=== Trevinho_ is now known as Trevinho [00:51] GunnarHj, is the chown you moved https://code.launchpad.net/~gunnarhj/lightdm/translate-guest-session-dialog/+merge/203091 related to the translations? [00:51] if not, it should be in a different MP [00:52] robert_ancell: No, it's not. I can split them, if you prefer so. [00:52] please do === Guest72010 is now known as balloons [02:02] robert_ancell: If I (now) understand it correctly, the lightdm trunk branch includes the upstream lightdm plus the Ubuntu debian/ folder. Is that correct? [02:03] GunnarHj, correct [02:03] robert_ancell: Good to know for next time... ;-) === TheMuso` is now known as TheMuso === duflu_ is now known as duflu [07:17] Good morning [07:19] hi pitti [07:19] hey darkxst, how are you? how did the DMB meeting go last night? [07:19] good, just missed out on MOTU, but got ubuntu-GNMOE [07:19] and desktop-extras [07:27] darkxst: ah, congrats [07:27] darkxst: (sorry, server reboot / fsck) [07:27] darkxst: that should hopefully reduce blocks on your side [07:30] pitti, yes it will help, once ubuntu GNOME package set gets some packages in it! [07:30] its currently empty [07:30] ah, heh [07:31] apparently no one ever noticed [07:32] pitti, anyway, any idea how I can inject code into gnome-shell to introspect the clutter stage? [07:32] I think that is all below gtk, so a gtk module probably wont work [07:32] darkxst: I think it rather needs to go into libautopilot-gtk [07:33] shell embeds gtk, I don't think it will have access to the clutter stage [07:33] darkxst: right, it needs to identify a widget base class (ClutterActor?) and then use its methods to iterate through its children [07:33] presuambly the property handling is identicall to Gtk, as they all (hopefully) just use GObject properties [07:34] darkxst: oh, you mean there is a big GtkWindow at the top, then some clutter widgets, and within those Gtk widgets again? [07:34] darkxst: I thougt the shell's .js only used clutter [07:34] the shell uses clutter, then embeds gtk within that for some cases [07:36] it doesnt like look I would get access to the global clutter stage from a gtk module (although I havent tried) [07:38] AP iterates "downwards", starting from gtk_window_list_toplevels() [07:38] so that wouldn't contain the clutter bits, supposedly [07:38] is there no such counterpart for clutter? [07:39] pitti, iterating the widget tree seems easy [07:39] but actually getting at it, not so [07:39] ClutterStage is the toplevel wiget [07:40] from there finding children is largely the same as gtk, just s/gtk/clutter/ [07:40] Hi desrt, any chance you can update your patch in gnome #708765? It is blocking my fix for bug 1189172 :) [07:40] Gnome bug 708765 in general "D-Bus-activated keyring-daemon remains even when its bus terminates" [Normal,Unconfirmed] http://bugzilla.gnome.org/show_bug.cgi?id=708765 [07:40] Launchpad bug 1189172 in python-secretstorage (Ubuntu) "Should run tests during build" [Wishlist,Triaged] https://launchpad.net/bugs/1189172 [07:44] pitti, mutter does know about the clutter stuff, but I don't believe the plugins there work anymore [07:49] pitti, I could in theory do it as a shell extension, but that would duplicate a lot of code [07:50] darkxst: you mean the whole bit of exporting the tree and its properties to D-BUS? [07:50] yeah, that woudl duplicate the entirety of the plugin, and be an one-project solution only (and the cost/benefit ratio is probably way too high) [07:50] pitti, yes, but really it would be better to somehow inject a g_module or something [07:51] pitti, I did not say I would actually do that! [07:51] (*phew* :) ) [07:51] darkxst: what's the kind of things you would want to test for the shell? [07:51] i. e. would that be possible with image-based matching? [07:52] the shell is always full screen, so certain elements should be at fixed coordinates for a particular environment [07:52] like xvfb with a fixed screen size [07:53] pitti, right now I just want to test that the shell loaded correctly, but it would also be nice to test things like overview search etc [07:53] right [07:53] the event injection probably works just fine [07:54] gnome-shell steals the gdk input events, so thats probably identical? [07:54] checking "the overlay comes up" is probably feasible with testing the average color of a particular rectangle, or something like [07:54] but ensuring that individual icons appear there sounds less robust, as they tend to change [07:55] darkxst: yes, and it doesn't matter; events shoudl be injected with uinput, so at the evdev level [07:55] that'll work with X, wayland, and every toolkit [07:55] ok [07:58] pitti, I probably should chat with walters again, he might have some ideas at getting to the clutter tree [07:59] darkxst: oh, does walters do clutter? [08:06] not specifically but he setup all the gnome c-i stuff [08:07] pitti, and he is very interested in having a gnome-shell installed tests, that test functionality (although no idea if they would take an autopilot solution) [08:30] darkxst: so how do the current gnome-shell tests work then? [08:31] darkxst: or they don't exist yet, and you mean walters would just like to have some tests? [08:50] good morning desktopers [09:03] pitti, right now, there tests are screenshots [09:03] morning! [09:03] Hi Laney [09:04] hey darkxst [09:05] bonjour seb128 [09:05] darkxst: ah, so they need to be updated with icon changes, etc. [09:06] pitti, I don;t think they even check the screenshots in any useful way [09:06] darkxst: ah, you got some upload rights [09:06] are you happy with that? [09:07] Laney, more or less, missed out on MOTU but probably Ubuntu GNOME + desktop-extras is enough. alteast once Ubuntu GNOME packageset has packages! [09:07] yeah I'm going to look at that stuff soon-ish [09:07] should be able to move it out of its silo [09:08] the code is quite jumbled atm so it's hard to make the change i mentioned [09:08] I think it'll come out mostly like desktop-extra, so that one probably will just go away [09:08] Laney, edubuntu has a lot of stuff that is not even seeded in edubuntu [09:08] build-deps get things [09:09] if you poke around in http://people.canonical.com/~ubuntu-archive/germinate-output/ you should be able to see things like that [09:09] Laney, really? things like gnome-shell and mozjs17 [09:10] some things will be directly on their image [09:10] gnome-shell-common (from gnome-shell) is seeded in: edubuntu: dvd [09:11] hey Laney pitti darkxst [09:11] hi seb128 [09:11] http://people.canonical.com/~ubuntu-archive/germinate-output/edubuntu.trusty/rdepends/gnome-shell/gnome-shell-common [09:11] for example [09:11] my point is that there's usually an explanation [09:11] Laney, oh I guess that is why they have mozjs17 then, although that will be gone with 24 [09:11] hi seb128 [09:11] how's it going? [09:12] good! you? [09:12] yeah pretty alright [09:12] went to a japanese restaurant and then a 'secret' (everyone knows about it really) cocktail bar for birthday celebrations [09:13] Laney, oh that is my fault then for adding a gnome-shell frontend to ubiquity? [09:13] * seb128 was replying to Jim from yorba about http://blog.yorba.org/jim/2014/01/shotwell-elementary-and-pantheon-photos-what-it-all-means.html (and about him calling our uoa-integration patch a fork) [09:13] Laney, oh, nice! [09:13] although tweak-tool is in there too [09:13] oh, nice, new gtk 3.10 [09:13] * seb128 adds to his todolist for the day [09:14] darkxst: dunno, anyway I'll make it so that the overlap is allowed [09:14] Laney, ok, thanks [09:14] Laney, also new gnome apps currently dont have any packagesets [09:15] might be a reason for MOTU in future if you don't want them in your iso [09:15] or maybe desktop-extra, dunno [09:15] darkxst: Laney: ubiquity under gdm/gnome-shell looks hot. [09:15] can talk about it [09:15] heh [09:15] Laney, gnome-maps and gnome-music so far [09:16] hoping to get gnome-photos and gnome-weather in also before feature freeze though [09:16] you didn't get MOTU? [09:16] seb128, not quite, apparently I haven't done enough non-GNOMEy packaging [09:16] k [09:17] xnox, :) [09:17] does it have csd and all that fun? [09:19] Laney, not in ubiquity, but it could! [09:19] totally [09:19] * darkxst not doing that though [09:20] Laney, its basically a stripped down shell, implemented through session modes, rather than the usual ubiquity DM [09:21] interesting [09:21] * Laney doesn't know about session modes [09:21] Laney: in essence, unlike lightdm, gdm allows to execute apps =) upon quiting, autologin is completed into full session. [09:21] Laney, thats gnome-shell specific [09:22] Laney: which looks like a full gnome-shell, but without e.g. launchers, multiple desktops, minimizing. [09:22] pix pls [09:22] also I'm jealous of your breakfast [09:22] Laney, breakfast? it dinner time here ;) [09:22] Laney: i've been asking ted, if I could launch unity7 without global-menu, dash, launcher, and some indicators. [09:23] Laney: and launch all of those, when "try ubuntu" is clicked for example. [09:23] Laney, is bdrung still around (in DMB context) [09:23] I haven't seen him very m uch [09:23] tried email? [09:23] xnox, what's the point of launching unity without any unity UI? [09:24] Laney, no, I'm going to try, I wonder if this year is the year of the libreoffice upload rights for Sweetshark :p [09:24] seb128: such that i don't have to maintain lightdm, and "try ubuntu" -> unity transition doesn't involve tearing down X, starting X, starting lightdm, completing auto-login and reloading indicators/panel which look the same as 3 seconds ago. [09:25] * ubiquity-dm that is. [09:25] (maintain ubiquity-dm...) [09:26] xnox, well, if you were using compiz you could probably just start it without the unity plugin and load that one later [09:27] gnome-shell has session modes, since gnome-shell draws the gdm greeter === vrruiz_ is now known as rvr [09:28] and also there is the gnome-initial-setup mode [09:28] its very easy to define any random custom mode, including whatever components are required [09:31] you have 2 gnome-shell running then? [09:32] one for your user and one for the greeter? [09:32] or is the greeter screen part of the user session? [09:33] seb128, lock screen is part of user session, but login greeter would be a seperate session [09:33] or maybe it all runs under a gdm session [09:34] the lock screen doesn't list other users then? [09:34] (e.g it bounce you to another screen to switch?) [09:34] it has a "switch user" link, which I believe would redirect to the login greeter [09:36] ok, similar to the "old way" then [09:36] I was wondering because we have been talking about using the greeter as lock screen with lightdm [09:37] but it creates issues when the greeter is running with another user/on another vt [09:37] you get your seat becoming unactive, which e.g stops audio playing for the user [09:38] seb128, not exactly, gnome-shell always renders the greeter [09:39] and I don't know exactly how it works, but there is no VT switch happening in there, so you don't get big glitches [09:42] so the greeter is on the same VT as the user but running as another user? [09:42] interesting [09:47] mitya57: i won't likely get much time to work on this soon [09:50] seb128, I believe there is a privileged auth channel between gnome-shell and gdm [09:51] seb128, halfline and/or robert_ancell probably know more of the gory details though [09:52] right [09:52] desrt: OK, I will try to look myself then [09:54] pitti, hey, do you have any idea what component is the right one for https://bugs.launchpad.net/ubuntu/+source/pygtk/+bug/1273380 ? [09:54] Launchpad bug 1273380 in pygtk (Ubuntu) "gtk.gdk.Window.get_origin() returns 3 values instead of the documented 2-tuple" [Undecided,New] [09:54] pitti, gtk itself? [09:57] seb128: yeah, tried that, was not reliable - compiz crashing or gtk-decorat [09:57] xnox, I doubt that's something we are going to fix/change for unity7 [09:59] xnox, right, despite your dire warnings, I haven't seen a single crash under gnome-shell ubiquity ;) [10:01] seb128: yeah. [10:04] seb128, what do you think of gnome-desktop transition? [10:04] (assuming g-s-d fork goes ahead) [10:07] seb128: yes, but it's not a bug; I'll follow up [10:09] pitti, danke [10:09] seb128: oh sorry, the way he wrote it it's indeed pygtk, not GI [10:09] seb128: *dankon ;) [10:09] nope, I think it's him being confused [10:09] pitti, oh, the example in there is pygi so I was wondering [10:09] desrt, good "night"? [10:09] seb128: dankon = thank you [10:10] desrt, yeah, I got that, it's not far from "danke", thanks :p [10:10] desrt, I'm just surprised to see you up at 5am [10:11] seb128: done [10:11] pitti, dankon ;-) [10:11] seb128: i'm in GMT [10:11] norwich, right? [10:11] yup [10:11] neat [10:11] I quite like it there [10:11] seb128: neniu problemo [10:11] desrt, oh ok, had a good flight? [10:11] pitti, ;-) [10:11] pitti: ;) [10:12] seb128: yes. very nice. [10:12] desrt, did you get promoted again? ;-) [10:12] 'no comment' [10:12] lol [10:12] but... [10:12] man, that's never happened to me [10:12] i went through halifax [10:12] which was awesome [10:12] because it means that the longest leg of the flight was 4h45 [10:15] hm, my train to Brussels will take longer than that :-/ [10:15] maybe I should ask them to go through Halifax! [10:15] * larsu heard everything's faster going through there [10:16] larsu: that may be longer for you :) [10:16] I'm not sure where the flaw in my logic would be... [10:16] there's a halifax in england [10:16] still took 2hr of flying to get to halifax from toronto [10:17] so the overall trip was much longer [10:17] try that one out [10:17] but being able to get out and stretch your legs is very nice [10:17] Laney: what's with your english guys? [10:17] naming all of your cities after ones in canada [10:17] can't improve on perfection [10:17] Laney: Berlin -> Halifax (England) -> Brussels still sounds like a detour to me :P [10:17] :P [10:18] i just don't understand why you picked London, Ontario (of all cities) to name your capital after [10:18] also it's a medium-sized West Yorkshire town [10:18] not exactly a bustling transport hub [10:19] seb128, my standalone display config daemon is working well, apart from one munched 'monitors-changed' signal, but I can fix that [10:19] https://en.wikipedia.org/wiki/File:London,_Ontario,_Canada-_The_Forest_City_from_above.jpg [10:19] inspirational [10:20] darkxst, oh, sorry, forgot to reply earlier ... that gnome-desktop transition makes me quite nervous [10:20] it's a non trivial change for mid-cycle in a LTS cycle [10:21] seb128, its a pretty big diff, but its really just code moving places [10:21] well, it's new code as well no? [10:21] and a new architecture [10:21] moving code from a library to a dbus service [10:22] is the new service using the exact same logic as the old gnome-desktop code? [10:22] or is that a new implementation? [10:22] its obviously not exactly the same, but it more or less similar [10:23] right "more and less" is the issue ;-) [10:24] probably the biggest change is the way monitor configs are stored [10:24] screen configs are non trivial, we had bugs during cycles about multimonitor and configs [10:24] like what config is applied when plugging a new monitor [10:24] what's the default for projectors, etc [10:25] oh, they changed the config format [10:25] is there a migration path ? [10:26] hmm possibly not [10:28] :-( [10:29] monitors.xml no longer exists, it is instead stored within the dameon somewhere [10:30] what do you mean? that config is supposed to persist between sessions/reboots/etc [10:31] seb128, it would be backed by g_settings or something, it does persist [10:32] ok, so it's stored in gsettings rather than in a .xml [10:32] well, one step at the time [10:32] it just wont migrate from the old .xml file [10:32] we are finishing the u-c-c transition, then g-s-d -> u-s-d should be next [10:32] then we can test the gnome-desktop transition in a ppa [10:32] but having config migration is going to be needed [10:33] we can't have people upgrading their LTS and having their screen config wrong on next reboot [10:33] tight [10:33] feature freeze is 02/20 [10:33] wtf, I did an American date [10:34] boooo [10:34] seb128, Laney, right if this is going to happen, probably needs to go in parallel [10:35] bah [10:36] QDBus and 'complicated' return types is hard [10:41] stop using qdbus :) [10:41] that'd be nice [10:41] it is by far the worst of the dbus bindings [10:41] and there is absolutely no reason it should ever be used..... [10:41] if we're exposing it as some app-facing API then we've deeply failed somewhere along the way [10:41] we ought to be using gdbus to create higher-level qml-based wrappers... [10:42] because if we're wrapping, then there is no reason to want to use it [10:42] I'd like it if that existed [10:42] if what existed? [10:43] qml dbus bindings in the distro [10:43] but why? [10:43] Laney, I think he's saying "use gdbus in your cpp backend code" [10:43] precisely. [10:43] dbus (in any form) has no business appearing at the application api level [10:44] therefore the precise flavour that dbus has is not important -- and you should use the best-featured and most convenient dbus implementation [10:49] I'd like if where it would avoid me having to write any C++ at all [10:49] that's not to say that exposing proper high-level QMLish interfaces isn't the best way [10:49] and yes, we should have chosen gdbus back at the start [10:49] ah... you hope to use pure qml to write the bindings? [10:51] that's an UI, not binding [10:51] that's the thing we discussed with larsu some time ago [10:51] it would be nice if you could binding e.g a checkbox in a qml to a dbus bool [10:52] seb128: no. that would be awful. [10:52] this is _exactly_ how you do not want to use dbus [10:52] at that level, it would [10:52] DBus shouldn't be exposed at that level, I agree [10:52] well, as a qml writer it would be nice, one like "checkbox.enabled: pulseaudio.muted" [10:52] it should be pushed down one more [10:53] Laney, well, then you force people to add cpp code and the build system glue, etc [10:53] that's lot of non obvious work to get one boolean value [10:53] I am saying that there ought to be an easier way to write bindings to QML [10:54] seb128: for all but the most trivial cases, writing a meaningful API around a dbus backend is going to be more complicated than binding up some properties [10:54] desrt, yeah, I was speaking about the trivial cases there [10:54] which is frankly what we have most atm [10:55] we want checkboxes to reflect on/off status of system components which export their status on dbus [10:56] and nothing complicated about that at all like policykit to change the value or anything? [10:57] yeah, writing is less easy [10:57] We don't have PK prompting for anything at the moment [10:57] ah [10:58] I just found my problem - I forgot to call qDBusRegisterMetaType ... [11:00] desrt: seb128 is talking about the DBusObject thing I wanted to write last year [11:00] they had the problem that they wanted to read a single property from some service and bind it to a checkbox [11:00] it was only ever intended for that trivial case [11:01] so that they wouldn't need to write wrappers around every service just to access one thing (this was about system settings iirc) [11:01] (correct) [11:01] I still think it's a great idea. It fell short because qml made it so hard that it stopped being worth it [11:01] the qml/cpp bridge is pretty bad [11:02] seb128: thanks :) [11:02] Would a code generation approach be better? [11:03] * larsu runs [11:03] From a feasibility standpoint ... [11:03] Laney: ya, probably (I forgot what the actual issues where tbh) [11:03] Something to do with QQMLPropertyMap iirc [11:04] ah right, signals wouldn't have worked [11:04] also, the hole asnyc fiasco [11:04] you can't instantiate qml object asnchronously [11:04] which is understandable, but annoying [11:05] well, at the end most app writers are not going to have to deal with that [11:06] and we can write the cpp code for u-s-s [11:06] it's a bit more work but we have things in place and are used to it by now [11:07] this seems to work now [11:07] * Laney MPs [11:36] our builders are too slow. The last rc is still building and I already have the next ready. [11:37] dont blame the builders for your lack of slack ! [11:44] ogra_: faster, ^&**^cat! kill, kill! [11:49] didrocks: c'mere when you're not busy [11:53] bah [11:53] seb128: can we do a gsettings-u-t-schemas release with the AS MPs from u-s-s? [11:53] you don't need to take the ringtone one yet [11:53] should be ok without that afaik [11:54] need to check the versions [11:54] on the Breaks/Depends === MacSlow is now known as MacSlow|lunch [12:01] Laney, yeah, it's on my list for today, in fact I was just waiting for the landing from yesterday/this morning to be merged back it to queue that one next [12:04] okay, so it's just releasing g-ut-s trunk with https://code.launchpad.net/~laney/gsettings-ubuntu-touch-schemas/auto-brightness/+merge/203302 and another MP to fix the Breaks version to the current u-s-s one [12:04] and then u-s-s with https://code.launchpad.net/~laney/ubuntu-system-settings/as-schemas-package/+merge/201470 [12:05] didrocks, sil2100, Mirv: is gsettings-ubuntu-touch-schemas still under autolanding or is it under CI train? [12:05] hmm, there is some obscure missing dependency in the LibreOffice buildsystem. Builds fine on a quad-core with hyperthreading but stumbles on 32 cores now and then. [12:06] what's "* None" in the u-s-s changelog? [12:07] Laney, a but in didrocks' CI script I guess [12:07] didrocks: ^ [12:07] ya [12:07] didrocks, https://launchpad.net/ubuntu/trusty/+source/ubuntu-system-settings/0.1+14.04.20140127-0ubuntu1 [12:08] also can we start running AP? sil2100? [12:11] Build needed 09:55:57, 27306784k disk space <- \o/ [12:11] O_O [12:12] Laney: its that awesome how small the LibreOffice build is once l10n is build in a separate package? [12:12] so tiny I can almost fit the bits in one hand [12:13] Sweetshark, how much was it before? (what arch is that) [12:14] seb128: that number is not really relevant as its the amd64 arch (PPA build). But good point, lemme check. [12:22] seb128, Laney: let me check [12:23] seb128: i386 on trusty -- 4.1.3-0ubuntu3 Build needed 07:24:00, 24700736k disk space -- 4.2.0~rc2-0ubuntu2 Build needed 04:37:34, 21572744k disk space [12:23] seb128, Laney: ok, so it seems I moved it out of cu2d today in the morning during the big batch of changes - so I guess it's ready for CITrain [12:23] (no more cu2d) [12:24] larsu: dbarth_: i want to join https://launchpad.net/~indicator-applet-developers to perform code reviews. I'm versed in C/GTK+ and some C++/Qt. [12:24] k [12:24] sil2100: what about the AP question? [12:24] seb128: so it safed some 3GB of discspace (plus probably another or two as 4.2 like grew itself some new code over 4.1) [12:25] sil2100, great [12:25] Laney: what do you mean exactly by running AP tests? [12:25] Sweetshark, nice ;-) [12:26] seb128: that should be in CI Train [12:26] (documented in the CITrain rampup doc) [12:26] sil2100: http://ci.ubuntu.com/smokeng/trusty/touch/mako/150:20140128:20140115.1/6308/ [12:26] Mirv, right, sil2100 just said so [12:26] in that list [12:26] yep, read that too now [12:27] seb128: ah wait, the 4.1 number is unreliable/likely too low as we deleted the objects files we build after linking and before installing in the mother-of-all-hack. [12:27] s/hack/hacks/ [12:27] seb128: so we likely benefit some more ... [12:30] Sweetshark, in any case it's a clear win ;-) [12:30] seb128: yep. [12:31] Sweetshark, btw, do you have another upload queued for the base fix? [12:34] Laney: ah, you mean, some specific new AP tests you want to be ran during smoketesting [12:34] ? [12:34] We have ubuntu-system-settings-autopilot these days [12:35] which should work (but please check it does before enabling it) [12:37] * didrocks looks for u-s-s MP [12:38] * didrocks bets on no commit messages [12:38] seb128: I have a hotfix queued up (this is this one I just for the ppa). And now that I found the rootcause, I fixed it properly. [12:39] actually not [12:39] seb128: (the lib was accidentally changed to a different library group, when debian moved it to another package in vendor patching. Thus the name changed and doesnt match with the expected one for dynloading from Java). [12:40] xnox: sure, done. [12:40] No commit message was specified in the merge proposal. Click on the following link and set the commit message (if you want a jenkins rebuild you need to trigger it yourself): [12:40] it was [12:40] the commit message was added after I think build was pressed [12:40] on https://code.launchpad.net/~seb128/ubuntu-system-settings/sound-match-design/+merge/203304 [12:41] shouldn't this check be enforced? [12:41] seb128: when do you want to upload (is this still timelocked with a poppler transisition?) [12:41] didrocks, right ... and what laney said, should a CI fail leads to a stop somewhee? [12:42] Sweetshark, the current proposed version is already build with the new poppler, if you are fine having that one moving to trusty proper no need of a rebuild/upload soon [12:42] Sweetshark, if you would like to fix base before libreoffice moves to release, it would be nice to do another upload soon to not block the poppler transitin [12:43] seb128: see on the code, I was wondering if I should fail or not [12:43] did you see why libreoffice is blocked? [12:43] there is # TODO: find a fallback or warn if no commit message attached to the MP [12:43] :) [12:43] funny, it's just the line before the one that failed on the unicode issue [12:43] xnox: ah, i see you've been added now [12:43] so maybe I'll fix both at the same time :) [12:44] dbarth_: I wouldn't mind joining too if you're adding people [12:44] Laney: seb128: do you mind if it failed and don't check the next ones? [12:44] Laney, one binary to promote to main, failing autopkgtest, some poppler rdepends still not rebuilt (last I checked) [12:44] because for all the rest, I tried to continue [12:44] dbarth_: We talked about adding ~ubuntu-core-dev in #ubuntu-devel a little while ago [12:44] check everything [12:44] what do you think about that? [12:44] and in the end failed [12:45] but it's a little bit more complex in that area of code [12:45] didrocks, ok, I've no strong opinion either way on the commit message thing, would still be nice to take the description as fallback :p [12:45] seb128: pffff, feature request, I was sure of this! [12:45]