Hi all, I understand that it Bryce's habit to give the Debian XSF a heads up on the plans that Ubuntu have for the X stack each release. As Bryce is off hacking on Launchpad this release, I'll be responsible for X this cycle. So here's my attempt at a “heads up” email! I'd like to ensure that we don't unnecessarily duplicate work. I get the impression that Squeeze will be freezing sometime relatively soon, so I'd guess that you don't want these mesa and X versions in Sid, but would experimental be appropriate? I've had good experiences in the past with the Debian-mono team, and while X is a bit more tied to platform specifics I'd like to get as much as possible done in Debian. I can easily test-build against Debian and do some limited testing, and encourage others to do so, but that might not be sufficient. If that isn't regarded as sufficient then we can try and ensure our changes are easily merged with git. I've somewhat been watching on the sidelines before now - what would you find most helpful? So, without further ado, here's what we're aiming for in the next Ubuntu release: * Planned versions: We aim for xserver 1.9 and mesa 7.9. These are fairly aggressive targets, and we'll want to track git and pre-releases quite closely. We'll be on a 2.6.35 kernel, so we'll have the new 0.0.16 nouveau ABI which will be reflected in our libdrm packages. For intel, we're looking at 2.11 or potentially 2.12. * Packaging changes: In support of our ARM friends we'd like to package mesa's GLES support. It looks like this would be split into a libegl1-mesa containing the library and a libegl1-mesa-drivers containing the dri, dri2 and glx drivers. I'd like to ship nouveau's mesa component in a separate libgl1-mesa-experimental package. Considering upstream's “no 3D support!” policy I've had relatively good experiences with these drivers myself, but I think we want them to be opt-in. Additionally, I understand that the r300g gallium driver is maturing rapidly, and there's a chance that it'll be the recommended 3D driver for mesa 7.9. If not, I think this would also be a good candidate for a libgl1-mesa-experimental package. Any comments, queries, complaints? What are the feelings for how we should collaborate most effectively?
Attachment:
signature.asc
Description: This is a digitally signed message part