[00:01] RAOF: pomke [00:01] *poke [00:02] Ow. [00:02] smspillaz: You rang, m'lud? [00:02] RAOF: hi [00:03] RAOF: oh that sounds so much more realistic now that I know what you actually sound like XD [00:03] :P [00:03] RAOF: anyways, I meant to ask - are damage events generally handled in the driver ? [00:03] eg, if you should get a damage event for something and then don't, is that the drivers fault, if you registered for damage at the right time ? [00:04] It could be the driver, it could be EXA. [00:04] Or the way that the driver interacts with EXA :) [00:07] RAOF: what if it's nvidia ? [00:08] Then it's probably the driver. [00:08] I don't *think* there's anything driver-independent that could make that not work. I could wander through the source if you wanted me to be more sure ☺ [00:09] well I'm basically just doing that myself now [00:10] it's just the thing is that this bug only happens with compiz 0.9 and not 0.8 [00:10] nvidia has had problems with damage in the past, IIRC. [00:10] (the one where menus don't appear at first but if you hover over them they work) [00:11] ISTR a patch in compiz for ubuntu which messed with damage though [00:13] RAOF: this one http://bazaar.launchpad.net/~compiz/compiz/ubuntu/revision/483#debian/patches/049-damage-report-non-empty.patch [00:22] Oh, incidentally: I hate you, and your -keypress stealing Unity plugin ☺ [00:22] Back to ++ for Do, then :( === james is now known as Guest98720 [01:16] smspillaz: that was a patch from maniac that went upstream, iirc [01:16] or a different solution went upstream [01:26] Amaranth: well it was added to the workaroudns plugin [01:26] force X to GLX sync [01:26] although the particular damage issue they were targeting was different [01:26] smspillaz: wait, I think we're talking about different things [01:26] oh ? [01:26] the force X to GLX sync was when I put a call to glXWaitX in core [01:26] ah right [01:26] yeah, I'm talking about this one [01:27] onestone implemented that in the workarounds plugin instead [01:27] http://bazaar.launchpad.net/~compiz/compiz/ubuntu/revision/483#debian/patches/049-damage-report-non-empty.patch [01:27] Right, that has nothing to do with the glXWaitX stuff, afaik [01:27] right [01:28] so what *was* that about then ? [01:29] the one I'm referencing didn't go upstream AFAICT [01:30] I dunno, someone dumped our changelog when merging with debian and the debian changelog doesn't say [01:30] Let me dig through old changelogs [01:32] smspillaz: oh, it was for damage issues with nvidia [01:32] I know that [01:32] I wonder why we stopped using it, I don't think the glXWaitX stuff was the same thing actually [01:32] but what were the issues people were getting ? [01:33] and more importantly, why do I only ever get this issue on compiz 0.9 and not 0.8 [01:33] bug 269904 [01:33] Launchpad bug 269904 in compiz-fusion-plugins-main (Debian) (and 9 other projects) "Screen refresh problems with nvidia on intrepid (affects: 64) (dups: 20) (heat: 182)" [Unknown,New] https://launchpad.net/bugs/269904 [01:34] Amaranth: I'll try applying this patch then once I confirm it is an nvidia related issue [01:35] smspillaz: now I'm all kinds of confused [01:35] I don't even remember this patch but apparently you were right [01:37] smspillaz: Well, do you see the bug with that workaround enabled? :) [01:37] I can't check right now [01:37] * smspillaz is recompiling his system [01:37] recompiling? gentoo? [01:37] arch [01:38] smspillaz: blasphemy [01:38] Amaranth: hrm I see it on intel [01:38] I would swear in here right now if I could [01:38] smspillaz: does the workaround fix it there? :) [01:39] Amaranth: erm, let me check [01:40] I don't have workarounds and friends enabled [01:40] available rather [01:40] maybe I might be able to get a hold of krh [01:41] in other news I wish there was a way to automate what yaourt does [01:41] smspillaz: I updated the plugins-extra package, if that's what you mean [01:41] having to consistently hit "y" "y" "y" all the time drives me nutes [01:41] *nuts [01:41] Amaranth: I had to compile from source [01:43] smspillaz: hmm, it seems I added my glXWaitX call in core after the workaround [01:43] Either that or in the bug report I'm referring to the patch you linked to [01:43] * Amaranth can't remember that far back [01:52] smspillaz: tbh the only drawing issue I see that isn't an app bug is xchat-gnome drawing on top of itself [01:52] although I think that happens with metacity too so that's just a generic driver bug, I guess [01:58] yeah === bjf is now known as bjf[afk] [02:18] /c/c [02:33] Amaranth: I just saw the bug on nouveau too [02:34] hmm, while I have nouveau up and running I should try and get unity working [02:35] smspillaz: What part isn't working currently? [02:35] aka: My netbook with nouveau circa mesa 7.9 happily runs Unity on nouveau. [02:42] RAOF: trigger a menu -> it doesn't appear in some cases -> hover over where the menu should be (eg do something to damage it) -> it does appear [02:43] RAOF: the problem is that when the window is mapped, we're supposed to get a magical "I'm damaged, ok feel free to draw me now" event, but in some cases we don't [02:43] and it's happening with nvidia, intel, nouveau [02:44] I don't believe that I've seen that behaviour, but that spread of drivers suggests it's not actually a driver problem :) [02:44] I know [02:44] I'm just wondering how on earth I could be using the xcomposite extension *wrong* here [02:45] it's not a race condition either - I'm registering for damage events on the client right after the CreateNotify [02:46] * smspillaz wonders if there's some magic he has to do [03:02] * RAOF doesn't know of any magic required. [03:09] that's what suprises me === Guest98720 is now known as dashua [03:19] Amaranth: btw, even with force-x-to-glx sync I still get the bug :/ [03:19] smspillaz: weird [03:19] smspillaz: perhaps it requires human sacrifice [03:20] * smspillaz sacrifices DBO [03:20] deciding to do this on arch linux didn't make my life any easier [03:20] they upgraded python to 3.0 and it broke ccsm [03:23] That was really silly. python 3.0 isn't python :) [03:27] silly AND annoying [03:30] Hello [03:31] I have a unique problem, I tried #ubuntu-boot, but no one was there. [03:31] Is anyone available to help? [03:31] Well, my CMOS/BIOS thinks I have a floppy drive. I disagree. [03:32] So, this of course spawns an infinite error when launching Ubuntu [03:32] I need to know the correct argument to pass to the kernel when starting Ubuntu, so it bypasses the floppy drive. [03:32] I tried nofloppy, but it didn't work [03:33] Also tried: floppy=0, floppy=0,16,cmos [03:33] Does anyone know the correct arg? === james is now known as Guest26881 [04:26] does anybody know where the gnome-shell package has gone in natty? [04:26] desrt: bug 690045 [04:26] Launchpad bug 690045 in gnome-shell (Ubuntu) "Please remove and blacklist gnome-shell (affects: 2) (heat: 128)" [Wishlist,Fix released] https://launchpad.net/bugs/690045 [04:27] cheers [04:27] desrt: I hope to have it in the gnome3 PPA by the end of next month [04:27] and hopefully in natty by the release? [04:28] desrt: probably not, since it'll most likely needs parts of the gnome3 stack not in natty, but will be in natty + 1 [04:29] * desrt doesn't look very impressed [04:31] i guess when jono said that there would be a first-class gnome-shell experience in natty he meant "if you run fedora in a virtualbox"? [04:32] * micahg can't speak for jono [04:32] desrt: gnome3 probably isn't landing, but if the pieces are there, I'll try to get it back in [04:32] so what's missing? [04:33] desrt: idk yet, I won't have time until next month to investigate [04:34] desrt: but we certainly won't have gnome3 landed without gnome-shell [04:34] * micahg shouldn't say certainly (since he has no authority), most probably [04:38] desrt: all I can say is I'll do what I can once the pieces are in the archive [04:39] you're going to lose quite a lot of users if it's not working, i think [04:39] desrt: I'm not on the desktop team, just a community member [04:42] From what I gathered, it's roughly a manpower problem. [04:44] RAOF: it always is.... [04:46] Well, some of the dependencies are also a bit annoying; I know gjs and the “this internal xul library is totally appropriate to link against” has been somewhat troublesome. [04:46] ya [04:47] i remember a cycle or two ago you guys put in a request for gnome-shell to switch to seed [04:47] and the response was "not really, opposed but we don't see the point" [04:47] desrt: we can't do stable updates for xulrunner, but hope to for webkit === james is now known as Guest74629 [07:54] RAOF: ping [07:58] RAOF: so two things [07:59] RAOF: the first is that I've compiled the whole stack from git and it seems that the sync fences seem to fix my damage issue [07:59] Do either of them involve porpoises? [07:59] so question: can we backport those ? [07:59] 2) [07:59] we're doing something weird in compiz where we register for damage events on both the window id drawable and the tfp pixmap [08:00] which seems, erm, wrong [08:00] since we only did it on the id in 0.8 [08:00] Aaah, yeah. And prior to the sync extension work there's no way to synchronise X<->GLX easily. [08:00] yeah [08:00] even our glXWaitX workaround doesn't quite work [08:01] although this was an issue where I don't seem to get a damage event on window maps, and I don't know if that's related or now [08:01] You also need XwaitGL or whatever it's called, I think. [08:01] *not [08:01] Well, they're in 1.10; xorg-edgers has that full stack built and I expect to be ready to upload 1.10 to natty proper by the end of the week. [08:01] hrm ok [08:02] I'm going to see if this "smart idea of ours" to register for damage events on both the named pixmap and the id was not so smart after all [08:02] :) [08:03] Also, stop eating :) [08:05] RAOF: not my fault :)\ [08:05] someone wanted the launcher to be shown on super, so we have to xgrabkey for super [08:06] RAOF: do you know what the commit was that introduced sync fences so that I can test without them to see if I still get my bug ? [08:07] I can grab the commit ID for you. [08:08] cheers [08:08] I just wouldn't know what the grep for :p\ [08:09] I don't think that the sync fences are actually hooked up to anything, though. [08:09] They're something that you need to use explicitly, IIUC. [08:12] hrm [08:12] I wonder why I'm not hitting this bug in xorg master [08:12] So, the candidate commits are: 02e18c9fb5, 9c0c7cc9a7ad, 397dfd9f87, and a couple of others leading up to 397dfd9f87 [08:12] ok [08:13] I think there have been some changes to Damage and EXA, but that shouldn't affect nvidia. [08:13] although I'm pretty sure this behaviour where we register for damage events on the same drawable is really a recipe for disaster [08:13] might want to check that one with krh [08:14] That might be a good idea. [08:15] it would be really handy if onestone was actually around during the time we merged his code :/ [08:15] then we could ask him, you know, like, why he did things the way he did .... [08:15] :) [08:25] dbarth: good morning [08:29] desrt: hi Ryan [08:29] do you know if seb is around this week? [08:29] or on holidays? [08:29] desrt: he is [08:29] desrt: he was there yesterday afternoon [08:29] Good morning [08:29] nice [08:29] he should connect soon [08:29] hi pitti [08:29] desrt: heeey [08:29] hi pitti [08:30] bonjour dbarth [08:30] Good morning pitti [08:30] dbarth: just about have a working set of packages... [08:34] RAOF: aaah ok, now I know why he does it [08:34] his solution was to nuke loose binding and only rebind textures when the underlying pixmap was damaged [08:35] Ah, so it's a nvidia performance hack? [08:36] well basically it was the solution to have a "half way between" strict and loose binding [08:36] because loose binding was the nvidia perf hack [08:37] although I wish he'd stop calling it "dynamic binding" and call it "damage binding" like it actually is :/ [08:40] * desrt gives up, goes to bed [08:41] RAOF: or rather, it looks like he added it for multitexture support [08:41] * smspillaz hates it that debugging compiz usually ends up in "guess what the other developer was trying to do" [08:42] Yeah, that game sucks. [08:43] especially when the other developer is some wizard like davidr or onestone but also suffers from "I don't document my code"-diasease [08:51] bonjour dbarth [08:51] oops, focus error [09:02] RAOF: btw, props to whoever figured out how to make plymouth look like not-garbage on nvidia [09:05] smspillaz: it wasn't driver specific [09:05] pitti: oh ? [09:05] smspillaz: it was a weird compiler bug, see bug 685352 [09:05] Launchpad bug 685352 in plymouth (Ubuntu Natty) (and 6 other projects) "libplymouth2_0.8.2-2ubuntu6 and later give ragged splash and text rendering (affects: 34) (dups: 3) (heat: 174)" [Medium,Fix released] https://launchpad.net/bugs/685352 [09:05] was a loong night to track that down :) [09:05] pitti: ahhhh compiler bugs :) [09:06] pitti: we had one where compiz would crash a few seconds after loading any plugin on gcc 4.5. Also took a loooooong night with onestone to track that one down [09:06] pitti: I see the one where VT text was offset to the lower right hand corner of the screen was fixed too. congrats :) [09:09] pitti: ah strange you got a bug with gcc 4.5 as well [09:09] gcc 4.5 is *really* buggy I've noticed [09:13] how to enable wifi on hp [09:13] laptop [09:29] morning [09:29] good morning rodrigo_ [09:32] hi didrocks [09:37] pitti: wow, impressive work on that bug. Nice job! :-) [09:37] thanks [09:37] hey didrocks [09:37] guten morgen pitti, how is the hackfest? [09:38] didrocks: going well indeed! making nice progress [09:38] nice :) [10:05] hey [10:05] hey seb128 [10:08] hey pitti [10:08] how are you? [10:08] salut seb128 [10:08] lut didrocks, ca va ? [10:08] seb128: ça va bien, et toi ? [10:08] ca va ;-) [10:49] pitti, oh, daily iso built [10:49] seb128: I'm great, thanks! beautifying gvariant handling in gi nw [10:49] seb128: yeah, at last! and it's small (I added a new langpack) [10:49] nice! [10:49] pitti, new langpack, great! [10:49] which one? just curious [10:49] we have french back on amd64 right? [10:49] ;-) [10:58] seb128: almost [10:59] seb128: tomorrow morning we shold have fresh langpacks (I need to build them manually) [10:59] then they should fit [11:02] pitti, ok great ;-) === njpatel_ is now known as njpatel [11:08] * didrocks wants a working Internet connexion :/ [11:09] didrocks, still no box? === zyga_ is now known as zyga-vaio [11:31] RAOF: ok, did some more investigating - there's some X bug that's causing us to not get damage events on MapNotify even with compiz 0.8 and it's fix in xserver git [11:31] do you think you could find commits that might have fixed them and backport them? [11:45] at last \o/ [11:51] didrocks: \o/ [11:56] hello everyone! [11:57] hey nessita, how are you? [11:57] didrocks: pretty good! you? [11:57] nessita: I'm fine, thanks :) [11:59] hey nessita, how are you? [11:59] hey pitti! [11:59] how was that return to the home? [12:01] nessita: went pretty well this time, no delays and I felt almost no jetlag [12:01] nessita: how about you? [12:02] the trip was more tiring that I thought... but once I was at home, it was great :-) === oubiwann is now known as oubiwann_ [12:09] hey nessita [12:10] hey seb128! how are you? [12:10] I'm great thanks! [12:10] I managed to sleep 6 hours in the 9 hours flight [12:10] then stayed up until midnigt, slept until 11:30 on monday [12:10] which get me out of jetlag in one day ;-) [12:10] what about you? [12:13] well, I slept pretty much nothing on my way back, I slept most of the sunday in my home. That made me recover pretty well :-) [12:16] ;-) === oubiwann_ is now known as oubiwann [12:39] hey didrocks, I hope you had a good flight back! Quick question: from which package did the "Classic Desktop" string that appears on gdm come from? You told me a couple of weeks ago, but I can't remember which one [12:42] seb128, pitti: can you please have a look at merging/uploading this: https://code.launchpad.net/~rodrigo-moya/ubuntu/natty/libcanberra/no-gtk-quit-add/+merge/46463 [12:42] seb128, pitti: not urgent until the new GTK3 is uploaded (robert is on it) [12:43] hi chrisccoulson, do you think you could merge Danilo's fixes to po2xpi to the trunk branch, as discussed on the rally? -> https://code.launchpad.net/~danilo/po2xpi/firefox4-support/+merge/46309 Thanks [12:44] dpm - yeah, sure [12:44] rodrigo_, will do, did you read robert_ancell's email about gtk? [12:44] sorry, i've not been too well this morning [12:44] seb128, about the building failures? [12:44] dpm, hey, the classic session should come from gnome-session [12:44] rodrigo_, yes [12:44] awesome, thanks chrisccoulson (the merging part, I mean :) - get well soon [12:44] seb128, yes, looking at building his last changes in a minute [12:45] seb128, great, thanks [12:45] seb128, I didn't get those when I built my branch though [12:45] rodrigo_, ok, I will try there as well [12:45] ok [12:47] dpm: hey, yes the flight back was nice. yes, it's coming from gnome-session [12:47] dpm: btw, A new upload create a new string to translate (the fallback message) [12:48] dpm: the ; are important to keep as it's a delimiter between fallback session message [12:49] didrocks, cool, thanks. Yeah, it seems it's there, but I'm not sure you mean by the delimiter. Is it this one we're talking about? -> https://translations.launchpad.net/ubuntu/natty/+source/gnome-session/+pots/gnome-session-2.0/ca/31/+translate [12:51] dpm, in the template it's this string [12:51] "";It seems you do not have the hardware requirement to run unity. Installing " [12:51] "new driver can maybe be a help.\\nYou should try the Ubuntu classic session " [12:51] "to directly get the traditional interface." [12:51] " [12:52] dpm, seems launchpad strips the ";" at the start of the line there [12:53] dpm: seb128 I just added the ; with an upload less than an hour ago, so maybe not refreshed yet [12:53] oh ok [12:55] seb128, didrocks, yeah, it seems that the template with the new string is not there yet: https://translations.launchpad.net/ubuntu/natty/+source/gnome-session/+imports?field.filter_status=all&field.filter_extension=pot [12:56] didrocks, has the new string got a translator comment re: the ";"? Because I think it is likely to confuse translators, and I'm afraid they might remove it in the translation [12:57] dpm: I wasn't sure how it worked in the .desktop file so I didn't put one. I can do that in a next upload if we are sure intltool isn't confused by that [12:58] didrocks, I don't think intltool supports comments in .desktop files - dobey, any ideas? [13:00] the \\n escaping does not seem to work, either on https://translations.launchpad.net/ubuntu/natty/+source/gnome-session/+pots/gnome-session-2.0/ca/+translate?batch=10&show=all&search=hardware [13:01] let me file a bug and see if any translations people comes up with an idea [13:02] ok, thanks dpm :) [13:18] rodrigo_, do you know why when g-s-d crashes the gtk theme reset rather than staying on what it was? [13:18] rodrigo_, is the xsettings dropped when g-s-d crashes? [13:18] seb128, hmm, no, it shouldn't [13:18] but it's true that it happens, I see it when logging out [13:19] seb128, 2.32. or 2.91? [13:20] rodrigo_: speaking of crashes, would you know why u1client is failing to build with: [13:20] dh_girepository -pgir1.0-ubuntuone-1.0 [13:20] dh_girepository: Could not find gir file for UbuntuOne-1.0.typelib [13:20] make: *** [binary-predeb/gir1.0-ubuntuone-1.0] Error 2 [13:20] nessita, you need the gir*ubuntuone package installed [13:20] nessita, is that on natty? [13:20] nessita, the gir version is 1.2 now [13:20] ah, right [13:20] rodrigo_: but that is not a local machine, that is u1client nightlies [13:20] nessita, where is your work in progress? do you want me to check? [13:21] seb128: I have no idea where this is built... rodrigo_, do you know? [13:21] nessita, dobey's machine? [13:21] nessita, where is the source I mean [13:21] seb128: log file is: http://launchpadlibrarian.net/62246782/buildlog_ubuntu-maverick-amd64.libubuntuone_0.3.9%2Br113~maverick1_FAILEDTOBUILD.txt.gz [13:21] or the vcs [13:22] nessita, I guess it needs to build depend on gir1.2-ubuntuone-1.0 [13:22] oh, it's a maverick build [13:22] rodrigo_, ^ [13:23] does it build depend on gir1.0-ubuntuone-1.0 then? [13:23] * rodrigo_ gets the source === oubiwann is now known as oubiwann_ [13:24] rodrigo_, it can't, the binary is coming from this source [13:24] so, is a 'which binary is built first' issue? [13:25] the binary is from libubuntuone, not u1-client [13:25] rodrigo_, the build log is a libubuntuone one [13:25] ah [13:26] hmm [13:28] nessita, can I get the source package or the vcs somewhere? [13:28] nessita, it's hard to say without seeing what is in the source [13:28] seb128, lp:libubuntuone/ubuntu [13:28] seb128: let me dig a deep to try to guess where that package builds from [13:28] ok, so the problem is that the package was renamed to gir1.2.... but debian/rules uses gir1.0 [13:29] rodrigo_: but that is maverick, is it correct the package to be gir1.2? [13:29] rodrigo_, seems they tried to make the build work with either abi version [13:29] rodrigo_, the build-depends are gir1.0 | gir1.2 for example [13:30] hum [13:30] the nightlies build from the same package branch afaik [13:30] I think we need to ask dobey in this case, he modified that, I think [13:30] seb128, where do you see that? in the branch I pasted, it just build-depends on gir1.2* [13:30] the issue is that natty and maverick have different gir versions [13:30] rodrigo_, http://launchpadlibrarian.net/62246782/buildlog_ubuntu-maverick-amd64.libubuntuone_0.3.9%2Br113~maverick1_FAILEDTOBUILD.txt.gz [13:30] the build log nessita gave [13:31] seb128, right [13:31] seb128, rodrigo_: I can't find a dedicated branch for this build... [13:31] "gir1.2-glib-2.0 | gir1.0-glib-2.0, gir1.2-json-glib-1.0 | gir1.0-json-glib-1.0" [13:31] nessita, we'd better ask dobey, this seems to use a different branch [13:32] ok, I will [13:32] thanks for looking into this! [13:32] nessita, ok, we need dobey to point us to the source used for those builds [13:32] yw ;-) [13:32] right [13:32] :-) [13:32] I'll change though the "real" branch to use gir1.2 everywhere [13:33] rodrigo_, https://code.launchpad.net/~ubuntuone-control-tower/libubuntuone/packaging-dailies [13:33] nessita, ^ [13:33] that's the one used for the daily build [13:33] r6 has the dual gir thing [13:34] yes [13:36] me browses [13:36] nessita, I'll build locally the natty version to see if that gir* stuff needs fixing [13:36] so will let it building while I get lunch, bbiab [13:36] rodrigo_, you also need to rename libwebkit-dev to libwebkitgtk-dev [13:36] in natty [13:36] ok [13:37] rodrigo_: awesome, thanks! [13:37] what was that webkit rename for? [13:37] Laney, upstream changed the library name [13:37] it's to have the binaries matching the soname [13:38] seb128, libwebkitgtk-3.0-dev ? [13:38] rodrigo_, no? [13:38] libwebkitgtk-dev [13:38] no, sorry, libwebkitgtk-dev [13:38] or do you want to build with gtk3? [13:38] didn't see it [13:38] ok [13:38] ok, lunch now, will do that later [13:39] rodrigo_, enjoy [13:39] Laney, debian did the same rename in experimental [13:41] seb128: yeah I see now, didn't catch that in the changelog [13:41] do they Provide: the old names? [13:42] Laney, no, but since the soname changed that would be misleading [13:42] maybe for the dev [13:42] they don't provide it, things would crash on a no found lib [13:42] for BDs [13:44] bazaar.launchpad.net must be down.. can't push or ping it [13:44] good morning all [13:44] bcurtiswx: Works for me. [13:44] bcurtiswx, works there [13:44] hey btw [13:45] hmm, odd [13:45] Laney, hum, I guess it would work for this one...do we still have lot of things not transitioned? [13:45] bcurtiswx: works for me too, I just pushed a branch [13:46] * bcurtiswx wonders why it doesn't ping [13:46] we have a small ice storm here, closed everything down [13:47] works now yay [13:48] kenvandine, are you back today? [13:49] its not 9 yet, i keep forgetting i got up at 5am ET :\ [13:49] nessita, rodrigo_: ok, so [13:49] you need to add to debian/libubuntuone-dev.install [13:49] "-debian/tmp/usr/share/gir-1.0/UbuntuOne-1.0.gir" [13:50] ups [13:50] "debian/tmp/usr/share/gir-1.0/UbuntuOne-1.0.gir" [13:50] nessita, to fix your build issue [13:50] seb128: http://paste.ubuntu.com/555392/ === oubiwann is now known as oubiwann_ [13:50]