[00:00] mlankhorst: I can fix that. [00:01] bryceh: I added a 14*.diff to ubuntu branch for newer libdrm, but it's intended for debian too, I wanted to upload the whole x1.13 stack tomorrow and upgrade all video drivers on debian [00:01] ok [00:01] in debian-experimental [00:02] but after that back to dmabuf stuff, I want a working solution preferably soon so it can live in linux-next long enough for merge window hopefully [00:02] That'd be good. [00:03] there are some changes that would be nice to have in -next like deferred fput [00:05] but yeah I completely underestimated how much work it was going to be to really complete this :) [00:07] RAOF, how's system compositor going? I saw the status updates on the blueprint, which sounded like it's got a long row to hoe. However I also read in the slashdot comments that it's going to maim kittens, level small towns, and generally mark an end to computing as we know it? [00:08] not to mention drown puppies [00:08] And cause freak hurricanes. [00:08] :-D [00:08] but phoronix seemed to have picked up on my dmabufmgr code quickly [00:08] The bits that are left shouldn't be *too* hard. [00:09] Phoronix: everyone's favourite mailing list aggregator! [00:09] nah it's lacking the feature where it links to the source [00:10] but it makes up for it with beer references [00:10] can't drink alcohol any more due to meds, not sure if I miss it.. [00:15] actually I do miss the taste, none of the alcohol free beer here is any good, in germany however :) [00:16] It's probably a technical challenge to create good tasting alcohol free beer. [00:16] I think the more likely answer is that there's less of a market for it [00:17] I had a coffee roaster describe what the process is for making decaf beans. It's amazing there's any taste at all. [00:17] Probably a combination of the two. [00:17] bryceh: Using crazy solvents, or submerging them in supercritical CO₂? [00:18] (Incidentally, supercritical CO₂ is pretty awesome) [00:18] I forget the exact process but yeah they extract all the stuff from the bean, chemically remove the caffeine, then smoosh it back into the bean. Something like that anyway. [00:22] 6.14.99~really6.14.5+git20120712.gitg8dc07e6-0~systemcompositor1. It's the world's longest version string! [00:25] RAOF: backported ~precise1~ppa1 [00:26] W: xserver-xorg-video-ati-dbg: package-has-long-file-name 103 (112) > 80 [00:26] It has officially triggered a lintian warning :) [00:28] I think this would be longer still: xserver-xorg-video-ati-lts-quantal-dbg 6.14.99~really6.14.5+git20120712.gitg8dc07e6-0~precise1~ppa1 [00:30] Yes. Yes it would. [00:30] but if you want i can try to append ~systemcompositor1 to that [00:35] Aaand everything should now be system-compositor-capable. Back to base for debriefing and cocktails. [00:35] Or possibly coffee and mesa SRU review. [06:03] duh, there was a security update to xorg-server, which doesn't have the change for bug 921236. could someone roll out a new version for -proposed? [06:03] Launchpad bug 921236 in xorg-server (Ubuntu Precise) "[12.04 Xorg, xserver 1.11.3] Dual monitor, after entering password, mouse pointer stuck on LHS of screen, no desktop." [High,Fix committed] https://launchpad.net/bugs/921236 [07:32] morning [09:32] tjaalton: do you still need a new version rolled out for xorg-server? [09:56] RAOF: ok I updated ubuntu-precise with the security update again, can the xserver sru be pushed out again? [10:52] yikes, 10 hour queue on xserver.. === seb128_ is now known as seb128 [18:01] ricotz: great, that newer mesa is busted on intel [18:02] craploads of DRM_IOCTL_I915_GEM_CONTEXT_CREATE failed: Invalid argument and the unity panel doesn't work [18:02] Sarvatt, i havent looked into it, but i noticed some areas not refreshing [18:02] ^ with gnome-shell [18:03] it is working though [18:03] although it broke xserver master build ;) [18:07] Sarvatt, grr, weston already needs a newer mesa :\ [18:08] Sarvatt, http://cgit.freedesktop.org/mesa/mesa/diff/include/EGL/eglmesaext.h?id=e6a33570b73aa56c87818d7f67a122d4427b7841 [18:09] great DRM_IOCTL_I915_GEM_CONTEXT_CREATE is only in drm-intel-next-queued [18:10] Sarvatt, uh, but it is proposed for 3.5? [18:12] note to self: stop updating crap when it's time for an intel quarterly release [18:17] hmm :\ [18:17] so i guess i'll go back to a checkout from the 8th, then cherry-pick those wayland commits.. [18:20] Sarvatt, reverting the introduction of DRM_IOCTL_I915_GEM_CONTEXT_CREATE, is not an option? [18:21] i dont know where it broke and its a freaking nightmare patching this to build out of tree, but apparently it wasn't broken in an 0708 checkout RAOF did for the system compositor ppa :) [18:25] hmm, i see [18:27] Sarvatt, i dont see any reference to DRM_IOCTL_I915_GEM_CONTEXT_CREATE [18:29] Sarvatt, it is probably caused by libdrm [18:30] http://cgit.freedesktop.org/mesa/drm/commit/?id=a5b2946889471f6075852949f90f660e43b68532 [18:35] Sarvatt, http://cgit.freedesktop.org/mesa/mesa/commit/?id=860d5bdf984730f69cd19b4f7145f3c84b57d33d [18:37] Sarvatt, try just to revert this ^ [18:54] Sarvatt: hm should we try to sru all the X packages that aren't core X but used by X to reduce breakage when switching stacks? [20:01] will adding the x-swat ppa get my old nvidia card's HW acceleration going again ? running ubuntu 12.04 with a nvidia geforce 4 fx 5200 [20:01] I understand this is an issue with the newer version of xorg not playing nice with the 173 driver [21:40] RAOF: does that 0708 mesa in the system compositor ppa actually work for you on intel? because it doesn't here, unity panel is blank [21:43] Sarvatt, RAOF, to be specific unity3d ;) [22:43] osiris, have you tried the nouveau driver? [23:40] Sarvatt: You mean, unity doesn't display any icons? That's a unity bug, I think. [23:40] Sarvatt: At least, people who weren't (AFAIK) using the system-compositor PPA were also complaining about it. [23:45] RAOF: we should have a discussion about sru'ing libdrm at one point :) [23:46] What needs to be SRUd there? [23:47] i think it would make more sense if we sru all the libraries that would need an update explicitly surrounding X so the core X packages + mesa package itself can be swapped easier [23:50] so all the X drivers + X server + Xorg package itself would be renamed [23:56] I don't think that's likely to happen. [23:57] Or, rather, policy explicitly forbids it, and unless there's a really good reason I don't think the policy will be amended. [23:57] mlankhorst: I though libdrm was easy enough to rename (adding in a bonus plymouth-rename, which is annoying, but not critical)? [23:58] RAOF: it's actually easy to rename, but hard to upgrade [23:58] things break apart rather quickly [23:58] Hm. What upgrade paths were we supporting, again? [23:59] I would guess upwards mostly, but it just breaks too fast leaving a chance of leaving system unbootable [23:59] since it's core