Systemd debate continues

I wrote about 'eudev', a fork of 'udev' recently (udev has been merged into systemd):
http://bkhome.org/archive/blog2/201303/eudev-fork-of-udev.html

The debate about whether systemd is a "good thing" or a "bad thing" continues:
http://www.linuxadvocates.com/2013/04/systemd-new-pulseaudio.html

And this:
http://bbs.archbang.org/viewtopic.php?pid=17583#p17583

---

Comments

Systemd again
Username: BarryK
More... http://www.linuxadvocates.com/2013/04/developer-dissatisfaction-looms-with.html

puppy's own initd in C
Username: technosaurus
"I have written an init replacement (just for learning purposes) that uses only data passed from the kernel command line, so it will work without mounting anything and can start X in ~1s (0.8s of that is kernel expansion) ... its really not that difficult and threading is stupid when the binary is <1Mb (each thread allocates 2Mb minimum AFAIK) so it is better to fork() ... chalk that up to application developers trying to do embedded programming, but then again with all the superfluous dependencies, it probably is 2Mb+ With X starting without a file system mounted, that means most of the startup can happen in a terminal window in X (everything except mounting /proc, /sys, /dev and /dev/pts + some ENV vars) If someone writes up a wish list, I could probably extend it for Puppy. I have code snippets for mounting loopback'd squash with aufs/unionfs which is the most difficult part. This would eventually need to be integrated with the installer process for non-dev-types. I would also like to finish my patch for vga=max command line to automatically load the highest resolution and bpp without the user having to deduce a bunch of video codes (for use with Xfbdev - for 1st run and failsafe)

Tags: general