Gentoo Developers Unhappy, Fork udev

[Teho]

With that said, Volunteer developers are more than capable of competing with professional developers when there is an order of magnitude difference in code.

That kinda requires that those developers actually know what they are doing but I seriously doubt that being the case here. Kay Sievers then again has developed udev for about nine years. These Gentoo users/developers seem to be only "good" at bashing Lennart, Red Hat and systemd. That certainly rises expectations.

[curaga]

Eh, at first the kmod removal looked bad, but with the rationale above, it certainly makes sense.


BTW, yes I hack init scripts often, and yes, I have edited udev source.

[Adarion]

Things become interesting. Sometimes a fork is a good thing and will accelerate development or overcome things and maybe one day it could be merged back. Though I'd still prefer one solution, not fragmented for such a core element of any Linux distribution.
But they also got to fix some issues in gentoo Diego "flameeyes" mentioned. I also noticed that there were complications on some big libs or core elements. (boost, glibc, udev, udisks...).

That is one of the few downsides of Gentoo. If upstream fucks up you're really soon to notice it - unless you go stable arch only. I still remember libpng breakage, incompatible with all before and I had to (revdep-)rebuild 80% of my system. Not even funny on a quadcore when your main system is unusable for more than a day.

[89c51]

The principle systemd developers are two guys and we have the relevant pre-existing work in our repository. We yet to do our first tag and our top priority is making progress toward our first tag. The plan is to reintroduce libkmod support when we have it rewritten as an optional dependency.

With that said, Volunteer developers are more than capable of competing with professional developers when there is an order of magnitude difference in code. The sheer amount of bloat in systemd is staggering and we are removing it from udev.

Just out of curiosity. You have a few seasoned linux devs -excluding Kay and Lennart- that tell you that what you are trying to do is quite wrong. Don't you feel a bit weird about that???

[EoD_]

Apparently systemd is coming faster to Gentoo than I feared. Here an excerpt from #gentoo-udev:

14:02:29 GMT    nenolod | will this kill-split-/usr stuff effect openrc next?
14:02:43 GMT  @Chainsaw | From what I've seen on the list a very vocal minority is running the show.
14:03:11 GMT    nenolod | the no-split-/usr stuff will force our hand to either fork openrc or something else
14:03:14 GMT  @Chainsaw | nenolod: It could affect openrc, if the silent majority does not speak up and let's
                        | this project die out, the systemd steamroller will be upon us about two months from now.
14:04:13 GMT  @Chainsaw | So if anyone is reading this that felt they shouldn't respond to gentoo-dev... 
                        | I think you should. ryao is bearing the brunt of this, and if you do not show your support 
                        | now he may think he's alone in this.

So, speak up if you support it: Let's redesign the entire filesystem

[EoD_]

Apparently systemd is coming faster to Gentoo than I feared. Here an excerpt from #gentoo-udev:

14:02:29 GMT    nenolod | Chainsaw: will this kill-split-/usr stuff effect openrc next?
14:02:43 GMT  @Chainsaw | nenolod: From what I've seen on the list a very vocal minority is running the show.
14:03:11 GMT    nenolod | the no-split-/usr stuff will force our hand to either fork openrc or something else
14:03:14 GMT  @Chainsaw | nenolod: It could affect openrc, if the silent majority does not speak up and let's 
                        | this project die out, the systemd steamroller will be upon us about two months from now.
14:04:13 GMT  @Chainsaw | So if anyone is reading this that felt they shouldn't respond to gentoo-dev...
                        | I think you should. ryao is bearing the brunt of this, and if you do not show your support now
                        | he may think he's alone in this.

So, speak up if you support it: Let's redesign the entire filesystem!

[ryao]

Just out of curiosity. You have a few seasoned linux devs -excluding Kay and Lennart- that tell you that what you are trying to do is quite wrong. Don't you feel a bit weird about that???

No opinion formed before our first tag has any value. The only weird thing is when end users decide to be parrots that echo such opinions to developers.

[ArchLinux]

Just out of curiosity. You have a few seasoned linux devs -excluding Kay and Lennart- that tell you that what you are trying to do is quite wrong. Don't you feel a bit weird about that???

Don't you feel a bit weird writing this..?

[XorEaxEax]

Like I said before when this came up, I feel uncomfortable when core functionality like udev is looking like it will no longer be developed as a separate component. As an Arch user I've switched to systemd and I like it, a bit faster at boot and much faster at shutdown, everything works (for me).

That doesn't mean I will always like it, or that something better would not come along. And in the case of the latter, tying udev to a specific init will only make it harder to experiment with alternatives as I see it.

So I think it's good we are seeing a non-systemd dependent udev fork, I just wish we could have friendly competition or better yet cooperation rather than what seems like lots of hostility.

[asdx]

Like I said before when this came up, I feel uncomfortable when core functionality like udev is looking like it will no longer be developed as a separate component. As an Arch user I've switched to systemd and I like it, a bit faster at boot and much faster at shutdown, everything works (for me).

That doesn't mean I will always like it, or that something better would not come along. And in the case of the latter, tying udev to a specific init will only make it harder to experiment with alternatives as I see it.

So I think it's good we are seeing a non-systemd dependent udev fork, I just wish we could have friendly competition or better yet cooperation rather than what seems like lots of hostility.

I couldn't agree more. Udev should be its own project. Period.