=== EphraimMB is now known as cutie === cutie is now known as cute === bschaefer_ is now known as bschaefer [04:48] hello [05:03] Good morning [05:09] good morning pitti [05:10] hey hikiko [05:33] * TheMuso -> EOD, later folks. [05:54] good morning [06:01] bonjour didrocks [06:05] ça va pitti ? [06:06] didrocks: ça va bien, merci ! j'ai dormi à 6h30, c'est bon ! [06:07] "bien" (no?) [06:07] et toi ? [06:08] "bient", oui :) [06:08] "bien"* [06:08] much better than yesterday :) [06:08] moi, ça va ;) [06:34] Laney: previously we only have Chinese simplified IME preseeded in the iso, other languages are supported via m17n stuff, ibus-m17n (or we can do the same with fcitx-m17n), but people generally think the m17n IMEs does not work. [06:35] ah seb is not around [06:46] good morning desktopers [06:47] bonjour seb128 [06:48] salut pitti, wie gehts? [06:48] seb128: prima, danke! und Dir? [06:49] pitti, auch gut, danke ;-) [06:50] brb, trying to debug bug 1508075 [06:51] bug 1508075 in policykit-1 (Ubuntu) "crash during network setup in ubiquity wily" [Critical,Confirmed] https://launchpad.net/bugs/1508075 [07:07] bonjour! [07:08] hey larsu ! [07:08] hi didrocks, how goes? [07:09] larsu: fine! doing some image testing, already spotted 2 bugs :p [07:10] one being… decoration icons being on the right, ZOMG! [07:10] uh oh [07:10] how can that happen? [07:11] larsu: ubiquity only mode [07:12] where it handles window decoration [07:12] the other one is really weird [07:12] fresh install, first login [07:12] open apps -> menus are exported [07:12] included gtk3 apps [07:12] but for the terminal! [07:12] starting it with ctrl + alt + t or the dash [07:12] moin [07:13] didrocks: it has a "show menubar" key which might be off by default [07:13] moin Sweet5hark [07:13] larsu: under "general"? [07:13] (sorry, I was testing the system in French, obviously :p [07:13] didrocks: right click menu I think [07:14] last menu item [07:14] larsu: it's on here [07:14] * larsu thinks didrocks understands french though [07:14] weird [07:14] than that's a different bug :) [07:14] larsu: and so, I can see the menu bar [07:14] and it's not exported [07:14] unity gtk module loaded? [07:15] GTK_MODULES=unity-gtk-module [07:15] in its env [07:15] I have the "Terminal" entry exported [07:16] I do as well, until I click "Show Menubar" in the context menu [07:16] but you say this is already on?! [07:16] yeah [07:16] toggle it! [07:16] did turn it off [07:16] then, I only have the "Terminal" entry [07:16] (and it's exported ofc) [07:16] but on my main laptop, the option is on [07:17] and I have Terminal + other menus all exported [07:17] as I would expect [07:18] hm... [07:18] so you have nothing until you turn that setting off, then you only have "Terminal" [07:18] seb128, willcooke: so, we found bug 1508177 yesterday from testing the ISO. Opinions? 0-day SRU worthy or wait for upstream 5.0.3 (expected to be tagged upstream this week, released next week)? [07:18] bug 1508177 in libreoffice (Ubuntu) "Starting LibreOffice with an empty profile from the ISO creates a profile selecting galaxy, not human as theme" [Undecided,In progress] https://launchpad.net/bugs/1508177 [07:19] didrocks: what happens when you turn it back on? [07:19] larsu: let me wrap that up on a bug report [07:19] yeah thanks [07:19] I'll be back in 15 [07:21] hey larsu [07:22] good morning desktop [07:22] hey desrt [07:23] Sweet5hark, so it's being fixed in our packages not upstream? If so, can the fix go in today? [07:23] morning all [07:23] Sweet5hark, how different do they look? [07:23] willcooke: morning. [07:24] * willcooke -> school. brb [07:25] larsu: bug #1508338 [07:25] bug 1508338 in gnome-terminal (Ubuntu) "terminal doesn't export its own menu on a new install" [Undecided,New] https://launchpad.net/bugs/1508338 [07:25] hey desrt, willcooke, Sweet5hark [07:26] Morning [07:26] forced air > radiators [07:26] seb128, willcooke, didrocks, Sweet5hark, Trevinho, larsu, others: hihi =) [07:27] hey Trevinho [07:27] * desrt enjoys the n*m thing [07:28] willcooke: I have a fix pushed to my PPA, Its still building. No need to wait for upstream for that -- just that LibreOffice needs to build for ~one day anyway and the new upstream version is around the corner. [07:30] didrocks, http://people.canonical.com/~seb128/ubiquity.png [07:30] didrocks, lol, ignore that [07:30] seb128: see :p [07:30] seb128: I think I found something [07:30] I had the decoration on the left at first, they moved on the right while I was opening the screenshoter [07:30] seb128: :p [07:30] was it like that in vivid? [07:31] I don't think it was [07:31] anyone having a vivid iso around to try? [07:31] seb128: confirmed btw [07:31] it's when you open another window that they are shifted [07:31] * didrocks updates the bug [07:31] seb128: they look different -- galaxy looks more dated than human, then again human isnt looking like a master piece of modern design anymore these days either, so I dont know how noticable it really is in the end. [07:32] Sweet5hark, screenshots? ;-) [07:32] iso tracker is buggy :/ [07:33] * Sweet5hark creates screenies [07:34] seb128: description updated, feel free to rephrase [07:34] didrocks, unity-settings-daemon is "" in the ps list [07:34] "nice" :/ [07:35] seb128: confirmed [07:35] oh [07:36] journalctl has a u-s-d segfault in libpower.so [07:36] indeed /var/crash has it [07:36] * seb128 gets bt [07:36] ok, retargetting the bug meanwhile [07:37] seb128: http://people.canonical.com/~bjoern/galaxy.png vs. http://people.canonical.com/~bjoern/human.png [07:38] Sweet5hark, urg! [07:38] yeah, SRU-0 seems like worth it [07:38] willcooke, ^ opinion? [07:39] I would almost argue it's worth a respin :p [07:39] Sweet5hark, note for next cycle, test libreoffice on the iso at beta time [07:39] seb128: I would argue that, basically, if you select a network from the indicator, you have this [07:39] or are you on the libreoffice thingy? [07:39] didrocks, my comment to Sweet5hark was about libreoffice, sorry [07:39] ok ;) [07:39] that was the next bug I was about to open :p [07:40] grrrr at that laptop [07:40] we should have done some iso testing in London :-/ [07:40] touch screen keeping giving bad inputs… [07:40] so hard to test… [07:40] ok, timezone is whenever is it… [07:41] seb128: yeah :/ [07:41] didrocks, http://paste.ubuntu.com/12884084/ [07:41] hard to say if it's in glib or power plugin without debug symbols [07:41] seb128: on the upside, I didnt work to get the fix in until 2am yesterday in vain then ... [07:42] but yeah, it's a signal it doesn't like it seems [07:43] didrocks, it seems similar to bug #1436861 that davmor2 reported previous cycle [07:43] bug 1436861 in unity-settings-daemon (Ubuntu) "unity-settings-daemon crashed with SIGSEGV in engine_update_composite_device()" [Medium,Confirmed] https://launchpad.net/bugs/1436861 [07:43] so maybe it was already in vivid [07:44] Sweet5hark, yeah, good work! [07:45] * didrocks fights with his screen [07:45] * desrt finds the best phone on earth [07:45] * desrt discovers that it is not available in canada, nor on the frequencies used by her carrier [07:45] bleh. [07:45] seb128: so, yeah, confirmed that it always crash after a while [07:46] desrt, what phone is it? [07:46] didrocks, [07:46] array = 0x0 [07:46] moto g 3rd gen dual sim 16gb (xt1550) [07:46] for (i=0;ilen;i++) { [07:46] ie: no more swapping sim cards when i need a data plan in europe [07:46] seb128: ah, I can see how this can go wrong :p [07:46] of course that doesn't work out well :p [07:47] seb128: I dont think we would have found this on the beta: FWIW I did see screenshots of LibreOffice on the beta IIRC. But because of tdf#93145, the human theme was _incomplete_ on beta. That was a known issue. [07:48] Sweet5hark, oh, ok :-/ [07:49] seb128: The fix for that was also shipping galaxy, thus adding the missing icons and everything look good -- _except_ on a LibreOffice you only started the first time ever on a virgin image _after_ that fix was in. [07:50] seb128: so: only really discoverable with testing the final ISO from scratch from what I see. :/ [07:50] grgrggr [07:51] didrocks, if I run u-s-d from vt1 it works, no segfault and decorations back on left :-/ [07:51] with DISPLAY=:0 [07:51] will be nice for debugging… [07:52] yeah, and I had a look at that segfault before [07:52] it's not as easy as putting a null check [07:52] the devices array is supposed to never be null [07:52] seb128, Sweet5hark - If a re-spin is going to happen anyway, let's get it in there. Otherwise SRU0 will be fine. Thanks davmor2 for spotting it [07:53] morning seb128 desrt willcooke [07:53] thanks didrocks [07:53] hey larsu [07:54] willcooke: k. [07:54] desrt: which one is the best one? [07:54] moto g 3rd gen dual sim 16gb (xt1550) [07:54] biggest battery life for ingress? [07:54] ie: no more swapping sim cards when i need a data plan in europe [07:54] willcooke: keep in mind that the fix is touching the LibreOffice source package, thus 24 hours build time for armhf ... [07:55] sounds appealing [07:55] nah.. it's just a cheap phone with a solid build, 2GB of ram, sdcard slot, quadcore, good camera, etc. [07:55] but the dualsim version is only made for use in india, it seems :( [07:55] and indeed: kudos to davmor2 for bringing this on the radar ... [08:00] Sweet5hark, ahhh, arm builds. Yeah, fair point. SRU it is [08:01] desrt: you need to run ingress in the cloud and have the phone only as a client. Everything is going back to 1960ies style big iron mainframes, so why shouldnt you? [08:03] willcooke: kk, will make a SRU version then and ping seb128 once the ppa build is finished and the fix verified. [08:03] thx Sweet5hark [08:04] morning! [08:05] hi Laney [08:05] hey Laney, good morning [08:06] oh hi pitti :) [08:06] hey larsu & pitti [08:06] how goes? [08:06] hey larsu! [08:07] Laney: quite fine, thanks; usual release test install/look at bugs mode [08:07] Laney: how is the sprint? [08:07] Laney: good good. Much better than last week ;) how are you? [08:08] Sweet5hark: nice you got to the bottom of it then \o/ [08:14] hey Laney [08:17] pitti: good thanks! heard last night that Ubuntu GNOME doesn't boot though :) [08:17] and there's some polkit thingy which cyphermox is looking at [08:17] larsu: great! [08:17] there's rain today! [08:17] Laney, what's the Ubuntu GNOME issue? [08:17] dunno yet [08:17] finally London looks like London?! [08:17] is that gdm not coming up? [08:18] Laney: I followed up to bug 1508075, I can't reproduce it [08:18] bug 1508075 in policykit-1 (Ubuntu) "crash during network setup in ubiquity wily" [Critical,Incomplete] https://launchpad.net/bugs/1508075 [08:18] cyphermox made it happen [08:18] maybe he can post steps to the bug [08:18] didrocks, k, I commented some more on the bug [08:19] it looks similar to what mterry debugged back then https://bugzilla.gnome.org/show_bug.cgi?id=673007#c17 [08:19] Gnome bug 673007 in general "[power]: gnome-settings-daemon crashed with SIGSEGV in engine_update_composite_device()" [Normal,Resolved: fixed] [08:19] but added back since the change for new upower, http://launchpadlibrarian.net/191328719/unity-settings-daemon_15.04.0%2B15.10.20141030-0ubuntu1_15.04.1%2B15.04.20141127-0ubuntu1.diff.gz [08:19] seb128: yeah, I saw that [08:19] I can't see what is wrong though, the g_signal_handlers_disconnect_by_data() is still there [08:19] needs debugging from somebody who understands signals&co better than me [08:20] * seb128 looks at larsu ;-) [08:20] * larsu looks somewhere else [08:20] apparently it is randomly fixed [08:20] awesome [08:21] hm? no need to look at it or are you talking about a different thing? [08:21] oh no, the gnome bug [08:22] larsu, speaking about bug #1508327 [08:22] bug 1508327 in unity-settings-daemon (Ubuntu) "ubiquity only mode: close/minimize/restore are on the right edge" [High,New] https://launchpad.net/bugs/1508327 [08:22] larsu, would welcome help [08:23] larsu, the source is http://bazaar.launchpad.net/~unity-settings-daemon-team/unity-settings-daemon/trunk/view/head:/plugins/power/gsd-power-manager.c [08:23] backtrace https://launchpadlibrarian.net/201298582/Stacktrace.txt [08:24] it's similar to https://bugzilla.gnome.org/show_bug.cgi?id=673007#c17 which mterry fixed and seemed to have come back in 15.04 with the new upower [08:24] Gnome bug 673007 in general "[power]: gnome-settings-daemon crashed with SIGSEGV in engine_update_composite_device()" [Normal,Resolved: fixed] [08:24] http://launchpadlibrarian.net/191328719/unity-settings-daemon_15.04.0%2B15.10.20141030-0ubuntu1_15.04.1%2B15.04.20141127-0ubuntu1.diff.gz was that new upower changeset [08:24] seb128: missing NULL check/ [08:24] larsu, ^ summary of the situation [08:24] array = NULL [08:25] larsu, well, it's not that simple, the devices array should never be null [08:25] I bet this function is called after cleanup or something [08:25] which is why the disconnect() thing was needed back then [08:25] back then it was happening because signals were not disconnected when the object was freed [08:26] right [08:26] I just don't see what changed that make the disconnect() not be enoguh [08:26] or what can trigger the callback after the cleanup [08:28] seb128: the disconnect is from the client... [08:28] the connect() to a device [08:31] ya this is wrong [08:31] and again a stupid plugin takes down all of u-s-d [08:31] yeah :-/ [08:33] I'll try to reprocude this locally [08:33] but I think I know the patch already [08:33] I think I have another transparency missing for larsu: http://people.canonical.com/~didrocks/tmp/contextmenu.png (I wonder how we missed it) [08:33] didrocks: → Trevinho [08:34] Trevinho: ^ [08:34] * larsu hopes this works [08:34] I guess there is a bug opened for that? :) [08:34] larsu: it's only gtk ones, but I guess that's due to the API change, right? [08:34] dunno [08:34] ♥ iso testers [08:34] larsu, right, g_signal_handlers_disconnect_by_data (manager->priv->up_client, manager) ... the up_client is wrong there, right? [08:34] didrocks: api change? this is just a theme issue, no? [08:35] seb128: no we need that one as well, because we connect to some of up_client's signals, too [08:35] larsu: yeah, well, behavior change as the previous version did have transparency [08:35] we need an additional one when removing devices [08:35] didrocks, what's the issue on that screenshot? [08:35] so something broke backward compatibility (as for what happened in the greeter) [08:35] seb128: look at the corners, they are black [08:35] not transparent [08:36] didrocks: ah right ... lots of css stuff changes all the time :/ [08:36] not considered api sadly [08:36] what corners? [08:36] should, but you know my opinion on that [08:36] the tooltip you mean? [08:36] didrocks: could be a compiz issue as well [08:36] seb128: yep [08:36] in which case Trevinho is the *even* better match [08:36] didrocks, that was rounded before? [08:37] seb128: it's rounded on the corners [08:37] larsu: well, it's all related [08:37] but then, you have black space between the rounded corners and the screen [08:37] so... I've the fix for unity, I need the change in GTK in order to be able to get this working [08:37] look at the gradient dude :p [08:37] Trevinho: do you have a bug for tracking this? [08:37] not sure, larsu? [08:38] Trevinho: ya, somewhere on unity [08:38] Trevinho: are you sure this is the same issue? This has always worked for tooltip-windows, no? [08:38] didrocks, my view is not good enough to see that, to me it just looks like the tooltip is rectangular ;-) [08:39] seb128: you have an internal rounded corner [08:39] with gradients [08:39] I believe you, I just can't see it [08:39] as said my view is not good enough for those details [08:39] zoom in! [08:39] doing so! [08:39] yeah, zooming may help :) [08:40] they should be rounded corners [08:40] now I see it ;-) [08:40] right [08:40] party! [08:40] larsu: it used to work, but maybe they've moved the shadow away from the server too [08:40] ok, opening a bug anyway, Trevinho, don't lose that one (affecting unity and gtk thus) [08:41] does that has something to do with unity? [08:41] compiz [08:41] I mean even for overrideRedirect windows [08:41] I bet that if we advertise to support _GTK_FRAME_EXTENTS, we'd get that right [08:42] Laney: hey... do you know anything about this empty.mp3 business? [08:42] haha [08:42] "told you so" [08:43] i got a comment from someone who provides nothing more than a theoretical reframing of the problem without telling me anything that actually broke [08:43] a canonical employee, no less [08:44] * desrt is not too impressed by that (yet) [08:44] Trevinho: bug #1508357 just for your pleasure :) [08:44] bug 1508357 in gtk+3.0 (Ubuntu) "Tooltips have a black squares outside of its rounded corners" [Undecided,New] https://launchpad.net/bugs/1508357 [08:45] desrt: is this about 0 byte files again? [08:45] yes [08:45] >( [08:45] desrt: hahahaha [08:45] stop caring [08:45] no I don't know anything about that [08:45] I didn't send him honest [08:45] Michi Henning is complaining that empty.mp3 is no longer an mp3 file [08:45] ...which it wasn't before, either [08:45] ice man! [08:45] ya... [08:45] ice him? [08:46] waiting to hear back on the bug [08:46] no violence, please :P [08:46] i'm getting annoyed.... [08:46] desrt: ignore. This is stupid. [08:46] probably some testsuite [08:46] testsuite is stupid, then [08:46] not saying if it is or isn't [08:46] be calm [08:46] I AM CALM [08:47] coffee machine is fixed [08:47] seb128: any idea if this bug can be reproduced in another way? [08:47] at least something works [08:47] Laney: for now... [08:48] larsu, no, I would go for the theorical fix and get that landed and see if it works [08:48] seb128: +1 [08:48] quite easy to reproduce anyway on the live [08:49] unsure why the plugin is stopped though? [08:49] but that's not new [08:50] I also wonder why we connect to these signals when colplugging [08:50] but not when hot [08:51] buggy maybe? [08:51] what is upstream doing? [08:52] the right thing [08:53] https://git.gnome.org/browse/gnome-settings-daemon/tree/plugins/power/gsd-power-manager.c#n281 [08:53] good, let's do that then! [08:53] calls the same function [08:53] we should probably backport some of the work done there next cycle [08:54] though there was quite some refactoring so not to do for this cycle [08:54] it also doesn't disconnect from the signal that I think is causing the issue :/ [08:55] seb128: ok let me try fixing this one issue first [08:55] all of this is blind anyway :L/ [08:55] :/ [08:56] larsu, thanks [08:56] didrocks, that terminal bug... seems fine here [08:57] willcooke: yeah, I couldn't reproduce after another install [08:57] ahh, kk [08:57] but the bug is still there, I didn't fake that screenshot :p [08:57] race! [08:57] :D [08:58] larsu: well, "session race" even [08:58] larsu: because I killed all terminal [08:58] desktoppers: for what it's worth, this should be a list of things to care about today: http://pad.ubuntu.com/6aWG0SagNg [08:58] and retry to start, and in that session at least, this was always the case [08:58] If it's helpful, please keep it updated, if not, please ignore [08:59] * didrocks finishes the non english use cases first [09:08] didrocks: mind trying this patch? https://code.launchpad.net/~larsu/unity-settings-daemon/lp1436861 [09:10] larsu, can you mp it so we can put it in a silo? [09:10] would make easier to test and put us closer from a landing [09:10] sure [09:10] yeah, that would be easier (I'm in a middle of an install with Internet connexion, so won't be able to test before some times) [09:11] like I said, not sure if this even works [09:11] larsu: you can easily test as well :) [09:11] just the first guess [09:11] theres a couple of other problems in this code [09:11] just download the iso, start ubiquity [09:11] but I'd need to reproduce myself [09:11] (ubiquity only mode) [09:11] ah, ok [09:11] in a vm? [09:11] should be the same, check first [09:11] you can play like opening text entry from the indicator [09:12] seb128 confirmed it easily [09:12] ok will do once I'm home (can't download the iso in the cafe I'm at) [09:12] actually let me go there right now :) [09:12] but please test in between if you're doing that anyway right now [09:12] bbiab [09:13] larsu: I'm afraid I won't be able before leaving for lunch though, but afterwards, should be possible [09:13] * didrocks waits on current install to download and finish [09:14] didrocks, is there a way to modify the iso to install the deb before starting ubiquity? [09:14] I couldn't get the issue by restart u-s-d by hand [09:14] so it needs to be the one in place on boot [09:15] didrocks, I couldn't reproduce. ARe you using 32 or 64bit? [09:15] jibel: u-s-d crash? 64 bits [09:15] jibel: ubiquity install mode (not live session) [09:16] didrocks, jibel, larsu, the issue was already there in vivid, like a race or depending of memory state so not a surprise it's not happening for everyone/consistently [09:16] seb128: hum, I guess then stopping at casper bottom, mounting an usb [09:16] seb128: I don't see better way :/ [09:17] didrocks, let me try usb-creator/key with persistant storage [09:17] maybe cyphermox would have a better clue though, as I guess he's testing also the ubiquity only mode when making modification [09:17] cyphermox, Laney, ^ [09:17] moo? [09:17] seb128: yeah, if that works, would be the easiest, obivously [09:17] obviously* [09:18] cyphermox, what's the best way to install a test deb on ubiquity install-only mode before it starts? [09:18] seb128: add break=casper-bottom to the command-line, chroot in and install the deb [09:19] I was right \o/ [09:19] either that or roll your own iso by modifying the squashfs, extracting files over it and re-squashing it, re-doing the iso [09:19] why does it need to be in ubiquity-only? [09:20] cyphermox: it's the only way we found to crash u-s-d [09:20] * didrocks goes for an early run, just updated the iso tracker with use cases I tested [09:20] if it crashes in ubiquity-only it should crash as much in the live session -- ubiquity-only just has less crap around it, because it doesn't start the session completely. [09:21] cyphermox: well, maybe this "less crap" is what results in u-s-d crashing [09:21] cyphermox: you are still talking about the polkit issue? [09:21] cyphermox: can you reproduce it? [09:21] in other words, my worry is just that if you're trying to fix a bug that was found in the session, yet you test your bug/fix in the ubiquity-only session, you might not hit quite the same thing, but you'd know better, depedning on the bug. [09:21] didrocks: exactly [09:22] cyphermox: but happy if you can reproduce it in a live session, we couldn't, please check the bug report :) [09:22]