=== maclin1 is now known as maclin [02:52] Bonjour [03:17] hey pitti :) [03:19] hey sarnold, how are you? [03:20] pitti: pretty good, just got back from dinner :) [03:21] pitti: how's your day going so far? [03:21] sarnold: it shouldn't even have begun yet, but I can't sleep any more :) [03:21] pitti: completely agreed :) [03:21] started as usual, wading through -proposed mess; due to the uninstallable kernel we got tons of failures at night [05:39] good morning === nudtrobert1 is now known as nudtrobert [07:22] good morning! [07:25] hey larsu [07:25] hi pitti! Wie gehts? [07:26] larsu: ganz okay, war nur viel zu frueh wach [07:26] larsu: how is your new flat? [07:26] pitti, amazing! Loved waking up there this morning :) [07:26] but no internet yet [07:26] working at dholbach's office with him and desrt [07:27] hello [07:31] hey larsu! desrt [07:31] hey pitti, /me se sentait seul ce matin :) [07:31] bonjour didrocks! ça va? [07:31] larsu: ça va, and you? [07:31] slept for the night in your new accomodation? [07:31] ça va tres bien [07:31] yep [07:31] didrocks: yup. it's sleepy. [07:31] nice! [07:32] larsu tried to wake me at 7... then at 7:30, then at 8.... [07:32] ahah [07:32] didrocks: ah, désolé -- j'ai me léve déjà à 5h, mais j'ai perdu à dire bonjour [07:32] then dholbach came to pick us up :) [07:32] larsu has the repeat function :p [07:32] hehe [07:32] handy feature [07:32] the "get up now, daniel is here" worked [07:32] heh, that was like "no other way now" [07:32] pitti: pas de problème ;) [07:32] there was also promise of coffee and a chocolatine [07:33] hey desrt, kiel vi fartas? [07:33] pitti: bonege, dankon :) [07:33] ni estas ĉe daniel :) [07:34] hi guys [07:34] ya. this one :) [07:34] can somebody take a look at https://code.launchpad.net/~attente/unity-control-center/non-gnome-unity-ibus-support/+merge/238293 and https://code.launchpad.net/~brunonova/software-properties/lp1381050_2/+merge/264634 and https://code.launchpad.net/~binli/unity-settings-daemon/dont-scale-out-of-screen/+merge/268881? [07:36] dholbach: even if that's not really my role, but I'll patch pilot this afternoon, so I can have a look [07:36] * dholbach hugs didrocks [07:36] * didrocks hugs dholbach back [07:37] * didrocks would have like a "pain au chocolat", but it seems it's only for people in Berlin :p [07:39] didrocks, it's a "Schokoladencroissant" over here :-P [07:39] didrocks, or "Schokocroissant" rather [07:39] dholbach: we don't need to add more friction to this particular argument... [07:39] also, disappointed that there is no ¨ in there somewhere [07:40] desrt: so you want an Überraschungsei ? [07:41] * didrocks lalalalala, "croissant" and "chocolate" in the same line [07:41] larsu: https://code.launchpad.net/~binli/unity-settings-daemon/dont-scale-out-of-screen/+merge/268881 sounds like a hack, do you know if there is a better fix upstream? [07:42] didrocks: love you too :) [07:43] didrocks, this is ridiculous... [07:44] larsu: mind commenting? :) [07:44] larsu: of course, using same exact words :p [07:58] morning all [08:00] hey willcooke [08:02] didrocks: hey. i just met someone who looks just like you [08:02] maybe a brother? [08:03] morning! [08:03] Laney: hi! [08:04] Morning all, today is Setting Orange, the 26th day of Bureaucracy in the YOLD 3181! [08:06] hey desrt [08:07] good moving yesterday? [08:07] yes [08:08] slept well :) [08:08] (although i didn't do that much, really) [08:12] desrt: arent you canadians supposed to only move on July, 1st? (https://en.wikipedia.org/wiki/Moving_Day_%28Quebec%29) [08:12] Sweet5hark: canada exists outside of quebec [08:12] Sweet5hark: furthermore, i'm moving in berlin... [08:12] desrt: oh, cool! [08:12] not me. [08:12] helping larsu [08:13] he needed some extra hands, so i flew over [08:14] he had to leave the country to move outside of the designated day [08:14] didrocks, finally commented [08:14] nice [08:14] Laney: so all Canada is just Quebec? Thought so! [08:18] * Sweet5hark should be careful not to troll _everyone_ today who has rights to sponsor his package. [08:18] :D [08:20] didrocks, commented on the bug as well. Probably we want to raise the dpi limit instead of hard coding a minimal resolution [08:20] 192 is *not* hidpi imo [08:20] but that's what u-s-d uses as a limit now [08:21] * larsu can prepare a patch if needed [08:21] Laney, is there something like `apt-get build-dep` for `apt`? [08:22] don't think so [08:23] * dholbach hugs didrocks [08:23] Laney, I don't understand your comment about blocking the constructor [08:23] * didrocks hugs dholbach back [08:23] what comment? [08:23] oh [08:24] larsu: no harcoding? so sad :p [08:24] more context please :P [08:24] larsu: agreed btw :) [08:24] Laney: hey! [08:24] Laney: //code.launchpad.net/~binli/unity-settings-daemon/dont-scale-out-of-screen/+merge/268881? [08:24] no [08:24] ah sorry, didn't backlog fully [08:24] larsu: we need to know the initial value for the property before giving an object [08:24] Laney, indicator-datetime MR [08:25] you block monitor_timezone_property(), not the constructor, right [08:25] yeah, that's only called from the constructor [08:25] so it essentially is it [08:25] ah, right, sorry [08:25] larsu: I'm going to smooth a little bit your answer :) [08:25] I kept the structure of the original code [08:26] yeah this is very weird [08:26] don't do that [08:26] there is another main loop still running as well, right? [08:26] ya [08:27] desrt, did you ever write a blog post "recursive main loops considered evil"? [08:27] haha [08:27] if not, please do and give Laney the link :D [08:27] Laney, gotta run for ~30 mins. Delivery to the new place [08:28] will have a further look after [08:28] tell me how to do it better [08:32] Laney: ah, if you piloting before me (I'll probably do tomorrow then, as I'm sure it will empty), do you mind looking at Daniel's request ~1h ago? (larsu commented on one already) [08:32] if not, I'll handle them in my shift [08:45] didrocks: I looked at the attente one and the other seems to have been reviewed by bdmurray 8 hours ago? [08:54] ah, my new keyboard just came [08:54] blank keycaps [08:55] Laney, wot? On purpose? [08:55] hahaha [08:55] yep [08:55] I thought it would look cool [08:56] and actually it does [08:56] also it's got clicky keys \o/ [08:56] WFH benefits [08:57] I guess not that WFH: http://www.wfh.org/ [08:57] sorry about your blood but have this neat keyboard [08:58] ;) [08:58] http://www.urbandictionary.com/define.php?term=WFH [08:58] * didrocks likes the definition [08:58] "Working from home - when your slacking off and not wanting to do any work." [08:58] sounds right [08:59] might have to move IRC onto the desktop now so that I can click clack more [09:00] or set up synergy [09:00] or develop a robot answering for you [09:00] schnap it [09:00] heh === greyback__ is now known as greyback === alan_g is now known as alan_g|llunch [09:56] Laney, do you need to fetch the value on startup? Can't you signal that it's ready after the result comes back? [09:56] it was ok before because local file io is generally considered ok to do blocking [09:56] not so for d-bus calls (imo, desrt thinks otherwise) [09:57] otherwise it looks good to me [09:57] will give it a spin now and then comment. Thanks for working on this! [10:04] larsu: I think the object should be ready from the start (also not having it so makes one of the old testcases fail) [10:05] Laney, yeah ideally it would be, but it's slowing down the rest of the program [10:05] and might cause issues with the other main loop (which uses the same main context) [10:06] it's weird to start up in (say) UTC and then reconfigure later [10:06] could I make a new main context? [10:07] in a new thread... [10:08] and it's a bit overkill... [10:08] is it just a test case that fails or does the real program depend on this too? [10:09] OH MY GOD STOP [10:09] dude [10:09] (sorry -- larsu just told me) :) [10:09] I did not! [10:09] I simply explained the problem to you [10:10] i'm actually okay with the sync dbus call [10:10] just don't recurse the global mainloop -- ever [10:10] it's just slightly more annoying to write all that shit again [10:10] it's not optimal [10:10] but in your situation it's probably something like the optimal solution [10:10] when I just want to say hey, wait here until you get the value [10:11] * larsu gives in becasue it's probably not that big of a deal [10:11] but keep in mind that datetimed is not normally running [10:11] so this sync call is going to be blocking on a dbus activation [10:11] ugh [10:11] which is sort of a good point [10:11] why are you using dbus for this? [10:11] instead of what? [10:11] read /etc/localtime? [10:12] someone broke gfilemonitor [10:12] this is what he was trying to get rid of [10:12] ya... [10:12] so use dbus to watch for the change signal [10:12] and then someone else said you should use timedated instead [10:12] but read the file [10:12] yes [10:12] use timedated but only for the signal [10:12] s/someone/larsu? [10:12] do the reading locally [10:12] jesus christ [10:12] seriously [10:12] this is how it's "supposed" to work with dconf-like services [10:12] which is more or less everything these days [10:12] writing and change monitoring = dbus [10:12] reading = do it directly yourself [10:13] meh, this is overkill [10:13] what is propertieschanged for then [10:13] you'd still use that [10:13] there are all kinds of dbus services.... [10:13] but then ignore the value that it gives you [10:13] and go read it off the FS [10:13] that is some kind of bullshit [10:13] * desrt shrugs [10:13] * larsu shrugs harder and is hungry [10:13] activating a dbus service for no good reason is some kind of bullshit [10:14] use the filesystem [10:14] adding a matchrule doesn't activate the service [10:16] * ricotz isn't really following, but gfilemonitor is really broken currently [10:16] Laney: do you have a specific bug for this case i can look at? [10:16] iirc you only sent trash ones yesterday [10:16] https://bugs.launchpad.net/ubuntu/+source/indicator-datetime/+bug/1485659 [10:16] Ubuntu bug 1485659 in indicator-datetime (Ubuntu) "Date/time indicator doesn't update after changing time zone" [Low,In progress] [10:17] we're going to lunch now. i'll look at this as soon as i'm back [10:17] I could maybe get behind reading for the initial value [10:17] but ya in general -- i agree with larsu about watching dbus for the change [10:17] but please use the disk for reading [10:17] since activating the service is really expensive [10:17] but if we get a new one in PropertiesChanged [10:17] then.... [10:17] ya [10:17] this does make sense [10:17] i mean, why read the disk again if you just got told, right? [10:17] yes [10:17] that is the "bullshit" I mentioned earlier [10:18] i'm 100% neutral on this point [10:18] i'd probably read the disk just because this tends to be how i structure my code [10:18] (and since the read is almost free) [10:18] but no harm in using the info from the signal, either [10:18] and luckily the original code had a function to read this /etc/timezone thing [10:18] but on startup.... ya... please don't activate the service [10:18] so can probably just bring that back [10:18] this makes sense to me [10:18] also: watch for the obnoxious race [10:19] gdbus doesn't make this easy :( [10:19] (honestly, it's probably safe to ignore it) [10:19] which is? [10:20] 1) send AddMatch via worker thread [10:20] 2) read from disk [10:20] 3) timezone is changed, signal is sent, match rule wasn't handled by dbus-daemon yet, you miss the change [10:20] 4) AddMatch is delivered and processed [10:21] gdbus actually kinda makes this situation shitty [10:21] but again, it's unlikely to hit you [10:21] you can work around it if you must by pinging the daemon [10:21] as a sync call, after adding the matchrule [10:22] I need to do what I am supposed to be doing instead of this bug [10:22] will look later [10:22] G_DBUS_SIGNAL_FLAGS_NO_MATCH_RULE got added for exactly this reason [10:22] since dconf wants to send its own match rules so it knows when the match is actually active [10:22] → food [10:23] you can fix GFileMonitor too if you want :) [10:23] that comes after food :) [10:23] * desrt kicks larsu away from irc [10:56] BinLi: hi, I hope you're the right BinLi. Can you look to fix wily for https://code.launchpad.net/~binli/ubuntu/trusty/modemmanager/lp1441095/+merge/255935 per seb128's comment please? [10:57] If it's cherry-picks from upstream this can probably go via Debian and maybe cyphermox can help you there [11:38] Laney: i think this bug was the atomic replacement bug, already fixed [11:39] HI [11:39] fixed where? [11:39] master? [11:39] gvfs-monitor-file /etc/localtime out of jhbuild works [11:39] can you test it with the version you have? [11:48] oh yeah, this works now [11:50] larsu: https://code.launchpad.net/~octoquad/software-center/899878-hardcoded-colors-dark-theme/+merge/268849 is this sane? [11:51] Laney, why does software center care at all about the theme it's running under [11:52] larsu: it uses css (and it's gtk2, so no upstream support), no? [11:52] didrocks, it can't be gtk2 and use css [11:52] or is that for the web view? [11:52] web view [11:52] nah it is gtk3 [11:52] (some parts are) === alan_g|llunch is now known as alan_g [11:53] but yeah, Laney is right, it's gtk3 now [11:53] clearly this is gtk css [11:53] it does have its own css though [11:53] weird [11:53] so if I run my desktop in a theme that software-center doesn't know about, it looks like shit? [11:54] anyway: fine by me if that's what it's doing. Lacking rationale for the transparency adjustment and includes whitespace fixes which should be a separate commit [11:54] (or just remove them, it's not that important) [11:57] larsu: I think I know looking at the code [11:57] as it has its own widgets, using cairo surfaces to render [11:57] ah. weird [11:57] I guess it tries to grab the value from the theme + additional properties [11:57] like in softwarecenter/ui/gtk3/widgets/buttons.py [11:58] makes sense [11:59] didrocks: you want to handle the upload? :) :) [11:59] Laney: that would be tomorrow for me, but sure [11:59] aren't you PPing? :P [12:00] Laney: when I saw your name this morning, I think I'll be better to handle that tomorrow, as there are too many times there is no-one [12:00] dholbach's mail really helped, eh? Now everyones PPing [12:00] I always pee pee [12:00] so better at least to have at least one dude every day than 2 one day and then no-one [12:00] is it? [12:00] there's still 99999 items [12:01] I guess it's better for online presence [12:03] * Laney shrugs [12:03] dude! [12:04] duuuuuuuuuuuuuude! [12:05] up to you man, you do it when you think is best [12:07] yep, hence tomorrow ;) [12:07] * didrocks opens moaarrrr sdk bugs meanwhile [12:07] then, I'll consider my example app "done" [12:07] with nice margin for Laney :) [12:15] \o/ === MacSlow is now known as MacSlow|lunch [13:26] MacSlow|lunch: enjoying your meal? [13:52] Laney: the conflicts on that branch should be resolved now === MacSlow|lunch is now known as MacSlow [14:25] * qengho afk a bit. [14:26] attente: nice [14:26] you're my favourite! [14:28] :* [14:30] Laney: wait a second. can i check something before you approve [14:30] ummmmmMMMMMMmmmmmmmmmmmmmmmmmmmmM [14:35] this is going to take a bit of getting used to [14:35] * Laney got a trackball maus [14:37] git bzr push lp:~/notify-osd/crisper-on-hidpi [14:37] phew [14:37] mmm crisps [14:37] also ++ for using bzr-git-bzr-git [14:38] :) [14:38] get someone who knows this shit to review that [14:39] Laney, who would that be? [14:39] can you give it a try? [14:39] not reviewing, just seeing how it works for you [14:39] dezrt? :) [14:40] or MacZlow himzelf? :P [14:40] she's spelled with an s [14:40] like so: desrt [14:40] :P [14:40] * Laney iz avoiding highlighting people [14:40] * larsu is destroying Laney's plan [14:40] at least I avoid the blame [14:40] :) [14:40] * didrocks is destroying larsu's plan to destroy Laney's one [14:40] yes I will try this [14:40] uh oh [14:40] this is getting meta [14:41] Laney, thanks :) [14:41] MR is up https://code.launchpad.net/~larsu/notify-osd/crisper-on-hidpi/+merge/269930 [14:43] and with this last bug opened, I consider my example app done! [14:43] Laney: yeah... should be ok :o [14:47] didrocks: get it in ze store! [14:51] and make $$$$$ [14:51] larsu: it sure looks crisp to me [14:52] glad to hear :) [14:52] thanks for testing! [14:53] thanks for fixing! [14:53] was a bit annoying indeed [14:53] * larsu sat on a particular issue all morning [14:53] we should redo this code to use proper widgets and gtk css (which wasn't available back then) [14:53] and i thought you sat next to dh [14:54] *dholbach [14:54] I did, and still do [14:54] * larsu hugs dholbach [14:54] man, he's not even in here anymore [14:54] yeah [14:54] snappy ate him [14:54] [14:54] haha [14:57] * Laney puts away the second keyboard [14:57] * Laney is getting into good habits [15:43] this file thing fucks up the testsuite [15:44] * qengho back [15:44] bah [15:44] be strong Laney [15:44] it calls a mocked SetTimezone [15:44] which doesn't write to /etc/timezone of course [15:44] :| [15:45] but should emit the signal now? [15:45] we don't wait for that any more [15:45] right [15:45] it's like oh hi I got an object [15:45] oh it's Europe/London not the thing I just set it to [15:45] this is annoying [15:46] good thing is that the old testsuite had some shit for this [15:46] I'm basicaly undoing half of my work [15:46] :( [17:04] bye! [17:04] I got the tests to pass again [17:04] not sure if some shit is weird or not [17:05] ready for more abuse tomorrow :)