=== jono is now known as Guest68697 [06:43] didrocks: btw was the bamf/precise ok for uploading now? [06:44] Mirv: well, I think you saw the activity about the other issues that I was kind of alone to deal with :) so not on the bamf thingy yet until we are fine on doing one unity release [06:45] didrocks: sure, sure, no hurry [06:45] Mirv: don't worry, you are in my opened tabs! Can you ask me again at the EOW please for safety? :) [06:45] the big milestone is most important to achieve [06:45] I hope your browser never crashes in a way that it loses tabs ;) [06:45] EOW ok [06:46] Mirv: well, it happens sometimes ;) [06:46] Mirv: but TBH, I think I can reopen most of them by the history [06:47] of course, there are maybe some collateral damage ): [06:47] hence pinging back is nice! [06:47] Mirv: oh, can you install ~ubuntu-unity/daily-build ppa please? [06:47] (on raring of course) [06:47] and shout if anything ugly happens? [06:56] didrocks: will do [07:07] smspillaz: FYI, there is no pressure from distro at all about this libmetacity thing [07:30] mmrazik: I think not all projects are restarted yet, some approved stalled [07:30] like on unity (https://code.launchpad.net/~3v1n0/unity/ups-empty-menu/+merge/136828) [07:30] didrocks: not sure what you mean by restarted but there was some power outage yesterday (again) [07:30] maybe the things approved during the shutdown? [07:30] didrocks: its building [07:30] ok :) [07:30] just the backlog was too big? [07:30] after the power outage [07:30] or did you have to kick it to restart after this? [07:31] didrocks: I had to kick it after the restart. Something got suck there blocking everything else [07:31] didrocks: right now nux and compiz and unity are building [07:31] and there are 3 more builds for compiz queued [07:31] mmrazik: excellent, then I can try to do a manual release :) [07:32] which should unblock the universe and nexus 7 :p [07:32] didrocks: the compiz branches need to land in sequence so I think it will take a few hours... [07:32] mmrazik: ok, no worry [07:38] didrocks: k, thx === mmrazik is now known as mmrazik|afk [08:21] didrocks: so far so good with ppa:ubuntu-unity/daily-build on raring, not seeing anything out of place or crashing [08:22] Mirv: sweet! do you see close button when you scale an app? [08:22] like open 2 nautilus windows [08:22] click on the launcher [08:22] normally when you hover the 2 windows opened, you can see the close button [08:22] didrocks: yes, I see it [08:22] and it works [08:22] interesting, for me, they work but I don't see them :) [08:23] thanks Mirv! Do not hesitate if you see anything bad [08:23] sure, I'll let you know immediately if I notice something bad [08:46] didrocks, i get a close button, but on the foreground window I also see the titlebar with the buttons in the titlebar of the hovered app.. sometimes.. hard to trigger http://popey.com/~alan/titlebar.png [08:46] popey: ah, even funnier :) [08:46] popey, confirming [08:46] I wonder why I don't get the close button at all [08:46] ever never ever [08:46] hey davidcalle! [08:46] davidcalle: can you please install ~ubuntu-unity/daily-build [08:46] thats a guest session though.. dunno if that makes a difference? [08:46] and warn/poke me if anything bad is happening :) [08:46] popey: let me try [08:47] open two nautilus, then open a terminal and close the terminal, then click nautilus in launcher [08:47] seems to reproduce it [08:48] popey: it is indeed working in a guest session [08:48] didrocks, hey. Ok. [08:48] popey: I tried what you say, but doesn't reproduce [08:48] davidcalle: thanks! (on raring ;)) [08:49] yeah, it doesn't always happen [08:49] let's hunt for Trevinho once he's around :) [08:49] roger roger [08:49] but not a first daily-release blocker [08:49] nah [08:49] (I'm not relying on autopilot for it yet though, hence the dogfooding) [08:49] * popey hunts for more [08:49] heh ;) [08:51] didrocks, http://popey.com/~alan/unreadable_close.png [08:51] bottom right window, the close button overlay is unreadable [08:52] ah, so not he back black fake title bar [08:52] yeah [08:53] seems not to resize properly [08:53] there is really something… some elements don't appear randomly [08:53] popey: oh it doesn't resize [08:53] popey: it's a fake, it overlays :p [08:53] ah, thats why you can see through on my first screenshot then, yes [08:53] yeah [08:53] its just not appearing [08:53] I thought it was opaque [08:53] but it's not :) [08:54] whenever i want to open shotwell from the dash, it wants me to buy shot glasses, this can't be good for me [08:54] popey: ahah, you can argue "but but, it was proposed to me, I needed them for sure!" :) [08:54] true! [08:54] blame mark [08:55] there's a lawsuit in there somewhere [08:55] heh [09:02] didrocks, open two apps, one full screen (e.g. shotwell) and another windowed (e.g. gimp). Click shotwell in launcher, get shotwell full screen. Want to go to a menu in gimp, click gimp in launcher, gimp comes to front. Now without clicking on the gimp window at all, hover over the menu bar, it flicks back to shotwell.. [09:02] (if you click on the menu) [09:02] annoying but not a showstopper IMO [09:03] popey: convoluted enough to not be a showstopper I guess :) but worth a bug and a ping to bregma :) [09:28] FYI all: unity stack rebuild for first release into raring in progress :) [09:29] didrocks: My preview of it looks like the old one, minus bugs. But that's good :) [09:30] duflu: heh, sweet! :) === mmrazik is now known as mmrazik|lunch === _salem is now known as salem_ === mmrazik|lunch is now known as mmrazik [13:15] popey: mhmh... weird [13:18] popey, didrocks: never got it here... Mhmhm... the first one it's just probably an opacity issue... But the 2nd... [13:19] Trevinho: I never have the close buttons here [13:19] popey: do you get issues also on the Alt+Tab spread? (i.e. doing Alt+` or alt+tab + down-arrow)? [13:19] Trevinho: clicking on where they are supposed to be works though [13:19] didrocks: mhmh... it could be a problem related to textures loading, I've heard something related to that some time ago [13:19] Trevinho, no, spread is fine [13:20] this is an intel hd2000 video card, so low-end i7 [13:20] popey: wierd, it's really the same code-path [13:20] Trevinho: how can I give you the debug info you need? :) [13:23] didrocks: try to put a SetupSharedTextures() above the line with "switch (close_icon_state_)"... [13:24] Trevinho: will do, not right now as I'm releasing unity, but after, I'll give it a shot :) === dandrader_ is now known as dandrader|afk [13:39] mmrazik: sil2100: hey, do you know where the autopilot tests results are stored? === dandrader|afk is now known as dandrader [14:23] didrocks: srry.. missed this again. What do you mean where they are stored. The physical xml files? [14:23] mmrazik: yeah, once autopilot ran [14:24] on the machine [14:24] didrocks: ~jenkins/results/testresults [14:24] mmrazik: thanks :) [14:26] fginther: hey! oh, it seems you implemented the fact to merge the changelog-change only without a rebuild, right? [14:27] didrocks: I don't think so as I didn't see any MP.... [14:28] mmrazik: hum? weird that's it's merging all my MP for the unity release blazing fast then [14:28] mmrazik: even if I sent 10 of them [14:28] didrocks, no, that was put on the todo list as it was not a blocker... [14:28] didrocks: I was surprised too [14:28] so you have maybe another issue where the package isn't built? :p [14:28] didrocks, this is for lp:unity? [14:29] the whole stack [14:29] fginther: I checked the logs (randomly). The build is there [14:31] didrocks, the lens generally build in under 10 minutes [14:31] yeah, but seeing the amount of requests I had, I'm *shocked* (in the good way ;)) [14:31] :-) [14:46] fginther: mmrazik: seems the bamf test is flacky again [14:46] I'll merge the changelog by end if you don't mind [14:47] didrocks, grumble, I'll ping somebody about that bamf bug again [14:48] fginther: thanks, workaround meanwhile :) [14:56] Trevinho, ping [14:56] fginther: pong [14:57] Trevinho, can you find someone to take a look at https://bugs.launchpad.net/bamf/+bug/1079329? It's become a blocker for autolanding bamf [14:57] Launchpad bug 1079329 in BAMF "bamf autolanding builds sometime fail with "kill: (nnnn) - No such proces"" [High,New] [14:59] fginther: I see there was someone working onit [15:00] fginther: ah, probably you was :) [15:00] hello everyone. im trying to figure out how unity works on a broad scale, which processes start the others, how they communicate etc. currently im digging through the src-package of unity and try to understand what "UBusMessage.h" does, or the ubus-server [15:00] naming reminds one of dbus, does ubus work the same way? or does it something entirely else? [15:00] fginther: isn't your way working? [15:01] Trevinho, I tried a couple of experiments, but ran into another problem [15:01] I'll try to find a log === dandrader is now known as dandrader|lunch [15:04] Trevinho, here's a log from one of my tests: http://paste.ubuntu.com/1412623/ [15:38] fginther: can't we just ignore the return value of kill? [15:40] Trevinho, there shouldn't be any harm. I can give it a try [15:42] Trevinho, do you know what causes the "WARNING **: Couldn't register with accessibility bus: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken."? [15:42] fginther: I think it's caused by the fact that running the test in a fake xserver the atk bus is not present [16:06] didrocks: FYI -- the nux build takes ~4h (due to arm). Your branch is building, it will just take some time [16:06] mmrazik: no worry, thanks! [16:11] I'll move the builds to pandas tomorrow === gatox is now known as gatox_lunch === dandrader|lunch is now known as dandrader [16:31] didrocks: Hi. Are you sure all build-dep are now in lp:unity ? The last time I tried to get the build-dep, the build was failing, because libxcb-dri2-0-dev was missing... [16:31] MCR1: well, it's building on the ppa :) [16:31] yeah, sure... but still strange... [16:36] smspillaz: Hi :) 1. Thanx 4 the reviews. 2. Q: Do you think it might make sense to use register int iterators for huge for loops ? [16:37] mterry: I just pushed that to distro, quick review? https://code.launchpad.net/~didrocks/bamf/remove-unexisting-dep/+merge/138242 [16:46] mterry: oh, btw, you need to change the global status for https://code.launchpad.net/~didrocks/unity/refresh-build-deps/+merge/138227 [16:46] approving in a comment isn't enough [17:02] sil2100: can you change the global status as well? :) [17:02] didrocks: ok ;) [17:02] thanks! [17:02] * sil2100 wanted to give some time for someone else to ACK it as well ;p [17:03] But this is an obvious change [17:03] yep :) [17:03] Something that was bothering everyone since looong time === gatox_lunch is now known as gatox [17:58] MCR1: not really, as daneil said micro-optimizations don't help much [17:58] yeah, probably you are right [17:59] MCR1: the best thing to do if you want to improve performance is to do profile guided optimization [17:59] I saw your new branch (actually I am testing it) [17:59] Top news ! [17:59] eg, run it through callgrind [17:59] 20->30 == + 50% [17:59] that sounds great [17:59] MCR1: actually, I just found the cause of the extremely poor performance when moving opengl windows around [17:59] yeah [18:00] 1) unfortunately its not easy to fix and 2) I really think nvidia should just fix it on their end, there's no reason why it should be like that [18:00] I am using ATI (currently with latest fglrx)... [18:01] but window moving is not fully smooth here either [18:01] yeah I think fglrx sends theiir command stream over the protocol too [18:01] MCR1: try commenting out line 848 in plugins/composite/src/screen.cpp too [18:02] I think that's the only other point where do a flush/wait lots [18:02] with framebuffer object enabled it stutters more than without, where it is almost fully smooth (with some minor pulsing) [18:02] MCR1: which card did you have ? [18:02] with both fglrx and gallium [18:02] HD 5750 [18:02] MCR1: resolution ? [18:02] 1920x1200+1280x1024 [18:03] or 1920x1200+1920x1080 [18:03] wouldn't be suprised if we were maxing out the fillrate [18:04] unfortunately there's not much we can do - getVideoSync / waitVideoSync require you to block the graphics pipeline and glXSwapBuffers necessarily means that you have to redraw the entire backbuffer on every frame [18:04] DRI3 will help that a lot [18:04] How can I force the compiled Compiz in the staging dir force to use my config ? [18:04] MCR1: did you actually install it anywhere ? [18:04] or are you just running it from /src/ [18:04] yes to staging [18:04] cause that won't work [18:04] compiz --replace ccp ? [18:04] installed [18:04] or at least [18:05] PATH=/home/user/staging compiz --replace ccp & [18:05] might need to set your LD_LIBRARY_PATH too [18:06] ah, maybe that is what I'm missing as CCSM opens with all my settings correctly enabled, but the freshly compiled Compiz runs without them... [18:06] so it is hard to compare the performance [18:06] #1 reason why I hate fixing performance things [18:06] "I made it go from 20FPS to 30FPS!" [18:07] "still too slow not fixed" [18:08] 20->30 sounds awesome [18:08] time for me to sleep though [18:09] I'll have a look into the opengl related thing tomorrow, I have an idea for that maybe it will work maybe it wont [18:09] hmm, strange - I've set LD_LIBRARY_PATH... [18:10] I will retry forcing compiz to load all plugins === dandrader is now known as dandrader|afk [18:16] haha, I've started the wrong CCSM... === dandrader|afk is now known as dandrader [19:53] MCR1: hi! What would it take to bribe you into fixing one of my pet-peeve compiz bugs? :) [19:53] MCR1: bug 1037164 [19:53] Launchpad bug 1037164 in compiz (Ubuntu Quantal) "Clicking on snapped windows in a different workspace produce unexpected results" [Medium,Triaged] https://launchpad.net/bugs/1037164 [20:18] mdeslaur: How do you "snap it" ? Via keyboard or mouse ? [20:18] oh, I just drag the terminal to the right edge of the screen, and it sticks to the edge [20:19] it's the default compiz snapping windows plugin [20:19] but even if I disable that plugin, as long as the window is a few pixels close to the edge of the screen, it happens [20:19] MCR1: are you not able to reproduce it? [20:19] hmm, I was not aware that there are still problems with mouse snapped windows... [20:20] I have to try in a VM [20:20] it is Grid you are talking about [20:21] no, not grid [20:21] MCR1, it's trivial to trigger, go to a ws, snap something to the right of the screen, go back to ws1, click on it in the unity launcher [20:21] the win is moved a bit over [20:21] as long as a window is within a few pixels from the right edge of workspace 2, the problem happens [20:21] so you have a border on ws1 and the remaining part where you were [20:23] mdeslaur, seb128: Uh, not nice indeed... [20:25] Indeed, grid is not needed to trigger the bug... [20:31] MCR1: there seems to be some details in one of the duplicate bugs here: https://bugs.launchpad.net/unity/+bug/834248/comments/29 [20:31] Launchpad bug 755842 in compiz-plugins-main (Ubuntu Precise) "duplicate for #834248 Non-maximized windows which sit on the border of a workspace move when called" [Medium,Triaged] [20:32] mdeslaur: Yeah, that seems very useful indeed. :) [20:40] mdeslaur: I can reproduce the issue running yesterday's Raring in a VM, but not when using my 4x1 config. [20:41] interesting [20:42] But I can also reproduce it using Raring and 4x1 desktop config [20:44] haha, I never saw this bug before, because I've deactivated Desktop Wall [20:50] I've been hitting it 20 times a day since natty :P [20:50] it's diving me insane :P [20:51] Sure, the worst things are windows that jump around on their own... [20:52] You could ofc workaround this special bug, by disabling Desktop Wall and working with the Cube and Expo instead [20:53] but you would have to change your config quite a bit and change workspaces from 2x2 to 4x1... [20:53] ofc fixing this bug would be a lot better [20:53] hrm, yeah...it's bad that it's been broken this long in the default configuration [20:56] the problem is that bugs of this kind seem to be extremely hard to fix - but I'll try (I am currently working on bug 1082001 which is also making windows move across workspaces) [20:56] Launchpad bug 1082001 in Compiz "Grid/Expo: Via keyboard-shortcut resized Grid windows confuse Expo and follow the user from workspace to workspace" [Undecided,Confirmed] https://launchpad.net/bugs/1082001 [20:57] This one kills grid keyboard functionality if one wants to work with different workspaces... [21:11] ah, another nice one :) === salem_ is now known as _salem [21:34] bschaefer, can you recommend someone to review this: https://code.launchpad.net/~fginther/nux/add-code-coverage/+merge/138007 [21:35] fginther, I can look at it...I don't think any other people are around for nux...possibly andyrock :) [21:36] bschaefer, it's not urgent, so if it's better that someone review it tomorrow, I'm fine with that [21:37] fginther, alright, well if I can test it, and check things are still working with nux I can approve it :) [21:37] bschaefer, cool! [21:38] mdeslaur: Time for a test ? [21:38] MCR1: sure! [21:39] please create a directory named compiz-1 in ~ [21:39] fginther, so this is for jenkins to get reports from nux pretty much? [21:39] MCR1: ok [21:39] * bschaefer hopes he has gcov [21:39] in compiz-1 create another directory called plugins [21:39] bschaefer, that's the goal, a developer can also run coverage-html to get their own results [21:40] fginther, cool, ill do that to test it out, and run some other tests [21:40] fginther, it doesn't seem like this touches other code, or really effects nux unless you have --enable-gcov which is good [21:42] mdeslaur: Download this file: https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/1037164/+attachment/3452533/+files/libwall.so and put it in ~/compiz-1/plugins [21:42] Launchpad bug 1037164 in compiz (Ubuntu Quantal) "Clicking on snapped windows in a different workspace produce unexpected results" [Medium,Triaged] [21:43] mdeslaur: Now open a terminal and run setsid unity (be sure you are using 12.10 ?) [21:44] yes, I'm on 12.10 [21:44] ok [21:45] please test and report [21:45] same thing, unfortunately...how do I tell it actually loaded that file? [21:46] fginther, Scanning . for .gcda files ... [21:46] geninfo: ERROR: no .gcda files found in .! [21:46] should I be doing something else before doing coverage-html? [21:46] if you ran setsid unity it will load the file (be sure that it is in yourhomedir/compiz-1/plugins) [21:47] yes, it's there [21:47] bschaefer, did you do 'make check' or 'make check-headless' first? [21:47] fginther, nope ... let me do that [21:48] named libwall.so, yes ? [21:48] fginther, possibly adding make-check-headless to the coverage-html? [21:48] bschaefer, I open to suggestions if there is a smarter way to do that... [21:49] MCR1: yep [21:49] or I suppose that would get run how you'll be doing that [21:49] grmpf [21:49] fginther, hmm [21:49] bschaefer, I didn't want to force 'make check' or 'make check-headless' [21:49] yeah [21:49] fginther, maybe adding another option that doe force it? [21:49] but it's not obvious how to collect coverage either [21:50] make check-coverage-html? [21:50] hmm [21:51] bschaefer, I can add some make targets to make it easier [21:52] fginther, yeah, as make check or make check-headless needs to be ran before doing the coverage [21:52] mdeslaur - it is .compiz-1 !!! [21:52] bschaefer, ok, I'll figure something out [21:52] fginther, or possibly some how adding an error message [21:52] he's gone... [21:52] if no .gcda files were found... [21:53] fginther, or you can make a comment in there, Im not the best with makes files haha [21:54] * bschaefer needs to fix the make check FTBS [21:54] mdeslaur: I was giving you wrong directions [21:54] mdeslaur: Please rename compiz-1 to .compiz.1 ;) [21:55] and run setsid unity again - sorry [21:55] .compiz-1 or .compiz.1? [21:55] .compiz-1 no? [21:55] .compiz-1/plugins [21:55] mdeslaur, just copy over the system .so and restart compiz [21:55] no need to f*ck up the system [21:56] you can always --reinstall the package to get back to normal state [21:56] true that [21:56] MCR1, it's easier to cp away and back the system one that to figure the .compiz [21:58] mdeslaur: Now tell me that it worx [21:58] hah, compiz didn't enjoy me copying over the .so while it was loaded :P [22:01] bschaefer: Hi, btw ;) I also have (easy) work 4 u... [22:01] MCR1, hello, link me [22:01] if its a review :) [22:01] yep, 1. https://code.launchpad.net/~mc-return/unity/unity.merge-reduce-scope-of-variables.2/+merge/137949 [22:02] 2. https://code.launchpad.net/~mc-return/nux/nux.merge-reduce-the-scope-of-various-variables/+merge/134787 [22:04] MCR1: copiz just dies when trying to open that .so...could you please just give me the patch to try? [22:04] MCR1, hmm well I just approved the first one but..hmm [22:04] MCR1, 111 === modified file 'services/panel-service.c' [22:05] you are messing with C files, and the standard for C files are to define things at the top [22:05] I think it depends if it is C89 or C99, but I can revert the .c file changes [22:06] but if it complies it worx [22:06] MCR1, I think it would be best for the C, and its ansi I believe [22:06] but i would prefer to keep to C looking like C [22:06] bschaefer: ok [22:06] MCR1, thank you :) [22:06] bschaefer: I'll change that then... [22:07] thanx 4 the review [22:08] mdeslaur: More or less I tried his solution: https://launchpadlibrarian.net/89746102/fix-762335.patch [22:08] MCR1: and it worked for you? [22:08] MCR1, your second one looks good and is approved :) [22:09] mdeslaur, I did not try yet, as it would mess up my whole config (I am using the cube, not wall those conflict with each other), but I'll try later... [22:10] MCR1: ok, I'll build it locally and try it too [22:10] bschaefer: For nux not all of the scopes are reduced yet, so there will be a Part 2 ;) [22:10] mdeslaur: Cool, please report [22:17] bschaefer: Should be fixed [22:17] MCR1, awesome thank you! [22:18] your fast help is awesome ;) [22:18] MCR1, nux needs a bit more help in parts :) [22:18] you know - I am just a newbie struggling with all this C++ complexity :-D [22:19] haha, it takes time [22:20] MCR1, also what you think should happen...when you grab a window then use the keyboard to change workspaces? [22:20] what do* [22:20] just the normal Ctrl+Alt+ [22:21] the window should move to the new workspace together with the mousepointer and the new workspace should be selected [22:21] imho [22:21] MCR1, yeah, I find it odd that a huge gab is now present [22:22] gap* [22:22] * bschaefer looks if a bug for that exists [22:23] the real problem is that Ctrl+Alt+Cursor + grabbing a window with the mouse is hard to achieve with 2 hands... :P [22:23] I am really struggling to get this done ;) [22:23] haha [22:24] i would never imagine it really happening, I was just looking at that other wall/gap bug [22:24] ah, yeah - the gap [22:24] its fixed in 0.9.8, but not in 0.9.7 so im digging through to find what fixed it [22:24] to see if it can be backported [22:25] it is still present here, btw... (without the wall) [22:25] MCR1, in 0.9.8? [22:25] and with out wall how are you dragging it to another workspace [22:25] when I grab a window and switch from workspace to workspace to workspace to workspace... [22:25] I can make the cursor move from its position [22:25] MCR1, just dragging it? [22:26] I am using the cube instead [22:26] shoot...I never use that plugin [22:26] no, dragging on the same workspace is fixed 100% [22:26] MCR1, that is what the bug is talking about [22:26] open another bug if you could :) [22:27] maybe this +/- 10 pixel thingy is hidden somewhere else also... [22:28] MCR1, hmm well it hasn't reached 0.9.8 yet [22:28] so if you bzr branch lp:compiz/0.9.8 [22:28] the -10 is still there, or at lease it was yesterday [22:28] I am running trunk of everything (unity staging PPA) [22:29] but the problem in 0.9.7 is waaay more then just a -10 [22:29] it seems to mess up on screen->vp().x() [22:29] or it seems to be missing that somewhere [22:30] I think it was just backported this morning (for me at lease), the -10 fix [22:30] I can reliably reproduce it when moving the window to the left from workspace to workspace [22:30] but not when moving it to the right [22:30] just the -10? [22:30] just a little at a time [22:31] MCR1, it is also UP but not DOWN [22:31] that you can do it [22:31] (if you are using a 2x2) [22:31] yes, I am using 4x1 and it is minor and does not happen every time I move to the left... [22:31] strange indeed [22:32] MCR1, it has to be the -10, as after that was removed it was perfect for me [22:32] if you grab the top left corner of the window [22:33] MCR1, well either way, Im looking for the larger problem :) [22:34] the one in 0.9.7 [22:34] * bschaefer goes off to hunt for it [22:42] good luck [22:48] thanks [23:24] hi, people, have unity some special automatic resolution setting?