[07:35] morning [07:52] https://launchpadlibrarian.net/127916308/HookError_source_xorg.txt [07:52] wth? [08:06] o.0 [08:06] guess we could switch to sna once 3.8 hits raring [08:06] and I'm preparing mesa master now [08:06] sure, makes the intel devs happy [08:06] you think? [08:07] I'm not sure it makes anyone else happy though! [08:07] asked ickle about it a month ago or so, he thought it was maybe 80% ready back then. they've fixed some issues since then, might be 9/10 by now :) [08:08] er, 90% [08:09] going to do 1 more testbuild of pixman, then push the thing [08:15] always fun merging new mesa upstream.. [08:20] :> [08:20] it's becoming more and more boring though [08:20] but hey tf2 fps increase, count me in [08:21] I'm hipster and knew about that before phoronix reported it [08:21] a ton of conflicts that need to be manually resolved [08:21] easy though, always take what's the new [08:24] when doing syncs of pre-release git, new maintainer entry should be Ubuntu Developers right? [08:25] i'd prefer ubuntu-x [08:25] dpkg-source is happy as long as the address has *ubuntu.com on it [08:25] i thought pre-release like that was special when there were no ubuntu specific changes [08:28] it's not a big deal [08:28] ok incoming pixman then [08:52] meh, robert hasn't kept up with wayland updates.. [08:55] or my mirror.. wtf again [09:00] right, mirror cleanup & sbuild-update fail, duh :) === yofel_ is now known as yofel [09:34] builds, with an additional uncommitted build fix.. [09:59] apw: so, were you going to upload 3.8 to raring soon? [10:05] tjaalton, hi, i there a nvidia 313.09 patch for 3.8 available? (i guess, i havent looked hard enough yet) [10:05] ricotz: tseliot is on them [10:05] great, thanks [10:05] :) [10:57] can someone sponsor pixman? it doesn't appear to be part of xorg package set [11:46] hi, i like to ask about a possible MIR of libxkbcommon. we will need libxkbcommon in main if we enable the wayland backend in gtk (bug #954352) [11:46] bug 954352 in gtk+3.0 (Ubuntu Raring) "Enable wayland backend" [Wishlist,Triaged] https://launchpad.net/bugs/954352 [11:47] bdrung: go for it [11:47] tjaalton: the description says that this library is experimental [11:47] ah [11:48] the current version of this package is a git snapshot [11:48] there is a 0.2.0 of it [11:48] ok, I'll update it [11:49] http://cgit.freedesktop.org/xorg/lib/libxkbcommon/log/ [11:54] tjaalton: the warning in the description came from bryce. bryce, did you add the warning, because the package was a git snapshot? [11:54] it was at some point [11:55] things have changed since [11:55] bryce: the versions-current page doesn't seem to track libxkbcommon upstream? [12:14] mesa snapshot builds properly now, doing libxkbcommon & weston next [12:33] ok first attempt at mesa done [12:36] what's the diff like? [12:40] bdrung: libxkbcommon 0.2.0 uploaded [12:41] few lines [12:41] show it :) [12:41] oh, it's in git already [12:41] not quite :) [12:41] it was the security update [12:42] http://paste.ubuntu.com/1512694/ [12:42] tjaalton: thanks [12:46] mlankhorst: so basically, mesa-common-dev should be backwards compatible? [12:46] pretty much [12:48] unless a package depends on glu_*.h :) [12:48] which got removed in 9.0 [12:49] but it shouldn't matter [12:50] bdrung, hi, is it possible to enable broadway too which doesnt need further depends and won't interfere either? [12:51] ricotz: what is broadway? [12:51] bdrung, the html5 backend of gtk [12:52] how much will the package size increase? [12:52] hmm, quite a question [12:52] ricotz: broadway will not need any new build dependencies? [12:52] no [12:52] ricotz: do we have an open bug report for it? [12:52] no [12:53] i was asked to do so in my ppa, and did for some time now [12:53] ricotz: and there were no regression? [12:53] it is a separate backend like wayland [12:53] ricotz: can you evaluate the size change? [12:54] i'll try [12:55] tjaalton: that should be pulled to libglu.*-dev then [12:55] mlankhorst: it's not in precise [13:00] weston needs to be updated in raring - see bug #1097685 [13:00] bug 1097685 in weston (Ubuntu) "libwayland and weston package versions in raring are incompatible" [Undecided,Confirmed] https://launchpad.net/bugs/1097685 [13:00] updating as we speak [13:03] bdrung: of course libxkbcommon is experimental. so's wayland... [13:03] just not api-wise, not anymore [13:05] jcristau: i was referring to the API (not how often the API will be extended) [13:07] erm afaict libglu1-mesa-dev from precise includes GL/glu.h and GL/glu_mangle.h, so installing it will work even on renamed stack [13:09] mlankhorst: of course, dunno what I was looking at.. [13:16] sigh, why on earth doesn't the ubuntu cairo package build the egl backend? [13:17] bdrung, the stipped gdk lib would grow by ca. 90kb [13:17] ricotz: and in percentage? [13:18] 20 [13:18] 550 to 640 [13:19] 90 kb is not that big if there are user wanting that backend [13:19] ricotz: feel free to extend the debdiff and state the dependency/size impact of the change in the bug report [13:20] it is getting in shape in 3.7.x, while staying on 3.6 it might be not the great to use [13:21] looks like an ancient cairo merge left the egl backend disabled, nice [13:22] tjaalton, it is intended to be disabled due nvidia blob memory problem [13:22] ricotz: ah, meh [13:23] ricotz: that was on oneiric? [13:24] tjaalton, still valid afaik :( [13:24] ok tehn [13:24] then [13:25] tjaalton, feel free to check [13:25] nah [13:25] the nvidia kernel module will massivly increase in size [13:25] yeah i remember now [13:27] ricotz: then a bug report with the current findings would be nice as reminder once we get >= 3.7.x into the archive [13:27] it isnt likely it will land in raring [13:28] 3.7.x i mean [13:28] therefore the bug report [13:28] tjaalton: have you seen http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=691699 ? [13:28] Debian bug 691699 in src:libxkbcommon "libxkbcommon: [PATCH] upgrade to 0.2.0" [Wishlist,Open] [13:29] bdrung: no [13:29] tjaalton: it would be nice to review that and reduce the lintian output [13:29] http://paste.debian.net/223110/ [13:29] ah great, pixman 1-0 and input update is enough for new xserver :) [13:30] bdrung: huh, I only got one warning [13:30] about out-of-date-standards-version [13:31] tjaalton: he runs lintian with the pedantic option [13:31] tjaalton: are you using lintian 2.5.11ubuntu1? [13:31] which outputs lots of irrelevant stuff [13:32] bdrung: whatever is in raring [13:32] jcristau: ah [13:33] tjaalton: i use --pedantic -IE [13:35] duh, I wasn't aware it didn't use debhelper 9 [13:36] bdrung, ah there is a bug https://bugs.launchpad.net/precise-backports/+bug/933641 [13:36] Ubuntu bug 933641 in gtk+3.0 (Ubuntu) "Please backport/enable HTML5 (broadway) gdk backend for gtk+3.0" [Wishlist,Confirmed] [13:36] who uses that backend anyway? [13:36] the latest I heard was that it's still very much WIP [13:36] some kind of webapps [13:37] yeah, that is why i said 3.8 will be more appropriate to have it in a usable stat [13:37] e [13:37] ricotz: can you comment that bug report? [13:40] done [13:43] bdrung, is it possible to remove the "Precise Backports" reference completely? [13:44] from the mentioned bug [13:44] ricotz: it doesn't seem to be possible [13:48] tjaalton: will you apply the libxkbcommon patches? [13:48] bdrung: ones that matter [13:49] tjaalton: the patches 004 - 008 makes sense to me [13:49] the patches 002 and 003 were already covered by you [13:49] not 5 & 6 [13:51] mlankhorst, hi, btw, what is holding back wine getting into raring? [13:53] no idea? [13:56] tjaalton: why not? what speaks against using 3.0 (quilt)? why is the tarball generation still needed? [13:56] mlankhorst, hmm, so scott is not comfortable with 1.5 yet? [13:56] I don't upload wine to distro, just maintaining ubuntu-wine ppa :-) [13:57] i know, i am just hoping you would know why it isnt there yet ;) [13:57] so you use the ubuntu-wine ppa, of course! [13:58] i do, yes [13:58] been maintaining audio there mostly, the ubuntu official release is probably lacking all these changes [14:00] ok, keep that up! :) [14:01] the bug about supporting pulseaudio is >5 years old at this point. :P [14:10] ok guess I'll do a 'proper' rework of all patches to xserver 1.14 now, botched up when testing airlied's stuff [14:11] use ubuntu+1, and reset --hard it to the current branch [14:12] was thinking of a debian-experimental+1 first [14:12] or just bump d-e there [14:12] jcristau: opinions? [14:13] about? [14:13] whether to bump xorg-server debian-experimental branch to git master [14:17] shouldn't be an issue, i don't think we need to go through 1.13 [14:17] yeah [14:17] ah k, x1.14 it is then [14:17] branch history will be a mess, but oh well :( [14:18] because of xi2.3? [14:18] it's git, it will resolve the conflicts ;-) [14:19] * ogra_ wonders why nobody had the idea to send git to syria tzhen [14:20] :P [14:20] they don't use version control there [14:20] mlankhorst: that's not what i'm worried about [14:21] just having lots of merges of stuff that was never used/uploaded [14:23] on the d-e branch? [14:23] it was used, just not on debian :) [14:45] wow, that was easy, just needs inputproto / pixman version bump [14:45] and removal of input barrier patch [14:51] bit suspicious that all other patches just apply, though [14:53] heh [14:53] not much has changed though [14:53] feature wise anyway [14:54] true [14:55] and I guess the 1.13 rc's grabbed most patches we cared about anyway [15:40] sdksyms.o:(.data.rel+0x3578): undefined reference to `PanoramiXSaveXFixesVector' [15:40] weee [15:45] ♥ sdksyms [15:49] well fine, wonder if a hack is enough, if it's a configure option not like that stuff should be exported anyway.. [18:27] bdrung, yeah the warning can be removed from the description. Back when I packaged it, it was pre-release experimental, but not anymore. [18:27] bryce: yep, done [18:27] 0.2.0 is in raring now [18:27] bryce: thanks [18:28] and pushed a further description update to debian-unstable branch [18:29] tjaalton, yeah there's several upstreams it doesn't track. It's on my todo list. [18:29] cool [18:30] weston update nearly there [19:08] bryce: fwiw x1.14 compiles, don't know about pointer barriers yet, seems to have been upstreamed? [19:24] mlankhorst, great [19:25] mlankhorst, yeah I saw peter put in some of those changes for 1.14. Enough for us to drop our patch? [19:25] well it fails to apply now [19:27] maybe raof can illuminate us [19:36] * bryce throws in the towel on #1056039. WFM. [19:44] the upstreamed version of p-b is functionally equivalent to the old one, aiui [19:46] tjaalton, ah good. So perhaps just a matter of testing. [19:46] but it was a rh/gnome guy who made it happen there [19:46] yeah, needs testing to be sure [19:47] maybe unity/compiz needs to adjust [19:47] no idea [19:50] sounds like a job for the TODO page [19:51] btw, I think that for bugs just assigning them to people/the team should be sufficient :) [19:51] tjaalton, how do you mean? [19:53] bryce: meaning that maintaining a list on a wikipage takes effort :) [19:53] I so how would I upload xserver-xorg, without it immediately getting pushed to raring directly? [19:53] mlankhorst: why? [19:54] was curious [19:54] there is no -proposed for -proposed ;) [19:55] tjaalton, certainly does, but I'm fine to do that, it's helping me keep track. Some people have a *lot* of assigned bugs, so just looking at the raw list would have a lot of noise. [19:56] bryce: right, I don't mind you looking after them :) [19:56] and it's true that trying to find a decent bug to work on from a list of 500 is "hard" [19:57] without getting burned out [19:57] tjaalton: the wheezy rc bug list is only 300! [19:57] * jcristau hides [19:57] jcristau: hehe :) [20:13] tjaalton, maybe some day I'll write a web tool that pulls assignments from launchpad. [20:13] bryce: yeah that might be handy [22:28] bryce, tjaalton: The upstream pointer barrier code is now sufficient, but the API's changed; unity will need to be ported. [22:54] RAOF, happen to know if there is a bug # or WI for that work? [22:54] bryce: Let me check my filed bugs; if I haven't filed it, I don't think there's one. [22:55] I had a feeling it would be something like that [22:57] bryce: No; it looks like I haven't filed that. [23:00] RAOF, ok, subscribe me to it once it's filed if you don't mind [23:58] bryce: Done.