September 6th, 2006
So congratulations to the GNOME team on the release of 2.16, a nice shiny release indeed. And it was in Cooker as soon as the .0 tarballs were tagged this morning, even before the release announcement, which is a refreshing return to our old ways after the slightly behind-the-times 2006!
I’ve been working like a dog for the last week or so (uh, not that I don’t USUALLY…), doing stuff which ain’t exactly in the job description but which I do on the basis that it’ll make the stuff that IS in the job description much easier in a short time when I have to start dealing with people actually _using_ this new ball o’ stuff we’re calling 2007. This basically means I’ve been reading all 500+ mails on the Cooker list each day, checking the forums, and searching through Bugzilla to find particularly serious bugs that haven’t been zapped yet. I then mark them as release-critical and proceed to bug the hell out of the entire development chain until they get fixed. This makes me cranky and unloved and makes the final product better, so yay for that.
So far we have soundwrapper (the script that makes sure sound apps get run through the appropriate software mixing service and you get sound from all your apps at once) being majorly broken, desktop bluetooth being in a right mess, the network service being completed on boot before (in some cases) the network is actually *up*, causing subsequent network dependent services trouble, nvidia proprietary drivers not installing, some mice failing to work on One and other fun stuff. Most of which has been fixed, but not all yet. It is my fond hope that all this running around will allow us to release an RC1 that we can describe as a Release Candidate without crossing our fingers behind our backs and falling about laughing as soon as everyone stops looking. Still, Linux distros are rather like laws and sausages – if you want to keep enjoying them, it’s best not to watch them getting made. I’m confident we’re going to get it sorted before release time. And frankly beta 3 was already better than 2006 at the same stage of testing.
If you want to help out fixing some of this stuff or are simply morbidly curious, here is the list of open bugs on Cooker that are tagged as release-critical. Go fix!
A couple of things we need help testing – does anyone remember the problem 2005 and 2006 (i586 only, x86-64 was okay) had with certain SATA chipsets? The install would fail when it actually came to access the hard disks, and the workaround was to boot the installer with the noapic kernel option. I’d like to make sure this problem is squished for 2007, so if anyone whose system was affected by this bug could test the conventional installer for either beta 3 (out now) or RC1 (due very soon) and let us know via a comment on the bug report whether it’s fixed now, that’d be great.
The same applies for anyone who had to boot the 2006 installer with ide=nodma , except I can’t find a bug report for that one, so just let me know in the comments, on the forum or by email (awilliamson A T mandriva D 0 T com).
The final one is the omnibook kernel issue. This one manifested itself as a regression in certain laptop functions between 2005 and 2006. If your laptop’s power monitoring (battery level etc), and/or shutting down the system, worked fine in 2005 but not in 2006, you’re likely affected by this bug. It affects mostly HP Omnibook and Pavilion models, also some Toshiba Satellites and a few miscellaneous models from other manufacturers – there’s a fairly full list in the bug report. The problem was that the support for these features in these models was moved from the kernel (in 2.6.9) to an externally maintained module called omnibook (in 2.6.10). This module never made it into our stock kernel. Unfortunately it’s now unmaintained and in rather poor shape, so it’s not going to be in the stock kernel for 2007 either. It WILL, however, be in kernel-multimedia (as it was in 2006) and also in contrib as a dkms package, dkms-omnibook. It would be great if someone with an affected laptop model could test both kernel-mm and the dkms-omnibook package from current Cooker and report if the module works (you’ll have to load it manually), as we’ve just built it blind since we don’t have an affected model to test on. Please report results in the bug report. Thanks!