[00:00] yay, working now :) [00:21] ok origin/ubuntu builds fine now on i386, other arches are still waiting for the libxfont rebuilds [00:58] mind if I make mesa stop building mach64 in git? theres no point to it, needs external drm modules that don't even exist in libdrm git anymore [00:59] just wasting livecd space :D [01:02] sure [01:19] RAOF: back from UDS yet? xserver-xorg-video-nouveau needs some lovin pretty badly :) [01:27] he's probably recovering from jetlag [01:29] bryceh: did you get all my messages about UDS? it looked like you didn't get the first on monday since you asked a few days later and i'm not sure you got the second since you signed off after [01:44] Sarvatt, no my irc has been real spotty this past week [01:44] anyone have a description of what exactly libkms is so I can add it to the new package description? :) [01:46] has something to do with vmwgfx [01:49] kms abstraction library or such [01:54] http://www.pagetable.com/?p=324 [01:55] ^ new logo for -intel? [02:41] fitting with how broken it's been this week :) [02:52] going to attempt to fix this accidental epoch in xorg-edgers libdrm.. [02:52] things may be broken for a bit :) [02:53] i'll have to post everywhere telling people to downgrade too, fun stuff [02:54] worried about leaving it since it's screwing with the symbol versioning, maybe I should just leave it.. [03:55] phew looks like xserver-xorg-video-intel is the only thing screwed up by the bumped epoch libdrm, i rebuilt everything i uploaded since the 12th though just incase [03:56] scrapped the libkms idea for now, it requires libdrm to build... lol [04:25] weird, libgl1-mesa-dri-gallium pulls in oprofile [04:26] libgl1-mesa-dri-gallium libllvm2.7 libopagent1 llvm llvm-dev oprofile [04:27] ah llvm-dev depends on it.. [04:53] xorg-server from origin/ubuntu currently brings in nvidia-current... :D [04:55] The following NEW packages will be installed: [04:55] dkms nvidia-current nvidia-settings xserver-xorg-video-radeonhd [04:55] and xserver-xorg-core held is back [04:55] is held back rather === Bernardo|away is now known as Bernardo [05:25] good morning === Bernardo is now known as Bernardo|away [08:46] Sarvatt: how's that possible? [13:23] hi, i am using an old tablet (compac tc1000) with 10.04 installed. [13:23] now i want to get the touchscreen working: [13:24] installed fpit. i works but i cannot configure it. [13:25] does anyone know what values for maxx and maxy needed to be set in xorg.conf? [15:25] oc13: man fpit should give you some default values to play with, since you said it installed and worked i'm guessing you're on karmic or older? [15:26] i uploaded an fpit that works with lucid to the x-updates ppa since the one in lucid/debian doesn't work with xserver 1.7, but some people are having configuration problems with it still [15:38] wow thats a change, I have to install gallium dri to use GL with xserver master, mesa classic i915 just segfaults all over the place [15:40] http://pastebin.com/kNN5MVL0 [16:49] thats odd, SIS doesn't compile against xserver master here because it needs mibank.h thats been removed, but it compiles fine on tinderbox? [16:51] tinderbox may have old headers still around [16:57] had to disable udeb building for xserver master until i have time to look into it, fails like this http://launchpadlibrarian.net/48590426/buildlog_ubuntu-maverick-i386.xorg-server_2:1.8.99.0%2Bgit20100517.345eb171-0ubuntu0sarvatt_FAILEDTOBUILD.txt.gz [17:02] i keep getting into weird situations I can't figure out with the new xserver abi stuff since 1.7.6.901, the breaks make it odd. like tslib needed to be updated in the archive before i could even install xserver-xorg-core even though i didnt have it installed [17:03] we have an older version in the archive than the breaks: in xserver [17:05] Sarvatt: you only get that in the udeb build? [17:06] yeah [17:06] disabled udeb works fine [17:06] builds fine rather [17:22] Sarvatt: http://people.freedesktop.org/~jcristau/0001-Fix-build-without-XACE.patch [17:26] oh nice! will add that and rebuild in a minute when I test this SIS change out to be sure it works. thanks for that jcristau, sorry to bug you with problems all the time! I'm not pointing them out asking for them to be fixed, just trying to get the issues I run across out there :) [17:27] yeah. i should work instead. but instead i jump on the chance to not work ;) [17:28] anyway, sent this one to the list [17:28] should I just drop the mibank.h header completely or conditionalize it for #if XORG_VERSION_CURRENT >= XORG_VERSION_NUMERIC(1,8,99,0,0) ya think? [17:28] it builds fine without the header against 1.8.99.0 at least [17:28] oh whoops I meant <= [17:29] i couldn't find anything in the driver that needed that header [17:29] so i'd say drop it [17:49] http://pastebin.com/RNM0ZPqC works, phew.. logs sure are ugly these days though [17:50] * Sarvatt needs to get a MTA set up for git one of these days [17:51] building the udeb with your patch now [17:52] ugly because of the timestamping you mean? [17:54] yeah [17:54] breaks a lot of the driver messages [17:59] xserver with your patch is building here btw if you're interested - https://edge.launchpad.net/~sarvatt/+archive/xorg-testing/+build/1742884 [18:00] i'll reply on xorg-devel with a tested-by and the details if it works [18:01] cool :) [18:02] 2 concurrent configure runs makes understanding what's happening quite hard.. [18:03] yeah, for sure [18:04] it builds quite a deal faster than 1.7.x though even with the extra udeb build since theres no tests [18:05] oh you disabled the tests? [18:05] i guess it's not very useful to run them for the ppa [18:05] ../../dix/gc.c: In function 'ChangeGC': [18:05] ../../dix/gc.c:152: error: dereferencing pointer to incomplete type [18:05] ../../dix/gc.c: In function 'ChangeGCXIDs': [18:05] ../../dix/gc.c:439: error: dereferencing pointer to incomplete type [18:06] same deal i guess.. [18:06] only in udeb build [18:06] oh the tests still run? i thought they didn't since it was so fast, guess i didnt look hard enough [18:06] didnt explicitly disable them [18:07] the main build has xace enabled [18:17] whoops, sorry to spam your mail there, replied to just you there and resent it :) [18:21] ok now i've actually tested the build... [18:43] RAOF - could you ping me when you're available? i would like to discuss bug 546578 with you at some point [18:43] Launchpad bug 546578 in xserver-xorg-driver-ati "black screen after a few user switches" [Unknown,Confirmed] https://launchpad.net/bugs/546578 [18:44] thats related to the fade problem isn't it? :( [18:45] Sarvatt, yes. i don't know if there are multiple issues or not, but there's definately a race in lucid thats contributing to some peoples issues [18:45] if the VT switches before the fade is finished, then the gamma doesn't get reset [18:45] its happening on other distros too and not limited to ati [18:46] so, i was going to disable fade-on-lock in gnome-screensaver for now (it will still fade on idle though) [18:46] i've talked to people on arch and gentoo using intel nouveau and ati having the same problem and disabling fade fixed it [18:46] but i wanted to discuss with RAOF first. i'm not sure if you have an opinion too though [18:46] yeah, i'm leaning towards switching off fading for now [18:47] it's a 1 liner in gnome-screensaver [18:47] have you checked other distros patches by any chance? [18:47] not yet [18:47] but i imagine they will probably be pretty similar [18:48] i see one reporter already posted a patch to one of the bug reports, but i'm not going to do it like that, as that globally disables fade [18:48] i still want fading to work on idle ;) [18:50] Sarvatt - btw, the nvidia binary driver doesn't seem to exhibit the issue. i can still use xgamma to control the gamma on X servers that aren't on the active VT [18:50] although that probably doesn't help much [18:51] binary drvier uses vidmode gamma fading though not xrandr ramps [18:51] Sarvatt - yeah, that's also what xgamma is doing though [18:51] both methods fail in lucid [18:55] chrisccoulson: do you think this is related? I dont understand the patch completely - http://git.debian.org/?p=pkg-xorg/xserver/xorg-server.git;a=blob;f=debian/patches/121_only_switch_vt_when_active.diff;h=e35467bf84a2f8327bc907246c92b26b42da19a6;hb=refs/heads/ubuntu [18:56] i can't find any context on why we have that patch, was added *years* ago [18:57] Sarvatt, i'm not too sure really. i don't understand the patch completely either [19:04] making it bug me even more, arch carries that patch too [19:07] Sarvatt, hi, what do you think about copying the kernel to edgers? https://edge.launchpad.net/~ricotz/+archive/unstable/+sourcepub/1137152/+listing-archive-extra [19:08] i kind of would prefer to wait a few hours and just use the ones straight from maverick if we're not changing any options but go ahead [19:09] ricotz: btw did you see the notice at the top of the xorg-edgers ppa page? i screwed up and added an epoch to libdrm for a few days and people have to manually downgrade back or it wont update :( [19:10] keeping the epoch would suck more than that pain IMO so i got rid of it, it screwed up the symbols and stuff [19:10] ia32-libs might be screwed up by that :( [19:10] not sure if the update script checks for a newer version and grabs that since the newer one will be older [19:10] oh, didnt noticed is, let me look [19:12] i can upload a new ia32-libs source package, the bandwidth isnt an issue here (got 20Mbps up and no quota) [19:12] not sure how though, need to look at it [19:12] Sarvatt, i have updated the ia32-libs just now [19:12] should be fine [19:12] thanks ricotz, sorry about all the trouble [19:13] i'm working on xserver master for maverick :D [19:13] sounds like trouble ;-) [19:13] but i'm going to hold off on it for at least a month or two, learned my lesson following master too closely in lucid [19:15] people actually expect xorg-edgers to work :) [19:16] chrisccoulson: you know, i should boot up a fedora 13 livecd and see if test-fade has the same problem there [19:16] Sarvatt - could do. remember though that test-fade has the additional issue where the buffer if not flushed properly, and the last SetCrtcGamma command never gets to the server [19:18] theres no way thats related to this problem? [19:18] Sarvatt - i'm fairly sure it's unrelated. people could still recreate it even with the fix for that in [19:18] seems like it could be to me but i dont understand it that well, people saying it happens when you switch users in quick succession [19:18] ah, test-fade is fixed in ubuntu? [19:19] Sarvatt - not yet. i uploaded a test build to my PPA a while ago, but it's not there any more [19:19] didn't know ya fixed that, i haven't looked since i dont even use gnome-screensaver and just reported the bug to condense a bunch of duplicates [19:19] ahh ok [19:49] gnome-screensaver in karmic didn't have xrandr fade right? i know it was upstreamed in november but not sure if it was added in -backports or something [19:49] or a point release later [20:42] Sarvatt, yeah, xrandr fade was only introduced in lucid [20:47] chrisccoulson, RAOF usually comes on in about +2 hrs [20:47] bryceh, thanks [20:48] chrisccoulson, maybe drop him an email in the meanwhile [23:03] sheesh, put off setting up git send-email for so long and it was so easy :D [23:04] ricotz: gconf was updated so your mutter ppa works again \o/ === BUGabundo is now known as BUGabundo_Bones