=== edwardc_ is now known as edwardc [02:47] xi patches won't make it upstream ? [03:15] hicham: What Xi patches? [04:07] RAOF, i have a complaint [04:08] seeing as your the person I complain to about x11 related crap [04:11] DBO: Complain away :) [04:12] I'd also like to complain that people don't make 13" laptops I want to buy. [04:13] Watch me: “People don't make 13" laptops that I want to buy! Grrrr!” [04:13] RAOF: the ones to XI2.h [04:14] RAOF: found a link, thanks [04:14] hicham: Oh, you mean the multitouch patches? [04:14] RAOF: yes [04:14] hicham: They're works-in-progress upstream. [04:14] RAOF: thanks, I am trying to build unity on fedora [04:15] so far so good [04:15] Yeah. Both new shells use X protocol that's not upstream! [04:15] now I think I am at the last step [04:16] alls deps built [04:16] I've mentioned overlay-scrollbar on firefox's mailing list [04:16] and I'm talking to their ux team about it [04:16] RAOF, so here is my complaint [04:16] xcb deadlocks for no obvious reason [04:17] Who is actually in the design team that is on IRC? [04:17] * hicham liked overlay-scrollbars [04:17] Omega, we have been sworn to secrecy to protect those responsible [04:18] I must hunt them down, one by one, and talk to them calmly! [04:18] DBO: ISTR an xcb deadlock that should be fixed in the new libx11 upload. [04:18] DBO: You're not doing something strange, like using threads or anything? :P [04:19] no threads [04:19] but several connections in one process [04:19] and if you fixed the xcb deadlock [04:19] But it'd really help if they could get in touch with firefox's ux team, or me, I'm in contact with them a lot. [04:19] I will gladly donate my ovaries [04:19] (I keep them in a jar) [04:20] Are you doing anything explicity prohibited? Like, say, calling any X function at all in an error handler? [04:20] no [04:20] its so weird [04:20] like sometimes we'll get a situation where calling XGrabPointer just deadlocks in xcb_wait_for_reply [04:20] but if we call XSync before then [04:20] oh man [04:20] it works [04:20] WTF??? [04:21] Funness! [04:21] yeah [04:21] having your window manager deadlock is AWESOME [04:22] RAOF, do you have any info on the deadlock fix [04:22] I would love to look [04:22] solution: call XSync before every function [04:22] sometimes that will fail though! [04:22] its *retarded* [04:22] xcb should be cast back to the hellish pit it came from [04:27] * DBO pokes RAOF [04:27] * RAOF is looking, and eating lunch. [04:27] lunch? [04:27] its 23:30 at night man [04:28] why would you be eating lunch [04:29] Hm. We've got the libx11 that the fix is in (1.4.2) [04:29] I come from the future, where it's 1:30pm [04:29] no, all times are the same [04:31] RAOF, yes but do you have more details? [04:32] * Omega is in the same spacezone as DBO [04:33] Omega, I subscribe to the flat earth theory, thus you are ALL in the same timezone [04:33] I do however believe the (flat) earth revolves around the sun [04:34] But you see, there is one thing you did not acocunt for, the sun revolves around me! [04:34] account* [04:34] Omega, blasphemy, I shall burn you ate the stake [04:34] prepare to be cooked, witch [04:34] hmmm, stake, delicious. [04:35] DBO: Oh. The first was to drop the Display lock when calling the X error handler, the other was to ignore some locks [04:35] RAOF, bug reports... [04:35] or commit logs [04:36] arghh, dbusmenu is missing from the repos [04:38] DBO: You're looking for http://cgit.freedesktop.org/xorg/lib/libX11/commit/?id=fd85aca7a616c595fc17b2520f84316a11e8906f [04:38] Ahem. Let's put that in the right channel, shall we? [04:43] DBO: Yeah, I didn't think you were using threads. That was the fix that I was thinking of, though, which clearly isn't a fix for your problem (because otherwise you wouldn't *have* your problem ☺) === _LibertyZero is now known as LibertyZero [08:02] good morning [08:34] good morning === API is now known as Guest31307 [11:45] good day al;l === MacSlow is now known as MacSlow|lunch === daker_ is now known as daker === yofel_ is now known as yfoel === yfoel is now known as yofel [12:56] http://openmatt.wordpress.com/2011/04/06/how-to-work-open/ Maybe we can learn a few things from our friends at Mozilla :) === rodrigo__ is now known as rodrigo_ [13:13] unity for opensuse - When ?? [13:16] anyone use Twitter ?? [13:19] Hello! I have a usability suggestion for unity [13:19] When i go to applications dash 4 out of 5 recently used applications are also a part of my launcher on the left [13:20] I think it would be better if apps, which are permanently on the launcher, don't appear in most commonly used === MacSlow|lunch is now known as MacSlow === jackneill is now known as Jackneill === bregma_ is now known as bregma [14:52] bug #700757 is marked as 'fixed' but I see no change in behaviour with todays bits [14:52] Launchpad bug 700757 in unity (Ubuntu Natty) "unity aborts when you plug in an external monitor" [Medium,Fix released] https://launchpad.net/bugs/700757 [14:53] plug a small monitor (1024x768) into a bog standard netbook and blammo unity dumps core, spitting out "Size too big width=2304height=1024" [14:53] kamstrup: do you have the askubuntu place installed? [14:53] i would point out that the machine ought to be about as supported a platform as exits, being an intel netbook [14:53] jcastro: I only think I have an olde copy lying around, but not running [14:54] kamstrup: I am starting to have some startup issues with lenses, but I'm not sure if it's the lens or unity [14:54] jcastro: you mean speed wise or other? [14:54] the daemons run and everything, just nothing displays, I have to go back and restart each daemon, and then restart unity [14:54] so like when I log in at first [14:54] oh, that sound sodd [14:54] my AU lens doesn't work [14:54] but apps and files does [14:55] FYI, ppa:askubuntu-tools/ppa [14:55] I don't think we're doing anything different than before [14:55] jcastro: did this start happing with a recent upgrade? [14:55] last week sometime [14:56] I don't recall exactly when [14:56] jcastro: right, the only thing that has changed recently, afaik, is that Unity will try and keep the apps place on the top spot in the dash [14:56] (which is great btw) [14:56] so it sounds like maybe a race which' result depends on whether the disk cache is warm or cold..? [14:57] ok [14:57] I'll see if I can at least find someone who can reproduce it [14:57] jcastro: kenvandine had a similar issue last week, but that was confined to a gwibber issue afair [14:58] kamstrup, i fixed my memory issues [14:58] kenvandine: zomg!?!?11one [14:58] kamstrup, at least somewhat [14:59] it wasn't Dee related... just python bloat [14:59] kenvandine: so what was the issue? [14:59] imports [14:59] wow [14:59] how to fix this bug...I can't use firefox 4 in unity becoz of this --> Bug #751067 [14:59] Launchpad bug 751067 in fglrx-installer (Ubuntu) "Colors are mangled in Natty" [Undecided,Confirmed] https://launchpad.net/bugs/751067 [14:59] importing gwibber.microblog.utils [14:59] apparently is huge [14:59] kamstrup, well, that was the virt mem [14:59] i shaved ~200M off virt [15:00] kenvandine: right... should give a slight startup boost too I expect? [15:00] yeah [15:00] RSS was doing things multiple times [15:00] which i still kind of have a problem with [15:00] but i made it much better [15:00] sweet [15:00] RSS is now about 60M [15:00] i have to stop watching the active property though [15:01] i need to figure out how to make it refresh the data on both a section change and when it becomes active [15:01] kenvandine: the "active" prop on Unity.PlaceEntryInfo? [15:01] without it doing both [15:01] yes [15:01] i was updating the data when active became true [15:01] but, if the section was also changing... it would update twice [15:02] because i also do that on section change [15:02] oh... I wonder if I do that too now you mention... [15:02] it's a problem... [15:02] yeah, there's something I don't like about the current protocol in this regard [15:02] i couldn't find a good way around it :/ [15:03] jcastro, if the AU lense starting watching for the active property to change, that could cause some problems [15:03] kenvandine: you only need to reload data on active=true if you are "dirty" [15:03] jcastro, last i looked it didn't [15:04] how can i tell if i am dirty? [15:04] that didn't sound good :) [15:04] kenvandine: lol [15:04] kenvandine: I just tricked you into saying that [15:04] ;-) [15:04] indeed [15:04] sneaky sneaky :) [15:04] kenvandine: in the files place (or is it apps) I just have a bool that I keep around [15:04] ugh [15:04] yeah :-/ [15:04] i tried that [15:05] wasn't very reliably [15:05] the active and section change props seem to change irradically [15:05] not always in the same order [15:05] oh... [15:06] i can't say that for certain [15:06] but it sure seemed that way [15:06] it was very annoying [15:10] kenvandine: yeah - you should have seen the 10.10 series... I had some *very* defensive coding in there because Unity was spamming the places :-) === hggdh_ is now known as hggdh === Cimi__ is now known as Cimi [15:12] kamstrup, unrelated question [15:13] is there any way to sort the data in a model? [15:13] or do i have to clear and populate it again? [15:14] kenvandine: you have to clear and populate I am afraid [15:14] ok [15:14] would also be a nice feature :) [15:14] kenvandine: yeah... sorting is a bit tricky for wire data like this [15:14] yeah [15:15] kenvandine: possibly we could use the Dee.Filter API to define some sort of sorting API [15:15] it would be nice to change the order in the view side [15:16] so perhaps apply a filter to the model on the client side [15:16] kenvandine: dee already ships with a collating filter that sorts on a string column according to the locale rules [15:16] ? [15:16] kenvandine: exactly [15:17] kenvandine: maybe dee could ship a set of "sorting rules" (implemented as DeeFilters) with canonical names and then you could remotely tell a DeeSharedModel which sorting it should apply or something [15:17] yeah [15:17] that might do it [15:18] i want to be able to sort by time [15:18] what priority should bugs wherein unity core dumps be [15:18] kenvandine: in DBus ala Sort(in u col_index, in s sort_name) [15:19] kamstrup, common problem for gwibber, the time column will come from twitter, facebook, etc... but will be appended as the results come from each service [15:19] apw: depends on how user visible it is [15:19] happens when you plug in an external monitor on intel atom netbooks [15:20] should make UDS fun [15:20] kenvandine: right, that works well if you put each sub-resultset in a group, but if you put them under the same group you currently have to collect all results on the daemon side, sort them, and then add them [15:21] apw: hehe, have you checked for dupes? I think I recal a report ala what you mention... then again, maybe my mind is just still in weekend mode :-) [15:22] kamstrup, yep, that was mine, which was closed fix released :/ but its still there exatly the same [15:22] starting to think the unity team needs to test h/w which isn't apple [15:23] perhaps we can get somthing really cheap for yas [15:23] apw: lol, I knew I saw it somewhere :-D only few use apple products for hacking on dx [15:23] :) [15:24] but i think if we get to UDS an noones netbooks will let a projector plug in ... we are going to get a slapping! [15:24] apw: but if you still see it, just reopen and state the case. There are quite a few DXers with multimonitor setups so rest assuered it's a priority :-) It's just that there are *many* priorities right now :-) [15:24] heh of that i am very sure :) [15:25] apw: hehe, if it comes down to that on UDS I am pretty sure someone will sit down and write the patch on the spot :-) [15:25] kamstrup, the biggest problem i have with reordering the data in the daemon and adding it again is all the row-added signals... lots of spam and then the client gets lots of churn [15:26] kenvandine: as long as you do the workout without entering the mainloop dee will batch all the transformations into one Commit() on the bus [15:26] kenvandine: so the row-added signals is only a local thing in each process [15:26] ok [15:26] not so bad i guess [15:26] still a whole new set of data [15:27] yes :-/ [15:27] ok [15:27] is this something that could happen for oneiric? [15:27] kenvandine: ok, I am adding this to my list of things in the Unity API we need to fix [15:27] or should i not get my hopes up? [15:29] kenvandine: oh wait... [15:29] kenvandine: why can't you just do a sorted model.insert()? [15:30] on the daemon side? [15:30] kenvandine: yes [15:30] i'd have to get the iter to insert after right? [15:30] * kenvandine hadn't even seen the insert api... just prepend/append [15:30] kenvandine: that is most efficient I guess, but you can also insert by offset [15:31] * kenvandine doesn't like offsets [15:31] i also haven't looked at how you do lookups [15:32] kenvandine: you can either just traverse the row iters until you find what you need, or you can use Dee.FilterModels or Dee.Indexes === zyga is now known as zyga-food [15:33] depends a bit what you need [15:33] kamstrup, guess i'll have to play around... that could be tricky if i want to insert it in the right place based on timestamp [15:33] or ticks [15:33] iterate over them all until it is no long < [15:33] then insert after -1 [15:33] i guess [15:34] kenvandine: yeah [15:34] kamstrup, should that be fast? [15:35] kenvandine: well, fairly fast, but you're still doing linear inserts of course, but not too bad unless your models grow very big [15:35] define "very" [15:35] ? [15:36] 100s of thousands of rows? [15:36] or millions? [15:36] kenvandine: i guess "very" depends on target hardware and your latency and cpu time restrictions [15:37] kenvandine: but a good guess would be below 500 don't worry [15:37] it'll always be over that [15:37] and then 500 < N < 5k, test it [15:37] and then > 5k worry [15:39] kenvandine: it was a long standing plan that one could use Dee.Index for this, but I just never got around to it this cycle [15:39] understand [15:46] om26er, cdbs good day [15:46] lamalex, hey there. [15:47] lamalex: Howdy! good morning [15:48] om26er, cdbs can either of you get this to happen? https://bugs.launchpad.net/unity/+bug/749940 [15:48] Ubuntu bug 749940 in unity (Ubuntu) "Unity application launcher does not hide when firefox is full screen and desktop is switched" [Low,New] [15:49] * cdbs checksz [15:50] lamalex, dodge active window is not the default option btw, so could be specific to it only [15:50] as the reporter is using it [15:50] yeah i know i changed my hide behavior === jono_ is now known as jono [15:51] lamalex: I managed to reproduce it [15:51] lamalex: though in a slightly different way [15:53] * cdbs comments [15:54] cdbs, and that is..? [15:54] lamalex: see my comment, can you reproduce that way? [15:57] does anyone know if there is any way to ask banshee to exit once it is started from the indicator menu ? [15:57] cdbs, got it [15:57] thanks [15:57] apw, i think mpt just filed a bug on that [15:57] and banshee is the default now right, i can remove rhythm box ? [15:58] apw: yes you can [15:59] You could remove banshee too if you don't like it. [16:05] Oh, is that a list of upgrades? :( [16:05] * cdbs 'll have to wait for Unity to upgrade before he could get down to bugs [16:09] sladen, is the scrollbar supposed to look different in gnome-terminal now? [16:10] sladen, the changelog says add dark theming to tweak around resize grip === zyga-food is now known as zyga-grocery [16:10] kenvandine, it's black now [16:10] but it seems pretty weird to make it look different than everything else [16:10] when _does_ the day in the date time widget get updated? it seem to stay as the day it was when i suspended until about midday and then change [16:11] It doesn't, because the whole gnome-terminal is black & purple, except for the white-silver scrollbar. [16:11] But I wish they used black instead of silver in quite a few other apps, like they were testing in maverick [16:13] kenvandine: I asked Cimi about this; it's something to do with the Gtk+ patch not picking it up because it's not a GtkScrolledWindow [16:13] kenvandine: and not actually related to the theme AFAIK [16:14] humm, weird that it just changed in today's upload [16:16] JohnLea, know if this was fixed? https://bugs.launchpad.net/unity/+bug/750471 [16:16] Ubuntu bug 750471 in unity (Ubuntu) "Ensure Unity prioritises icon loading from Unity-icon-theme" [Undecided,New] [16:24] smspillaz|z, what exactly do you mean by 'terminal output' ? === davidcalle_ is now known as davidcalle [16:54] gawd, has anyone tried actually moving windows from one monitor to another in unity? half the time it complely mangles the window [16:55] apw, I did when I had it i nstalled on this system with dual monitors [16:55] apw, I have nvidia [16:55] apw, it worked fine [16:56] apw, although I have it installed on single monitor system now [16:56] hmmm, not very well on intel graphics [16:59] lamalex: Please reopen the Unity bug task and set it to Triaged in bug #726033 . [16:59] Launchpad bug 726033 in unity (Ubuntu) "does not display icons until hovered" [High,Triaged] https://launchpad.net/bugs/726033 [16:59] apw, ah I am not up on intel issues ,, although I did hear there was a bug that popped up erors every 10 seconds about the cpu temp even though it wasnt hot and that's all I know about it [17:05] cdbs, done [17:05] lamalex: thanks === zyga-grocery is now known as zyga [17:12] coz_, ok i can reproduce the same kind of 'wipping' behaviour by grabbing the title bar of a window and ramming it onto the title bar ... then sliding left and right, the window decorations dissappear and the window gets erased at the edges [17:12] any unity devs in here ? [17:13] apw, yikes,, that sounds painful to visualize [17:13] hicham, I believe this is the place,,, not sure who is actually at their systems however [17:14] coz_ : I am just trying to figure out how to override some default stuff when building/installing [17:14] coz_, yeah grab a non-full screen window and slide up till you hit the top of the screen; as the cursor enters the menu bar the decorations go and then slide left and right (all holding the click) and you'll lose the edges of the window [17:14] hicham, I would stick around here and also log onto #ubuntu+1 [17:15] normal service is restored when the cursor exits the global menu panel [17:15] apw, wait ,, moving a window to the sides or center or bottom edge might be kicking in the windows like thing where they maximize in a certain way [17:16] coz_, yeah the action would trigger maximise when i let go, but as i haven't done that, but during the time the cursor is in the menu bar the main window is not updated correctly [17:17] coz_: thanks [17:17] apw, ooo mm you may want to speak with smspillaz|z about this then [17:18] * apw files a bug [17:22] bug #757714 [17:22] Launchpad bug 757714 in unity (Ubuntu) "when moving a window and the cursor enters the global menu the window is no longer updated correctly" [Undecided,New] https://launchpad.net/bugs/757714 [17:25] i assume we know that none of libre office is using the global menubar ? [17:28] apw, there's a package available. [17:28] But not default yet because of bugs. [17:29] ok will ignore that [17:42] does anyone know if we expect the main text window in xchat to not get the new menubar [17:42] menubar? the menubar is showing up in the panel in here, as expected. [17:43] If you're talking about the scrollbar, it isn't working in a lot of places. [17:44] i maent scrollbar, doh [17:44] ok# === API is now known as Guest2711 === daker is now known as daker_ [18:37] lamalex, Hey! I have not done triage today, been helping Sam debug the invisible window thingy. Still on. [19:02] I need a Unity Expert [19:03] tremolux, you here? [19:04] ?? [19:04] DBO: yep, heya, what's up? [19:04] sure [19:04] whats up pasjr [19:04] tremolux, sorry I meant to ping someone whose not here and you got tab completed [19:04] my mistake [19:04] ah, no problemo :) [19:05] I have always been good a tweaking Ubuntu, but I can not find the file to change the colors for the launcher bar in 11.04 can you help? [19:06] you cant change the launcher bar colors without recompiling pasjr [19:06] its hardcoded [19:06] I understand that but there is a file with that info is there not? [19:08] afaik, the color is picked up from the icon itself [19:08] kamstrup, clearing the model in python isn't freeing memory [19:09] ??? [19:09] pasjr, sorry [19:09] pasjr, the colors for the launcher background are in Launcher.cpp [19:10] Thanks [19:10] launcher icon colors are determined in LauncherIcon.cpp [19:11] Any clue what folder that file is in? My system cant find it. [19:21] om26er, cdbs when you guys triage autohide bugs, could you tag them with autohide? [19:21] ai ai sir [19:21] thanks [19:21] i've got a hunch that autohide is one of our buggiest areas [19:22] want to see how many autohide related bugs are just in the new queue [19:24] cdbs, are you triaging? [19:26] lamalex, seems like a problem with qt drag n drop has rose [19:30] launcher does not hide after that [19:37] good evening [19:39] after an update (few mns ago) the launcher looks empty but it is still working icons are not visible > http://i.imgur.com/AHsPn.png < i hope this is already known [20:09] kenvandine: I just plugged a pretty hefty leak in Dee, https://code.launchpad.net/~kamstrup/dee/commit-transaction-leaks/+merge/57219, but I don't think that's what you are seeing since this one was only on the receiving end... [20:10] i can check though [20:10] if i remove the home icon on the launcher, how do I get it back? [20:11] kamstrup, basically everytime i do a clear and append the memory usage grows [20:11] kamstrup, and it never goes down [20:11] append come after clear normally? i'd think you'd start fresh after a clear? [20:12] kenvandine: the most likely problem is that pygi somehow keeps a ref around on the gvariants stored in the row [20:13] bcurtiswx, basically empty all the data then append [20:13] kenvandine: as this is definitely not something we see in C. I just did a ~30min torture run adding 1000 rows and calling clear() each second [20:13] no leaks at all [20:13] kamstrup, ok [20:13] * kamstrup gotta run [20:14] so my original question, if i remove the home icon from launcher, how do I go about getting that back ? [20:16] bcurtiswx, maybe via dconf-editor ? [20:18] If I wanted to try to fix bug 740515, where should I start looking? [20:18] Launchpad bug 740515 in unity (Ubuntu) "Window controls in Unity panel do not extend to top of display" [Low,Confirmed] https://launchpad.net/bugs/740515 [20:24] hmm, i don't have dconf-editor nor dconf [20:24] install dconf-tools [20:28] bcurtiswx, you can either reset the launcher favorite list using 'unity --reset-icons' or use dconf-editor to add 'user-home.desktop' to the favorite list (desktop → unity → launcher → favorites) [20:29] htorque, zniavre much thanks :) [20:29] both suboptimal solutions imo [20:29] yw :-) [20:30] htorque, good evening do yu ahv any idea about my worrie > http://i.imgur.com/AHsPn.png < icons are missing in the launcher [20:31] jcastro, the au lens also seems to suffer from the same leak [20:31] you hav * [20:31] but on a much lower scale [20:31] smaller data set [20:32] no clue how to track down memory leaks in GI [20:32] bcurtiswx: sorry, the .desktop file is nautilus-home.desktop (user-home is the icon :)) [20:33] kenvandine: I can't even get my lens to run on login [20:33] only when I run it by hand [20:33] really? [20:33] weird [20:33] working here === htorque_ is now known as htorque [20:34] kenvandine: with the ppa and everything up to date? [20:34] no luck with unity [20:34] everything is built but when enabling unity from ccsm it crashes [20:34] any debugging hints ? [20:35] jcastro, yes [20:35] well [20:35] i was running an old version actually [20:35] i just updated it and killed the daemon [20:35] unity restarted it automatically [20:35] which daemon ? [20:35] ok [20:35] next time you login fresh [20:35] try the lens [20:35] and tell me if it loads [20:36] hicham, the askubuntu lens [20:36] jcastro, i'll try in a moment [20:37] kenvandine: it's been killing me all week, lmk if you find anything [20:38] jcastro, reproduced [20:38] i had the same problem in gwibber [20:38] let me try to fix this for you :) [20:38] oooh [20:38] I hope you have a fix. :) [20:38] i fixed it a couple weeks again in the gwibber lens [20:41] * hicham was hoping for some debugging help :( [20:42] hicham: there is a bug we know about, it's very demanding for fixes all the over the place, we'll try to tackle it for next release [20:42] jcastro: you should help me champion bug 740515. :) [20:42] Launchpad bug 740515 in unity (Ubuntu) "Window controls in Unity panel do not extend to top of display" [Low,Confirmed] https://launchpad.net/bugs/740515 [20:43] whoa, never noticed that before [20:43] nice catch! [20:43] yay [20:44] jcastro: it's slowly been driving me crazy. :) [20:54] bug 757990 [20:54] Launchpad bug 757990 in unity (Ubuntu) "missing but working icons in unity-launcher" [Undecided,New] https://launchpad.net/bugs/757990 [20:56] kenvandine: any luck? [20:56] not yet [20:57] looking through my commits, it looks like my fix was to make initiallizing the gwibber connection handle errors [20:57] i guess the problem was the daemon didn't start quick enough [20:57] in the case of failures [20:58]