Udev and systemd to merge

Posted Apr 3, 2012 17:14 UTC (Tue) by JEFFREY (guest, #79095)
Parent article: Udev and systemd to merge

This seems logical. My only concern is at what point does init become a super-duper program with as many lines of code as Microsoft Word, and become equally unmaintanable -- as opposed to keeping systems small, and doing only one thing well to isolate bugs.

All I can do is trust that they know how to edit and prevent feature bloat. The various Linux contributers have done quite well, thus far.

Should the project be renamed 'Udevstemd?'

Udev and systemd to merge

Posted Apr 3, 2012 17:30 UTC (Tue) by cesarb (subscriber, #6266) [Link] (1 responses)

Zawinski's Law.

Quick, someone write an email reader for systemd, to stop the progression before it's too late. ;-)

Udev and systemd to merge

Posted Apr 4, 2012 7:24 UTC (Wed) by dgm (subscriber, #49227) [Link]

Mail or Emacs are poor matches for systemd. The Linux kernel on the other hand...

Udev and systemd to merge

Posted Apr 3, 2012 17:30 UTC (Tue) by mezcalero (subscriber, #45103) [Link] (2 responses)

Note that this doesn't mean that udevd is merged into PID 1. udevd and systemd stay two separate binaries, which are just built from the same tree.

The net result will be less lines of code, simply because we can reduce duplicate code (i.e. code which exists in both projects will now be replaced by the same implementation)

Make it a library?

Posted Apr 3, 2012 17:44 UTC (Tue) by Pc5Y9sbv (guest, #41328) [Link] (1 responses)

Hopefully, this effort will include thinking about how to factor that common part out into library form? Or perhaps a core event service and client bindings for interop? It would seem unfortunate if it was just internally shared code between two monolithic daemons.

I have never dug into the Android side of things, but wonder if there is a common ground that could serve the needs of server, desktop, laptop, mobile, router, and embedded environments. I realize that is asking a lot, particularly in terms of flexibility/performance/efficiency tradeoffs, but this feels like the kind of thing that could help keep Linux from branching off indefinitely into different niche platforms... I'm not even sure if it needs kernel side effort, given the latency and efficiency issues on mobile or embedded?

Make it a library?

Posted Apr 4, 2012 14:58 UTC (Wed) by RCL (guest, #63264) [Link]

When you start to reuse code it becomes unmaintainable. Everyone is afraid of touching the common part lest they break some functionality they cannot test for themselves (or even know about).

So unless the common part is really simple and plain and will *never* need any refactor/improvement, don't do that. The world is built upon redundancy, which helps parallelization of efforts.

Udev and systemd to merge

Posted Apr 3, 2012 17:39 UTC (Tue) by marduk (subscriber, #3831) [Link] (3 responses)

So I take it you are opposed to merging udev/systemd with emacs?

Udev and systemd to merge

Posted Apr 3, 2012 17:44 UTC (Tue) by JEFFREY (guest, #79095) [Link] (2 responses)

emacs doesn't already implement these?!?

Udev and systemd to merge

Posted Apr 3, 2012 18:15 UTC (Tue) by SEJeff (guest, #51588) [Link] (1 responses)

It likely does, but google will need to re-invent it from scratch poorly as it is GPLv3 and GPLv3 has cooties.

Udev and systemd to merge

Posted Apr 4, 2012 8:36 UTC (Wed) by k3ninho (subscriber, #50375) [Link]

Well, if they checked the package repository, they'd have found systemd.el and udev.el and thus would only need to fork the pre-GPLv3 edition.