[00:27] 17 hours and Apport retracing service still hasn’t retraced the crash report. :-P [00:30] Ah, missed seb128’s message. [05:31] My issue had already been reported. LP #686698. [05:31] Launchpad bug 686698 in Nux "compiz crashed with SIGSEGV in nux::IOpenGLSurface::UnlockRect()" [Undecided,Triaged] https://launchpad.net/bugs/686698 === ssj6akshat is now known as akshat|school [06:51] morning [06:52] that. [07:06] ion: moro [07:06] hi [07:31] good morning === smorar__ is now known as smorar_ [07:44] mornin [07:45] didrocks: could you check something for me? [07:45] hey smspillaz [07:45] smspillaz: sure, not right now, but I can [07:45] didrocks: have you got two screens ? [07:46] smspillaz: no, only one there… [07:46] hmm ok [07:46] didrocks: I'm wondering whether this thing that gnome-panel does is a feature or a bug [07:47] basically if you have a panel on the left side of screen 1, and then there is another screen directly to the left of that panel (eg screen 2 is left of screen 1) [07:47] err, the panel needs to be on the right hand side, rather [07:47] actually, on the left (slightly dyslexic) [07:47] anyways [07:47] if you have that, and then maximize a window on screen 1, it will go underneath the panel [07:47] this happens with metacity and compiz [07:48] I checked with xprop, it seems that the panel does not set the _NET_WM_STRUT for that part of the screen [07:48] it seems like a bug, but there'd be some specific code path that would make it not set that strut property, so I'm wondering if it is the intentional behaviour === smorar_ is now known as smorar__ [07:52] smspillaz: ok, so if you have two screens [07:52] and gnome-panel is on the left [07:53] you mean that it's doesn't set WM_STRUT and so, apps go underneath? [07:53] didrocks: yes [07:53] but only on panels that are adjacent to other screens [07:54] it would seem like this is not the desired behaviour ... [07:54] smspillaz: weird, WM_STRUT is set screen by screen, isn't it? [07:54] so there is no reason to not restrain the area on the one with the panel [07:58] didrocks: it should be set by the output [07:59] I'm just compiling docky now to see what the behaviour should be [08:27] greetings everybody [08:29] morning MacSlow [08:30] didrocks, salut here too :) [08:30] :) [08:33] reasons I hate autotools #24533452: It only tells you which dependencies you are missing one dependency at a time [08:34] smspillaz, :) [08:35] smspillaz: reasons I hate cmake #${num_smspillaz}+1: cmake ignores it on some case and tell "let's take a chance" :-) [08:35] smspillaz, annoying indeed [08:38] Heh. Reason to hate cmake #whatever: CFLAGS? What's that? ☺ [08:40] hey RAOF! [08:40] RAOF: you mean CMAKE_CXX_FLAGS_DEBUG ? [08:40] Hey didrocks! Looks like that segfault in mipmapping on r600 might be fixed in git. [08:41] * smspillaz w00ts [08:41] RAOF: oh nice! [08:41] smspillaz: One of the rather endeering parts of autotools is that setting the CFLAGS/CXXFLAGS environment variable DTRT. [08:42] I guess. Although cmake was really designed to be multiplatform [08:42] it is not a full buildsystem, it only generates native makefiles [08:42] so it doesn't do things like setting env vars [08:43] (and you have to do a COMPIZ_DESTDIR which conflicts with the env one :p) [08:43] * RAOF is not going to get diverted into a build system discussion. They all suck! [08:43] really someone at debian needs to sit down and write proper macros for cmake [08:43] RAOF: shell + cp ftw :) [08:43] why? [08:44] because there are none? [08:44] we have our packaging system :) [08:44] so, no interest in writing something specific to cmake [08:45] grrr, too much flickering when switching ws in compiz :/ [08:45] really? [08:45] smspillaz: yeah, the thing we discussed the other day [08:45] smspillaz: like if you had a "magnet", not sure what's the best way to describe it [08:46] intel ? [08:46] nvidia [08:46] so it's like a vsync issue ? [08:46] exactly [08:46] quite disturbing [08:47] that's nvidia being stupid. You need to go into nvidia-settings, -> opengl settings -> sync to vblank, then ccsm -> opengl -> sync to vblank and restart compioz [08:47] that will fix it [08:47] and I have X damage not triggered when scrolling sometimes [08:47] smspillaz: hum, let me see if it fixes it [08:47] didrocks: ccsm -> workarounds -> force X to GLX sync [08:47] smspillaz: the thing is that we should do a distro choice there [08:48] perhaps [08:48] we can't ask our users to do that :) [08:48] not perhaps "certain" :) [08:48] didrocks: it should be enabled by default [08:48] smspillaz: even for non nvidia drivers? [08:49] didrocks: in the detection module we can distro patch workarounds to enable "force x to glx sync" if an nvidia card is detected [08:49] smspillaz: agreed [08:49] ccsm -> opengl -> sync to vblank is already checked [08:49] let me see the workarounds plugin [08:49] didrocks: check nvidia settings [08:49] that's where it counts [08:49] smspillaz: yeah, I checked it [08:50] what's the difference between the workarund setting and the opengl one? [08:50] workaround setting fixes damage issues [08:50] because the nvidia driver is stupid [08:50] so we have to slow down compiz a bit [08:51] smspillaz: heh, you are telling intel is stupid, nvidia is stupid, which next? :p [08:51] smspillaz: all checked, we'll see [08:52] smspillaz: should the workaround plugin be enabled by default? [08:52] smspillaz: i mean, I enabled it, and so, I have other settings now that I didn't have before [08:54] didrocks: sure [08:54] didrocks: disable by default the "show previews of minimized windows" though, that has random bugs that I'd rather not hit for no reason right now [08:55] (since it tells all kinds of lies to X11 to work correctly) [08:56] didrocks: hm, ok with docky I don't see the maximization issue [08:56] smspillaz: it's not activated by default [08:56] didrocks: ah ok good :) [08:56] smspillaz: what I have are: [08:56] OOo Menu Fix [08:56] Java Window Fix [08:56] Java Taskbar Fix [08:56] AIGLX Fragment Parameter Fix [08:57] and of course, the Synx X and GLX :) [08:57] you can disable AIGLX Fragment Parameter fix [08:57] it's not necessary any more [08:57] IMO I should write a patch for workarounds where we will enable Sync X and GLX only if there is and NVIDIA card [08:58] smspillaz: right [08:58] smspillaz: and the others? [08:58] didrocks: the others are find [08:58] *fine [08:58] what's the OO menu fix? [08:58] didrocks: OO menus don't register as type "menu" [08:59] smspillaz: so, why OO is enabled and not firefox/thunderbird? [08:59] didrocks: *shrug* [09:00] which means? ;) [09:00] didrocks: maybe the person who wrote them didn't enable them by default [09:00] smspillaz: so it should as well, I guess? [09:00] didrocks: only if you see problems [09:00] I didn't see any issue with OOo before enabling the plugin [09:01] ok :) [09:03] grrr, tomboy crashes… hard to take a note this way :) [09:04] didrocks: hmm, have you been getting random crashes with an error like "call to empty boost::function ?" [09:04] smspillaz: no, I've been quite lucky like that, all is pretty stable to me (just had a crash this morning) [09:05] s/like that/on that side/ [09:08] smspillaz: it didn't fix the sync issue [09:08] hum… maybe I should restart X to take the nvidia settings into account [09:08] didrocks: perhaps [09:08] didrocks: it fixed it here at least [09:08] didrocks: although it doesn't work on multiple monitors [09:09] well, not my case there :) [09:09] arghhhh. spent over 30 mins trying to find out why automake doesn't install a script to pkgdatadir and it was because I had _SCRIPT, not _SCRIPTS [09:15] didrocks: have you got this commit in the packages ? http://git.compiz.org/~dbo/compiz-with-glib-mainloop/commit/?h=glibmm-experimental&id=3e1c09f13eb99af88bf0bd8d73be8bbcf533ded7 [09:15] didrocks: You shouldn't need to restart X; in fact, restarting X will probably revert the changes :) [09:16] RAOF: smspillaz: ok, in that case, it doesn't fix it :) [09:17] RAOF: smspillaz: for once, I'll bring my heavy laptop at the Rally, maybe we can sit and look at it if it can help you? [09:18] smspillaz: I think I have, let me check [09:22] didrocks: about the symlink from /usr/bin/ to the python script in /usr/share/app. is it better to create the symlink with autotools or with dh_link? with autotools it's a bit tricky [09:23] kvalo: depends if you only care about the package or also the upstream tarball + make && make install [09:24] didrocks: good point [09:24] smspillaz: I'm still opening a bug for that issue (nvidia/sync), so that we can get it tracked [09:24] didrocks: I'll try to make the autofoo magic work. thanks [09:30] smspillaz: the commit is in [09:45] didrocks: huh, ok [09:45] didrocks: I thought that commit fixed it :/ [09:46] smspillaz: did you check the bug report mentionned the last ubuntu compiz package? [09:46] smspillaz: alpha1 version has it (1:0.9.2.1+glibmainloop2-0ubuntu4) maybe even ubuntu3 [09:46] jcastro, didn't njpatel present a diagram like the one you ask for in ayatana-dev at uds? [09:47] or ted at gnome summit [09:47] didrocks: no not yet, but I am still getting that crash [09:47] let me dist upgrade [09:47] lamalex, well then to start this diagram should be on the unity.ubuntu.com webpage [09:48] ping ted when he's awake [09:48] lamalex, the diagram is also not enough, we need things like "if you want to use appindicator you should check those apis" [09:48] with the list of libraries that you might have to use to write an indicator [09:48] and a short explanation of what they do [09:49] seb128, right [09:49] we need awesome developer docs [09:49] Hm. While I wait for mesa to build, is there any particular reason why unity-panel-service should take up 200MB RSS? [09:50] that's being worked but will take a bit [09:50] (documentation) [09:50] until now we could do a better job to at least point the sources used and what they do [09:50] RAOF, out of leaks you mean? ;-) [09:51] seb128: That thought had crossed my mind, yes :) === smorar__ is now known as smorar_ [09:58] lamalex: hey, seeing your comment on the merge, what's the difference between LAUNCHER_ICON_QUICK_STARTING and LAUNCHER_ICON_QUICK_LAUNCHING ? [10:02] didrocks, there is no LAUNCHING [10:02] didrocks, I just updated that comment, thanks [10:02] I'm sick :\ [10:03] lamalex: oh? the grey weather of London hit you? [10:13] salut seb128 [10:13] seb128: about https://bugs.launchpad.net/unity/+bug/686603 [10:13] Ubuntu bug 686603 in Unity "valgrind "Conditional jump or move depends on uninitialised value"" [Undecided,New] [10:13] lut dbarth [10:13] seb128: you get that on the console, i mean it's a debug log? [10:14] I get that by running test-panel under valgrind [10:14] ah right [10:18] smspillaz: hey sam, have you seen mvo bug with the link.py testcase? [10:19] smspillaz: that sounds like a scary regression [10:20] dbarth: no, could you link it to me ? [10:20] dbarth: also, could you post your config profile somewhere? I'm trying to track down the crash in windowInitPlugins [10:55] kamstrup: "Why can't I ever get the email API for LP working!?!?11one" <- same here :) [10:55] kamstrup: thanks, will remove the "there" :) [10:56] kamstrup: exec didn't want if you want to embeeded in gdb probably [10:56] didrocks: ah, right [10:56] didrocks: I found the answer on https://help.launchpad.net/Code/Review I think [10:56] "review approve" needs a whitespace prefix [10:56] kamstrup: oh? [10:56] " review approve" [10:56] :-S [10:56] kamstrup: sorry dude, but I have a long merge request again: https://code.launchpad.net/~kvalo/indicator-network/settings-merge/+merge/43183 [10:56] kamstrup: yeah, that's what I'm doing [10:57] kamstrup: all commands need a whitespace prefix [10:57] kamstrup: but for unknown reason, it hates me :) [10:57] kamstrup: but this is just a start, so I think you don't need to review it in so detail [10:58] kvalo: I think I may have to push it to monday [10:58] I really need to get the dee gdbus port out of the door [10:58] and friday is sprint [10:58] kamstrup: monday is ok for me :) [10:59] * kamstrup rolls up sleeves to get those sodded unit tests running again === njpatel is now known as njpatel_ === akshat|school is now known as ssj6akshat [12:12] when someone has a minute? https://code.launchpad.net/~alexlauni/unity/xids-introspection/+merge/43193 === smorar_ is now known as smorar__ === MacSlow is now known as MacSlow|lunch [12:17] njpatel_, didrocks, with CMake can we have unityshell.xml.in have variables in it that get set on build? [12:18] lamalex: yes, look at what I've done some commits ago with unity.cmake in tools/ (it's the binary) [12:18] didrocks, sweet thanks [12:19] hello everyone! don't know when it started, but window buttons now have tooltips again - should this be filed as bug? [12:19] lamalex: so that to replace VERSION and tools/CMakeFile* [12:19] htorque: it's already filed again compiz, I dropped my patch and I'm just working on it again :) [12:19] htorque: I dropped it when going to 0.9 [12:19] didrocks, nice, thanks! [12:19] htorque: thanks for your excellent reporting and testing job :) [12:20] yw! thanks for all the fixes. ;-) [12:25] njpatel_, didrocks: so my launcher is acting weird [12:25] seb128: oh weird? [12:25] like the launchers do a start animation when clicked [12:25] but they don't start anything [12:26] if I run firefox manually it doesn't match the pinned icon [12:26] but add a new one in the launcher [12:26] seb128: is it an application you already started? [12:26] same for nautilus or gedit [12:26] well firefox and gedit have no running process [12:26] but I used them today yes [12:26] like, you started/closed them 4 times? [12:26] could be, I didn't count [12:26] seb128: I think you're striken by the ubuntu starter edition :) [12:27] seb128: should be fixed in trunk btw :) [12:27] is there a way to reset the counter? [12:27] seb128: killing the bamf service and restarting unity [12:27] we didn't merge the fix last week because it added others issues IIRC [12:28] ok, compiz --replace did it [12:28] is that known that compiz crashes on --replaec? [12:28] well apport triggers [12:28] seb128: bug #683623 [12:28] Launchpad bug 683623 in unity (Ubuntu) "Starting a launcher only works four times" [High,Triaged] https://launchpad.net/bugs/683623 [12:28] yeah, I've read this one last week [12:29] seb128: yeah, compiz isn't good when you give unknown args and unknown plugins [12:29] IMHO, if it can't find a plugin, it should just ignore it [12:29] I just though it was when clicked 4 times in a row [12:29] seb128: not in a row :) [12:29] didrocks, thanks [12:29] seb128: yw [12:30] seb128: btw, we will have the unity binary from now on [12:30] yeah, I noticed, great work [12:30] so, will make things easier to reset the config and such [12:31] well, not very difficult, just that I hate python with making some kind of tee-like behavior :) [12:36] smspillaz: are you changing gtk/window-decorator/gtk-window-decorator.c with your decorator work? [12:49] didrocks: no, I am writing my own decorator [12:49] didrocks: well actually, I forked gtk-window-decorator (temporarily) and I am adding inactive/active shadow work [12:49] smspillaz: ok, so I should maybe wait or set it to you the task of the option to hide tooltip? [12:50] smspillaz: it was a distro-patch we had (that I submitted upstream as well some month ago). [12:50] didrocks: Our unity-window-decorator can neuter tooltips [12:50] smspillaz: nice! so yeah, please do that :) [12:50] * didrocks drops his task then ;) [12:50] could you link me to the patch so I don't have to hunt around in the code ? ;-) [12:51] smspillaz: well, you removed the gconf part in 0.9 so you will have to add it again [12:51] didrocks: did we want it as an option? I thought design said no tooltips ever [12:51] smspillaz: but the distro-patch was http://paste.ubuntu.com/541431/ for 0.8 [12:52] smspillaz: well, I added an option to get it accepted upstream [12:52] smspillaz: but you never answered! [12:52] :) [12:52] didrocks: must have missed it [12:52] smspillaz: so if it's our own, yeah, no worry, no option [12:52] didrocks: yeah, that's what I was thinking [12:52] smspillaz: http://bugs.opencompositing.org/show_bug.cgi?id=1264 [12:53] didrocks: ah right. I never checked the opencompositing bugs, I was too busy porting plugins :p [12:53] smspillaz: no worry, patch tagging is useful to refind bugs even if the bugtracker is down :p [12:53] smspillaz: so I'm assign you the relevant compiz bug, one sec [12:54] didrocks: yeah, the opencompositing bugtracker is (still) down [12:54] smspillaz: you should use launchpad for upstream! ;) [12:54] we probably could [12:54] bugzilla is a PITA [12:55] smspillaz: +1 [12:55] smspillaz: so, activating the workaround plugin with upstream defaults for now minus AIGLX one === MacSlow|lunch is now known as MacSlow [13:04] njpatel_, how does the launcher's autohide get set initially? [13:04] does optionChanged get called once at start? [13:06] lamalex, I'm not sure about the compiz options stuff, sorry [13:07] njpatel_, who is that? Jason? [13:07] yep [13:07] or smspillaz [13:07] ok [13:07] DBO, wake up === oubiwann is now known as oubiwann_ [13:18] njpatel_, could you please review https://code.launchpad.net/~alexlauni/unity/xids-introspection/+merge/43193 [13:19] ronoc: hi. time again for a review: https://code.launchpad.net/~kvalo/indicator-network/gdbus-cellular-crash/+merge/43204 [13:26] ok guys, added C++ to the style guide: https://wiki.ubuntu.com/Unity/CodingStyle [13:28] lamalex, will take a look in a bit [13:30] jcastro, you can remove everything from "Be Verbose" to before "Commit Messages", as it's all stated at the start of the C++ bit [13:31] ok [13:32] done [13:32] njpatel_: what else is C++ after nux and compiz? [13:47] njpatel_: don't you mind if I move unity_support_test from examples to tools (I'll add the directory and such) as it's not really an example :) === njpatel_ is now known as njpatel [13:47] jcastro, that's it for now [13:51] lamalex, approved with a comment [13:51] please fix before merging [13:52] lamalex: njpatel: any new people in merges yet? [13:52] jcastro, waiting for one to sign CA [13:52] I see 2 pending merges from people I don't know, so crosses fingers [13:52] yes a few, but I'm not sure their success rate, looking now [13:52] rock [13:52] er, waiting for one to fix his branch and sign ca [13:52] lamalex: hey, let's do this, when they're new and asking you for CA when you reply to them CC me, I like to get a little bio on each person and a photo [13:52] so I can put them on planet [13:53] lamalex: just a fwd or CC is fine, I can handle introducing myself, etc. [13:53] jcastro, sure [13:55] nooooooo, I'm starting to google "cmake" when I want to google "autotools" [13:55] that's the end /o\ [13:57] njpatel: did you see my question about unity_support_test location and nux? === oubiwann_ is now known as oubiwann [13:59] jcastro, I am always getting interrupted when I am about to write a post on the Bitesize bugs [14:00] I guess someone doesn't want to it to appear :P [14:01] ssj6akshat: it's ok, it's a marathon not a sprint [14:02] didrocks, has anyone considered changing the workspaces settings to be 2x2, so expose would look more like it did in the mutter based unity? [14:02] and imho it is a better use of space [14:02] dbarth: speaking of, getting branches on these, I will need more! https://bugs.launchpad.net/unity/+bugs?field.tag=bitesize [14:02] kenvandine: it will do that later the cycle, but currently compiz doesn't listen to gconf gnome keys [14:02] didrocks, no, sorry. Erm, I don't know what that does so feel free to move it where you feel right :) [14:03] didrocks, humm... i have mine set to do that now [14:03] njpatel: ok, it's the detection module at start, so definitively not an example :) [14:03] right [14:03] kenvandine: right, but it's not the GNOME keys, that means, it's not share with your desktop classic session [14:03] njpatel: ok, doing and commiting [14:05] njpatel: didrocks: while I have you here, post release I need more bugs kthx. :) [14:06] jcastro: well, I'm tagging everything I can… i've reviewed all the bugs as you already know and add them :) [14:06] I can't believe I am disappointed that there aren't enough bugs [14:06] yesterday or so neil fixed like three of them and I was like "no! Don't take them, let someone else do them!" then I came to my senses. [14:07] jcastro: I guess once the dash/places will be there, we can have more of them [14:07] yeah [14:07] how many are still there? [14:08] ~10 [14:09] but with these 2 pending ones we've had one per day since I announced the report on tuesday [14:09] didrocks, i see [14:09] so I think it's off to a good start considering we're still just around A1 [14:09] didrocks, ok, thats fine... glad it is planned [14:09] :) [14:10] jcastro: right, little issues will come later, when it doesn't need refactoring everything :) [14:12] njpatel, I don't think I should free the list from bamf [14:12] I'm pretty sure bamf owns that list (granted bamf should probably be duping the list and then returning it) [14:17] lamalex, the docs say it owns the contents but not the list, that's what I was going on [14:18] lamalex, the code seems to suggest the same [14:18] kvalo, done [14:20] ronoc: thanks! [14:21] kvalo, np [14:21] agateau, got a sec? [14:22] ronoc: give me 5 minutes [14:22] agateau, no rush [14:29] ronoc: ok, I am back [14:30] agateau, one sec [14:34] agateau, can see anything obvious why the submenu in this dbusmenu is not appearing on the indicator [14:34] http://pastebin.ubuntu.com/541475/ [14:35] haha so which is worse, potential race conditions, or making a member variable public [14:35] * agateau looks [14:35] lamalex, race condition any day of the week [14:36] man but making a member variable public feels so wrong [14:37] ronoc: menu looks good [14:37] lamalex, i dont like waking up [14:37] lamalex, I know, but sure isn't every python obj wide open [14:37] lamalex: making a member variable public can be a good way to get race conditions :) [14:37] DBO, I think the same thing about you waking up, but I didn't want to say anything ;) [14:37] ronoc, yeah and it's one of the things I hate about python [14:38] tedg, can you turn that into a policy please [14:38] lamalex: did you get any further with atk? [14:38] lamalex, I take ruby over python any day of the week [14:38] ronoc: I think tedg built a tool to load a menu from json, you should try if your dump shows there [14:38] I'm wondering about the event redirection issues i'm facing, but they may be the result of offscreen rendering [14:38] still trying to work it out :/ [14:38] * ronoc must stop saying 'any day of the week' [14:39] tedg, good morning -> http://pastebin.ubuntu.com/541475/ [14:39] ok [14:39] so is appmenu maintained and by who? [14:40] dbarth_, ^ [14:40] tedg, the submenu is not being rendered in the indicator [14:40] tedg, can you see any problems there [14:40] doh! ;-) [14:40] agateau, thx - might try this [14:41] seb128, indicator-appmenu is me, appmenu-gtk is bratsche [14:41] ronoc, looking [14:41] tedg, I guess with unity none is yours? ;-) [14:41] tedg, things like clicking on a menu activating another item than the one selected [14:41] seb128, ? indicator-appmenu is used by unity [14:42] tedg, I confused with indicator-applet-appmenu [14:42] ignore me ;-) [14:42] seb128, If it's an item on the panel, that's njpatel, if it's in the actual menu that's me. [14:42] it's an actual menu item [14:42] I select "show conversations history" and it sends a "share desktop request" [14:42] it's quite embarassing [14:43] I just shared my desktop with someone I didn't want to talk to [14:43] twice [14:43] Hmm, that's odd. [14:43] Consistently? [14:43] when it happens yes [14:44] * tedg can't find anyone to share his desktop with :( [14:44] now the menu item just doesn't work [14:44] tedg, well, open a buddy list [14:44] double click on someone [14:45] go to contact, and select conversation history [14:45] Yeah, share my desktop is consistently greyed out. [14:45] does it work? [14:45] or does it do nothing? [14:46] It did nothing, but now I've got a menu from somewhere else? [14:46] I had that before as well [14:46] I got the nautilus menu at some point [14:46] but otherwise it does nothing [14:46] then I picked "info" [14:46] which send a share desktop request to dholbach [14:47] seb128, And you didn't want to talk to dholbach more than once? ;) [14:47] no, I'm just testing on dholbach now :p [14:47] Hmm, okay. I'm not sure what's up there. [14:47]