[03:25] thomi, ping [03:25] yo [03:26] hey I saw your awesome work using valgrind! [03:26] heh. It's not finished yet :) [03:26] and was trying to see what I could find but I kept getting these python 2.7 errors [03:26] and nothing relating to unity :( [03:26] I'm starting to think I've gotten most of the easy ones. [03:27] There's a bug in pango / fontconfig that leaks every time we cann pango_get_text_extents... I haven't had time to follow that up with the gtk folks yet [03:27] I keep getting this when I run valgrind [03:27] ==7515== by 0x815749B: PyEval_EvalFrameEx (in /usr/bin/python2.7) [03:27] bschaefer: how are you running valgrind? [03:28] alias valgrind-unity='G_SLICE=always-malloc G_DEBUG=gc-friendly valgrind --tool=memcheck --num-callers=50 --leak-check=full --trac k-origins=yes --log-file=unity-valgrind.20120409T160245.txt unity --reset 2>&1 | tee ~/logs/unity-valgrind.20120409T160245.log' [03:28] heh, exactly the same as me [03:28] tim provided that a few months ago [03:28] bschaefer: and unity is in your PATH and everything? [03:29] actually, mine is slightly different - I make it output an XML log so I can view it with valkyrie [03:29] where I install it? [03:29] yeah [03:29] hmm I don't see /opt/unity there [03:29] Ill add it and see if that helps [03:29] bschaefer: if you do 'which unity' does it tell you the correct one? [03:29] and BTW: ewwww... installing to /opt is icky :) [03:30] I just take the output and use a python script to parse it :) I love parsing for some reason haha [03:30] haha, it was something I followed a while ago stuck with it! [03:30] does your python parsing script show you the source code for each line in the stack trace? [03:30] ;) [03:31] pfft, no. It just parses useless stuff from the log file haha nothing fancy yey [03:31] yet* [03:43] thomi, hmm I think it was running the wrong one. The top of the log says "/usr/local/bin/unity --reset" [03:43] that could be the issue [03:43] but Im still not getting anything around unity :(. Ill take a better look at this tomorrow. Thanks! [03:44] good luck with the ones you have found! [04:35] bschaefer: still around? === sil2100_ is now known as sil2100 [09:14] if not mistaken, bash meta key is used by the HUD and/or by the dash. any way to set another meta key in gnome-terminal? [09:15] or change the meta key used by the HUD and/or dash [09:36] davidcalle, hey [09:37] seb128, hey [09:37] davidcalle, how are you? [09:37] seb128, fine and you? [09:38] is the dash not normally resizable ? [09:38] davidcalle, I'm good thanks [09:38] davidcalle, sorry for assigning you a music lens rhythmbox scope bug, I though you were the one who wrote the rb scope this cycle for some reason [09:38] davidcalle, do you know who did? ;-) [09:40] didrocks, mhr3: https://bugs.launchpad.net/ubuntu/+source/unity-lens-music/+bug/964347 is in the top3 duplicates from april, could you have a look (there is some dee function in there) [09:40] seb128, I did, with mhr3. We have a divergence of opinion on how to solve the bug in question. And as he has more experience than me... [09:40] Error: ubuntu bug 964347 not found [09:40] davidcalle, so I should assign the bug to mhr3? [09:40] mhr3: do not hesitate to unblock that ^ [09:40] seb128, mhr3, the bug in question : album art is present, but not for everything. [09:41] davidcalle, well in my case album art is present for everything embedded in the mp3s and not showing [09:41] mhr3, you know, the awkward tdb database. [09:41] davidcalle, I did add the art to all the albums in the dir I use using itunes by then for ipod enjoyment [09:42] davidcalle, i see a crash bug, no mention of album-art [09:42] seb128, that was fixed in latest dee [09:42] seb128, rbox uses several ways to store album art. And there is still one place we are not looking for. [09:42] mhr3, good, can you close it? [09:42] mhr3, sorry cross discussion [09:42] seb128, sure, looking for the original bug # [09:43] mhr3, 963991 [09:43] ? [09:43] seb128, yep [09:43] good [09:43] davidcalle, mhr3: so yeah, the second bug is bug #976067 [09:43] Launchpad bug 976067 in Music Lens "unity music lens pictures" [Undecided,New] https://launchpad.net/bugs/976067 [09:44] davidcalle, mhr3: in my case all my mp3 have the artwork embedded in the file (done by itunes back then) so it's nothing specific to rb [09:44] the artwork should show [09:44] it does show on an ipod in rb or in banshee [09:44] but not in the lens [09:44] can I help debugging? [09:46] davidcalle, so wrt to the album art bug, the issue was that there is nothing i had known of to parse that thing, last time we talked about it you said you'll check with rb devs to see what they use to parse it, correct? [09:46] seb128, the artwork would show... if we were looking for it. We don't, we rely on where the music player stores it. [09:46] ok, well currently the lens is a bad experience for me, all my mp3 have the artwork showing everywhere and the lens is a full grid of no artwork icons :-( [09:47] mhr3, correct, I've seen a few ml discussion of people having the same question, without answers. There is a reliable way to extract the info... But it's kind of hack-ish. [09:48] davidcalle, it's either reliable or hackish, so which is it? :) [09:49] mhr3, heh, hackish because we parse the raw db file. Reliable, because it's been working for the python scope for six months. [09:50] parsing raw db files cannot be reliable, db systems have locks and we're just lucky that it works most of the time [09:52] mhr3, if we parse the file, there is no lock issue. We can still duplicate it before messing with it. [09:52] i'm sure we'd see issues if we unleashed that to all ubuntu users [09:53] ranging from crashes to subtle incorrect string termination bugs [09:57] mhr3, crashes can be fixed, we need the data. And from what I've seen, there are no vala bindings for tdb. [09:58] davidcalle, what does rb use to work with it? [09:59] the crashes will be bound to the way you're getting the data, so the only fix is to use a library that can work with those dbs [10:01] mhr3, it has a dep on libtdb1. And I've just noticed tdb-tools. [10:03] davidcalle, then a proper fix is to write a quick libtdb binding and use that [10:06] mhr3, you? [10:08] me what? :) [10:08] mhr3, I'm obviously unable to do this kind of magic, so I'm asking if you are? [10:09] there go my plans to work on libunity docs [10:23] if not mistaken, bash meta key is used by the HUD and/or by the dash. any way to set another meta key in gnome-terminal? or change the meta key used by the HUD and/or dash ? === kiwinote_ is now known as kiwinote [10:52] JohnLea: ping === MacSlow is now known as MacSlow|lunch [11:12] gord, mhr3, jaytaoko: is it known that the lens selection doesn't react well to mouseover in the dash? [11:12] like often the right "rectangle" doesn't get highlighted when the mouse enter it and clicks don't work [11:12] it gives a buggy feeling [11:13] seb128, selection of what? [11:13] the items in the search resultset? [11:14] seb128, "clicks don't work", are you clicking on the rectangle or on the icon? looks like clicking on the rectangle doesn't work here, but the icon does [11:14] mhr3, no, the lens icons at the bottom of the dash [11:14] oh [11:14] gord, I'm clicking on the rectangle :p [11:14] yea, i can reproduce what gord says [11:14] well sometimes the rectangle doesn't display as it should on mouseover too [11:14] can't reproduce that though [11:15] like if I do left-right with the mouse at a moderated speed I can "enter" areas that don't get highlighted [11:16] yea, i see that too, if you enter mouse from the top but you go to the middle of two icons, there's no rect [11:16] although it appears once you move over an icon [11:17] right [11:17] well the behaviour is not consistant, sometimes the selection change when entering the "rectangle" [11:18] mhr3, ok, you are right, it just change when mouseovering the icon [11:18] but in most my "normal" moves I overshot enough that I touch the icon it seems [11:19] where sometimes I stop just before [11:19] nonetheless i have zero idea of that parts of the code :) [11:19] gord, do you know if it would be complicated to make the clicks work in the selection rectangle out of the icon as well? the current behaviour feels buggy since often you are not over the icon and the click seems to not work [11:20] potentially, i never actually did any work with those buttons so i'm not sure, would have to take a look later in the week [11:21] gord, should I ping jaytaoko or somebody else rather? ;-) [11:21] well it's not a blocker in any way [11:21] one of those lack of polish that it would be nice to fix for the lts :p [11:22] i think everyone has a bunch of priority stuff to fix today, i know jay does ;) [11:22] he may not know it yet but he does ;) [11:22] hehe, indeed [11:23] I wil come back in a few days with all the small nitpick :p [11:23] but yeah i agree it would be good to fix, just need to get a few scary bugs off our plate first [11:23] gord, SRU0 or SRU1 will do [11:23] seb128, andyrock could fix this [11:24] Cimi, what? [11:24] Cimi, that would be nice ;-) [11:24] seb128, it was supposed to work as andyrock implemented by design [11:24] andyrock, hey [11:24] seb128, I agree with you though, keep JohnLea in the loop [11:24] andyrock, run! work is coming :p [11:24] seb128, cool :) [11:24] andyrock, we were discussing how selecting a lens with the mouse feel buggy because you have to click on the icons [11:25] andyrock, i.e you can't click in the rectangle out of the icon [11:25] seb128, got it [11:25] andyrock, do you think it's something you could look at once you are out of blocker bugs? [11:25] seb128, of course [11:26] andyrock, thanks, I also assigned you bug #973386 not sure if you noticed [11:26] Launchpad bug 973386 in unity (Ubuntu) "[regression] The keyboard shortcuts overlay bottom is truncated." [Medium,Confirmed] https://launchpad.net/bugs/973386 [11:26] Cimi, andyrock: [11:26] https://launchpadlibrarian.net/99959004/bug2.png [11:26] the bottom part of the frame is missing [11:26] that is weird [11:27] seb128, Cimi I've not changed anything in the overlay [11:27] andyrock, maybe it's not vertically centered [11:27] maybe the bug is in Nux [11:27] andyrock, or does not fit in [11:27] andyrock, sorry to intrude but I was the one who submitted that bug... [11:27] well no, the screenshot is clipped [11:27] Cimi, if it doesn't fit we don't show it [11:27] ok [11:28] andyrock, it started to happened after a compiz update [11:28] Cimi, andyrock: I can confirm that on my 1920x1080 screen, it's not a space issue [11:28] andyrock, krnekhelesh I can reproduce === om26er__ is now known as om26er-away [11:28] I can reproduce too === om26er-away is now known as om26er [11:28] Cimi, yeah it is not a space issue...the bottom is clipped on all resoultion monitors [11:28] krnekhelesh, thanks for filing the bug [11:29] krnekhelesh, I have serious difficulties in spelling your nickname :D need to teach me how at the UDS :) [11:29] Cimi, seb128 ok i'll fix the lens bar bug and then the overlay one [11:29] Cimi, I keep changing my nickname to nik90...but it keeps changing back :P [11:29] andyrock, you rock, thanks ;-) [11:29] np [11:29] lunch time :) [11:29] seb128, only didrocks really rocks [11:29] ahah [11:29] hahaha === andyrock is now known as andyrock|lunch [11:31] andyrock|lunch, I've opened bug #977961 for the lens clicking issue for tracking [11:31] Launchpad bug 977961 in unity (Ubuntu) "lens selection should work when clicking in the rectangle outside of the icon" [Undecided,New] https://launchpad.net/bugs/977961 [11:31] Cimi: well said! :) [11:32] Cimi, ^ [11:34] seb128, can you give the bug low or medium priority? [11:34] krnekhelesh, which one? yours? [11:35] seb128, no the bug you filed just now [11:35] krnekhelesh, ok, put it medium, though it might be high [11:35] bug #977961 [11:35] Launchpad bug 977961 in unity (Ubuntu) "lens selection should work when clicking in the rectangle outside of the icon" [Medium,Confirmed] https://launchpad.net/bugs/977961 [11:36] it's very disturbing from an user perspective, it's one of those details that make you think the system isn't responsive or is buggy [11:40] yeah...I was a bit surprised to see it happen [11:41] though I am personally so happy to see so many multimonitor bugs being fixed in Unity 5.10 [12:16] Cimi: I redid the ubuntu-docs screenshots after all with the new default wallpaper [12:20] jbicha, wohoo [12:20] jbicha, awesome [12:20] Cimi: I wrote a script to mostly automate taking the screenshots so it's a bit easier to re-do them or to translate them [12:21] jbicha, on a different task, I don't need updated screenshots for this, but I'd like to have your +1 if possible [12:21] https://bugs.launchpad.net/ayatana-design/+bug/977974 [12:21] Ubuntu bug 977974 in notify-osd (Ubuntu) "Improve background colorization" [High,In progress] [12:21] oh just seen [12:21] thanks jbicha === MacSlow|lunch is now known as MacSlow === andyrock|lunch is now known as andyrock === _salem is now known as salem_ === yofel_ is now known as yofel [13:41] seb128, could you give lp:~mhr3/unity-lens-music/rb-parse-tdb a try? [13:41] mhr3, sure [13:50] mhr3, not perfect but much better! [13:51] seb128, what's missing? [13:52] mhr3, I think I need to play the songs once in rb for the artwork to be listed [13:52] mhr3, so half my songs have no artwork [13:52] so it seems to fix itself by playing stuff and restarting the lens [13:53] seb128, sucks that restarting the lens is needed though :/ [13:53] yes [13:53] mhr3, well sucks to start that I need to play stuff to have their artwork in the dash :p [13:54] like new install, plug your ipod, import your music -> not artwork [13:54] no [13:55] seb128, i dont think we're trying to turn the dash into a music player that fetches artwork for everything you have, generates thumbnail etc. [13:56] mhr3, ok, well then consider that vcs of yours to fix the issue ;-) [13:56] mhr3, it's stable and works great for stuff I played [13:56] mhr3, you would get a bonus point if it was picking new artwork without a lens restart though ;-) [14:04] seb128, that reminds me, rb is stupid and it updates its xml only on close, so your ipod use case wouldn't work properly anyway [14:05] seb128, we should push for some more sane data store [14:05] mhr3, oh ok, well get your merge request in and I'm happy ;-) [14:05] xml... ffs [14:06] mhr3, so the .vapi file is what generates the vala bindings, right? [14:06] yes [14:07] Ok, I'm trying to understand what you did. [14:07] also the .vapi *is* the binding [15:14] mhall119: hey [15:14] gotwig: hey, congrats on membership! [15:14] mhall119: thanks ^^ === Pici is now known as Guest42174 === Pici` is now known as Pici [16:37] greyback: ping. [16:38] hi all! is there already a bug report about an invisible launcher popping up when (re)loading flash content (i think it's labeled "exe")? short clip: http://www.youtube.com/watch?v=pqQpOL61Kwk [16:47] seb128, https://code.launchpad.net/~andyrock/unity/fix-977961/+merge/101397 [16:48] andyrock: \o/ thanks [16:48] yw [16:53] andyrock: hey, while you are working on small issues, do you work on the overlay being cut off? [16:53] andyrock: the shortcut dialog is cut of at the bottom [16:53] elopio: pong [16:53] didrocks, i'm working on it right now but I've no idea why we have this regression [16:54] ok [16:55] greyback: hey. I'm copying a lot of things from your tests. But testhelper.rb has no license. Can I copy it without issues? === krnekhelesh is now known as nik90 [16:59] elopio: you can yes, it's made of 2 chunks of ruby code I found. There are links to original sources, both of which are made available with a CC-type licence [17:00] a few changes were made to suit our needs [17:01] andyrock, nice, thanks! [17:05] greyback: great. Also, is there a way to use the same startup and teardown for tests in separate files? I would like all my tests to open the u1 control panel and log in. But as a ruby newbie, I don't know how to do it. [17:06] greyback: in fact, I've just understood half of your code. But, now I have things running: http://ur1.ca/8z54h :) [18:22] elopio: hey, sorry but I had to go afk. [18:23] elopio: right now we use 'startup' and 'shutdown' methods, which run before and after (resp.) the whole test suite [18:24] elopio: to have them in separate files, hmmm, certainly possible, just need to figure out how [18:24] elopio: I'm no Ruby guru either [18:45] greyback: well, if I find something I'll let you know. In the meantime, I'll be putting everythin in the same file. [18:45] greyback: just another question. Is tdriver_visualizer working for you on precise? I get Aborted (core dumped) [18:53] elopio: curses, it's crashing on me too [18:53] elopio: I'll push a fix to the ppa [18:54] :) thanks greyback. [18:54] elopio: yw [18:55] elopio: It'll become an official package in Q-series [19:26] thomi, ping, hey! Im on now [19:46] bschaefer: uhhh, I forgot what it was. [19:46] what timezone are you in again? [19:46] umm 21 hours behind you :) [19:46] -8 [19:46] oh ok [19:47] I'm still not used to Americans going to sleep while I'm eating my breakfast [19:47] It was about 11:30pm my time, which im on sometimes [19:48] It's only 12:48 pm here now, so if Im going to sleep when your eating breakfast then that would be odd [19:50] thomi, o yeah I wanted to ask you how about using valgrind. [19:50] so you just have that alias and type "valgrind-unity" [19:50] yup [19:51] I ran the autopilot tests after that get as many things used as possible [19:51] bschaefer: for me the autopilot tests don't run since unity is too slow [19:51] so I do things manually [19:51] open the dash, hud, switcher etcx [19:51] hmm nothing really slows down when Im running valgrind [19:51] that could be a problem, but I know my .log file is getting updated [19:51] bschaefer: ok, you're doing it wrong then :) [19:52] I have a blinging laptop, and it slows to a crawl for me [19:52] I suspect you're running valgrind over the python script, not the compiz child [19:52] o and im on a VM sooo yeah. Im guessing valgrind might be crashing [19:52] can you paste your valgrind alias again? [19:52] yeah [19:52] alias valgrind-unity='G_SLICE=always-malloc G_DEBUG=gc-friendly valgrind --t ool=memcheck --num-callers=50 --leak-check=full --track-origins=yes --log-fi le=unity-valgrind.20120409T160245.txt /opt/unity/bin/unity --reset 2>&1 | te e ~/logs/unity-valgrind.20120409T160245.log' [19:53] thomi, that is what it seems like! [19:53] ahhh, I see the issue [19:53] You need this instead: [19:53] alias valgrind-unity='G_SLICE=always-malloc G_DEBUG=gc-friendly valgrind --tool=memcheck --num-callers=50 --leak-check=full --track-origins=yes --xml=yes --xml-file=unity-valgrind.`date +%Y%m%dT%H%M%S`.xml compiz --replace 2>&1 | tee /home/thomi/unity-valgrind.`date +%Y%m%dT%H%M%S`.log' [19:53] obviously change the paths [19:53] I'm running compiz, you're running unity. [19:53] ooo [19:53] unity is just a python script that restarts compiz. [19:54] I changed it because I had been using that command [19:54] There may be an option to tell memcheck to trach child processes... [19:54] duh! [19:54] bschaefer: at some point (maybe @ UDS) I'd love to get you to show me how to run Unity3d inside a VM. I've never been able to get it working [19:55] That would be awesome, I have to talk to thumper about that [19:56] o and Ill be getting my laptop today, sooo no more VM yay [19:57] sweet! [19:57] Thanks for your help! Now hopefully I can fix something :) [19:58] you seemed to have fixed a lot of things yesterday haha [20:09] bschaefer: yeah, there's plenty more [20:31] thomi: \o/ thanks for fixing so many leaks! [20:32] htorque: uuhhh, you're welcome. [20:32] there's a few more left, I hope to get them today. [20:32] ...unless bschaefer beats me to it ;) [20:32] :-) [20:33] I have the timezone advantage though, so I think I'm safe [20:36] thomi, haha I hope, I just had to go get some coffee [20:36] plus I have class later! [20:37] thomi, o I also stole your alias and will learn how to parse xml with valkyrie [20:37] bschaefer: just open valkyrie, and inside the UI get it to open the XML file. [20:38] it somehow automatically finds the right source code to laod as well.... [20:39] awesome, thanks. [21:29] mhr3, for some unidentified reason, tonight I have rb updating the xml without restarting it. [21:30] davidcalle, lol, that's nice of it :) === salem_ is now known as _salem [22:39] bschaefer: still around? [22:39] thomi, yeah [22:39] man there is a lot of stuff in these logs... [22:39] could I get you to review another valgrind fix please? https://code.launchpad.net/~thomir/unity/panelmenuview-fix-invalid-read/+merge/101452 [22:39] sure! [22:39] bschaefer: yes, most of it is pango / fontconfig though [22:41] thomi, looks good [22:41] cheers [22:42] np! If you have anything else that needs review let me know [22:42] there's something similar happening with the hud controller, but I can't work it out [22:42] :-/ [22:42] I think I saw that, I was working through the uninited vars [22:43] I found that there are A LOT of new icons made in LauncherController and views and they don't seem to get delete [22:43] well the model_ it self doesn't seem to get delete but I have to figure out if it gets done some how... [22:47] bschaefer: views are usually deleted by nux [22:55] hmm well thats good, but valgrind seems to complain about some views not getting deleted