Month: July 2005

OpenAFS gets a much needed face lift

I blog so rarely these days. So, in the beginning (100 000 bugs ago) there was Daniel Robbins who got all the incoming bugs and he assigned them. Then Gentoo grew to 15-20 developers and I came along and we created bug-wranglers to handle the majority of incoming bugs (that are not recruiters, devrel or infrastructure related). As of a few weeks ago we created two new aliases: maintainer-wanted and maintainer-needed. The first is to put all the “this is a new package, I want it in portage, and optionally I’ve attached an ebuild” bugs. The second is for packages that are pretty much orphaned and unloved in the portage tree. Among the second list was openAFS. Then along came Stefaan De Roeck (don’t click, you probably can’t see that bug anyway). And he spent a few weeks combing through all the openAFS issues, and came up with two brand new ebuilds (no packages.gentoo link yet because I just finished committing them to portage for him) to fix all those issues.

A few things of note about the new openafs stuff (1.2.13 for the 2.4 kernel users and 1.3.85 for the sexy users (2.4 or 2.6, doesn’t matter). There are now three packages. That’s right, I said 3. We have openafs-kernel, which serves the purpose of not having to recompile openafs everytime you install a new kernel. Now, you have a much smaller compile of just the kernel module. Additionally, openafs itself now installs into FHS-correct locations (instead of crap like /usr/afs and /usr/visi or whatever that was). There are those who vehemently disagree with any sort of fhs-correctness for packages like heimdal and openafs, but I personally am not totally convinced. I’d rather see patches and correct ebuilds and problems with current ebuilds — like actual problems caused by current ebuilds.

So, that brings us to this: what about those legacy binary-only type programmes that expect openafs in its old locations???? Well, that’s easy: Stefaan created openafs-legacy to install all sorts of symlinks to bugger your filesystem up with /usr/afs and the like.

In sum, please test, please report.

This is why the lefties keep losing

Because they would do the same things as their counterparts on the right. Yep, they’d even sell their souls for the right price. Honestly, what do you do in a situation like this?

Here’s a guy (Brian’s ex-friend) who claims to be anti-establishment and wanting others to tell the truth, but he is unwilling and completely unable to do exactly that his own self. He can’t even write as himself (follow the links in Brian’s post, I’m not giving the faker any direct publicity, sorry). It’s exactly the sort of behaviour that Pete rants on about and then turns around and does. Which makes him what? A money-grubbing hypocrite in my book. It’s apparently not about the truth. It’s about advertising. And you know what else? It gives him as much credibility as the idiot commenters.

I am not a Jedi.

But I could be with your vote!! All kidding aside, Gentoo’s organisational structure is undergoing a change. This is the culmination of a lot of proposals from a lot of different people. The old TLP management structure simply did not scale well enough to retain its effectivity. It’s no secret that the Gentoo project has grown: very large very fast. It’s been doing this for about 3 years now, with no real signs of letting up.

Now, slarti did nominate me to be on the council. I’d like to thank him for thinking of me. I was a part of the initial metastructure: I believe I headed up devrel and qa and something else at the time. Devrel was quickly handed off to avenj who had been my parter-in-crime in being devrel before it officially existed. The QA project never really took off, despite my best efforts. Let me put that into perspective. I was heading up base, x11, and devrel which took most of my time. I took on QA with the express purpose of getting someone else to take it over. I tried a bunch of people (some of whom already actually _do_ qa), but nobody wanted it. Finally, out of the blue, Sven approached me and qa exists and is in process. The page I’ve linked to is just a placeholder, so don’t take it too seriously. In the next week watch for massive updates!

As for devrel, we did some things right and some things wrong. The wrongs are being fixed currently and I believe that the new and improved devrel under Deedra’s leadership will be a stronger and better one. It’ll probably never be a popular one though (at least not for the right reasons).

I’ll take QA as my failure to build a team. However, that’s been my strong point and the recurring theme throughout my time as a Gentoo developer. It’s what I do, it’s my M.O.: find what’s broken, find someone who can fix, put the two together, back away slowly.

Also, I’m the head (once again) of bug-wranglers, heading a team of two! Jakub of course is the primary bug-wrangler currently. So anyway, without getting too long winded — I’m not *highly* technical in the coding sense, but I have some pretty good ideas (cascading profiles, runlevels etc) and I can put together teams to make them come to pass (even if it’s two years after the fact).

May the force..