It can be sustainable as long as it can work with multiple communication methods or at least be buildable for either, and Gentoo has a good track record.

It would probably end up sporting patches to the parent udev code, but until kdbus is kernel included, we'll never know. Only Linus knows when if it will be added, and ever since the Linus/Kay squabble things have been silent. It could eventually be added or there is a chance it could end up as reiser4 ended up. Only time will tell.

As far as building colord...

Do you have these or similar build flags in the SlackBuild set?

After problems building with udev-182, I was looking for the simplest set of configure options that would result in a successful build. For me, those options were --disable-bash-completion and --disable-systemd-login. Your other options may be useful but, having succeeded with a basic build, I won't now bother with a full SlackBuild until there's more indication about which way Slackware will go with regard to systemd, udev, eudev etc.

chris

I know those options work against eudev-1.6 at minimum, so try Didier's eudev build first.

eudev is not meant for production so why bother with it ?

The Gentoo eudev's devs explained that in a conference where they invited also Mr. Poettering.

eudev was meant to be a toy udev replacement, nothing else.

It was started as a novelty, but it's become a viable project used by a lot of distributions especially Gentoo for OpenRC usage. The code is actively maintained and applied from the systemd code with enough changes to allow it to be a separate project that keeps it's own code hence it's name eudev simply means extracted udev.

Many lighter weight Linux distributions use it because the older methods of trying to extract udev grew to be too much work to maintain. Ask Bruce Dubbs of LinuxFromScratch about the painful problems they had from systemd-udev extraction using their script set and why eudev became a more sane and easier to use choice for the classic book.

Plus please post proof of claims where this statement applies to the modern implementation. The Gentoo developers are some of the most respected and active developers in GNU/Linux contributing to numerous projects. I would hardly consider it to be a toy after 10+ releases.

Gentoo does not use eudev by default, it uses plain udev, extracted from the systemd build.

Yes, but at the moment that implementation isn't the most up-to-date. Gentoo still uses extracted udev version 208 which is the last attempted extraction version most distribution developers have tried to use. If you want anything newer you either have to choose systemd itself, or eudev. Other than that you have classic udev like Slackware uses.

Bruce even wrote on the LFS mailing list that extracting udev from systemd-209 was very problematic and the scripts they had been using we're not effective in doing the job correctly. They attempted both systemd as a hybrid install but it was just impossible to use as they wanted for a minimalism style system, so they chose eudev.

Nope:Note that I currently do not have it installed because this is a system I use for evaluation of systemd, so I have the udev installed that comes with Gentoo's systemd package on this machine.

Interesting. Maybe they wrote a new extraction method or something, but that's actually nice to see. Good find Tobi.

How about this project ?

https://www.google-melange.com/gsoc/...Speed=noscript

Yes, a "free" alternative to systemd's logind would be very nice... It's been in development for a while now from the openbsd guys so lets keep our fingers crossed..

Has the OpenBSD project even put the official code repository up publicly yet?

The link to systembsd is two pages back in this thread, posted by lems and helpfully marked with '2 members found this post helpful'

Edit: Damn. The damn thing is down at the moment, or has moved...
2nd Edit: There is a code drop here
3rd Edit: ... which is exactly what Nh3xus posted in #100, three posts ago.