[00:22] robert_ancell: do you know why there are no armhf ddebs for lightdm after 1.2.1? [00:22] http://ddebs.ubuntu.com/pool/main/l/lightdm/ [00:22] mfisch, no, they're done server side though [00:23] I wonder if xnox knows [00:24] randomly picking another one, lsof, and no armhf except for from 2010 [00:25] RAOF, I emailed you the list of experimental/update drivers awaiting sru processing. all should be basically wave-thru's, and all are relatively urgent due to valve. [00:26] bryce: Excellent, thanks. The decaying husks at the end of the processing queue were getting depressing :) [00:26] RAOF, there are also several xserver srus that have been waiting a while. But less urgent. [00:26] RAOF: Hi Chris! Since you asked, I'd say that the mail-notification item in the Quantal queue is urgent, since many users are disinclined to use the app without SSL working. [00:48] can i use tray icon in 12.10 gnome fallback? [00:50] i'm trying to show how upstream ibus 1.4 tray icon work [00:52] maxiaojun: you may need to add the system tray to the panel, and the ibus packaging may disable the ibus tray icon [00:53] i'm using upstream tarball directly, thank you for your advice, let me check the panel [00:55] right-click the panel has no action, it's not gnome 2 :-( [00:56] maxiaojun: you need to hold down the Alt key (maybe Super+Alt) while right-clicking the panel [00:58] thank you very much [00:59] mfisch: hmm? [00:59] xnox: the armhf ddebs seem to be out of date [01:01] xnox: for example, the ones for lightdm are a couple revs old: http://ddebs.ubuntu.com/pool/main/l/lightdm/ [01:01] From the build log: dpkg-deb: building package `lightdm-dbgsym' in `../lightdm-dbgsym_1.4.0-0ubuntu2_armhf.ddeb' [01:01] https://launchpadlibrarian.net/119223399/buildlog_ubuntu-quantal-armhf.lightdm_1.4.0-0ubuntu2_BUILDING.txt.gz [01:01] so they were built. [01:02] yes, there is a mystery here [01:02] pitti: why did lightdm-dbsym package from ^^^ buildlog did not get uploaded / published on ddebs?! [01:02] lsof, randomly picked, has similar [01:30] done video demo vanilla ibus: http://dl.dropbox.com/u/45139465/vanilla-ibus.ogv [01:31] i demoed two mode, the first mode don't work in ibus indicator, second mode (most user including me currently using) is dropped in 1.4.99+ [01:33] jbicha: shipping a PPA is a TB problem, not a release team problem [01:34] RAOF: i think i have one in queue about libgpod. [01:37] micahg: what is TB sorry? [01:37] maxiaojun: Tech Board [01:42] maxiaojun: a bug report would have been a lot clearer; after watching your screencast I still don't understand what feature the indicator is missing [01:43] you can try indicator if you have a box nearby [01:43] i can record another video for indicator case, if you need [01:44] maxiaojun: I don't know if I need another video, could you just explain what the indicator can't do? [01:45] you cannot see the input method context menu, if you are using indicator and select 'Embedded in menu' (default) in ibus-setup [01:47] what context menu can't you see? and which specific input method are you using? [01:48] can you recognize that i'm using ibus-pinyin in the video? [01:49] maxiaojun: perhaps you'd be better served by taking screenshots of what specifically you like (or dislike?), draw arrows pointing to something specific, and explain what about that thing you like or dislike [01:50] maxiaojun: I just spent 3:24 watching someone typing on their machine without any idea why I watched it or what I was supposed to notice. :) [01:51] please wait for my next video, i'm sorry that i'm not savvy in video tutorial business [01:51] it _was_ fun seeing how ideograms were presented when the latin equivs were presented, it seems a _lot_ nicer than the old wapros I used to use, but I couldn't tell if you liked or disliked them. :) hehe [01:51] maxiaojun: ok, I think I finally understand what you're saying [01:51] maxiaojun: the nice thing with screenshots + text is that it can be read in seconds, not watched in minutes.. hehe. [01:51] ibus-pinyin has its own preferences menu (not even just its own preferences button) which isn't exposed in Unity unless you change that option to "Always" [01:52] the text to describe the pictures wasn't there in the 'always' mode -- is _that_ what was missing? [01:52] anyway, if you can file a bug or find an already reported bug we can use that for tracking this release cycle [01:53] When Active is very similar to Always (both dropped in 1.4.99+), but some people find that panel may disappear using 'When Active' [01:54] jbicha: i will do that [01:54] sarnold: i guess you don't understand the whole thing yet, i'm sorry for my poor video, :-( [01:55] maxiaojun: that's why I'd like screenshots with the important bits highlighted :D [02:07] my machine is encoding my second video, hopefully you can understand this time [02:08] thanks :) [02:12] http://dl.dropbox.com/u/45139465/ibus-indicator.ogv [02:14] I agree with sarnold, a few screenshots and some text explanation would be easier to understand [02:15] let me do that later [02:16] maxiaojun: definitely better, I think I almost get it. :) [02:16] ok, i'd leave my computer a while and find a time to report/reuse bug report later [02:17] have fun! :) [02:51] jasoncwarner_, still downloading from steam? [03:05] robert_ancell: heh, finally finished! [03:06] jasoncwarner_, and then the transformed blew up? Did you overload it ;P [03:06] robert_ancell: :) theyt just can't handle the kind of speeds I need ;) Or bandwidth! the conversation at telstra when I said I needed more than 50g a month...classic [03:07] they just doubled our cap to 100G. No more problems round here [03:09] * RAOF sometimes hits the throttling at 150G here; and the Internode Ubuntu mirror doesn't count against that! [03:10] how do you gusy even hit your bandwidth caps ... [03:10] *guys [03:10] * smspillaz is on 50/50 and never goes above 10 [03:20] jasoncwarner_: Afaik thats the problem with business connections, if you were on a consumer connection you would have more options. [03:21] Quota wise at least. [03:34] smspillaz...NFL games. I stream those things after the kids go to bed. Gotta keep up with my football ;) [03:35] TheMuso: yeah, I heard that as well, though the only real connection I have here is business. I already looked in Victoria. Connections there are much better than where I am here. Looking forward to those 50-100meg down and 20 meg up connections again ;) [03:36] the world changed in the 2 years I was here. town had just gotten 3g when I moved in...seems rest of the world went all 4g and LTE on me ;) I won't know what to do with all that speed! [03:36] ROAF: Isn't internode and iinet one in the same now? And the iiNet Ubuntu mirror is in the iiNet freezone [03:37] ROAF: Sorry, misinterpreted your comment. Ignore that [03:50] jasoncwarner_: The downside of living in regional Australia. [03:50] Or should I say, one of the downsides. [03:51] * ajmitch accidentally left debmirror running during the day, currently got 2.5GB left until he hits cap [03:51] In australia [03:51] Have much insects? [03:52] not many, the snakes & dropbears eat most of them [03:52] :) [03:52] Ow snakes? [03:55] ajmitch: you... you... You know what I'm talking about. === attente is now known as attente_zzz [05:39] xnox: the ddeb generation process unfortunately fails over every so often :-( it's a shared machine with sometimes a high load, sometimes the apache servers on the buildds are broken, etc. [05:39] Good morning [06:14] good morning [06:21] Good Morning. [06:22] morning didrocks :) [06:22] hey smspillaz, how are you? ;) [06:22] hey BigWhale [06:22] good good. I've finally finished my exams, so I'm just chilling out [06:23] oh excellent! enjoy some quiet time :) [06:23] yeah :) you'll see me around [06:23] great! [06:24] hello didrocks! [06:28] Will Python3 come on the installation CD? [06:28] in 13.04? [06:43] BigWhale: it's already on the CD [07:21] didrocks, excellent! [07:22] I guess I'll jump the ship and make Kazam python3 only ... [07:24] BigWhale: sounds good! Good luck :) [07:25] didrocks, yeah, I ported it last week, so now it needs some testing. [07:25] Hmm, where can I check what goes on CD? [07:31] BigWhale: look at the manifest: http://cdimage.ubuntu.com/daily-live/current/raring-desktop-amd64.manifest [07:32] those are the packages installed on the live system [07:32] python3-xdg is missing [07:32] :'( [07:34] I'll have to add this to the dependencies [08:23] pitti, ping? [08:23] hey mhr3 [08:23] pitti, morning, i'm having an odd issue with pygi, see http://paste.ubuntu.com/1372808/ [08:24] any ideas why the column param isn't properly "out"? === agateau_ is now known as agateau [08:26] (tried with both caller- and callee-allocates, neither works) [08:26] mhr3: no off-hand idea, no; the gir looks fine to me [08:27] * pitti looks for a similar case [08:27] maybe the tuple returns are broken for virtual methods? [08:27] it's not even getting that far [08:28] it's already breaking in the marshalling to C, i. e. before even calling the method [08:28] i meant that perhaps because it's a virtual method and not a normal one breaks the marshalling [08:29] right, that could be the difference [08:29] in the tests we have plenty of caller-allocates="0" transfer-ownership="full" (out) int arguments [08:29] and we also have caller-allocates="1" / transfer none args [08:29] (out args, I mean) [08:29] virtuals? [08:29] so at first sight it looks like the virtuals indeed [08:30] * mhr3 would like a fix in R :) [08:30] yep, I can work on that [08:30] would you mind filing a bug about it? [08:30] awesome, thx [08:30] sure [08:32] we also already have and call a virtual-method method_int8_out() with a caller-allocates="0"/transfer-ownership="full", that works as well [08:32] so it shouldn't be too hard to fix [08:33] the caller-allocates=1 was a test, =0 was default [08:34] right, so if it's not that, I'll try some other variations [08:37] pitti, https://bugzilla.gnome.org/show_bug.cgi?id=688783 [08:37] Gnome bug 688783 in general "Marshalling of out parameters not working properly" [Normal,Unconfirmed] [08:38] mhr3: thanks [08:40] good morning everyone [08:40] hey chrisccoulson [08:41] hi pitti, how are you? [08:41] chrisccoulson: quite fine, thanks! [09:13] pitti: =( about ddebs === Sweetsha1k is now known as Sweetshark === vrruiz_ is now known as rvr_ === tkamppeter_ is now known as tkamppeter === duflu is now known as duflu|dinner [10:17] hrhr, the server of the pirate pary freiburg was burned down by their PR wrt to migrating away from openoffice [10:24] to? [10:24] Tm_T: s/to// [10:24] Tm_T: they had to backup on pastebin [10:25] somebody burned down a server because they wanted to stop using openoffice? [10:27] migrating from openoffice to ... === Quintasan_ is now known as Quintasan [11:04] back [11:18] hi, what's will happen to gnome classic in ubuntu when upstream drops it? [11:21] cyphermox: are there any common pitfalls for a application in userspace wrt suspend/resume with bluetooth?\ [11:35] bdrung, the code will not go away, it basically means GNOME stop adding feature there or maintaining it [11:36] bdrung, so it will stay the way it is, or improve through new upstream if somebody out of GNOME itself steps up to do the work [11:47] seb128: so we will keep it in ubuntu and it will be an option in 14.04 [12:04] mhr3: hm, I have some trouble with reproducing this; I followed up in https://bugzilla.gnome.org/show_bug.cgi?id=688783 [12:04] Gnome bug 688783 in general "Marshalling of out parameters not working properly" [Normal,Needinfo] [12:05] pitti, it's in lp:~mhr3/dee/easy-hints === tiagohillebrandt is now known as tiagoscd [12:18] pitti, perhaps try with guint param instead of gint8? otherwise i have no idea why it'd work in tests for not in dee [12:19] but not in dee* [12:19] mhr3: currently looking at it [12:19] mhr3: the annotations as they are in the branch are wrong anyway, but I'll play with some variations [12:19] now I need to find a minimal piece of code which can call this [12:19] do you have one? [12:20] bdrung, dunno what will be Ubuntu in 3 cycles so it's hard to say, but no reason to drop it as long as it keeps working [12:21] bdrung, so likely it will still be an option but who knows, xorg might be replaced by wayland and gnome-panel not working with it at some point or similar [12:21] pitti, this should do - http://paste.ubuntu.com/1374697/ [12:21] bdrung, just making that one up, bottom line is "hard to say what will happen over 3 cycles" [12:22] okay, but "keeping it as long as possible" is better than just dropping it in the next release :) [12:22] yeah [12:22] mhr3: ah, get it, thanks [12:24] pitti, should have been get_field_schema, but i guess you noticed :) [12:25] mhr3: as it doesn't even get that far, it doesn't matter yet :) [12:25] actually [12:25] * mhr3 should have tried it first :) [12:25] TypeError: set_schema_full() takes exactly 2 arguments (3 given) [12:25] m.set_schema("s", "u", "a{sv}") [12:25] this is even before [12:26] pitti, it's not using the overrides from the branch [12:26] no, it's not [12:26] then use set_schema_full(["s", "u", "a{sv}"]) [12:27] same for set_column_names [12:28] mhr3: hmm: http://paste.ubuntu.com/1374709/ [12:29] wrong .so [12:29] ah [12:29] ok, got it [12:30] mhr3: thanks, I'll take it from here (after lunch) [12:30] pitti, glad you're able to rep now ;) === Chucrute301 is now known as sabdfI [12:55] Hi guys :) === gatox is now known as gatox_lunch [12:56] hey sabdfI, how are you? === sabdfI is now known as Chucrute301 [12:56] Iam me [12:56] :) === Chucrute301 is now known as sabdfI [12:58] Sabdfl is mark ? [12:59] sabdfI: of course [12:59] note the "I" instead of "l" [13:00] Yeap === sabdfI is now known as Chucrute301 [13:01] much better, lets avoid confusion :) [13:01] :) [13:04] Finally :) [13:04] I finished my quantal download:) === attente_zzz is now known as attente === broder is now known as broder_ [14:11] w00t http://bazaar.launchpad.net/~mozillateam/firefox/firefox-trunk.head/revision/1420 \o/ [14:12] chrisccoulson, wouhouh, great stuff [14:14] chrisccoulson: wow! [14:15] chrisccoulson: that's awesome, I'm hoping we can do the same for Thunderbird, I'll also try to add some of those disabled tests to QRT at some point so we get some coverage [14:16] mhr3: hah! now, that took me way longer than it ought to have to figure out what's wrong in your branch :) === gatox_lunch is now known as gatox [14:31] pitti, so what is/was it [14:31] ? [14:31] mhr3: I followed up on the bug [14:34] oh ffs [14:35] being this sensitive to param names is counterproductive === Chucrute301 is now known as sabdflI [14:37] :) === sabdflI is now known as sabdfI [14:37] :) [14:41] sabdfI: good morning [14:41] Sweetshark: I'm not sure what pitfalls you're expecting. one thing that comes to mind would be delay in the device being ready after resume, or broken drivers failing to set the device back in the right mode [14:47] cyphermox: /win 19 [14:47] ups [14:47] really? :) === m_conley_away is now known as m_conley [15:44] pitti, I'm looking into using python3-dbusmock, and I'm curious... once I set up a mock, what's the best way to validate that the arguments the mock receives are the expected ones? (the code string dbusmock lets me provide in AddMethod can't seem to percolate up an exception) [16:26] charles: are you there? [16:55] seb128: ok I think we can push g-c-c/g-s-d 3.6 with --disable-ibus and switch with ibus 1.4 for now [16:56] jbicha, hey, great, did you see that I upload g-s-d 3.6 to raring? [16:56] oh you did? [16:56] jbicha, but I went for the "revert keyboard code to 3.4" patch way rather [16:56] which [16:56] - let us our keyboard indicator (until the new one is written) [16:56] - let us the "layout per win" option [16:57] - don't require ibus 1.5 [16:57] the --disable-ibus would have broke the first 2 items [16:58] ok I used rev. 370 of the g-s-d branch but with --disable-ibus and separate-layout-per-window seems to work ok [16:58] well, they dropped the ui in g-c-c no? [16:58] we can still revisit that part in a later update [16:58] but at least we don't regress on the indicator meanwhile [16:58] it was the safest bet [16:58] yeah that's probably fine for now [16:59] looking at g-c-c atm [16:59] there is still an input sources tab in g-c-c but it doesn't set ibus methods, just alternate keyboard layouts [17:02] seb128: can I push g-c-c then? [17:03] jbicha, once I get the patch to apply and build sure ;-) [17:03] working on it atm [17:03] cool stuff! [17:04] seb128: what patch is that? [17:04] jbicha, reverting the keyboard stuff to 3.4 so it works with g-s-d [17:04] it wouldn't work if the control center was writing the new keys and gsd loading the old ones [17:05] it's also temporary but it's the easiest way to get g-c-c 3.6 in [17:05] that narrows the problem space to keyboard then [17:05] seb128: ah, I just pushed rev. 513 so you'll need to rebase then [17:05] ok [17:05] jbicha, thanks === Ursinha is now known as Ursinha-afk [17:23] seb128, jbicha: Saw your discussion; then there is no reason to postpone language-selector => g-c-c, right? [17:23] cyphermox, find . -name <...> -delete btw [17:24] rather than -exec rm ... [17:24] didrocks, ^ [17:24] seb128: good catch, not a biggie, but would be nicer to avoid spawning it in the future [17:24] GunnarHj, what do you mean "postpone"? so far we stay on the GNOME 3.4 stack again so it's we might keep language-selector for another cycle [17:24] seb128: .. how long has -delete been there? :) [17:25] sarnold, years? [17:25] seb128: cool. :) [17:25] I've been using it for years at least [17:27] seb128: Thought you just said that g-c-c 3.6 without ibus support is about to be uploaded to raring. Maybe I misunderstood. [17:27] GunnarHj, that's what I said [17:27] GunnarHj, there are issues with the new ibus stack, cf my email on the ubuntu-desktop mailing list for details [17:28] GunnarHj, but if we roll back that code we need to roll back the region capplet updates as well ... which is what we wanted to replace language-selector with [17:28] GunnarHj, so we might end up not replacing it this cycle [17:28] GunnarHj, does that make sense? [17:29] seb128: the only reason you need to rollback the keyboard/region stuff is for the keyboard indicator rewrite, right? [17:29] seb128: Yes, I've read that. Might make sense... But what exactly is missing in the 3.4 code that would have been available in 3.6? [17:29] seb128: ah, right [17:29] GunnarHj, we need to build support for language pack there and some other improvement and we don't want to do it on an outdate codebase [17:30] jbicha, the main reason to roll back so far is my lack of understand of what issue will come with upgrading those [17:30] understanding* [17:30] as gnome 3.4 didn't have any ibus integration either [17:31] jbicha, so I want to get it 3.6 out this week and I do what is easiest for that (it also has the benefit to keep the indicator) [17:31] jbicha, well, I'm more concerned about stuff like the "layout per window" [17:31] seb128: Aha, that does make sense to me. The install/remove of languages is one of the things pending, I see. [17:31] jbicha, they also dropped the "options" from keyboard layout to add an ui for some specific ones, dunno if that works when ibus is turned off [17:32] GunnarHj, right [17:32] jbicha, as said it's mostly me taking the easiest path, I need to have an another look at what is available and missing exactly if we take 3.6 and turn off the ibus code [17:32] jbicha, but I don't want to block the updates on that [17:34] jbicha, I'm also reluctant to write a new indicator based on the "without ibus" stack if we go with ibus soon [17:34] seb128, jbicha: In any case we can start preparing l-s => region. For instance replacing im-switch with im-config is desirable. [17:35] seb128, jbicha: Can't help feeling that the ibus support they plan will have issues in the beginning. No need to hurry IMHO. [17:36] GunnarHj, right, that's why we are somewhat changing plan, it has issues ;-) [17:36] seb128: And it's not live yet, right? [17:36] GunnarHj, GNOME 3.6? it is [17:37] we didn't include that piece of code in ubuntu yet though [17:37] seb128: I mean in one of the larger distros. [17:37] no [17:37] opensuse next release is in march and they don't have the new ibus yet and fedora18 will release in january [17:38] GunnarHj: it looks like im-config needs an MIR, have you ever done one of those? [17:39] jbicha: No, I don't even remember what MIR stands for. :( [17:40] GunnarHj: https://wiki.ubuntu.com/MainInclusionProcess [17:40] jbicha: Yeah, I just looked it up too. [17:42] didrocks, is that ok if I do an unity upload to recommends g-c-c-unity? [17:42] jbicha: Will read up on it. I'd like to work on that switch. It requires some l-s changes that we need to do in any case, since l-s needs to be left to Lubuntu and Xubuntu in a usable state. [17:42] seb128: sure, but please propose a MR as well [17:42] didrocks, will do [17:42] including your changes and changelog [17:42] thanks [17:42] yw ;-) [17:42] well it might be tomorrow [17:42] still some issues with g-c-c [18:11] Hi I am using Ubuntu 12.10. I like to create a stand alone grub, i.e. a usb storage disk with only Grub content. And install the Ubuntu on a different storage device. Is this possible?thx [18:13] sounds like you want /boot mounted to that usb disk.. [18:14] yes that is what I want. [18:14] probably need to do some custom partition,ing [18:14] any idea please??? [18:15] hrm, you should be able to place grub on one device, /boot on a second device, rootfs on a third device... [18:16] .. and if your /boot and rootfs on are the same device, that just saves some hassle :) hehe [18:17] * didrocks waves good evening [18:22] seb128: how would I go about having gtk2/gtk3 builds with dh9? do you have an example package that does this? [18:22] cyphermox, libxml2 [18:22] ah, thanks [18:23] cyphermox, you basically override the targets to configure, build, install [18:23] so you can do [18:23] cd gtk2; ../configure --enable-gtk2 [18:23] cd gtk3; ../configure --enable-gtk3 [18:23] same for build and install [18:24] in fact libxml2 doesn't do exactly that === Ursinha-afk is now known as Ursinha [18:24] well, I understand the idea though with TARGETS=... [18:27] cyphermox, http://lists.debian.org/debian-mentors/2012/06/msg00598.html [18:48] seb128: thanks. I think I mostly got it, but using logic similar to libxml2... [18:49] cyphermox, yw! [18:49] because I can easily just call the dh_auto_* most cases, just need to pass parameters [18:49] finishing up with the install that is a bit more tricky now... [18:49] fwiw, that was about updating the packaging for libdbusmenu, to go with the indicators [19:00] Is anyone familiar with passing file descriptors over dbus? Apparently it has support for it, but I'm not sure how it is supposed to work [19:10] mterry: I don't think you can pass one directly, it needs to get passed using a socket [19:37] mterry, try asking desrt or davidz on the GNOME IRC === mterry is now known as do_not_highlight [19:42] do_not_highlight: hi! [19:43] poor guy... [19:43] desrt, yeah, I dropped him an email to ask if that was a segfault on highlight ;-) [19:43] seems to be [19:43] that's my guess too [19:44] i didn't figure it out until i did that again :p [19:44] aww [19:44] desrt, he did it 3 times in the afternoon already, that started being a lot ;-) [19:46] lol [19:51] graaa [19:51] g-c-c build fix conflict on push [19:52] jbicha, thanks for fixing my upload but please if you jump on a fix ping on IRC [19:52] jbicha, I was fixing it as well ;-) [19:52] * seb128 uncommits [19:53] jbicha, I think the libxkbfile-dev build-depends can be dropped as well [19:54] seb128: ok [19:54] jbicha, thanks for fixing it ;-) [21:01] cyphermox, hi :) [21:03] just wanted to mention that thunderbird crashes with the new indicator-messages snapshot (running with glib 2.35.3 here) http://paste.debian.net/plain/211360 [21:09] oh, cool, thanks! [21:25] Laney, hi, just wanted to mention a gst1.0 port of rhythmbox is available https://launchpad.net/~ricotz/+archive/experimental/+sourcepub/2800650/+listing-archive-extra [21:25] (the visualizer plugin is disabled upstream) === m_conley is now known as m_conley_away [21:51] cyphermox: http://pastebin.ubuntu.com/1375857/ [22:29] seb128, what was depending on guile 1.8? [22:29] achiang: ad-hoc right? [22:30] cyphermox: i don't think so [22:30] robert_ancell, is that the old version? [22:30] seb128, yeah [22:31] achiang: because the line in that trace doesn't match what I have here :P [22:31] cyphermox: this is on 12.04 [22:31] :) [22:31] robert_ancell, not sure, I didn't look to in, everything but aisleriot by then iirc [22:31] robert_ancell, seems to still be the case, the new version has no rdepends [22:31] achiang: err, then the name of the file looks wrong [22:32] seb128, yeah, I'm just trying to track down what there was... [22:32] cyphermox: hm, i have a full log in our favorite bug [22:32] cyphermox: you interested in seeing it? [22:32] achiang: I'll look at the bug again -- except later, gotta go dinner now [22:32] cyphermox: ok [22:33] robert_ancell, reverse-depends is your friend ;- [22:33] ;-) [22:33] seb128, yeah and it's not showing anything interesting [22:33] achiang: ok, i see what the issue is, I think. [22:34] cyphermox: oh yeah? what do you see? [22:34] robert_ancell, yeah, could be, but at the same time there is nothing interesting justifying the work to do the update,transition [22:37] cyphermox: this is interesting... http://pastebin.ubuntu.com/1375958/ [22:37] cyphermox: and the log - http://pastebin.ubuntu.com/1375959/ [22:37] i wonder if that is a false positive [22:38] any glib and/or valgrind people want to tell me if my test program is wrong? ^^ [22:40] achiang, you should call g_string_free() on it once done? [22:40] seb128: yeah, i think my test program confused valgrind [22:41] well it's leaking [22:41] seb128: i mean, it's not strictly necessary to free pointers if your program is going to exit, but valgrind would still consider it a leak [22:41] valgrind is right [22:41] well, it's still a leak [22:41] the fact you exit doesn't change the fact that you didn't free the memory correctly [22:41] agreed that it wasn't proper [22:42] with you reasoning valgrind couldn't ever list leaks since it does the summary once you stop the run and at this point the program exited and there is no leak remaining [22:42] in fact adding in the g_string_free() makes the error go away [22:42] yeah, it fixes the leak ;-) [22:43] yeah, i was just testing to see if g_string_assign() leaked anything or not [22:43] or rather, seeing if we might be calling it improperly [22:43] alrighty, thanks [22:44] yw [22:45] the glib documentation could be better there [22:48] in fact the document is better in the current version [22:48] http://developer.gnome.org/glib/2.34/glib-Strings.html#g-string-assign [22:48] "the destination GString. Its current contents are destroyed." [22:49] "destroyed" == "freed" ? because it's not clear to me [22:50] heh [22:50] g_string_assign does: g_string_truncate ; g_string_append() [22:50] yes [22:52] i guess destroyed makes sense now, but you really have to read the code to figure that out [22:53] right, the documentation could be a bit more detailled on the topic [23:06] cyphermox: i found a definite bug in set_menu_item_accessible_desc(), doing some more research, but will send an MP soon i hope [23:15] cyphermox: oh, i see now. r364 from trunk was never backported to 12.04 :-/