[00:12] ROFL at bug 861710 [00:12] Launchpad bug 861710 in unity "[regression] smspillaz fails to sleep at proper hours, despite solar status" [Critical,Confirmed] https://launchpad.net/bugs/861710 [00:35] lol...nice bug [00:37] heh [00:48] Yeah I thought so too. [00:55] jasoncwarner_, you here? [00:58] *anyone* from desktop team? [00:58] Yes? [00:59] I have a patch for compiz I think [00:59] to fix this [00:59] https://bugs.launchpad.net/ubuntu/+source/unity/+bug/861793 [00:59] Ubuntu bug 861793 in unity "compiz crashed with SIGSEGV in PrivateWindow::addWindowStackChanges()" [Critical,Confirmed] [00:59] what do I do? [01:02] You can attach that patch to the bug. [01:02] DBO, bzr clone lp:compiz ; add your patch and a changelog entry ; upload a branch ; propose for merging [01:02] And ping didrocks when he comes online, I think. [01:02] yeah or just tack it to the bug :-) [01:02] He'll be awake and in a position to upload before final freeze. [01:03] okay [01:03] testing the fix now [01:06] c === Shadow1 is now known as SoulShadow [01:34] DBO: yeah, back...keep losing my internet today. effing telstra [01:41] bryceh cyphermox RAOF anyone able to lend DBO a hand? being time sensitive, it owuld be great if we can get him some package help before he goes to sleep [01:41] jasoncwarner_, yeah we gave him some pointers [01:42] yeah [01:42] I do apt-get source compiz-core [01:42] then I go into the source [01:42] make my changes [01:42] apt-get build-dep compiz-core [01:42] and then I run [01:42] jasoncwarner_, jason smith has commented on the bugs [01:42] dpkg-buildpackage -rfakeroot -us -uc [01:42] thats me [01:42] but it always fails at: Linking CXX shared library ../../lib/libkde.so [01:43] DBO, oh there you are. odd username [01:43] bryceh, this is my username from the days of yore [01:43] I only switched to Jason on Canonical IRC [01:43] becuase it was available [01:44] hmm, I don't usually use dpkg-buildpackage but that seems like it ought to work [01:44] DBO, try instead debuild -uc -us [01:44] k [01:44] note you may want to make a second copy of your tree [01:44] since that'll do all the building locally in-tree [01:45] its a linking issue of some sort... [01:45] seems odd it would link against libkde.so [01:45] it makes a lib called libkde.so [01:46] ah [01:46] luckily that won't cause any confusion [01:47] why use debuild [01:47] just curious [01:47] actually for my workflow I generally only do debuild -S, and then do the actual package building using pbuilder [01:48] dunno if that's the best method but that's what I was taught [01:48] bryceh, http://paste2.org/p/1678077 [01:48] multiarch issue? [01:49] I have no idea [01:49] it sort of looks like it's failing when trying to link against 32-bit libs, when it should be using the 64-bit ones [01:49] RAOF might have more clue than I here [01:49] you mean the glib thing? [01:50] that happens every time it tries to link [01:50] it happily builds 100 other things against glib :/ [01:50] hmm, well it looks like in this case you're linking with stuff like /usr/lib/x86_64-linux-gnu/libQtCore.so which I'm guessing isn't compatible with 32-bit stuff [01:51] why would I be linking against 32 bit stuff at all? [01:51] no idea [01:51] yeah [01:51] I dont think thats the case :/ [01:52] well, more correct question is, it's not linking against that, it's skipping it, but then it's returning as if it's failed [01:52] so, why isn't it finding and linking against /lib/libglib-2.0.so [01:52] does that file exist on your system? [01:52] the skipping isn't the error [01:53] ok, then what's the error? [01:53] Huh. Yeah, there's no error in that log. [01:54] http://paste2.org/p/1678080 [01:54] it doesn't report an error [01:54] you know, it's linking in a whole mess of kde gunk [01:55] DBO: Can you throw your source tree anywhere? [01:55] like /usr/lib/libkdeui.so.5.7.0 /usr/lib/libkdecore.so.5.7.0, etc. [01:55] DBO: Alternatively, just throw your patch somewhere? :) [01:55] We'll happily build packages for you. [01:57] RAOF, http://paste2.org/p/1678088 [01:57] I wish I knew why my packages wont build... [02:08] RAOF, so uh, you building me debs? [02:08] Yes. [02:08] Just fighting with paste2 a little :) [02:08] You want amd64 or i386? [02:09] amd64 [02:16] RAOF, how much longer on those packages buddy :) [02:16] Less time if compiz hadn't been rewritten in C++ :P [02:16] 35% done. [02:17] (So, not long) [02:26] DBO: Debs now in people.canonical.com/~chrishr Sorry for the delay! [02:26] RAOF, gratzi [02:27] how did you build that btw [02:27] debuild -S -us -uc; sbuild -d oneiric-tmpfs -j5 [02:40] yay the fix works [02:42] Woot! [02:45] DBO: How safe is that fix? [02:45] RAOF, safe as hells [02:45] but hold up on pushing that further just a moment [02:45] I'll upload it now? oh, ok. I won't upload it now then :) [02:47] jasoncwarner_, you still around? [02:48] jasoncwarner_, ok, will email instead [03:03] RAOF, https://code.launchpad.net/~compiz-team/compiz-core/compiz-core.fix_861793/+merge/77431 [03:03] we are going with that delta I am pretty sure [03:05] DBO: Ok. I'll head off for lunch, and upload a new compiz with that applied once I get back, unless (a) someone picks it up first, or (b) I hear otherwise from you. [03:21] http://people.canonical.com/~bryce/ExternalMonitor/results.html [03:23] bryceh, is there a TL;DR for that? [03:23] DBO, "external monitor configuration sucks" [03:24] bryceh, just FYI [03:24] in this next unity release [03:24] unity will ALWAYS go to the left most monitor [03:24] if there are 2 left most minitors [03:25] it will go to the lowest one [03:25] bryceh, are any of those issues you are reporting our problem really? [03:26] other than the launcher being afraid of the monitor identifiers [03:26] DBO, hard to say what exactly is at fault for some of these problems [03:27] some are the config tool, some unity or compiz [03:27] can you maybe try to identify cases where we (compiz/unity) are likely the problem? [03:27] DBO, that first one where you get double indicator menus feels unity-ish [03:28] DBO, how do you mean? [03:28] that one is fucking impressive [03:28] I have no idea whats happening there [03:28] from the looks of it [03:28] it's quite reproducible [03:28] bryceh, you know what would *really* help this [03:29] take a physical photo each time [03:29] but also a screenshot [03:29] so we can see the framebuffer layout [03:29] ok, can do [03:29] it looks like those monitors are overlapped [03:29] somehow [03:29] so when the smaller monitor paints, it clobbers the top monitors paint [03:31] I was surprised I never repro'd that bug using the gui tool [03:34] bryceh, are you going to the pre-uds thing? [03:34] no, I'm not [03:36] bryceh, when do you arrive at UDS? [03:37] late sunday I think [03:45] * bryceh -> EOD + vacation. cya next week. [04:18] Good morning [04:18] dobey: just by pointing at a core file, no; interesting idea, though; you'd have to make a couple of assumptions for this, like "the package did not change underneath" [04:19] kirkland: pong [04:24] jbicha: retrying gnome-shell powerpc build, mutter wasn't built at that time yet [04:27] pitti: thanks [04:34] cyphermox: oh, so you tracked down the libnl issue? congrats! [04:52] I got past the KDE linking thing [04:52] but now everytime I try to package its looking for a desktop file [04:52] thats frankly not there [04:52] (in fact we explicitly patch out) [05:24] pitti: could you sync anjuta-extras 3.2 from sid? [05:26] jbicha: done [05:40] good morning [05:41] didrocks: it's all your phault! [05:41] * pitti hugs didrocks [05:41] pitti: snif, what happened? :p [05:41] didrocks: now both DBO and I have an extra knot in our brain [05:41] :) [05:42] * didrocks doesn't understand, I'm either out of context or not awake :) [05:43] didrocks: sorry, talking in other channel [05:43] didrocks: so, we were wondering about /usr/share/applications/compiz.desktop [05:43] didrocks: compiz-core installs that [05:43] didrocks: but debian/patches/091_no_use_gnome_but_desktop_file.patch actually seems to make an effort of patching the upstream build system to not install it [05:44] hum, I was thinking that this one is deactivated now [05:44] * didrocks checks [05:44] didrocks: but then again, debian/compiz-core.install does install it [05:44] didrocks: the package built fine on the builders, but fails for DBO as dh_install doesn't find the file [05:44] didrocks: so it seems the patch worked on DBO's box, but failed somehow on teh builders [05:44] didrocks: so, some questions: [05:45] - do we need the file or not? (It has NoDisplay=true) [05:45] - is it meant to be built (then the patch probably souhld be dropped), or not? (then it should be dropped from *.install) [05:45] we need the file otherwise gnome-session can't start the wm [05:45] - can I have some hot chocolate? [05:46] - this patch was there because compiz cmake insane build system was failing at some point with the GNOME integration [05:46] I didn't receive any FTBFS? :/ [05:46] and it's built? [05:47] pitti | didrocks: the package built fine on the builders, but fails for DBO as dh_install doesn't find the file [05:47] didrocks: welcome to our confusion :) [05:47] didrocks: I was suspecting a build system race condition [05:48] It also builds fine here. [05:48] didrocks: I have no idea about cmake, so is it possible that the package already ships pre-made Makefiles which sometimes are rebuilt and sometimes aren't? [05:48] RAOF, I think because you used a different build method [05:48] he needs [05:48] gnome-window-settings-2.0 [05:48] gnome-desktop-2.0 [05:48] so that USE_GNOME is set to true [05:48] and then, to go to the patch path [05:48] similar to patches in the past which patched Makefile.in and Makefile.am at the same time, and depending on whether or not the file system supported sub-second resolution the build would fail because of missing automake or not? [05:48] he also need [05:49] xrender>=0.8.4 [05:49] gtk+-2.0>=2.18.0 [05:49] libwnck-1.0 [05:49] pangocairo [05:49] didrocks: but dpkg-buildpackage complains if there's a missing build dep, no? [05:49] pitti: nothing with that kind of thing with cmake [05:49] yeah, normally, it should [05:50] and debian/rules set the needed bits [05:50] pitti: anyway, so to sum up: [05:50] 1. this patch is because compiz cmake is totally broken, in different way [05:50] didrocks, shouldn't apt-get build-dep give me all those things [05:50] 2. the file is needed [05:50] DBO: yeah, it should [05:51] because I have them all [05:51] I still have to remove the patch to build [05:52] DBO: removing the patch fails on buildds [05:52] Yeah, the compiz build seems quite fragile; it's failed while likning libkde.so here (outside a build chroot) [05:53] yeah you have to uninstall ia32-libs to make it work [05:53] That actually fails the build? What a crock. [05:53] yep [05:53] its pretty cool [05:53] Extoling the virtues of Cmake... Um, what did I just say? Excuse me. :) [05:54] wow, you still have ia32-libs installed? Out of interest, what for? [05:54] cmake: we hated autotools, so we did something *even worse* [05:54] (i. e. should we make an effort to multiarch-ify a particular library to stop the need for it) [05:54] pitti, skype [05:54] you cant install skype without it [05:54] DBO: doesn't need ia32-libs any more [05:54] how? [05:54] You shouldn't need it for skype - install skype:i386 [05:54] DBO: that's what multiarch is for! [05:54] DBO: I thought flashplugin and skype are pretty much the only two things people ever need ia32-libs for [05:55] pitti: wine still needs ia32-libs [05:55] RAOF: ah, ok; three then (I don't have wine installed) [05:55] coool [05:55] unfortunately the amd64 skype package still exists, and hasn't been updated yet [05:55] it shoudl be removed, or made so that it magically installs skype:amd64 [05:55] erm, skype:i386 [05:56] but that'd require another package name [05:56] And libgl1-mesa-dri:i386 isn't properly installable due to libpciaccess0 not being multiarched in Ubuntu. :/ [05:56] DBO: pitti: ah got it! [05:56] HEUREKA! [05:56] so building with USE_GNOME had been broken for 3 months now [05:56] so… [05:56] what I do [05:56] is to remove the GNOME build-dep [05:56] then, I added this patch initially to still ship the needed .desktop file with ELSE(HAS_GNOME) [05:57] and have the desktop file installed if you are or not in the HAS_GNOME part [05:57] this was committed upstream differently [05:57] with repeating the same chunk of code in else(HAS_GNOME) [05:57] fun [05:58] so, basically, we don't build with HAS_GNOME in the package [05:58] (which rely on build-dep, I guess you have a build-dep instaled that we don't have in the package) [05:58] now, on the patch refresh, let me check [05:59] so dbarth refreshed the patch in 1:0.9.4+bzr20110606-0ubuntu6 [05:59] I guess he didn't see that the upstream patch was partially committed and make this patch useless [05:59] but anyway, don't build with HAS_GNOME, it's bringing you issues :) [05:59] *phew* here is the story :-) [06:02] didrocks: thanks for the heads-up; I nearly poked my eyes out this morning :) [06:04] pitti: heh, no worry, here was the initial patch FYI (http://paste.ubuntu.com/698913/) [06:04] pitti: made more sense, didn't repeat a chunk of code but seems upstream wants it differently… [06:09] So, now that we've got that little nugget out of the way, is anyone handling adding the patch from https://code.launchpad.net/~compiz-team/compiz-core/compiz-core.fix_861793/+merge/77431 , or shall I? [06:16] RAOF: no please, don't add smspillaz's patch without a release [06:16] RAOF: we had issues on the past, we only accept tarballs for compiz from dx [06:19] didrocks, hey, do you know about this? "update-alternatives: error: alternative path /etc/lightdm/lightdm-gtk-greeter-ubuntu.conf doesn't exist" [06:20] robert_ancell: hum, interesting, is it when reinstalling the same version of lightdm-gtk-greeter? [06:21] didrocks, from no version. I was wondering if I had some bad files around, so I removed it completely and tried to reinstall [06:21] robert_ancell: let me try to purge and install [06:22] didrocks, I just tried a purge then and it worked. So there must have been something left over from a previous version [06:22] robert_ancell: I'll try installing twice the same version, purging and not purging, 3 min! [06:22] false alarm :) [06:23] robert_ancell: oh, you found it? [06:23] no, I can't reproduce since I did the purge [06:23] hum, better to check if it's really a local issue that can be handled in postinst [06:23] robert_ancell: is it a bug or a feature that when I select a different session in lightdm to login to but enter the wrong password, my session choice is reverted [06:24] I did a double install then uninstall and still no problem [06:24] robert_ancell: yeah, the three cases worked, I guess you had some local cruft [06:24] jbicha, no, it should stay the same [06:25] jbicha, unity-greeter? [06:25] robert_ancell: yes [06:25] jbicha: thanks a lot for your documentation UIFe ack and for your tremendous patience :) [06:25] robert_ancell: do you want a bug report for it then? [06:25] jbicha, yes please [06:25] brb... [06:27] didrocks: well hopefully next cycle will be better with UI freeze [06:27] jbicha: I think we need to discuss the process indeed [06:27] we shouldn't have any big transitions unless we do something crazy like port to mutter [06:52] good morning [06:52] bonjour jibel [06:53] guten Morgen pitti [06:53] mm, Pleasant Pheasant sounds good [06:55] it feels a bit late this year for the next name announcement [06:55] oh, is it out yet? [06:56] yes, traditionally it comes out after beta [06:56] no, I've been spending the past few weeks making up names though [06:57] * pitti votes for perky penguin [06:57] how many "P" changes will we still have? [06:58] it can't be penguin, that's too obvious! [07:00] pitti: do you need packages to show up in syncpackage first to sync them? [07:01] jbicha: I need them to be on the UK debian mirror [07:01] we can sync from unapproved, but it's quite a lot of overhead [07:02] ok, then how about fonts-cantarell? [07:02] jbicha: new package, needs FFE [07:02] ok, I'll do that [07:06] pitti, if you're wondering the lightdm upload I did fixes bug 861398 [07:06] Launchpad bug 861398 in lightdm "the unity-greeter g-s-d is not stopped when login in since 1.0" [High,Confirmed] https://launchpad.net/bugs/861398 [07:06] oh, nice [07:07] robert_ancell: accepted, please close [07:15] bug 862086 [07:15] Launchpad bug 862086 in ubuntu "[FFe] Please sync fonts-cantarell from Debian sid" [Wishlist,New] https://launchpad.net/bugs/862086 [07:15] hey didrocks care to update me on compiz? dbo said he fixed the latest crasher.... [07:15] its being pushed [07:16] god help us all [07:16] jasoncwarner_: I'm just restarting on it [07:16] bring'em on! thanks DBO [07:16] didrocks: awesome... [07:16] DBO still up? [07:16] jasoncwarner_, its final freeze [07:16] where the hell else am I going to be? [07:16] DBO: thanks for jumping on that right away... [07:17] jasoncwarner_: seems a lot of dx commits backport and work the last couple of days: pushed 3 Qt, 2 compiz, 2 libunity, 1 nautilus… :-) [07:17] DBO: wait, today is final freeze? I thought we had like 3 more months ? ;) [07:17] me too [07:17] they sprung it on me [07:17] like a trap [07:17] like a spider monkey? [07:17] who knew today was going to be Sept 29th [07:17] TOTALLY unpredictable [07:17] did not imagine this would happen at all really.. [07:18] the mayans knew [07:18] DBO: oh, that's because you couldn't mentally prepare yourself as you couldn't click before last week on the calendar to change month! :) [07:18] didrocks: ouch! [07:18] :) [07:18] bazinga [07:18] BAHZING [07:19] didrocks, with the headshot [07:19] ETOOMANYITCROWDREFERENCE [07:19] oupss ETOOMANYBIGBANGTHEORYREFERENCE [07:19] rather :) [07:19] ok, I can still make bad jockes, it's a proof that compiz is working :-) [07:19] even after a restart [07:19] didrocks: nice! [07:20] ok, let's run it for 10 minutes and then push the crack out of it! [07:20] didrocks: you want us to be testing from a PPA or something? [07:21] jasoncwarner_: yeah, that would be nice, one sec, pushing [07:21] jbicha: at it [07:22] pitti: did the "system proxy settings != session proxy settings" warning get lost when the code moved from the old control center to the new one? [07:22] good morning btw :) [07:22] mvo: hey [07:22] mvo: even worse, seems the latest release drops the entire proxy thing [07:22] we have a bug, assigned to rodrigo [07:22] mvo: but possible, yes [07:22] ok, thanks [07:23] didrocks, so uh, if we decide to do a plugins release [07:23] what does that mean to you [07:23] DBO: jasoncwarner_: compiz should magically appear in the ubuntu-desktop ppa shortly for you, yours sincerly, didrocks [07:23] my dearest didrocks [07:23] didrocks: <3 you [07:23] im already running it [07:23] love [07:23] DBO [07:23] DBO: that I will of course hate you more and more, but that's fine, I'll get a revenge in just a month, mwahahahahaha [07:23] DBO: seriously, do you think there are needed fixes? [07:23] needed? not needed no... [07:24] certainly some nice to have shit though [07:24] * Sweetshark wonders at all the high emotions (must be postbeta anexiety) ... [07:24] DBO: do you have the bug # so that we can asses risk/gain? [07:24] uhm [07:24] sure let me look... [07:25] DBO: if it can wait for a SRU, that have more testing (one week before pushing to users) [07:26] * didrocks add his waiting list today, a libunity, unity-2d, nux, unity, nautilus, unity-place-files and counting! (already done qt, compiz…) [07:26] didrocks, https://bugs.launchpad.net/unity/+bug/827560 [07:26] Ubuntu bug 827560 in compiz "unmaximizable windows still show orange glow but fail to maximize" [Low,Triaged] [07:26] didrocks, https://bugs.launchpad.net/unity/+bug/827560 [07:26] there is another one but I suffer from too much brain dumb right now to find it [07:27] DBO: the bot is smart, it avoids the repetition :) [07:27] oops [07:27] https://bugs.launchpad.net/compiz-grid-plugin/+bug/796594 [07:27] Ubuntu bug 796594 in compiz-grid-plugin "Window behaviour - pressing the 'restore' window indicator on a semi-maximised window should return it to the restored state" [High,In progress] [07:27] outsmarted by a bot [07:27] excellent way to start my day [07:27] jbicha: there, shepherded through binNEW as well, enjoy :) [07:27] DBO: "start", really? :) [07:28] well it is 3:30am [07:28] pitti: thank you! [07:28] its closer to starting than ending at this point [07:28] DBO: ok, I guess it's easily testable for regression [07:29] DBO: do you know to make dist in the oneiric branch or is it better to cherry-pick the patches there? [07:29] DBO: urgh, the branch seems to not have previous fixes [07:29] of? [07:30] DBO: I have a tarball from yesterday in c-p-m [07:30] i dont know what to do about that [07:30] DBO: and I can ensure you I have more fixes than what is in https://code.launchpad.net/~compiz-team/compiz-grid-plugin/oneiric [07:31] yeah I know [07:31] DBO: I hope the tarball from compiz have all fixes from yesterday as well [07:31] dont do c-p-m [07:31] yeah, seems hairy [07:31] I hope so too [07:31] DBO: let's check for compiz before pushing it [07:31] do you have any of the bugs so I could check? [07:31] DBO: you took: https://code.launchpad.net/~compiz-team/compiz-core/oneiric [07:31] yeah, one sec [07:31] bryceh, RAOF, lightdm has been using -background none for some time so we can probably drop the -nr option [07:32] patch rather [07:32] DBO: http://paste.ubuntu.com/698947/ here is the full list of fixes [07:32] robert_ancell: Know about the other ?DMs? [07:33] didrocks, I can immediately confirm several of those fixes are present [07:33] RAOF, no, but I suspect they would have been carrying matching patches? [07:33] DBO: for instance, on this one: https://bugs.launchpad.net/compiz-core/+bug/860397 I don't see a merge in ~compiz-team/compiz-core/oneiric [07:33] Ubuntu bug 860397 in unity "Windows which are marked transients of docks should be treated like docks" [High,Fix committed] [07:33] didrocks, yes that one is present [07:33] onboard still works iwth unity [07:33] huzzah [07:33] great :-) [07:33] *phew* [07:33] RAOF, robert_ancell: our gdm has -background [07:34] thanks for confirming DBO :) [07:34] DBO: let's forget about c-p-m now, not going into trouble and plan those for a SRU [07:34] sure thing [07:34] yep [07:34] * RAOF checks kdm and xdm. [07:34] RAOF: checking kdm.. [07:34]