced the EXE, DEV, DOC, NLS checkboxes in the Puppy Package Manager (PPM) to recognise Mageia package naming.

I took out some common code to new script /usr/local/petget/postfilterpkgs.sh.

I discovered a few bugs that were introduced in recent changes to the PPM, one causes the Configure window to crash.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/f6f14d019b

Woof: Mageia build fixed

The Mageia "Drake Puppy" build for the upcoming Mageia version 2, was broken. Forum member scsijon requested that this be fixed sometime ago, I have only now got around to it.

The version 2 repositories do not yet exist, have to build from the 'cauldron' repos. No problem, Woof can be updated after Mageia 2 is released.

The problem was that the support/mageia2ppm script did not work properly. Now fixed, and works nicely for 'core', 'nonfree' and 'tainted' repositories.

I have only tested it as far as running the '0setup' script, and I updated the package list.

Woof commits:
http://bkhome.org/fossil/woof2.cgi/info/e737c34dd0
Latest: http://bkhome.org/fossil/woof2.cgi/info/163a8b8e64

Linux kernel 3.2.17

I have compiled the Linux kernel 3.2.17. I used 3.2.13 in the alpha2 build of Precise Puppy. Pemasu sent me some recommendations for the pre-compile configuration, which I have incorporated into the 3.2.17 compile.

Note, 3.2.17 is using Aufs, whereas 3.2.13 used Unionfs.

The PETs (23.4MB, 832KB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_kernel-3.2.17-pae-i686-aufs-w5.pet
http://distro.ibiblio.org/quirky/pet_packages-common/linux_headers-3.2.17.pet

The build scripts and sources, including kernel SFS file, are here:
http://bkhome.org/sources/kernel-3.2.17/

I put the PET into the 'common' repo, for convenience of all builds to use it. I intend to use it for the alpha3 build of Precise Puppy (Upup), but note that it was compiled in Wary so if any extra modules are compiled using the kernel source SFS, it should be done in Wary.

shutdownconfig: minimum size of save-file

Jemimah and Karl Godt observed a problem at first shutdown. It is possible to choose the size of the save-file smaller than the size required to save all the files created in the session.

Karl Godt created a patch for /usr/sbin/shutdownconfig to fix that:
http://murga-linux.com/puppy/viewtopic.php?t=60678&start=120

Thanks, Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/59d09dfc1a

PPLOG 1.1.3

L18L has enabled non-English text in PPLOG, Puppy's personal blog:
http://murga-linux.com/puppy/viewtopic.php?t=78229

Iwill provided a patch to fix the problem of "day 31":
http://bkhome.org/blog/?viewDetailed=02754

Over the years, we have made little improvements to PPLOG, whereas the official project has remained effectively dead. The author 'fedekun' has moved the official project to here:
http://code.google.com/p/pplog/

Interesting, 'sc0ttman', who is one of our Puppy fans, has enhance PPLOG and named it SJPPLOG:
http://pplog.scottjarvis.com/?viewDetailed=00021

...I will need to check that one out!

For now though, the two fixes by L18L and Iwill are applied and the version number bumped to 1.1.3:
http://distro.ibiblio.org/quirky/pet_packages-noarch/pplog-1.1.3.pet

Remaster script: support xz squashfs

01micko posted a patch for /usr/sbin/remasterpup2, to support xz compression of sfs files:
http://murga-linux.com/puppy/viewtopic.php?p=623475#623475

Thanks to scsijon who posted a link to that patch.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/8cb34303da

Woof: support .rar archives

We do encounter these .rar files every now and again. I had to expand one a few days ago, which reminded me that we don't support them in most flavours of Puppy. Wary and Racy don't.

It is not just a matter of installing the 'unrar' PET, of which there are a few lurking around.

Some people have done that, and then set the run action in ROX-Filer to exec xarchive "$@", or to run xarchiver if that is installed.

Woof-built puppies already have correct run action for mime type application/x-rar, which is exec pupzip "$@".

/usr/local/bin/pupzip is a wrapper around whatever archiver utility is installed, be it xarchive, xarchiver, or various others.
pupzip had to be fixed to recognise .rar files, which I have done.

I have also added the unrar package into the package-lists of most of the flavours in Woof, that is, Wary, Racy, Slacko, Upup, Dpup, etc.

Here is the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/193471fa09

I have uploaded an unrar PET to the 'common' repo (88KB):
http://distro.ibiblio.org/quirky/pet_packages-common/unrar-3.7.8-x86.pet

I have just now (re)discovered some very interesting work done by abushcrafter:
http://www.murga-linux.com/puppy/viewtopic.php?t=56651

...I have posted this link here for future reference.

Woof: buglets fixed

A few small fixes, see:
http://murga-linux.com/puppy/viewtopic.php?t=77303&start=105

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/cf31283d88

Staying synced with online Woof

My online Woof repository has instructions on how to download to your PC and get started with using Woof.

However, my online docs did not explain how to stay synced with the latest Woof.

01micko and pemasu have been doing this, but it should be documented for all to be able to easily do. 01micko has created a tutorial on this, which has inspired me to add to my own online docs.

My 'Getting started with Woof' wiki page now has information on syncing with latest Woof:
http://bkhome.org/fossil/woof2.cgi/wiki?name=Getting+Started+with+Woof

This is 01micko's tutorial:
http://www.smokey01.com/01micko/woof2/fossil-update.html

nVidia SFS files

Shinobar has been supporting nVidia video hardware with SFS files. This is for many versions of Puppy, back to 4.3.1, also including the latest Wary and Racy 5.3 and Slacko 5.3.3.

Forum thread:
http://www.murga-linux.com/puppy/viewtopic.php?t=72405

Shinobar's package hosting site:
http://182.164.210.61/puppy/opt/

For the convenience of Wary/Racy users, I have uploaded the 'nvidia-glx-sdl-295.49.sfs' to the SFS repo (downloadable via the 'install' icon on the desktop):
http://distro.ibiblio.org/quirky/sfs/

HOSTNAME updated if hostname changed

An issue with the network when hostname is changed, was discussed recently on this blog, and I applied a fix in Woof:
http://bkhome.org/blog/?viewDetailed=02813

There has been ongoing discussion on the Puppy Forum:
http://www.murga-linux.com/puppy/viewtopic.php?t=77743

...in which it was mentioned one problem is that the global variable HOSTNAME does not get updated until a reboot. Some applications read this variable.

HOSTNAME is exported in /etc/profile at bootup. However, as QuickSetup recommends to restart X after hostname is changed, there is the opportunity to export the updated HOSTNAME variable.
This is already done in the case of the LANG variable, which is exactly the same situation, exported in /etc/profile, and may be changed by QuickSetup.

The code that exports the updated LANG and HOSTNAME variables is /usr/bin/xwin. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/c440074e1c

Mele: SD-card image, "Lui"

"Lui" is my codename for this build. Puppy built from Ubuntu Lucid Lynx armel binary packages. It is alpha quality. This is an SD-card image, for a 4GB card.

Download the image from here. The image file is only 87.1MB, includes all applications:
http://distro.ibiblio.org/quirky/arm/releases/alpha/

Further explanation:
http://distro.ibiblio.org/quirky/arm/releases/alpha/README-mele-sd-4gb-lui-5.2.90.htm

PPM: Ziggy user interface fixed

01mick posted a bugfix here:
http://murga-linux.com/puppy/viewtopic.php?p=625843#625843

Thanks, I have put this into Woof. Commit:
http://bkhome.org/fossil/woof2.cgi/info/02dc746288

OK, packing up. Have family-visiting duties today. Someone's birthday.

Linux_firewall internationalised

L18L has internationalised the Linux_firewall install script:
http://murga-linux.com/puppy/viewtopic.php?t=77882

I have updated the PET:
http://distro.ibiblio.org/quirky/pet_packages-noarch/linux_firewall-2.0rc9-patched3.pet

Mele: SeaMonkey compiling, and compiling...

Ha ha, what a laugh. In my Mele A1000 ARM box, I am running Ubuntu Lucid Lynx armel. Yesterday I tried, for the first time, to run SeaMonkey.

It aborted at startup with "illegal instruction".

Goodness gracious, this is a Cortex-A8 CPU! My thinking is that whoever compiled SM must have done so on a higher-level CPU and not configured the compile options properly. Or, could they have configured for some instruction in a lower-level CPU that is not in the A8?

So, I decided to compile SM myself, on the Mele. I am doing it on a USB 2.5 inch hard drive, as my SATA drive does not work with the Mele yet.
One partition on the hard drive is for swap, essential as the Mele only has 512MB RAM.

Last night I left it running, went to bed. Got up this morning, it is still compiling.

Man, I would hate to do this on a RasPi! Only 256MB RAM. I would of course use my USB hard drive, with swap partition, so it probably is doable, if you have a few days to spare and don't mind thrashing the hard drive.

Cdburner Wizard internationalised

L18L has internationalised /usr/sbin/cdburner-wizard. Plus a small improvement was made.

L18L posted the code here:
http://bkhome.org/blog/?viewDetailed=02770

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/788b0e7a60

Xgamma-gui fix

L18L posted a fix for Xgamma-gui. Earlier on, L18L internationalised this script, now there is a fix for the range settings of red, green and blue:
http://bkhome.org/blog/?viewDetailed=02770
http://208.109.22.214/puppy/viewtopic.php?t=76182&start=201

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/59ed978392

A10 .fex and .bin config files

I am posting this for my own future reference

On the Mele A1000 there is a compiled configuration file named 'evb.bin'. This has been a mystery to me, but the fog is gradually clearing.

This is useful information:
http://www.cnx-software.com/2012/05/06/editing-allwinner-a10-board-configuration-files-script-bin/

There is another utility to convert a .fex file to .bin, available as binary executable only. This page has a link to it:
http://elinux.org/Hack_A10_devices
Direct download link of 'script' executable:
http://goo.gl/Zwhbk

3builddistro now supports ARM build to SD

This completes the basic support of building for an ARM board target with Woof.

I have modified the script '3builddistro' to create an SD card image. This uses a skeleton image, as I described earlier:
http://bkhome.org/blog/?viewDetailed=02808

3builddistro writes the skeleton image to an SD card, then populates it with kernel, modules, and Puppy filesystem. Then a xz-compressed image file is created that can be distributed for others to use.

Good news, as xz compression is being used, the compressed image file is currently only 88MB. That is my Ubuntu Lucid Lynx based build, with all of the usual applications, including SeaMonkey.

Woof commits:
http://bkhome.org/fossil/woof2.cgi/info/2bd2a350d2
http://bkhome.org/fossil/woof2.cgi/info/9f7c222f6e

Configuration, 1st boot full-hd installation

Well, "full-SD" installation really. When doing a cross-build in Woof, running on a x86 machine for an ARM target, and creating an SD-card image which is effectively a "full installation" of Puppy, certain things were not setup properly.

When you use the Universal Installer to do a full installation to hard drive, a chroot to run rc.update is done, which performs required pre-configurations.
But that is not happening in the case of building the SD image in Woof, as a chroot cannot be done from a x86 host to a ARM target filesystem.

So, I have modified /etc/rc.d/rc.update to perform these missing configurations at the first boot.

In my earlier announcement of Puppy Linux running on the Mele A1000, I mentioned that the JWM tray was missing -- this is one of those missing pre-configurations, the file /root/.jwmrc was not created.
That problem should now be fixed.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/4f789d3df9

We are still not there though, with Woof. To build an SD image that will bootup on the Mele A1000 (or other ARM system), I have to enhance '3builddistro' -- will do that soon.

Slacko Puppy 5.3.3

This year, Slacko Puppy project coordinator Mick Amadio ('01micko' on the Puppy Forum) became a full-time college student, so we expected considerable slowdown of development of Slacko. However, like the rest of us, Mick is a Puppy addict, and has now released version 5.3.3. Mick, I hope that your college studies are not neglected!

Mick has created a nice new Slacko page:
http://01micko.com/slacko.html

Here are the Release Notes:
http://distro.ibiblio.org/pub/linux/distributions/puppylinux/puppy-5.3.3/release-Slacko-5.3.3.htm

Extracted from the Release Notes, a brief overview:

Slacko Puppy Linux 5.3.3 is an update release of the recent 5.3.1. It has binary compatibility with Slackware-13.37, which simply means that it is a Puppy built with packages from the Slackware, Salix and Slacky repositories.
The main version has been upgraded with kernel 3.1.10 compiled with Aufs layered file system support, in the typical Puppy manner. There is also a PAE version to cater for machines with large amounts of RAM. Both iso images have SCSI boot support.
The Seamonkey-2.9.1 suite is the default browser and email suite but Firefox, Chromium, Opera, Netsurf, Iron, Dillo and Links are only a few clicks away.
Minimum system requirements for the main version are an i686 processor, 900MHZ, with 256MB RAM. For the PAE version a slightly newer processor may be needed as most but not all processors support PAE.
Slickpet is a cut down version of Quickpet to get a few handy apps without diving into the well stocked Puppy Package manager. Slickpet also supports installing SFS packages, which are modules that are loadable as layers in the Puppy filesystem. Packages such as Libreoffice-3.5.3, JRE and Samba3.5.4 are available.
While on the Puppy Package Manager, apart from Puppy's native PET package format it can also fetch Slackware, Salix and Slacky packages in tgz or txz format.
Puppy's sfs format is supported by a program called sfs_load. Sfs are a way to load layered filesystems so that the filesystems can be unioned and act as one.
Slacko has Abiword and Gnumeric, late versions to open those documents and spreadsheets to keep you working.
Lots of types of multimedia are supported so you can play hard.
Slacko is aimed at reasonably modern machines up to six or seven years old. It may run on even older hardware.
Much work has gone into fine tuning packages such as Samba and the light weight alternative Samba-TNG. There is also a netboot add on that allows you to boot Puppy over your LAN.

If you are interested, here is my blog announcement for Slacko 5.3.1:
http://bkhome.org/blog/?viewDetailed=02621

Ibiblio download link:
http://distro.ibiblio.org/pub/linux/distributions/puppylinux/puppy-5.3.3/

...note, Ibiblio may be slow. Go to my Puppy download page for faster mirrors:
http://puppylinux.com/download/

Disconnect from network if hostname changed

Gcmartin started this thread:
http://www.murga-linux.com/puppy/viewtopic.php?t=77743

The script in Woof that is used to change the hostname is /usr/sbin/hostname-set.

I have modified this, so that if the hostname is changed, any network connection is brought down, and information is displayed that it must be reconnected for the new hostname to take effect.

From before, the post-change window also says that X must be restarted for the new hostname to take full effect, but I don't remember why that is required, if at all.

The new script can be viewed here. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/886f20e9b5

PPM: enhanced splitting of exe,dev,doc,nls packages

Continuing with enhancing of the Puppy Package Manager, see previous post:
http://bkhome.org/blog/?viewDetailed=02811

The PPM main window has checkboxes labeled "EXE", "DEV", "DOC" and "NLS".

The "EXE" checkbox was broken. The "exe" packages (anything other than doc, dev or nls) were not filtered out when the checkbox was unticked. Fixed.

I have enhanced these checkboxes. Previously they only worked on PET packages, and only those packages conforming to the "_EXE", "_DEV", "_DOC" and "_NLS" naming scheme, for example 'abiword_DOC-1.2.3.pet'.

This has been extended to recognise the package naming used in Debian and Ubuntu. DEB packages have these strings in the names:

-doc_
-docs_
-dev_
-langpack
-lang-

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/dc585512f0

PPM: Multi-keyword search

There were some comments posted to this blog awhile back, about improving the search capability of the Puppy Package Manager:
http://bkhome.org/blog/?viewDetailed=02716

I have only just started to look at this.

Mavrothal has posted some patches, one of which is to allow multiple keywords in the search box, and to search both package name and description:
http://www.murga-linux.com/puppy/viewtopic.php?t=77564

I have implemented this part of Mavrothal's patch only, and increased it from 3 to 4 keywords:
http://bkhome.org/fossil/woof2.cgi/info/afe5f65a84

I experimented with Mavrothal's patch to search all repositories, however found some issues. One was a bug, the button to choose the current repo can in fact have two lines of text displaying in the button. The second issue is that search results should identify which repo the found package is in.

I will take a look at implementing it with these issues addressed.

Blog under attack

Some idiot in Spain is trying to crack my blog. My host has warned me. The ip address is 87.219.27.60.

Unfortunately, the attack has caused my host to place some restrictions on my site, for awhile.

I will make this blog read-only, while I consider this.

gtkdialog: widget ordering wrong

This is a new one. I am using gtkdialog 0.8.0. Here is the SNS (Simple Network Setup) main window on my laptop:

With Puppy running on the Mele A1000 ARM-based system, GTK 2.20.0, running SNS:

The content is slightly different, as these snapshots are taken on different machines. But see how the rendering order of the widgets is reversed.

So, what causes this? Other applications are displaying ok on the Mele. Is there some GTK setting that needs changing, or a problem in gtkdialog?
I will contact thunor, the current maintainer of gtkdialog, perhaps he will know what the cause of the problem is.

Puppy running on Mele A1000!

Yes! Here is the very first screen-shot, to prove it:

This is built from Ubuntu Lucid Lynx armel binary packages, plus an assortment of PETs that I compiled (see recent blog posts).

Two things that are broken, that I am aware of:

1. urxvt
At first I used the 'urxvt-lite' DEB, but it complained could not find a font and aborted. Then I compiled my own rxvt-unicode, but that reported that it cannot "initialize pseudo-tty" and aborted.

2. jwm tray
Jwm is running, but not the tray. Probably the .jwmrc file is broken.

These will be fixed, for sure. But I thought that I would announce this significant event.

The way that I intend to distribute this (when get to releasing it), is as an SD card image. Compressed, it will be quite small.

Creating an SD image from scratch is rather difficult, so I have taken the easy way out. If you look here:
http://distro.ibiblio.org/quirky/arm/sd-skeleton-images/

...the Woof '1download' script will download the one you want. These skeleton images are intended to be for whatever ARM board you want to build for, you choose the appropriate one.

Then, the '3builddistro' script only has to add the kernel and Puppy files to the SD card, and it is all ready to go.

Apart from the above bugs, a bit of coding to do in Woof, but I hope to have something released very soon.

RasPi killer?

I have been expecting one or more boards to appear on the market with superior specs to the Raspberry Pi but about the same price.

The most likely will be something based on the Allwinner A10 or it's cut-down version the A13. Which is partly why I am using this architecture myself.

don570 has posted about the first of them, though it is preliminary, not yet for sale:
http://murga-linux.com/puppy/viewtopic.php?t=77917

The company is Olimex and they are claiming the board will be about 30 Euros:
http://olimex.wordpress.com/2012/04/24/cortex-a8-in-tqfp-sure-allwinner-a13/
http://www.olimex.com/dev/a13-olinuxino.html

One significant difference is that it will have VGA only, not HDMI output. My Mele has both (both provided by the A10 chip). The RasPi only has HDMI.

More packages compiled for armv7

This is a continuation of my previous post:
http://bkhome.org/blog/?viewDetailed=02805

I have added more PETs:
http://distro.ibiblio.org/quirky/arm/pet_packages-armv7/

These are the extra packages:

getcurpos
glipper-0.89
gwhere-0.2.3-patched_4
libcddb-1.3.2
nenscript-1.13.3
picpuz-2.4
rubix-1.0.6
setvol
squashfs-3.3
squashfs-4.2
streamripper-1.64.6
xload-1.1.0

Some sources have been added to my sources repo.

I have made some notes on how to compile, see comments below.

armv6-compiler-person wanted

I have compiled all of this lot manually:
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/
http://distro.ibiblio.org/quirky/arm/pet_packages-armv7/

These are "common" packages that are likely to be in all puppies and aren't available in the host-distro's packages (or the packages in the host-distro are unsuitable, such as too many dependencies).

Iguleder is doing some great development to get this all automated, something that we should look at in the future. Getting started with ARM, I have been feeling my way, and discovering a few issues with compiling, so perhaps doing it manually this time was good -- except for the amount of time it took.

I have compiled some armv6 PETs, at least I think they are, that should work in the Raspberry Pi. Due to backwards compatibility, they will work in the armv7-based Mele, which is why I haven't compiled those PETs for armv7.

I can't say that I'm keen on the thought of doing all that again, recompiling all those armv7 PETs as armv6. That's where I am seeking help. If anyone has a armv6-based box of any kind, or emulator, if you want to compile any of those packages, please do. Probably best if we went for a base armv6 with software floating point.

All of the source packages are here:
http://bkhome.org/sources/alphabetical/
Us#er#na#me: pu#pp#y Pa#ss#wo#rd: li#nu#x

After compiling a package, all that you have to do is open up the corresponding armv7 PET and bung in the replacement binaries, then change all occurrences of "armv7" to "armv6" in pet.specs file, and rename the package. Like this:

# tar -zxf dbus-123-armv7.pet
...put in armv6 binaries, edit pet.specs, change dir name.
then put it together again...
# dir2tgz dbus-123-armv6/
# tgz2pet dbus-123-armv6.tar.gz

Most packages are pretty straightforward to configure and compile. If running on armv6 hardware then is probably not even necessary to specify the '--build' parameter.
I have made some notes for some packages, that I will post as comments to this blog thread.

There are a couple of BaCon apps in Puppy, that you will find in the devx: /usr/sbin/pngoverlay.bac and /usr/sbin/welcome1stboot.bac. I compiled those for armv7, so they need to be redone for armv6.

Mele A1000 testbed

Here is a photo of my testbed:

It is an improvisation, due to lack of space. The "table" is a piece of chipboard on top of my camping fridge.

The Mele is on the right. It has three USB sockets, which is good, but I have plugged in my USB hub for convenience to plug in more devices.
Notice the little daughter-board hanging off a cable -- that is a socket for a SATA 2.5 inch laptop drive -- definitely on my to-buy list, when the budget allows.

The only thing that I am finding inconvenient is that I have to set time/date at every bootup, as I am working off-line. I don't know if the motherboard has any capability to add a battery-backed clock, or whether all it needs is a battery.

There is a snapshot of the motherboard. This was posted to the arm-netbook mail-list and the full-size photo is 2590x1943:
http://i.imgur.com/1xd2j.jpg

Here it is reduced:

Some of the sockets on the motherboard are a mystery, but I do know that the one at the bottom is a UART, and Tom (a guy on the arm-netbook mail-list) kindly provided me with a serial-to-usb converter so that I can monitor kernel bootup on my laptop.

Ubuntu: Want to mount tmpfs on /tmp

I thought perhaps there might be an Ubuntu guru reading this...

I am running Ubuntu Lucid Lynx off an SD card on my Mele A1000 ARM box.

It is a normal "full" install, and I think that it is hitting the SD card really hard, that is, lots of writes to it. In particular, a lot of stuff happens on /tmp, and I would like to mount a tmpfs (in RAM) onto /tmp at bootup.

However, I am unfamiliar with the init/bootup scripts used in Ubuntu -- is it the 'launchd' system? Anyway, I will need to mount the tmpfs on /tmp as early as possible in the bootup, and I cannot see what script to put it into.

In Puppy, I would do that in /etc/rc.d/rc.sysinit, right after remounting '/' as rw.

ext4 support for save-file

01micko posted patches to add support for the ext4 filesystem in the save-file:
http://bkhome.org/blog/?viewDetailed=02798

Thanks, I have implemented this, see commit (the relevant files are rc.shutdown and shutdownconfig):
http://bkhome.org/fossil/woof2.cgi/info/52c69e3b0e

Another A10-based board

My Mele A1000 has an Allwinner A10 CPU/GPU SoC chip, as do a few tablets, and at least one upcoming netbook. Well, some of them have the A13, a slightly cut-down version.

Here is another board, that was reported by Krasi on the Arm-netbook mail-list:
http://micdigi.com/2012/04/hyston-will-release-the-google-tv-box-based-on-a10-solution/

I don't know, but my guess is that blank space on the board is for a hard drive?

Perhaps the board will not be available stand-alone.

ROX-Filer for ARM CPU

This has been very educational. I compiled ROX-Filer in the Mele A1000, which has a A10 chip with Cortex-A8 ARM CPU. Operating system is Ubuntu Lucid Lynx.

Without specifying any CPU-tuning, it seems to have automatically compiled for a 'armv7' CPU, which is what the Cortex-A8 is. Fine, I got a stripped executable of 361KB. Works great.

However, as the Cortex-A8 is backwards-compatible with the arm1176jzf-s CPU (armv6) used in the Raspberry Pi, I had another go at compiling ROX-Filer. This time I did this before the configure step:

# export CFLAGS='-mcpu=arm1176jzf-s'

Note: the 'configure' script in ROX-Filer reads the global CFLAGS and prepends to it's own options. So, CFLAGS is the same in both cases except the latter case has the additional -mcpu option.

The good news is that this works fine on the Cortex-A8 CPU. The bad news is that the stripped executable is 489KB!

Yikes. I think that this is because the armv7 has the 'thumb2' instructions, which greatly reduce executable size. For the first time I am seeing just how much!

I have created a PET package. This is for Lobster to test, as he is the only one who has a RasPi. Lobster, grab this:
http://distro.ibiblio.org/quirky/arm/pet_packages-armv6/rox_filer-20111022-patched2-armv6.pet

Expand it:

# tar -zxf rox_filer-20111022-patched2-armv6.pet

You have to get the files over to your RasPi somehow, and manually install the files. Do not put them in a fat or ntfs partition, only a Linux partition -- might be better to get the pet into the RasPi then run the tar operation.

ROX-Filer is launched via /usr/local/bin/rox, so just open a terminal and type "rox". Please be running as 'root' when you test rox.

Note, I am greatly enhancing the Ubuntu Lucid installation on the Mele -- well, it isn't really an "installation" as it just runs off the SD card. I am thinking of making an image soon and upload it somewhere -- but it will be about 1GB compressed, and I will have to go to a friend's place to do the upload.

For reference, if you want to see the GNU compiler ARM parameters:
http://gcc.gnu.org/onlinedocs/gcc/ARM-Options.html

Earlier posts about Ubuntu on the Mele:
http://bkhome.org/blog/?viewDetailed=02799
http://bkhome.org/blog/?viewDetailed=02796

Mele: installing -dev packages

I posted recently about getting Ubuntu Lucid Lynx running on my Mele A1000 ARM box:
http://bkhome.org/blog/?viewDetailed=02796

I haven't got Internet access from it yet, and it needed a heap of '-dev' packages to turn it into a compile environment. So I had to chase up all the DEBs and dependencies manually and download them on my laptop, then transfer them over to the Mele and install them. Which took a few hours, well, more than a few.

Ubuntu no longer have the ARM DEBs for Ubuntu Lucid Lynx in the official repos, but I found them here:
https://launchpad.net/ubuntu/lucid/armel/

If you know a package name, you can append it to the URL, for example:
https://launchpad.net/ubuntu/lucid/armel/libxrender-dev

Time to sleep. In the morning I plan to compile some apps.

Ha ha, today I bought a USB keyboard. This one appeals to the child in me! Actually, it is practical, as I need it to fit into a small space. It is a Migear Junior keyboard:

Price was AU$14.95 from Dick Smith:
http://dicksmith.com.au/product/JT3004/migear-junior-learning-qwerty-compact-keyboard

The keyboard has nice big letters, very easy to see. It is a delicate construction though -- the Mele is sometimes a bit slow to respond when I hit a key (running Lucid with Gnome2!) and when I first started to use the keyboard, I thumped a key a bit too hard (thinking that it hadn't worked) and it stopped working entirely -- however I pried the key-cap off and examined underneath -- there is a delicate little rubber mechanism, which must have got squashed. I put it together again and the key works. I find that typing lightly on the keys is the best, which is ok by me -- I have been using it for a few+ hours and like the feel of it.

/dev/mouse fix in full hd install

Pemasu reported this, here is Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/b5721241f3

Backlog!

I have a lot of pm's waiting for me at the Puppy Forum. Then there are many pages of posts in the Wary/Racy 5.3 thread that I have not read.

Then there's Precise Puppy, that I started, see this blog and Forum thread:
http://bkhome.org/blog/?viewDetailed=02766
http://murga-linux.com/puppy/viewtopic.php?t=76961

Forum member pemasu has taken an interest in this and has applied fixes and upgrades and started a new thread:
http://murga-linux.com/puppy/viewtopic.php?t=77697

I have got my nose down working on getting Puppy running on the Mele A1000. Getting up to speed on a completely difference non-PC architecture is taking awhile, plus there is a lot that has to be compiled. Right now I am experimenting with compiling the 2.6.36 kernel, such that it will be Puppy-compatible.

I can see this ARM stuff keeping me busy for a long time. I have not yet decided to aim for getting a Upup running on the Mele, or compile everything in T2.

There is quite a bit of interest in the x86 Precise Puppy, and pemasu is doing some good work with it. So, do you think that we should ask him to take it on as an "official" Puppy?

I might be building a ARM port of Precise Puppy, so will provide some upgrades in Woof that will also apply to the x86 build. So, the condition would have to be that pemasu stays synced with the latest Woof.

I haven't even asked him yet, he might not want to do it!

I notice that pemasu is having trouble with printing (not working). Yeah, it may be necessary to roll cups, gutenprint, ghostscript and foomatic-filters back to the PETs used in Wary/Racy.

Your thoughts on this are welcome...

Running Lucid Ubuntu on Mele A1000

This is very exciting! I have got Ubuntu running on my Mele A1000.

Tom posted a video, see my earlier blog post:
http://bkhome.org/blog/?viewDetailed=02794

To save me the hassle of cross-compiling u-boot, kernel and installing Ubuntu on an SD card, Tom uploaded his own SD card image. This made it so easy -- I just used 'dd' to copy the image to the card, plugged the card into the Mele and turned it on.

For anyone else who has a Mele and wants the SD image, see the 'arm-netbook' mail list:
http://lists.phcomp.co.uk/pipermail/arm-netbook/
Or:
http://blog.gmane.org/gmane.comp.hardware.netbook.arm

The image is 2GB compressed, and uncompressed is 4008706048 bytes. You need a 4GB SD card that has that many or more bytes capability -- some are a bit less -- run 'disktype' on it to find out.

I did find that the compressed image can be made smaller, I got 931MB, but we haven't uploaded a smaller one yet.

We currently don't know what Tom set the root password to, but I just edited /etc/shadow and inserted the "woofwoof" password from my Puppy system.

No GPU acceleration yet, but I am surprised how snappy the responsiveness is. Tom's image has a lot of apps, including OpenOffice. Also gcc compiler stuff, so it is ready for me to compile. Good, I will start creating some ARM PETs -- such as that gtkdialog PET that Lobster wants

But first, I have to go out today and buy a USB keyboard. For now, I am using it with mouse only.

Sound works. Video is set to 1920x720 (720p) 50Hz, I have to find out how to change that.

Anyway, having fun!

Human Devolution deleted

Yesterday I posted about the book "Human Devolution" that I am currently reading, which as I expected, invited some strong responses.

I have given this some thought, and decided that it was too Off-Topic for this blog.

I posted awhile back about the book "The China Study" that documents massive "knowledge filtering" in the health sciences.
That is one thing, but Human Devolution, although it documents knowledge filtering in the archeology/history fields, also goes heavily into spiritual/paranormal fields, and that latter is too OT.

This blog is primarily for reporting on Puppy Linux development, with the occasional foray into some interesting OT areas. However, heavy OT topics really belong in some other blog. If I had "Barry's Philosophy Blog" for example, it would be appropriate for such postings.

My apologies to those who replied to that thread and had their post deleted.

Mele A1000 running Ubuntu video

Tom has made a video of a Mele A1000 running Ubuntu:
http://www.youtube.com/watch?v=BhnupMPCHSU&feature=youtu.be&noredirect=1

There is not yet any GPU video acceleration. The page that explains how to install Ubuntu is here, but is out-of-date and Tom will be updating it soon:
http://rhombus-tech.net/allwinner_a10/hacking_the_mele_a1000/

Here is my blog post that introduces the Mele A1000:
http://bkhome.org/blog/?viewDetailed=02754

Sunlike UMPC-1021
A netbook will soon be available with the Allwinner A10 chip:
http://www.cnx-software.com/2012/04/16/sunlike-umpc-1021-85-usd-allwinner-a10-based-10-2-android-4-0-netbook/

It is not yet in mass production, the developers are still debugging the camera.

Apparently it will also be available with a 4000mAH battery which will power the netbook for 10 hours.

Here is a video of the Sunlike UMPC-1021. Note that "Boxchip A10" is the same as the Allwinner A10:
http://www.youtube.com/watch?v=MbL7rymDuE8

On holiday

As from today, I will be on holiday for seven days.

Just to let everyone know that I won't be very responsive to any questions or reports.

I might keep looking into the USB uevent detection stuff in the initramfs, but won't have time for much else.

USB drives kernel uevents

I was recently communicating with a couple of the kernel developers on one of the kernel mail-lists, about the 'usb-storage' patch that I use in Puppy kernels. Certain 'printk' lines were removed from the kernel at version 2.6.33, that the 'init' script in the initramfs needs to ascertain when all USB drives are available for use.

I developed a patch for later kernels. I raised this problem with the kernel developers and asked how to achieve the same goal without the patch. I wasn't given a solution as-such, but was informed that I should be able to glean the required information from monitoring the kernel uevents.

Uevents are status messages that the kernel generates. The usual scenario is that a daemon such as 'udevd' monitors these and is able to respond appropriately when hardware is detected, added or removed.

In the initramfs I don't use udevd, however I do have various other means for monitoring uevents in the initramfs. Anyway, first things first. I conducted an experiment to examine what uevents we need to look out for. Using Wary 5.1.2 with 2.6.32.x kernel, I modified /sbin/wait4usb to log certain information, using the 'hotplug2stdout' utility ( a uevent monitor daemon). It was a frugal HD installation, and I booted with two USB Flash drives plugged in.

This is the log just after the 'usb-storage' kernel module is loaded:

<7>usb-storage: device found at 4

<7>usb-storage: waiting for device to settle before scanning

<7>usb-storage: device found at 5

<7>usb-storage: waiting for device to settle before scanning

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6 SUBSYSTEM=scsi DEVTYPE=scsi_host SEQNUM=1165 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/scsi_host/host6 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/scsi_host/host6 SUBSYSTEM=scsi_host SEQNUM=1166 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7 SUBSYSTEM=scsi DEVTYPE=scsi_host SEQNUM=1170 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/scsi_host/host7 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/scsi_host/host7 SUBSYSTEM=scsi_host SEQNUM=1171

...the above shows what my 'usb-storage' patch is logging, as well as the relevant uevents.
...what the above indicates is that if we look for two kernel uevents for "scsi_host", that would indicate two drives are detected.

After a few seconds, the 'wait4usb' script reports this:

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0 SUBSYSTEM=scsi DEVTYPE=scsi_target SEQNUM=1173 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0 SUBSYSTEM=scsi DEVTYPE=scsi_device MODALIAS=scsi:t-0x00 SEQNUM=1174 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0/scsi_disk/6:0:0:0 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0/scsi_disk/6:0:0:0 SUBSYSTEM=scsi_disk SEQNUM=1175 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0/scsi_device/6:0:0:0 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0/scsi_device/6:0:0:0 SUBSYSTEM=scsi_device SEQNUM=1176

And then about another second later:

<7>usb-storage: device scan complete

<7>usb-storage: device scan complete

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0 SUBSYSTEM=scsi DEVTYPE=scsi_target SEQNUM=1177 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0 SUBSYSTEM=scsi DEVTYPE=scsi_device MODALIAS=scsi:t-0x00 SEQNUM=1178 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0/scsi_disk/7:0:0:0 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0/scsi_disk/7:0:0:0 SUBSYSTEM=scsi_disk SEQNUM=1179 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0/scsi_device/7:0:0:0 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0/scsi_device/7:0:0:0 SUBSYSTEM=scsi_device SEQNUM=1180 

change@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0 ACTION=change DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0 SUBSYSTEM=scsi SDEV_MEDIA_CHANGE=1 DEVTYPE=scsi_device DRIVER=sd MODALIAS=scsi:t-0x00 SEQNUM=1181 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0/block/sdc ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0/block/sdc SUBSYSTEM=block MAJOR=8 MINOR=32 DEVNAME=sdc DEVTYPE=disk SEQNUM=1182 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0/block/sdc/sdc1 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.3/2-1.3:1.0/host7/target7:0:0/7:0:0:0/block/sdc/sdc1 SUBSYSTEM=block MAJOR=8 MINOR=33 DEVNAME=sdc1 DEVTYPE=partition SEQNUM=1183 

change@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0 ACTION=change DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0 SUBSYSTEM=scsi SDEV_MEDIA_CHANGE=1 DEVTYPE=scsi_device DRIVER=sd MODALIAS=scsi:t-0x00 SEQNUM=1185 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0/block/sdb ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0/block/sdb SUBSYSTEM=block MAJOR=8 MINOR=16 DEVNAME=sdb DEVTYPE=disk SEQNUM=1186 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0/block/sdb/sdb1 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0/block/sdb/sdb1 SUBSYSTEM=block MAJOR=8 MINOR=17 DEVNAME=sdb1 DEVTYPE=partition SEQNUM=1187 

add@/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0/block/sdb/sdb2 ACTION=add DEVPATH=/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.2/2-1.2:1.0/host6/target6:0:0/6:0:0:0/block/sdb/sdb2 SUBSYSTEM=block MAJOR=8 MINOR=18 DEVNAME=sdb2 DEVTYPE=partition SEQNUM=1188

...once again, the "device scan complete" is from my patch. Now, the uevents identify all the drives and partitions.

This is good. It looks like enough information is provided to do away with my patch. I have not yet implemented a solution, thinking about it. I might modify the 'hotplug2stdout' daemon to detect when all USB drives are ready, and signal the main 'init' script.

Anyway, I have documented the test results so far, as it will probably interest other Puppy developers. Note, the original 'hotplug2stdout' C source is in my sources repo.

Lucid Puppy 528 -005 released

Although I have posted about Lucid Puppy 528 being de-emphasized, not an "official" release anymore, it continues to be actively developed and will interest many people.

There are various reasons why Lucid might suit you:

1. One is compatibility with the Lucid Lynx DEB package repositories.
2. Another is that it has many enhancements provided by some keen developers.
3. Another reason is that it has become very mature.
4. Being mature, it has also accumulated a good collection of PETs that you can install.
5. Also, it is still being actively supported, so if you have a problem, there are people to help you.

Some reasons why you might not be interested in Lucid:

1. Built from an old version of Woof, so is missing many of the fixes and improvements that I have announced on my blog over the last couple of years -- though Lucid has been going on an independent parallel path with it's own development, so there will be many Woof-level improvements since the original fork from the official Woof (plus some ported from the official Woof). So, this is not necessarily a reason against.
2. The Lucid Lynx repos are getting old. I don't know when Ubuntu will drop them. Also, as Lucid Puppy gets upgraded with it's own PETs, I don't know about compatibility with (some of) the DEBs on the Ubuntu repos.

If you are having issues with the puppy that you are currently running, try Lucid.

On April 5, 2012, playdayz released the 005 sub-version release of Lucid Puppy 528. See Forum announcement:
http://murga-linux.com/puppy/viewtopic.php?t=70855

Discussion of the 005 release starts on page 94 of the Forum thread:
http://murga-linux.com/puppy/viewtopic.php?t=70855&start=1395

Minit: alternative to init

I have implemented Minit in Puppy. This is a process-1 service manager alternative to the Busybox 'init' system. Minit does not use /etc/inittab, it has its own /etc/minit directory with a quite different structure.

Minit home page:
http://www.fefe.de/minit/

I have tested in Racy, booting off a USB Flash drive, and created a PET package. The PET installs and co-exists with the Busybox init system. At bootup, if /sbin/minit is detected, then Minit is used. If the PET is uninstalled then everything goes back to init, as before.

...although I haven't thought through what will happen if 'minit' is deleted while in use. I didn't really envisage it being uninstalled, more a choice made in Woof when building a Puppy.

The PET, compiled statically (16KB):
http://distro.ibiblio.org/quirky/pet_packages-common/minit-0.10-i486.pet

There is also a documentation PET, that as well as man files also has a couple of intro PDFs (159KB):
http://distro.ibiblio.org/quirky/pet_packages-common/minit_DOC-0.10-i486.pet

The catch: Minit needs this very latest Woof, with modifications to some scripts:
http://bkhome.org/fossil/woof2.cgi/info/6ddf124baf

See also the previous commit:
http://bkhome.org/blog/?viewDetailed=02789

Will I use Minit in a future Puppy? Hmmm, I don't know if it is really giving me anything extra that I am already doing with Busybox init. Potentially, it could be used to provide more logical structure to the provision of services at bootup.

It was a good exercise, clarified some things that I had not fully understood before. I have setup Woof to seamlessly use Minit, and created PETs, so it could be easily adopted if myself or anyone else takes a fancy to it in the future.

Do not delete /tmp/* at shutdown

I was experimenting with 'minit' a simple alternative to the 'init' process-1 startup service manager. Note, we currently use the Busybox 'init' system.

I noticed something rather odd. I found that when /etc/rc.d/rc.shutdown executes, it somehow kills 'mingetty', which promptly restarts.

With the Busybox init system, /etc/inittab specifies the start of mingetty, which performs a login on /dev/tty1, and thus we have our console. Then X can be launched from the console.
Here is the content of /etc/inittab, as we currently use in all puppies:

::sysinit:/etc/rc.d/rc.sysinit

tty1::respawn:/sbin/mingetty --autologin root tty1

tty2::respawn:/sbin/getty 38400 tty2

tty3::respawn:/sbin/getty 38400 tty3

::ctrlaltdel:/sbin/reboot

The "respawn" means that if mingetty gets killed for whatever reason, the 'init' daemon (that runs from start to shutdown in the background) will restart it.

I do not really understand why we need "respawn", as we only kill mingetty once, at shutdown. Anyway, my discovery, while testing minit (which does not use /etc/inittab, but does something equivalent to run mingetty), was that rc.shutdown was killing mingetty and then the 'minit' daemon immediately restarted it.

I have not yet confirmed whether this happens with Busybox init, but it would go toward explaining why I have sometimes been very confused by some of the stuff appearing on the console while shutting down.

Anyway, I isolated the offending line in rc.shutdown:

rm -rf /tmp/*

...but I still haven't determined just what inside /tmp causes mingetty to be killed.

I have commented this line out. Some PUPMODEs run a tmpfs mounted on /tmp, so the contents of /tmp get automatically removed at shutdown, so the above line makes no difference. However, some PUPMODEs, such as 2 (full hd installation) and 13 (booting off flash drive) do not mount a tmpfs on /tmp -- in the latter case, the 'init' script in the initrd.gz wipes /tmp, so again no need for that line in rc.shutdown.

However, PUPMODE=2, a full hd installation, does require /tmp to be wiped, and I have put that into /etc/rc.d/rc.sysinit -- though, I should probably think about putting a tmpfs for that as well.

Although this might seem like a small change, one never knows what odd repercussions there might be, hence I have announced it here and done a Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/cb5ec7d978

Booting from mmc (SD) card

There is some discussion about this on the Forum. I posted recently about recognising an mmc card at bootup:
http://bkhome.org/blog/?viewDetailed=02771

...Wary and Racy 5.3 have that fix.

There is something important that testers must understand here. My latest laptop has an SD card slot, but internally it is recognised as a USB device. Similarly, many people plug an SD card into an external USB card-reader adapter. These are no problem in Puppy, as the SD card behaves just like any other USB Flash drive.

However, many laptops (including my own older Acer laptop) have internal card readers that have their own special interface hardware. The most common type are what we call the "mmc" card readers. The kernel loads special device drivers for this. These are the type that we are having issues with at bootup.

Forum member riccardo started a Forum thread about the problem of booting from an SD card, using Wary 5.3:
http://murga-linux.com/puppy/viewtopic.php?t=77417

...I have posted some tests to that thread. If anyone else has the same problem, you are welcome to try these tests.

One-click editing of initrd.gz

This has been such a long time coming! I think that someone did develop a tool for editing an initrd.gz file, but anyway, I decided that it is about time this feature was built-in to Woof.

Just as we can click on any .sfs, .2fs or .3fs file in a ROX-Filer window, and have them opened up, now the same can be done with a 'initrd.gz' file.

initrd.gz is the initramfs, that gets loaded by the kernel. Developers often want to modify this file, perhaps to just quickly test a change. Or when remastering and want to make a permanent change.

Now, just click on a 'initrd.gz' file and it is recognised as mime-type 'application/initramfs-gz' and ROX-Filer will run /usr/sbin/edit-initramfs. The latter is a new script. The initrd.gz gets expanded at /root/initrd-expanded. After any modifications, clicking on the initrd.gz file again will update it. Dialog windows clearly explain each step.

This will work in all future Woof-built puppies, but for anyone who wants to implement it right now, you need to do these things:

1. Mime type
Append this line "application/initramfs-gz:initrd.gz" into /usr/share/mime/globs.

2. ROX-Filer
Get the file 'http://bkhome.org/fossil/woof2.cgi/finfo?name=woof-code/rootfs-skeleton/root/Choices/MIME-types/application_initramfs-gz'.

3. Handler script
Get the script 'http://bkhome.org/fossil/woof2.cgi/finfo?name=woof-code/rootfs-skeleton/usr/sbin/edit-initramfs'

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/5c05a1fe16

Note, it is supposed to automatically recognise gzip, bzip2 and xz compression of intrd.gz, however I have only tested gzip.

Mele A2000 Internet set top box

I posted recently about the Mele A1000, of which I have ordered two:
http://bkhome.org/blog/?viewDetailed=02754

The manufacturer has now released the A2000, which appears to have almost the same specifications. Details:
http://www.cnx-software.com/2012/04/04/mele-a2000-android-2-3-media-player-powered-by-allwinner-a10/

The A2000 can be ordered from here:
http://www.smalltao.com/product/14546035231/Hard%20disk%20player%20Zhuo%20Ying%20A2000

Woof commit 9d81d0919b

This is the version of Woof used to build Wary and Racy 5.3-final.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/9d81d0919b

The "last minute" changes to Woof before building Wary/Racy 5.3 was deciding on what version of the Flash Player to download (though in future that needs more work, such as a choice) and a fix for alsaconf.pot submitted by L18L.

Wary (and Racy) 5.3

Wary Puppy 5.2.2 was released on 18 November 2011 (also Racy Puppy made his debut), see my blog announcement:
http://bkhome.org/blog/?viewDetailed=02602

Since then, there has been the usual feverish pace of development. We now have a significant upgrade, hence the version jump from 5.2.x to 5.3. A brief announcement:

Wary Puppy is our build of Puppy Linux targeting older hardware. Racy Puppy debuted at the previous release (5.2.2) and is "Wary on steroids" -- identical except with later X.org and kernel, to suit more recent hardware. Version 5.3 has a fairly small number of application upgrades, however there are fundamental changes at the infrastructure level, that is, the "Woof" level -- in particular, pervasive support for internationalization, plus a "multitude" of bugfixes. There are now "langpacks" available for many languages, and these have translations for the system-level and most of the applications used in Wary and Racy.

For this release, I would like to send a special thank you to the faithful testers, and in particular the guys who volunteered to create the langpacks. The latter are (Format is language : Forum-name : real name):

French (fr): esmourguit : JJ Moulinier
German (de): L18L : Lutz-Ulrich
Polish (pl): robwoj44 : Robert Wojew�dzki
Russian (ru): rodin.s : Sergey Rodin
Spanish (es): vicmz : V�ctor Mart�nez

...bug reports or improvements for these translations should be directed to these guys on the Puppy Forum http://murga-linux.com/puppy

Here is the detailed announcement and release notes:
http://distro.ibiblio.org/quirky/wary-5.3/release-Wary-5.3.htm

Download Wary from here:
http://distro.ibiblio.org/quirky/wary-5.3/

Download Racy from here:
http://distro.ibiblio.org/quirky/racy-5.3/

Other faster download links can be found on the main download page:
http://puppylinux.com/download/

I have started a Forum thread for bug reports and any other feedback:
http://murga-linux.com/puppy/viewtopic.php?t=77303

"Old" ideas for a new lightweight pup

I would like to draw attention to what I think is a fascinating thread on the Puppy Forum:
http://www.murga-linux.com/puppy/viewtopic.php?t=51478

Goingnuts and technosaurus (and some more guys, see the thread) are investigating some very interesting ideas. There is so much "stuff" out there that has been left behind, yet some of it is really good, with much promise for building a small pup that runs fast on slow and memory-constrained hardware -- I am thinking of the fairly slow ARM CPUs for example.

I remember back when I dropped GTK1 from Puppy, I was sad about that, as apps ran noticeably faster than the GTK2 equivalents. Plus, GTK1 is much smaller. One of the main problems with GTK1 was the visually unappealing non-antialised font rendering, but there was one partially successful fix to use Xft that I used -- but, technosaurus has posted about this other old project:
http://sourceforge.net/projects/gtk-xcb/
...that also uses Xft, plus Pango.

To build a Puppy with GTK1, one thing that we need is gtkdialog. Well goingnuts has backported gtkdialog to GTK1:
http://murga-linux.com/puppy/viewtopic.php?t=75708

Anyone know other links to revitalized old GTK1 projects? Or even dead projects that should be revitalized? Any other relevant Forum threads?

Wary and Racy 5.3RC3

If anyone wants to give this a spin, you are welcome. If there are no disastrous bugs, this will become 5.3-final.

Download:
http://distro.ibiblio.org/quirky/wary-5.3/
http://distro.ibiblio.org/quirky/racy-5.3/

Regarding recent discussion about which Adobe Flash Player to use, I chose 10.3.183.18. However, I added a warning window that alternative versions may work better, depending on hardware and website.

Flash Player PETs

Thanks to Terryphi who created some PETs:
http://www.murga-linux.com/puppy/viewtopic.php?t=77257
http://www.murga-linux.com/puppy/viewtopic.php?p=616867#616867

I have uploaded Terryphi's PETs plus one other version here (5.1MB, 6.3MB, 6.4MB):
http://distro.ibiblio.org/quirky/pet_packages-common/flashplayer10-10.3.183.18.pet
http://distro.ibiblio.org/quirky/pet_packages-common/flashplayer11-11.1.102.63.pet
http://distro.ibiblio.org/quirky/pet_packages-common/flashplayer11-11.2.202.228.pet

Note, we also have these older versions. Interesting how much smaller they are (2.9MB, 3.9MB):
http://distro.ibiblio.org/quirky/pet_packages-common/flashplayer9_pretend10-20110327.pet
http://distro.ibiblio.org/quirky/pet_packages-common/flashplayer10-10.0.42.34.pet

Kernels with Aufs

I have compiled 2.6.32.59 for Wary and 3.0.25 for Racy, with Aufs. Both of these were compiled in a full hd installation of Wary 5.1.2.

For Woof builders, here are the PETs (28.8MB, 23.5MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/linux_kernel-2.6.32.59-uni-i486-aufs-w5.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/linux_kernel-3.0.25-pae-i686-aufs-w5.pet

I have uploaded the kernel source SFS files to the Wary/Racy SFS repo (81MB, 95.4MB):
http://distro.ibiblio.org/quirky/sfs/kernel_src-2.6.32.59-patched.sfs
http://distro.ibiblio.org/quirky/sfs/kernel_src-3.0.25-patched.sfs

Schedule for Wary/Racy 5.3

I was planning to release Wary/Racy 5.3 toward the end of this week. Then a serious bug in Unionfs emerged, see post yesterday:
http://bkhome.org/blog/?viewDetailed=02777

Aufs
So, I am going to change the kernels over to using Aufs. This means that there will have to be an RC3, but I am still going to target near-end-of-week to release 5.3-final.

There is going to be tester burnout if I drag this on much longer. So, 5.3 is coming very soon! I am even toying with the idea of taking a bigger risk and not releasing RC3, just test it myself, and go straight for the final.

After all, we have plenty puppies that use Aufs without much trouble. Wary, with 2.6.32.x kernel, will have to use an older version of Aufs, as the developers stopped supporting that kernel 8 months ago. However Racy will have the latest Aufs3, which does introduce the possibility of some new bug.

Note, I have already compiled these kernels and will upload the PETs and sources soon. Wary will have 2.6.32.59 and Racy will have 3.0.25.

Adobe Flash
A note about the Adobe Flash Player, Woof has code (in /usr/sbin/delayedrun) that offers to download the Flash Player after an Internet connection is detected.
Unfortunately, there are many reports of the latest Flash Player not working on older CPUs, even not-so-old CPUs.

So, I plan to create a PET package and that will be installed. I have feedback that 11.1.102.63 works ok -- is there already a PET of this version? -- does anyone know what URL to download this version from? I even wonder, there are some 10.x versions PETs in the repos, is 10.x too old?

Langpacks
If any of the translators would like to provide an updated langpack, today is the deadline, as I intend to build RC3 (which could become 5.3-final) several hours from now.
These are the ones I currently have:

langpack_de-20120325
langpack_es-20120328
langpack_fr-20120330
langpack_pl-20120327
langpack_ru-20120318

Getting Aufs3

Oh. Aufs2 is hosted in a git repository, with convenient links to download a tarball. However, the aufs3 git repository does not have tarball links.

Curses. My knowledge of git is extremely rudimentary. The Aufs site is extremely unhelpful:
http://aufs.sourceforge.net/

Well, I have done some research, this seems to work:

# git clone git://aufs.git.sourceforge.net/gitroot/aufs/aufs3-standalone.git 3.0

# cd 3.0

# git checkout origin/aufs3.0

And do this somewhere else, as it will create the same directory name:

# git clone git://aufs.git.sourceforge.net/gitroot/aufs/aufs-util.git 3.0

# cd 3.0

# git checkout origin/aufs3.0

Unionfs letting us down?

I used Unionfs in the kernels in Wary and Racy 5.2.2, and I thought that all was well.

However, Forum member vicmz reported a mysterious problem with the desktop drive icons. For example, with 'sda1' icon, click on it and sometimes an error message came up that /root/.pup_event/drive_sda1/AppRun is missing.

It took me several hours, but I eventually traced the original cause of the problem. /sbin/clean_desk_icons runs this line:

rm -rf /root/.pup_event/drive_* 2>/dev/null

However, a test line inserted immediately after;

ls -1 /root/.pup_event

...sometimes, lists the drive_* folders as still present. They are still present later, in the /sbin/pup_event_frontend_d script, and their presence, when they shouldn't be there, upsets the script very much.

Most of the time those folders do get deleted, just sometimes they do not. This is a gross failure, and the only thing that I can think of is Unionfs.

I am running Wary, with 2.6.32.55 uniprocessor kernel, booting off a USB stick. Vicmz has also reported the problem with Racy. I think that vicmz is also using a USB stick for the savefile.

I have been using Wary and Racy for many months without any problems, but most of that time it has been a frugal installation on hard drive.
The USB case would have the top layer of Unionfs in RAM, I don't know why that would make any difference.

Anyway, tonight I will recompile the kernels with Aufs.

Fix language change in initrd

Wary/Racy RC1 testers reported a problem with setting the locale. The user selected a locale in QuickSetup, but forgot to tick the UTF-8 checkbox. The user opted to change the language in the initrd, which means that early boot messages are in the user's language.

However, it also means that the LANG variable, for example "de_DE" carries over from the initrd to the main filesystem.

When the user later reran QuickSetup and ticked the UTF-8 checkbox, upon reboot it was gone, no UTF-8.

The LANG variable has to be updated in the initrd, in this example to "de_DE.UTF-8". I have fixed the scripts 'quicksetup' and 'lang2initrd' to do this.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/0c1a1fe45e

Woof commit 8974b31b2c

This is the Woof used to build Wary/Racy 5.2.91 (5.3RC2), plus a couple of small tweaks.

One of those tweaks was a bug when changed to a non-English locale, but the initrd is not translated -- rc.sysinit runs 'clear' to clear the screen after a switch_root, as the console font changes which messes up the display prior to the switch_root.
The 'clear' need redirection to /dev/console.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/8974b31b2c

Wary/Racy 5.2.91 (5.3RC2)

This is the announcement for 5.2.90 (5.3RC1):
http://bkhome.org/blog/?viewDetailed=02759

Download 5.3RC2 from here:
http://distro.ibiblio.org/quirky/test/wary-5.2.91-5.3RC2/
http://distro.ibiblio.org/quirky/test/racy-5.2.91-5.3RC2/

The only known issue for now is ffmpeg. It was reported that ffmpeg lacked network support, which is needed under certain circumstances by Pmusic, so I recompiled it and that was in 5.3RC1. However, then someone reported that ffmpeg did not work for X11-grab (although it was compiled with that enabled) -- I asked that person to install the older ffmpeg to confirm that X11-grab worked with that. However, that person did not respond in time, so I have reverted ffmpeg anyway. That is, 5.3RC2 has the older compile of ffmpeg.

Forum thread for feedback (page 14 of thread):
http://murga-linux.com/puppy/viewtopic.php?p=615981#615981

Everything has to be pretty much frozen at this point. Essential bug fixes only.
I know that language translation can be taken up a notch, however consider the current situation as frozen, until after 5.3-final.

Pmusic 2.5.2

I have updated to zigbert's latest Pmusic.

The PET (92KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/pmusic-2.5.2.pet

A quick note, while I think of it:
The RC1 release of Wary/Racy has Pmusic 2.5.0. I think that someone reported that music played in a continuous loop as default, which they reported as a fault -- and I asked them to report that direct to zigbert. I don't know the outcome of that.

Polish langpack

Robert, known as robwoj44 on the Puppy Forum, has created a Polish (pl) langpack for Puppy.

This is great, welcome Robert to the translation team. Now we have these guys:
German (de): L18L
French (fr): esmourguit
Polish (pl): robwoj44
Russian (ru): rodin.s
Spanish (es): vicmz

I have uploaded the pl langpack (1.9MB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/langpack_pl-20120327.pet

Robert, note that I inserted this code into your 'pinstall.sh' script:

#120315 maybe have hunspell dictionaries in langpack (see also momanager)...

for ONEHUN in `find ./usr/share/hunspell -mindepth 1 -maxdepth 1 -type f -name '*.dic' -o -name '*.aff' | tr '\n' ' '`

do

 HUNBASE="`basename $ONEHUN`"

 [ -e ./usr/lib/seamonkey ] && ln -snf ../../../share/hunspell/${HUNBASE} ./usr/lib/seamonkey/dictionaries/${HUNBASE}

 [ -e ./usr/lib/firefox ] && ln -snf ../../../share/hunspell/${HUNBASE} ./usr/lib/firefox/dictionaries/${HUNBASE}

done

L18L has provided an updated German langpack (2.4MB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/langpack_de-20120325.pet

mmc drive available at bootup

Jemimah reported recently a problem with mmc cards needing a bit more time at bootup to become available:
http://bkhome.org/blog/?viewDetailed=02767

Also, there is a Puppy Forum thread where someone is trying to save the session to a mmc card, booting off a CD. That thread proposes a solution involving a boot parameter "pdev", however, there is no such parameter and "pdev=..." will make no difference. More likely, it works because the person has a USB drive plugged in.

I have now taken into account the extra time needed by a mmc drive. In other words, it should now be fixed.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/a4a910b45f

Xdialog Yes/No buttons bug

L18L discovered this. The documentation for Xdialog states that the "--yesno" window will display buttons labeled "Yes" and "No". The proof is here:
http://xdialog.free.fr/doc/box.html

Perhaps this did work correctly with an earlier version of Xdialog? Anyway, now it doesn't, and the buttons display as "OK" and "Cancel".

This is where L18L reported the problem:
http://murga-linux.com/puppy/viewtopic.php?t=76182&start=105

I have fixed a bunch of scripts in /usr/sbin in Woof, haven't checked anywhere else.

These are gettext'ed scripts that will need to be updated:
alsaconf, burniso2cd, cups_shell, modemtest, pupdial, puppyinstaller, puppypdf, remove_builtin

I didn't notice this one before. It was created by 01micko I think, and had t12s internationalization. I have converted it to gettext:
nouveau_unload

These are not internationalized:
Pudd, remasterpup2, xgamma-gui

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/bcf91a7cc8

'rc.country' rewritten

I have completely rewritten /etc/rc.d/rc.country.

Rodin.s reported that the console font was not loaded for ru_UA locale, which got me to re-examine the relevant code. rc.country was not properly handling non-English locales, and it really needed to be rewritten, so I did.

Fixes were also required in quicksetup and xorgwizard-automatic scripts.

I am reluctant to be making such big changes when Wary/Racy are at RC1, but the fixes are required.

This Woof commit also has BootFlash fixes, for booting off a non-optical iso9660 filesystem, (init script), and booting from USB drive with separate partition for the save-file (shutdownconfig script).

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/aeb3a81a7a

mo files in Woof

Woof has some mo files in rootfs-skeleton/usr/share/locale.

As we now have the langpack system, there is no point in keeping these mo files, so I plan to delete them.

This is a notice to the langpack maintainers. If they are not in your langpack, would you please put them in.

It is also going to save space in a build, not having lots of mo files.

TODO shortlist

This is my current shortlist:

1. Precise Puppy: PPM initially has 'puppy-precise' radiobutton ticked, however the packages in 'ubuntu-precise-main' repo are displayed. This bug has been there for awhile (not just for Precise Puppy), I must finally fix it.

2. BootFlash. Apparently it is broken.

3. Kernel 3.2.13 does not have the usb-storage patch, that is normally required for usb drives detection at bootup. I need to fix drive detection without the patch, probably by monitoring uevents.

Precise Puppy alpha2

This is working great, using it right now!

Download from here:
http://distro.ibiblio.org/quirky/test/precise-5.2.52-5.3alpha2/

Previous announcement, for alpha1:
http://bkhome.org/blog/?viewDetailed=02761

There is one known issue:
ePDFView does not work, liblmcs2.so.2 is missing. To get this, run the Puppy Package Manager, choose the "ubuntu-main" repository, then type "liblcms2" into the search box.

Forum thread for feedback:
http://murga-linux.com/puppy/viewtopic.php?t=76961

3.2.13 kernel

I have compiled the 3.2.13 kernel, for both Precise and Wary.

As usual, the sources, including a SFS file, patches and build scripts, are at my sources repo:
http://bkhome.org/sources/kernel-3.2.13/

I compiled the kernel under Wary, but have not added the PET to the 'wary5' repo.

I compiled the kernel under Precise Puppy, here is the PET (for Woof developers only) (22.9MB):
http://distro.ibiblio.org/quirky/pet_packages-precise/linux_kernel-3.2.13-pae-i686-up.pet

My latest build of "Precise Puppy LTS" is working very well, thanks for the feedback on the first alpha. Another build with the 3.2.13 kernel is coming very soon.

Simple Network Setup fixed

This affects all Woof-built puppies, since the internationalization effort.

The script /usr/local/Simple_Network_Setup/sns has a syntax error introduced when it was internationalized. Causes a crash.

I didn't realise this, as it is not encountered for wireless connection, only for wired ethernet.

Anyway, fixed. You can download the fixed script:
http://bkhome.org/fossil/woof2.cgi/info/c1dfc7bd5b

Note, we have to expect this. It was a mammoth effort internationalizing all of the scripts, and the occasional typo is bound to happen.

'pupmessage' partial replacement for 'xmessage'

I have knocked up a very simple partial replacement for 'xmessage', named 'pupmessage', and converted some of the scripts in Woof to use it.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/cf8d967fcf

Multiple Sound Card Wizard improved

Forum member npierce has improved the Multiple Sound Card Wizard, for the situation where no cards are detected:
http://murga-linux.com/puppy/viewtopic.php?p=614133#614133

Note, this Wizard was originally written by Forum member kirk, who was the 'wizard' behind FatDog64 (x86_64 Puppy).

Thanks for the improvement to the Wizard. I have uploaded the PET (1.5KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/mscw-2.pet

Note, all architecture-independent PETs now live in the pet_packages-noarch folder.

There are some still in pet_packages-common, for historical reasons. Eventually I will clean that folder up, remove all architecture-independent PETs.

Note: "architecture independent" means PET has no binary executables.

Precise Puppy progress

I introduced Precise Puppy here:
http://bkhome.org/blog/?viewDetailed=02753

Well, I now have it booting up to the desktop, working remarkably well.

There were some issues, such as the xorg-server segfaulting, so I recompiled it, used version 1.11.4 instead of 1.11.3 used by Ubuntu.

Other issues were mainly due to Debian/Ubuntu habit of "messing up" packages by putting files into strange places. Then there is the habit of splitting packages into many small ones, so I have to hunt around and put them together again.

There are still some issues, some missing libraries, for example 'cupsd' is missing libgssapi.so.3. I will fix those next.

Gnome-mplayer segfaulted when I selected a video to play. In that situation, it is usually best to compile it ourselves.

I suppose that when I have worked on it a little bit more, I should build an ISO and upload it. There are probably quite a few people who would like to play with this.

As far as coordinating it to release status, I am interested, but worry whether I can give it sufficient time. Well, one step at a time -- I'll get the ISO out, and see what people think of it.

Woof commit df38d3f499

This is the version of Woof used to build Wary and Racy 5.2.90, which is 5.3 Release Candidate 1.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/df38d3f499

If you are interested in Woof, see my notes yesterday:
http://bkhome.org/blog/?viewDetailed=02753

There is a Woof intro page, though it is getting a bit out-of-date. For example, Woof is using the Fossil version control system now, not Bones. Also, Woof is now at version 2 -- see my blog for announcements on that. Intro page:
http://bkhome.org/woof/

Wary/Racy 5.2.90 (5.3RC)

I have been minding someone's place for a few days, and they have a fast Internet connection, with high monthly transfer allowance. They will be home this evening, so I decided to go for it and upload Release Candidate 1 of Wary and Racy 5.3.

You can download from here:
http://distro.ibiblio.org/quirky/test/wary-5.2.90-5.3RC/
http://distro.ibiblio.org/quirky/test/racy-5.2.90-5.3RC/

There are delta files to upgrade from the previous beta releases, also from the previous official 5.2.2 releases.

Relative to 5.2.2, there are many bug fixes and incremental improvements. There is one major structural change, and that is massive non-English support, in the form of "langpacks".

Langpacks are to be found in the Puppy Package Manager, in the "puppy-noarch" repository in the "Setup" category. Some of them are still being added to and should be ready by the time 5.3-final is released.

When I refer to non-English support as a "major structural change", I mean that it is a complete solution in Woof. It will even be possible to use Woof to build an "out of the box" variant of any Puppy for a specific language.

One thing that I think is a bit awkward, is the replacement of 'xmessage' with 'gxmessage', as the latter is required to display non-English text. However, in my opinion, the gxmessage windows look awful. I do intend to replace use of xmessage with pupdialog in many Woof scripts, but not sure how far I will get with that before 5.3-final.

There is no Release Notes page yet. To find out what has happened since 5.2.2, scan my blog posts.

Still the same old theme as used since version 5.2! It takes me a long time to come up with a new theme that I like, so I am hanging onto this one a bit longer. Probably the ARM-Puppy will get a new theme though.

ROX-Filer multiple-jwm-menus fix

To understand the bug, and the fix, read this forum thread:
http://murga-linux.com/puppy/viewtopic.php?t=72258&start=345

Argolance reported the problem, npierce created a patch to fix it, and argolance has reported that the fix works.

So, I have updated the PET, in the 'common' repo for all puppies to use (376KB):
http://distro.ibiblio.org/quirky/pet_packages-common/rox_filer-20111022-patched2-w5c.pet

The patched source is 'rox-filer-2011-10-22-patched2.tar.gz', uploaded to my source repo:
http://bkhome.org/sources/alphabetical/r/
(npierce's patch is there too)

X keyboard layout fixed

We were discussing a problem with Brazilian keyboard layout here:
http://bkhome.org/blog/?viewDetailed=02751

Forum member shinobar had this all figured out, and I have his 'keymap-set' utility in Woof. I call this from 'quicksetup' and it correctly modifies 'xorg.conf'.

However, 'xorgwizard' does not modify 'xorg.conf' correctly for the Brazil case. The problem crops up in Wary, and anytime that 'xorgwizard' is run.

So, I have modified 'xorgwizard' to call 'keymap-set --xopts' and then modify 'xorg.conf' keyboard layout, model, and variant correctly.

You can get /usr/sbin/xorgwizard from the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/b6fc849376

Notice that the Woof commit has some other fixes also, such as one for 'fido' from L18L, a fix for the 'pman' script, and improving the langpack creation by 'momanager'.

ffmpeg 20111002svn recompiled

There was a report, if I can recall it right, that ffmpeg does not work properly with Pmusic as network support is disabled.

Hmmm, there is a '--enable-network' configure option, but it seems to be the default. But, the 'ffserver' executable is not in the PET package.

Anyway, I have recompiled with that option, this time have put 'ffserver' into the PET. Interesting, the PET is a bit smaller this time.

PETs (2.5MB, 2.9MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/ffmpeg-20111002-1-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/ffmpeg_DEV-20111002-1-w5c.pet

These will be in the upcoming Wary/Racy 5.3 Release Candidate 1.

Pburn, Pmusic

Thanks to zigbert's tireless enthusiasm, we have updates for these:

Pburn 3.5.1
Pmusic 2.5.0

I have uploaded these to the 'noarch' repo.

Mele A1000 Set Top Box

I have ordered one of these:
http://www.dealextreme.com/p/mele-1080p-android-2-3-internet-tv-set-top-box-w-wifi-optical-3-x-usb-hdmi-av-lan-sd-119913

However, be warned, if this item interests you. I placed my order on March 7, and now it is the 17th and it still hasn't shipped. To read more about the slowness of dealextreme, type "dealextreme slow" into a search engine.

US$100 postage-free worldwide, superb value. There are other set-top-boxes that do something similar, but this one is special as it has the Allwinner A10 CPU (Cortex A8, with Mali GPU) and has been hacked so that we can install Linux on it (it comes with Android 2.3).

I have been following the Rhombus Tech email list with extreme interest, as they are planning an alternative to the Raspberry Pi based on the Allwinner chip:
http://rhombus-tech.net/allwinner_a10/

Their email list archive:
http://lists.phcomp.co.uk/pipermail/arm-netbook/

There is a web page that describes how to hack the Mele A1000 (more info in the email list):
http://rhombus-tech.net/allwinner_a10/hacking_the_mele_a1000/

About 14 of the guys volunteering under the Rhombus Tech umbrella have purchased these Mele A1000's to use as development system for their planned tiny RasPi-alternative board. One of the guys already has his Mele and wrote the above Linux-hack page. Another guy is waiting his, with a plan to port T2 to it (he is also planning a RasPi port of T2).
Note, those guys did not order from dealextreme, they had a special arrangement where someone in China physically obtained them from the manufacturer.

I did purchase a board awhile back, that uses the same CPU/GPU as the Raspberry Pi, for US$143, as reported on my blog:
http://bkhome.org/blog/?viewDetailed=02596
http://bkhome.org/blog/?viewDetailed=02605
http://bkhome.org/blog/?viewDetailed=02606

...I struggled with it for awhile, then I went on an extended camping holiday. After getting back home, I never got back to using that board, as I launched into internationalization of Puppy. It really was an uphill battle with not just the Chinese-only documentation but all of the software in Chinese. Now, I have no interest in it, and I consider the Mele to be far better as a development system for both the RasPi and the proposed Rhombus Tech board. So, if anyone wants it, give me a small donation say $50 and I'll post it to you -- contact me at bkaulerATgmailDOTcom.

I am thinking of using the Mele a1000 as the compiling platform for both RasPi and the proposed Rhombus Tech Allwinner-based board.

Here is an earlier blog post about the Rhombus Tech initiative:
http://bkhome.org/blog/?viewDetailed=02645

Some useful links about Linux and ARM:
http://www.linux-arm.org/Main/WebHome
http://www.arm.com/community/software-enablement/linux.php

Cortex A8 CPU used in the Allwinner A10 chip (in the Mele A1000):
http://www.arm.com/products/processors/cortex-a/cortex-a8.php

Mali 400 GPU used in the Allwinner A10 chip:
http://www.arm.com/products/multimedia/mali-graphics-hardware/mali-400-mp.php
Mali 400 open-source kernel, X11 and openGL software:
http://www.malideveloper.com/

Allwinner A10 development info:
http://www.cnx-software.com/2011/12/28/allwinner-a10a1x-processor-resources-development-board-and-sdk/

If you want to understand a bit more about ARMv6 versus ARMv7, this blog post is an excellent read:
http://wanderingcoder.net/2010/07/19/ought-arm/
...it seems that the two most significant features of ARMv7 are Thumb-2 and NEON.

Precise Puppy boots, no X

I mentioned yesterday that I was thinking of upgrading Woof to build from Ubuntu Precise Pangolin debs. I am mostly interested in an ARM build, but it follows that an x86 build gets a "free ride" (part of the way).

I posted about the problem of Multiarch:
http://bkhome.org/blog/?viewDetailed=02752

The problem is now solved in Woof, in the scripts '2createpackages' and '3builddistro'.

I have built Precise Puppy, did a frugal install and it boots, but only to the commandline as X fails. I think that the problem is the missing 'kbd' Xorg driver -- Puppy is setup to use it, however I think that these days we are not supposed to be using the 'mouse' and 'kbd' drivers. I know there was some discussion about this on the Forum awhile back, and some puplets have "moved on".

Anyway, that might be all that is preventing X from running. Maybe, maybe not.

I am not inclined to take on bringing an x86 (or x86_64) Precise Puppy to release status, it is just too much work, and I am going to get overloaded.

I wanted to update Woof to build a Precise Puppy, and I will probably do a bit more work on it, then will probably move on. Here is the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/e2e7ebd871

For those who don't know, Ubuntu Precise Pangolin LTS is due to be release April 26, with 5 years Long Term Support.

Ubuntu Multiarch

I have started to update Woof to build from Ubuntu Precise Pangolin binary packages, and have run into an awkward problem.

Libraries that were previously in /lib and /usr/lib, are now in /lib/i386-linux-gnu and /usr/lib/i386-linux-gnu.

Or, in the case of x86 64-bit libs, /lib/x86_64-linux-gnu and /usr/lib/x86_64-linux-gnu.

The rationale behind this is explained here:
https://wiki.ubuntu.com/MultiarchSpec

This change is not even a standard. It causes a lot of problems in Woof.

I would rather everything stayed in standard locations, so I am hacking it so that the libraries get placed in /lib and /usr/lib, with these symlinks:

# ln -s ./ /lib/i386-linux-gnu
# ln -s ./ /usr/lib/i386-linux-gnu

...that is, /lib/i386-linux-gnu is a link to /lib, and /usr/lib/i386-linux-gnu is a link to /usr/lib.

I think that the symlinks will fix some cases where Ubuntu packages expect stuff to be in /lib/i386-linux-gnu and /usr/lib/i386-linux-gnu.

I am still working on some details of this.

MoManager scoops up more translations

I posted about MoManager being able to build a bigger langpack:
http://bkhome.org/blog/?viewDetailed=02748

I also posted about non-English Hunspell dictionaries:
http://bkhome.org/blog/?viewDetailed=02750

I have further refined this support in MoManager. The relevant files, that translators may download if they wish, are /usr/sbin/momanager and /usr/share/doc/langpack-template/pinstall.sh, see Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/ffbc31d3e5

Non-English spell-checking

A message to all translators! (L18L, esmourguit, rodin.s and vicmz).

Puppy uses Hunspell for spell-checking, for most apps. Currently there is only one language-specific dictionary PET package:

http://distro.ibiblio.org/quirky/pet_packages-noarch/hunspell-en-us-2.pet

Please download this and open it up:

# tar -zxf hunspell-en-us-2.pet

This will give you the template for a non-English dictionary.

You can obtain a non-English dictionary by going here:

http://packages.ubuntu.com/

and do a search for package names "hunspell". You will find, for example "hunspell-de-de". Download it, right-click on it and choose to open with pupzip, and expand it. Then you can get the dictionary files out of it.

What I suggest though, is that we don't create 'hunspell-de-de-1.pet', instead put it into the langpack_de-.pet. That is, put the template and dictionaries into it.
We will then be keeping all required translations in the one langpack, which simplifies things for users.

Note that the template, that is, hunspell-en-us-2.pet, has a 'pinstall.sh' script, you will need to copy the code out of that and put it into the langpack's pinstall.sh -- actually, I should modify MoManager to have that already in pinstall.sh -- will do that now.

Roar-ng

Finally I am getting some time to have a bit of a play with Iguleder's fork/redesign of Woof, sufficiently different to be a separate project with it's own name, Roar-ng.

Here is the Forum thread:
http://murga-linux.com/puppy/viewtopic.php?t=75753

The project is hosted here:
https://github.com/iguleder/roar-ng

MoManager builds bigger langpack

I think it was vicmz who reported that some translation files were getting left out when a langpack PET was built in MoManager.

Yes, MoManager only puts in those translation files that it manages.

However, I have added the option to search for all mo, po, pot and t12s translation files in the system and put them all into the PET (only those of the target language of course).

This will also search for zigbert's translation files. So, if you had installed zigbert's Pburn, Pmusic, etc., translation PETs, the appropriate translations would also be put into the langpack.

This can be accumulative. In a new pristine Puppy, you can install your langpack, then put in any more translations that you can find, then rebuild the langpack via MoManager main window.

If the new code leaves out some language translations, let me know. Currently, it searches all of /usr for mo and po files that have "/de" somewhere in the path/name (using 'de' as the example), and all pot files, and t12s files named as de_[A-Z][A-Z]:*.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/6a17e54d18

QuickSetup, MoManager, PPM fixes

Various improvements and fixes. See Woof commits:
http://bkhome.org/fossil/woof2.cgi/info/19b2ca58ff
http://bkhome.org/fossil/woof2.cgi/info/0d7b69d187

if you want to download the scripts and put them into Racy 5.2.2.9, go for it.

Acrobat Reader SFS

Forum member sszindian reported that the Acrobat Reader (PDF reader) from the Lucid repo on ibblio.org works fine in Racy. It will probably work in Wary too, so I have uploaded it to the Wary/Racy SFS repo.

Or rather sszindian tested the PET on the Lucid repo, but they also have a SFS file, so I got the latter.

Here it is (61MB):
http://distro.ibiblio.org/quirky/sfs/Acrobat_Reader-9.3.3-Lucid.sfs

One area where you would want this rather than the ePDFView Poppler-based viewer, is for editable PDF files, those that have fields that can be filled in.

Mini-translation for QuickSetup

When someone chooses a non-English locale in QuickSetup, a window comes up advising about availability or otherwise of a langpack in the chosen language.

To improve the user experience, I would like that window to display in the user's language. So, I would like to obtain translations for as many languages as possible, for these:

AVAILMSG="There is a suitable \Zb${AVAILPACK}\ZB package in the \Zbnoarch\ZB PET repository, in the \ZbSetup\ZB category. After you have an Internet connection, click the \Zbinstall\ZB icon on the desktop to run the Puppy Package Manager, then choose the repository and category.

Note: Before downloading the langpack, you will have to make an Internet connection and use the Puppy Package Manager in English. If you think that you cannot read English sufficiently, an alternative is to run another operating system, go to http://distro.ibiblio.org/quirky/pet_packages-noarch and download the langpack. Then, boot Puppy, click the partition icon on the desktop to mount the partition (where you downloaded the langpack to), then click on the langpack PET to install it."



AVAILMSG="Official langpacks are available in the \Zbnoarch\ZB repository, however there is not one for your language. Perhaps there is one under development in the Puppy Forum (be careful, some older versions of Puppy have incompatible langpacks). Or, perhaps you might volunteer to help out and create a langpack -- click the Menu -> Help -> HOWTO Internationalization (you will of course need a good knowledge of English)."



pupdialog --background '#8080FF' --colors --backtitle "Language pack needed" --msgbox "The locale will be changed to ${LANGCHOICE}, but you need to install a language pack to translate Puppy to your language. You need a PET package named \Zblangpack_${LANG12}\ZB or \Zblangpack_${LANG1}\ZB. We have not yet automated this, so you will have to download and install the package. You may do this any time in the future. Officially supported langpacks are available by running the Puppy Package Manager (click the \Zbinstall\ZB icon on the desktop), in the \Zbnoarch\ZB repository.

${AVAILMSG}" 0 0

What I was thinking of doing is put a case statement in the script, to display the correct translation. Note, as the translations are being placed directly into the script, gettext is not needed.

So, for anyone reading this, if are fluent in another language, you are invited to respond!

Note, when you post a comment, click the 'code' button to get the code tags, then paste your translation in between the tags.

Precise Puppy planned, maybe

Just posting my thoughts, so that I don't clash with what someone else is thinking of doing.

I am thinking of building a Upup based on the upcoming Ubuntu Precise Pangolin 12.04 LTS release.

I have various reasons for this. One is that the Ubuntu ARM port is based on the ARMv7 architecture, and I will be needing a pup based on that architecture -- reason for that, I should be able to announce in a day or two.

Another reason is that this is the first Ubuntu where LTS means the desktop is to be supported for 5 years (previously it was 3 years). So, this could be a LTS pup also.

Precise Pangolin is scheduled for release on 26 April, supported until April 2017.

This website has a very interesting comparison of a board with ARM Cortex dual A9 CPUs, and various x86 CPUs, running Ubuntu 12.04 beta:
http://www.phoronix.com/scan.php?page=article&item=ubuntu_1204_armfeb&num=1

...notice in particular how well it performs against the Intel Atom CPU. Also, it was not measured by that comparison, but the ARM chip consumes considerably less power.

However, I must add that the above is only my intention at this stage. It could emerge that another ARMv7 distro will better suit my needs. I should be able to make a final decision about 3 - 4 weeks from now.

Racy 5.2.2.9 (5.3beta3)

If you have tested Racy 5.2.2.8, 5.2.2.9 does not bring much new, just a few bug fixes. Most of those fixes are to do with locale -- so the language translators are recommended to use 5.2.2.9.

For language translators, I still recommend do not upgrade from a previous save-file.

Download from here:
http://distro.ibiblio.org/quirky/test/racy-5.2.2.9-5.3beta3/

...note, there are delta files to upgrade the iso and sfs files from 5.2.2.7 or 5.2.2.8.

For language translators, there is a new script to translate, 'xdelta_gui'. Also, 'quicksetup' translation will need to be reviewed.

L18L has pointed out some possible limitations in creation of a langpack by MoManager -- some things left out that should really be included. I will have a look at that.

'xdelta_gui' fixed

/usr/sbin/xdelta_gui is a GUI for the delta difference file utility, available in the Utility menu.

The t12s conversion of this script has broken an if-then operation. However, rather than fix it, I have changed the script to a gettext conversion done by rodin.s.

MoManager does not manage t12s'ed scripts, so now xdelta-gui is being managed. xdelta_gui was the only t12s'ed script in Woof.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/8c805ffb85

de, fr langpacks

A quick note:

L18L, esmourguit, thanks very much, I have updated your langpacks.

BootFlash fix

Forum member bigpup reported that BootFlash (install Puppy to USB) utility is failing with Lupu. Bigpup provided this link:
http://www.murga-linux.com/puppy/viewtopic.php?t=76274

Yes, for some modes of installation, a 128M first partition is created, which is too small for some puppies (such as Lupu).

I have bumped it up to 256M, but haven't tested it. Here is the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/bfbf56bdde

Get the latest script from here (notice the Download button):
http://bkhome.org/fossil/woof2.cgi/artifact/62d099be35c5f5a53757c44cecf98c678aea0abd

New firmware for 3.x kernel

Thanks to pemasu who did the hard work here, finding new firmware needed for the 3.x kernel:
http://murga-linux.com/puppy/viewtopic.php?t=76182&start=45

I have put these into Woof, with consideration that Woof should still be able to build puppies with older kernels.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/353d46505f

Universal Installer fixed

It is amazing how these little bugs can be in there for years. Most users, including myself, never encountered these bugs, but they can potentially stuff up the install. npierce has isolated and fixed them:
http://murga-linux.com/puppy/viewtopic.php?t=76182&start=30

The problem occurs if you have more than 9 partitions in your hard drive. The Installer cannot distinguish between sda1 and sda10 for example. Now fixed. The script is /usr/sbin/puppyinstaller.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/625e796717

ROX-Filer language translation.

Thanks to L18L who posted here:
http://murga-linux.com/puppy/viewtopic.php?t=76182&start=30

Which linked to this post by LaRioja:
http://www.murga-linux.com/puppy/viewtopic.php?t=76281

I need a file ROX-Filer.pot to put into Woof, however the ROX-Filer source package does not have a .pot file!

However, it does have lots of .po files, ex de.po.
However, Puppy Forum member LaRioja has done some work on de.po, improved it.

However, i need a .pot file, to put here:

/usr/share/doc/nls/ROX-Filer/ROX-Filer.pot

...MoManager will find this and a translator can then create a .po hence .mo file.

I found some info how to convert po to pot:
http://www.commandlinefu.com/commands/view/10042/empty-a-gettext-po-file-or-po2pot

This line does it (plus a bit of editing by me):

# msgfilter --keep-header -i ROX-Filer.po -o ROX-Filer.pot awk -e '{}'

Here is the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/ecbf88309e

Parts for Raspberry Pi

It could be some time before I get my hands on a Raspberry Pi, but I thought that I might as well order some parts to be able to get it going.

I have purchased these:
http://www.oo.com.au/mbeat-7-Port-Powered-USB-Hub-_P21980.cfm (AU$19.95 + P&H)
http://www.oo.com.au/Micro-USB-5-Pin-to-USB-Cable-_P113935.cfm
http://www.oo.com.au/mbeat-Full-HD-1080p-HDMI-to-HD_P42027.cfm

The way I am intending to put it together is the USB hub will power the RasPi, via the micro-USB cable. The RasPi has a micro-USB socket that is only for 5V power. There is one other, and only one, normal USB socket, and I will connect that back to the input-USB socket on the hub.

There are cheaper powered USB hubs on the market, however as far as I could find out, the power adaptors are quite low current, for example two different 7-port hubs had 1A and 400mA 5V adaptors -- 400mA is ridiculous.

The one that I have purchased supplies 2A. It also has an Australian power plug (I think).

This gear adds up to more than the cost of the RasPi!

A note on the side: I was surprised that Farnell and RS had been chosen to sell the RasPi, as my memory of them 10+ years ago is that they are very expense. Everything has a huge markup. I looked around in the Farnell online catalog, and the big markups still seem to be the situation.

So why would those companies agree to sell the RasPi at low profit margin? Well, those guys ain't fools, and my guess is that they know that the RasPi is a barebones board, and people are going to want a lot of extra stuff to get them going.

Sylpheed_portable 3.1.3

I posted yesterday about compiling Cryptofs:
http://bkhome.org/blog/?viewDetailed=02734

And the day before that about the need for a "portable" email client:
http://bkhome.org/blog/?viewDetailed=02732

Well, I was up to 3am working on it, but the sylpheed_portable PET has arrived (1MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/sylpheed_portable-3.1.3-cryptofs0.6.0-bogofilter1.2.2-w5c.pet

To make it completely self-sufficient, requiring no dependencies, it has Cryptofs and Bogofilter inside it. The only deps are 'fuse' and 'sqlite3', that all puppies have. I haven't tested Bogofilter.

When you install this PET, it installs to /usr/local/sylpheed_portable, but doesn't actually run from there. When you run it, it asks to be copied to a partition, from whence it can be run into the future. Any partition, USB Flash drives also.

The reason that it installs first into /usr/local/sylpheed_portable is so that Woof-built-puppies can have it in their package list. It will be there, looking like any normal Sylpheed, runnable via the menu or the 'email' desktop icon, but the first time you run it, it gets relocated wherever you want it.

When it is finally installed in a mounted partition, it is just a RoxApp that you click to run. All configuration files and downloaded email are encrypted, and when you run Sylpheed you are asked for a password. Everything gets decrypted for the duration that you are using Sylpheed, then locked up again when you quit Sylpheed.

I'm using it with all the emails that I downloaded from Gmail, all 6,500 of them. Strangely, it seems faster accessing them via cryptofs than before.

Can we give the same "portable" treatment to Thunderbird/SeaMonkey-Mail? I don't know, will have to experiment -- sometime, don't know when.

Today will take some time off for family stuff -- someone's birthday, plus the old folks like to be visited.

cryptofs 0.6.0

I posted yesterday about creating a portable Sylpheed. But, the issue was raised that the email login passwords are stored in plain text.

Well, so are all the emails. I was thinking, what if someone stole my laptop? This got me thinking that I could implement an encrypted filesystem inside the Sylpheed installation.

Toward that end, I have compiled Cryptofs, which can encrypt an entire directory.

Here is the PET (230KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/cryptofs-0.6.0-w5c.pet

The only dependency is libfuse, which is in all puppies. Cryptofs does require libgpg-error and libgcrypt, however I linked those statically into the 'cryptofs' executable.

Cryptofs home:
http://reboot78.re.funpic.de/cryptofs/

French language pack

This is great, Forum member esmourguit has created a language-pack for Puppy. This is a large PET, with all the translations you will need for the 'fr' locale.

The language pack is created using MoManager, plus other translations found, mostly via the Forum. It is designed to work with any Puppy built from Woof after Feb. 27, 2012. We don't actually have any such puppies officially released yet, only alpha/betas.

I would like to congratulate esmourguit on this, it is not a small task. Translating so many files takes a long time, and the scripts need particular care so as not to break their syntax. I think that esmourguit, L18L (de) and rodin.s (ru) are three of our many unsung heroes.

Here is the PET (1.5MB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/langpack_fr-20120301.pet

Please note that I intend to update Woof in a day or two, with an enhancement for QuickSetup such that when someone chooses a non-English locale, a check is made automatically about availability of a suitable langpack, and some kind of offer is made to download it.
I will also build a Wary/Racy to test it.

Truly portable Sylpheed

I have created a truly portable Sylpheed. That is, I have a directory on a hard drive partition, named appropriately 'sylpheed', and it has everything, including all configuration files and mailboxes.

No symlinks are required in /root, it is entirely independent, and the 'sylpheed' directory can be copied or moved anywhere.

I have set it up as a RoxApp, so just click on the 'sylpheed' directory to start Sylpheed.

I think that this could be put together as a PET if there is any interest.

It solves my problem of managing my emails independently of whatever Puppy I am currently running. I reckon that I should incorporate x86 and ARM executables into it and automatically run the correct one -- thinking ahead.

Woof: 'find_cat.c"

In Woof, there is a binary executable support/find_cat. This is written in Genie, source is support/find_cat.gs.

Forum member noryb009, who has made a lot of contributions to Woof to support building from Arch Linux packages, has rewritten find_cat in C, and it is 4 times faster than the Genie version.

Woof has a new file, support/find_cat.c, and the binary woof-arch/x86/build/support/find_cat has been updated.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/6fa6394179

Sylpheed versus SM Mail

What a difference! It is comparing something created by one or two people with something created by a committee.

I have been using online (cloud) email with Gmail for so many years and have not used a local email client in all that time. I have had a wake-up call about the dangers of the cloud, see my previous post:
http://bkhome.org/blog/?viewDetailed=02729

I have been complacent, have too much stuff in my online email account. Which of course Gmail encourages.

I have decided to POP it all. First I setup SeaMonkey Mail and News. What convoluted logic! -- I was unable to move the 'inbox' to my mounted partition.

Enter Sylpheed. Easy to setup my mounted partition as the mailbox -- simple, logical, like it should be. Then I was able to import the emails that I had already downloaded with SM (about 500). Now I am downloading the rest.

But, as Lobster says, with us changing puppies so often, we need somehow to automate the setup of the mail client. Some kind of "remember all of this" which creates a script on the mounted partition -- then when running a new pup all you have to do is click on the script. It could even install Sylpheed if it isn't already!
Possible?

Email account hijacked

Ha ha, it seems to be in my "stars" that this is my time for mini-crises

There was the failure of my wi-fi mobile broadband module a couple of days ago, then my apparent loss of $150.

This morning I decided to checkout an email account that I haven't used for a long time, puppylinux@gmail.com, but I was unable to login. It informed me that I was using an old password that had been changed 5 months ago. Fortunately I was able to go through the recovery process and change the password.

But, after logging in, everything was in Russian!!!

This was a bad situation, as I could not even figure out how to logout. The thing is, I cannot login to my regular gmail account while the other one is still open. I figured out how to get the Account Management page in English, then deleted the account.

I don't know how anyone could have got the password. was Google itself compromised?

One lesson from this, I need a backup email account that is not with gmail.com.

In fact, this whole online email thing is a worry.

Reloading console font

L18L has been investigating this. It seems that we need to reload the console font every time there is an exit from X. This never bothered us before, as we were only using the kernel's inbuilt font -- and 'en' locale continues to use it. This issue affects those non_English locales that are using a different console font.

I think that I have got it all figured out. See latest Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/7d52a93d7b

The scripts that pertain to this console font issue are:

rc.sysinit, xorgwizard, xwin, quicksetup, load_consolefont

That last one, /sbin/load_consolefont, is a new script that is now called by a few other scripts.

There is bit of a cosmetic problem when booting. If the initrd.gz has not been converted to the non-English locale, then rc.sysinit changes font and the earlier o/p made by the 'init' script changes into partial garbage. It might be necessary to detect that situation in rc.sysinit and clear the screen before loading the font.

Raspberry Pi launched

Ha ha, I received an email that the Raspberry Pi board had been released, so went to their site about 10pm last night (Western Australian time, 8 hours off GMT).

They are selling the boards through Farnell and RS Components. The Farnell site was down, the RS site was only taking "registrations". So, I registered my interest.

It seems that all 10,000 board have already sold. The next lot probably won't be available until the end of next week.

Sites:
http://www.raspberrypi.org/
http://uk.rs-online.com/web/generalDisplay.html?id=raspberrypi&cm_mmc=UK-PPC-0212-_-02_Raspberry_PI-_-Raspberry_PI-_-Raspberry_Pi
http://www.element14.com/community/groups/raspberry-pi

Bug fixes

I have been off-line all day and haven't had a chance to go to the Forum to pm these guys. I have found and fixed a couple of bugs, and while I remember, before I go to bed:

rodin.s,
This applies to all translators, a simple mistake when using gettext. Script /lib/modules/all-firmware/pgprs/pgprs-setup was broken. There were two lines, for example:

xmessage -center -title "$(gettext 'PGPRS SETUP')" "$(gettext \"Connect to the Internet using 'PGPRS Connect' in the 'Network' menu.\")"

...you do not back-slash the " around the gettext string, it causes a syntax error. Corrected:

xmessage -center -title "$(gettext 'PGPRS SETUP')" "$(gettext "Connect to the Internet using 'PGPRS Connect' in the 'Network' menu.")"

This was something that I pointed out in a recent blog post, a difference compared with using "`gettext ..... `"

L18L,
A small thing that you need to fix in your langpack_de. /usr/local/apps/Connect/AppInfo.xml is broken. You have put in a lot of which should be tag ends.

Wi-Fi crisis today

This morning my Telstra Mobile Wi-Fi module stopped working. I took it to a Telstra Shop and they asked me when I bought it, which I could not recall. Well, here is my blog post:
http://bkhome.org/blog/?viewDetailed=02121

...ha ha ha! I bought it on Feb. 1, 2011, and it has a one year warranty -- the designers have sure done a good job there!

The big problem for me is that yesterday I spent $150 on a prepaid recharge, which I activated last night.

I have a Telstra USB wireless modem that someone gave me a year or so ago. So, I put the sim into that, but unfortunately PupDial reports:

AT+CGDCONT=1,"IP","telstra.internet"
ERROR
--> bad init string

Note, "telstra.internet" is my guess at the APN, which I think worked with this same modem when I first acquired it. Dialup number I set to "*99#".

For the record, the chip in this modem is ZTE MF626i. It switches to modem-usage ok and works on /dev/ttyUSB0.

Anyway, I also have an Optus USB wireless modem, that I have reported about previously on this blog. I haven't used it for a long time, so the Optus Shop told me that I had to buy a new sim card -- for $30, which I did. Dialup number "*99#" and APN "preconnect", and it works.

I am using it now. I went to the designated Optus site to activate it, and it told me that I have $0.25 credit, which was a bit disconcerting considering that I am supposed to have 1GB for 30 days. Anyway, it is working. But, more expenditure, as I will have to go back to the Optus Shop in a few days and purchase a recharge.

I have got to try and get that Telstra modem working, otherwise I will have to throw away $150.

Langpack for 'de'

L18L has posted a German langpack, though be advised that it is a work-in-progress.

I have uploaded the PET to the 'noarch' repo (79KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/langpack_de-20120228.pet

The only change I made to the PET was I deleted /usr/share/sss/doc_strings/doc_strings.de. A note about this for all translators:

/usr/share/sss/doc_strings/doc_strings. is only an identical copy of /usr/share/sss/doc_strings/doc_strings. At least for now.
L18L had an older doc_strings.de which had less documentation files in it.

I have modified MoManager so that it will automatically update /usr/share/sss/doc_strings/doc_strings., but for now if you just delete /usr/share/sss/doc_strings/doc_strings. and then run MoManager, it will also get updated.

glew, llvm, clang, libxml++, xulrunner

I have compiled these in Wary and created PETs:

glew-1.7.0 OpenGL Extension Wrangler Library
llvm-2.8 Low Level Virtual Machine
clang-2.8 C/C++ compiler
libxml++-2.35.1 C++ interface to XML files
xulrunner-10.0.2 Embed Mozilla technologies into apps

http://glew.sourceforge.net/
http://llvm.org/
http://clang.llvm.org/
http://libxmlplusplus.sourceforge.net/
https://developer.mozilla.org/en/XULRunner

The PETs:
http://distro.ibiblio.org/quirky/pet_packages-wary5/

Wary has all the dependencies, except libxml++ needs glibmm.

Now that might seem like a strange combination of PETs to create! I was compiling the dependencies for Lightspark, an open source Flash player. Then I found that Lightspark needs gcc 4.4+, whereas Wary has 4.3.4 -- so I compiled 4.4.6. However, Lightspark still needs Boost -- I installed libboost DEBs from Debian Lenny, but Lightspark compile failed, doesn't seem to like that version of Boost.

Lucid will lose "official" status

Playdayz, the coordinator of Lucid Puppy, and I have agreed that by the time of the next official releases of Racy, Wary and Slacko, Lucid will be dropped from the list of "official" puppies.

Lucid is based on Ubuntu Lucid Lynx binary packages and is getting on to two years old. Even though there have been upgrades, the time has come to "de-emphasise" Lucid. However, Lucid will likely continue as another puplet, for those interested.

The idea of "official" is those puppies that we list in puppylinux.com as the first-stop for newcomers. And of course they get announced on distrowatch.com. But of course there are many other puplets, and even separate projects that promote themselves separately, such as on Distrowatch.

As you guys know who have followed Lucid development, playdayz (Larry) has done an incredible job, and has "hung in there" for a long time. Of course there are many others who have assisted.

Set hostname in QuickSetup

Gcmartin will be pleased with this! QuickSetup now has an entry box to change the hostname, but it will only display if there is an automatic Internet connection at first boot.

In the case of wired Ethernet, /etc/rc.d/rc.sysinit may determine that 'eth0' is already available and can connect to the Internet/network, so it is done automatically at the very first (PUPMODE=5 or PUPMODE=2) boot.

I do not want a hostname-entry box in QuickSetup, the more appropriate place for it is in the Internet Connection Wizard, however for the above specific situation in which the ICW is not run, then it would be appropriate to set the hostname in QuickSetup.

Woof commit creating new 'check_internet' script:
http://bkhome.org/fossil/woof2.cgi/info/a356cc92a7

Woof commit, hostname-entry-box in QuickSetup:
http://bkhome.org/fossil/woof2.cgi/info/1fc00656bd

Racy 5.2.2.8 (5.3beta2)

Wary and Racy 5.2.2.7 (5.3beta1) was only released a few days ago:
http://bkhome.org/blog/?viewDetailed=02705

5.2.2.8 is recommended for language translators. It has some important bug fixes and enhancements for better and more complete translation.
Run MoManager and review translations, also you might find more to do in the 'doc_strings' category.

Get 5.2.2.8 from here:
http://distro.ibiblio.org/quirky/test/racy-5.2.2.8-5.3beta2/

We have three guys who I would like to acknowledge, who have completed or almost completed, "langpacks":

de L18L
fr esmourguit
ru rodin.s

I am expecting that when Wary/Racy beta3 arrives at least two of these PETs will be available and ready for use. I intend to do some work on making it particularly easy to install these.

Latest Woof commits uploaded

Here they are:

[d86fd715a0] Leaf: typo in rc.update. just about to build racy 5.2.2.8
[ef37a66ebc] 01micko: /usr/sbin/bcrypt_gui converted to gtkdialog4 (0.8.0+)
[c6d2373191] 01micko: sound tweaks, /usr/sbin/alsawizard, /etc/init.d/10alsa
[0aaf69b981] 01micko: bugfix for xorgwizard, call new script nouveau_unload
[ba9e4830d4] MoManager bugfixes, fixed PuppyPin going back to en when layered f.s. changes
[51cd381c2f] MoManager now manages all of woof doc files, in sss doc_strings

Online, 2012-02-25 and 2012-02-26:
http://bkhome.org/fossil/woof2.cgi/timeline

VLC 2.0.0 SFS

I announced recently that Billtoo has compiled Thunderbird and created an SFS for Wary and Racy. Now Billtoo has added VLC 2.0.0 (video and audio media player).

Right now I am uploading the SFS to ibiblio.org:
http://distro.ibiblio.org/quirky/sfs/

This is where Billtoo posted to the Forum:
http://murga-linux.com/puppy/viewtopic.php?t=76182&start=15

MoManager: documentation files

Continuing to enhance MoManager to be a one-stop-shop for language translation of everything in Puppy, I have added management of documentation files.

Forum member esmourguit, who is working on a 'fr' langpack, was wondering how to handle /usr/local/petget/help.htm. This is a help file for Puppy Package Manager.

An application usually has help or other doc files with naming to identify the language, for example 'help.htm', 'help-de.htm', 'help-fr.htm', etc.

So, I have modified /usr/local/petget/ui_Classic and ui_Ziggy to use a translation file if it exists, for example 'help-fr.htm'. In fact, I have put 'help-fr.htm' into Woof as esmourguit has already created it and it is a good example.

Translation of such documentation files is now handled by MoManager via /usr/share/sss/doc_strings/doc_strings -- the easiest way to understand it would be to look at that file.

The end result is that it is now in MoManager, you choose 'doc_strings' and click 'Edit' or 'Create', and an appropriate editor will open with the file to be translated -- in the case of HTML files, the 'defaulthtmleditor' is used, which is usually Composer.

Currently only the PPM help file is being managed by MoManager. I will have to identify other help files and add them to /usr/share/sss/doc_strings/doc_strings.

To see the details, and grab the latest files if you want, here is the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/49d5144356

PPM Ziggy interface fixed

Forum member esmourguit reported that the Ziggy user interface of the Puppy Package Manager, does not fully translate to another language, also a couple of icons are missing.

The file is /usr/local/petget/ui_Ziggy, and it is fixed. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/3df37e36e2

Also in this commit is a fix for /usr/sbin/quicksetup to load a console font when the locale is changed. This should fix a problem reported of wrong font after exiting from X.

'sfsget' broken for Racy

Testing Racy 5.2.2.7, L18L has reported that the SFS downloader (via 'install' icon on desktop) is broken:
http://murga-linux.com/puppy/viewtopic.php?t=76182&start=15

Hmmm, I found that it has always been broken for Racy. Fixed.

If you want the fixed /usr/sbin/sfsget, here is the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/03856c3ba7

Retrovol 0.12

This is the latest version of pizzasgood's superb sound mixer and tray applet:
http://www.murga-linux.com/puppy/viewtopic.php?t=50744

I have uploaded the PET to the 'common' repo (35KB):
http://distro.ibiblio.org/quirky/pet_packages-common/retrovol-0.12.pet

Note, I used the 'retrovol' executable compiled by pizzasgood for Puppy 5.x, but I don't know which build exactly that he compiled it in. It should work fine in all our 5.x pups.
I have kept /root/.retrovolrc same as before (0.10 PET).

Unionfs versus Aufs

The old battle between these two is continuing. Some brief notes here.

Aufs currently has a problem with "udba=reval" in latest kernels (and more than one CPU), which is what we use in /usr/local/petget/installpkg.sh when installing packages. The "problem" is that Aufs hangs sometimes. This may be resolved soon.
http://sourceforge.net/mailarchive/forum.php?thread_name=16084.1329797367%40jrobl&forum_name=aufs-users

Unionfs has now reached version 2.5.11 and works on the latest kernels.
http://download.filesystems.org/unionfs/stable/

Fixed desktop drive icons redrawing

I reported this as number 2 on my to-do list:
http://bkhome.org/blog/?viewDetailed=02706

The bug was introduced with the internationalization of /sbin/clean_desk_icons. It causes the desktop drive icons to be wiped and then redrawn every time X starts. This imposes an unnecessary workload on the CPU just at a time when it is needed for many other things.

I fixed it, also removed a dependency on a matching language translation:
http://bkhome.org/fossil/woof2.cgi/info/0e35f0e27e

Sound settings not remembered after reboot

This was something dumb that I did late in 2011. Thanks to broomdodger for bringing this to my attention. Fixed.

if you go here and do an anonymous login, you can download the latest /etc/init.d/10alsa:
http://bkhome.org/fossil/woof2.cgi/finfo?name=woof-code/rootfs-skeleton/etc/init.d/10alsa

Check-dependencies bug fix

Forum member npierce is an extremely helpful guy! We have had a bug in the check-dependencies script for a long time -- ever since I wrote it. You will find this in the menu at "Setup -> Check dependencies installed pkgs", and the script is /usr/local/petget/check_deps.sh.

npierce has fixed it here (thanks to bigpup for drawing this to my attention):
http://www.murga-linux.com/puppy/viewtopic.php?t=75071

I have put the fixes into Woof. There are two, the problem of too many installed packages, and a bug if a package has a "_" in it's name.

npierce did identify a potential third bug but I did not apply the fix as it cannot be assumed that shared library files have their executable flag set.

Here is the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/bde8c7a2e5

Early-boot splash help screens

If booting Wary/Racy 5.2.2.7 from CD, you will see that the help splash screen now has this entry:

plang= Locale -- not normally required as asked after bootup.

To override the default locale (usually en_US).

Now, I have added this into the Advanced help splash screen:

underdog=sda9 Mount an entire Linux distro (in sda9) underneath Puppy.

This is the "underdog" feature, that has made a return to Puppy:
http://bkhome.org/blog/?viewDetailed=02668

gettext_translator 0.9

gettext_translator is another po editor. It uses gtkmm libs, which I compiled statically, so the final 'translator' executable is 2.7MB stripped uncompressed, a trifle on the large size.

I have uploaded the executable, for anyone who would like to play with it (715KB):
http://distro.ibiblio.org/quirky/test/translator.gz

I haven't put it into the Wary/Racy 'devx' SFS, as it won't start a po file from the commandline -- that would be a good project for someone!

Compiling it statically with gtkmm was a painful experience, if anyone wants to do it, contact me and I will send you the steps that I went through. I was wise enough to record them.

gettext_translator appears to be a dead project. I got it from here:
http://www.sourcefiles.org/Editors/GNOME/

Slacko 5.3.3 RC2

01micko has released another Release Candidate for Slacko 5.3.3:
http://murga-linux.com/puppy/viewtopic.php?t=75209

Mick has very little time these days, as he has become a full-time college student. However, he will attempt to get this pup to a final release!

One of the constraints is that he is staying with Woof of January 25, 2012, as since then Woof has been in heavy transition with the internationalization. So, you have to bear in mind that all of my blog posts about Woof after Jan. 25 do not apply to Slacko.

Of course, if our very latest builds of Wary and Racy, announced in my blog today, prove to be very stable and reliable, then I wonder if Mick will be tempted?... but, that is always the quandary, you are at RC stage, you shouldn't change things much, otherwise you find yourself in another series of testing/fixing cycles.

To-do asap

There are a couple of things that I want to work on ASAP.

1. langpacks
When you boot Wary/Racy 5.2.2.7 for the first time, you get the QuickSetup window and can choose a locale -- if you choose a non-English locale, there is a window that advises to install a "langpack" for your language, if one is available.

However, it is a bit vague. Doesn't tell you if a langpack is available, nor exactly where to get it from. I want to fix that, maybe even ask again if you want to install the langpack when you have established an Internet connection.
Note, so far we only have a Russian langpack.

For anyone else interested in creating a langpack, we have MoMananager -- please read the posts in my blog. Shinobar has mentioned an existing collection of translations, which would be a useful resource for building a Woof-compatible langpack.

2. Desktop drive icons
I will need to check this, but I have observed that it seems the desktop drive icons are getting redrawn every time X starts. This should not be happening, only when a change in the drives is detected. The drive icons are specified in /root/Choices/ROX-Filer/PuppyPin, and if no drive-change then the icons should appear virtually instantly on the desktop, just like the other desktop icons.

Wary/Racy 5.2.2.7 (5.3beta1)

It seems that the last couple of alpha builds got out some bugs, mostly to do with internationalization changes in Woof, so I am going for beta1 builds.

The Wary build is here:
http://distro.ibiblio.org/quirky/test/wary-5.2.2.7-5.3beta1/

The Racy build is here:
http://distro.ibiblio.org/quirky/test/racy-5.2.2.7-5.3beta1/

If you are not familiar with what has been going on, that is, what has changed since earlier pup releases (Slacko, Wary, Racy), please browse down through my blog.

You will notice that the Wary ISO is very big. This is due to all of the analog modem drivers -- in Wary 5.2.2 I left out the Intel 53x modem drivers due to their enormous size, this time left them in. So, this time it really is the "kitchen sink" included, more SCSI drivers too.

Here is the Woof commit, dated 2012-02-21:
http://bkhome.org/fossil/woof2.cgi/timeline

...note the change in how 'xwin' is executed. Karl Godt raised this issue in the Puppy Forum and I took another look at it. Here is a relevant earlier blog post:
http://bkhome.org/blog/?viewDetailed=02381

Regarding the problem of how to launch 'xwin', this is a very ancient problem right from the early days of Puppy, and I think exposes a fundamental problem with Busybox and it's lack of support for runlevels in /etc/inittab. Unfortunately, the Busybox developers never provided an alternative to the runlevels. That is, there is no way to run 'xwin' after /etc/profile and /root/.profile (or .bashrc) have run (they are only supposed to set environment variables, not run executables).

Pmusic 2.4.4

I have updated to the latest from zigbert:

pmusic 2.4.4
pmusic_tray 0.1

...the latter is a binary executable that was previously inside pmusic PET.

I have also uploaded streamripper, optionally required by Pmusic, version 1.64.6.

PETs (3KB, 55KB, 91KB):
http://distro.ibiblio.org/quirky/pet_packages-common/pmusic_tray-0.1-x86_32.pet
http://distro.ibiblio.org/quirky/pet_packages-common/streamripper-1.64.6-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-noarch/pmusic-2.4.4.pet

Aaaah...

We get lots of Puppy success stories, but I particularly liked this email that I received recently, from "Beau". I asked his permission to put it on my blog:

I live in south-west Victoria and I'm a keen young computer user. I have been using Puppy for quite a few years now, and may I say how excellent it is!

Anyway, the main purpose of this email was to share with you a little story. I realise you must receive a lot of emails like this, but I'm so appreciative for Puppy that I felt compelled to write to you.

Back in 2008, a severe storm hit our region. Our family's computer (containing many years worth of memories which hadn't been backed up) failed to work afterwards. We feared that all was lost.
Even taking the system into our local Harvey Norman store, we were told that the entire system was corrupt and unless we wanted to spend $20,000 and send the system to a specialist company in Canada, there was no way that we'd get our files back. We decided that there was no chance of us recovering our files and we purchased a new system.

Fast forward to 2011. I received a very old computer from school. They said that it was "stuffed" and that I could do what I wanted with it. So, just to see what would happen, I plugged in the hard drive containing our files. I then made a "live USB" of Puppy (version 4.x, I believe) and booted it up. It recognised the hard drive, so I mounted it expecting to see an error message. How wrong I was. Within seconds I was looking at the contents of the hard drive - pictures of me as a toddler, pictures of my first day of primary school, a picture of my very first piano lesson from more than ten years ago!

I called my parents into the room and to say that they were "happy" would be a gross understatement. They were initially skeptical of Linux but after showing them what was possible with not much at all, they have become converts! And all of this thanks to an old PC, a USB stick and the magic that is Puppy Linux.

'new2dir', 'download_file'

Jemimah proposed a fix for /usr/bin/new2dir so that it will work with cmake:
http://murga-linux.com/puppy/viewtopic.php?t=71767&start=420

...there is uncertainty if this will break anything else, but I guess we will find out!

Jemimah has improved the user interface for /usr/sbin/download_file:
http://murga-linux.com/puppy/viewtopic.php?t=71767&start=420

I have put both of these into Woof:
http://bkhome.org/fossil/woof2.cgi/info/70753efeee
http://bkhome.org/fossil/woof2.cgi/info/fb69ed779d

'filemnt' bug fix

Ah, it is great to see another long-standing bug squashed! With Puppy you can click on a SFS or .2fs or .3fs file to mount it, but if another file with the same basename is already mounted, then you can't mount the second one. That might seem like a "corner case", but it does come up with SFS files sometimes.

Forum member npierce has fixed it:
http://murga-linux.com/puppy/viewtopic.php?t=71767&start=450

I applied the patch manually, as the latest /usr/sbin/filmnt has internationalization changes.
Here is the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/d8d8b428eb

Thunderbird SFS

Forum member Billtoo has created a SFS of Thunderbird email client, compiled in Racy, tested ok in Wary, should work ok in other puppies too.

I have uploaded it (18.4MB):
http://distro.ibiblio.org/quirky/sfs/thunderbird-10.0.2-racy.sfs

Russian langpack PET

rodin.s has completed the langpack for 'ru', which I have uploaded (1MB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/langpack_ru-20120219.pet

This is a good example for other language translators to study, as rodin.s has put in many more translations than those managed by MoManager. There are many applications in Puppy that have translation files that can be found by searching on the Puppy Forum.

Of course, this langpack can be installed after someone has booted Puppy and connected to the Internet, which means of course that they would have to use English-language Puppy until then.

However, the great news is that with this PET, someone using Woof will be able to create a Russian Puppy, that starts up first time in Russian.

There is already a variant of Puppy in Russian, known as PuppyRus, that was started in 2007 I think:
http://uco.puppyrus.org/
http://puppyrus.forum24.ru/
http://www.murga-linux.com/puppy/viewtopic.php?t=66909

Now, the PuppyRus team have the opportunity, if they want, of building a Russian Puppy from latest Woof, and always from now on being able to sync with latest Woof.

If anyone uses the langpack_ru and finds a problem with any translation, rodin.s is the maintainer, so please send a pm on the Puppy Forum.

ROX-Filer focus problem fixed

The problem with icons (files, directories) remaining selected after an operation has been performed, such as an "Open With...", has been a big annoyance for us. This problem has occurred ever since GTK >= 2.18.0.

The workaround employed in some Puppy builds of using GDK_NATIVE_WINDOWS=true is a very bad solution, as it can cause ROX-Filer to crash. As I have documented here:
http://murga-linux.com/puppy/viewtopic.php?t=72258&start=315

I filed a bug report and Thomas Leonard partly fixed it, back in April 2010:
http://repo.or.cz/w/rox-filer.git/commit/c2232d5075342347a8ff814ced5ce8b9e1cf64b9

However, it still wasn't fully fixed, as I reported here:
https://sourceforge.net/tracker/?func=detail&aid=3147977&group_id=7023&atid=107023

Well, a couple of days ago I posted a "whinge" to the 'rox-devel' mail list, and oh joy, a very kind person by the name of Nicola Fontana posted a patch ...which fixes it!
Rox-devel mail list thread:
https://sourceforge.net/mailarchive/forum.php?forum_name=rox-devel&max_rows=25&style=ultimate&viewmonth=201202

I will upload a PET very soon.

Unmounting network shares

Not my area of expertise! Recently, Forum member ldolse posted a patch for /etc/rc.d/rc.shutdown to properly unmount network shares. You will find that in Wary/Racy 5.2.2.

However, Forum member zekebaby reports that does not work for him:
http://murga-linux.com/puppy/viewtopic.php?p=605451#605451

I have thought about it, and patched rc.shutdown, inserted zekebaby's code just below ldolse's code:

#111107 ldolse: unmount network shares before taking down the network

#(see 111106, need to do it sooner, but 111106 will remount read-only if failed to umount here)

for MOUNTPOINT in `mount | grep '^//' | cut -d  ' ' -f 3 | tr '\n' ' '`

do

 umount -f $MOUNTPOINT

done 

#120219 hmmm, zekebaby reports above does not work for him. i looked up an ex: mount -t nfs boffo.clowns.org:/home/bozo /mnt/bozo

#so this is zekebaby's patch: (see http://murga-linux.com/puppy/viewtopic.php?p=605451#605451)

for MOUNTPOINT in `mount | grep ':' | cut -d  ' ' -f 3 | tr '\n' ' '`

do

  umount -f $MOUNTPOINT

done

Pupdial empty password/username

This problem has been around for such a long time! I have often thought of fixing it, but never got around to it.

Rerwin has proposed a fix:
http://murga-linux.com/puppy/viewtopic.php?t=72258&start=315

Thanks very much for that, I have put in the fix, but I didn't changed the "Stupid mode" name as I think it is too well known amongst wvdial users and the many online docs.

You can see the change to 'pupdial' here:
http://bkhome.org/fossil/woof2.cgi/info/8ae4b53c31

'gxmessage' PET

I have compiled this in Wary and put it into the 'common' repo (10KB):
http://distro.ibiblio.org/quirky/pet_packages-common/gxmessage-2.12.4-1-w5c.pet

The post-install script of this PET creates a symlink of 'xmessage' to 'gxmessage', thus overwriting any existing xmessage in /usr/bin.

I have modified all of the (x86) package lists in Woof to use gxmessage, as this is needed for internationalization. All (x86) builds should be able to use this PET.

MoManager uses poedit

MoManager now recognises the presence of 'poedit' or 'poeditor' and will display a checkbox to choose between poedit and the default text editor.

Note, I had a play with poedit, maybe I didn't really appreciate it's capabilities, as I still prefer to use Geany to edit po files.

Here is the momanager file:
http://bkhome.org/fossil/woof2.cgi/artifact/d5812623fa650d95fe1deca39bf9bcabd4973ab9

Translation of early-boot splash screens

When you boot Puppy from CD, and in some cases from USB or HD, a splash screen appears, to give an opportunity to enter kernel boot parameters. There are also two help screens, available by pressing or .

Side note: we probably should look into having this splash screen with frugal HD install and GRUB boot, it is useful.

Working towards being able to build a custom Puppy that boots first time in a non-English language, and is fully translated, I have added translation of the initial splash screen (file boot.msg) and the first help screen (file help.msg).

It is still a one-stop-shop for all translations, and this also is managed in MoManager. When someone creates a 'langpack' PET it may also now have the splash screen translations. If a pup is built in Woof with the langpack, 3builddistro will perform the translations.

To see how this was accomplished, look at the files in the latest Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/130b4c0ec0

Plans

I have been working so intensely on the internationalization stuff these last couple of weeks, I have got out of touch with reality.

Yesterday, I was thinking, is it Monday or Tuesday... then I looked and much to my surprise I saw it is Thursday!

Anyway, the momentum will continue for awhile, before I can get back onto "normal" Puppy things.

I will continue to work on various aspects of the internationalization, for example, the initial bootup splash screens -- aiming for Woof to be able to build a truly language-specific Puppy. The splash screen translations will also be managed by MoManager.

Racy 5.2.2.6-alpha build

This is a build to test the internationalization features of Woof.

I think that I have fixed some bugs that were in the Racy 5.2.2.5-alpha build, plus exciting new features such as modification of the initrd.gz to permanently set the chosen language at early bootup -- meaning that the on-screen messages right from early bootup will be in the correct language.

You can test by choosing 'de_DE' locale. For the initrd.gz to be modified, Puppy will have to be installed, not running off a CD. Puppy should detect the right conditions and ask if you want to modify the initrd.gz.

You will see the German boot messages at early bootup, but then going back to English, as the scripts in /etc/rc.d/ do not yet have 'de' translations -- run MoManager (Utility menu) and look in 'script_strings' SSS-domain, you will see the simple sed expressions awaiting translation.

Note to rodin.s, the SSS translation files have changed, the old ones have to be deleted.

Download Racy 5.2.2.6 from here:
http://distro.ibiblio.org/quirky/test/racy-5.2.2.6-alpha/

I am hoping now that things have stabilised enough, so that translators can start translating. MoManager is highly automated, and now has the lovely feature of creating a 'langpack' PET that you can email to me (or Forum pm) for uploading to the ibiblio 'noarch' repo.

MoManager is still only using Geany to edit po files. rodin.s did report being able to use poedit with MoManager, so it is still on my to-do list to detect poedit if it is present and optionally use it.

Another great development is that a Puppy build can now default to a non-English language at the very first boot. This is for those using Woof -- if you have a langpack PET for your language, include that in the build. The 3builddistro script will ask what language you want to set as default, and you choose your language. Hey presto, a custom Puppy for your language!

French keyboard broken in Racy

A few people have reported this to me, and I haven't known what to do about it.

Fortunately, Forum member 'npierce' has tracked down the cause of the problem:
http://murga-linux.com/puppy/viewtopic.php?t=75728

I have put the patched 'fr' keyboard file into 'x_xorg76_mega_pkg-7.6-4-w5c.pet' and it is uploading right now.

initrd internationalized

Well, I have done an awful lot of coding today, I'll be surprised if there isn't a typo in there somewhere.

I have internationalized the 'init' script in the initrd, but not with t12s. Instead, I have an extremely simple system much like the original idea from L18L, a file of simple sed expressions. I have done a 'de' translation to demonstrate, based on L18L's existing translations.

Translations for the 'init' script are managed in MoManager. They are stored at /usr/share/sss/initrd_strings/. The translations are copied into the initrd by '3builddistro' in Woof.

I have done something very ambitious with 'chooselocale' (which is now inside quicksetup). The user is asked if they want the initrd to be converted to their chosen locale, and if they do, then the initrd.gz is opened up and a file /PUPPYLANG is created in the initrd, plus latest translations for 'init' copied in. Then the initrd.gz is put together again.

Thus, on next boot, the early boot messages are automatically in the correct language. Note, a kernel 'plang' parameter is also recognized, which will override whatever the initrd is set to.

I have to tidy up a few details, then will do a Racy build.

MoManager generates 'langpack' PET

This is wonderful for the maintainer of a particular language translation for Puppy. MoManager can now create a PET of all relevant translations. I have updated the snapshot to show the new langpack button:

This is very convenient, as the language maintainer just has to send the PET to me and I will upload it to ibiblio.org for all Puppy builds to use.

MoManager handles binary executables

MoManager now manages translation for binary executables. I have updated the snapshot:

Langpacks for Puppy

I have been thinking how the output from MoManager can be improved. So far, MoManager creates individual tarballs of translations for each application, that can be sent to package authors or to me if they are Woof translations.

However, I have been rethinking that. Instead, MoManager can accumulate all the translations for a particular language and automatically create a "langpack" PET. That way, the translator won't just send off translations to app authors and hope they get put into the packages. The translator can still do that, but now will keep them all together as a PET also.

I have worked out a basic design for such a PET, in the case of Russian translation, as rodin.s has provided a lot of translations. The PET is uploaded (189KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/langpack_ru-20120213.pet

This PET can be used in two ways:
1. When a Puppy is built in Woof, so you would have a Russian Puppy out-of-the-box.
2. Installed after first boot and the user chooses a locale.

In the latter case, I have not yet automated that step, just putting up an advice window to run the Puppy Package Manager to install the langpack for a more complete translation.

The example Russian langpack has some interesting features, that translators will want to examine -- other translators can use it an a guide to what their own PET should look like, although I intend that MoManager will generate the PET automatically.

The langpacks will be in the 'noarch' repository, so must not have any binary executables.

It is possible that rodin.s will look inside the ru PET and might want to put more stuff into it. Which he is free to do as the maintainer of the ru langpack.

Anyway, I will do a Racy build tonight to do some sanity testing of the langpack. If it looks good, I will do some more work on MoManager to generate the PET, then upload a test Racy build.

A note to everybody reading this. The ru langpack PET is designed to work with very latest Woof-built Puppy. It will not work fully on older puppies. It really needs the Racy build that I will upload soon.

'xlock_gui' translations

The PET package 'xlock_gui' is ancient. It puts the screen-lock icon on the desktop.

When you right-click on it, you get a menu, in English, that is defined by an XML file in the package. I have added fr, ja and ru translations for the XML file, the latter two supplied by shinobar and rodin.s.

The PET package has been uploaded:
http://distro.ibiblio.org/quirky/pet_packages-noarch/xlock_gui-0.2-5.pet

Note, this package has a help file which currently only has a fr translation. If anyone want to download this package and add some more translations, they are welcome -- then send it to me.

'pup_event_icon_change_flag' path changed

This is a post for developers/coders. I have changed the path of file 'pup_event_icon_change_flag' from /tmp to /var/local.

This file is read by /sbin/clean_desk_icons, which in turn is run from /root/.xinitrc when X starts. If the file exists, the desktop drive icons are wiped and redrawn.

There are three files in Woof that write to this file: /usr/sbin/eventmanager, video-wizard and xorgwizard.

I changed the path because /tmp/pup_event_icon_change_flag does not survive a reboot. On the other-hand, /var/local is designated as a directory that Puppy will save.

For any developer who has an app that writes to this file, please make the change -- though of course you can write to both /tmp and /var/local, for compatibility with earlier puppies.

MoManager now manages XML files

MoManager has two more drop-down lists, for management of translations in XML files.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/1b114f7fd0

EDIT Feb. 13, 2012:
I updated the snapshot:

Racy build with MoManager and SSS

These are recent Woof commit comments for 2012-02-09:

[2b7e120e96] SSS bugfixes: fixdesk, rc.update
[0958a9147f] bugfixes of SSS infrastructure, add /usr/sbin/fixscripts symlink and /usr/share/sss/script_strings SSS-domain, reverted /etc/rc.d/functions4puppy4 from gettext to sss translation method
[5415937e5e] added /usr/sbin/fixdesk and desk_strings sss-domain, plus rest of infrastructure

Something that is very interesting about the SSS technique is that it handles any kind of text files, including scripts. I created a special "SSS domain" for scripts, /usr/share/sss/script_strings/. The file /usr/share/sss/script_strings/script_strings is the template, and I have put in one target-script to demonstrate it:

[general]

#this SSS domain is intended for any scripts that need translation.

#this is an alternative method to gettext and t12s, as they are runtime translations, hence script runs slower.

#the translations in this SSS-domain, that is, this file script_strings*, are performed by /usr/sbin/fixscripts,

#which in turn is called from quicksetup/chooselocale whenever locale is changed, also by rc.update whenever a version upgrade.

#the section-ids are a full path, for example _etc_rc.d_functions4puppy4 means /etc/rc.d/functions4puppy4

# ...i originally imposed this substitution for / as i thought there might be a problem with having / in the section-id

#    however, now i think not. anyway, fixscripts will accept any substitution, ex XetcXrc.dXfunctions4puppy4

#note, the text with \n in it (in the target file), this must be double-escaped to properly translate: \\n



[_etc_rc.d_functions4puppy4]

#target file is /etc/rc.d/functions4puppy4

CODESTRING='MOUSEINTRO="INSERTMARKER"'            ; ENGLISH='Only very old PCs may have a serial mouse. These are recognised by the fairly large rectangular plug, technically known as a DB9 or the even larger DB25 (where the 9 or 25 refers to the number of pins in the plug). Press the UP and DOWN arrow keys to choose your mouse type, then press the ENTER key:'  ; TRANSLATION=''

CODESTRING="MOUSEMENU1b='INSERTMARKER' #ttyS"     ; ENGLISH='(mouse in very old computers)'                 ; TRANSLATION=''

CODESTRING="MOUSEMENU2b='INSERTMARKER' #ttyS"     ; ENGLISH='(mouse or touchpad in all modern computers)'   ; TRANSLATION=''

CODESTRING="MOUSEMENU1b='INSERTMARKER' #usb-ps2"  ; ENGLISH='(mouse or touchpad in all modern computers)'   ; TRANSLATION=''

CODESTRING="MOUSEMENU2b='INSERTMARKER' #usb-ps2"  ; ENGLISH='(mouse in very old computers)'                 ; TRANSLATION=''

CODESTRING='INSERTMARKER" 0 0 0 "$MOUSEMENU1a" "$MOUSEMENU1b"'     ; ENGLISH='Please choose mouse type...'  ; TRANSLATION=''

CODESTRING='--menu "INSERTMARKER" 0 0 0 ttyS0'    ; ENGLISH='Serial port mouse plugged into...\\nUse UP/DOWN ARROW keys, ENTER key to finish'  ; TRANSLATION=''

CODESTRING='"COM1 INSERTMARKER" ttyS1'            ; ENGLISH='(most likely)'  ; TRANSLATION=''

If you want to create say a German translation, you make a copy of that file named script_strings.de and fill in the TRANSLATION variables. That's it.

The target script, in that example, /etc/rc.d/functions4puppy4, is the original unmodified script. Note, I had earlier implemented gettext in that script, but I changed it back to it's pre-gettext state. The SSS mechanism translates the script to the required language, so there is no run-time translation, so the script runs faster.

This mechanism works for any script. MoManager is the GUI tool for creating and editing these SSS translations, and the actual translation of the scripts occurs whenever there is a locale change or a version upgrade.

I don't envisage using the SSS technique on all scripts, as we have 100+ of them, which might take rather a long time. Or maybe not, I haven't timed it. I was thinking of applying it to a few of the time-critical scripts.

Anyway, I have built a Racy Puppy for anyone who wants to try this out (113.5MB):
http://distro.ibiblio.org/quirky/test/racy-5.2.2.5-alpha/racy-5.2.2.5-alpha.iso
http://distro.ibiblio.org/quirky/test/racy-5.2.2.5-alpha/racy-5.2.2.5-alpha.iso.md5.txt

I have only done German translations, so at first boot, choose the de_DE locale, with UTF-8 enabled, then restart X. You will see the desktop icons and JWM menu in german.

To study how it is done, look at MoManager in the Utility menu, and for the SSS stuff look in /usr/share/sss and the scripts /usr/sbin/fixmenus, fixdesk and fixscripts. The scripts quicksetup and rc.update call these.

If anyone wants to add translations of any text data-files or scripts, just add them into /usr/share/sss/desk_strings/ or script_strings/.

Testing can be done easily. For example, say that you add a new file into /usr/share/sss/desk_strings/desk_strings, and a translation file /usr/share/sss/desk_strings/desk_strings.de, you can test translation of the target file by running this:

# fixdesk de
# fixdesk en

...to convert the target data-files to de then back to en.

Note that if you use MoManager to edit an existing desk_strings.de file, there is automatic syncronisation with the master (template) file desk_strings, so the latter is the only one that you need to change if the target files are changed, added or removed.

Anyway, it is better if I don't overload the reader with lengthy explanations. It is quite simple, as you will see.

Note, the the Racy build is "alpha", meaning that it hasn't had much testing and bugs could show up.

MoManager: SSS domains

MoManager is my GUI application that intends to make it easy for anyone to translate Puppy for their (non-English) language. See earlier blog posts:
http://bkhome.org/blog/?viewDetailed=02676
http://bkhome.org/blog/?viewDetailed=02678

There has been much progress. Here is a snapshot:

I have been studying shinobar's technique of using gettext translation for all types of text files. Very impressive, but I decided to play around with a different technique, that I have named SSS, Simple String Substitution -- to make it seem much more fancy than it is . I even introduced the term SSS Domains to make it seem even more clever!

The SSS files are at /usr/share/nls. So far I have only created one SSS-domain, /usr/share/nls/menu_strings/, and the way to understand what it is all about is to look at the files in that directory. Then look at script /usr/sbin/fixmenus to see how it reads the SSS file -- for example, if running with LANG=de_DE.UTF-8, the appropriate translation-file for the 'menu_strings' SSS-domain is /usr/share/nls/menu_strings/menu_strings.de.

It really is very simple. I have only created 'menu_strings' so far, and that even incomplete (only handling JWM), as people might tell me it is a really dumb method.

Yes, it is a dumb method, but does have some great advantages. Firstly, it is very easy to understand and create translations. Secondly, it handles any type of text files -- absolutely anything that needs English text substituted for another language. No knowledge of the syntax in the text file is required.

To view the appropriate files, here is the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/cc6861c1a2

I plan to build a Puppy from latest Woof in a couple of days, for those who would like to play with this.

Note, there is one change I am thinking of making, change /usr/share/nls to /usr/share/sss.

xdg_puppy 0.7.6-13

xdg_puppy is a PET used in all puppies. It defines the structure of the XDG menu. It was forum member rarsa who originally created this PET, and rarsa wrote some utilities in C for translating the XDG specifications into actual menu files for some of our window and tray managers such as JWM and IceWm.

The PET contains folder /usr/share/desktop-directories, which has files that define the menu hierarchy, that is the top-level categories and subcategories. Early in 2011, shinobar added translations to these files, so that the top-level and sub-level titles in the menu will display in the appropriate language. Shinobar created a PET, puppy-desktop-20110101.pet, see Forum thread:
http://www.murga-linux.com/puppy/viewtopic.php?t=63533

The PET also does have some other stuff, but I have taken just the /usr/share/desktop-directories and put it into xdg_puppy-0.7.6-13-p4.pet, which will now be the PET used for all future Woof builds. Here it is, in the 'common' repo (15KB):
http://distro.ibiblio.org/quirky/pet_packages-common/xdg_puppy-0.7.6-13-p4.pet

What this means for future Puppy builds is that now the entire JWM (or IceWM or whatever) menu will be translated. Well almost, there are some menu titles in .jwmrc that still need to be translated -- shinobar worked out a system for that too, that's on the to-do list.

MoManager has creation/editing of menu entries, the next thing I am going to do is add creation/editing of the categories/sub-titles as specified in /usr/share/desktop-directories.

Simplifying translations

I have a question for non-English Puppians. I would like to simplify many of the gettext translations in Woof scripts, leaving out any XML or HTML or any other formatting, also as much as possible leaving out substitutions for example "${VAR1}".

I really do not like having formatting in a translation string, they should be simple text strings. That is my ideal anyway.

I would also like to avoid use of eval_gettext as much as possible, again because I would like the strings to be simple easy-to-understand plain text.

The biggest problem for the latter though, is the semantics of some non-English sentence construction, or ordering of the words. An English string:

"The script $FILE1 has changed"

I would like to translate this in two pieces:

"$(gettext 'The script') $FILE1 $(gettext 'has changed')"

I don't know much about non-English languages, but I understand that simply translating the above might not be very good for some languages, due to order of words having to be re-arranged.
Perhaps there is a better example of this problem than the above.

Anyway, I was wondering if I can "get away with it". Even if the translation is a bit strange. or, is it possible to work around the problem, change the wording in those two gettext translations such that the translated sentence is quite sensible?

MoManager handles .desktop files

The first release of MoManager manages gettext'ed scripts, announced earlier:
http://bkhome.org/blog/?viewDetailed=02676

Now I have added support for creation and editing of .desktop menu files. Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/47509d2301

But, I need some help. For testing purposes, I have got LANG=de_DE.UTF-8 in /etc/profile. Applications/scripts that have a 'de' .mo file translate ok. However, the JWM menu does not translate. Even though there are some .desktop files with Name[de]= entries, only the English Name= entry is displaying.
Have any of you guys who use non-English locale encountered this problem, and more importantly know the cause?
...I wonder if I have stuffed up the LANG setting in 'fixmenus'?
EDIT: yes, I had a moment of dumbness. MoManager was calling 'fixmenus' with LANG=C. Fixed, will upload to online Woof repo soon.

Author info in pet.specs file?

There has been some discussion recently on the Forum about the PET package format:
http://murga-linux.com/puppy/viewtopic.php?t=75383

It is very healthy to have this kind of analysis. Yes, the PET format started life in Puppy 2.0, but it has evolved since then. In late 2011 there was a significant improvement in the format of information kept in the "database" entry for a PET package, that is, the content of 'pet.specs'.

In particular, there is now versioning support for dependencies:
http://bkhome.org/blog/?viewDetailed=02368
http://bkhome.org/blog/?viewDetailed=02363

I am currently working on MoManager, which is a tool to internationalize Puppy. When a translation file for a script is created or updated by MoManager, the package to which the script belongs is identified, and if a Woof script, MoManager supplies my email address and gzips the new .mo file ready to send. However, a consistent mechanism is needed to do that for PET packages -- if pet.specs contains author and email address, that would solve that problem.

The way that I implemented the "database", one line per package, fields that are delimited by a | character, it is easy to extend with more fields.

Any thoughts on this are welcome.

Introducing MoManager

I am very mindful of the unsatisfactory situation with internationalization of Puppy. We have guys in the past who have done translations but they never made it into later Puppy builds.

Also, there hasn't been a consistent methodology, so developers have tended to do bits of translation here and there.

What we need is a single tool, a GUI application, that someone can use to completely translate Puppy to their language. Furthermore, that tool must produce translations that can readily be sent to me for inclusion in Woof and to the maintainers of packages.

This is a magical tool, and I am dreaming somewhat, but I have made a start, it is named MoManager.

As we now have a huge number of scripts in Woof (as well as in many PET packages) that use the 'gettext' translation method, this is what I have implemented for the first release of MoManager. This is a one-stop-shop for creating and maintaining translations for all the gettext'ed scripts in Puppy.

To use MoManager, you really need to be running a Puppy built from the very latest Woof that I have just uploaded. This has the big collection of gettext'ed scripts created by rodin.s and myself (mostly rodin.s) -- and importantly, each script has a unique TEXTDOMAIN, which is what MoManager requires.

MoManager will not only allow you to create new translations but also edit existing ones -- and synchronise with the latest script. And it is all done in GUI windows, clicking buttons, no need to type commands.

Note that translations are to be kept in Woof (and in PET packages) as .mo files. .po files are not required.

I have updated the Internationalization HOWTO with information on MoManager:
http://bkhome.org/fossil/woof2.cgi/artifact/60cd815002b4de659f1c37459b4974512793a015

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/6e0f3af7f1

There are probably going to be some people keen to use this, so I will do a build of Wary or Racy in a few days.

Tomorrow I intend to start extending MoManager a bit, probably add management of non-English entries for the menu and for ROX-Filer. I haven't thought about integrating t12s'ed scripts into it yet.

T2-raspberrypi, Musl

For coders/developers, this stuff is interesting.

There is a port of T2 that aims to support the RaspberryPi:
https://github.com/arete/t2/commits/raspberrypi

Dougal sent me a link to 'musl', an interesting alternative to uclibc:
http://www.etalabs.net/musl/

Android on the PC

This is interesting:
http://www.muktware.com/news/3275/soon-you-will-be-running-android-your-pcs

I have never used Android, so don't have any concept of how it's usability stacks up against our usual Linux distros. I don't know how well it will perform with the conventional keyboard/mouse input.

Very interesting though, if it does start to appear on laptops/desktops.

Internationalization progress

I have finished putting rodin.s's internationalized files into Woof.

In the Internationalization HOWTO page I mentioned technosauruses/L18L's 't12s' method and that it might be our preferred method. However, the DoOcracy has been demonstrated, with rodin.s submitting a huge collection of files using the 'gettext' method.

However, the t12s method is good and any existing non-internationalized scripts or future scripts will be welcome to use it.
Note, Woof currently only has one t12s'ed script, /usr/sbin/xdelta_gui.

Internationalization HOWTO:
http://bkhome.org/fossil/woof2.cgi/artifact/375a350ce4363addfd87592bd883b00d7f9fe112

Blog post on t12s:
http://bkhome.org/blog/?viewDetailed=02620

Earlier blog post describing localization:
http://bkhome.org/blog/?viewDetailed=02652

What I am planning on working on next is a GUI tool to make it easy to create translation files. We have the i18n'ed scripts (and some XML and .desktop files) but now need to make it very easy for users to create non-English translations. This will be for the gettext'ed scripts.

I might be a bit unresponsive to other to-do requests for a few days

Puppy is a DoOcracy?

There is seeming disorder in the way that the Puppy project is run, for which I sometimes get criticized as a weak leader. However, what those people don't understand is that my strategy is deliberate. Although I am also sometimes viewed as the "benevolent dictator", that does not really fit the model -- I like to see individuals empowered to "run with the ball".

A little while ago Jemimah posted to the Puppy Forum about "DoOcracy", which kind of fits how the Puppy project is run. Forum thread:
http://www.murga-linux.com/puppy//viewtopic.php?t=55028&start=75

A page introducing DoOcracy:
http://www.communitywiki.org/en/DoOcracy

Yet more scripts internationalized

I have applied rodin.s's internationalised scripts connectwizard, connectwizard_2nd, cups_shell, delayedrun, dotpup and download_file to Woof. See Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/4978ad1d29

I decided to translate /usr/sbin/eventmanager myself, as I wanted to try doing it slightly differently.

That slight difference is just a small thing... in gtkdialog XML code, rodin.s translates strings by inserting directly, for example;

export DIALOG=" ... ... "

However, if the string has a ' in it, rodin.s does it this way:

export DIALOG=" ... ... "

Which is fine, however I wanted to consistently use only the $(gettext ....) syntax. So my translation is:

export DIALOG=" ... ... "

This also works, and the " does not need to be back-slashed, in fact must not be back-slashed. It confuses Geany though, but we are accustomed to that It works because everything inside $( ... ) gets evaluated first, and only the resulting string inserted into DIALOG.

It is interesting that the back-tick works differently.

More scripts internationalized

I am working through the scripts that rodin.s has sent me. It is taking awhile as I am editing the Woof scripts line-by-line.

These are the latest internationalized scripts:
http://bkhome.org/fossil/woof2.cgi/info/40efc714d4

Xorg Wizard internationalized

Rodin.s has supplied me with a lot of internationaized scripts for Woof. I am starting to work through them.

I have already upgraded 'dialog' to support wide-characters, see earlier post:
http://bkhome.org/blog/?viewDetailed=02667

I have applied rodin.s's changes to /etc/rc.d/functions4puppy4, /etc/rc.d/rc.shutdown and /usr/sbin/xorgwizard.

The last one was a lot of work, as rodin.s's script is somewhat different from that in Woof. So, I went through line-by-line making the changes.

Anyone who wants to create a .mo translation file for 'xorgwizard' is welcome! Note, download 'functions4puppy4' also, as it is included inside 'xorgwizard'.

You can get the files from the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/229cbfcb8d

The return of Underdog Distro

This is a feature that was last seen in Puppy 2.x, it was gone in 3.x. I was reminded as tronkel started a thread about it on the Forum:
http://murga-linux.com/puppy/viewtopic.php?t=13750

It is an interesting concept, that is introduced here:
http://puppylinux.com/development/howpuppyworks.html

There are possible problems, mostly to do with mismatched libraries. The Puppy libraries would be in a higher layer, which might mean that an application that you want to run in the Underdog Linux layer might not work. With some care though, it should be ok, for example with Wary/Racy, use an older version of Ubuntu not the latest as the Underdog. Or, better, use a matching distro, example Slacko and Slackware 13.37.
Another possible problem is all the extra bootup stuff in an Ubuntu (or whatever) Underdog.

I do recall though, back in the Puppy 2.x days, Underdog did work surprisingly well.

I have modified the 'init' script in the initramfs, in Woof, so this feature will be in the next Puppy built with latest Woof. Here is the Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/e82716c2f7

What you do is create a file 'underdog.lnx' in /mnt/home or in the subdirectory of a frugal installation. The file contains the name of a partition, for example "sda9", nothing else.

However, you will also be able to do it from the kernel boot commandline, ex: "underdog=sda9". The really interesting thing about this is it will work even on first boot before you have created a save-file. The partition is mounted read-only, and will be the very bottom layer.

I haven't tested it yet.

dialog wide-char enabled

Rodin.s has sent me some files for improving internationalization in Puppy. One of those files is 'dialog' linked against 'libncursesw', for wide-character support (such as UTF-8).

There is some history with regard to 'dialog' and wide-char support. See these two blog posts:
http://bkhome.org/blog/?viewDetailed=02201
http://bkhome.org/blog/?viewDetailed=02204

Despite having created a PET with wide-char support, I backed off and recent Puppies have used 'dialog' linked against 'libncurses' only. I don't recall why I backed off.

Anyway, I have compiled dialog version 1.1-20111020 with wide-char support, configured as per the second link above. Here are the PETs (72KB, 15KB, 17KB):
http://distro.ibiblio.org/quirky/pet_packages-common/dialog-1.1-20111020-widechar-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-common/dialog_DOC-1.1-20111020-widechar-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-common/dialog_NLS-1.1-20111020-widechar-w5c.pet

I have put these into the 'common' repo, so that all Woof-built puppies can use them. Note, it does not require 'libtinfo'.

It is unfortunate that UTF-8 has to be enabled before dialog will display wide-characters properly. It really should recognise wide-characters regardless, so this perhaps should be reported as a bug to the dialog developer.

Pfilesearch 1.28

Zigbert's Pfilesearch is a helper-script called by some of zigbert's other apps. I have updated to the latest version.

PET (9KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/pfilesearch-1.28.pet

Snap2 4.21

This is Lloyd Standish's backup utility. project home:
http://standish.home3.org/snap2

The version I had was getting a bit old, so I have updated to version 4.21. This requires 'rsync' but is itself architecture-independent, so I have put the PET into the 'noarch' repo (31KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/snap2-4.21.pet

Asunder, Homebank, Isomaster

I have compiled these for Wary (and Racy):

Asunder 2.1 (audio CD ripper)
Homebank 4.4 (personal finances, supports OFX)
Isomaster 1.3.9 (edit ISO files)

These are included in Wary and Racy builds.

Asunder PET (46KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/asunder-2.1-w5c.pet

Homebank PET (208KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/homebank-4.4-w5c.pet

Isomaster PET (76KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/isomaster-1.3.9-w5c.pet

Woof: synclient, shutdown fixes

Jemimah and Karl Godt have posted some little fixes, put into Woof, see commit 2012-01-29:

http://bkhome.org/fossil/woof2.cgi/timeline

ROX-Filer focus problem still there

This problem has been with us for sometime, and I have reported on it in this blog. Recently, a guy (Sourceforge member 'n8o') posted a patch, which raised my hopes and I have just tried it -- it doesn't work.

So, I have posted a description of the problem to the ROX mail list, in the hope that Thomas Leonard, the founder of ROX-Filer, will tackle it. Here is my post:

Probably this email needs attention by Thomas Leonard, thanks.

I originally reported an icon focus problem, that Thomas fixed, see patch from Thomas:
http://repo.or.cz/w/rox-filer.git/commit/c2232d5075342347a8ff814ced5ce8b9e1cf64b9

...which did fix the problem, partly. However, even that fix no longer works.

It may be because Puppy Linux has progressed to GTK 2.24.8. Back then, I think we were using GTK 2.18.x or 2.21.x.

Back then, after applying Thomas's patch, I found that there was still a problem, that I reported:
https://sourceforge.net/tracker/?func=detail&aid=3147977&group_id=7023&atid=107023

Note, not many ROX-Filer users are reporting this problem, as some binary package maintainers, Debian included I think, also some releases of Puppy Linux, have put in a work-around. If this is put into /etc/profile:

export GDK_NATIVE_WINDOWS=true

...this does fix it, however, I have had reports that this is a hack that can upset some applications, can even upset ROX-Filer, cause a
crash.

I really don't want to have to resort to that hack. This problem has been going on for such a long time, it would be great if it could be
finally fixed.

Sourceforge member 'n80' has recently posted a bug report about this bug, and a patch:
https://sourceforge.net/tracker/?func=detail&aid=3463291&group_id=7023&atid=107023

...the patch is for ROX-Filer 2.11. I have tested it, it doesn't work. Perhaps again, it is an issue with GTK 2.24.x.

Regards,
Barry Kauler

FFconvert 1.2

Shinobar has updated his video/audio file format converter (that uses ffmpeg) to version 1.2:
http://murga-linux.com/puppy/viewtopic.php?t=54056

I have uploaded the PET to the 'noarch' repo (51KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/ffconvert-1.2.pet

mtPaint 3.40

The mtPaint home page:
http://mtpaint.sourceforge.net/

Wjaguar's latest version, 3.40 download:
https://github.com/wjaguar/mtPaint/archives/master

I have compiled version 3.40 in Wary and uploaded the PET (311KB, 360KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/mtpaint-3.40-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/mtpaint_NLS-3.40-w5c.pet

Pmusic 2.3.0

Zigbert has updated his music player application to version 2.3.0:
http://murga-linux.com/puppy/viewtopic.php?t=31206

I have uploaded the PET to the 'noarch' repo (88KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/pmusic-2.3.0.pet

Pburn 3.4.3

Zigbert has released version 3.4.3 of his CD/DVD burning application:
http://www.murga-linux.com/puppy/viewtopic.php?t=23881

I have uploaded the PET to the 'noarch' repo (65KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/pburn-3.4.3.pet

Grub4dos 0.4.4.v1.7.2

Shinobar has upgraded Grub4dos, see Forum:
http://www.murga-linux.com/puppy/viewtopic.php?t=51697

And shinobar's download site:
http://shino.pos.to/linux/puppy/

I have uploaded the PET to the x86 'common' repo (211KB):
http://distro.ibiblio.org/quirky/pet_packages-common/grub4dos-0.4.4.v1.7.2.pet

Wary: SeaMonkey 2.6.1, Sqlite 3.7.10

I have compiled these for the next Wary and Racy.

Sqlite (368KB, 427KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/sqlite-3.7.10-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/sqlite_DEV-3.7.10-w5c.pet

Seamonkey (16.4MB, 4.4MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/seamonkey-2.6.1-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/seamonkey_DEV-2.6.1-w5c.pet

Linux kernel 2.6.32.55

Wary 5.2.2 has kernel 2.6.32.45, configured with uniprocessor (not SMP) and aufs driver. I have now compiled kernel 2.6.32.55, uniprocessor, and intend to use that for the next release of Wary. I also used Unionfs instead of Aufs, and enabled "Fusion MPT SCSI device support".

The sources and build scripts are here:
http://bkhome.org/sources/kernel-2.6.32.55/

The PET (28.8MB) (this PET is for Woof users only):
http://distro.ibiblio.org/quirky/pet_packages-wary5/linux_kernel-2.6.32.55-uni-i486-unionfs-modems-w5.pet

I also compiled nVidia kernel drivers (3.3MB, 3.5MB, 1.2MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/nvidia_kmodule-195.36.31-k2.6.32.55-wary5.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/nvidia_kmodule-260.19.12-k2.6.32.55-wary5.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/nvidia_kmodule-71.86.14-k2.6.32.55-wary5.pet

Linux kernel 3.0.17

Racy 5.2.2 has the 3.0.7 kernel. I intend to use the 3.0.17 kernel for the next release of Racy, rather than a 3.1.x or 3.2.x kernel, as I wish to keep using Unionfs -- for which there is only a RC release for those kernels.

So, I compiled 3.0.17, patches, build scripts, patched sources, including SFS, are here:
http://bkhome.org/sources/kernel-3.0.17/

The PET is here (this is for Woof developers only) (23.4MB):
http://distro.ibiblio.org/pub/linux/distributions/quirky/pet_packages-wary5/linux_kernel-3.0.17-pae-i686-unionfs-w5.pet

Woof fixes

Fixes have been submitted by Jemimah, npierce and mavrothal, see commit date 2012-01-25:

http://bkhome.org/fossil/woof2.cgi/timeline

Puppy localization

I just posted the following to one of the Akita Puppy threads on the Forum, thought that it would be informative here also:

Regarding localisation of scripts, Woof now has /usr/share/doc/HOWTO-internationalization.htm -- not in Wary/Racy 5.2.2.

This describes the three techniques employed by Puppy, as well as technosauruses technique -- the latter has been used by L18L and L18L has created an online help page will more details.

View HOWTO file:
http://bkhome.org/fossil/woof2.cgi/artifact/375a350ce4363addfd87592bd883b00d7f9fe112

If anyone wants to contribute to internationalization of a script in any of the puplets, get the latest version of the script out of the Woof Fossil online repository. It is easy to grab individual files out of the Woof repo -- go to http://bkhome.org/fossil/woof2.cgi/index, login as anonymous, click on 'Files' in menu, locate file, then click on 'Download'. Or, download the entire Woof repo, which is explained in online help page.

After changing the script to support localization, send it to me at STARTbkaulerATgmailDOTcomEND -- please put "woof" in the message title. Or, if adding translations, send those to me also.

b43 firmware loading maybe fixed

I don't have the hardware to test, but I might have found the solution.

The problem is that the firmware for the b43 module does not load for the 3.x kernels.

I examined the udev logs that peebee posted:
http://www.murga-linux.com/puppy/viewtopic.php?t=75209&start=150

Running "modinfo b43" told me that this is an 'ssb' device, and it seems that the appropriate kernel uevents are not getting replayed. This gave me the hint. I looked into /etc/rc.d/rc.sysinit and found a bug in version comparison -- a test for kernel versions under 2.6.24 is broken for 3.x kernels, causing 'ssb' uevent replay not to occur.

Fixed, Woof commit here:
http://bkhome.org/fossil/woof2.cgi/info/c74bc1e0cf

Back to full-speed soon

I have been off-line the last couple of days, so haven't responded to posts. However, I am no longer living in a tent, back into a real house!

I will be home by this weekend and looking forward to getting up to full-speed working on Puppy.

'ndiswrapper' written in C, not Perl

I am still on holiday, and don't have the hardware to test this. Help requested!

In this thread, PANZERKOPF posted an ndiswrapper.c that was obtained from the GeexBox Linux project:
http://murga-linux.com/puppy/viewtopic.php?t=51478&start=405

I am looking again at being able to build Puppy distros without Perl. Currently, we have a cutdown Perl in all puppies, needed by ndiswrapper, pplog, one of the partview variants, don't know if anything else.

The ndiswrapper package has /usr/sbin/ndiswrapper, a Perl script. This was the original reason that we were forced to introduce Perl into Puppy.

Note, if we take out cutdown-Perl, the full Perl will still be in the 'devx' SFS, as required for compiling source packages.

I would like to know if the ndiswrapper executable written in C, that PANZERKOPF found, works. Could you test it with your hardware?
There is one issue apparently, the *.inf file must not have multi-byte characters (unicode) -- it would be good if the program could detect that rather than just fail. But, are there any such .inf files out there?

Alternatively, we can pre-process such a .inf file, remove multi-byte chars. I posted how to do that here:
http://bkhome.org/blog/?viewDetailed=02620

The compiled binary is posted to the Forum. Gunzip it, make sure the executable flags are set, and place it at /usr/sbin. Remove prior files at /etc/ndiswrapper then try your wireless hardware:
http://murga-linux.com/puppy/viewtopic.php?t=75352

Feedback will be greatly appreciated!

Partview rewritten

I posted recently about a patch created by Karl Godt for Partview:
http://bkhome.org/blog/?viewDetailed=02632

That version of Partview is a Perl script, created by Forum member ecube. Ecube contacted me, also unsure what Karl's patch does, and ecube sent an alternative modification. I have applied it and the PET is now 'partview-2.2-2.pet'.

I used ecube's Partview in Wary and Racy, however, Woof does already have /usr/sbin/partview, an Ash script that generates an xpm image and uses gtkdialog. Using ecube's PET replaces that. There were some other guys who also had a go at writing a Partview.

I want to have another look at removing Perl from the base Puppy system, so I have removed ecube's PET from the Wary/Racy package list. I checked the package-lists for some other Woof puplets, they use the internal partview.

I have also rewritten the internal Woof partview, it now generates a nice SVG image. It was technosaurus who got me thinking about this, as he created 'text2xpm' and in the Forum also suggested that it might be easier/simpler/nicer to use SVG to generate bar graphs with text in them. Yes indeed, it is very nice.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/d738537e7f

"Puppyfied" gFTP

Ttuuxxx has "puppyfied" gFTP:
http://bkhome.org/blog/?viewDetailed=02644

I compile gFTP for Puppy with a couple of patches, so I made ttuuxxx's changes into a patch also, 'gftp-ttuuxxx.patch'. Original source, patched source, and patches are at my source repo:
http://bkhome.org/sources/alphabetical/g/

I have compiled in Wary and uploaded PET (248K):
http://distro.ibiblio.org/quirky/pet_packages-wary5/gftp-2.0.19-patched_2ubuntu1_expandedpathsigsegv_ttuuxxx-w5c.pet

JWM 574: '-reload' icons fixed

We reported about the icons getting washed-out when "jwm -reload" is executed:
http://bkhome.org/blog/?viewDetailed=02629

I posted a bug report and after I supplied some snapshots to Joe and found the exact revision where the bug came in (rev. 514), Joe was able to identify the cause, and rev. 574 fixes it.

New PET (75KB):
http://distro.ibiblio.org/quirky/pet_packages-common/jwm2-574-w5c.pet

Woof commit, fixed ppm scripts:
http://bkhome.org/fossil/woof2.cgi/info/f8ead420c2

Allwinner A10 linux board coming

This is very interesting. A UK company is planning to create a board with a similar price to the RaspberryPi, but with a much more powerful CPU.

Yeah, that is going to be the main thing that will hinder adoption of the RasPi, after the hype has died down. The 1.5GHz A8 ARM CPU in the Allwinner A10 is three times faster than the ARM11 CPU in the RasPi (that's what they claim anyway).

Here is the news report:
http://www.itwire.com/opinion-and-analysis/open-sauce/52054-british-company-looks-to-create-cheap-open-platforms

A couple of very interesting points about this initiative: it seems the chip will be GPL-compliant (not really sure about what they are claiming here), and their first board will be a PCMCIA board.

That last one seems odd, but they are talking about a new-generation PCMCIA, which, it seems from further reading that it serves as a compact expansion bus. You can have a CPU board, with lots of I/O on one end, PCMCIA plug on the other, and add all kinds of interfaces. More info:
http://elinux.org/Embedded_Open_Modular_Architecture/PCMCIA

Here is a page on the Allwinner A10 chip:
http://rhombus-tech.net/allwinner_a10/

This page shows a cut-and-paste of what the board is going to look like:
http://rhombus-tech.net/allwinner_a10/news/

If they really can bring this out at a RasPi-price, well even a bit more, then they are onto a winner. I'll be watching progress!

Small 'fido' fixes

I fixed a couple of things running Puppy as user 'fido':
http://bkhome.org/fossil/woof2.cgi/info/b7b440326c

It is still broken somewhat. I plan to do a bit more work on it.

GTK "gseal" fix, take 2

I reported on this saga here:
http://bkhome.org/blog/?viewDetailed=02618

Well, today I built Wary with GTK 2.24.5, as reported in above link, but the slow/hanging BootManager (Xdialog SFS-chooser window) is back again.

So I tried something that I had not done before, I compiled GTK 2.24.8 with this:

# ./configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var --build=i486-t2-linux-gnu --with-xinput=yes --enable-debug=minimum --enable-introspection=no

I had not previously compiled GTK 2.24.8 with '--enable-introspection=no'. Well, that fixed it!

There were three problems:

1. Slow/hanging BootManager Xdialog window.
2. "gseal" errors when compile applications.
3. Long drop-down lists in GTK apps very slow.

Number 2, I compiled 'gtktetris', which previously exhibited the "gseal" errors.

Number 3, 'quicksetup' exhibits the slow and cpu-hogging drop-down lists.

I built wary with the new GTK 2.24.8, the above 3 problems have gone away.

The PETs are uploaded:
http://distro.ibiblio.org/quirky/pet_packages-wary5/gtk+-2.24.8_no_introspection-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/gtk+_DEV-2.24.8_no_introspection-w5c.pet

My package-selection entry:

yes|gtk+|gtk+-2.24.8_no_introspection*,gtk+_DEV-2.24.8_no_introspection*|exe,dev,doc,nls|

Fossil 1.21

I have created a Fossil 1.21 PET, in the 'common' repo (1.2MB):
http://distro.ibiblio.org/quirky/pet_packages-common/fossil-1.21-x86.pet

In the next upload of Woof, the package-selection files will include this to go into the 'devx' SFS.

usb-modeswitch 1.2.1

I have upgraded usb-modeswitch to version 1.2.1.

The PET is in the 'common' repo for all Woof-built puppies to use (28KB):
http://distro.ibiblio.org/quirky/pet_packages-common/usb-modeswitch-1.2.1-w5c.pet

The data is in the 'noarch' repo (15KB):
http://distro.ibiblio.org/quirky/pet_packages-noarch/usb-modeswitch-data-20111023.pet

The usb-modeswitch PET has a couple of modifications so as to work with two scripts that are in Woof, /usr/sbin/usb_modeswitch_status and usb_modem_special_status. Those two scripts have also been modified, a couple of improvements. See Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/c6277794a9

USB 3G modems that do not require usb-modeswitch
My Optus 3G modem is in this category. With those that do require switching, I have set things up so that when you plug it in, a message comes up saying "USB 3G modem detected, configuring..." followed by another message of success and ready to use.
However, for a USB 3G modem that does not require a mode-switch, the messages do not pop-up. I have created special detection of my Optus modem, so the messages do popup.
If you also have such a modem, let me know, post the output of 'lsusb' (running Wary or Racy, which has the full lsusb and usb.ids file -- I don't know if any other Puppy builds do). Then I can get the id numbers and put in the special handling.

Conia TV 1920x1080 with HDMI

Well, I am getting ready for when I buy my first RaspberryPi!

Just before Christmas I purchased this:
http://www.dealsdirect.com.au/p/conia-22in-full-hd-led-tv-w-dvdpvrusb-hd-tuner/

My old CRT analog TV has been retired. The poor thing got damaged by lightning and blew the infra-red to the remote control, plus it randomly won't turn on when I push the power button. It is now replaced by this cheeky young fella.

I know that there are cheaper LED TVs available, I saw a 16 inch LED TV recently for AU$88, but I want 1080p HD resolution. The 22 inch Conia is the smallest that does that -- it has 1920x1080 pixels. It also has a DVD player builtin.

So, I am being quite economical, getting the one screen for both TV and computer monitor.

I am still camping, and brought this TV with me. Yes, it runs off 12 volts, draws 60 watts (5 amps). Works great, except that my rabbit-ears antenna doesn't fit too well inside the tent! -- so, it is going back into it's box and back into the car. After all, what kind of holiday is it if I just sit there watching TV? -- might as well be home.

On the subject of the RasPi, one guy has started a series of video tutorials:
http://www.youtube.com/raspberrypitutorials

Lucid Puppy 5.2.8 Updates

Lucid Puppy 5.2.8 is the last in the "Lucid series" of official Puppy releases, however the developers have continued to develop it and have released updates. These are available as PET packages, or an ISO (live-CD) file built with the update.

It is recommended that Lucid with update be downloaded, so I have inserted a notice to that affect into the original Lucid 5.2.8 announcement on my blog:
http://bkhome.org/blog/?viewDetailed=02418

I have also modified the download page to inform of the updates:
http://puppylinux.com/download/

PPM tweaks

Rerwin posted some little fixes for my recent changes to PPM scripts.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/0cc4a30c89

SNS: Profile management

Simple Network Setup now has profile management. It was simple enough to do, just a morning's work, but greatly improves usability of SNS.

Each time that you setup a network/Internet connection with SNS, it is saved as a "profile", a one-line entry, in file /etc/simple_network_setup/connections. Anyone with this knowledge, and knowledge of the fields in the profile, could directly edit the 'connections' file, however most users will not be in that situation.

Previously, when SNS is started, the main GUI window doesn't tell you about any existing profiles, and you have to start from scratch.
Now, there is a new frame in the main GUI with basic profile management features. More detail to it could be added later, but for now there is basic connect/disconnect and deletion of invalid profiles. I also added a Help window explaining about profiles.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/2203405fe9

Arch Linux -4 patch

Forum member noryb009 sent me another patch for Arch Linux support in Woof. Quoting the pm:

This is a small patch to speed up arch linux in 0setup, along with a sorting fix.

- make the awk command get the category and path (doesn't need to be parsed the second time), which speeds up the script a bit.
- removes using the sort command, ls already sorts the folders by name

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/f5b36f77c2

Firmware for rtl8192se wireless driver

Forum member nedsnow posted that the rtl8192se wireless driver loaded but did not work in Racy 5.2.2:
http://www.murga-linux.com/puppy/viewtopic.php?t=74819

Good thing that I saw that post, as I am mostly off-line these days and not reading the Forum much, though I am trying to follow the Wary 5.2.2, Racy 5.2.2 bug threads, plus the Slacko 5.3.1 bug thread (not quite up to date on that one though).

Anyway, Forum member pemasu replied to that thread and posted firmware for testing, as required by the driver. I have put this into Woof. Pemasu, thanks for that!

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/df52b04e76

Desktop icons mounted-status bug

Karl Godt reported about this problem. It can occur when there are more than 9 partitions in a drive. Karl was getting the 'sda1' icon showing as mounted when in fact it wasn't. I also have had this happen to me a couple of times over the past year, infrequent enough so as to be puzzled but not do anything about it.

Karl's comment that it has to be a problem in the /bin/umount script prompted me to try and find out what the problem is. Yes, the bug is in that script, now fixed.

Note, this is an ancient bug, going right back to Puppy 4.x, whenever I first introduced the desktop drive icons into Puppy Unleashed.

Hmmm, I see that /bin/umount is still using 'fusermount' to unmount an ntfs partition. This is also ancient history, I think that ntfs-3g recommends not to use that anymore. Well, I will leave that one for another day, as I don't think anyone is reporting difficulty with unmounting ntfs partitions.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/df52b04e76

removepreview.sh, snapmergepuppy, rc.shutdown fixes

Shinobar improved script /usr/local/petget/removepreview.sh, for case of a file getting exposed in a lower Aufs/Unionfs layer when removed from the top. I have applied some of it into Woof.

Rerwin sent me a pm about some further changes to /usr/sbin/snapmergepuppy. I have applied rerwin's suggestion to screen out /.XLOADED so that it doesn't end up on the Flash drive save-file. I am leaving out rerwin's suggestion to screen out /dev/snd, as I want snapmergepuppy to mimic the behaviour of normal saves (on hard drive, that do not use snapmergepuppy), in which all of /dev is currently "saved". This is pending a review of the whole save of /dev concept.

Karl Godt posted some more fixes to /etc/rc.d/rc.shutdown, regarding to killing of stray processes and zombies. Applied.
I got this from: http://www.murga-linux.com/puppy/viewtopic.php?t=73122

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/a94d3c19f4

NTFS, Pmount, Partview, rc.shutdown, 10alsa fixes

In the Racy 5.2.2 Forum feedback thread, member otropogo reported error messages when copying files to/from an NTFS partition. Member shinobar replied that this is due to file permissions being lost in the transfer and the file is still copied ok. Shinobar showed a fix for /bin/mount to remove the error messages:
http://www.murga-linux.com/puppy/viewtopic.php?t=71767&start=375

...I have implemented shinobar's fix in Woof. Note that older versions of ntfs-3g do not have this problem, which is why it has occurred in recent puppies.

Forum member otropogo reported that the 'preferences' in Pmount is broken:
http://www.murga-linux.com/puppy/viewtopic.php?t=71767&start=375

I have fixed /usr/sbin/pmount. In the case of 'mut2', this is an alternative to 'probepart' and 'probedisk' that was developed by Jesse Liley. If missing, the preferences window now puts up text that the 'mut2' package is required.

mut2 has not been developed for a long time, and I don't know how well it works. It is not included in Puppy anymore.

otropogo also wrote this, after failing to get F-Prot working:

So then I tried to install clamav, which I've read Barry prefers to the "broken" f-prot. There's only one package in the repos clamav-0.91.2.pet, in the Puppy 3 repository. I installed it "successfully", but no tab or icon for it appears anywhere, and I have no idea how to access it to run it or update its sigs file.

Incorrect, I recommended Avast, for which there is a PET package for Wary/Racy:
http://bkhome.org/blog/?viewDetailed=02494

Karl Godt reported a problem with the 'sort' utility in Partview:
http://www.murga-linux.com/puppy/viewtopic.php?t=71767&start=390

...I have modified the Partview package, it is now 'partview-2.2-1.pet', in the 'noarch' repo.

...I should have read further down the thread, Karl determined the problem to be something else and posted a patch. I don't really know what problem is being fixed, anyway I have applied the patch. Rather, patch failed, did it manually.

In same thread link above, Karl Godt reported a bug in /etc/rc.d/rc.shutdown, unmounting stray partitions. Fixed.

In same link, Karl Godt posted a fix for unloading sound modules, in /etc/init.d/10alsa. Applied. Actually, patch failed, did it manually.

Woof commit:
http://bkhome.org/fossil/woof2.cgi/info/80b4545f72

Installing a pkg with symlink-to-directory

In the Racy 5.2.2 feedback thread on the Puppy Forum, member Tman reported:

ttuuxxx's gtkhash would not install for me with woof. I later discovered that his pet contains a .config/rox.sourceforge.net/SendTo folder that is not symlinked while, in recent puppies, that folder is symlinked to the .config/rox.sourceforge.net/OpenWith folder. This is what is causing the breaking of other Right-click options in Rox-Filer. I have been informed, by another forum member that the SendTo folder should not be a symlink. However, woof builds recent pups with this directory-structure.

The offending PET is here:
http://www.murga-linux.com/puppy/viewtopic.php?search_id=711057642&t=32700

Yeah, this is a problem. I did solve it in older versions of the Puppy Package Manager, but when I moved to support multiple types of packages, not just PETs, this problem came in.

The 'tar' extraction utility defaults to not following symlinks, which is what I want for files, but not for directories. However, tar has no option to distinguish between treatment of file or directory symlinks, not that I can see anyway.
Consequently, when ttuuxxx's PET is installed, it has a file in /root/.config/rox.sourceforge.net/SendTo which just gets installed as-is. The pre-existing symlink /root/.config/rox.sourceforge.net/SendTo is removed.

I have modified /usr/local/petget/installpkg.sh to fix this. It is a bit of a klutzy method of examining the directories after installation and finding if a symlink-to-directory has been overwritten.

Committed:
http://bkhome.org/fossil/woof2.cgi/info/58a1e67dd2

Still on holiday

Sorry, I am still not very responsive to Puppy issues when they are reported. I am still on holiday, living in a tent, and this is likely to continue for awhile.

The place that I am staying at does not have any Internet wireless signal, so I have to drive to the top of a nearby hill, which I do most days. On the hill, I prop my laptop in front of the steering wheel, and it is not really conducive to settling down for long sessions. So, I read my blog, email, and a few other things, and maybe download some things for working off-line later.

In a few days I hope to be at another place that will have a Telstra 3G signal -- in fact, that will be my main criterion for choosing where to pitch my tent next. The South West corner of Western Australia has forests of giant trees, very pleasant, and there is one lovely campsite where only a nominal fee is charged, set deep in a valley of giant Karri trees, but I have to rule that out for the next stop as there is no wi-fi signal (I think they are Karri's will have to confirm that).

The last few days have been overcast, with a bit of rain, so I have put away the solar panel and just using the large deep cycle battery to run light and laptop without recharging the battery -- probably would be good to get an isolator thingy so can charge the battery via the car alternator, but probably not committed enough to this camping scene to do that. With fairly light drain, the battery should hold out for several days, but I expect the sun to return soon.
But, perhaps if there was a simple isolator that just plugs into the cigarette lighter socket? -- um no, I think that I read somewhere that the wiring to the socket might not be able to handle the current.

Darn, the only downside to this camping is my back is having trouble with the camp bed -- will pickup another layer of mattress in the next big town. Camped next to me are a group of young French girls and boys, and I see them jumping around and crawling in and out of their tents with no trouble at all. Ah, to be young and flexible!

JWM reload menu without flicker

01micko reported that JWM now has the '-reload' option, that reloads the menu only, without restarting jwm itself. This avoids the screen flicker.

01micko reported on modifying scripts in the PPM to take out the "screen will flicker" message if JWM version is greater than 547:
http://www.murga-linux.com/puppy/viewtopic.php?p=590227#590227

Don't use 01micko's 'jwm-update-553.pet' as it has bugs -- I fixed it and posted the scripts to Woof:
http://bkhome.org/fossil/woof2.cgi/info/c90cc1a6f0

I compiled JWM version 562 in Wary, uploaded here (75KB):
http://distro.ibiblio.org/quirky/pet_packages-common/jwm2-562-w5c.pet

snapmergepuppy, ppm and whiteouts

I am looking at rerwin's "patch6", changes to script 'snapmergepuppy'. The changes of 111210 and 111220 I don't think are necessary -- these are changes to /dev/snd handling and /pinstall.sh, /puninstall.sh and /pet.specs files. However, I have applied the change of 111205, which fixes Jemimah's patch of 110212.

There has been some discussion about a problem with whiteout files when booting off Flash and installing/uninstalling packages. When a package is installed, /usr/local/petget/installpkg.sh installs the files direct to the "save layer" (on the Flash drive), but installpkg.sh then forces aufs/unionfs to perform a layer re-evaluation, which makes the files visible on top. There is also code that ensures no whiteout files on the top "ram layer" will hide the installed files underneath.

One the otherhand, package uninstall works on the top of the layered filesystem, and does not require layer re-evaluation. It does have code that makes sure that if an uninstalled file reveals an earlier file of same name (in an SFS layer), it becomes visible again (by the crude method of copying it to the top, rather than doing a layer re-evaluation).

It should not make any difference, that I can see, whether a session-save occurs before or after a package install/uninstall, or at shutdown. However, there is a situation that could occur...

If I install a package, a file of which replaces one of same name in underlying SFS layer, then I delete package, underlying file gets copied to top of layer f.s. If I then install same package or another, that file gets installed direct to save-layer, leaving older copied-from-SFS-layer file on top, thus the system sees the older file only.
I have fixed this situation in /usr/local/petget/installpkg.sh.

See Woof commit '75e3845e8b', 2011-12-29:
http://bkhome.org/fossil/woof2.cgi/timeline

snapmergepuppy launch

Thanks to playdayz who posted some links on the "disappearing sound" and other issues being worked on by rerwin and shinobar:
http://bkhome.org/blog/?viewDetailed=02623

Rerwin has done some in-depth analysis of some issues in Lucid Puppy, one of which is "disappearing sound" that some users have reported when running from a Flash drive. I have downloaded rerwin's "patch6" and have just started to take a look into it.

My first tentative step is to re-examine the timing of execution of 'snapmergepuppy' when a package is being installed or uninstalled...

When the Puppy Package Manager is running, the script /sbin/pup_event_frontend_d will not call snapmergepuppy, so you will never get a save of session during a package install or uninstall.

The converse is also true: if snapmergepuppy is in the midst of running when you decide to unstall a package, then the script /usr/local/petget/installpkg.sh will wait for the save to finish before installing the package.

One thing that seemed to be missing, is when you uninstall a package, there should also be a test to ensure that snapmergepuppy does not run in the midst of an uninstall operation. I have added code to /usr/local/petget/removepreview.sh to take care of this.

See commit 'c85e9aba0a' 2011-12-28:
http://bkhome.org/fossil/woof2.cgi/timeline

On the South Coast

I am currently camped on the South coast of Western Australia. The little coastal village must be one of the few remaining undeveloped holiday destinations left on the Australian coastline. The village is composed mostly of beach shacks, no electricity, no water, no telephone, no wireless phone or Internet.

The caravan park is run by the local Shire, only charges a nominal fee. They have a small shop, with solar and wind power, and rainwater tanks.

Near the village is a massive limestone bluff, at the top of which there is a Telstra wireless signal -- I get three bars on my phone, a good signal.

I have a 110ah deep cycle battery in my car, running the laptop off it. It is 7.25am, drove up to the top of the bluff. There's another guy parked beside me doing the same thing. Ha ha, we need our Internet fix!

Internet is incredibly slow though. I might try it late tonight.

On holiday

Guys, sorry that I am not responsive to posts. Have been offline a few days, currently living at a caravan park and have a site without power -- Right now, running my laptop on it's battery. the Telstra 3G wi-fi module is also running on it's own battery.

It doesn't cost much more for power, might do so tomorrow.

Anyway, I'm having fun living in a tent!

Puppy running on Pandaboard (ARM)

Forum member dim-kut has beat us all to it! He has got PuppyRus (Russian Puppy) running on a Pandaboard, which has a Cortex-A9 ARM CPU:
http://uco.puppyrus.org/novosti/portirovanie-na-arm-zagruzka-konsoli.html

Dim-kut is using a 3.0.9 kernel, U-boot, and booting Puppy off a 2GB SD card.

So far, it is booting to the commandline, no X.

I posted about the Pandaboard recently:
http://bkhome.org/blog/?viewDetailed=02601

Disappearing sound?

Bigpup posted to this blog about "disappearing" sound in Lucid 528+ and Slacko 5.3.1, when booting off a USB Flash drive:
http://bkhome.org/blog/?viewDetailed=02621

Report of disappearing sound in Slacko:
http://www.murga-linux.com/puppy/viewtopic.php?p=588564#588564

Reports in Lucid start here:
http://www.murga-linux.com/puppy/viewtopic.php?t=70855&start=585

Bigpup has suggested that it might be a problem with 'snapmergepuppy', the script that handles saves from RAM to the save-file on the Flash drive.

If so, then the problem would apply to Wary and Racy also. So, I installed Racy 5.2.2 to a 1GB USB Flash pen drive, and I have been using it for awhile. Trying all sorts of things, installing packages, clicking 'save' button on the desktop, powering-off, rebooting. Have restarted about a dozen times so far, sound continues to work flawlessly.

Running on my Acer Aspire 5742G laptop.

Some points to note:

1. Alsa Wizard: I did not have to run this.
2. Shutdown without saving. ???

Regarding point 2, sylvander reported (http://www.murga-linux.com/puppy/viewtopic.php?t=70855&start=600):

Opened and closed retrovol...[OK]
Browsed some web pages...[OK]
Manually saved the session...[OK]

Rebooted, and chose to not save the session [already saved]...[OK]

Once back into the desktop, the speaker icon was GONE! Sad

The thing is, booting from USB, save session, reboot, how is it possible to then choose not to save session? Does Lucid have some extra tool/facility to shutdown without saving the session from RAM to save-file?
If so, this is not a standard Woof feature, and I suggest that you guys look into that.

Note, the Puppy Event manager does have a save-interval value, if changed to 0 then won't save periodically, but will still save at shutdown.

There was a problem with "disappearing" sound some time ago, which I fixed. Essentially, it involved deleting /dev/snd at shutdown, I think that is done in /etc/rc.d/rc.shutdown. If Lucid (and/or Slacko) has any other shutdown mechanism that does not perform that operation, that would be the cause of the problem.

rc.shutdown also does some other correct shutdown of sound drivers.

Gold metal detector

Ha ha, better late than never! I wrote about wanting to go off and do a spot of prospecting back in 2007 (http://puppylinux.com/development/project-statement.htm):

Possibly the hectic release schedule will slow down sometime, after all there are other things in life -- I want to go off fossicking soon (gemstones, gold).

Heh, heh, it is finally happening. I have bought myself a cheapie metal detector:

link removed, on request from dealsdirect.com

...they offered buy all you want for $10 total postage & handling, so I got a few other camping items as well.

The "outback" of Western Australia has a lot of iron in the ground, making gold detecting very difficult. It is possible to spend a lot of money, for example the top-of-the-range gold detector here in Australia is the Minelab GPX 5000, with rrp of AU$6,595 (about the same in US$):
http://www.minelab.com/aus/products/gold-detectors/gpx-5000

I just enjoy being in nature, finding interesting rocks, plus it is healthy exercise, so I will be quite happy fossicking around with an ultra-cheapie detector. My $79 detector has the controls that I think are needed for detecting in the goldfields -- two discrimination modes, ground balance, plus various other dials, switches and buttons to play with. I have assembled it, and it looks quite well made. Hopefully the internal electronics is "up to it".

Slacko 5.3.1

Slacko Puppy 5.3, our flagship Puppy, premiered on October 25, 2011:
http://bkhome.org/blog/?viewDetailed=02556

There are many flavours (builds, editions) of Puppy Linux, but "Flagship" means basically that we recommend this pup as a "first-stop" for newcomers (unless you have older hardware and might want to go for our Wary Puppy, or some particular need that attracts you to one of our other Puppy builds).

Slacko has had a very good reception, but as always there are things that can be improved, and Project Coordinator Mick Amadio has now released Slacko 5.3.1, a minor upgrade and bugfix release.

Here is the release announcement:
http://distro.ibiblio.org/pub/linux/distributions/puppylinux/puppy-5.3.1/release-Slacko-5.3.1.htm

Here is the announcement on the Puppy Forum:
http://www.murga-linux.com/puppy/viewtopic.php?t=72806

Download from here:
http://distro.ibiblio.org/pub/linux/distributions/puppylinux/puppy-5.3.1/
Fast mirror in Europe:
http://ftp.nluug.nl/ftp/pub/os/Linux/distr/puppylinux/puppy-5.3.1/
Mirror in Australia:
http://mirror.internode.on.net/pub/puppylinux/puppy-5.3.1/

Here is a Forum thread for bug reports and general feedback. Discussion specific to 5.3.1 starts on page 23:
http://www.murga-linux.com/puppy/viewtopic.php?t=72805&start=330

...please report any issues there, also read that thread for further fixes/improvements.

Script internationalization by 't12s'

This is a brilliant technique developed by technosaurus and implemented for Puppy by L18L.

Forum discussion:
http://murga-linux.com/puppy/viewtopic.php?t=73440

Online documentation on how to do a translation:
http://extra-inter.net/puppy/t12s_DOC/t12s.html

I have put version 0.4.5.1 of 't12s' into /usr/sbin in Woof. I also created a help file /usr/share/doc/HOWTO-internationalization.htm that explains the three methods -- BaCon code, using 'gettext' in scripts, and using 't12s' in scripts.

The above two files are in Woof, which has been uploaded. Timeline:
http://bkhome.org/fossil/woof2.cgi/timeline

...do an anonymous login and you can view the source of 't12s' and also read the 'HOWTO-internationalization.htm' file rendered as a web page (not source code) -- a nice feature of Fossil.

I was going to put 'xdelta_gui' that L18L had created, that has the 't12s' method of translation, however when I tried to run it, it output this error message:

# ./xdelta_gui
./xdelta_gui: line 23: : command not found

...the application works, but that error worries me, so I did not put it into Woof.

So, so far there are no scripts using 't12s' method in Woof.

LibreOffice 3.4.4

I got the SFS from gray's NOP Puppy site, and modified it slightly -- fixed icons in menu, added desktop icons.

Added to the SFS collection here (129.2MB):
http://distro.ibiblio.org/quirky/sfs/libreoffice-3.4.4.sfs

I have only tested it with Racy 5.2.2, works ok.

GTK "gseal" fix

Wary and Racy 5.2.2 users have reported applications failing to compile, with "gseal" error messages.

I had gone through a rather tiring process of getting GTK to work properly prior to releasing 5.2.2, and didn't feel like getting into that again. Here are some of the pre-5.2.2 deliberations:
http://bkhome.org/blog/?viewDetailed=02594
http://bkhome.org/blog/?viewDetailed=02595
http://bkhome.org/blog/?viewDetailed=02598

Well, the "gseal" problem needed to be fixed. I have rolled back to GTK 2.24.5, compiled with this:

# ./configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var --build=i486-t2-linux-gnu --with-xinput=yes --enable-debug=minimum --enable-introspection=no

After installing it, I tested 'quicksetup' and the dropdown lists are fast (that was a big problem before). I compiled gtktetris, that ttuuxxx had reported gave the "gseal" errors -- now it compiles.

Here are the PETs (2.3MB, 266KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/gtk+-2.24.5-1-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/gtk+_DEV-2.24.5-1-w5c.pet

The next releases of Wary and Racy will be built with these.

Menu icon when install a package

When a package is installed, and it has a menu entry, in some cases the icon is absent. Fixed.

This came to my attention yesterday when I installed the Parole media player .deb (Debian Squeeze) package, and an entry appeared in the Multimedia menu category without an icon.

This problem appears in two situations, when building a Puppy in Woof, and when installing a package in a running Puppy.

In the former case, I have fixed script '2createpackages'.

In the latter case, I have fixed script /usr/local/petget/installpkg.sh.

You can see the changes in the online Woof timeline, commit '6fbecf9dca' (2011-12-07):
http://bkhome.org/fossil/woof2.cgi/timeline

Woof: Dpup Debian Squeeze progress

This post is some notes on what I'm currently doing.

I am using Woof2 to build an x86 Dpup, and using it every day to discover and fix any issues.

I am not doing anything fancy, just a build that uses as many Debian Squeeze .deb packages as possible rather than my own PETs -- getting ready for when I do an ARM-target build.

I know that there are others building Dpups. My Dpup would be a good starting point for other Dpup developers who will substitute many of their own PETs for more compact builds.

I do not intend to develop my x86 Dpup to an official release Puppy. The other guys who are building Dpups can do that (and have done). But, for the curious, I might upload one of my Dpup builds soon.

I am discovering some interesting things, for example the Debian Squeeze 'tar' utility gives error reports when it should not, so I replaced it with my tar from Wary.

Something that I will be looking at very soon is how developers can contribute to Woof2.

Woof: build Slacko 5.3

When 01micko released Slacko 5.3, he also published some instructions to fix Woof to build Slacko 5.3 exactly as his release version:
http://murga-linux.com/puppy/viewtopic.php?t=72806

I have only just now got around to putting those fixes into Woof, well Woof2 now. If you go to the Fossil timeline and do an anonymous login, you will be able to see the changes:
http://bkhome.org/fossil/woof2.cgi/timeline

There are a couple of differences in my upgrade:

1. lsmod
I left out 01micko's fixup to move lsmod from bin to sbin, as I cannot see anything wrong with my module-init-tools template. However, I then discovered a bug in 2createpackages -- the lsmod problem should now be fixed, see last commit (c49450a62d).

2. slacko-splash.gif
This already exists at woof-code/boot/boot-dialog/. 01micko has it in support/slacko/.

I have not done a build to test that these changes are sane. It will actually build 5.3+, as I have put in the latest Packages-puppy-slacko-official.

Woof: Arch Linux patch -2

Forum member noryb009 sent me a '-2' patch for Arch Linux support in Woof. This is patching file '0setup'.

When I apply changes to Woof, I am "pushing" it online frequently, so won't announce it every time as I did before. If you look at the timeline, you can see the latest commit that has been pushed:

http://bkhome.org/fossil/woof2.cgi/timeline

Midori crashes

I reported earlier that I was unable to compile Webkit, needed for Midori web browser:
http://bkhome.org/blog/?viewDetailed=02577

Jemimah replied that where it was failing is a utility application that can be removed from the makefile. Yes indeed, I did that, and it compiled. Then I compiled Midori 0.4.2.

At startup, Midori partly loaded the Google home page, then crashed.

Weird, Midori is loading libraries inside /initrd/pup_rw/usr/lib/ -- that is so incredibly weird, as that is definitely not in the library search path. It must be something peculiar that the 'waf' build system does.

Jemimah has also compiled Midori 0.4.2, for Racy 5.2.2:
http://murga-linux.com/puppy/viewtopic.php?t=73799

...one tester so far, reports it crashes.

So much for Midori.

Woof2 now online

I announced this in the previous blog post:
http://bkhome.org/blog/?viewDetailed=02611

I did a test download and build a x86 Dpup, it worked, so it all looks sane. I have modified some scripts since then, but hopefully it is still sane.

Woof2 is hosted in a Fossil Version Control System (or Content Management System):
http://bkhome.org/fossil/woof2.cgi/index

This wiki page explains how to download and install Woof2:
http://bkhome.org/fossil/woof2.cgi/wiki?name=Getting+Started+with+Woof

If you login as an 'anonymous' user, you can click on 'Files', then 'README-FIRST' to read some introductory notes about Woof2.

Woof2: Fossil is back

Those who have been around for awhile might recall that I had a brief fling with Fossil, a simple project Version Control System (VCS), toward the end of 2009.

Fossil has some limitations, that I managed to work around, however the lack of support for symlinks was too difficult to work around. At that stage, I developed my own very basic Version Control System named Bones.

However, Fossil now supports symlinks, which prompted me to reconsider using Fossil.

Woof2 is a complete re-think of Woof, with two main advantages over Woof1:

1. Supports multiple architectures (exs: x86, arm)
2. VCS (Version Control System) friendly

It is a work in progress, but I have now got it online with Fossil. Finally, we will have a bug-reporting system, that comes with Fossil!

I won't give out the URL to access it just yet, as it needs some more sanity testing. Hopefully very soon though.

Kernel for RaspberryPi

I am gradually getting things together for building a Puppy that will run on the upcoming RaspberryPi. Toward that end, I have compiled a kernel. It is version 2.6.35.14, with a special patch for the RaspberryPi. It also has the usual Puppy-patches, and Unionfs patch.

I have create a PET and uploaded it here (5.3MB):
http://distro.ibiblio.org/quirky/arm1176jzf-s/pet_packages-squeeze/linux_kernel-2.6.35.14-armv6-raspi.pet

I have uploaded the source and patches here:
http://bkhome.org/sources/kernel-2.6.35.14-raspi/
Us#er#na#me: pu#pp#y Pa#ss#wo#rd: li#nu#x

...that link contains 'bk.txt' that explains where I got the patches from, and how the kernel was cross-compiled.

I think that the kernel .config file needs more work. It is provided by the patch from raspberrypi.org, and I modified it slightly. However, there are a lot more drivers that could probably be enabled, to cater for things being attached to the RaspberryPi board.

I am still debating with myself whether the RP board will be anything more than a toy. For most of us that is. Definitely it can be used in a wide range of embedded applications. For more general use, we may find ourselves adding on so much external stuff (with difficulty due to very limited I/O features) that it might have been better to go for a more feature-rich board to start with.
But, it is going to be a lot of fun working within the constraints and see what we can achieve.

I am personally quite happy with the HDMI output. My aging CRT TV is going to be retired and I will buy a small LED-LCD TV with HDMI input. They are getting cheaper, $150 - $200.

Alternate Puppy forums

Our main Puppy Linux Forum is "down". It may be awhile before it is restored. In the meantime, Linux Questions also have a Puppy Forum:

http://www.linuxquestions.org/questions/puppy-71/

Also, Raffy has setup another forum:

http://puppylinux.info/

Woof: improved Arch support

Forum member noryb009 sent me a patch for Woof, to improve support for Arch Linux. Here is the pm from noryb009:

Over the last few days, I've been playing around with woof and archlinux. I changed a few scripts for better archlinux support.

Changes:
- support for .pkg.tar.xz, while keeping .pkg.tar.gz support (some packages are still .pkg.tar.gz for backwards compatibility)
- better reading from database files - supports split packages
- changed some repository locations, I couldn't connect to a few of the previous ones, and a few of the previous ones were really old snapshots
- added support for arch .INSTALL files (arch's .pinstall.sh)
- changed/removed some renamed packages in DISTRO_PKG_SPECS-arch

With the patched DISTRO_PKG_SPECS-arch, you can make an ISO, but X doesn't start for some reason.

That last point will need attention. noryb009 has modified the package selection. Back when I first created Arch support in Woof, X did start.

I have applied the patch, and it will be in the next upload of Woof.

Gtkterm rs232 serial terminal

I posted about the Cutecom GUI serial terminal in the previous blog post:
http://bkhome.org/blog/?viewDetailed=02606

There is a later version of Cutecom, that requires Qt4:
http://cutecom.sourceforge.net/

There is also a GTK-based GUI serial terminal, named Gtkterm. I discovered that it is still being developed. It needs VTE. Site:
https://fedorahosted.org/gtkterm/

I tried to compile these apps statically against Qt4 and VTE, but failed. Probably it could be figured out. In the latter case there were undefined symbols, but applying what I thought were logical steps to fix it, did not. So, left it and just compiled Gtkterm again shared VTE library. Here are the PETs:

Gtkterm (37KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/gtkterm-0.99.7-rc1-w5c.pet

VTE (356KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/vte-0.28.2-w5c.pet

Serial link to OK6410-B

Some notes, as I progress. I have found that trying to apply the instructions for the FriendlyARM mini6410 board to my OK6410-B board is not smooth sailing.

Anyway, I have the serial connection with the board working. The board has a RS232 port. My laptop does not have a RS232 serial port, however the board is supplied with a (separate) "USB2.0 to RS232 Convertor" -- and it is truly a no-name brand, no manufacturer identification anywhere.

When the board starts up, it outputs to the serial port. For example, if I were to get Linux booting, then the bootup log would go to the serial port. It is, effectively, the operator console.

But, this requires a RS232 serial communications application running in Linux on my laptop. Wary and Racy have 'microcom', a commandline application.
However, there is a GUI app called 'cutecom', that requires Qt 3.x. In the Package Manager, you will find this in the 'puppy-3' repository.
When it installs Qt, some libraries are reported missing (libmysqlclient*) -- just ignore that.

The 'pl2303.ko' kernel driver is required, and this will load automatically when the usb-rs232 adaptor is plugged in. However, I had to manually load 'usbserial.ko':

# modprobe usbserial

Then, run 'cutecom' and type "/dev/ttyUSB0" into the interface-port box, "115200" into the baud-rate box, then click the "Open port" button.

Now we are in business. Turn on the OK6410 board, and bootup messages should appear in the cutecom window.

Debian, Dpup, mini6410

I think that I have decided "which way to go" for getting Linux running on my OK6410 ARM board: Debian Squeeze.

Despite having downloaded 12 Debian Squeeze i386 CDs, I don't think that I will install Debian. There is no need, when we have Dpup.

So, I am running Woof to create a Dpup of Debian Squeeze 6.0.3. This is i386, to install on my laptop. I am doing this partly to check that Woof can still build a Debian Squeeze puplet.

The ARM thing though, is something else...

To create a small Debian system for the Flash card in my OK6410 board, I am following the instructions posted for the mini6410 board, here:
http://code.google.com/p/mini6410-debian/wiki/Tutorial

The OK6410 and mini6410 look very similar, so I anticipate the instructions can be applied to my OK6410 board without too much difficulty.

The tutorial starts by requiring the 'debootstrap' DEB, which is here:
http://packages.debian.org/squeeze/debootstrap

I am running on Racy 5.2.2, and debootstrap seems to have worked nicely.

Next step, I have to go out and buy a couple of SD cards...

Debian: 52 CDs

I am visiting my daughter. They have super-fast Internet, so I am taking advantage of the situation, downloading anything that I can think of.

I was thinking of installing Debian Squeeze on my laptop, so went to the download site... it is fifty two CDs! Yes, 52.

Getting them from an Australian mirror, they download at about 1 MB/sec, very fast:
http://mirror.cse.unsw.edu.au/debian-cd/6.0.3/i386/iso-cd/

But, I doubt that I will need all 52. I have stopped at 12.

There is online support to find out which CD has a file:
http://cdimage-search.debian.org/

Woof uploaded, Nov. 18, 2011

This is commit '20111118155739', used to build Wary and Racy 5.2.2.

The last upload of Woof was Nov. 10:
http://bkhome.org/blog/?viewDetailed=02591

There was a blog post awhile back on how to download a recent snapshot tarball of Woof and use that as the reference to download Woof changes since then, without the full history:
http://bkhome.org/blog/?viewDetailed=02306

I have uploaded a snapshot tarball of commit '20111118155739'. This can be downloaded by itself, but it is recommended to also download the Bones infrastructure as described in above link. Get tarball:
http://bkhome.org/bones/woof/archive/woof-20111118155739.tar.gz

To download and use Woof, please read the Bones intro page:
http://bkhome.org/bones/
And the Woof intro page:
http://bkhome.org/woof/

Wary Puppy 5.2.2

I am pleased to announce the release of Wary Puppy version 5.2.2, our build of Puppy Linux that targets older hardware. This is a minor upgrade and bug-fix release of Wary 5.2. Though, if you look in the release notes you will might not think it so "minor".
This is also the debut of Racy Puppy, which we think of as "Wary on steroids". Racy enhances Wary to run on recent hardware, with Xorg 7.6 and 3.0.7 kernel.

Full announcement and release notes:
http://distro.ibiblio.org/quirky/wary-5.2.2/release-Wary-5.2.2.htm

Download Wary (130.5MB):
http://distro.ibiblio.org/quirky/wary-5.2.2/

Download Racy (111.6MB):
http://distro.ibiblio.org/quirky/racy-5.2.2/

Forum thread for Wary feedback:
http://murga-linux.com/puppy/viewtopic.php?p=583947#583947

Forum thread for Racy feedback:
http://murga-linux.com/puppy/viewtopic.php?p=583948#583948

Pandaboard, Beaglebard, Hawkboard

I received an email from a guy who is working on porting Puppy to a Pandaboard. Dougal also suggested the Pandaboard might be a good way to go.

But, just what is the Pandaboard, and what are these others, Beagleboard, Hawkboard?

Pandaboard
http://www.linuxfordevices.com/c/a/News/PandaBoard/
http://www.linuxfordevices.com/c/a/News/PandaBoard-runs-ICS/?kc=rss
http://www.pandaboard.org/

Beagleboard
http://www.linuxfordevices.com/c/a/News/BeagleBoardxM/
http://beagleboard.org/

Hawkboard
http://www.hawkboard.org/

In terms of performance, it would seem to be in order Hawkboard, Beagleboard, Pandaboard, with latter being fastest. The Pandaboard has a 1GHz dual-core Cortex-A9 CPU, the Beagleboard-XM has a 1GHz single-core Cortex-A8 CPU.

CPU specs:
http://www.arm.com/products/processors/cortex-a/cortex-a8.php
http://www.arm.com/products/processors/cortex-a/cortex-a9.php

I have just noticed something on the Cortex-A8 page "binary compatibility with the ARM1176 family". That would be backwards compatibility.
Assume that also applies to the A9. That would mean I could compile code optimised for the RaspberryPi on a Pandaboard or beagleboard?

CodeSourcery experiment

I want to get hands-on experience with CodeSourcery and Buildroot, even though it might not be what I ultimately use for development. Documenting here...

CodeSourcery is a toolchain for cross-compiling and debugging:
http://www.mentor.com/embedded-software/sourcery-tools/sourcery-codebench/

The free Lite edition is at this link. I grabbed the "GNU/Linux" download:
https://sourcery.mentor.com/sgpp/lite/arm/portal/subscription?@template=lite

I installed by following these instructions:
http://www.arm9board.net/wiki/index.php?title=Codesourcery_toolchain

Except that I added the installation to the PATH variable in /etc/profile, then rebooted. The toolchain uses GCC 4.5.1.

Wary and Racy 5.2.2 "TESTING"

If anyone is interested in conducting a "sanity check", here they are:

http://distro.ibiblio.org/quirky/racy-5.2.2TESTING/

http://distro.ibiblio.org/quirky/wary-5.2.2TESTING/

About several hours from now I intend to remove the "TESTING" from the directory names and formally announce the releases.

'welcome1stboot' fixed

A few days ago ...14th?... I had Wary and Racy at what I thought was release quality. Then I found that 'welcome1stboot' would not start after I had upgraded GTK from 2.24.5 to 2.24.8.

(On the other-hand, other testers reported that welcome1stboot did work.)

So, I played with rolling GTK back to 2.20.1.

But, just now I have discovered why welcome1stboot does not work. I am both relieved and very exasperated.

'welcome1stboot' is written in BaCon. Now, BaCon does not link the GTK libraries as shared libraries in the normal manner. Instead, BaCon has it's own code to load the libraries. This is where the problem occurs -- and it sure is awful, I consider it to be a serious bug in BaCon.

Wary/Racy has these in /usr/lib, where the first is a symlink:
libgtk-x11-2.0.so.0 libgtk-x11-2.0.so.0.2400.5

When the 'devx' is loaded, there is also the symlink:
libgtk-x11-2.0.so

Now this is the awful part: at runtime, the Bacon app tries to load 'libgtk-x11-2.0.so'. It cannot, so crashes. There is no error message identifying the cause of the crash.

It is standard, engraved in stone, practice that *.so files are used during compiling, not at runtime. The BaCon application should really be using libgtk-x11-2.0.so.0.

God, I have spent days working on this problem, on the verge of "developer fatigue". Oh well. I will put the *.so symlinks into the gtk PET, and hopefully should be able to put up another Release Candidate of Wary/Racy.

Wary (and Racy) 5.2.1.91

I tried an experiment. Although it was the "wrong" thing to do, I built Wary/Racy 5.2.1.91 with the GTK 2.20.1 PETs from Wary 5.1.x. The reason that it is "wrong" is that the core packages in 5.2.x were recompiled in T2, with GTK 2.24.5 and some later libraries (ex: pixman, cairo, glib). However, as reported, there are problems with GTK 2.24.x. To roll GTK back by just plunking in the PETs from Wary 5.1.x is just asking for trouble.

Or so it would seem. I built it, and everything seems to work. Well, I haven't checked printing, that is one area where there might be some difference between the GTK versions.

Terryphi posted about "tester fatigue". Yes, I am getting "developer fatigue" with these troubles with GTK. I really want to get Wary and Racy 5.2.2 released, then move on to other stuff -- multi-architecture Woof and ARM Puppy. So, give 5.2.1.91 a run, if it looks ok, I'll go with that for the final.

I will be off-line for approximately the next 48 hours, maybe slightly less. Starting from about now. Racy 5.2.1.91 has been uploaded, Wary is still uploading, eta 15 minutes from now:

Racy: http://distro.ibiblio.org/quirky/test/racy-5.2.1.91/

Wary: http://distro.ibiblio.org/quirky/test/wary-5.2.1.91/

As soon as the upload has completed, I will be going off-line. Got other family-related stuff to do for a couple of days.

Please go here to provide Wary feedback:
http://murga-linux.com/puppy/viewtopic.php?t=72258&start=255

Please go here for Racy feedback:
http://murga-linux.com/puppy/viewtopic.php?t=71767&start=285

ARM11 board has arrived

The OK6410 board that I ordered from here has arrived:
http://www.aliexpress.com/product-fm/503734975-Free-Shipping-ARM11-OK6410-S3C6410-Android-2-3-4-WinCe-7-0-development-board-4-3-wholesalers.html

It comes with 14 CDs, including with Ubuntu, WinCE and Android. There is a catch though, most of the documentation is in Chinese. For a product that is sold internationally, and advertised in English, this is very disappointing. The main 178-page PDF file for the board, that introduces the board and explains how to use it, is in Chinese.

I looked at the website of the manufacturer of the board, but that also is Chinese only:
http://www.witech.com.cn/

So, I am going to also purchase one of the "FriendlyARM" boards, which have all documentation in English. I know this as the docs are available online. These boards are very similar, in fact I should be able to apply the same techniques and get both boards working.

Now, I am going to have to think why I would want two boards. Testing a LAN maybe.

Actually, what I will do first is download the FriendlyARM documentation and see if I can apply it to the board I already have. If so, then I might not need to order the FriendlyARM board.

GTK trouble

See my previous post "GTK success!":
http://bkhome.org/blog/?viewDetailed=02594

Yeah, well, now the pendulum has swung the other way. I built Wary and Racy 5.2.1.99, even uploaded them. However, after doing a frugal install, I found that 'welcome1stboot' segfaults.

I recompiled it in Wary 5.2.1.99, it then worked. I used that new executable in Racy 5.2.1.99 ....aaaargh, it still segfaulted!

I regret moving up from GTK 2.20.1. With that version the ROX-Filer focus problem that came in at GTK 2.18.x, was partly fixed. Well, with GTK 2.24.x that "part fix" is gone again.

I wonder if it would break anything if I did roll back to GTK 2.20.1? The problem is, I did a complete compile of all the core packages in T2, against GTK 2.24.5. Rolling GTK back would be asking for trouble.

GTK success!

The guys testing the latest Wary and Racy have reported that large drop-down lists are very slow and hog the CPU. This has been observed in the 'quicksetup', 'bootmanager' and 'remove_builtin' scripts.

I was going to recompile GTK anyway, due to the insert-key-crash bug in SeaMonkey/Firefox. This problem has been reported with recent Wary/Racy, after I upgraded to GTK 2.24.5 (compiled in T2). The bug is documented in an earlier blog post:
http://bkhome.org/blog/?viewDetailed=02253

So, I compiled GTK 2.24.8, with this config;

# ./configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var --build=i486-t2-linux-gnu --with-xinput=yes --enable-debug=minimum

Once again, a big jump in size for the PET, from 1621KB to 2588KB. Once before, I think that I isolated which library was the culprit, then recompiled all the others without debugging, and that worked. Will have to put that on the to-do list.

Anyway, a very pleasant surprise. Running this new GTK, the slowness and CPU-hogging of large drop-down lists has gone away. Now they are snappy. at least this is the case on my laptop.

If anyone would like to confirm this (2.5MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/gtk+-2.24.8-w5c.pet

Flash player download broken

Woof has semi-automatic download of the Adobe Flash player (script /usr/sbin/delayedrun). In Slacko, 01micko wrote 'getflash'.

However, both are broken. The download URLs for the 10.x and 11.x players are both broken.

Does anyone know a URL for downloading the latest 11.x player?

Script delayedrun has this:

cd /usr/lib/mozilla/plugins
download_file http://fpdownload.macromedia.com/get/flashplayer/current/install_flash_player_10_linux.tar.gz

In the forum, getflash was changed to download the 11.x latest, using this URL:

http://fpdownload.macromedia.com/get/flashplayer/current/install_flash_player_11_linux.i386.tar.gz

However, neither URL works anymore.

Planner project manager

Our master of hacking and doing extremely clever stuff with code is technosaurus, and I continue to be surprised by the innovative and interesting things he does. Like this latest...

technosaurus has hacked Planner, project manager software, to compile without Gnome dependencies. This makes it small enough to be included in Puppy.
Project management has always been a gap in the Puppy application suite, and for a very long time I have been on the lookout for something to fill this gap.

Anyone interested in project management, please try this Planner PET. If it gets the thumbs up, I will put it into the next Wary and Racy. And maybe other devs might put it into their builds also.

We will of course have to get technosaurus to post the hacked source, which I will upload to the official source repo, to fulfill our legal requirement.

Forum link:
http://murga-linux.com/puppy/viewtopic.php?t=73371

Woof uploaded, November 10, 2011

This is commit '20111110213851', used to build Wary and Racy 5.2.1.90 (5.2.2 Release Candidate).

The last upload of Woof was October 27:
http://bkhome.org/blog/?viewDetailed=02571

There was a blog post awhile back on how to download a recent snapshot tarball of Woof and use that as the reference to download Woof changes since then, without the full history:
http://bkhome.org/blog/?viewDetailed=02306

I have uploaded a snapshot tarball of commit '20111110213851'. This can be downloaded by itself, but it is recommended to also download the Bones infrastructure as described in above link. Get tarball:
http://bkhome.org/bones/woof/archive/woof-20111110213851.tar.gz

To download and use Woof, please read the Bones intro page:
http://bkhome.org/bones/
And the Woof intro page:
http://bkhome.org/woof/

Wary/Racy 5.2.2 Release-Candidate

I have uploaded both Wary and Racy, with same version numbers, 5.2.1.90, which I consider to be 5.2.2 RC.

They are pretty good. I was tempted to bump them immediately to release versions. But, decided it would be wise to release an RC first. Anyone who wants to give these guys a run, please do.

Of course, if it turns out that these are pretty perfect as-is, then you can consider that you have got the release, then no need to download the actual 5.2.2 when I release that!

Anyway, here they are:

Wary: http://distro.ibiblio.org/quirky/test/wary-5.2.1.90/

Racy: http://distro.ibiblio.org/quirky/test/racy-5.2.1.90/

Wary is built with 2.6.32.45 uniprocessor kernel. Racy is built with 3.0.7 SMP (multiprocessor) PAE kernel. Also, the former uses Aufs, the latter Unionfs.

Grub4dos, Pmusic, Psync

I have upgraded these PETs in the 'common' repo:

Grub4dosconfig 0.4.4v1.7.1 (shinobar)
Pmusic 2.2.1 (zigbert)
Psync 2.7 (tasmod)

I have suddenly realised that I am ready to create RC's for the next Racy and Wary. Since the release of Wary 5.2 there have been a lot of bug fixes and incremental improvements. I need to bring out 5.2.2 very soon, however I will do an RC first.

Encrypted save-file fix

Encrypted save-files are partly broken in the latest Wary/Racy/Slacko.

I did a first shutdown in Racy and created a save-file with light encryption. I then opened it, and found that nothing had been saved in it.

I have extended /usr/sbin/filemnt so that it now mounts encrypted .2fs files (save-files). That's one useful outcome.

I examined /usr/sbin/shutdownconfig, and found that it was not exporting variable CRYPTO in file /tmp/shutdownconfig_results, needed by /etc/rc.d/rc.shutdown. Consequently, the encrypted save-file was not getting mounted at first shutdown, so the session was not saved. Fixed.

Racy bugs fixed

I haven't been lurking on the Puppy Forum much lately, but have been gradually reading through the Racy feedback thread, and fixing bugs. Read from page 7 of thread:
http://murga-linux.com/puppy/viewtopic.php?p=576964#576964

These are mostly Woof-level fixes, so will apply to all future puppies.

I have reached the end of that thread (page 16), and now will go on to reading the Wary 5.2.1 feedback thread.

Busybox 1.19.3

I have upgraded to the latest Busybox, PET uploaded to the 'common' repo (394KB):
http://distro.ibiblio.org/quirky/pet_packages-common/busybox-1.19.3-w5c.pet

The source and .config file are at my source repo:
http://bkhome.org/sources/alphabetical/b/

technosaurus has made a number of suggestions to reduce the size of the 'initrd' (initramfs):
http://www.murga-linux.com/puppy/viewtopic.php?t=73045&sid=1305fbcaf56f7d2285772a32328f55d3

One of those suggestions is the replacement of the 'guess_fstype' utility with the Busybox 'blkid' applet. The extra capability of blkid must be turned on at the configure stage, which I have done with this latest Busybox PET. Here is the commit detail:
http://git.busybox.net/busybox/commit/?id=90615a0c5c326fa3cf78fc719f7b16207f47395a

Yeah, it works nice:

# blkid /dev/sda9
/dev/sda9: UUID="9d1e4211-578a-45d3-9227-24acc4cbc8a8" TYPE="ext3"

I intend to work on implementing technosaurus's suggestions in Woof.

ARM architecture
I have used Buildroot to compile Busybox statically with uClibc, both for the arm1176jzf-s and i486 CPUs, for inclusion in the initrd in Woof.
I am gradually getting Woof setup to build an ARM Puppy (RazPup?). Compiles are otimised for the arm1176jzf-s CPU.

Sound Wizard improved

This seemed like a good thing to do:

http://murga-linux.com/puppy/viewtopic.php?p=580303#580303

all-firmware now directories

The so-called "firmware tarballs" in /lib/modules/all-firmware are no longer tarballs. They are expanded directories.

I have done this as they contain binary executables. These need to be separated out, as Woof moves toward supporting multiple architectures.

A side benefit is that the .iso is likely to be smaller, as Squashfs can better compress the firmware directories when they are not already compressed.

This required a small modification to script rootfs-skeleton/sbin/pup_event_backend_modprobe.

SoC Broadcom BCM2835

The RaspberryPi has a most intriguing Soc (System on Chip) -- CPU plus GPU plus more on the one chip. It is a Broadcom BCM2835, and has the ARM1176JZF-S CPU running at 700MHz.

Raffy recently posted a link to a PDF for the CPU:
http://infocenter.arm.com/help/topic/com.arm.doc.ddi0301g/DDI0301G_arm1176jzfs_r0p7_trm.pdf

However, the GPU remains a mystery. I can't find any details specs on-line, only this:
http://www.broadcom.com/products/BCM2835

Here is another board that has the BCM2835, the Roku2:
http://www.mycablealternatives.com/2011/07/roku-2-xs-teardown/

I would love to locate Xorg driver source for the GPU. I presume that the RaspberryPi people have such a thing, and will release the source.

T2 cross-compiling

I took off a couple of days to do some house renovating. Apart from that, I have been experimenting with doing a cross-compile build in T2.

So far, I have only used T2 in the "normal" way, which is to compile the source packages on a x86 PC, for an x86 PC. However, compiling on a x86 PC for an ARM CPU board is a "cross-compile", which is very different, and much more difficult for T2 -- T2 can only compile a sub-set of it's package collection.

What I soon discovered, after hitting many compile failures, is that it has been awhile since anyone has done a cross-compile with T2. I have been reporting bugs and supplying fixes to the T2 email list:
http://news.gmane.org/gmane.comp.t2.devel

...I have been a busy boy!

In the T2 development system, there is a package/*/*/*.desc file for each package, and this has an entry that states whether the source can be cross-compiled. When I say "can", it is a matter of the T2 developers having ironed out the wrinkles if any, and possibly other packages not-so-marked could be also made cross-compilable.

An example that cannot currently be cross-compiled is Firefox, but as I see it, the main purpose of the T2 cross-compile is to create a working toolchain that can be booted into, then apps can be compiled.

Woof: cups optimisation

When Woof imports a 'cups' package from another distro, there is a 'help' directory, that I retained as I thought it might be needed for the CUPS web interface. However, it seems non-essential, so I have now removed it.

The location of the 'help' directory varies with different distros:

T2: /usr/share/doc/cups/help
Slackware: /usr/doc/cups/help
Debian: /usr/share/cups/doc-root/help

I could have added code into script '2createpackages' to split the 'help' directory out to *_DOC, however I decided to treat this as a special case, and have done it in the 'cups' template.

I have added code into packages-templates/cups/FIXUPHACK to handle this. It will work for all distros. It locates the 'help' folder and moves it to cups_DOC in the packages-* target directory (ex: packages-slacko/cups_DOC).

These help files are about 1MB uncompressed, so a significant size saving.

Guy Dog 5.0.1

Forum member Iguleder is one guy who is leading innovation in the Puppy world. He is exploring new ways of doing things. One such outcome is Guy Dog, a fresh new take on a Dpup. Small size, with an interesting and rather different selection of applications and utilities.

Find out more:
http://www.murga-linux.com/puppy/viewtopic.php?t=72576

There are many ideas here that may be paving the way for future puppies.

No difference Woof vs Unleashed

I was reading something that Iguleder posted in his "Next Puppy" forum thread. It says something that I should reply to, as it is incorrect.

Quoting Iguleder, from http://www.murga-linux.com/puppy/viewtopic.php?t=70758:

Puppy gained big amounts of fat during the transition from the 4.x series to the more automated 5.x series, which is built using Woof. Many packages that were included in the early Puppy 4.x series made their way into recent Puppy versions, although they were totally unneeded.

Additionally, although the high degree of automation gave Puppy developers more control over the choice of packages shipped with each release, it also pulled in lots of unneeded files and packages from other distributions, such as distribution-specific documentation. The intimate knowledge of each and every package included in the release was gone.

These statements are very misleading. There is no difference between how Woof builds a Puppy live-CD and how Unleashed (the pre-Woof build system) does it.

Puppy 4.3.1 was built with Woof. How? No difference from earlier 4.x puppies.

There are just a collection of PET packages. The build script, either in Woof or Unleashed, puts them together, and voila a live-CD. No difference.

The difference is in the extra functionality of Woof. Whereas Unleashed had minimal mechanisms for creating the PET packages, Woof adds extra automation. In Unleashed, I did have some automatic tools for converting binary packages compiled in T2, to PET packages. Woof inherited those same mechanisms, just extended it to handle binary packages from other distros.

Spelling this out again:
The Puppy 4.x series is built from base packages originally compiled in T2. In Unleashed, I used scripts to convert them to PETs, exactly as is done in Woof. The PETs can then be hand-crafted for smaller size -- regardless whether you are using Unleashed or Woof.

Now, for a difference with Woof. Importing binary packages from another distro such as Debian, Woof does not convert them to PETs, just applies automated conversion. So, without any hand-crafting, the packages may be bigger. However, they can still be hand-crafted by means of the template directories -- that is the new way of doing it.

Wary Puppy is an interesting case study. I compiled the sources in T2. I used scripts in Woof to convert the T2 binary packages to PETs. I did some hand-crafting of the PETs. I built Wary (and Racy) from those PETs. In other words, a process exactly like Unleashed, resulting in the same small size.
And indeed, Wary and Racy are very small.

So, Woof has extended the functionality of Unleashed, and in no way taken away from it.

Now, if someone building say, a Dpup, wants to make it smaller, they have various choices.

1. Tweak the templates is the recommended way, as that sets in place something that others can use.
2. Compile some PETs, hand-craft them. Which is what we already do. In fact, the Dpup that I created was built with some of my own PETs instead of the Debian packages, as mine had less dependencies and were much smaller.
3. Convert the Debian binary packages into PETs, then hand-craft them. This is what I did for Wary (from T2, not Debian). Woof has scripts in the 'support' folder to do this -- 2compat2pets, 3compat2pets -- they so far have only been used on T2 pkgs, might need some mods for Debian. Or, you can convert individual Debian packages to PETs then hand-craft them -- which is very easy, as the package is to be found in folder packages-dpup (or whatever name is assigned to the folder that holds the packages after they have been converted from the original .deb packages) -- just grab the package, hand-craft it, make it into a PET (dir2pet).

Quick Setup translations

I introduced Quick Setup here:
http://bkhome.org/blog/?viewDetailed=02550

I posted on the Forum requesting translations:
http://www.murga-linux.com/puppy/viewtopic.php?t=72868

...the 'quicksetup' script that I posted there is the latest and is slightly later than the one in Racy 5.1.110.

I received an email from Forum member 'esmourguit' with a fr (France) translation. Thanks for that, it has gone into Woof.

WebKit compile fail

So far, anyway. I did compile WebKit successfully once before. This time, I compiled all the dependencies, no problem there. Then started WebKit compiling, and it went on for ages.

Finally, stopped with:

make[1]: Entering directory `/mnt/sda10/downloads/input216/0-midori/webkit-1.7.1'

  CXXLD  Programs/DumpRenderTree

./.libs/libWebCoreInternals.a(libWebCoreInternals_la-Internals.o): In function `WebCore::Internals::createShadowContentElement(WebCore::Document*, int&)':

Internals.cpp:(.text+0xcba): undefined reference to `WebCore::ShadowContentElement::create(WebCore::Document*)'

collect2: ld returned 1 exit status

make[1]: *** [Programs/DumpRenderTree] Error 1

make[1]: Leaving directory `/mnt/sda10/downloads/input216/0-midori/webkit-1.7.1'

make: *** [all] Error 2

This is WebKit version 1.7.1. That error message is double-Dutch to me. I will try a different version of WebKit.

Multi-architecture Woof

Woof has quite a lot of binary executables in it. For a long time I have known that I would eventually have to do something about these, as Woof moves to support other architectures such as x86-64 and ARM.

The way I was thinking of implementing it, is move all the binary executables into a directory. For example, support/printcols. Move it to:

binaries/native/x86/support/printcols

Then there would be, for example:

binaries/native/arm11/support/printcols

There would have to be two variables:

DISTRO_BUILD_NATIVE_ARCH=arm11
DISTRO_BUILD_TARGET_ARCH=arm11

One of the early build scripts would read those variables, then scan the appropriate binaries sub-directory and create symlinks to all the binary executables.

So, there would still be woof-tree/support/printcols, but it would be a symlink.

An example of a "target" binary executable would be rootfs-skeleton/usr/sbin/pngoverlay. That would be moved to:

binaries/target/x86/rootfs-skeleton/usr/sbin/pngoverlay

This mechanism is quite simple to implement, and Woof will work without any changes, except will have to recognise "target" symlinks and copy the destination file into the build, not just the symlink.

Any thoughts on this from other developers welcome.

Cross-compile toolchains

I thought that I would take a look at some of the alternatives out there. I have used T2 (many times) and Buildroot:

T2: http://t2-project.org/
Buildroot: http://buildroot.uclibc.org/
Crosstool-NG: http://crosstool-ng.org/
OpenEmbedded: http://www.openembedded.org/wiki/Main_Page
PTXdist: http://www.pengutronix.de/software/ptxdist/index_en.html

Racy: glxgears?

On my laptop, running glxgears hangs the system. A black window appears, but no gears inside.

Mesa was compiled in T2, so I had a go at recompiling it myself. The T2 configure options:

./configure --prefix=/usr/X11R7 --sysconfdir=/etc --localstatedir=/var --build=i686-pc-linux-gnu --enable-gallium-swrast --enable-gallium-svga --enable-gallium-i915 --enable-gallium-i965 --enable-gallium-radeon --enable-gallium-r600

...after I recompiled, glxgears still hangs.

So I tried just this:
./configure --prefix=/usr/X11R7 --sysconfdir=/etc --localstatedir=/var --build=i686-pc-linux-gnu

...this time, glxgears does work. A problem though, the DRI drivers are much bigger, like up from about 2MB each to 3MB, bumping the entire package up by 10MB (uncompressed).

I am wondering if the problem is just with either my particular laptop, or the Intel DRI driver. I would greatly appreciate feedback from Racy testers, try 'glxgears'. Let me know what video you have (Intel, nVidia, etc., if you know).

If you don't have the glxgears binary handy (it isn't in Racy), it is here:
http://distro.ibiblio.org/quirky/test/glxgears.gz
...gunzip, be sure to set the execute flags.

Wary Service Pack 5.2 - 5.2.1

This Service Pack upgrades Wary from version 5.2 to 5.2.1.

Download, then click on it to install it. It might not seem to do anything, but fixes various minor bugs -- and one big bug, printing was broken when try to print from some apps such as Abiword.

Download PET (134KB):
http://distro.ibiblio.org/quirky/wary-5.2/service-pack-wary-5.2-5.2.1.pet

Congrats for Slacko 5.3

I sent a pm to Mick congratulating on the release of Slacko 5.3, but would also like to do it publicly.

Personally, I was impressed by Mick's even temperament, patience and persistent enthusiasm.

During the alpha/beta/rc development phases, I noticed the huge amount of traffic on the Slacko Forum threads -- many enthusiastic contributors! of course it is not just Mick who needs to be congratulated, it is also the many other developers and testers, many of whom Mick has listed in the Slacko 5.3 Announcement page:
http://distro.ibiblio.org/puppylinux/puppy-5.3/release-Slacko-5.3.htm

I have just looked at Distrowatch, for the last 7 days Puppy is at number 6, and heading upward!

Woof uploaded, Oct. 27, 2011

This is commit '20111026211915', used to build Racy 5.1.110.

The previous upload of Woof was Oct. 14, 2011:
http://bkhome.org/blog/?viewDetailed=02542

Racy 5.1.110 (5.2 beta2)

The latest and greatest Racy has arrived! Download from here:

http://distro.ibiblio.org/quirky/test/racy-5.1.110/

This has the new "first run" Quick Setup. If you would like to create a language translation file for it, please get the latest 'quicksetup' script from here (tiny changes from the one in Racy 5.1.110), gunzip it and place at /usr/sbin:
http://murga-linux.com/puppy/viewtopic.php?t=72868

One difference from the previous Racy, is this time I put the Xorg 'nv' driver into the alternatives directory and left 'nouveau' in the main drivers directory. So, the latter will get used, and if you want to use 'nv' then you will have to run the Xorg Wizard.

Here is a Forum thread for feedback:
http://murga-linux.com/puppy/viewtopic.php?p=576964#576964

Translations for Quick Setup

If anyone is interested in contributing a translation for Quick Setup, I have started a Forum thread:

http://murga-linux.com/puppy/viewtopic.php?t=72868

Small SeaMonkey addons

This is interesting. This guy has repackaged many Firefox/SeaMonkey addons, making them much smaller:
http://code.google.com/p/firefox-extension-shrink/source/browse/trunk/?r=6

To give you an idea how much they are shrunk, look at this URL which has the full addons:
http://releases.mozillaonline.com/pub/mozilla.org/addons/
Or here:
http://ftp.acc.umu.se/Public/mozilla.org/addons/

Video_download_helper is here:
http://releases.mozillaonline.com/pub/mozilla.org/addons/3006/

...version 4.9.4 is 666KB, whereas the repackaged .xpi is only 349KB.

Wary/Racy printing fix

Wary 5.2 users have reported that printing works from some applications, not from others.

The one guy who knows all about printing is forum member rcrsn51, and awhile back he advised me that the file /usr/lib/cups/filter/pdftops is missing in Wary.

Fine, except that the cups compiled in T2 does not have that file. I checked, no previous compile of cups in T2 has it either. However, the file does exist in the Poppler package -- and there is a comment in the Poppler docs that their utilities may work better than those of the same name in the cups package (at least, that is the comment I remember). So, I applied my own reasoning, created a symlink from /usr/lib/cups/filter/pdftops to the Poppler file at /usr/bin/pdftops.

When I tested printing prior to releasing Wary 5.2, it worked. However, I just happened to only test printing from ePDFView and maybe one other app that does work. It turns out that printing from Abiword and many other apps is broken.

So, I was left puzzled. This is where rcrsn51 has come to the rescue, and we are indebted to his knowledge and helpfulness. Here is his response and fix:
http://murga-linux.com/puppy/viewtopic.php?t=72258&start=120

Just install the 'cups_pdftops_filter-1.3.11.pet' that rcrsn51 has posted.

This leaves me wondering what went wrong in T2. Perhaps there is a pre-compile configure option missing.

Psync 2.6

This is an application to synchronise the computer clock to an Internet time server, created by tazmod. Forum thread:
http://murga-linux.com/puppy/viewtopic.php?t=42978

I have uploaded the latest PET (31KB):
http://distro.ibiblio.org/quirky/pet_packages-common/psync-2.6-ul.pet

PuppyPhone 1.2, take 2

I posted previously about Psip (PuppyPhone):
http://bkhome.org/blog/?viewDetailed=02500

Forum member Sylvander reported that PuppyPhone freezes in Wary 5.2, so smokey01 has recompiled it for Wary (see post 21 Oct.):
http://murga-linux.com/puppy/viewtopic.php?t=70867&start=405

I have added this to the 'common' repo (697KB):
http://distro.ibiblio.org/quirky/pet_packages-common/puppyphone-1.2-psip32-1-w5.pet

Pfilesearch 1.27, Pmusic 2.1.9

I have upgraded zigbert's apps in the 'common' repo.

Forum threads:
http://www.murga-linux.com/puppy/viewtopic.php?t=26764
http://www.murga-linux.com/puppy/viewtopic.php?t=31206

PETs (9KB, 71KB):
http://distro.ibiblio.org/quirky/pet_packages-common/pfilesearch-1.27.pet
http://distro.ibiblio.org/quirky/pet_packages-common/pmusic-2.1.9.pet

Grub4DosConfig 1.7

Up until today the 'common' repo had Grub4DosConfig 1.6.3, although shinobar had released 1.7 in February 2011. Shinobar recently advised me about this, so I have now upgraded. Forum thread:
http://www.murga-linux.com/puppy/viewtopic.php?t=51697

I have uploaded to the 'common' repo (204KB):
http://distro.ibiblio.org/quirky/pet_packages-common/grub4dos-0.4.4.v1.7.pet

PeasyPrint 1.8

Forum member rcrsn51 has created this neat little GUI app for printing PNG/JPG files:
http://murga-linux.com/puppy/viewtopic.php?t=70133&search_id=142719086

He included the 'gphotofs' utility in his PET, however I took that out and made it as a dependency. I have uploaded the PET to the 'common' repo (2KB):
http://distro.ibiblio.org/quirky/pet_packages-common/peasyprint-1.8.pet

...well, that is 2.3KB actually!

I had a need yesterday to print some PNG images in Wary, and I realised it is not so straightforward. Neither Viewnior nor mtPaint can print. The solution was to import the image(s) into some other app, such as Abiword. Rcrsn51 also has peasyPDF which would do it.

Anyway, PeasyPrint makes printing images much more straightforward, and I will include this app in the next Wary/Racy.

Pcdripper 3.9.3

Forum member plinej has upgraded Pcdripper, his Audio CD song ripper:
http://murga-linux.com/puppy/viewtopic.php?t=24157

I have uploaded the PET to the 'common' repo (21KB):
http://distro.ibiblio.org/quirky/pet_packages-common/pcdripper-3.9.3.pet

...21KB, ah, I do so like our tiny little (yet powerful) Puppy apps!
That PET even has locale translation files, I didn't bother to separate them to a _NLS PET package.

ath9k_htc firmware

Forum member pemasu sent me a pm with firmware for the ath9k_htc.ko network driver. Pemasu also provided these links:

About ath9k_htc firmware:
http://www.murga-linux.com/puppy/viewtopic.php?p=568931#568931
http://www.murga-linux.com/puppy/viewtopic.php?p=550645#550645
Seen more of these needs.....
/lib/modules/2.6.39/kernel/drivers/net/wireless/ath/ath9k/ath9k_htc.ko

I have put the firmware at kernel-skeleton/lib/modules/all-firmware/ath9k_htc.tar.gz and updated kernel-skeleton/etc/modules/firmware.dep. The post-install script kernel-skeleton/pinstall.sh has a test added to delete the firmware if the ath9k_htc.ko driver is not present.

More information:
http://linuxwireless.org/en/users/Drivers/ath9k_htc

Gpptp 0.2

Forum member jafadmin has done some work on the source code for Gpptp, a GUI for pptp VPN client:
http://murga-linux.com/puppy/viewtopic.php?t=41929

Jafadmin has sent me the source, which I have uploaded to my source repo:
http://bkhome.org/sources/alphabetical/g/
(Us#er#na#me: pu#pp#y Pa#ss#wo#rd: li#nu#x)

I have compiled the source in Wary and added the files /etc/ppp/vpn_servers, /etc/ppp/vpn_userids and /usr/share/doc/gpptp.txt (was 'readme') provided by jafadmin (8KB):
http://distro.ibiblio.org/quirky/pet_packages-common/gpptp-0.2.pet

Note, I don't know how, if at all, this source differs from the one I had before. Anyway, I have assigned this latest as version 0.2.

Intel iwlwifi wireless firmware

Thanks to tempestuous who has provided extra firmware for the Intel iwlagn/iwlwifi wireless network chips:
http://murga-linux.com/puppy/viewtopic.php?t=70736&start=255

I have put this into Woof, at kernel-skeleton/usr/lib/all-firmware/iwlwifi.tar.gz.

I am currently running Wary, with 2.6.32.45 kernel. Running "modinfo iwlwifi":

firmware:       iwlwifi-4965-2.ucode

firmware:       iwlwifi-5150-2.ucode

firmware:       iwlwifi-5000-2.ucode

firmware:       iwlwifi-6050-4.ucode

firmware:       iwlwifi-6000-4.ucode

Tempestuous has supplied those, except for 'iwlwifi-6050-4.ucode' -- he has supplied 'iwlwifi-6050-5.ucode' -- gee, I wonder if a symlink would work?

I want the firmware tarballs in Woof to work for all kernels from 2.6.32 to latest. But, this iwlwifi tarball is becoming a monster.

Network Wizard 20111016

I posted recently about Network Wizard 20111015:
http://bkhome.org/blog/?viewDetailed=02544

01micko has made some more changes and now we have 20111016. I have uploaded this to the 'common' repo (61KB):
http://distro.ibiblio.org/quirky/pet_packages-common/net_setup-20111016.pet

8-bit has reported that it spits out an error message that "scanpci" is missing. Yes, but that doesn't prevent the application from working. Sometime, we need to put in a test for the existence of scanpci.

Puppy Linux 5.3 (Slacko)

Our new flagship Puppy Linux has been released, version 5.3, named Slacko Puppy. This is an edition of Puppy Linux built with binary compatibility with Slackware 13.37 packages, and access to the Slackware package repositories. The Project Coordinator is Mick Amadio.

Announcement and Release Notes:
http://distro.ibiblio.org/pub/linux/distributions/puppylinux/puppy-5.3/release-Slacko-5.3.htm

Download live-CD (124.2MB):
http://distro.ibiblio.org/pub/linux/distributions/puppylinux/puppy-5.3/

A Forum thread has been started for feedback on Slacko 5.3:
http://www.murga-linux.com/puppy/viewtopic.php?t=72805

Slacko has been developed over some months, with a very enthusiastic team, and hectic pace. To read more about the development:
RC2: http://www.murga-linux.com/puppy/viewtopic.php?t=72744
RC1: http://www.murga-linux.com/puppy/viewtopic.php?t=72361
B5: http://www.murga-linux.com/puppy/viewtopic.php?t=72030
B4: http://www.murga-linux.com/puppy/viewtopic.php?t=71951
B3: http://www.murga-linux.com/puppy/viewtopic.php?t=71787

Slacko 5.3 is now our recommended "first stop" for all newcomers. Try Slacko!

Be aware however, that Puppy Linux development is an incredibly rich environment, with many individuals exercising their creativity, and we actively encourage lots and lots of different puppies. We do however currently have three "official" puppies, Slacko, Lucid and Wary.

Lucid Puppy
Note that our previous flagship release of Puppy Linux was 5.2.8, Lucid Puppy, built from Ubuntu binary packages, and for those who would prefer to have compatibility with Ubuntu packages, this edition is still being developed and has upgrade packages and some experimental later releases codenamed "Three Headed Dog". Read more here:
http://www.murga-linux.com/puppy/viewtopic.php?t=70736

Wary Puppy
For older hardware, there is Wary Puppy 5.2:
http://bkhome.org/blog/?viewDetailed=02543

Please don't read too much into the version numbers, 5.3 (Slacko), 5.2 (Wary), 5.2.8 (Lucid). Our puppies follow parallel but independent paths, with different version numbering.

More puppies
This News Page does not cover all of the great puppies out there, but is a good starting place to find out more:
http://puppylinuxnews.org/
We also have some excellent puppies being developed based on Debian binary packages, look in the Puppy Forum:
http://murga-linux.com/puppy/

Gtkdialog dimensions fix

Thanks to thunor, an issue that I have been having with gtkdialog for a very long time has been resolved.

I have often had a problem with window dimensions not fitting snuggly around the widgets, with blank space on the left or bottom of the window.

I posted about the problem here, with snapshots, and thunor replied with the solution:
http://murga-linux.com/puppy/viewtopic.php?t=69188&start=375

Inserting resizable="false" into the tag works like a charm!

The other bit of advice, to use wrap="false" on the top text widget, has also been very helpful.

Viewnior 1.1

I announced back in January 2011, compiling Viewnior 1.1 in Wary:
http://bkhome.org/blog/?viewDetailed=02106

I have now copied those PETs into the 'common' repo. That makes them accessable for all Woof-built puppies (58KB):
http://distro.ibiblio.org/quirky/pet_packages-common/viewnior-1.1.pet

3.0.7 kernel

I recently reported about compiling the 3.0.4 kernel:
http://bkhome.org/blog/?viewDetailed=02546

Thanks to links posted by 'RandomVisitor' I have now downloaded the 3.0.7 source and compiled it.

The kernel source, and SFS file, are here:
http://bkhome.org/sources/kernel-3.0.7/

For Puppy builders, the 3.0.7 PET is here (23.4MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/linux_kernel-3.0.7-pae-i686-unionfs-w5.pet

Yes, this one also has Unionfs instead of Aufs. It started as an experiment, but as I have had no trouble whatsoever with Unionfs so far, it is going to stay on. I expect to use this PET in the next build of Racy.

Slacko 5.3-RC2

Release Candidate 2 of the upcoming Slacko Puppy 5.3 is out:

http://murga-linux.com/puppy/viewtopic.php?t=72744

If you want to test it, do so today, as the final, if no big bugs emerge, is imminent.

Set Hostname, etc.

Quick Setup does not have the facility to set the hostname. Instead, I have placed it in a separate script, /usr/sbin/hostname-set.

'hostname-set' is called via button in the Internet Connection Wizard, which is the network setup frontend "glue" tool that runs when the 'connect' icon is clicked on the desktop.

So, when will all this new stuff be released? I am expecting to bring out a beta of Racy about mid-week, built from latest Woof and this will have Quick Setup and Set Hostname.

I think also a bugfix "Service Pack" PET for Wary 5.2 should be released soon -- not with new features, just some fixes. I will try and do that later in the week.

Quick Setup

I posted about Quick Country here:
http://bkhome.org/blog/?viewDetailed=02548

'quickcountry', or 'countrywizard', are symlinks to 'quicksetup' that bring up a subset of the full Quick Setup.

I think that I have now got the full Quick Setup working nicely, and this is what runs when there is automatic X startup at first boot. Snapshot:

Next thing that I will do is implement hostname selection, but this will be in the Internet Connection Wizard (the default app that runs when click on the 'connect' desktop icon).

'mousecheck'

I introduced this concept in Woof awhile back, then took it out. Now I have worked on it some more and it is back. For the situation of X starting automatically at first boot, a new script 'mousecheck' runs in the background.

If you don't move the mouse within a certain time, it puts up a request to please move it, if still no movement then puts up a dialog for configuring the mouse -- that may be keyboard driven.

The script is internationalised. If anyone is interested in creating a translation for it, please do. It is here:
http://distro.ibiblio.org/quirky/test/mousecheck.gz

...down and gunzip to /usr/sbin, make sure execute attributes are set.

Instructions for internationalising:
http://distro.ibiblio.org/quirky/test/mousecheck.pupdev.gz

Contents of instructions file:

i created the mousecheck.pot file like this. first I renamed script temporarily

to mousecheck.sh, then:



xgettext -o mousecheck.pot --no-wrap mousecheck.sh



Now convert to a .po file in your language, example 'de':



# msginit --no-wrap --locale=de_DE --output-file=de.po --input=mousecheck.pot



...the generated de.po has a line with "charset=", please change that to

"charset=UTF-8"



Also, edit this line to look like this:



"Project-Id-Version: mousecheck 1.0\n"



Then insert your country translations into de.po.



Then create a compiled mousecheck.mo file:



# mkdir -p /usr/share/locale/de/LC_MESSAGES

# msgfmt --check --output-file=/usr/share/locale/de/LC_MESSAGES/mousecheck.mo de.po



If you want to contribute this file so it will be distributed with Puppy, email de.po

to me: bkaulerATgmailDOTcom



note that an existing .mo file can be uncompiled like this:

# msgunfmt mousecheck.mo

It is easy to test. Open a terminal window. Put the mouse pointer in centre of screen. Type "mousecheck" in the terminal, then wait.

Quick Country

A few days ago, I started to implement shinobar's firstrun into Woof, but I was customizing it so much that I realised that I would be better off writing it from scratch. So, I created Quick Setup, script 'quicksetup'.

These are symlinks to 'quicksetup':

quickcountry
timezone-set
chooselocale

The last two replace the scripts of the same name, and work with or without X.

I implemented the country settings, same as shinobar's firstrun, but when I started on the X setup and hostname, I had second thoughts. I implemented resolution selection by means of xrandr, however I don't think it appropriate to be offering the xrandr solution as the first choice. Also, I don't know if this is the right place to be setting the hostname.

There is a variable, DISTRO_XORG_AUTO in /etc/DISTRO_SPECS. If 'yes', then X starts automatically at first boot, otherwise the Xorg Wizard is run. Wary is built with 'no', so behaviour is as before, locale, keyboard and Xorg Wizard text-mode dialogs.

However, Racy has a recent version of Xorg and is built with 'yes', so now at first boot execution goes straight to the desktop, no text-mode dialogs at all. However, instead of bringing up the full Quick Setup, the symlink 'quickcountry' is executed, which only brings up the country settings:

After clicking the 'OK' button, a restart of X is performed if necessary, then the welcome1stboot is offered -- I intend to redesign that to offer the Xorg Wizard more prominently. Setting hostname will be offered somewhere in the network setup Wizard.

Script /usr/sbin/quicksetup is internationalised, and /usr/sbin/quicksetup.pupdev has full instructions for creating a language translation file.

numlockx

'numlockx' is a utility for turning keyboard numlock on and off. Project home:
http://home.kde.org/~seli/numlockx/

Shinobar's 'firstrun' PET has the numlockx binary executable in it. Firstrun is used in many puppies, as it offers a nice X GUI localization interface instead of the first-boot text-mode dialog windows of the traditional Puppy.

I am considering using firstrun, maybe even merging it into Woof, however, the first step is to extract 'numlockx' as a separate PET. It is uploaded to the 'common' repo and usable by all builds (3KB):
http://distro.ibiblio.org/quirky/pet_packages-common/numlockx-1.2-w5c.pet

3.0.4 kernel

I have compiled the 3.0.4 Linux kernel, configured for PAE (supports up to 64GB RAM), SMP (multiple cores) and Unionfs -- not Aufs.

I have tested with a frugal installation on hard drive, and booting off a USB pen drive -- I also tested a save-file and saving to entire partition. I also tested installing a PET package, in which the files install direct to the save-layer.

Success in all cases. So, this has been uploaded. Here is the PET (23.4MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/linux_kernel-3.0.4-pae-i686-w5.pet

The source, patches, 3rd-party drivers, build scripts, instructions, and source SFS file are here:
http://bkhome.org/sources/kernel-3.0.4/

Perhaps the next Racy will use this kernel.

Racy progress report

Racy is now getting his turn...

I have compiled the 3.0.4 kernel. Some scripts in Woof need to be fixed for a 3.x kernel, so far I have fixed 3builddistro support/findpkgs, and rootfs-skeleton/usr/sbin/bootmanager.

Shinobar recently reported a problem with Aufs when booting from USB drive and installing a PET. This coincided with an announcement of a new version of Unionfs. We have tried Unionfs many times in the past, and have had problems, and ended up going back to Aufs. But, I decided to give Unionfs another go. This 3.0.4 kernel is using Unionfs.

I have done a frugal install of Racy built with 3.0.4 kernel and Unionfs. Running it now, so far so good!

Next thing to do is modify /usr/sbin/snapmergepuppy to work optimally with Unionfs, then I will test booting off a USB Flash drive. Fun!

Network Wizard bugfix

Not my Simple Network Setup, the big fella, that we have dubbed "Dougal's Network Wizard", as Dougal had a hand in a major rewrite of it at one stage. But, this Wizard has had many contributors.

A bug has just been reported by 'repos':
http://murga-linux.com/puppy/viewtopic.php?t=72543

I have fixed it, and uploaded the PET to the 'common' repo (61KB):
http://distro.ibiblio.org/quirky/pet_packages-common/net_setup-20111015.pet

Wary Puppy 5.2 released

The last official release of Wary Puppy, our build or "edition" of Puppy Linux that targets older hardware, was on September 1, 2011, see announcement:
http://bkhome.org/blog/?viewDetailed=02454

Since then, development has continued at an intense pace. As this extraction from the Wary 5.2 Announcement page states:

This is a massive upgrade relative to the 5.1.x series. All of the base packages were recompiled in T2. Certain choices were made in T2 with the plan of seamless upgrading from Xorg 7.3 to 7.6 -- that is, the default Wary system has Xorg 7.3, but it is planned that Wary can be upgraded to Xorg 7.6 by installation of a single PET package, and all applications will work before and after. This required some very careful configuration. The idea is to "have our cake and eat it too" -- Xorg 7.3 for old hardware, easy upgrade to 7.6 for newer video hardware. The plan actually seems to be working.
As usual, huge changes yet only a small version-number change. Many bug fixes, upgrades, new packages.

There are fascinating new features in this Puppy. Take the time to peruse the Announcement and Release Notes page:
http://distro.ibiblio.org/quirky/wary-5.2/release-Wary-5.2.htm

Also, for an introduction to what Wary is all about, read this page:
http://bkhome.org/wary/

The live-CD ISO file is 123.6MB, which might seen pretty small, yet it could have been much smaller -- the build has the "kitchen sink" in terms of kernel drivers, for example very large 3rd-party analog modem drivers. Download:
http://distro.ibiblio.org/quirky/wary-5.2/wary-5.2.iso

All the tools needed for C/C++/Vala/BaCon coding and compiling are in this SFS file (109.7MB):
http://distro.ibiblio.org/quirky/wary-5.2/devx_wary_5.2.sfs
(SFS files are introduced here:
http://puppylinux.com/development/package-management.htm)

Extra release notes

1. The Puppy Package Manager can install PET packages from the 'wary5' repository. However, we need to test all of the packages in that repo, that they still work. One issue is that 'ffmpeg' got updated in latest Wary, which might have broken a couple of the media player PETs (Gnome-mplayer, Smplayer). This testing will be done ASAP.

2. The pup has the 2.6.32.45 kernel, configured with SMP (multiple CPU cores) support enabled. I did also compile a uniprocessor kernel, however in alpha/beta testing no-one reported a failure with the SMP-enabled kernel on a uniprocessor system that was fixed by using the uniprocessor-enabled kernel. Hence, I have not built a uniprocessor version of Wary 5.2.

3. I am still not 100% certain that an old save-file will upgrade correctly. If you have a save-file from an earlier version of Wary, booting up with 5.2 should upgrade it ok -- if X fails, try rebooting -- if X still fails, try with boot parameter "puppy pfix=purge".

4. Wary 5.2 has a new theme. I would especially like to thank forum member 'dejan555' for the lovely wallpaper. I posted about the theme here:
http://bkhome.org/blog/?viewDetailed=02538

Woof uploaded, October 14, 2011

This is commit '20111014045659'. This version was used to build Wary Puppy 5.2.

The previous upload of Woof was October 5:
http://bkhome.org/blog/?viewDetailed=02521

There was a blog post recently on how to download a recent snapshot tarball of Woof and use that as the reference to download Woof changes since then, without the full history:
http://bkhome.org/blog/?viewDetailed=02306

I have uploaded a snapshot tarball of commit '20111014045659'. This can be downloaded by itself, but it is recommended to also download the Bones infrastructure as described in above link. Get tarball:
http://bkhome.org/bones/woof/archive/woof-20111014045659.tar.gz

To download and use Woof, please read the Bones intro page:
http://bkhome.org/bones/
And the Woof intro page:
http://bkhome.org/woof/

Low-cost Aakash tablet

Some of the guys are enthused about the $35 Raspberry Pi Arm board:
http://murga-linux.com/puppy/viewtopic.php?t=72407

Which reminded me of something else that I came across recently, a $35 tablet:
http://www.techdrivein.com/2011/10/indias-35-tablet-aakash-launched-runs.html

More info:
http://pib.nic.in/newsite/erelease.aspx?relid=76476

That $35 price is the discounted price to students in India. Regular price is $50, and about $60 when it goes on sale to the public.

Fatdog64 version 521 released

Kirk has announced the release of Fatdog64 521:
http://www.murga-linux.com/puppy/viewtopic.php?t=72483

Fatdog64 home:
http://distro.ibiblio.org/pub/linux/distributions/fatdog/web/

If you have a computer with 64-bit x86 CPU, which is most cases these days, then you might be interested in this pup!

Why?
Well, as with all of our puppies, try it, it might have something that you like. Maximum speed of course. Then there's the extra security of web browser and some other apps running as user 'spot'.

Dialup will be dead soon?

Forum member mrbubl3s asked "Am I the only one still using dialup?":
http://murga-linux.com/puppy/viewtopic.php?t=53621

The USA has been rather slow on the uptake of broadband, but they are getting there. The latest estimate, as far as I can determine, is that about 10-12% of Internet users in the US are still using analog modem dialup:
http://www.webmasterworld.com/webmaster/4087174.htm

That is a very high figure. In most parts of the world, including third-world countries, broadband in some form has become the norm.

I don't personally know anyone here in Australia who is still using dialup. I recently let my dialup account with Netbay lapse, as last year I only used it about twice.

Which raises the question, what about the future of Wary? Wary currently uses the 2.6.32 kernel and supports a range of old analog modem cards. However, I might only keep this up for another year, after that, advance to a later kernel -- which will still support some analog modems -- anyone still on dialup will just have to get a compatible modem. You can buy a brand new USB hardware analog modem for $40-$50, so for a few people still lingering on dialup I don't think it is asking too much for them to upgrade their modem hardware.

Here is a blog post awhile back, with link to such a modem (US$45.95):
http://bkhome.org/blog/?viewDetailed=02265

Wary Puppy 5.1.106 (5.2beta2)

Thanks
This seems pretty good to me, was tempted to call it an RC. But, you never know what you tester guys will discover -- and I have just been thinking recently how much I appreciate the effort you guys put in, testing one alpha/beta after another. With so much parallel development going on, you must sometimes think, "oh no, not another one!"

PupCamera
Well, this one should be an interesting variant, as it has PupCamera. I would like to find out if this works with cameras other than my own! If your digital camera has a USB cable, plug it in (after Wary has booted), find out if, firstly, it gets automatically detected, and secondly, can PupCamera access the photos.

...in case PupCamera does not automatically popup, you will find it in the Graphic menu.

usbutils
The build also has the 'usbutils' package, which includes the 'lsusb' utility and 'usb.ids' chip database. I have finally succumbed to putting this in. PupCamera uses it (if present) to display extra information.

Numerous bugfixes, some packages upgraded since the last Wary (5.1.103, 5.2beta1).

Slow down
There's a lot of stuff pending. My mind is full of thoughts about what I would like to do to keep on improving Puppy. Like automatically detect when you plug in a printer for example. Then there's some interesting bootup internationalisation work that L18L is doing. And so on... however, I want to get Wary to 5.2final soon, probably also get Racy to release status, so will put a go-slow on changes/improvements for now, unless they are essential bugfixes.

Theme
This Wary has yet another theme:
Wallpaper: puppy_lights_blue2_dejan555.jpg
Desk icon set: LaGaBluenight
JWM: DeepBlueBold
GTK: FlatBlueContrast
ROX-Filer: desktop font DejaVu Sans Bold 12

...thanks for Forum member dejan555 for the very nice wallpaper. This theme will be in the final, or something very close, I might tweak it a tiny bit. I intend that Racy final will have the same theme.

Download
Anyway, get Wary 5.1.106 from here:
http://distro.ibiblio.org/quirky/test/wary-5.1.106/

...there are small delta files if you already have downloaded 5.1.103.

Universal Installer bugfix

Wary testers reported that the Puppy Universal Installer is broken when doing a full hard drive installation. I found that frugal also was broken.

The script is /usr/sbin/puppyinstaller, and I think it was 01micko who recently modified it to work with gtkdialog3 (it previously required gtkdialog2). I tested it at the time and it seemed ok, but now gtkdialog3 is choking on the the XML being fed to it by stdin -- however, gtkdialog4 works.

So, I have bumped the script to use gtkdialog4, plus fixed some other minor things.

Well, the latest Wary is looking good. I have done some more work on the theme. It seems good enough to be an RC...

pupdialog, bootmanager

I have extended /usr/bin/pupdialog, my simple replacement for 'dialog' when X is running, adding '--checklist' dialog. Now, these dialogs are offered:

--msgbox --yesno --menu --inputbox --checklist

I also added another parameter that is not available in 'dialog', '--helpwindow', that brings up a help window without exiting the current pupdialog window.

I was prompted to do the above as I wanted to remove Xdialog from the BootManager, at least the window that offers which SFS files to load. I have not been able to fix Xdialog hanging, as reported previously in this blog. When the Xdialog window appears, clicking any button hangs all applications, although mouse and keyboard still work, but eventually control returns -- Xdialog seems to be in a waiting loop hogging the entire system, then it times out -- waiting for what I don't know. I know that the source could be traced and the problem isolated, but i want a quick fix now.

Curiously, if I run any other application first, then BootManager, the problem does not occur. It is probably something to do with GTK 2.24.x.

By extending pupdialog, I was able to quickly change BootManager from Xdialog to pupdialog. The modified scripts are /usr/bin/pupdialog and /usr/sbin/bootmanager.

BaCon highlighting in Geany

Forum member big_bass has done some ground-breaking work to get BaCon syntax highlighting in Geany.
This is highly relevant for Wary/Racy as I have taken out Nicoedit and gtksourceview -- the only builtin text editors are now Geany and Leafpad.

big_bass has posted about it here:
http://murga-linux.com/puppy/viewtopic.php?t=48901&start=345

It is a hack, as the Freebasic highlighting has to be chosen in Geany. There is a modification that is supposed to make that automatic, but that doesn't work for me. Not all BaCon keywords are highlighted. Anyway, this is a very good start, and I have put big_bass's files into a PET (3KB):
http://distro.ibiblio.org/quirky/pet_packages-common/z_geany_bacon_hack-1.pet

This PET is not necessarily suited to all Geany installations, as it overwrites /usr/share/geany/filetype_extensions.conf.

SeaMonkey 2.4.1 ...nup

I have compiled SeaMonkey 2.4.1 in Wary, running it now. I looked through the list of bug fixes since 2.3.2, not much at all. The main bug in Composer, the greyed-out "Image properties" in the right-click menu is not fixed.

So, I am not going to upgrade the PET for Wary/Racy.

Can't say that I'm impressed with the big version jumps with Firefox and SeaMonkey these days. They made hardly any changes and went from 2.3 to 2.4, now 2.5 is in beta. Probably not as bad as what Firefox is doing though.

Psync 2.3

I have upgraded tasmod's Psync, which synchronises local time to an Internet time server. I will upload the PET to the 'common' repo soon.

Forum thread:
http://murga-linux.com/puppy/viewtopic.php?t=42978

Pmusic 2.1.8

I have created PETs for zigbert's latest Pmusic. There is a bit to explain about this...

The last version of Pmusic that works with gtkdialog3 is 1.8.3, though that has some missing icons when used with some versions of GTK. I have renamed this as 'pmusic1' and left it in the 'common' repo for anyone who needs the older version, plus I have put in the missing icons. I also changed tmp/.config/rox.sourceforge.net/OpenWith to root, for reason explained below. PET (74KB):
http://distro.ibiblio.org/quirky/pet_packages-common/pmusic1-1.8.3-1.pet

Pmusic 2.x series has many more features, and needs gtkdialog >= 0.8.0, also known as 'gtkdialog4'. A minor correction in pet.specs, again changed tmp to root. PET (71KB):
http://distro.ibiblio.org/quirky/pet_packages-common/pmusic-2.1.8-1.pet

There is another PET that Pmusic 2.x can use, I presume that it is optional, it is a bit big so I probably wouldn't have it builtin to my Puppy unless required. The pet.specs file was somewhat broken, fixed. Again, root instead of tmp. PET (137KB):
http://distro.ibiblio.org/quirky/pet_packages-common/pmusic_index_radio-2011.06.26-1.pet

There are possible complications with the files being in tmp/.config/rox.sourceforge.net/OpenWith, in the Woof build environment. Woof copies all of the packages into sandbox3/rootfs-complete, and concatenates all the pinstall.sh scripts into one big script, which is then run afterward. If files are to be relocated at installation, better to have them somewhere unique, so that they can be moved (not copied), for example it would be okay to use ./tmp if the files were somewhere unique, say ./tmp/pmusic_install. It would be tidy for the pinstall.sh script to delete ./tmp/pmusic_install. It might cause trouble if the pinstall.sh was to delete /tmp/.config. Also, some other PETs could be using the same technique of having files in tmp/.config/rox.sourceforge.net/OpenWith, so the files of all the PETs would be in there together.

Forum thread for Pmusic:
http://www.murga-linux.com/puppy/viewtopic.php?t=31206

Slacko RC is out

01micko has announced that The Release Candidate for Slacko is available. Slacko is an "spup", that is, built from Slackware binary packages (with lots more PETs compiled especially for Slacko).

http://www.murga-linux.com/puppy/viewtopic.php?t=72361

Testers welcome! last chance to report a bug!

Convert hex to decimal

This is quite interesting, thought that I would post it. I wanted to convert a hexadecimal number to decimal in a shell script. I vaguely recall doing it once before, might have used the Busybox 'printf' applet -- but can't recall how I did that.

So, I did it using 'bc', the commandline math processor:

VARHEX='12'

VARDEC="`echo "ibase=16; ${VARHEX}; obase=10" | bc`"

echo "$VARDEC"

Note, 'bc' is not a Busybox applet. Busybox has another simpler maths applet called 'dc'. Most puppies should have 'bc'.

PupCamera digital camera access

I reported about automatic detection when a digital camera is connected to the computer via a USB cable, and a basic GUI for accessing the photos:
http://bkhome.org/blog/?viewDetailed=02527

Dougal has already done some work in this direction, and Jemimah sent me a tarball of what Dougal had created. It was very useful to study what Dougal has done.

Dougal created udev rules for each type of camera that Gphoto understands, but that will need to be updated regularly -- so, I have retained the single-line udev rule as described in my previous blog post.

Dougal used the 'gphoto2' utility to access the photos, however 'gphotofs' works well, at least on my camera, so I have stayed with the latter.

I have added some "bells and whistles" to my application, and now it's name is PupCamera. It pops up when a camera is connected to the computer, but can also be run from the Graphic menu. Here is a snapshot:

It works so nicely, and if it is also successful with other cameras then I will be mighty pleased. I don't want to pull out the files separately, please wait, the next Wary or Racy will be out very soon with PupCamera in it.

pupRadio fixed

A Wary tester reported that pupRadio is missing some files. He obtained /root/.puradio from another Puppy and all was well.

I seem to recall this happened once before. I just checked, pupradio-0.13.pet and pupradio-0.14.pet that I downloaded, do not have /root/.pupradio. However, pupradio-0.11.pet that I downloaded, does have it, so I copied that over, and have built a new PET (6KB):

http://distro.ibiblio.org/quirky/pet_packages-common/pupradio-0.14-1.pet

Autodetect digital camera

Just a quick note, I have success with auto-detection of my camera. This udev rule does it:

ACTION=="add", SUBSYSTEM=="usb_device", ENV{INTERFACE}="6/1/1", RUN+="/usr/sbin/pupautodetect ptpcamera"

The "6" is bInterfaceClass, which is "Imaging"
The "1" is bInterfaceSubClass, which is "Still Image capture"
The last "1" is bInterfaceProtocol which is "Picture Transfer Protocol"

...all of this is found in /sys.

Gtkam --- useless!!!!!!!

I have got to let off some steam about this. Gtkam is a GTK GUI for libgphoto, for accessing photos in digital cameras. Some cameras have standard USB-storage interfaces, which does not need libgphoto of Gtkam to access. However, many cameras do need libgphoto.

I have only got one digital camera, a Fujifilm Finepix A610, that I have owned for a few years. When I first acquired it, Gtkam kind of worked -- but it crashed at the drop of a hat. Many others reported the same experience.

I think it was about a year ago that Gtkam actually became fairly stable and usable with my camera.

Now, the latest version 0.1.18 and the previous 0.1.17, won't detect my camera at all -- perhaps it doesn't like libusb-compat and libusb1. If I try and run the option to choose which camera, it hangs.

Yet, the CLI app, gphoto2, works like a charm. This lists my camera name and USB port:

# gphoto2 --auto-detect

Model                          Port                                            

----------------------------------------------------------

USB PTP Class Camera           usb:002,014

Various other commands can access the photos, for example this downloads them all:

# gphoto2 --get-all-files

Then I tried 'gphotofs', a FUSE filesystem for gphoto2. I just did this:

# mkdir /mnt/photos
# gphotofs /mnt/photos

...hey presto, there were all my photos in /mnt/photos!

I don't know why no one has done this yet. A simple GUI for gphoto2 and maybe gphotofs. Well, if no one volunteers to do this, I'll knock up a little GUI and give Gtkam the kick.

Generalized logo handling in Woof

Currently 01micko has a hack-script for inserting Slacko logos into Woof. I have now generalized the mechanism, to cater for whatever pup is being built.

The pup logos are kept in rootfs-skeleton/usr/share/doc/puplogos, and are named by this format:

${DISTRO_FILE_PREFIX}48.png
${DISTRO_FILE_PREFIX}96.png

The correct logos are copied up to rootfs-skeleton/usr/share/doc and 'puplogos' directory deleted in the build. This is done by script '3builddistro'.

The boot splash has also got this treatment. A file boot/boot-dialog/${DISTRO_FILE_PREFIX}-splash.gif will be automatically used as the boot splash image.

So, if any pup builder sends me their logos for example for Racy I have racy48.png, racy96.png and racy-splash.gif, I will bung them into Woof.

To assist 01micko, I created a splash file from dejan555's Slacko logo:
http://distro.ibiblio.org/quirky/test/slacko-splash.gif
http://distro.ibiblio.org/quirky/test/slacko-splash.svg

...The .svg image can be opened in InkscapeLite, export total drawing to .png at 640x388, open in mtPaint and reduce palette to 16 colors (last one unused) and save as slacko-splash.gif.

Leafpad fixed

Terryphi reported that the icon for Leafpad is missing in the menu. Fixed.

I have not changed the package name, just reuploaded it (35KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/leafpad-0.8.18.1-1-w5c.pet

A note about this PET. Leafpad has an optional dependency with libgnomeprintui. This is actually a deprecated library, but I have it in Wary/Racy, as some older apps use it. InkscapeLite can also optionally use it.
The above PET has that dependency, it gives better printing capability.

Telstra 3G twist

I have reported on this blog that my Telstra 3G connection speed goes south every now and again, then comes good again for no apparent reason. I thought that it might be at times of high demand, like in the evening, which is sometimes the case, but sometimes it can be unusable all day.

Three times recently when it has happened, I have tried an experiment. I booted Windows, ran Internet Explorer, and hey, full speed was back! On one of those tests, I booted Windows then ran SeaMonkey, super slow, then ran Internet Explorer, suddenly fast.

The third time I have run this experiment is right now. Pages were not even loading, timing out (running Wary). So, I booted up Windows XP on my old laptop, which connects to the same Telsta wi-fi wireless module (which allows up to 5 computers), ran IE, got fully speed. Then, over on my new laptop running Puppy, suddenly found that SM is loading fast again.

Coincidence?

I am wondering if Telstra is using some kind of prioritizing system that favours IE. The next experiment, when this problem occurs again, will be to try SM with user-agent set to pretend it is IE.

/root/.local/share

When I started Notecase in Wary -- just checking that I have built with the one that doesn't need gtksourceview -- got these error messages:

# notecase



(process:14164): Gtk-WARNING **: Locale not supported by C library.

	Using the fallback 'C' locale.



(notecase:14164): Gtk-WARNING **: Attempting to store changes into `/root/.local/share/recently-used.xbel', but failed: Failed to create file '/root/.local/share/recently-used.xbel.NGQG2V': No such file or directory



(notecase:14164): Gtk-WARNING **: Attempting to set the permissions of `/root/.local/share/recently-used.xbel', but failed: No such file or directory

The locale problem I have always had with Notecase, nothing new there. The missing /root/.local/share occurs at exit of the application.

Here is some discussion about this error message:
http://ubuntuforums.org/archive/index.php/t-1631433.html

Get past someone pontificating about running gedit as root, and the solution is to just create /root/.local/share. However, even leaving it out doesn't seem to matter.

Anyway, I have created rootfs-skeleton/root/.local/share in Woof.

Woof uploaded, October 5, 2011

This is commit '20111005065933'. This version was used to build Wary Puppy 5.1.103 (5.2beta1).

The previous upload of Woof was September 17:
http://bkhome.org/blog/?viewDetailed=02483

There was a blog post recently on how to download a recent snapshot tarball of Woof and use that as the reference to download Woof changes since then, without the full history:
http://bkhome.org/blog/?viewDetailed=02306

To download and use Woof, please read the Bones intro page:
http://bkhome.org/bones/
And the Woof intro page:
http://bkhome.org/woof/

Wary Puppy 5.1.103 (5.2beta1)

Finally got this one uploaded. This is the first beta of the upcoming Wary 5.2. Download from here:

http://distro.ibiblio.org/quirky/test/wary-5.1.103/

The link also has xdelta files to upgrade from Wary 5.1.99 (5.2 alpha).

There are some known issues with this release:

1. Ogle control panel
If you plug in a DVD, and then click either the 'play' icon on the desktop or the 'sr0' drive icon on the desktop, Ogle will automatically start. However, in the latter case the GUI control panel is non-responsive. I don't yet know why the difference in behaviour.

2. BootManager hangs
This bug is still there. At second bootup, after having created a save-file, if any suitable .sfs files are found at /mnt/home, the BootManager will automatically start -- but it hangs if try to use it, requiring CTRL-ALT-BACKSPACE to get out. On the other-hand, click the tray and choose 'kill' will close the BootManager. The BootManager can be started later and it will work. I have not yet found out why this is happening.

3. Media players incompatible?
Ffmpeg has been compiled, latest from SVN. This may have API changes from the previous version in Wary, which may compromise the media player PETs such as Smplayer and Gnome-mplayer -- we probably will need to recompile them. And anything else that uses the ffmpeg libraries. I have already recompiled mplayer which is builtin to Wary.

There is another experimental theme:

Wallpaper: wary_spangles_futwerk.jpg (thanks to futwerk)
Desk icon set: LaGaDesk Bluenight (with slight mods)
JWM: blue variety (my creation)
GTK: Delicate Blue (my creation)

...I don't know if the JWM and GTK themes go very well with the icon/wallpaper themes.

Woof, Wary coming soon

I am getting back up to speed after taking a couple of days off.

Recently, two significant bugs were fixed in Woof, one dating back to 2009, where incorrect usage of 'fuser' in rc.shutdown caused a reboot instead of a shutdown (when the latter was desired), the other a more recent problem with kernels compiled with builtin USB drivers, causing save-files to be found twice at bootup.
Both of these are documented in this blog.

Woof is looking pretty good now. In the morning I'll have a bit of a read of Wary/Racy/Slacko threads in the Forum. Either later in the day or early next day I will probably upload Woof and Wary.

Racy Puppy logos

I think that these logos were experimental designs by afgs for Wary Puppy. But afgs came up with another design for Wary, that I settled on.

These logos seem ok for Racy, as it looks like the pup is way ahead, looking back to see how far behind the competition is...

NicoEdit retired

Terryphi has reported that NicoEdit crashes when choose File -> Open in Racy. I have confirmed that. I recompiled NicoEdit in Racy, still crashes.

NicoEdit is written in Vala, but is unmaintained. I did ask Nicolas, the author, about maintaining it awhile back, but he replied that he has moved on.

So, I have taken it out of Wary and Racy. Geany is all we need, and I think that I saw on the Forum that the guys have got syntax highlighting for BaCon working in Geany, so Geany now covers everything.

Note, I have compiled the latest libgee, 0.7.1, for Wary/Racy. However, I will leave it out of the build as I think only NicoEdit needed it.
Libgee is a library that adds extra data types to Vala.

This raises another question: what do we need gtksourceview for? NicoEdit used it. Notecase can be optionally compiled to use it I think. Hmmm, I think that I will take that out too.

Media packages upgraded

I have upgraded some multimedia packages, that did not get upgraded in the last round through T2. This is in Wary. In some cases I just recompiled the package without a version upgrade, to ensure 100% compatibility in the latest Wary. This is them:

x264 20110930
speex 1.2rc1
opencore-amr 0.1.2
libvpx 0.9.7
ffmpeg 20111002
mplayer 20110929
ogle 0.9.2 (patched)
ogle_gui 0.9.2 (patched)

Busybox rollback 1.19.2 to 1.17.2

Raffy was testing the /usr/sbin/root2user script in the latest Racy Puppy. This is called from /usr/sbin/shutdownconfig at first shutdown, if the user chooses to run as 'fido'. Raffy reported that the Busybox 'chpasswd' segfaulted.

Yes, it does segfault, but after it has done it's job. Can't say that I'm happy with that though. Racy is using the latest Busybox 1.19.2, reported recently in this blog. I tested 1.17.2, that we have been using for sometime in Puppy, and that works without the segfault.

So, I have rolled Busybox back to 1.17.2 in the 'common' repo. I know that 1.19.2 has more features, and some bug fixes, however 1.17.2 works well. I just want to quickly restore to a known working and well-tested Busybox. I recompiled 1.17.2 and enabled the 'sha1sum', 'readahead' and 'rfkill' applets.

The PET will be uploaded to the 'common' repo soon.

Bug finding USB drives at bootup

Forum member ldolse sent me a pm explaining a problem with double-finding of save-files at bootup. Here is part of ldolse's pm:

I've been working on a puplet which is using a 2.6.39 kernel, and this issue seems to be common across 2.6.39 kernels as Pemasu's puplets behave similarly.

Basically the same save file gets discovered multiple times during bootup with the latest kernels and USB flash. This forces the savefile selection menu to pop up during all reboots.

On a related note, also discussed in the thread, Woof only seems to be supporting kernels that have USB built into the kernel, as the keyboard doesn't work until after the init script. I don't know if this is intentional.

ldolse posted a fix here:
http://www.murga-linux.com/puppy/viewtopic.php?t=72087#568841

I have put this fix into Woof, however, I am surprised as I do not have this problem with my 2.6.39.4 kernel. Puppy kernels used to be built with USB drivers as modules, my latest kernels have them builtin to the kernel -- however the 'init' script should work for both.

Anyway, I will do a test build with an older kernel to verify.

Moblin->Meego->LiMo->Tizen

Operating systems and the user interface and application structure is a very volatile area on mobile devices. A few of them have come and gone.

Now let's see, I think Intel backed Moblin, which became Meego. Hewlett Packard was backing WebOS, then recently dumped it. Some of these are also intended to work on desktop systems. Then there's Google's Android.

Anyway, the latest news is that Intel is dropping Meego, in favour of Tizen. Yes, yet another one! Tizen apparently, is intended to run on anything from mobile phones to netbooks.

Now, Tizen seems to actually be based on LiMo (though it looks like bits of Meego will be in it also), and one of the LiMo executives described it thus:

LiMo 4, under LiMo�s proven collaborative governance model, enables flexible disaggregation of the device platform and the service propositions such that operators and device manufacturers can more freely shape attractive user propositions and secure sustainable long-term value

...hmmmm

Here are links:
https://www.tizen.org/
http://limofoundation.org/

I have been trying to find technical details. Very few, but allied to Gnome -- that means GTK, good, I wouldn't like it if they had gone for Qt as that tends to lock developers into coding in C++.

Interesting bunch of companies backing LiMo:
http://limofoundation.org/component/option,com_limomembers/Itemid,134/

powerapplet_tray 2.4

I reported that the battery status tray applet is not working in Racy. I found the problem. The 2.6.39.4 kernel does not have /proc/acpi/info, which causes the applet to abort. I have fixed it to work with old and new kernels.

The PET (5KB):
http://distro.ibiblio.org/quirky/pet_packages-common/powerapplet_tray-2.4-w5c.pet

The source is in my source repo:
http://bkhome.org/sources/alphabetical/p/

Short break

Just a note, so no one wonders why I am not responding to posts.

I have other stuff to do for a couple of days.

After the mini-break, I intend to upload Woof and Wary 5.2beta1.

I intend to work on a couple of issues in Racy today, such as the non-functioning battery-status tray applet.

Tiny Racy 5.1.101 (5.2beta1)

I have built another Racy, this time I wanted to see if I could get it below 100MB, just to show that we can still do it. Well, this one is 94.5MB, and it could even have been made a bit smaller. This is done without sacrificing any applications, it even has the monster SeaMonkey 2.3.2.

What I have done to achieve the reduction is as follows:

1. The Mesa DRI drivers removed, except for the software-rasterizer driver.
2. Some of the more exotic kernel modules removed. The big analog modem drivers removed. SCSI drive driver modules removed.
3. No Adobe Flash Player.
4. xz compression of .sfs file.

None of the above will disadvantage the typical user. As soon as an Internet connection is made, a window pops up asking if you want to install the Adobe Flash Player, and that is no trouble to do.

I have not provided the Mesa DRI X drivers as a separate PET, but could do so. Most users will not notice their absence. If the Tiny-Racy build becomes popular, this could be provided as a PET.

Other notes about this build:

1. The Xorg 'radeonhd' (ATI) and 'nouveau' (nVidia) drivers are in the 'drivers-alternate' directory, meaning that they are not used. Instead, Xorg will try and use the 'radeon' (ATI) and 'nv' (nVidia) drivers, as they are the ones most likely to work. If you want to try the alternative, run the Xorg Wizard.
2. The power-tray-applet (battery status) is not working on my laptop. I'll check that out today.
3. The theme is once again temporary. The surfing dog is back for this release only, seems appropriate for "racy".
4. I need to do a bit of work on the Puppy Package Manager, from a Racy perspective. In the 'puppy-wary5' repo it offers various Xorg PETs, but they are for Xorg 7.3, so they should be filtered out. Racy has Xorg 7.6.
5. The new GUI shutdown has 'de' and 'ru' translations, courtesy of L18L and rodin.s. Anyone else is invited to contribute a translation, see http://bkhome.org/blog/?viewDetailed=02489 http://bkhome.org/blog/?viewDetailed=02506
...note though, the latest rc.shutdown and shutdownconfig scripts are in Racy, not posted on the Forum.
6. Racy has PuppyPhone 1.1 (Psip) Internet telephone. As Skype has gone over to the Dark Side, maybe PuppyPhone is new killer app? Try it out! Read more: http://murga-linux.com/puppy/viewtopic.php?t=70867

Is it worthwhile offering Tiny Racy in the future? As a marketing exercise perhaps. I could push it down a bit more, say 92 or 93MB. But, considering that Racy is targeting modern hardware, with heaps of RAM, and usually fast Internet connection, the exercise is pointless. Leaving everything in, all the drivers, all DRI, with xz .sfs compression, but no Flash, it is 113MB, which is pretty small anyway. Flash would add another 5MB.

But then, perception is the point. To show how tiny Puppy is, yet with the kitchen sink inside it, is a massive publicity plus.

Download Tiny Racy from here:
http://distro.ibiblio.org/quirky/test/racy-5.1.101/

Peasyscan 1.7, PeasyPDF 1.4

I have upgraded the Peasyscan PET in the 'common' repo, from 1.4 to 1.7. This is rcrsn51's simple scanner application, see Forum:
http://murga-linux.com/puppy/viewtopic.php?t=61046

I have also uploaded the PeasyPDF application for creating multi-page PDFs out of individual images:
http://www.murga-linux.com/puppy/viewtopic.php?t=70680

PETs (4KB, 47KB):
http://distro.ibiblio.org/quirky/pet_packages-common/peasypdf-1.4.pet
http://distro.ibiblio.org/quirky/pet_packages-common/peasyscan-1.7.pet

gtkdialog4 0.8.0

I have previously posted about thunor's enhanced gtkdialog, for example:
http://bkhome.org/blog/?viewDetailed=02455

Thunor has now got it to a release-version, 0.8.0. This is, in thunor's words "for the masses". I am very impressed, the documentation is superb. One other thing that thunor commented on has also impressed me, that he started off knowing nothing about GTK -- but he has done such a professional job. He has also been very responsive to requests on the Forum.

Above link has link to thunor's project page. Forum thread:
http://www.murga-linux.com/puppy/viewtopic.php?t=69188

I have created PETs (73KB, 48KB):
http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog4-0.8.0-w5c.pet
http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog4_DOC-0.8.0-w5c.pet

I will be uploading the next Wary and Racy in a day or two, and it will have these. The docs will be in the 'devx' SFS, in /usr/share/doc/gtkdialog4.

JWM window manager 505

Yes! Joe has taken up development of JWM again. We were on version 500 (from SVN) for a very long time, now Joe has released 505.

See Joe's JWM page:
http://joewing.net/programs/jwm/

The last change was two days ago, described as:

Fix issues with dragging windows to new desktops and fix a fullscreen problem.

I have uploaded a PET (74KB):
http://distro.ibiblio.org/quirky/pet_packages-common/jwm2-505-w5c.pet

shutdownconfig, rc.shutdown i18n

I have tweaked things a bit. I separated out the translation files for these two, so that you would have 'shutdownconfig.mo' and 'rc.shutdown.mo'.

In Woof, there is no /usr/share/doc/nls. Instead, there is 'shutdownconfig.pupdev' and 'shutdownconfig.pot' alongside the script -- these get automatically placed in the 'devx'. Ditto for rc.shutdown.

Note, .pupdev files are anything, usually just text with explanation relevant to developers.

As there is some doubt about translating for 'dialog', I have set LANG=C in 'shutdownconfig' when X not running.

Regarding a schedule for next Wary and Racy, I have a fairly small shortlist, and I hope to upload tomorrow.

Shutdown/reboot bug fixed

01micko reported that in Slacko, when poweroff was chosen, a reboot happened instead. This only happened when a partition was left mounted.

Well, I found the problem. It turns out to be a very old bug, first observed late 2009. If you look in /etc/rc.d/rc.shutdown, there is a comment dated 2009-11-17:

#091117 weird shutdown bug. would not shutdown when a partition mounted, x restarted.

Testing in Wary, with a partition mounted, the shutdown process went through, but at the point where it should power-off, it rebooted.

Back in late 2009 I seem to have put in a workaround, without understanding what caused the bug. This time, I have tracked it down.

The code 'fuser -m /dev/sda5' (for example) does not work as I expected. I wanted all of the processes that are running in /dev/sda5 when it is mounted, but this example returns a heap of processes, some of which if you kill them, upset things very much, hence the wonky shutdown.

Instead, it has to be done by the mount-point 'fuser -m /mnt/sda5'.

Here is fixed code for rc.shutdown:

#091117 110928 if partition mounted, when choose shutdown, pc rebooted. found that param given to fuser must be mount-point, not /dev/*...

STRAYPARTL="`echo "$MNTDPARTS" | grep ' /mnt/' |grep -v -E '/dev/pts|/proc|/sys|tmpfs|rootfs|on / |/dev/root|usbfs|unionfs|aufs|/initrd'`"

STRAYPARTD="`echo "$STRAYPARTL" | cut -f 1 -d ' ' | tr '\n' ' '`"

STRAYMNT="`echo "$STRAYPARTL" | cut -f 3 -d ' ' | tr '\n' ' '`"

for ONESTRAY in $STRAYMNT

do

 echo "Unmounting $ONESTRAY..."

 xFUSER="`fuser -m $ONESTRAY 2>/dev/null`"

 [ "$xFUSER" != "" ] && fuser -k -m $ONESTRAY 2>/dev/null

 killzombies #v3.99

 sync

 umount -r $ONESTRAY

done

jwm_tools 0.7

Technosaurus has created some interesting utilities for managing JWM. See Forum:
http://www.murga-linux.com/puppy/viewtopic.php?t=70804

I have uploaded the PET (12KB):
http://distro.ibiblio.org/quirky/pet_packages-common/jwm_tools-0.7.pet

Note that I modified the pinstall.sh script so that the PET can also be used in Woof.

To get the ball rolling, this will be in the next Racy and Wary.

Sys-info 0.6

Tazoc is the creator of a Puppy derivative named Lighthouse Puppy. Read more here:
http://www.lhpup.org/

Tazoc has also created a PET named 'sys-info', that reports on all of the hardware in a computer. This is very handy when someone reports a bug, and instead of the developer having to ask one question after another about what hardware the person is running Puppy on, sys-info provides it all in one hit. See Forum thread:
http://www.murga-linux.com/puppy/viewtopic.php?t=60462

I have uploaded the PET to the Puppy 'common' repo, for all Woof-built puppies to use. Note, I had to modify the 'pinstall.sh' script slightly, to suit installation of the PET in the Woof build environment (that is, not in a running Puppy) (44KB):
http://distro.ibiblio.org/quirky/pet_packages-common/lhp_sys_info-0.6.pet

This will be in the next Wary and Racy.

Pburn, Pfind, Phash

I have upgraded the 'common' repo to zigbert's Pburn 3.3.7. This is the last version to use gtkdialog3. The previous was 3.3.5.

I have also upgraded zigbert's Pfind, from 4.23 to 4.24.

Jamesbond created a very simple little script for ROX-Filer, that adds an item in the right-click 'Open With...' menu, to calculate the md5sum or sha1sum of a file.
I have added phash-2.pet to the 'common' repo.

Xorgwizard touchpad fixes

The Xorg Wizard had some issues with configuring for a touchpad. This is script /usr/sbin/xorgwizard.

In Racy, there was a report of the touchpad working ok when automatic setup of X was used, but not after the Xorg Wizard was run.

I went through the relevant section of code and cleaned it up somewhat. There was a problem that caused the touchpad code to get inserted twice into /etc/X11/xorg.conf, fixed.

There was messy stuff, like "modprobe evdev" getting written to /etc/rc.d/rc.local when that wasn't needed.

PuppyPhone 1.1

I couldn't resist it, I think that the name 'PuppyPhone' is very catchy, compared with 'Psip'. So, I changed the menu entry to "PuppyPhone Internet telephone" and the PET to 'puppyphone-1.1-psip32r-ul.pet'.

I think that the plain-language name and description "PuppyPhone Internet telephone" is more likely to get us many users. Hey, maybe we could even hope for an avalanche of new users!

The PET (696KB):
http://distro.ibiblio.org/quirky/pet_packages-common/puppyphone-1.1-psip32r-ul.pet

Thanks to jamesbond for this, and the other guys who are helping, lobster and smokey01.

Oh, while I think of it, I will need the source tarball to put into my source repo, for legal reasons. Smokey01, do you have that?

For earlier discussion on Psip/PuppyPhone:
http://bkhome.org/blog/?viewDetailed=02496

Internationalized shutdown

As I already posted, L18L has converted my script to support international languages. Actually, two scripts, /etc/rc.d/rc.shutdown and /usr/sbin/shutdownconfig.

I have tweaked the latter a little bit, found some typos, and put it into Woof. L18L, I greatly appreciate your help here -- I only did i18n conversion to a BaCon app before, so am unfamiliar with using gettext in shell scripts. I learnt a lot studying what you did.

The situation now in Woof is a bit lopsided. We have internationalization of the 'welcome1stboot' window (which pops up the first time that the desktop is started), and now at shutdown. However, the progress messages at bootup and the first-boot pre-X dialog windows are not internationalized.

Anyway, summarizing the current situation:

welcome1stboot
This is a compiled executable written in BaCon. Currently we have conversions for fr, de and ru. To find out how to add your language, see files /usr/sbin/welcome1stboot.pot and welcome1stboot.pupdev (these are in the devx).

shutdown
Currently we only have 'de' conversion, created by L18L. L18L has put instructions for creating other conversions in /usr/share/doc/nls/shutdownconfig.

I will be building Wary and Racy soon, from latest Woof, and I will be inviting everyone to contribute more language conversions.

I guess that I should be looking at converting those bootup scripts...

envsubst, gettext.sh

L18L has very kindly internationalized a script in my shutdownconfig GUI shutdown system, /usr/sbin/shutdownconfig. L18L posted to this Forum thread:
http://murga-linux.com/puppy/viewtopic.php?t=71911

The method that L18L has used requires two other executables from the 'gettext' package, 'envsubst' and 'gettext.sh'. Up until now, Woof-built puppies have those in the 'devx' only.

I have modified the 'gettext' template in Woof so that now those two (as well as the 'gettext' executable) are in the main SFS. This will apply to all puppies built from the next upload of Woof.

Here is a nice tutorial that explains how a Bash/Ash script can be internationalized:
http://www.gnu.org/s/hello/manual/gettext/Preparing-Shell-Scripts.html

Flash player automatic download

Back in October 2010 I setup semi-automatic downloading of the Adobe Flash Player, see blog report:
http://bkhome.org/blog/?viewDetailed=01897

However, that is broken. I have fixed it. The script /usr/sbin/delayedrun, which runs after X has started, if it sees that the Flash Player is not installed, it goes into a loop waiting until Puppy has established an Internet connection. It then offers to download the player.

Note that the latest player has an executable 'flash-player-properties', for setting preferences. The delayedrun script also fixes the .desktop file for this. Note that the executable crashes on Wary, but only when one of the tabs is selected. Haven't tried Racy yet.

This is in Woof, so affects all Woof-built puppies. I am planning to build the next Wary and Racy without the Flash Player builtin, as it saves about 5MB. I am planning to build a very trim Racy, about 100MB.

Psip 0.26

Hmmm, I seem to be a bit "behind the times". The Psip voip project has had a resurgence of interest recently, special thanks to lobster, smokey01 and jamesbond. A few days ago I downloaded Psip 0.26, and I have just now uploaded it to the 'common' repo. But, I was just looking at the Forum, and things seemed to have progressed somewhat since that version.

Anyway, here it is (598KB):
http://distro.ibiblio.org/quirky/pet_packages-common/psip-0.26-ul.pet

The Forum thread for Psip:
http://murga-linux.com/puppy/viewtopic.php?t=70867&start=285

There is now version 1.0! Wow, ok, maybe I will wait a day or two, you guys can let me know if 1.0 is working nicely or whether there needs to be a 1.1. Then I will upload to the 'common' repo -- and it will be in next Wary/Racy. I think 01micko has already put 1.0 into Slacko.

Window background colours

I was offline all day yesterday. Today I will get stuck into Puppy development.

While driving yesterday, I was thinking that there is something that I might not have posted about anywhere. Puppy users will have noticed that some of the popup windows that I have created for Puppy have a background colour (that's "color" if you are in the USA!).

Most recently, the shutdownconfig GUI windows, that 01micko has put into beta4 of Slacko, has bright yellow windows (note, I have now changed those to pale yellow, see the zz_shutdownconfig-20110924.pet, also see L18L's internationalized zz_shutdownconfig-20110925.pet -- I will post about that separately soon).

I originally conceived that background colour for windows could be a defacto feature, or convention, for Puppy. Regardless of themes, these colours could be standardized, so always convey the same information to the viewer. I envisaged the colours to be:

Yellow: next-step
Violet: announcement
Green: success, completion
Pink: warning
Red: error
Orange: pending, in-progress
Blue: information

However, the primal red, green and blue are a bit dark, and lighter shades are probably better. For shutdownconfig, instead of yellow and orange, I am now using #FFFF80 and #FFDC32.

Popup GUI dialog applications that allow the background colour to be set are 'gtkdialog', 'yaf-splash' (both the new and old versions), 'xmessage' and my new 'pupdialog'. No for Xdialog and gxmessage. I don't know about 'yad'.

Avast! virus scanner

Great, this has a GTK-based GUI scanner, as well as a commandline scanner. Avast! is a commercial product, but free for personal use. It requires a license key to use, which is sent via email.

Here is the project Linux-edition home:
http://www.avast.com/linux-home-edition

Here is a review, comparing virus scanners for Linux:
http://www.tuxradar.com/content/get-best-virus-scanner-linux

I have created a PET. This does not have the signature file. At first startup, it will ask for the license key, and has a button to go to the web page to obtain the key. Once that is provided, the program then sees that there is no signature file and offers to download it -- it is big, 44MB.

When I first did this download of the signature file, the program spat out an error message then quit:

An error occurred in Avast! engine: invalid argument

However, I found the solution here:
http://www.facebook.com/topic.php?uid=38282497425&topic=14021

I inserted this code into /usr/lib/avast4workstation/bin/wrapper-script.sh:

SHMMAX=`sysctl -n kernel.shmmax`
[ $SHMMAX -lt 128000000 ] && sysctl -w kernel.shmmax=128000000

Avast! PET (2.5MB, 2.7MB):
http://distro.ibiblio.org/quirky/pet_packages-common/avast-1.3.0-puppy.pet
http://distro.ibiblio.org/quirky/pet_packages-common/avast_NLS-1.3.0-puppy.pet

Xfprot broken

Thanks to technosaurus who advised that there is anew download URL for f-prot, which needs to be fixed in Xfprot. The latter is a GUI for f-prot, and f-prot is a command-line virus scanner.

I fixed the URL. I noticed that Wary is using a very old Xfprot PET from the puppy-4 repo. I download version 2.4 source, had to hack it to compile, due to a symbol not supported by GTK (at least from version 2.20). Example fix in src/mygtk.c:

/*BK if(GTK_WIDGET_VISIBLE(widget)) replace with...*/
if(gtk_widget_get_state(widget))

Xfprot does not recognize the version of f-prot, and aborts at startup with:

Cannot detect f-prot's version. Aborting

I hacked around that as well, and got Xfprot working. Yes, it scans, however the commandline parameters for f-prot's 'fpscan' seem to have changed and Xprot does not control it properly. For example, archives are scanned even though the checkbox to do so is not ticked. Scanning is only down about 4 directory levels and Xfprot offers no way to change that. And so on.

I was going to put Xfprot into the 'common' repo, however it is too broken. Someone needs to put some time into updating it to work properly with latest f-prot.

I have removed it from Wary and Racy.

Some guys using Lucid Puppy have recently been trying to use Xfprot:
http://www.murga-linux.com/puppy//viewtopic.php?t=71679

...the suggestion of using f-prot directly, on the commandline, is the only viable choice for now.

Busybox 1.19.2

I have compiled the latest Busybox and put it into the 'common' repo (387KB):
http://distro.ibiblio.org/quirky/pet_packages-common/busybox-1.19.2-w5c.pet

However, this is not a PET that you can install into Puppy as-is. It is for Woof-built puppies and will need the Woof that I will upload soon.

I have resolved a long-standing problem with Busybox, the conflict between the naming of Busybox applets and "full" utilities. Now, the Busybox PET can have all the applet symlinks set to their proper names, and script 3builddistro discovers any conflict with full utility names, and automatically renames a conflicting Busybox applet with post-fix "-BB-NOTUSED".

Great, the Busybox developers have fixed the "grep -f EMPTY_FILE" bug!

These are some of the new applets, relative to earlier 1.17.2 PET:

setserial
Request or set Linux serial port information

sha1sum
Print or check SHA1 checksums

rfkill
Enable/disable wireless devices

readahead
Preload FILEs to RAM

pstree
Display process tree, optionally start from USER or PID

pwdx
Show current directory for PIDs

gHasher md5sum/sha1sum generator

Ttuuxxx discovered this originally, and back in December 2009 I compiled it for Quirky -- unfortunately that PET segfaults in the latest Wary/Racy.

So, I have recompiled it for Wary, now works. I made one little change in the source, to stop it from putting out an error message that 'libgnome.so' failed to load.

Here is the PET (12KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/ghasher-1.2.1-patched1-w5c.pet

My patched source is here:
http://bkhome.org/sources/alphabetical/g/

The homepage for gHasher is no more.

Note, I have a vague memory that someone once created a little drag-and-drop hash generator using gtkdialog. Just like my little Xdelta app (see Utility menu). I searched the forum though, couldn't find it. Drag-and-drop would be a big improvement on gHasher, also adding right-click menu in Rox would be good.

Would anyone be interested in this as a little project? Maybe use my /usr/sbin/xdelta_gui as a starting-point.

Note, many puppies, including Wary/Racy, don't have 'sha1sum', though they do have 'sha256sum' and 'sha512sum'. This is an oversight, I will recompile Busybox today to build sha1sum.

Another hash utility that is in Wary/Racy is 'cksum', which is part of 'coreutils'. So, perhaps a nice little drag-and-drop hash generator could automatically test what is installed, md5sum, sha1sum, cksum, and generate accordingly. Such a utility should also offer testing of a file, that a downloaded hash is correct.

ePDFView 0.1.8 crashes in Wary/Racy

Testing in both Wary and Racy 5.1.99, ePDFView 0.1.8 crashes if try to print.

I recompiled poppler 0.16.7 and ePDFView, no difference. I rolled ePDFView back to version 0.1.7, now printing works.

I looked at the ePDFView changelog:
http://trac.emma-soft.com/epdfview/wiki/ChangeLog#Version0.1.82011-05-28

...I see that support for poppler >= 0.15 is added. So, the older ePDFView is working ok with poppler 0.16.7, presumably just not using it's newer features. Which should be fine, as long as they haven't changed the API in poppler.

I'll leave it at that, ePDFView 0.1.7 and poppler 0.16.7. If there is any trouble, I can roll poppler back also.

GUI shutdown

Most readers would know that shinobar developed 'pupsaveconfig', which provides GTK dialog windows to create a save-file (to save session at first shutdown), rather than the text-mode dialog windows that have always been offered in Puppy.

Here is the Forum thread on pupsaveconfig 1.9.9:
http://www.murga-linux.com/puppy/viewtopic.php?t=60678

Excellent work, but I never put that into Woof, as I wanted to do it a bit differently. Over the last few days I have implemented it the way I want.

For the sake of efficiency, I wanted all of the code in /etc/rc.d/rc.shutdown that presents the dialog windows to work with or without X -- without then use 'dialog', with then use ...well, Xdialog would be the usual choice.

I took all of the code concerned with saving the session out of rc.shutdown, and created /usr/sbin/shutdownconfig. This is called from rc.shutdown when shutting down without X running. However, when shutting down from the desktop, the two scripts that handle exiting from X, /usr/bin/wmpoweroff and /usr/bin/wmreboot, now call shutdownconfig. So, the same code is called, with or without X running.

I wrote 'pupdialog', which has simple compatibility to 'dialog', but runs in X with GTK dialog windows. Currently it supports the --msgbox, --yesno, --menu and --inputbox dialogs. It uses gtkdialog3 or gtkdialog4.

What happens now, is at the first boot (or "pfix=ram", that is, PUPMODE=5), when power-off or reboot is chosen from the Shutdown menu, dialog windows now appear on the desktop, before X exits. So, you get a nicer interface and can use the mouse for selecting.

If you still want the old non-X shutdown dialogs, just exit from X to the prompt and type in "poweroff" or "reboot".

This will be in the next upload of Woof. It is not yet internationalized, I want to be sure that it works properly first.

'pupdialog': 'dialog' with X

I am writing a little replacement for 'dialog', for when X is running. It uses gtkdialog, and is intended to be a drop-in replacement for 'dialog'.

This came about yesterday when I started to convert one of my scripts that uses 'dialog', to have nicer interfaces. Rather than a piece-meal approach of replacing each call to dialog with a block of code for gtkdialog, which is what I started to do, I thought it would be simpler in the long run to write a replacement for dialog.

Hence 'pupdialog'. This will be in Woof, so available for all scripts. I will probably get it to call 'dialog' if no X, so "pupdialog" can be simply put into scripts replacing "dialog", and those scripts will then work with or without X.

Last night I was playing with code to parse the commandline. I came up with this, which is very compact:

#evaluate the passed parameters...

#ex: --timeout 90 --extra-button --ok-label "SAVE TO A FILE"  --extra-label "SAVE TO ${xPDEV}" --cancel-label "DO NOT SAVE" --yesno "etetetete" 0 0

AVAR='START'; AVAL=''

for AFIELD in ${@}

do

 case $AFIELD in

  --timeout|--extra-button|--ok-label|--extra-label|--cancel-label|--yes-label|--no-label|--yesno) eval "${AVAR}='${AVAL}'"; AVAR="`echo -n "$AFIELD"| tr -d '-' | tr '[a-z]' '[A-Z]'`"; AVAL='' ;;

  *) AVAL="${AVAL}${AFIELD} " ;;

 esac

done

AVAL="`echo "$AVAL" | tr '\t' ' ' | tr -s ' ' | sed -e 's% $%%'`"

WIDTH="`echo -n "$AVAL" | rev | cut -f 1 -d ' ' | rev`"

HEIGHT="`echo -n "$AVAL" | rev | cut -f 2 -d ' ' | rev`"

AVAL="`echo "$AVAL" | rev | cut -f 3-999 -d ' ' | rev`"

eval "${AVAR}='${AVAL}'"

udev 167 with all rules

Mavrothal put in a request for this. The 'common' repo has a udev 167 PET that has cut-down and customised rules especially for Puppy.

I have created PET packages of the original udev 167 with all of the rules. These are for experimentation purposes.

PETs (169KB, 3KB, 32KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/udev-167-patched_t2-w5.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/udev_DEV-167-patched_t2-w5.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/udev_DOC-167-patched_t2-w5.pet

'scanpci' broken in Racy

the Xorg 7.6 Mega Package post-install script is supposed to delete any pre-existing 'scanpci' executable, but failed to do so.

Note, for Xorg 7.5+ we use 'lspci' instead.

I have fixed the package and uploaded it:
http://distro.ibiblio.org/quirky/pet_packages-wary5/x_xorg76_mega_pkg-7.6-1-w5c.pet

Meebo chat client

Linuxcbon posted about a typo in /usr/local/bin/meeboview, which I fixed.

'meeboview' is just a simple script that opens with site in the default web browser:
http://www.meebo.com/

But, it was back in Puppy 3.0.1 that we actually used this script, with a menu entry under the "Internet" category. (though, I don't know how we used it, as it has had a typo all these years)

There is really no reason (that I know of) why we should not have this menu entry by default in Puppy, as it gives us a multi-protocol chat client for free (no size added to Puppy). So, I grabbed the /usr/share/applications/meebo-chat.desktop file out of Puppy 3.0.1. Now in Woof.

Broadcom brcm80211 firmware

Thanks to the post by jim3630, this came to my attention:
http://murga-linux.com/puppy/viewtopic.php?t=71777
And thanks to tempestuous for creating the firmware PET.

The firmware is now in Woof. In Woof, kernel-skeleton/lib/modules/all-firmware/brcm.tar.gz is the firmware tarball. The file kernel-skeleton/etc/modules/firmware.dep specifies that the brcm.tar.gz tarball will load into /lib/firmware when the brcmfmac.ko and brcmsmac.ko modules need it.

A new feature in Woof is script kernel-skeleton/pinstall.sh. This is executed by script 3builddistro and will delete brcm.tar.gz if the modules are not there. This is to save space, as the firmware is rather large.

This will be in the next upload of Woof, so puppies built from that Woof and with the 2.6.39.x+ kernels will support those Broadcom drivers out-of-the-box.

Woof uploaded, Sept. 17, 2011

This is commit '20110917224225'. This version was used to build Wary Puppy 5.1.99 and Racy Puppy 5.1.99.

The previous upload of Woof was August 29:
http://bkhome.org/blog/?viewDetailed=02450

There was a blog post recently on how to download a recent snapshot tarball of Woof and use that as the reference to download Woof changes since then, without the full history:
http://bkhome.org/blog/?viewDetailed=02306

To download and use Woof, please read the Bones intro page:
http://bkhome.org/bones/
And the Woof intro page:
http://bkhome.org/woof/

Wary 5.1.99 (5.2 alpha)

This is the first alpha of what is to become Wary 5.2.

Download:
http://distro.ibiblio.org/quirky/test/wary5.1.99/

You might not see any difference from previous Wary versions. Lots of packages upgraded though, so it is mostly changes "under the bonnet".

There is one known bug. I have not yet determined the cause of this or exactly what circumstances cause it. After the first shutdown and creating a save-file, on next bootup the BootManager runs -- clicking any button in the BootManager hangs it, in fact hangs all applications, although the mouse and keyboard still work.

The workaround, if this happens, is to press CTRL-ALT-BACKSPACE, then "xwin" to restart X, then when the BootManager starts don't try to use it -- instead, right click on the BootManager item in the taskbar and choose "Kill".

The weird thing is, if you start the BootManager via the "System" menu, it works ok.

It's a weird bug, and I am letting it rest in my brain for awhile. Will feel inspired to solve it soon.

Racy 5.1.99 (5.2alpha)

I stopped uploading earlier, as it got too slow. However, I did manage to upload Racy. Wary is still uploading.

Racy 5.1.99 is "Wary on steroids", the same as Wary except built with Xorg 7.6 (X server 1.11.0) and 2.6.39.4 PAE-enabled i686 kernel.

So, like Wary, but designed for very modern hardware.

Note that for now I have left the "file prefix" as "wary", for example 'warysave.2fs'. A bit later I intend to change that to "racy" so as to differentiate it some more from Wary.

The intention is that the entire collection of apps for Wary will also run in Racy. I have been aiming for very close binary compatibility, despite the bump in the Xorg and Mesa libraries.

The Xorg 7.6 Mega Package can also be installed in Wary, which means that you will get support for later video hardware. But, very important, you will still be using the 2.6.32.x kernel, which might not have drivers for the latest hardware.

For example, my new laptop -- the 2.6.32.x kernel does not have a driver for the ethernet interface, the 2.6.39.x kernel does.

The 2.6.32.x kernel does not have the 'nouveau' nVidia kernel driver, so although the Xorg 7.6 Mega Package has the 'nouveau' Xorg driver, it won't work without the corresponding kernel driver (the post-install script detects the inappropriate kernel and deletes the nouveau Xorg driver so it won't cause any conflict).

Regarding nVidia, with Wary you do have the Xorg 'nv' driver, or you could install the commercial nVidia driver.

With Racy, you will have a choice of all three.

Wary and Racy have my new 'stark' theme. An experiment, I have been looking at it for almost a week now and I am already tired of it. So, it will probably only have this one public appearance.

Anyway, anyone who would like to play with Racy and provide feedback, get it from here:
http://distro.ibiblio.org/quirky/test/racy-5.1.99/

Pfilesearch, Pfind, Pschedule, pupRadio, wcpufreq

I just did a round of upgrading, getting things together for next releases.

Zigbert's apps:
Pfilesearch 1.25 -> 1.26
Pfind 4.22 -> 4.23
Pschedule 1.0.1 -> 1.0.2

01micko's app:
pupRadio 0.11 -> 0.14

Tazoc's app:
wcpufreq 0.7 -> 0.8

These will all be uploaded to the 'common' repo soon.

Wary "5.2" and Racy progress

Just a quick note. Things are looking good. I built "Racy", which is Wary except with Xorg 7.6 and 2.6.39.4 PAE kernel, using it now.

A bit on the chubby side, 134MB, but it has SeaMonkey 2.3.2 and the full Mesa (with all the DRI drivers, which are enormous -- many puppies are built with only the software-rasteriser DRI driver). I chose gzip compression for the SFS file in the live-CD, but this kernel supports xz compression of Squashfs files so that could be done if smaller size is desired -- Woof asks which one to use at build time.

Anyway, I have a small to-do list, probably need one more day. Then there are a huge number of PETs to upload, likely to take several hours. After that, expect Wary 5.2-beta and Racy uploads.

Woof has lots of fixes too, so I will also be uploading that. In 1-2 days.

xf86-video-vesa 2.3.0 crashes

When I recompiled everything in T2, I used the latest xf86-video-vesa, 2.3.0. I used this for both the Xorg 7.3 compile and the Xorg 7.6 compile.

Running the standard Wary, that is with Xorg 7.3, I was getting strange behaviour with the vesa driver. On two out of three computers tested, the desktop came up okay, but clicking anywhere crashed X. X did log a message to /tmp/xerrs.log that it was "waiting to exit" or words to that effect, but no information anywhere about why it was waiting to exit.

I found that rolling back to xf86-video-vesa 1.3.0 fixed it, at least on one PC tested so far.

I just checked Debian Lenny, they use xserver-xorg-video-vesa 1.3.0. Lenny is an Xorg 7.3+ system.

However, the Xorg 7.6 Mega package has vesa 2.3.0 in it.

usb-modeswitch 1.1.9

The last upgrade was 1.1.7 back in March 2011:
http://bkhome.org/blog/?viewDetailed=02165

I have upgraded to 1.1.9, and the data to 2011-08-05 (27KB, 14KB):
http://distro.ibiblio.org/quirky/pet_packages-common/usb-modeswitch-1.1.9-patched-w52.pet
http://distro.ibiblio.org/quirky/pet_packages-common/usb-modeswitch-data-20110805.pet

The PET is not quite a pristine usb-modeswitch, I have patched a line into /lib/udev/usb_modeswitch to call my script /usr/sbin/usb_modeswitch_status.
Also, logging is turned on in /etc/usb_modeswitch.conf, required by my script.

Usb-modeswitch is partly broken on my laptop, and I have been discussing this with Josh on the usb-modeswitch forum:
http://www.draisberghof.de/usb_modeswitch/bb/viewtopic.php?t=743

However, I implemented a workaround in usb_modeswitch_status, that will create /dev/gsmmodem if it has not been auto-created. This is just a temporary fix until Josh fixes it officially.

2.6.39.4 kernel

I have compiled the 2.6.39.4 kernel, to use in what I have tentatively named "Racy Puppy".

These days, we have complete automation of patching and compiling the kernel and 3rd-party drivers, plus automatic generation of the kernel source SFS and kernel binary PET package.

I have uploaded everything needed to compile the 2.6.39.4 kernel:

http://bkhome.org/sources/kernel-2.6.39.4/
Us#er#na#me: pu#pp#y Pa#ss#wo#rd: li#nu#x

...that site has a readme file that explains usage.

A question was asked recently about my web page that explains how to compiled the kernel, whether it was still relevant. No, it is very dated. I have edited it:
http://puppylinux.com/development/compilekernel.htm

Touchpad etc., fix for Xorg 7.6

This now running so beautifully on my laptop. I have fixed some things with the automatic setup of X.

When the user installs the Xorg 7.6 Mega Package, a message pops up to advise exit from X and run either "xorgwizard" or "xorgwizard-automatic", the latter recommended as first attempt.

I have fixed some things in xorgwizard-automatic and in /etc/x11/xorg.conf-auto-pc (that the Wizard reads for skeleton xorg.conf file). These are fixes in Woof.

Touchpad can be configured now, Flsynclient works.

I'm running with the Intel Xorg driver, at 1366x768. Regarding speed, I was earlier today playing with Fatdog64, and did some rough timings. Running Wary with Xorg 7.6, the timings are about the same. So, I can't see any difference in speed.

I have tentatively decided to name this enhanced Wary "Racy Puppy".

Next thing on the to-do list, I want to compile the most recent 2.6.x kernel to use with Racy, mostly because I want the 'nouveau' Xorg driver support (nVidia), which the 2.6.32 kernel doesn't have. I cannot yet go up to a 3.x kernel as there are some scripts in Puppy that are hard-coded to expect "2.6".

And, looking ahead, I will probably build Racy in Woof, all ready to go with Xorg 7.6 and recent kernel.

New "stark" theme for Wary

The Wary "5.2 alpha" will have an experimental theme, collectively named "stark".

It is composed of:

Wallpaper: wary_stark.png (rootfs-skeleton/usr/share/backgrounds in Woof)
Icon set: desk_icon_theme_stark042-1.pet (based on area042)
JWM theme: jwm_theme_stark-1.pet
GTK theme: gtk_theme_stark-1.pet

The PETs will be uploaded later.

This theme will not necessarily be in 5.2-final, it is an interesting experiment with, well, starkness.

Xorg 7.6 Mega Package for Wary

I have created a "Mega Package" to upgrade Wary from Xorg 7.3 to 7.6. Note, I was earlier referring to it as an upgrade to 7.5, but in fact it is 7.6.

This upgrade PET will be made available when the first alpha of Wary 5.2 is released. Soon.

The good news, on my new laptop (Acer Aspire 5742G) the Intel driver now works. With Xorg 7.3 I had to use the 'vesa' driver, and as '915resolution' cannot recognise my Intel hardware, I cannot run at the native 1366x768', instead have to use 1024x768.
But, with the proper Intel driver I now have 1366x768 resolution.

Another bit of good news is that /usr/sbin/xorgwizard-automatic (the fully automatic Xorg Wizard) got me to the desktop with correct Intel driver and resolution.

There is also some bad news...

The 'vesa' driver is incredibly slow. Everything is sluggish, the mouse pointer is jerky, scrolling is painful. Unusable.

Over the last three years I have come to expect the Intel Xorg driver to be less than satisfactory. Unfortunately, my expectation was once again fulfilled.
It works ok, seems fairly snappy but I haven't stress-tested it yet. What is annoying is the bits of cruft left behind on the screen -- Intel can't seem to get that right. It is only sometimes, we can live with it, well, we have to.

More things to do. Next-up will be to fix Flsynclient, which also needs to be upgraded to work with Xorg 7.6.

JWM menu icon

I was playing with a new theme for my experimental Wary "5.2", when I noticed that the 'area042' icon set has a file 'start-button.png' in it (I think that theme was put together by forum member steve_s, but not sure).

That triggered my memory. Back in the Community Edition Puppy 4.1.2, it was popular to have a special icon for the start-menu, instead of the little dog picture with text "Menu" alongside it that is in current Woof-built puppies. I think that they hard-coded the name of this as /usr/local/lib/X11/pixmaps/start-button.png.

If any icon theme has this same 'start-button.png', then it will be used by JWM if that theme is chosen. Well, that was the case back in 4.1.2.

I have decided to support this in Woof. When a theme is chosen in '3builddistro' when building a Puppy, if that theme has 'start-button.png' then rootfs-complete/root/.jwmrc-tray is modified to use that icon instead of the little-doggy-and-Menu-text.

I also modified rootfs-complete/usr/sbin/icon_switcher and icon_switcher_cli to do the same. If the chosen theme has 'start-button.png' then is is used, if not then reverts to the little doggy.

I have only done this for JWM.

gcc was broken

I was exasperated last night, in my latest experimental "5.2" build, gcc was broken. It complained:

error: stdarg.h: no such file or directory
error: stddef.h: no such file or directory

It took me hours to fix, I was getting fed up. But eventually I found a post that explained exactly what the problem was. Some files were missing from my gcc package.

My T2 build environment still has the bug that has been reported before in this blog, it requires that I have to recompile gcc in a running Puppy and create new PET packages. Yeah, I did that, but the "new2dir make install" left out some files. Those two, 'stdarg.h' and 'stddef.h', plus some more were missing, they were supposed to be here:

/usr/lib/gcc/i486-t2-linux-gnu/4.3.4/include

The post that solved it for me, thanks to that guy. He/she explained that if you run:

gcc -v hello.c -o hello

...the "-v" shows whereabouts gcc is looking for it's inbuilt include files, and that is how I discovered the missing files.

Yay, gcc is working!

Fatdog64 520 final

Fatdog64 is a 64-bit x86 version of Puppy Linux. Kirk has announced a new release version, 520:

http://murga-linux.com/puppy/viewtopic.php?t=71387

So, you guys with 64-bit CPUs might want to give this pup a run! I also, finally, have a PC with 64-bit CPU, my recently acquired Acer Aspire 5742G, so I will be downloading fatdog64 also.

Update package for Lucid Puppy 5.2.8

Playdayz has sent me this info:

Instant Update 001 for Lucid Puppy 5.2.8 is now released.
It is recommended that everyone using Lucid Puppy 5.2.8 download and install the Instant Update.

Announcement on LupuNews
http://www.diddywahdiddy.net/LupuNews/

Main Lucid Puppy 5.2.8 release page
http://www.murga-linux.com/puppy/viewtopic.php?t=70855

Instant Update 001 for Lucid Puppy 5.2.8 adds two new programs, CloudApps and PupSysInfo in PupControl. It also enables use of the latest Chromium-based browsers without having to install anything else, other than the browsers. There are also assorted bug fixes and upgrades.
Details of Instant Update 001 for Lucid Puppy 5.2.8
http://www.murga-linux.com/puppy/viewtopic.php?p=560890#560890

Backup X drivers fixed

There is something that I have been meaning to fix for sometime, but don't have the problem with Xorg 7.3 so it got put on the back-burner.

Puppy has a mechanism whereby Xorg drivers can be "hidden" by putting them in /usr/lib/xorg/modules/drivers-alternate. The place where all the "available" drivers is is in /usr/lib/xorg/modules/drivers. The '3builddistro' script in Woof offers to put any X drivers into 'drivers-alternate' -- for example, if you have the 'intel' driver and also are building with the old 'i810' driver then you would probably want to put the latter into 'drivers-alternate'. Another example is the 'nouvea' and 'nv' drivers.

The problem is though, that the Xorg 7.5 server scans everything in /usr/lib/xorg/modules, and tries to use the modules that are supposedly hidden.

The same problem occurs in the Xorg Wizard when a specific driver is chosen, say 'vesa' -- to prevent the server from trying to use any other driver, all other drivers are moved to /usr/lib/xorg/modules/drivers-backup. But with Xorg 7.5 this mechanism is rendered useless.

Therefore, I have moved the 'drivers-alternate' and 'drivers-backup' directories to /usr/lib/x. The scripts modified are '3builddistro' and rootfs-skeleton/usr/sbin/xorgwizard.

Wary "5.2" progress

A couple of days ago I reran the compile in T2, and then imported all of the T2 binary packages to PETs. I now have a big collection of PETs that will be uploaded soon.

I have built Wary, currently at version 5.1.94, running it now, looking real good.

Give me another 4-5 days, and I should have the first beta of Wary "5.2" ready to go.

I also expect to have an upgrade PET for Xorg 7.5, that is intended to be seamlessly compatible with Wary and all apps. We shall see.
...that really has been the whole point of my "5.2" experiment.

Woof '2compat2pets', '3compat2pets'

This is very specialized, only of interest to me and perhaps one or two others.

I reran the compile of packages in T2, with Xorg 7.3, mesa 6.5.2. I reran it because I wanted all package versions other than Xorg and mesa to be the same versions for the build with Xorg 7.5. Previously, there were different versions of pixman, which I thought might perhaps cause a conflict.

Anyway, having done the compile in T2, I created two new scripts in Woof to convert all of the T2 binary packages into PETs. It is a rather roundabout way of doing it. With file DISTRO_SPECS setup with:

DISTRO_BINARY_COMPAT='t2'
DISTRO_FILE_PREFIX='t2test'
DISTRO_COMPAT_VERSION='9'

I ran these scripts, in this order:

1 ./0pre
2 ./0setup
3 ./support/2compat2pets
4 ./2createpackages
5 ./3builddistro
6 ./support/3compat2pets

This has created a directory 'new-pets' with every binary package compiled in T2 converted to PETs. File 'z-new-pets-db' has all of the database entries.

Running 2createpackage and 3builddistro is not for building a usable Puppy. It only sets things up for 3compat2pets to perform the conversion to PETs with maximum optimizations.

Afterward, I can merge these PETs into directory packages-pet-wary5 and then change DISTRO_SPECS back to building a Wary Puppy.

SeaMonkey 2.3.2

I have just compiled SeaMonkey 2.3.2.

I don't know... the quality control seems to have slipped, they are bringing out new releases without proper testing.

I reported that the Composer component in SM 2.2 is broken:
http://bkhome.org/blog/?viewDetailed=02397

Ok, so now the image doesn't duplicate when I click on it. But, the right-click menu still has "Image properties..." greyed-out. I can select "Format/Image Properties..." from the main menu, and that works ok.

I never reported this bug myself, as I thought there are so many users of SM, and this is such an obvious bug, it is sure to be fixed.

It is not a show-stopper bug, just very confusing for someone using Composer for the first time.

SM 2.3.2 installs into /usr/lib/seamonkey-2.3.1, another sign of a rush job.

Was away

I was away from home all day yesterday, so nothing happened in the Puppy department.

Home this morning, and back onto the experimental Wary "5.2" (I really should have a codename for this, not just "5.2"!)

I am running the T2 builds again, for both cases, compiling with Xorg 7.3 then again with Xorg 7.5. I fixed a few things the first time around.

Once I have the T2 build compiling nicely, I intend to make it into a big tarball and upload it, so others will be able to reproduce my work. Although based on the official T2 from SVN, my T2 is heavily customised.

dhcpcd problem

Just a quick note, in case anyone else is experiencing this problem.

I am building experimental Wary's, and I find that when I connect to my wireless network, dhcpcd drops out.

However, if I install the dhcpcd PET (from Wary ibiblio repo) -- the same PET as used in the build by Woof -- dhcpcd then works.

I tested this a few times, got same result. That would indicate that some file in the dhcpcd PET is getting changed by some other package when built in Woof.

This is on my to-to list. But, for anyone experiencing same problem, try installing the PET.

ePDFView 0.1.8

I have compiled the latest. Project page:

http://trac.emma-soft.com/epdfview/

Poppler 0.16.7

I noticed that the version of Poppler (PDF rendering library, used by ePDFview) compiled in T2 is rather old, so I have compiled the latest, 0.16.7.

Hmmm, it complained about a conflict in zlib.h, it turns out that zlib.h in zlib 1.2.5 (the latest) is broken. Most of the distros patch it, but T2 doesn't. The stock zlib.h in zlib 1.2.5 works for most cases.

Anyway, I patched it:
http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/sys-libs/zlib/files/zlib-1.2.5-lfs-decls.patch?revision=1.3

Then Poppler compiled ok.

I also modified the template for Poppler in Woof, to create extra symlinks for those cases where apps are linked against older versions of libpoppler*.

This is happening in my experimental Wary "5.2".

Wary "5.2" massive recompile

Continuing my Wary "5.2" experiment, all of these packages have been recompiled in T2:

a52dec-0.7.4 abiword-2.8.6 apr-1.4.5 apr-util-1.2.12 atk-1.30.0 audiofile-0.2.7 autoconf-2.64 automake-1.10.2 bdb-4.7.25 bin86-0.16.18 binutils-2.20.51.0.2 bison-2.4.1 boehm-gc-7.1 bzip2-1.0.6 cairo-1.10.2-noxcb cdrkit-1.1.11 chmlib-0.40 cmake-2.8.2 cpio-2.10 cups-1.3.11 curl-7.21.7 cvs-1.11.23 dialog-1.1-20100428 dietlibc-0.32 diffstat-1.54 diffutils-3.0 dmidecode-2.11 dosfstools-3.0.11 dvdauthor-0.7.0 dvd+rwtools-7.1 e2fsprogs-1.41.14 eject-2.1.0 epdfview-0.1.7 ethtool-6 exiv2-0.21 expat-2.0.1 faac-1.28 file-5.03 findutils-4.4.2 flac-1.2.1 flex-2.5.33 foomatic-filters-4.0.7 freeglut-2.6.0 freetype-2.4.5 fribidi-0.19.2 fuse-2.8.5 gawk-4.0.0 gdb-7.2 gdbm-1.8.3 gdk-pixbuf-2.22.1 gettext-0.17 gftp-2.0.19-patched1 ghostscript-8.15.4 git-1.7.5.4 glade3-3.6.7 glib-2.28.8-noxcb glibc-2.10.1-2 glibc_dev-2.10.1-2 glibc_locales-2.10.1-2 gmp-4.3.1 gnumeric-1.10.16 goffice-0.8.16 grep-2.9 groff-1.20.1 gtk+-2.24.5-noxcb gtkam-0.1.17 gutenprint-5.2.7 gzip-1.4 hicolor-icon-theme-0.5 ifplugd-0.28 imake-1.0.2 inotify-tools-3.13 intltool-0.40.6 iptables-1.4.12 jasper-1.900.1 lame-398-2 lcms-1.19 libao-1.1.0 libart-2.3.21 libcap-2.17 libcddb-1.3.2 libcdio-0.82 libdaemon-0.14 libdv-1.0.0 libdvdcss-1.2.10 libdvdnav-4.1.3-1 libdvdplay-1.0.1 libdvdread-4.1.3-1 libexif-0.6.19 libexif-gtk-0.3.5 libgnomecanvas-2.30.3 libgnomecups-0.2.3 libgnomeprint-2.18.8 libgnomeprintui-2.18.6 libgsf-1.14.21 libid3tag-0.15.1b libidl-0.8.14 libidn-1.22 libieee1284-0.2.11 libjpeg-7 libmad-0.15.1b libmng-1.0.10 libogg-1.2.2 libpciaccess-0.12.1 libpng-1.2.44-patched_apng_t2 libpthread-stubs-0.1 libraw1394-2.0.7 librsvg-2.32.1-noxcb libsigc++-2.2.10 libsndfile-1.0.20 libtheora-1.1.1 libtiff-3.9.5 libtool-2.2.6a libusb-1.0.8 libusb-compat-0.1.3 libvorbis-1.3.2 libxml2-2.7.8 libxslt-1.1.26 m4-1.4.13 madplay-0.15.2b make-3.81 man-1.6f-1 mesa-6.5.2 mesa_base-6.5.2 mesa-demos-7.6 mktemp-1.7 module-init-tools-3.16 mpfr-3.0.1 mtr-0.80 nasm-2.07 ncurses-5.9 neon-0.29.6 netpbm-10.34 net-tools-1.60-1 normalize-0.7.7 ntfs-3g-2010.8.8 ntfsprogs-2.0.0 openssl-1.0.0d pango-1.28.4-noxcb parted-2.4 patch-2.6.1 patchutils-0.3.2 pciutils-3.1.7-ids_20110511 pcmciautils-017 pcre-8.12 perl-5.10.1 perl-digest-sha1-2.12 perl-extutils-depends-0.302 perl-extutils-pkgconfig-1.12 perl-html-parser-3.64 perl-uri-1.40 perl-xml-parser-2.36 perl-xml-simple-2.18 pixman-0.18.4-1 pkgconfig-0.26 popt-1.16 ppp-2.4.5 procps-3.2.8 psmisc-22.14 python-2.6.4 readline-6.0 rman-3.2 rsync-3.0.8 rxvt-unicode-9.12 sane-backends-1.0.22 sed-4.2.1 setserial-2.17 sqlite-3.7.5 startup-notification-0.9 subversion-1.6.17 sysfsutils-2.1.0 sysprof-1.1.8 tar-1.26 texinfo-4.13a unzip-60 util-linux-2.18 vala-0.13.1 vamps-0.9.5 vcdimager-0.7.23 vorbis-tools-1.4.0 wget-1.12 wireless-tools-29 wpa_supplicant-0.7.3 x11proto_noxcb-7.3 xcompmgr-1.1.5 xload-1.0.2 xorg73_base_t2_noxcb-7.3 xsane-0.997-patched xtrans-1.0.4 zip-30 zlib-1.2.5

These have been compiled in T2 and have upgraded versions:

Format: name new-version old-version
apr 1.4.5 1.2.12
aspell 0.60.6.1 0.60.6
aspell-en 7.1-0 6.0-0
audiofile 0.2.7 0.2.6
bin86 0.16.18 0.16.17
bzip2 1.0.6 1.0.5-1
cairomm 1.10.0 1.8.4
cdrkit 1.1.11 1.1.9
cmake 2.8.2 2.6.2
curl 7.21.7 7.19.6
dejavu-ttf-fonts 2.32 2.30
dhcp 4.2.0 4.1.1b1
diffstat 1.54 1.50
diffutils 3.0 2.8.1
dmidecode 2.11 2.10
dosfstools 3.0.11 3.0.6
dvdauthor 0.7.0 0.6.14
e2fsprogs 1.41.14 1.41.9
ed 1.5 1.3
enscript 1.6.5.2 1.6.1
exiv2 0.21 0.18.2
fontconfig 2.8.0 2.7.3
foomatic-filters 4.0.7 3.0-20090114
freetype 2.4.5 2.3.11
fuse 2.8.5 2.7.4
gawk 4.0.0 3.1.6
gdb 7.2 7.0
geany 0.20 0.19.1-1
gftp 2.0.19-patched_2ubuntu1_expandedpathsigsegv 2.0.19-patched1
git 1.7.5.4 1.6.5.2
glib 2.28.8 2.24.2-1
gnome-icon-theme 2.30.3 2.28.0
gnumeric 1.10.16 1.10.9
goffice 0.8.16 0.8.9
grep 2.9 2.5.4
gtk+ 2.24.5 2.20.1-1
gutenprint 5.2.7 5.2.6
gzip 1.4 1.3.11
hdparm 9.37 9.26
iptables 1.4.12 1.4.5
lame 398-2 3.98.4
lcms 1.19 1.18
leafpad 0.8.18.1 0.8.16
libao 1.1.0 0.8.8
libart 2.3.21 2.3.20
libavc1394 0.5.4 0.5.3
libdc1394 2.1.3 2.1.2
libdvdnav 4.1.3-7 4.1.3-1
libdvdread 4.1.3-9 4.1.3-1
libexif 0.6.19 0.6.18
libgnomecanvas 2.30.3 2.26.0
libgnomeprint 2.18.8 2.18.6
libgnomeprintui 2.18.6 2.18.4
libgsf 1.14.21 1.14.16
libidl 0.8.14 0.8.13
libidn 1.22 1.15
libogg 1.2.2 1.1.4
libpcap 1.1.1 1.0.0
libraw1394 2.0.7 2.0.2
libsigc++ 2.2.10 2.2.4.2
libtiff 3.9.5 3.8.2-1
libusb 1.0.8 0.1.12
libvorbis 1.3.2 1.2.3
memtester 4.2.1 4.1.2
mkinitrd 2010-08-18 2005-12-26
mktemp 1.7 1.6
module-init-tools 3.16 3.11.1
mpfr 3.0.1 2.4.1
mpg321 0.2.13-2 0.2.11
mtr 0.80 0.75
ncurses 5.9 5.7
neon 0.29.6 0.29.0
ntfs-3g 2010.8.8 2009.4.4
openssl 1.0.0d 0.9.8k
orc 0.4.14 0.4.9
pango 1.28.4 1.28.1
pangomm 2.26.3 2.26.2
parted 2.4 2.3
patch 2.6.1 2.5.9
patchutils 0.3.2 0.3.1
pcmciautils 017 016
pcre 8.12 8.00-1
pkgconfig 0.26 0.23
ppp 2.4.5 2.4.4
psmisc 22.14 22.8
qt4 4.7.2 4.5.3-1
rdesktop 1.7.0 1.6.0
rsync 3.0.8 3.0.6
rxvt-unicode 9.12 9.05-1
samba 3.5.9 3.4.2
sane-backends 1.0.22 1.0.21
schroedinger 1.0.10 1.0.9
sdparm 1.06 1.04
shadow 4.1.4.3 4.0.18.1
subversion 1.6.17 1.6.6
sylpheed 3.0.3 2.7.1
sysprof 1.1.8 1.0.12
taglib 1.7 1.6
tar 1.26 1.22
usbutils 0.91 0.86
util-linux 2.18 2.16.1-1
vala 0.13.1 0.9.8
valgrind 3.6.1 3.5.0
vorbis-tools 1.4.0 1.2.0
wpa_supplicant 0.7.3 0.6.9
xcompmgr 1.1.5 1.1.4
xf86-video-apm 1.2.3 1.1.1
xf86-video-ark 0.7.3 0.6.0
xf86-video-ati 6.9.0-patched_lenny4 6.7.195
xf86-video-chips 1.2.4 1.1.1
xf86-video-cirrus 1.2.1 1.1.0
xf86-video-geode 2.11.11 2.11.3
xf86-video-glint 1.2.5 1.1.1
xf86-video-i128 1.3.4 1.2.1
xf86-video-i740 1.3.2 1.1.0
xf86-video-neomagic 1.2.5 1.1.1
xf86-video-rendition 4.2.4 4.1.3
xf86-video-s3 0.6.0 0.5.0
xf86-video-s3virge 1.10.4 1.9.1
xf86-video-savage 2.2.1 2.1.3
xf86-video-siliconmotion 1.6.0 1.5.1
xf86-video-sis 0.10.3 0.9.3
xf86-video-sisusb 0.9.0 0.8.1
xf86-video-tdfx 1.4.3 1.3.0
xf86-video-tga 1.2.1 1.1.0
xf86-video-trident 1.3.4 1.2.3
xf86-video-tseng 1.2.4 1.1.1
xf86-video-vesa 2.3.0 1.3.0
xf86-video-voodoo 1.2.4 1.1.1
xfsprogs 3.1.5 3.0.5
zlib 1.2.5 1.2.3

I have built it, running it now. A couple of things need fixing, but it is looking good.

Wary "5.2" progress

I have got my nose down, busy with the experiment. The proper way to eliminate any phantom dependencies on libxcb is to recompile everything from source, which is what I have done in T2.

Yesterday afternoon I wrote a big script to update all of the Wary PET packages from the packages compiled in T2.

I also took the opportunity of upgrading lots of packages, a report on that will follow.

I have to manually check every PET package, especially when the script has performed a version upgrade on it.

I had previously named all Wary packages in the format --w5.pet, but the ones that have just been recompiled in T2, regardless of version upgrade or not, are now named --w52.pet.

About one-third of the way through the manual checking of the PET packages, just having a bit of a diversion right now.

GTKThemeMaker 1.4, JWMThemeMaker 1.5

These two are created by trio. I have updated JWMThemeMaker in the 'common' repo, and have add GTKThemeMaker.

Forum threads:
http://murga-linux.com/puppy/viewtopic.php?t=45397
http://www.murga-linux.com/puppy/viewtopic.php?t=45702

PETs (10KB, 6KB):
http://distro.ibiblio.org/quirky/pet_packages-common/gtkthememaker-1.4.pet
http://distro.ibiblio.org/quirky/pet_packages-common/jwmthememaker-1.5.pet

...ah, such lovely tiny size!

Pnethood, Pprocess, Wallpaper

I have modified these packages to call 'gtkdialog4' instead of 'gtkdialog3':

http://distro.ibiblio.org/quirky/pet_packages-common/pnethood-0.6.8-2.pet
http://distro.ibiblio.org/quirky/pet_packages-common/pprocess-2.2.4-1.pet
http://distro.ibiblio.org/quirky/pet_packages-common/wallpaper-0.6.1-1.pet

Forums threads:
Pnethood: http://murga-linux.com/puppy/viewtopic.php?t=70532
Pprocess: http://murga-linux.com/puppy/viewtopic.php?t=23024
Wallpaper: http://www.murga-linux.com/puppy/viewtopic.php?t=69658

...you will see in those threads that these require the new "development version" of gtkdialog by thunor. These are major applications released by zigbert, clarf and 01micko, that called 'gtkdialog3', and this was one factor that required me to upgrade gtkdialog to the development version, but ended up causing grief with the latest Wary 5.1.4 when we found old scripts were broken.

The clean break now is to name the development version 'gtkdialog4'.

The apps that caused grief in Wary 5.1.4 were zigbert's Pfind, Pburn, Pfilesearch, Pmirror, Pbackup:
http://bkhome.org/blog/?viewDetailed=02452

I have reverted Pmirror to version 0.4:
http://distro.ibiblio.org/quirky/pet_packages-common/pmirror-0.4.pet
http://www.murga-linux.com/puppy/viewtopic.php?t=13606

Pfind is ok, have left it at unmodified version 4.22:
http://distro.ibiblio.org/quirky/pet_packages-common/pfind-4.22.pet
http://www.murga-linux.com/puppy/viewtopic.php?t=15136

Pfilesearch, have left that unmodified at version 1.25, presume it still works with the old gtkdialog3:
http://distro.ibiblio.org/quirky/pet_packages-common/pfilesearch-1.25.pet
http://www.murga-linux.com/puppy/viewtopic.php?t=26764

Pburn, have left that unmodified at version 3.3.5, presume still works with old gtkdialog3:
http://distro.ibiblio.org/quirky/pet_packages-common/pburn-3.3.5.pet
http://www.murga-linux.com/puppy/viewtopic.php?t=23881

Pbackup, have left that unmodified at version 3.1.9, presume works with the old gtkdialog3:
http://distro.ibiblio.org/quirky/pet_packages-common/pbackup-3.1.9.pet
http://www.murga-linux.com/puppy/viewtopic.php?t=10975

In Woof, I have modified all of the package-selection files to build with both the old gtkdialog3 and the new gtkdialog4 PETs. That is, variable PKGS_SPECS_TABLE now has this:

no|gtkdialog2||exe
yes|gtkdialog3||exe|
yes|gtkdialog4|gtkdialog4|exe| #modify all pets that need this to explicitly call 'gtkdialog' or 'gtkdialog4'.

CORRECTION 14 Sept 2011:
It should be:

no|gtkdialog2||exe
yes|gtkdialog3||exe,dev,doc>dev,nls|
yes|gtkdialog4||exe,dev,doc>dev,nls| #modify all pets that need this to explicitly call 'gtkdialog' or 'gtkdialog4'.

Big experiment for "Wary 5.2"

I am conducting a big experiment for the next jump in Wary, which will be version 5.2.

1. Remove libxcb
The libxcb used with Xorg 7.3 has an incompatible API with the current libxcb. I think this may cause applications to fail if I want to apply a big Xorg upgrade PET to Wary with a later libxcb.
The solution is to remove libxcb entirely from Wary. An upgrade Xorg PET can introduce libxcb and there will be no clash.

2. GTK 2.24.5
This is the last of the 2.x series. I don't think that I will ever go up to GTK3, I'll retire before that. It makes sense to move Wary up to this version, and from then on, no need to upgrade ever again, unless they bring out a 2.24.x bugfix release.

With regard to point-1, the vast majority of applications can be compiled without dependency on any of the xcb libraries. There are some apps however, and I can't think which right now, that do require libxcb -- they cannot be used in the base Wary system, not without an Xorg upgrade.

What I'm trying to do is set up Wary to very easily lift itself up by it's bootstraps and run later Xorgs, so work on any video hardware. Without library conflicts.

So, how far have I got with this experiment? I'm running it right now. This is what I have been working on, when I very reluctantly got dragged back to do the emergency bugfix for Wary 5.1.4.

Packages that have been upgraded/recompiled/removed so far:

pango 1.28.1 -> 1.28.4

librsvg 2.32.1 -> 2.32.1

libX11 (inside xorg73_base_t2 mega-package) 1.1.3 -> 1.1.3

xcb-proto (inside x11proto mega-package) 1.0 removed

libxcb 1.0 removed

glib 2.24.2 -> 2.28.8

cairo 1.10.2 -> 1.10.2

gtk 2.20.1 -> 2.24.5

gdk-pixbuf 2.22.1 new

zlib 1.2.3 -> 1.2.5

glibc 2.10.1 -> 2.10.1

libxml 2.7.6 -> 2.7.8

There is a slight difference with glibc, I am using kernel 2.6.32 headers, upgraded from 2.6.30, hence needed to recompile glibc.

Not everything is working. Abiword and Gnumeric for example look for libxcb, so I will recompile those. SeaMonkey is working, using it now.

Interesting, I have just recompiled gtksourceview, as it wanted libxcb. During compile it wanted libxcb.la so I just created a dummy file and it compiled. So something in the system is still telling to look for libxcb.la and libxcb-xlib.la.

gtkdialog4

Due to the problem with the development version of gtkdialog breaking scripts, I have implemented a solution. Now that I know that my previous expectation that the new gtkdialog would not break existing scripts is wrong, the solution is to have a clean break.

I have reinstated the old 'gtkdialog3' PETs in the 'common' repo:
http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog3-0.7.20-patriot-e-1.pet
http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog3_DEV-0.7.20-patriot-e-1.pet

The latest development release from thunor's SVN site:
http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog4-246-w5.pet
http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog4_DOC-246-w5.pet

As we have now been able to retire 'gtkdialog2', we are now moving up to two executables 'gtkdialog3' and 'gtkdialog4'. The gtkdialog4 package has 'gtkdialog' as a symlink to 'gtkdialog4'.

To fully implement this, I will have to go through all the PETs that require gtkdialog4 and change the executable name from 'gtkdialog3' to either 'gtkdialog' or 'gtkdialog4'.

If any developers who use the development version of gtkdialog want to anticipate this, please put in a test for the existence of gtkdialog4 and use that instead of gtkdialog3.

The clear break also means that developers can put "+gtkdialog3" or "+gtkdialog4" in the dependencies field of pet.specs, so there it is definitely known which one is required. Furthermore, if a particular SVN revision is required, say 246: "+gtkdialog4&ge246".

Wary Puppy 5.1.4.1 released

This is an emergency bugfix upgrade of Wary 5.1.4. The problem is described in this blog post:
http://bkhome.org/blog/?viewDetailed=02452

Release notes:
http://distro.ibiblio.org/quirky/wary-5.1.4.1/release-Wary-5.1.4.1.htm

Download:
http://distro.ibiblio.org/quirky/wary-5.1.4.1/

If you already have downloaded 5.1.4, the upgraded PETs can be installed, so no need to upgrade to 5.1.4.1. The PETs are in the top (blog) link.

NOTICE TO ALL DEVELOPERS
If you maintain any script that uses gtkdialog, please test with the gtkdialog PET that is in the 'common' repo:
http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog-0.7.21-thunor224-w5.pet

This is the version of gtkdialog that is currently chosen by default in all Woof-built puppies. I would like to be sure that all the scripts out there run on it. Unless some big bug is found in it, I intend to leave it at that version (revision 224) for some considerable time -- therefore, if you are testing with a later revision of gtkdialog, please test that it still works with 224.

Note, as well as the bugfix for the file-select dialog, there was another thing broken in some scripts, the password-entry box:
http://bkhome.org/blog/?viewDetailed=02423

EDIT Sept 1 2011:
A solution for the gtkdialog problem: http://bkhome.org/blog/?viewDetailed=02455

PNG wallpaper

Up until now, the images in /usr/share/backgrounds had to be JPG. Woof script '3builddistro' asks you to choose which one you want as default, and renames it to 'default.jpg'.

ROX-Filer is started in /root/.xinitrc and uses /root/Choices/ROX-Filer/PuppyPin to define various things on the desktop, including background which is set as "default.jpg".

/root/.xinitrc also executes /usr/sbin/background_reshape, which defaults to operate on /usr/share/backgrounds/default.jpg, cropping it top and bottom for a widescreen.

I have modified Woof '3builddistro' to change the name to 'default.png' if a PNG image is chosen as default, and to modify 'PuppyPin' accordingly. Script 'background_reshape' now recognises either default.jpg or default.png.

Pburn, Pmirror, Pbackup, Pfind broken

These are all broken in the latest Wary 5.1.4. I'm annoyed about this. It is due to the upgrading of gtkdialog breaking scripts that use it. I reported about this to the gtkdialog development thread:
http://murga-linux.com/puppy/viewtopic.php?t=69188&start=270

Zigbert has fixed Pburn (now 3.3.5), Pfilesearch (now 1.25), Pmirror (now 0.5) and Pbackup (now 3.1.9) and I have uploaded them to the 'common' repo.

But, I wonder what other scripts are out there that are now broken? My understanding was, from comments posted by thunor sometime ago, was that existing scripts would not be broken. if I had known that this was going to happen, I would have named it 'gtkdialog4'.

It is a bit too late to do that now, however, I am freezing the version of gtkdialog that is available in the 'common' repo. It is revision 224 from thunor's SVN site.
http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog-0.7.21-thunor224-w5.pet

We will have to go through all the old PETs that we can find, any that are designed to use 'gtkdialog3', and thoroughly test them.

If you have Wary 5.1.4, you can install the fixed PETs:
http://distro.ibiblio.org/quirky/pet_packages-common/pburn-3.3.5.pet
http://distro.ibiblio.org/quirky/pet_packages-common/pfilesearch-1.25.pet
http://distro.ibiblio.org/quirky/pet_packages-common/pmirror-0.5.pet
http://distro.ibiblio.org/quirky/pet_packages-common/pbackup-3.1.9.pet

I will be releasing an emergency bugfix version of Wary very soon.

EDIT Sept 1 2011:
A solution for the gtkdialog problem: http://bkhome.org/blog/?viewDetailed=02455

Pcdripper, close-dvd+rwtools

I have upgraded to plinej's latest audio CD ripper, version 3.8.4.

Zigbert has a utility available, a binary executable named 'close', in PET package 'close-dvd+rwtools'. This is used by Pburn to fixate DVDs, if present.

I don't know where the source code of 'close' is.

I have put both of these in the 'common' repo, will upload soon.

Woof uploaded, August 29, 2011

This is commit '20110829142202'. This version was used to build Wary Puppy 5.1.4.

The previous upload of Woof was August 28:
http://bkhome.org/blog/?viewDetailed=02446

There was a blog post recently on how to download a recent snapshot tarball of Woof and use that as the reference to download Woof changes since then, without the full history:
http://bkhome.org/blog/?viewDetailed=02306

To download and use Woof, please read the Bones intro page:
http://bkhome.org/bones/
And the Woof intro page:
http://bkhome.org/woof/

Wary Puppy 5.1.4 released

This is a bugfix and minor upgrade of Wary 5.1.3. Wary is an edition of Puppy Linux that supports older hardware, including analog modems for those on dialup, and old video hardware.

The full announcement and release notes are here:
http://distro.ibiblio.org/quirky/wary-5.1.4/release-Wary-5.1.4.htm

Some extra notes:

1. Theme
The Wary 5.1.x series has had the same theme throughout and 5.1.4 is no exception. I do plan an interesting new theme for 5.2.

2. Kernel
5.1.4 has an SMP-enabled kernel that has a configuration change that fixes a nasty kernel crash and other mysterious behaviour. I have also compiled a uniprocessor-only kernel but not built 5.1.4 with it -- apparently there is still some old hardware that won't boot with a SMP kernel, so if you have such hardware there is a build of Wary 5.1.3.4 with the uniprocessor kernel: http://distro.ibiblio.org/quirky/test/wary-5.1.3.4/wary-5.1.3.4-uni.iso
-- I am interested in knowing of anyone with hardware that requires such a kernel.

The Wary 5.1.4 live-CD (120.5MB):
http://distro.ibiblio.org/quirky/wary-5.1.4/wary-5.1.4.iso

'devx' SFS file for complete compiling support (120.1MB):
http://distro.ibiblio.org/quirky/wary-5.1.4/devx_wary_5.1.4.sfs

Delta files to upgrade from 5.1.3:
http://distro.ibiblio.org/quirky/wary-5.1.4/delta_files

Forum thread for discussion:
http://murga-linux.com/puppy/viewtopic.php?t=67572&start=315

nVidia download fixed

Sage reported that the nVidia driver upgrade was broken in Wary 5.1.3.4. Yes, I found out why. Although I don't have an nVidia card, I can temporarily fool the script into thinking there is one, which is what I should have done before, to verify that the script works.

I have fixed scripts /usr/local/video_upgrade/video_upgrade_wizard and /usr/local/video_upgrade/which_nvidia_pkg.

Desktop drive icons fix

Those who are using a Puppy built from recent Woof will notice on the first boot that the drive icons draw along the bottom of the screen rather slowly.

One-second intervals in fact. I had put a "sleep 1" into the code to delay the icon changing from mounted to unmounted, as there is a problem with a USB activity-light flashing up to a second beyond the drive being unmounted (and 'sync' completed).

However, this 1-second is also delaying the initial drawing of the icons. Furthermore, the 1-second delay was only supposed to occur for USB drives but is happening for all.

So, I have changed the test for the 1-second delay in /etc/rc.d/functions4puppy4:

icon_unmounted_func() {

 #passed params: drivename category  ex: sda usbdrv

 DRVICON="drive48.png" #default.

 [ "$2" = "usbdrv" ] && DRVICON="usbdrv48.png"

 [ "$2" = "card" ] && DRVICON="card48.png"

 [ "$2" = "floppy" ] && DRVICON="floppy48.png"

 [ "$2" = "optical" ] && DRVICON="optical48.png"

 #find out if already done...

 if [ -f /root/.config/rox.sourceforge.net/ROX-Filer/globicons ];then

  GLOBLINES="`cat /root/.config/rox.sourceforge.net/ROX-Filer/globicons | tr '\n' '|' | sed -e 's%|%\n%g'`"

  diPATTERN='/root/.pup_event/drive_'"${1}"'"'

  [ "`echo "$GLOBLINES" | grep "$diPATTERN" | grep "$DRVICON"`" != "" ] && return

 fi

 #[ "`echo "${ATADRIVES}" | grep "${1}"`" = "" ] && sleep 1 #110722 usb drive, LED light briefly flashes after unmounted, so wait a bit. ATADRIVES is in /etc/rc.d/PUPSTATE.

 [ "$2" = "usbdrv" ] && sleep 1 #110828 above test doesn't work.

 echo "



 

  

   /root/.pup_event/drive_${1}

   /usr/local/lib/X11/pixmaps/${DRVICON}

  

 

"  | rox -R

}

I have no idea why the previous test does not work. This change should fix it, at least now the internal drive icons will draw fast.

Woof uploaded, August 28, 2011

This is commit '20110828152702'. Used to build Wary 5.1.3.4 (5.1.4RC).

The previous upload was on August 23:
http://bkhome.org/blog/?viewDetailed=02431

There is one slight uncertainty: I modified the downloading of nVidia X driver (in rootfs-complete/usr/local/video_upgrade) to handle split packages (nvidia_base and nvidia_kmodule) -- haven't yet verified that works -- but I think that I retained backwards compatibility (a single pkg) ok.

Wary 5.1.3.4 (5.1.4 RC)

Here it is:
http://distro.ibiblio.org/quirky/test/wary-5.1.3.4/

There are two builds, one with my modified-SMP configuration that I announced yesterday, the other with a uniprocessor-configured kernel.

rodin.s,
I am most keen to know if this fixes your USB crashing problem!

To anyone who has nVidia graphics, please test. I want to know that the Video Upgrade Wizard downloads both the 'nvidia_base' and 'nvidia_kmodule' packages. I wasn't able to test it myself.

Pcdripper 3.8.2

I have upgraded the PET on the 'common' repo to plinej's latest CD-ripper:
http://murga-linux.com/puppy/viewtopic.php?t=24157

Commercial nVidia drivers

My three scripts that compile the 2.6.32.x kernel have fully automated the whole thing, as reported earlier in this blog. I have now added compiling of the commercial nVidia kernel modules.

Furthermore, the script also creates the PETs.

I have split the nVidia PETs into two, 'nvidia_base' and 'nvidia_kmodule'. The former has all the library , executable, data and doc files, the latter has a kernel module.

Uploading them now, for both the 'smp' and 'uni' 2.6.32-45 kernel.

Next thing I will do is update the Video Upgrade Wizard to download both packages.

Probably tomorrow I will upload another Wary build, and Woof -- which has some important fixes.

2.6.32.45 SMP works

I reported about the kernel crashing when I plug in a USB drive, and rodin.s has the same problem:
http://bkhome.org/blog/?viewDetailed=02439

I compiled a uniprocessor kernel, and it did not crash:
http://bkhome.org/blog/?viewDetailed=02441

I decided to experiment with the SMP configuration choices. A little while ago, including the 2.6.30.x kernel that I compiled for Puppy 4.3.1, I configured with SMP enabled, but with these two disabled:

[ ] SMT (hyperthreading) scheduling support
[ ] Multi-core scheduler support

But people kept telling me that they should be enabled, so for the 2.6.32 kernel used in Wary, they are.

Well, guess what, I have just tried with those two disabled, and now I have been plugging and replugging a USB pen drive and just can't get the kernel to crash.

It is probably only one of those that is the culprit, but I will leave it at that. The next Wary will have SMP kernel but with those two items disabled.

You still get SMP support. My Intel i3 CPU is recognised as having 4 cores, and mkquashfs does its thing much faster than with the uniprocessor kernel.

Linux 2.6.32.45 uniprocessor

I have configured the 2.6.32.45 kernel for uniprocessor systems, and with 'tickless' disabled. otherwise, it is the same as the SMP version as used in Wary 5.1.3.

The PET (for Woof developers) (28.7MB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_kernel-2.6.32-uni-45-i486-w5.pet

The source:
http://bkhome.org/sources/kernel-2.6.32-uni-45/

The '1kernel.sh' compile script now allows setting of any string to name the kernel, and this one is named "2.6.32-uni", so the modules are in /lib/modules/2.6.32-uni, and that is what it identifies itself as when 'uname -r' is executed.

Drive icons missing border

Broomdodger reported this:
http://bkhome.org/blog/?viewDetailed=02432

I tested, ran Wary 5.1.3, created a save-file, frugal installation. Then upgraded to 5.1.3.2, and the mounted drive icons reverted to the old "balls" instead of the borders.

I have fixed it. Modified '3builddistro', 'rc.update', 'icon_switcher_cli'. This will be in the next Woof upload.

Wary: uniprocessor kernel?

Wary is targeting older hardware, most of which are uniprocessor systems. An SMP-enabled (multi-core CPU) kernel does not work on all old uniprocessor CPUs, even when the appropriate kernel boot command to turn off SMP is used -- I was reminded of this recently in a pm from someone.

The thing is, even if a uniprocessor kernel is run on a multi-core CPU system, I don't think that users would notice the difference in speed. Except for some multimedia uses such as playing videos. But even then, playing a DVD or whatever, you are going to get sufficient Frames Per Second running on one core, as evidenced by DVDs that I have happily played on various old systems.

So, I am thinking of releasing the next Wary with a uniprocessor kernel, but perhaps I should "put a toe in the water first" and release it alongside the build with SMP kernel. Users can then compare.

I could even roll the kernel back to the old IDE pre-PATA drivers (yes I know that IDE and PATA mean much the same thing, but we use these terms to designate two different generations of drivers in the kernel. PATA and SATA are the new generation).

FatDog64 RC2

Kirk's FatDog64 has reached Release Candidate 2:
http://murga-linux.com/puppy/viewtopic.php?t=70976

The thought has just occurred to me -doh- my new laptop has a Intel i3-core x86 64-bit CPU, so I can finally test FatDog64.

Puppy projects: Bright Ideas

I already announced the "Puppy Projects" web page yesterday:
http://bkhome.org/blog/?viewDetailed=02436

I received feedback with great ideas for some very ambitious projects, plus I thought of some myself, so I have added a "Bright Ideas" section to the web page, just scroll down:
http://puppylinux.com/development/projects.htm

So, in the future if anyone asks on the Forum how they can get involved on the developer side of things, point them to this page!

Puppy projects!

The page says it all:

http://puppylinux.com/development/projects.htm

SoftMaker Office 2008 PET

SoftMaker Office 2008 is free, but requires registration to use. I have made it into a PET package. This is fully integrated into Puppy, I worked on it all morning to make it so. Mime handling with ROX-Filer also.

The first time that you run it, my special 1st-start script runs, which takes you to the SoftMaker site to obtain the license-key.

SoftMaker Office has three components, TextMaker word processor, PlanMaker spreadsheet, and Presentations. They are all highly MS-Office compatible.

I split the original package into three. The main one, which is all you need for English users (12.1MB):
http://distro.ibiblio.org/quirky/pet_packages-common/softmakeroffice-2008-bk.pet

Documentation for English and German (13.2MB):
http://distro.ibiblio.org/quirky/pet_packages-common/softmakeroffice_DOC-2008-bk.pet

International language support (24.2MB):
http://distro.ibiblio.org/quirky/pet_packages-common/softmakeroffice_NLS-2008-bk.pet

...that last one is massive. It doesn't even have any extra documentation.

One small thing that I wasn't able to sort out: the menu font is not anti-aliased. Maybe there is some trick to it. If anyone can figure that one out, please let me know.

So, a Puppy could be built with SoftMaker Office builtin, instead of Gnumeric and Abiword. I haven't worked it out, but I think the latter two plus their support packages would total about 5-6 MB. So, the total size would grow a little bit, but maybe worth it?

InkscapeLite is good

I was just reminded, I read a post in the forum recently that InkscapeLite is no good for anything. I also read a similar post sometime ago.

Not so! I use InkscapeLite a lot. The main reason that I use it is to create composite bimap images. Vovchik has created an overlay application for PNG images, but all along we have had InkscapeLite that does this very well, and then some.

You can import PNG images, arrange them as required, insert text, line-art, then export as a PNG bitmap image. Works great. I used it today, wanted to overlay the SoftMaker icon on top of an empty document image, to create a mime-type icon for SoftMaker documents.

InkscapeLite is in most puppies. It is very small, one of the main reasons it is inbuilt.

'common' repo in PPM

I have added the Puppy PET 'common' repo as a default in the Puppy Package Manager in Wary.

The repos that are selected by default are chosen in script '3builddistro' in Woof, I modified this part that selects which PET repos:

 case $DISTRO_COMPAT_VERSION in #100911

  lucid)

   enabledrepos="Packages-puppy-lucid-official Packages-puppy-common-official "

  ;;

  wary5)

   enabledrepos="Packages-puppy-wary5-official Packages-puppy-quirky-official Packages-puppy-common-official Packages-puppy-4-official " #110824 added -common-

  ;;

  slacko) #110824

   enabledrepos="Packages-puppy-slacko-official Packages-puppy-common-official "

  *)

I also added the common repo to Lucid, and put in a case for Slacko, so Slacko will have the "puppy-slacko" and "puppy-common" repos as default (plus two or more Slackware repos).

Wary 5.1.3.2 (5.1.4beta)

This is a beta for the upcoming Wary Puppy 5.1.4, for testers only, not yet for public release.

Download:
http://distro.ibiblio.org/quirky/test/wary-5.1.3.2/

...note the delta files are very small, very easy to upgrade from 5.1.3 if you are on dial-up.

5.1.4 is intended to be a bugfix and minor upgrade from 5.1.3. More extensive upgrades of applications are targeted for 5.2.

Here is the blog announcement for the public-released 5.1.3:
http://bkhome.org/blog/?viewDetailed=02399

Forum thread for feedback, go to page 20 for discussion of 5.1.3.2:
http://murga-linux.com/puppy/viewtopic.php?t=67572&start=285

Woof uploaded, August 23, 2011

Commit number '20110823183411'.

Now Woof is sane! I fixed a bug in '3builddistro', also in script 'findwoofinstalledpkgs'.
I did a run-through, built Wary 5.1.3.2, running it now.

The previous build of Woof was August 22, 2011:
http://bkhome.org/blog/?viewDetailed=02427

There was a blog post recently on how to download a recent snapshot tarball of Woof and use that as the reference to download Woof changes since then, without the full history:
http://bkhome.org/blog/?viewDetailed=02306

To download and use Woof, please read the Bones intro page:
http://bkhome.org/bones/
And the Woof intro page:
http://bkhome.org/woof/

Muted sound at first start

pemasu reported a fix for muted sound, by adding this line 'set Speaker 75 % unmute' into /etc/init.d/10alsa:
http://murga-linux.com/puppy/viewtopic.php?t=70399&start=270

Yes, I have the same problem on my new laptop, have to unmute and raise the level in the "Speaker" slider in Retrovol. So, pemasu's fix seems appropriate to me.

I have inserted the new line into line 60 of /etc/init.d/10alsa.

I see also that 01micko has suggested some changes also, but there doesn't seem to be any feedback on that.

Note, the fact that I have picked up this fix on page 19 of the Puppy Exprimo thread does not mean that I have read all the pages of that thread! I just followed a link into it from somewhere, oh, yeah, I was looking for something don570 posted there. So, there might be other relevant fixes buried there that I should know about?

Woof: kernel selection

Selection of which kernel to use in a build needs to be done early-on, as some packages are kernel-version-dependent, so selection of the right packages to use in a build requires that the kernel is already selected.

The current uploaded Woof now has kernel-selection in the first tab of the woof_gui GUI.

However, if not using the GUI and doing the build by running the scripts directly, the kernel selection is in 3builddistro. I have now moved this to 1download.

I reorganized 1download so that online kernel PETs are downloaded near the beginning of the script, then the kernel is chosen. After that, the correct packages can be chosen and downloaded.

Science Fiction

Lobster found this:
http://murga-linux.com/puppy/viewtopic.php?t=70984

Fascinating! I love SciFi. I enjoy settling down in the evening with a SciFi book ...if anyone has surplus SciFi books that they don't want, I want them! I'll pay for the postage.

I inherited this from my dad. He had a collection, and I read every one. He subscribed to a couple of the short-story paperback monthlies of the time. He doesn't have those anymore, I never asked him what happened to them.

I do however have a few of his old books, for example "Adventures in Time and Space" published in 1946.

In a couple of years I will get rid of them, as I will probably be moving to a small unit, and will sell or give them away. If anyone wants to stake a claim on them, let me know!

Woof uploaded, August 22, 2011

Commit number '20110822214018'.

This one needs testing, as there is major infrastructure change, see blog posts over last 5 days. Maybe someone could try a complete build of Slacko in Woof, starting from 0setup, then 1download, etc.

The previous upload of Woof was August 17, 2011:
http://bkhome.org/blog/?viewDetailed=02413

There was a blog post recently on how to download a recent snapshot tarball of Woof and use that as the reference to download Woof changes since then, without the full history:
http://bkhome.org/blog/?viewDetailed=02306

To download and use Woof, please read the Bones intro page:
http://bkhome.org/bones/
And the Woof intro page:
http://bkhome.org/woof/

Pnethood 0.6.8-1

clarf's Pnethood is at version 0.6.8:
http://murga-linux.com/puppy/viewtopic.php?t=70532

However, it has a bug with latest gtkdialog, the password entry box does not show anything. This has been mentioned with regard to pupdial.

I hacked it and created 0.6.8-1, the PET is in the 'common' repo on ibiblio.org.

Pprocess 2.2.4, Precord 7.0.1

I have updated to zigbert's latest Pprocess:
http://murga-linux.com/puppy/viewtopic.php?t=23024

I have updated to mcewanw's latest Precord:
http://murga-linux.com/puppy/viewtopic.php?t=49907

...comment, it seems odd that Precord has a hidden directory in /etc. Why would you want to hide it?

These will be in the 'common' repo, I will upload soon.

Woof improvements for Slacko

01micko posted that KMS (Kernel ModeSetting) needs to be turned off for the Xorg 'radeon' driver in Slacko:
http://murga-linux.com/puppy/viewtopic.php?t=70844&start=90

I have modified script '3builddistro' to choose appropriate defaults for 'i915', 'radeon' and 'nouveau', with options to override.

Pupdial fixes

Scripts that use gtkdialog with "password" input boxes (typed-in text hidden as "x" or "*" characters) are broken with the latest gtkdialog from SVN. Nothing shows in the entry box, although it still works.

01micko has fixed this in /usr/sbin/pupdial, and has also rearranged the GUI window to fit a smaller screen. He also rearranged the window for /usr/sbin/modemprobe:
http://murga-linux.com/puppy/viewtopic.php?t=70844&start=150

I have put these two scripts into Woof (they are not in a separate PET).

As far as I know, the only other script that uses password-type input box is Pnethood, and that also needs to be upgraded.

Puppy package database format

I have updated the web page that introduces the package database format used in Puppy:

http://bkhome.org/woof/pkg-db-format.htm

gtkdialog 0.7.21 revision 224

Thunor is still working intensely on gtkdialog, and zigbert's latest Pschedule requires revision 224 from SVN (it fixes a bug).

I have compiled this and placed it in the 'common' repo (74KB, 42KB):

http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog-0.7.21-thunor224-w5.pet
http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog_DOC-0.7.21-thunor224-w5.pet

In Woof, this entry in PKGS_SPECS_TABLE will cause gtkdialog_DOC-0.7.21-thunor224-w5.pet to go into the 'devx':

yes|gtkdialog||exe,dev,doc>dev,nls

...at least it does after I have fixed it in Woof. The current uploaded and all earlier Woofs don't do the doc>dev redirection properly. Woof will be uploaded soon.

That 4th field is still a work-in-progress. It is supposed to define what components of a package end up in the Puppy build.

Woof: '2createpackages' updated

It has been a long day. I have updated the '2createpackages' script in Woof to use the new 'findpkgs' script. It required massive reorganisation of 2createpackages, but I seem to have got there.

I am just about to let it run right through building the generic packages for Slacko.

Woof: '1download' updated

I reported last night that Woof has a "magic" new script for finding packages to use in a build, 'findpkgs':
http://bkhome.org/blog/?viewDetailed=02414

It is a lot of work, I start about 6am and work for most of the day, evening too. I finally got 'findpkgs' to work properly, then I turned my attention to '1download'.

1download had to be modified to use findpkgs. The proof of the pudding is in the eating, so I am running 1download to download the packages for Slacko. So far, so good.

Unfortunately I have not availed myself of ttuuxxxes offer, as need to run 1download myself to make sure that it works as expected.

Heaps of packages downloading, looking good!

I still have about three days of solid work left I think to update the rest of Woof.

Lucid Puppy 5.2.8 released

Lucid Puppy is our "flavour" or edition of Puppy Linux that is built from Ubuntu binary packages, hence has compatibility with, and access to, the vast Ubuntu package repositories.

Coordinator "playdayz" (Larry Short) has made this short announcement:

Lucid Puppy 5.2.8 is the fastest and friendliest Lucid yet. It is the fastest because it is the first Lucid to use the C and ffmpeg libraries optimized for i686 computers rather than the older i386 computers. Lucid 5.2.8 has also received a thorough going-over under the hood. There is new and updated firmware and drivers for many devices, courtesy of forum member tempestuous, and the hardware detection and configuration routines have been extensively tested and enhanced by forum member rerwin. The overall look and feel has been influenced by experiments forum member pemasu, to make the user interface fun for Linux newbies and experienced users alike.

For a longer announcement and release notes:
http://distro.ibiblio.org/puppylinux/puppy-5.2.8/release-Lucid-528.htm

Download the live-CD iso file (129.2MB):
http://distro.ibiblio.org/puppylinux/puppy-5.2.8/lupu-528.iso

'devx' SFS file for complete compiling support (135.1MB):
http://distro.ibiblio.org/puppylinux/puppy-5.2.8/lupu_devx_528.sfs

UPDATE NOTICE JAN 8, 2012
Although Lucid 5.2.8 is the last official release, the developers have continued to improve Lucid and have issued 'updates', as PET packages, or you can download an ISO (live-CD) file built with the update. It is recommend to get the latest. Go to the ibiblio.org to find the latest to download:
http://distro.ibiblio.org/puppylinux/puppy-5.2.8/
Or, faster download mirrors can be found here:
http://puppylinux.com/download/
Go here for information about the updates:
http://www.murga-linux.com/puppy/viewtopic.php?t=70855

Lucid Puppy 5.2.5 was the previous release, on April 2, 2011, see announcement:
http://bkhome.org/blog/?viewDetailed=02203

Here is the Forum thread for discussion on development of Lucid 5.2.5:
http://www.murga-linux.com/puppy/viewtopic.php?t=65136
Here is the development discussion prior to release of Lucid 5.2.8:
http://www.murga-linux.com/puppy/viewtopic.php?t=69563
For discussion of the release version 5.2.8, please go here:
http://www.murga-linux.com/puppy/viewtopic.php?t=70855

...note, the above Forum link has more detailed release notes for 5.2.8 in the first post of the thread.

About Puppy "flavours" or editions
For the uninitiated, Puppy Linux releases are built with the Woof build system (http://bkhome.org/woof/), which can build a Puppy from the binary packages of many major distros. The result is binary compatibility with that distro, yet with all the speed and usability of Puppy. We currently have two official "flavours" of Puppy, Lucid (built from Ubuntu binary packages) and Wary (packages originally compiled from source in T2, so not binary-compatible with any particular major-distro). Wary targets older hardware, whereas Lucid has more "bells and whistles" and support for very modern hardware.

If you look on the Puppy Forum, you will find several other flavours of Puppy. For example, there is Fatdog64 for x86-64 CPUs, Slacko Puppy based on Slackware packages -- both of these are at beta status.
Correction: FatDog64 is at Release Candidate stage.

Netpbm rolled back

It looks like I have to get back to Bryan, the netpbm developer, a bug in latest Netpbm from svn. I reported recently that I had created a PET in the 'common' repo of latest from svn:
http://bkhome.org/blog/?viewDetailed=02404

However, Slacko beta1 testers have reported it is not working, and they fixed the problem by rolling back:
http://www.murga-linux.com/puppy/viewtopic.php?p=554413#554413

So, I have rolled back Netpbm in the 'common' repo -- that will be in next Woof upload.

PPM: syntax error in 'installpkg.sh'

rcrsn51 found an error in /usr/local/petget/installpkg.sh:
http://www.murga-linux.com/puppy/viewtopic.php?t=70844&start=60

Thanks, fixed.

Slacko 4.99.1 (beta1)

01micko's Slacko Puppy has reach the first beta:
http://www.murga-linux.com/puppy/viewtopic.php?t=70844

This pup is going to be one of our "official" releases. Testers are welcome!

Note to 01micko:
I have just downloaded the file 'DISTRO_PKGS_SPECS-slackware-13.37.gz' for beta. One little detail: could you edit it so that comments are now on the end of lines, like this:

Instead of this:



#util-linux-ng has libuuid,libblkid replacing those in e2fsprogs...

yes|e2fsprogs|e2fsprogs|exe,dev,doc,nls



Do this:



yes|e2fsprogs|e2fsprogs|exe,dev,doc,nls| #util-linux-ng has libuuid,libblkid replacing those in e2fsprogs.

The reason is that one of the Woof scripts has an option that sorts the table, and the comments can end up in the wrong place. If they are on the end of the line, they will stay in place. Be sure to append a '|' delimiter as I have shown.

Woof progress: magic 'findpkgs' script

I have written a new script support/findpkgs in Woof, that replaces a whole lot of klutzy code and adds great new features.

Using Slacko as the example, file DISTRO_PKGS_SPECS-slackware-13.37 has variable PKGS_SPECS_TABLE that has a list of entries that specify what packages are to be in the Puppy build, for example:

yes|glibc|glibc-solibs,glibc-zoneinfo|exe,dev,doc,nls

yes|glibc_locales|glibc|exe,dev,doc,nls>exe| #slackware glibc-i18n- does not have enough files, need to extract from full glibc.

no|glibc_i18n_en||exe

yes|glibmm||exe>dev

no|glipper||exe,dev>null,doc,nls

yes|gmeasures||exe,dev>null,doc,nls

If the 3rd field is empty, see gmeasures entry, it means find a package named 'gmeasures' in the PET repositories.
If the 3rd field is not empty, see glibc entry, then look for packages named 'glibc-solibs,glibc-zoneinfo' (glob wildcards allowed) in the compatible-distro repos (Slackware repos) and combine them into a package named 'glibc' for inclusion into the Puppy build.

My new script has backwards compatibility, but you can add more stuff to narrow or force the search. Generalizing, the format is, where [] designate optional:

yes|genericpkgname|[pkgnames]|[splitup]|[pet:[repo]]
yes|genericpkgname|[pkgnames]|[splitup]|[compat:[repo]]

Take the gmeasures example, we can extend like this:
yes|gmeasures|gmeasures-2*|exe,dev,doc,nls|pet:4

...the optional 3rd field can be used to narrow the search to a specific version. The 'pet:4' means only look in the Packages-puppy-4-official db file.

Actually, the 'pet' on the end overrides the older interpretation, forcing search of PET repos, regardless of whether anything in 3rd field. For example:
yes|gmeasures||exe,dev,doc,nls|compat
...means look for package in compat-distro repos.

In the case of Slacko, the compat-distro package db files are:
Packages-slackware-13.37-official
Packages-slackware-13.37-salix
Packages-slackware-13.37-slacky

If we wanted to force 'glibc' to be got from the 'salix' repo only:
yes|glibc|glibc-solibs,glibc-zoneinfo|exe,dev,doc,nls|compat:salix

Note, as before, as well as wildcards, the '-' exclusion prefix is still supported, for example:
yes|xorg-server|xf86-video-*,xf86-input-*,-xf86-video-cirrus|exe,dev,doc,nls

'findpkgs' has very rigorous searching, with well-defined rules. It takes into account the versioned-dependencies in the dependencies field of a pkg db entry, and there can now be more than one condition, for example:
|+linux_kernel&ge2.6.32<2.6.33|

This is a progress report. I have a couple of things to do to finish the script, then it has to be integrated into Woof.

Woof uploaded, August 17, 2011

This is commit '20110817071638'. Previous upload was August 15:
http://bkhome.org/blog/?viewDetailed=02410

01micko has provided files that he uses to build Slacko in Woof, plus I made some small mods. So, this Woof should be able to build the same Slacko that 01micko has posted on the forum.

I haven't tested it though. My Telstra 3G costs me AU$15 per GB, and I use it up so incredibly fast. I buy it prepaid, in 10GB lots for AU$150, and I'm lucky if that lasts me for 2 months.

To test Slacko would probably use about 1GB download. Anyway 01micko and anyone else who is interested can try it. I'll probably be doing a Debian build soon so will go through the whole download thing for that.

Note, just to let everyone know, I might be a bit unresponsive for the next few days. I am just about to embark on some major re-coding in Woof, which will require all my attention.

Xarchive minor bug fix

Rerwin posted a fix for a bug in Xarchive, our archive creator/extractor:
http://murga-linux.com/puppy/viewtopic.php?t=67572&start=225

I have implemented this. Here is the fixed PET (38KB):
http://distro.ibiblio.org/quirky/pet_packages-common/xarchive-0.2.8-8-p4.pet

Pupscan improvement

There has been a lot of activity with the /usr/sbin/pupscan script recently. 01micko converted it to work with gtkdialog3 instead of gtkdialog2, sc0ttman posted a fix for the exit button when running in Lucid.

Rerwin has posted an improvement to distinguish between kernel inbuilt-drivers and modules. That patch is for Lucid, and did not work for me, so I have implemented it a different way.

The output of Pupscan now displays whether a driver is builtin to the kernel or a module.

Rerwin also made a little modification to improve line spacing, which I have put in.

The improved Pupscan will be in the next upload of Woof.

Woof uploaded (again), August 15, 2011

I already uploaded Woof today:
http://bkhome.org/blog/?viewDetailed=02407

I have fixed the problem of booting from USB when there is no hard drive, so uploading Woof again.

The latest commit is '20110815174956'.

Reason cannot boot USB when no HD

I have found out why there are recent reports that Puppy will not boot from USB when there is no hard drive.

I did fix it awhile ago, which I reported on this blog. So, I was puzzled.

I found the reason, I had upgraded Busybox in the initrd, from 1.4.2 to 1.18.5:
http://bkhome.org/blog/?viewDetailed=02341

I tried to move over to using more Busybox applets, but ended up going back to the full versions -- except for 'grep', I kept the Busybox 'grep' applet as it seemed ok.

Until now, I have discovered it is not ok:

# echo 'abc' | grep -v -f emptyfile 

abc

# echo 'abc' | ./busybox grep -v -f emptyfile 

BusyBox v1.18.5 (2011-07-09 23:09:48 GMT-8) multi-call binary.



Usage: grep [-HhnlLoqvsriwFEz] [-m N] [-A/B/C N] PATTERN/-e PATTERN.../-f FILE [FILE]...



Search for PATTERN in FILEs (or stdin)



Options:

	-H	Add 'filename:' prefix

	-h	Do not add 'filename:' prefix

	-n	Add 'line_no:' prefix

	-l	Show only names of files that match

	-L	Show only names of files that don't match

	-c	Show only count of matching lines

	-o	Show only the matching part of line

	-q	Quiet. Return 0 if PATTERN is found, 1 otherwise

	-v	Select non-matching lines

	-s	Suppress open and read errors

	-r	Recurse

	-i	Ignore case

	-w	Match whole words only

	-F	PATTERN is a literal (not regexp)

	-E	PATTERN is an extended regexp

	-z	Input is NUL terminated

	-m N	Match up to N times per file

	-A N	Print N lines of trailing context

	-B N	Print N lines of leading context

	-C N	Same as '-A N -B N'

	-e PTRN	Pattern to match

	-f FILE	Read pattern from file



#

...the first line is using the full grep, the same operation with Busybox grep fails, even though it should work.

So why did I bother to upgrade Busybox? The old 1.4.2 worked, in conjunction with full utilities where required, and it added up smaller size.
I am going to roll Busybox back and put in the full 'grep', will upload Woof soon.

New "official" Puppy: Slacko

Of course we have many puplets, builds of Puppy that get announced on the forum. However, currently only two are "official", Wary and Lucid -- these are the ones we list at distrowatch.com and that get promoted as "Puppy Linux". Normally, newcomers would try one of these first.

Lucid is currently at version 5.2.5, and as announced recently there won't be another release, at least not in capacity as an "official" Puppy:
http://bkhome.org/blog/?viewDetailed=02394

However, there is another flavour of Puppy that has been coming along nicely, "Slack Puppy", or just "Slacko" (or even Spup), developed and coordinated by 01micko. This is currently at alpha, coming up to first beta soon:
http://murga-linux.com/puppy/viewtopic.php?t=69798

Playdayz and Lobster contacted me independently, recommending Slacko. I then asked 01micko (Mick) if he wanted to go for it, and he is keen.

Mick is a disciplined coordinator I think, from what I have seen so far, and has made sure that his Slacko releases have synchronised with the latest Woof. I don't recall him getting embroiled in any fights -- I mean discussions of a heated nature -- in the forum. He has also been around for quite a while and contributed regularly, some really good stuff -- script patches and apps (see recent blog posts, and pupRadio for example).

So, what we have to look forward to in the not-to-distant future is Wary 5.2 and Slacko 5.x "official" Puppy Linux releases. I started Wary at 5.0 as it is fifth-generation, so I suggest Slacko will also start at 5.something.

For Ubuntu fans, there probably will be another official Upup sometime. Apparently though, Ubuntu is becoming increasingly weird, and we might choose to back Debian instead. I intend to bring out another "reference" build of Dpup or Upup sometime soon. Probably another build of Drake Pup is also on the horizon.

Woof uploaded, August 15, 2011

This upload is commit '20110815093119'.

The previous upload was on August 13:
http://bkhome.org/blog/?viewDetailed=02400

There have been some small but useful improvements since August 13, hence this upload.

Fatdog64 520 Release Candidate

Kirk has announce that Fatdog64 has reach RC status:
http://murga-linux.com/puppy/viewtopic.php?t=70771

Fatdog64 is a pup built especially for 64-bit x86 CPUs.

Testers welcome!

'pngoverlay.sh'

The previous blog post announces the latest netpbm, that can overlay PNG images, with transparency:
http://bkhome.org/blog/?viewDetailed=02404

I have written a script, 'pngoverlay.sh', that performs the same as 'pngoverlay':

#!/bin/sh

# Barry Kauler 2011 GPL3 (/usr/share/doc/legal)

#pngoverlay.sh is an alternative to pngoverlay written by vovchik (in BaCon)

# (vovchik's pngoverlay requires X to be running, which may be a disadvantage)

#requires netpbm svn rev 1543 or later, with pamcomp -mixtransparency

#requires three params, 1st and 2nd must exist:

# bottom-image top-image output-image

#overlays the two images, with common areas of transparency in output image.



[ ! $3 ] && exit 1

[ ! -e "$1" ] && exit 1

[ ! -e "$2" ] && exit 1

[ "`echo -n "$1" | grep 'png$'`" = "" ] && exit 1

[ "`echo -n "$2" | grep 'png$'`" = "" ] && exit 1



pngtopam -alphapam "${1}" > /tmp/pngoverlay_${$}_1.pam

pngtopam -alphapam "${2}" > /tmp/pngoverlay_${$}_2.pam

#1st image on top, 2nd on bottom, 3rd is output...

pamcomp -mixtransparency /tmp/pngoverlay_${$}_2.pam /tmp/pngoverlay_${$}_1.pam > /tmp/pngoverlay_${$}_out.png 2> /dev/null

pamrgbatopng /tmp/pngoverlay_${$}_out.png > "${3}"

rm -f /tmp/pngoverlay_${$}_1.pam

rm -f /tmp/pngoverlay_${$}_2.pam

rm -f /tmp/pngoverlay_${$}_out.pam

I have put it into Woof, in /usr/sbin, however have still left in Vovchik's pngoverlay and the latter is still being used. The problem is that it can't be guaranteed that all Woof-built puppies will use this latest netpbm -- some may use an older version supplied by the compatible-distro.

Anyway, I have put the pngoverlay.sh script into Woof with a view to moving over to it one day.

netpbm 11.55 svn revision 1543

Netpbm is a collection of CLI (commandline, can be used in scripts) utilities for manipulating images. A subset of the full package has been in Puppy almost from the beginning.

The maintainer of Netpbm is Bryan Henderson, and I have been communicating with him with a view to enhance the utilities to perform a png overlay function, as we currently do with Vovchik's 'pngoverlay'. There were some iterations, but the latest in SVN, revision 1543, works beautifully.

I'll post about png overlay soon, for now, here is the PET (324KB):
http://distro.ibiblio.org/quirky/pet_packages-common/netpbm-11.55-svn1543-w5.pet

...compiled in Wary, but should work in all puppies.

Netpbm home:
http://netpbm.sourceforge.net/
User Manual:
http://netpbm.sourceforge.net/doc/
Documentation on the utilities:
http://netpbm.sourceforge.net/doc/directory.html

Retiring 'gtkdialog2'

This has been on my to-do list for a very long time, to convert some scripts that use gtkdialog2 to gtkdialog3. Thanks to 01micko, it has now happened:
http://murga-linux.com/puppy/viewtopic.php?t=70728

The affected files are dotpup, input-wizard, pmount, pupscan, xorgwizard, puppyinstaller, all in /usr/sbin.

01micko converted older versions of dotpup and pmount than in latest Woof, but I merged his changes ok. I fixed a bug in xorgwizard.

01micko had not converted puppyinstaller, so I did that.

Note that pupscan has a bug, the "USB devices" button does not work, at least not with the 2.6.32.x kernel. I'll look into that today. After fixing it, I will probably upload Woof soon after.

Unique hostname on network

Puppy is assigned the hostname "puppypc", see /etc/hostname and /etc/hosts.

GCMartin and others have been tackling the problem of more than one Puppy computer on the network, where they would require different hostnames. In fact, GCMartin has done a superb job in explaining the problem:
http://www.murga-linux.com/puppy/viewtopic.php?t=70427

01micko has devised a very neat solution, see page 2 of the above thread. I have implemented this in Woof, script /etc/rc.d/rc.sysinit. I have implemented it exactly as 01micko has done, so anyone using Woof can do that now if they want, but it will of course be in the next upload of Woof.

You guys are to be congratulated, GCMartin for so clearly explaining the problem, and persisting in making it known to me and others. 01micko for the nice automatic solution.

Worried about 2.6.39.x kernel

Drake Puppy 0.3 has the 2.6.39.3 kernel. It seems like a nice kernel, except for one thing...

A few times in Woof when I built a live-CD of Drake Puppy, I found that a file, and in one case a complete directory, were missing from the main SFS file, 'puppy_drake_0.3.sfs'. This is running on my new laptop, with Intel i3-core CPU.

I have no explanation for this, they just disappeared. They disappeared from sandbox3/rootfs-complete, the complete filesystem used to create the SFS, so it is not a problem with Squashfs. They should have been copied into rootfs-complete, somehow they never got there.

That is, the code-loop in '3builddistro' copies packages into sandbox3/rootfs-complete, and the files are in the packages, but don't appear in the destination. It is just a tight loop with the 'cp' command.

After this happened three times, I lost all confidence in the kernel, as I could not see any other reason for the disappearance. Since then I have been using Wary, with 2.6.32.x kernel, for all development work, and everything has been fine. Note, I am working in a partition with ext3 filesystem.

I am wondering if there is something in the configuration of the 2.6.39.x kernel -- cgroups maybe? -- that is causing this. I might try a recompile of the kernel with more conservative configuration.

Woof uploaded, August 13, 2011

The commit number is '20110813145444'.

This version of Woof was used to build Wary 5.1.3.

The previous upload of Woof was August 9:
http://bkhome.org/blog/?viewDetailed=02391

...please read link for further details about Woof.

For a summary of the changes to Woof, see the Wary 5.1.3 release notes page:
http://distro.ibiblio.org/quirky/wary-5.1.3/release-Wary-5.1.3.htm

Wary 5.1.3 released

Here is a brief announcement:

Wary is our 'flavour' of Puppy that targets older hardware. Version 5.1.3 is mostly bugfixes and minor upgrades to 5.1.2. There have been some significant improvements at the Woof level, such as a close-box on desktop drive icons and faster bootup. New applications: Default Applications Chooser, Urxvtset. New utilities: picscale, pngoverlay, yad. I decided to stay with SeaMonkey 1.1.18, as 2.2 has a badly broken Composer module. Also, 1.1.18 is smaller and better suited to older hardware -- FlashBlock and AddBlock plugins are included to suit those on dialup and download-limited Internet connection. A couple of projects that have been upgraded in Wary and are worthy of special mention: BaCon and gtkdialog. BaCon is a superb BASIC compiler, and gtkdialog provides sophisticated GUIs for shell scripts. Both of these are extremely active projects, with very keen developers and users.

Further announcement and release notes:
http://distro.ibiblio.org/quirky/wary-5.1.3/release-Wary-5.1.3.htm

Download ISO live-CD (120.6MB):
http://distro.ibiblio.org/quirky/wary-5.1.3/wary-5.1.3.iso

'devx' SFS for complete compiler environment (109.1MB):
http://distro.ibiblio.org/quirky/wary-5.1.3/devx_wary_5.1.3.sfs

For those on dialup and/or restricted download capability, we have xdelta files, smaller files for upgrading from an earlier Wary download:
http://distro.ibiblio.org/quirky/wary-5.1.3/delta_files
...ideally, you need a running Puppy, as the Utility menu has a simple drag-and-drop GUI for updating with an xdelta file.

'install' choices fixed

I you click the 'install' icon on the desktop, a window with further choices pops up.

One button opens a web page with information on package management in Puppy, however, that was not the best page, fixed. Script /usr/sbin/dotpup.

Another button opens a window to choose SFS files to download. This is broken if the SFS file has a '_' in it's name. Fixed. Script /usr/sbin/sfsget.

Another button opens the Puppy Package Manager (PPM). One of the download URLs is 'cc.gatech.edu', however, this no longer mirrors Puppy PETs from ibiblio.org. I have removed it from the list. Files /root/.packages/DISTRO_COMPAT_REPOS and DISTRO_PET_REPOS.

SeaMonkey 2.2 Composer broken

It's unusable. Open a web page, click on an image, the image duplicates in the web page. Try a right-click, "Image properties" is greyed out.

This is really bad. I had hoped to release Wary 5.1.3 with SM 2.2, however, I am going to stay with 1.1.18 as in Wary 5.1.2.

Hopefully I will be able to upgrade SM for Wary 5.2.

'common' repo download fixed

It has been reported that when an attempt is made to install a PET from the 'common' repo in the PPM (Puppy Package Manager), that it tries to download from the 'puppylinux' directory on ibiblio.org, instead of the correct 'quirky' directory.

This problem afflicts all recent Woofs. I have fixed it, will be uploaded with next Woof, but if you want to fix it now, go into /usr/local/petget/downloadpkgs.sh and edit about line 70 so that it looks like this:

 #100117 need a little hack for quirky... not the best, really should be a generic fix...

 #if [ "`echo -n "$ONELIST" | grep 'quirky'`" != "" ];then

 #if [ "$REPO_DEFAULT_SUBSUBDIR" == "quirky" ];then #100903

 

 #110812 hack for pets that are in quirky site at ibiblio...

 OFFICIAL_REPO='puppylinux'

 [ "$REPO_DEFAULT_SUBSUBDIR" == "quirky" ] && OFFICIAL_REPO='quirky' #100903

 [ "$REPO_DEFAULT_SUBSUBDIR" == "wary5" ] && OFFICIAL_REPO='quirky' #101014 wary5 pets also in quirky repo.

 [ "$REPO_DEFAULT_SUBSUBDIR" == "common" ] && OFFICIAL_REPO='quirky' #110812

 [ "$REPO_DEFAULT_SUBSUBDIR" == "drake" ] && OFFICIAL_REPO='quirky' #110812

 

 #if [ "$REPO_DEFAULT_SUBSUBDIR" == "quirky" -o "$REPO_DEFAULT_SUBSUBDIR" == "wary5" ];then #100903 101014 wary5 pets also in quirky repo.

 if [ "$OFFICIAL_REPO" = "quirky" ];then #110812

It would seem that the 'drake' repo had the same problem, so I fixed that too.

2.6.32-44 Linux kernel

I have compiled the latest in the 2.6.32 series.

The kernel PET (for Woof developers only) (29.1MB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_kernel-2.6.32-44-i486-w5.pet

Sources:
http://bkhome.org/sources/kernel-2.6.32-44/

Wary 5.1.2 has kernel 2.6.32-40 and this is built in such a way that it identifies itself as version 2.6.32 and the modules are in /lib/modules/2.6.32. This latest, 2.6.32-44, has also been configured like this so identifies itself as 2.6.32.

The above potentially causes a problem when upgrading a pupsave file. Upgrading from 2.6.32-40 to 2.6.32-44 is not intended to be any problem, however I have configured the latter with inbuilt HID and USB drivers, which may make other modules incompatible. In other words, if you had compiled or installed a module compiled for the 2.6.32-40 kernel, it might not work with the 2.6.32-44 kernel, or it might. In future, I will keep the same configuration, so older modules will continue to work when the kernel is upgraded (to 2.6.32-45 etc).

I have edited the 'init' script in the initrd, and /etc/rc.d/rc.update to take care of some possible old incompatible modules when there is an upgrade of a pupsave, as described in previous paragraph.

Upup, Spup, Dpup, Drake, Wary...

I was away for a couple of days, back this morning and logged in to the forum, very surprised to see a pm from playdayz cancelling Lucid Puppy. See forum:
http://www.murga-linux.com/puppy/viewtopic.php?t=69563

I am sure that playdayz has good reasons. As I understand from a quick look at the Lucid forum thread, Lucid has grown to be a bit more than just a bug-fix release of 5.2.5, so has got a bit out of hand?

Anyway, the current situation from my point of view, is that Wary 5.1.3RC is coming very soon, a couple of days, final soon after.

Then, I have great plans for Wary 5.2.

For Ubuntu-Puppy lovers, I am thinking of doing a build in latest Woof, using packages of the latest release of Ubuntu. This could be used as a base for anyone who wants to create the next Upup.

Drake Puppy (Mpup) is at version 0.3, seems quite sound. Still only a beta, and I have been tossing up whether I should press ahead myself to get it to release status.

01micko is working on Dpup I think. I haven't checked on the latest status.

Also, FatDog64 created by kirk is progressing, another release soon.

Lucid Puppy 5.2.7 RC2

Playdayz has announced the second Release Candidate for Lucid Puppy 5.2.7:

http://www.murga-linux.com/puppy/viewtopic.php?t=69563

Try it out!

Offline for couple of days

Just a note to let everyone know. I may not be responsive to posts for the next couple of days, as I have other non-Puppy stuff to do.

Woof uploaded, August 9, 2011

This is the Woof used to build Wary 5.1.2.7, commit number '20110808161144'.

The previous upload of Woof was July 22, 2011:
http://bkhome.org/blog/?viewDetailed=02361

There was a blog post recently on how to download a recent snapshot tarball of Woof and use that as the reference to download Woof changes since then, without the full history:
http://bkhome.org/blog/?viewDetailed=02306

To download and use Woof, please read the Bones intro page:
http://bkhome.org/bones/
And the Woof intro page:
http://bkhome.org/woof/

This release of Woof has many new features and fixes. Read this blog -- most posts are related to Woof, but for specific posts click the "Woof" Category on the left of this page.

Wary Puppy 5.1.2.7 (5.1.3beta)

Here it is, the latest in the Wary series. This is a beta for the upcoming 5.1.3, or maybe even an RC.

Download from here:
http://distro.ibiblio.org/quirky/test/wary-5.1.2.7/

The kernel is version 2.6.32.43, and if you need the source:
http://bkhome.org/sources/kernel-2.6.32.43/
Us#er#na#me: pu#pp#y Pa#ss#wo#rd: li#nu#x

Note, my Telstra 3G connection is so poor recently. I uploaded the .iso file yesterday morning, then the Internet became unusable from about 3pm until 10pm, after which I completed the upload. Right now it is 3.45am local time.

To find out everything in this release, scan my blog. Note, it has SeaMonkey 2.2 -- I have some reservations about this on old hardware.

If you want to read about the last official release, 5.1.2, here is the blog announcement:
http://bkhome.org/blog/?viewDetailed=02304

To read about the concept behind Wary:
http://bkhome.org/wary/

Telstra 3G unusable

Sometimes my Telstra 3G wireless connection gets slow in the evenings. Last night from about 6pm to 9pm is was unusable -- pages would not finish loading, connections timed out.

Then it came good and I was able to upload the 2.6.32.43 kernel stuff.

But all day today it has been unusable. Maybe they are having equipment problems. Even now, at 9.45pm, it is only just working.

Hmmm, myprepaid.telstra.com won't load. Nothing will load. I will wait awhile before posting this.

2.6.32.43 kernel

i have compiled the latest in the 2.6.32 series. There are some significant changes:

Fully automated
There are now three scripts, '1kernel.sh', '2modules.sh' and '3package.sh'. They are executed one after the other, and together they do everything. The first patches and compiles the kernel, the second compiles all the 3rd-party drivers, and the last creates the source SFS and PET.

2.6.32-43 versus 2.6.32.43
The first script offers the choice of retaining the version number as just "2.6.32" or "2.6.32.43". As there are major configuration changes, I chose the latter.

Inbuilt HID and USB
The kernel is configured with core USB drivers and HID (Human Input Device) drivers builtin to the kernel. I already reported on this in my blog. I think that it gives faster bootup.

The PET, compiled in Wary 5.1.2 (for use in Woof only) (29.1MB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_kernel-2.6.32.43-i486-w5.pet

The sources, including build scripts and the SFS:
http://bkhome.org/sources/kernel-2.6.32.43/

zenity symlink to yad

As I am intending that all Woof-built puppies will have yad, some applications that require zenity will also work. This requires 'zenity' to be a symlink to 'yad'.

Rather than rely upon yad PET creators to include that symlink, I have put a check into '3builddistro' in Woof:

#110806 if yad installed, no zenity, make sure zenity is symlink to yad...

YAD="`find ./rootfs-complete/bin ./rootfs-complete/sbin ./rootfs-complete/usr/bin ./rootfs-complete/usr/sbin ./rootfs-complete/usr/local/bin -type f -name yad`"

if [ "$YAD" ];then

 ZENITY="`find ./rootfs-complete/bin ./rootfs-complete/sbin ./rootfs-complete/usr/bin ./rootfs-complete/usr/sbin ./rootfs-complete/usr/local/bin -name zenity`"

 if [ ! "$ZENITY"  ];then

  YADDIR="`dirname $YAD`"

  ln -s yad ${YADDIR}/zenity

 fi

fi

What reminded me to do this is System_info created by 8-bit. It calls 'zenity' so requires the symlink. System_info looks very interesting:
http://www.murga-linux.com/puppy/viewtopic.php?t=70513

Wary plans

For those wondering about the next Wary, I plan to release a beta of 5.1.3 in a couple of days.

Wary 5.2 will have many more package upgrades, and that is a bit further away.

There have been many improvements to Woof since the last upload, so I intend to upload Woof in a couple of days also.

Pnethood 0.6.8

It is great to see Pnethood continue to be developed. Thanks to clarf for doing this. It was HairyWill who originally wrote Pnethood, but he moved on, and clarf took it on-board.

Early in 2010, Patriot developed an alternative, lameSMBxplorer, which I think was supposed to become the successor to Pnethood:
http://murga-linux.com/puppy/viewtopic.php?t=50901

However, the last post to that thread from Patriot was in August 2010.

Wary 5.1.2 has Pnethood, and this latest version is targeted for the next Wary. I will upload the PET to the 'common' repo soon.

Read about Pnethood here:
http://www.murga-linux.com/puppy/viewtopic.php?t=23464

Pmusic 1.8.3

I have upgraded zigbert's great audio player, see forum:
http://www.murga-linux.com/puppy/viewtopic.php?t=31206

The only thing that I changed was one line in 'pinstall.sh' needs relative paths so as to work when the PET is installed in Woof.

Uploaded PET (73KB):
http://distro.ibiblio.org/quirky/pet_packages-common/pmusic-1.8.3.pet

freememapplet_tray 2.3.1

I am running a full hard drive installation of wary 5.1.2, and I noticed that /root/Startup/freememapplet_tray crashes at startup. I had noticed this before, but this time decided to find out why.

01micko added code to test for PUPMODE=2 and bumped the version from 2.2.x to 2.3. However, 01micko's fix is broken.

However, it would have worked when 01micko worked on it. The problem is that Busybox 'df' and the full df return this:

# busybox df

Filesystem           1K-blocks      Used Available Use% Mounted on

/dev/root             10509560   2032264   7943436  20% /

shmfs                   594672         0    594672   0% /dev/shm

/dev/sda10           514972152  90904020 397909028  19% /mnt/sda10

# 

# df-FULL

Filesystem           1K-blocks      Used Available Use% Mounted on

rootfs                10509560   2032264   7943436  21% /

/dev/root             10509560   2032264   7943436  21% /

shmfs                   594672         0    594672   0% /dev/shm

/dev/sda10           514972152  90904020 397909028  19% /mnt/sda10

However, quite some time ago I created a script /bin/df, that is a wrapper around df and calls 'rdev' to substitute the actual partition name instead of /dev/root:

# df

Filesystem           1K-blocks      Used Available Use% Mounted on

/dev/sda9             10509560   2032264   7943436  21% /

shmfs                   594672         0    594672   0% /dev/shm

/dev/sda10           514972152  90904020 397909028  19% /mnt/sda10

All Woof-built puppies have this, quite some way back, but I don't recall exactly when I introduced the script.

Anyway, I have fixed freememapplet_tray and it is now version 2.3.1 (4KB):
http://distro.ibiblio.org/quirky/pet_packages-common/freememapplet_tray-2.3.1.pet

pupRadio/pupTelly 0.10

01micko has upgraded his streaming radio and tv player:
http://www.murga-linux.com/puppy/viewtopic.php?t=50016

This is a great app. Unfortunately it is broken in Wary 5.1.2, which was my fault, I left something out of the package.

I have uploaded the PET to the 'common' repo:
http://distro.ibiblio.org/quirky/pet_packages-common/pupradio-0.10.pet

Double-login "bug"

There is a curious little "bug" that has been in Puppy right from the beginning, that causes a double-login.

When you power-up your pup and get a desktop, you are logged in once. If you then exit to the prompt (via "Exit to prompt" in the "Shutdown" menu, or hit CTRL-ALT-BACKSPACE), you get dropped down to the console, but you also get logged in a second time.

This has never bothered us, well I don't think it has. But, I am thinking how to fix it...

The problem comes about by this line in /etc/profile:

exec xwin

This line transfers control to /usr/bin/xwin, which in turn launches the desktop. When there is an exit from X, there is then a return to xwin, which then exits and there is a return to /etc/profile.

Now this is the problem: /etc/profile is really being used in an unnatural way. It never completes, instead there is a transfer to xwin. When there is finally an exit from xwin, /etc/profile completes, but this somehow causes a second login to occur ( the auto-root-login runs, see /etc/inittab).

I recall that Patriot considered this and other bootup/shutdown issues in some depth awhile back.

Well, there is one possible "fix". Instead of "exec xwin" in /etc/profile, have "xwin &". This allows /etc/profile to complete.

However, when there is an exit to prompt from X, it looks strange as the cursor is there but no prompt before it. A fix for this is to have "echo -n '# '" just before exit from xwin.

I might implement this fix in Woof, still thinking about it.

Default Applications Chooser 0.8

Sc0ttman has created this great application:
http://murga-linux.com/puppy/viewtopic.php?t=56282

Yeah, this deserves to be in all puppies. I have repackaged it slightly (created pet.specs, modified default-chooser.desktop) and uploaded (5KB):
http://distro.ibiblio.org/quirky/pet_packages-common/defaults-chooser-0.8.pet

I have added this to the package lists for all builds in Woof.

Urxvtset 0.5

Forum member stu90 has created a great little GUI utility to configure the appearance of urxvt (rxvt-unicode), the terminal emulator (window) used in most of our puppies these days.

Here is the forum thread:
http://www.murga-linux.com/puppy/viewtopic.php?t=68624

I have repackaged it slightly and uploaded the PET to ibiblio.org, the 'common' repo.

I changed the 'urxvtset' executable location to /usr/sbin, modified /root/.Xdefaults, some changes in /usr/share/applications/urxvtset.desktop and pet.specs.

urxvtset-0.5|urxvtset|0.5||Desktop|44K||urxvtset-0.5.pet|+yad,+urxvt|UrxvtSet setup terminal emulator||||

A great little app, stu90 is doing a lot of work developing applets using yad.

Here it is (3KB):
http://distro.ibiblio.org/quirky/pet_packages-common/urxvtset-0.5.pet

I have entered this into the PKGS_SPECS_TABLE variable of all the builds in Woof.

gtkdialog 0.7.21 rev174

Thunor is working intensely, improving gtkdialog. Forum thread:
http://murga-linux.com/puppy/viewtopic.php?t=69188

I have downloaded the latest from SVN, revision 174, and created a PET (67KB):
http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog-0.7.21-thunor175-w5.pet

Note that the executable is /usr/sbin/gtkdialog, with gtkdialog3 a symlink to it.

Thunor has also been improving the examples, and they are in the _DEV PET (42KB):
http://distro.ibiblio.org/quirky/pet_packages-common/gtkdialog_DEV-0.7.21-thunor175-w5.pet

Nathan's Wallpaper Setter 0.6

Nathan wrote this a long time ago (Nathan created Grafpup, a very nice Puppy derivative). The Wallpaper app has been hacked on by various people, including me, and the last version was 0.5.6 -- which is buggy, so I dropped it in favour of Pwallpaper (now at 1.1.2-1) which is an app discovered by ttuuxxx.

01micko and zigbert have been hacking on Nathan's Wallpaper Setter and have released version 0.6, which I presume is in Lucid 5.2.6RC.

However, it has a bug, and I have been trying to refresh my memory about this... it has been awhile... the app has lost the ability to crop images for widescreens.

Woof has rootfs-skeleton/usr/sbin/set_bg, which came out of an earlier Nathan's Wallpaper Setter. It has a modification to crop the image for wide screens.

The original Pwallpaper C source was modified by ttuuxxx to call set_bg (then by me to specify path /usr/sbin/set_bg) -- see pwallpaper.c in pwallpaper-1.1.1.tar.gz in my sources repo.

The latest Nathan's Wallpaper is calling it's own internal /usr/local/apps/Wallpaper/set_bg, which doesn't have the image cropping.

Note also, the original image gets moved to /usr/share/backgrounds_original.

The fix seems to be simple, just get Nathan's Wallpaper Setter to call /usr/sbin/set_bg -- I did a quick test, seems to work ok.

I'll send a pm to 01micko to get the app upgraded with the fix.

Forum link:
http://www.murga-linux.com/puppy/viewtopic.php?t=69658

Pcdripper audio CD ripper 3.8

This is plinej's baby, now upgraded to version 3.8:
http://www.murga-linux.com/puppy/viewtopic.php?t=24157

I have upgraded this PET in the 'common' repo, for all Woof builds to use. Will upload soon.

'picscale' image scaler 0.1b

Vovchik is our master at BaCon coding -- it was vovchik who introduced us to BaCon. Vovchik has created some great little utilities written in BaCon, 'picscale' is one of them.

The original version 0.1a was posted here:
http://www.murga-linux.com/puppy/viewtopic.php?t=49129&start=15

A bugfix for spaces in filenames, version 0.1b:
http://murga-linux.com/puppy/viewtopic.php?t=69512

I have compiled picscale in Wary 5.1.2 and created a PET. This will be in the 'common' repo (will upload soon), as 'picscale-0.1b-w5.pet'. I will also upload the source soon to my source repo http://bkhome.org/sources/alphabetical/p/

picscale is used in Vovchik's Picona button-creator application, and in stu90's SimpleWALL wallpaper chooser. So, I reckon that picscale should be in all puppies and I have added it to the package-lists in all the builds in Woof (dpup, spup, wary, etc.).

SimpleGTK theme maker 0.7b

I have added SimpleGTK 0.7b to the 'common' repo. This is a simple little GUI utility created by stu09, using 'yad':

http://www.murga-linux.com/puppy/viewtopic.php?t=69312

I tweaked the 'pet.specs' file slightly:

simplegtk-0.7b|simplegtk|0.7b||Desktop|44K||simplegtk-0.7b.pet|+yad|SimpleGTK GTK theme maker||||

It isn't a big deal, but the convention in Puppy, T2 and many other distros is to keep package names lower-case.

The dependencies field must have package name preceded with a '+' character, and a ',' delimiter for multiple dependencies -- note, the 'petspec' utility (the GUI for creating pet.specs) has popup help for this and other fields in the latest Woof.

Note, 'yad' is already in the 'common' repo.

Pfind 4.22

I have just started to upgrade some of the packages in the 'common' repo. First up is zigbert's Pfind:

http://murga-linux.com/puppy/viewtopic.php?t=15136

I will upload it to ibiblio.org after I have done a few more packages.

Wary: upgrade glibc/gcc?

I am working on a recompile in T2 of the packages for the next major version jump of Wary, to "5.2".

Originally, I was thinking that glibc and gcc should stay at the same versions for the life of the Wary project, to help to ensure that packages stay working in later releases of Wary.

But, I'm not quite sure.

Glibc is currently version 2.10.1, latest in T2 is 2.13.

Gcc is currently 4.3.4, latest in T2 is 4.6.1.

The big question is, would bumping either of these packages to the latest version, break anything, in particular break any of the current collection of packages in the Wary repository?

I will send a pm to tempestuous, as I think that he know a lot about this topic.

Bumping glibc, and the libstdc++ in gcc, can break packages, especially libsdtc++. Yes, perhaps I should leave them at the current versions.

Luci 5.2.6 RC

Playdayz has announced that Luci (Lucid Puppy) has reached 5.2.6 RC (Release Candidate). Findout more here:

http://murga-linux.com/puppy/viewtopic.php?t=69563

Short-term plans

I'm trying to do too much! It is resulting in me getting a little bit tense, I need to back off. I might do another round of bug-fixing with Drake Puppy, but slowly, no rush -- after that, perhaps I will put it up as a project for someone, an old "Mandrake fan", who might like to do a bit more work on it and get it to release status.

There is so much other stuff waiting for me to do. Some of it is important infrastructure stuff. These are things that I plan to work on over the next week or two:

1. Pkg selection in Woof
The code that chooses what packages go into a build, based on a list (PKGS_SPECS_TABLE), has issues. I need to re-write it.

2. T2 to Wary
Wary is based upon packages compiled in T2. Unfortunately, some of that work was on the USB drive that was stolen (see my blog report recently, my house was burgled), but I am managing to get it setup again. I want to recompile Xorg 7.3 without libxcb, as the API is incompatible with the newer xcb (the functions in the shared library are incompatible). Removing dependency on xcb will improve compatibility when install an Xorg 7.5/6 upgrade PET that does have xcb.
There are also some other changes I want to do in the recompile.

3. Wary kernel
Recompile 2.6.32.x kernel with inbuilt USB and HID drivers.

4. Various Woof improvements/fixes:
4.1 better fix for pngoverlay workaround.
4.2 frugal: sfs not copying into ram.

5. Upgrade PETs
There are many packages that need upgrading, that our keen developers are always improving. I need to do a round of upgrading of gtkdialog, Pburn, etc. Most of these will go into the 'common' repo.

There have been some other good suggestions posted to this blog, plus some Forum pm's and emails, but they have mostly been neglected by me due to lack of time. I am reluctant to call for suggestions to be re-posted, I might post a message soon requesting reminders of things that need to be looked at in Woof/Puppy.

Drake Puppy 0.3

I am boldly announcing this as the first beta release of Drake Puppy.

Version 0.1, that I announced as an alpha, was released on July 3:
http://bkhome.org/blog/?viewDetailed=02330

Version 0.1 has a 2.6.34.8 kernel, 0.3 has a 2.6.39.3 kernel. The latter is also PAE-enabled.

This pup also has SeaMonkey 2.2.

Get it from here:
http://distro.ibiblio.org/quirky/drake-0.3

The ISO has a XZ-compressed SFS inside it, whereas the 'devz' SFS is only gzip-compressed, hence the larger size of the devx file.

Mandrake lovers, try this pup!

Note, this is not yet a "public release". Probably still has lots of bugs. It seems to be working ok though, I'm using it now.

Known issues:

1. Xorg Wizard needs work.
2. ?

Note 2: I'm not sure about the future of Drake Puppy. I mean, as far as myself managing it's development. I'm spreading myself too thin. Will post more about this soon.

'petspec' improved

/usr/bin/petspec is a little GUI application that can be run directly but is also called from 'dir2pet' -- the latter is our utility for converting a directory into a PET package.

I have put question-mark buttons alongside some of the fields in the GUI, to explain each field.

In particular, the dependencies box has help, which explains how to enter the dependencies of the package, in particular how to include versioning as introduced here:
http://bkhome.org/blog/?viewDetailed=02365
http://bkhome.org/blog/?viewDetailed=02363

Plans for Drake Puppy

I can crash SeaMonkey 1.1.18 in Drake Puppy consistently, in fact Xorg crashes and I get dropped back to the pre-X console. On my blog, type "fido" in the search box, then either left-click or right-click on the link-result of the search, then X crashes.

I have compiled SM 2.2 in Drake Puppy, it does not crash.

Planning ahead a little bit, I intend to bring out Drake Puppy 0.3 in a couple of days.

This will showcase the latest features of Woof, but also should have many bug fixes for Drake Puppy itself. I will get onto reading the Drake pup feedback thread in the forum this afternoon.

I am using Drake Puppy everyday, mostly because the video resolution is correct on my new laptop.

Gtkdialog: help?

I have been trawling through the forum, trying to find a solution for this. I am using thunor's revision-83 gtkdialog package.

I have started work on /usr/bin/petspec, and I want the main window to be able to open child windows. Yes, this works, but the only way to close the child window is to click the close-box on the window.

The original 0.7.20 official examples give this example of a child window:

export DIALOG='

  

    

      This is an other dialog window.

    

    

  

'

The main window launches the child by this:

I recall Patriot discussing this sometime ago, and it did apparently work. I have tried various things, can't get a button in the child window to close itself.

There was some discussion back in 2009:
http://www.murga-linux.com/puppy/viewtopic.php?t=50253

Is there a gtkdialog guru reading this who can help out?

Basic versioning support in Woof

I announced versioning support for package dependencies in the previous blog post:
http://bkhome.org/blog/?viewDetailed=02363

I have now implemented basic support when building a Puppy distro in Woof.

I have modified some kernel-version-sensitive packages accordingly:

dhcpcd
aufs2-util
squashfs-tools4

For example, in the 'common' repo, there are now two 'aufs2-util' packages. Here are their pet.specs entries:

aufs2-util-20090904-1-p4|aufs2-util|20090904-1-p4||BuildingBlock|996K||aufs2-util-20090904-1-p4.pet||utilities for aufs2 layered filesystem|puppy|4||

aufs2-util-20110110aufs21-1-w5|aufs2-util|20110110aufs21-1-w5||BuildingBlock|96K||aufs2-util-20110110aufs21-1-w5.pet|+linux_kernel&ge2.6.32|aufs 2.1 layered filesystem utilities, for kernels 2.6.32 and later|puppy|wary5||

Woof chooses the appropriate package to suit the chosen kernel.

I wrote a new script in Woof, support/1buildversiondeps, which does the version checking. It is very klutzy and inefficient how it fits in with the other scripts, but it does work.
I need to rewrite some big chunks of the Woof scripts sometime.

Luci Puppy 5.2.6 beta

Playdayz has announced that Luci (Lucid) Puppy has reached 5.2.6-beta (internal numbered as 266).

If you would like to do bug testing, or just play with it, go here:
http://www.murga-linux.com/puppy/viewtopic.php?t=69563

It is expected that after this there will be 5.2.6-RC (Release Candidate).

Progress on versioning support for dependencies

This has been a long time coming. I had it in my to-do list back in the early days of developing Woof, but it never happened.

When someone compiles a package and creates a PET, they might want to specify that the dependencies require certain versions.

Also, there are some packages that are kernel-version-sensitive, for example 'dhcpcd' seems to be. I recall a tentative list of such packages being made up some years ago, yes, in this page (search for the word "sensitive"):
http://puppylinux.com/news/news217.1-220a.htm

There is a method of choosing a package based on kernel version implemented in Woof, but it is a kludge.

I am now implementing versioning support for dependencies in a systematic manner. In the case of the latest 'dhcpcd' package that I have compiled in Drake Puppy, running a 2.6.39-3 kernel, I have created 'pet.specs' file with this content:

dhcpcd-5.2.12-i486-m1|dhcpcd|5.2.12-i486-m1||BuildingBlock|176K||dhcpcd-5.2.12-i486-m1.pet|+linux_kernel&ge2.6.39|network connection client daemon|mageia|1||

Here is another hypothetical package:

abiword-2.8.6-w5|abiword|2.8.6-w5||Document|7012K||abiword-2.8.6-w5.pet|+cairo,+enchant,+fribidi,+geany,+goffice&eq0.8.9,+gtk+,+libgsf,+libxml,+wv|The GNOME word processor|puppy|wary5||

It is very simple. You only need to specify a version for a dependency when you know or suspect the repo may have alternate packages, for example more than one 'goffice' and you want to make sure the correct version is installed.

The allowed operators are:

&le &ge < > &eq

I did think of implementing &or and &and to make the versioning more sophisticated, but I don't think that is needed.

I have already modified /usr/local/petget/dependencies.sh and /usr/local/petget/findmissingpkgs.sh to support versioning in the PPM (Puppy Package Manager).

I am now working on this in Woof, building a distro, which is a lot more work. The pet.specs GUI builder will also be upgraded.

More 'common' PETs

I was examining the Drake Puppy build and noticed where Woof had fallen back to getting PETs out of the 'puppy 4' repo. I have placed some of these into the 'common' repo, as they are packages that all Puppy builds are likely to need.

'common' repo:
http://distro.ibiblio.org/quirky/pet_packages-common

Latest database file:
http://distro.ibiblio.org/quirky/Packages-puppy-common-official

Woof uploaded, July 22, 2011

I want to add versioning support to the dependencies field in the PPM package database files, but as that might break something, I decided to upload Woof now.

This is commit '20110722103947'.

The previous upload was on July 12:
http://bkhome.org/blog/?viewDetailed=02344

There was a blog post recently on how to download a recent snapshot tarball of Woof and use that as the reference to download Woof changes since then, without the full history:
http://bkhome.org/blog/?viewDetailed=02306

Enhanced desktop drive icons

This is something that I have wanted to do for a long time. I always diligently unmount partitions by right-click on a desktop drive icon then choose to unmount from the popup menu. Sometimes however, the right-click doesn't get interpreted as a right-click, due to mouse movement, and it gets interpreted as a left-click or I have even had the icon disappear. Besides, it would be nice to eliminate having to do it via a popup menu.

Now, I have eliminated the need for the right-click menu. A simple left-click can unmount a partition. this snapshot shows what mounted icons look like:

There is no longer any need for icon themes to have the sets of drive icons representing mounted and unmounted, for example 'drive48.png', 'drive_mnt48.png' and 'drive_mnt_boot48.png'.
Now, all that is needed is 'drive48.png' and the other two are autogenerated using a modified form of Vovchik's PNG overlay utility, named 'pngoverlay', see previous post:
http://bkhome.org/blog/?viewDetailed=02359

This requires two template images, /usr/local/lib/X11/pixmaps/closebox_outline48.png and /usr/local/lib/X11/pixmpas/locked_outline48.png. These latter two images can also be included in a desktop icon theme if it is desired to overwrite the defaults in /usr/local/lib/X11/pixmaps.
'pngoverlay' now generates 'drive_mnt48.png' and 'drive_mnt_boot48.png'.

To achieve the above, the scripts /usr/sbin/icon_switcher and /usr/sbin/icon_switcher_cli were modified.

Now that there is a little "close box" on the icon, if the mouse-pointer position is determined when a left-click is made on a desktop icon, it is then possible to perform an unmount.
This is what is done, using the 'getcurpos' utility. The script /usr/local/bin/drive_all has been modified to run 'getcurpos' and detect if there was a click in the "close box".

All of the above depends on the existence of /usr/local/lib/X11/pixmaps/closebox_outline48.png and /usr/sbin/pngoverlay. If either are missing, then there is a fallback to previous behaviour (not supporting the "close box", and the 'drive_mnt48.png' and 'drive_mnt_boot48.png' images will be required in a theme).

This is in Woof, will be uploaded very soon.

'pngoverlay' utility

Vovchik developed a little program, written in BaCon, to overlay two PNG images:
http://murga-linux.com/puppy/viewtopic.php?t=48901&start=135

I modified it a little bit, added entry of image filenames on the commandline, CLI only (conversion only, no output window), and recognises sizes of original images.

The executable is /usr/sbin/pngoverlay (31KB), and I also included /usr/sbin/pngoverlay.bac. I put these straight into rootfs-skeleton in Woof, as I plan to use it in system scripts.

There is also a new image, /usr/local/lib/X11/pixmaps/closebox_outline48.png in Woof, the purpose of which will be revealed soon.

I will upload the code very soon, that is, Woof will be uploaded.

Desktop drive icon, session entire partition

When I was recently playing with the Edubook, with session saved to entire partition, which in my case is /dev/sda2, I noticed a problem.

When a user clicks on the desktop "sda2" drive icon, Rox opens at /initrd/pup_rw.

This is bad. Users should not be opening a Rox window at a layer inside the aufs. The layers look like this:

/
/initrd/pup_rw <</initrd/pup_ro2 <<
/initrd/pup_ro4 <<

It will likely cause great confusion to most users when they click the "sda2" icon and Rox opens with contents that look like that at "/" but not quite, less directories/files.

Besides that, users should not be reading/writing directly in the aufs layers.

So, I have modified /usr/local/bin/drives_all (which gets copied into /root/.pup_event/drive_sda2 etc.) and /usr/sbin/pmount, to open "sda" at "/".

This will apply to any mount point /initrd/pup_*.

2.6.39.3 kernel recompile, take-5
Here we go again! As pemasu advised that vga-switcheroo requires the debugfs to be enabled in the kernel, I have recompiled it. Same links as per previous post:
http://bkhome.org/blog/?viewDetailed=02354

I took the opportunity to extend my script 'bk-build.sh'. It is now totally automated right through, from patching the kernel, compiling, building the source .sfs, and now also builds the PET package. This is very convenient, considering how many times I have been recompiling the kernel.

Human Probiotic Infusion
Quote from Catalyst TV program:

Seven years ago, Coralie was struck down by a strain of C.difficile after taking a course of antibiotics for a simple dental procedure.

Over a year ago I also was prescibed an antibiotic by a dentist after having two teeth extracted. After taking the antibiotic, I got diahorea, and I looked up the antibiotic on the Internet and was alarmed at the warnings about its use, and the high risk of getting C.difficile.
After the treatment, my tummy did not feel right, and I went to a doctor, had stool tests, all clear. I asked the doctor, what if there had been C.difficile, and he said it would be treated with more antibiotics, really strong ones.

I'm wondering if my current tummy troubles are connected with that earlier problem. But, the latest tests have shown no C.difficile. But, there are hundreds of different flora in the tummy, and any imbalance can cause trouble.

The main problem is that C.difficile has mutated, and in the USA, 300 people are dying every day from it. It is an epidemic, created by modern medicine.

Anyway, for those who have contracted C.difficile, there is treatment that works (95% success), a very old one, known as Human Probiotic Therapy. An Australian Doctor has pioneered this, as reported last week in the Australia TV series Catalyst:
http://abc.com.au/catalyst/stories/3269844.htm

These are the places where the work is being done:
http://www.cdd.com.au/
http://probiotictherapy.com.au/

Apparently there are good flora that actively attack C.difficile and also destroy the spores. The Doctor thinks that one day these will be isolated and grown, so poo won't have to be injected, but that is still a long way off.

My new laptop running at 1366x768!
Oh boy, am I happy! I purchased an Acer Aspire AS5742G laptop, but then found that it has "hybrid graphics", that is, two GPUs, and Wary would only display at 1024x768 and only 'vesa' driver worked.

So, I installed Mageia 1, and alas, that also can only run at 1024x768 and vesa driver. I messed around in Mageia, couldn't get anything better ...but things get interesting...

pemasu has been working in this area, see helpful posts to this blog:
http://bkhome.org/blog/?viewDetailed=02354
Also forum thread on hybrid graphics:
http://murga-linux.com/puppy/viewtopic.php?t=67561

I have just now built Drake Puppy with the 2.6.39-3 kernel built as announced in above blog link. I let Xorg configure itself automatically, and lo-and-behold it came up with the 'intel' driver and at 1366x768 (the correct resolution)!

I didn't have to do any "VGA switcheroo" operation. That's why I'm so pleased. What is interesting is that Mageia couldn't do it, but Drake Puppy is built from Mageia's Xorg packages.

Anyway, it is important for the kernel to have VGA switcheroo enabled, so I am recompiling (take-5) with debugfs enabled, as pemasu has advised. I might then experiment with switching my laptop to use the nVidia GPU.

2.6.39.3 kernel recompile, take-4
Fourth time, fifth time? I forget. The previous compile and upload was reported here:
http://bkhome.org/blog/?viewDetailed=02349

The source link is the same, I have updated it:
http://bkhome.org/sources/kernel-2.6.39-3/

The latest PET (23.3MB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_kernel-2.6.39-3-pae-i686-1-w5.pet

The reason that I have recompiled the kernel, is that I have discovered that my new laptop has "hybrid graphics". That is, it has Intel graphics plus nVidia graphics chips, with a multiplexer so that either can access the screen.

To be able to switch betwen the GPUs, a kernel option has to be enabled "Graphics support -> laptop hybrid graphics - GPU switching support". This enables what is called "VGA Switcheroo". nVidia refers to it as Hybridpower, and ATI as Powerxpress.

Note, I also enabled ISA bus support, and all the ISA sound drivers.

Note also, I will need to do some work on the Xorg Wizard to support this, and hopefully will succeed.

New Acer Aspire AS574G-374G64Mn
I was in Perth yesterday, and I showed one of my Edubooks to the NGO person from India. After a short demo, she asked if I could install Windows on it.

Anyway, early this morning I bought the newspaper and saw a laptop advertised at Rick hart, just $529 and $99 cashback, bringing it down to only AU $430. Look at the specs:

http://www.cheapcheapdeals.com.au/retailer-deals/electronics/computers-laptops/acer-aspire-as5742g-374g64mn-laptop-on-sale-save-171-further-99-cashback/

I'm running it right now.

If the NGO person doesn't want my Edubook, I might donate my old laptop, an Acer Aspire 3681wxmi -- this has been my workhorse for years. I just did a quick search, I bought it back in November 2006:

http://puppylinux.com/news/news211-212.htm

It has though, become an increasing struggle. Slow 1.4GHz Celeron CPU, only 512MB RAM, 60GB HD. I do some very long compile jobs, like the T2 compile, that takes days. Just a few days ago when compiling SeaMonkey 2.2 I was listening to the hard drive thrashing, and thinking that I had to do something about it. Then there's storage, 60GB just doesn't cut it.

The old Acer has Windows XP, I will enlarge the partition, but probably leave a small ext3 partition with Puppy on it, plus a swap partition, and GRUB boot management.

My new laptop has Windows 7 Home Premium. I thought that I read somewhere that there is a problem with shrinking a HD partition that has Win7 on it? I need to do so, will have to research that.

The i3 370M CPU seems ok:

http://www.notebookcheck.net/Intel-Core-i3-370M-Notebook-Processor.32767.0.html

...curious, the CPU has on-chip graphics, yet the laptop has nVidia GeForce GT 420 graphics chip.

dhcpcd 5.2.12
There was some recent discussion on the Puppy Forum about Wary having an old version of dhcpcd:
http://murga-linux.com/puppy/viewtopic.php?t=68298

Yes, this is because Puppy 4.3.1 and Wary 5.1.2 were originally compiled from source in T2, and T2 is stuck on using a very old version -- 3.1.8. I just checked the T2 website, yep they are still using that old version (released in 2007).

I downloaded the latest, 5.2.12, and configured it like this (these are the recommended options for Linux):

# ./configure --libexecdir=/lib/dhcpcd --dbdir=/var/lib/dhcpcd --sysconfdir=/etc

Here is the PET (46KB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/dhcpcd-5.2.12-w5.pet

Something is tugging at my memory about this... yes, I did compile dhcpcd 5.2.5 in Wary late last year. But, it wasn't put into Wary 5.1.2, must have forgot about it.

Wary 5.1.2 on Gecko Edubook
I have Wary 5.1.2 running nicely on the Gecko Edubook netbook. See previous post
http://bkhome.org/blog/?viewDetailed=02350

The main problem with this netbook is the under-powered 1GHz RDC Vortex86MX CPU.

Everything in Wary is compiled for a i486 CPU, except for libvpx, ffmpeg and perhaps mplayer. However, back at Wary 0.3, I did have ffmpeg (and perhaps mplayer) configured for i486 CPU, so see above link for PETs that I installed (as used in Wary 0.3) to fix Mplayer in Wary 5.1.2 on the Edubook. Note though, ffmpeg does not use libvpx, these are pre-webM packages.

There are some issues. I couldn't get Wary 5.1.2 to boot off a USB stick, I had to hack the 'init' script, but I will probably have that fixed for the next release of Wary. However, if you have an external USB optical drive, you should be able to boot from that and install Wary to the Edubook.

I chose an ATA/IDE hard drive installation in the Universal Installer, even though it is actually flash media. I partitioned the media with sda1 as a small 256MB vfat, and sda2 as ext3. I ran Grub4DOS after installing Wary, which modified the MBR. I did a frugal install.

SNS (Simple Network Setup) recognised my wifi network, but could not connect. That is something I need to checkout. Anyway, Dougal's Network Wiizard did connect, but when I powered up this morning it did not connect -- I right-clicked the desktop 'connect' icon and chose "Autoconnect" and it then did connect. Why not automatically at bootup? -- another thing to checkout.

On first shutdown, I chose to save the session to the entire sda2 partition. This is a good choice for non-technical users I reckon, as they won't have to worry about resizing the save-file.

It all works, wifi (apart from issue above) (I am typing this message in the Edubook!), sound, powerof.

I am giving this Edubook away, but have another that I will work on, tackle the issues -- or rather, that will go into my to-do list.

Need Mplayer PET for Edubook
Sometime ago I donated an Acer netbook to an NGO in India. I also gave them one of my Intel Classmates (I had two Acers and two Classmates).

I also have two Gecko Edubooks, one given to me for evaluation, the other I purchased, and I was going to donate one of those also, except that I never got it working right.

Well, the lady from the NGO will be in Perth soon, so I am setting up one of the Edubooks, running Wary 5.1.2. There are some issues that have emerged, that I need to look into, but anyway it is working, running it now, via wifi. Sound works too.

One problem is the illegal instruction crash with libvpx, ffmpeg and mplayer. I recall someone posted about PETs from another puplet that do work in the Edubook, but I have hunted through my blog and the forum, can't find that info.
Does anyone remember that? It was 'pskin' who was posting here about it and he/she mentioned about a successful compile, but someone also mentioned about PETs available for download.

I will be handing this Edubook over in a couple of days.

Linux 2.6.39-3 kernel
I have previously announced compiling of this kernel:
http://bkhome.org/blog/?viewDetailed=02345
http://bkhome.org/blog/?viewDetailed=02342

There has been a lot of recompiling. Fortunately now it is automated (see script bk-build.sh).
Some notes:

1. PAE enabled
2. Minimum Pentium Pro (i686) CPU
3. ISA bus not supported
4. Voluntary preemption (not full preemption)
5. usb-storage patch

I have removed earlier PETs. The PET is now in the 'common' repo for use by all puppies (23MB):
http://distro.ibiblio.org/quirky/pet_packages-common/linux_kernel-2.6.39-3-pae-i686-w5.pet

The sources are uploaded here:
http://bkhome.org/sources/kernel-2.6.39-3/

A note on point 4. The .config file had full preemptible kernel, which is supposed to give snappier responses. However, using this kernel, I sometimes noticed hesitations, more so than with previous puppies. All previous kernels I have compiled have been configured with voluntary preemption -- which as far as I know will give full attention to the window that is in focus and be less likely to allow other apps to take time away from that.

Another difference with this .config is that it has "cgroups" turned on, and that I think affects scheduling, I recall reading somewhere. I don't know if that is causing the problem that I have observed.

Wrong ROX Filer PET
Oh, recent Woofs have been building puppies with an old version of ROX Filer, one that has the right-click focus bug.

This came about because I removed the 'pet_packages-quirky' repository, which was in the Woof search path. Instead, the Woof scripts have fallen right back to 'pet_packages-4' to find a ROX Filer PET package.

Ok, I have put the correct ROX Filer PET into the 'common' repo. This was compiled in Wary and should work in all puppies.

Download (374KB):
http://distro.ibiblio.org/quirky/pet_packages-common/rox_filer-201012git-patched.pet

The latest database file for the 'common' repo can also be downloaded, for use in Woof (39KB):
http://distro.ibiblio.org/quirky/Packages-puppy-common-official

Woof supports xz-compressed SFS
Building a Puppy with the 2.6.39-x kernel, the '3builddistro' script now asks if gzip or xz compression is wanted for .sfs files.

I did a test build. Wary with gzip compression gave me a 121.5MB .iso file, then with xz compression gave a 104.6MB .iso file.

Hmmm, should be easy to take out some less-useful kernel drivers and get that under the magic 100MB figure. This Wary has everything, including Flash player and SeaMonkey 2.2.

I booted off a usb stick to test if responsiveness is any slower, particularly opening applications. However, I can't tell. Maybe apps do open a tad slower, but perhaps it isn't something that you would really notice. I'll have to upload a build and see what you guys think.

squashfs-tools 4.2
I have compiled this in Wary. Biggest feature is that it supports xz compression, as does the squashfs.ko driver in the 2.6.39 kernel.

It is in the 'common' repo (83KB):
http://distro.ibiblio.org/quirky/pet_packages-common/squashfs-tools42-4.2-gzip-xz-w5.pet

There are three different squashfs-tools PET packages, for v3, v4.0 and v4.2 Squashfs. It is a bit involved to explain here, but the package list in Woof, for example in file DISTRO_PKGS_SPECS-puppy-wary5', has "yes" to all three:

yes|squashfs-tools||exe|
yes|squashfs-tools4||exe| #use instead of squashfs-tools42 for <2.6.39 (see Woof file PKGS_MANAGEMENT, variable PKG_PREFER_OLDER)
yes|squashfs-tools42||exe| #use with 2.6.39+ kernel, supports xz (see Woof file PKGS_MANAGEMENT, variable PKG_PREFER_OLDER)

But only the first and a choice between second and third is made by script '3builddistro'.

The first package has executables 'mksquashfs3' and 'unsquashfs3', the next two both have 'mksquashfs4' and 'unsquashfs4'.
'3builddistro' creates symlinks 'mksquashfs' and 'unsquashfs' to 'mksquashfs4' and 'unsquashfs4'.

2.6.39-3 kernel recompiled
I have recompiled the kernel. See post about previous compile:
http://bkhome.org/blog/?viewDetailed=02342

Some very small configure changes, but this time I put in my usb-storage patch. All patches and sources are uploaded:
http://bkhome.org/sources/kernel-2.6.39-3-patched2/

I have modified script /sbin/wait4usb in the initramfs in Woof to work with a kernel with builtin usb drivers and with or without the usb-storage patch. The current uploaded version of Woof will work also, as wait4usb just has fixed sleeps -- the new version has a test that may give more exact waiting for all usb drives to be ready (utilising the usb-storage patch).

Here is the kernel PET (23MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/linux_kernel-2.6.39-3-patched2-w5.pet

Woof uploaded, July 12, 2011
This is commit '20110712172303'.
Previous upload was on July 4:
http://bkhome.org/blog/?viewDetailed=02331

This Woof has modified /sbin/wait4usb and /init scripts in the initramfs, and modified /etc/rc.d/rc.sysinit in the main filesystem, to support the 2.6.39-3 kernel with hid and usb core drivers (including usb-storage) builtin to the kernel.

The scripts also support the kernel without my usb-storage patch. That patch was used in 2.6.34+ kernels, see 'usb-storage.patch.gz' at:
http://bkhome.org/sources/kernel-2.6.35.7/patches/

Without that patch, we don't have any means to determine when all usb drives have been fully scanned and are ready for use. Instead, I have put a "sleep 5" into /sbin/wait4usb. This is very crude, but the kernel developers took away the only means that I know of to know when all usb drives are ready, with 2.6.34+ -- this issue has been reported in this blog.

For experimentation, wait4usb writes to /initrd/tmp/usb_drvs_found_1sec_intervals each second, so we can get a report on how long it takes before drives appear in /sys/block.

Booting from CD on my Pentium4 desktop, with a usb pen drive plugged in, /initrd/tmp/usb_drvs_found_1sec_intervals has this:

USBPCI: 0C0320 0C0310 0C0310 0C0300 0C0300
USBDRVS:

USBPCI: 0C0320 0C0310 0C0310 0C0300 0C0300
USBDRVS:

USBPCI: 0C0320 0C0310 0C0310 0C0300 0C0300
USBDRVS: sdc

USBPCI: 0C0320 0C0310 0C0310 0C0300 0C0300
USBDRVS: sdc

USBPCI: 0C0320 0C0310 0C0310 0C0300 0C0300
USBDRVS: sdc

...which means that sdc appears after 2 seconds.

This might be fine for modern hardware, but what about USB1 interfaces? What about USB optical drives? -- unfortunately mine was stolen so can't test that.

Tonight I am going to recompile the 2.6.39.3 kernel, with a few small changes. I am also going to test with my "usb-storage" patch put back in.

Note, I have built Wary 5.1.2.3 with the 2.6.39-3 kernel, uploading it now. If anyone is interested in trying it, to see if it detects their usb ok, you're welcome. Posting that /initrd/tmp/usb_drvs_found_1sec_intervals would be most helpful. I'll post when it is uploaded.

SeaMonkey 2.2
I have compiled SeaMonkey 2.2 in Wary. I did report a lockup problem with 2.2beta3, but that did not re-occur, see report here:
http://bkhome.org/blog/?viewDetailed=02332

I'm running 2.2 right now. Starts fast, snappy performance. So, I reckon this will be in the next release of Wary.

But, if you want it now (16MB, 4.8MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/seamonkey-2.2-w5.pet
http://distro.ibiblio.org/quirky/pet_packages-wary5/seamonkey_DEV-2.2-w5.pet

I now consider Wary to be the "base system", and applications compiled in Wary should work in other puppies such as Lucid, Drake, Spup and Dpup.

Linux 2.6.39.3 compiled
I have compiled this kernel in Wary 5.1.2. It is configured somewhat differently from the usual Puppy kernels:

1. HID and USB core drivers built-in to the kernel.
2. PAE-enabled, can handle up to 64GB RAM.
3. Does not have my usb-discovery patch for usb.c.

Note that I have only compiled three 3rd-party drivers, aufs, ndiswrapper and ltmodem.

Points 1 and 3 mean that the kernel will not work with Puppy as-is. The 'init' script in the initramfs needs to be modified. I have partly done this and have Wary booting off a USB stick. However, more work is needed, and I will continue this promptly and upload a Woof that should work properly with this kernel (as well as earlier "classic" Puppy kernels). Stay tuned.

If you want to grab the files:

Here is the kernel PET (23.2MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/linux_kernel-2.6.39-3-w5.pet

And the sources, including an .sfs file:
http://bkhome.org/sources/kernel-2.6.39-3/

Busybox 1.18.5 in initramfs
Up until now, the initramfs has used Busybox 1.4.2, an old version. I intend to compile a recent kernel and investigate reported problems with USB, and some code I contemplate putting into the 'init' script in the initramfs needs the 'xargs' application -- which the old Busybox doesn't have (or rather, is not configured to have xargs), so I decided to upgrade Busybox.

The initramfs is using full versions of 'cp', 'find', 'grep', 'losetup', 'lsmod', 'modinfo', and 'modprobe', as the old Busybox applets are inadequate. With the upgrade, I have checked out whether the BB applets are now suitable. Unfortunately, mostly not.

For BB 1.18.5, I have had to keep these full programs:

cp
-u option is not supported.

find
-noleaf, -not options not supported.

losetup
needed options not supported.

modinfo
Broken, is supposed to treat 'usb-storage' and 'usb_storage' as equivalent. Also, does not return all informnation ('Depends:' field is missing).

modprobe
inadequate. modprobe still does not support a modalias on commandline.

However, I have promoted 'grep' and 'rev' to use the Busybox applets.

The new Busybox static executable is now in Woof, along with the '.config' file (also now have the 'bbconfig' applet, that prints out the configure settings).
I did a test build, seems to work ok.

I'm back!
I was away for a week, in Melbourne visiting my daughter.

Good to be home, my place wasn't burgled (again) while I was away, good!

I enjoyed excursions into the Melbourne CBD. As I am vegetarian, I scouted around for cheap vegetarian meals, and found these two:

Om Vegetarian: http://omvegetarian.com/
Crossways: http://www.iskcon.net.au/food/crossways

Om Vegetarian charges AU $6.50 for all-you-can eat, Crossways charges AU $7.50 (or $5.50 concession) for all-you-can-eat.
Fantastic prices.

I love Indian food, so these places suit me very much. However, Crossways use Ghee (clarified butter) for cooking, which conflicts with my trend toward becoming vegan.

BaCon 1.0.23
The BaCon BASIC compiler has reached version 1.0.23. I have compiled it in Wary and put the PETs into the 'common' repo for all puppies to use.

Download (63KB, 924KB):
http://distro.ibiblio.org/quirky/pet_packages-common/bacon-1.0.23.pet
http://distro.ibiblio.org/quirky/pet_packages-common/bacon_DEV-1.0.23.pet

The runtime PET has /usr/lib/hug.so, which BaCon apps using HUG (Highlevel Universal GUI) may optionally use to make executables smaller (ex: see /usr/sbin/welcome1stboot.bac).

The _DEV PET goes into the 'devx' and has the compiler, include files and documentation. It also has 'bacongui' which is a GUI IDE.

Changes:
http://basic-converter.org/stable/CHANGES

HUG has undergone significant development since BaCon 1.0.22, see docs:
http://basic-converter.org/hugdoc.html

For more info, see my BaCon intro pages:
http://bkhome.org/bacon/

BlueGriffon web page editor
I have compiled BlueGriffon 1.0, a web-page editor based on Mozilla code. I used the Firefox 4.0.1 source and the BlueGriffon 1.0 patch.

I had hoped that I might be able to get Firefox and BlueGriffon to work together, sharing many common files, but that isn't so simple.

I compiled in Wary 5.1.2. I did install my libpng with animation PET, however the Firefox/BlueGriffon configure complained that is not suitable, so BlueGriffon is using internal libpng still.

Download the PET (14.9MB):
http://distro.ibiblio.org/quirky/pet_packages-wary5/bluegriffon-1.0-w5.pet

Here is the BlueGriffon home page:
http://www.bluegriffon.org/