=== m_conley is now known as m_conley_away [00:52] Laney: can you add libkkc and ibus-kkc to input-method pkgset? I've sent the request email just now. === Trevinho_ is now known as Trevinho === meetingology` is now known as meetingology === Guest97591 is now known as Guest20083 === siel_ is now known as siel === broder_ is now known as broder === wdean_ is now known as wdean === laza_ is now known as laza === Guest20083 is now known as Ursinha === duflu_ is now known as duflu [06:47] good morning desktopers [06:47] dpm, hey [06:48] dpm, the fixes for the indicator-location and content-hub translations landed on the current rtm image, if you fancy testing/confirming those later [06:51] seb128, cool, I've already seen the indicator-location fix working, let me test the content-hub one [06:51] morning all [06:52] dpm, the content hub might need the next langpack update, the template was incomplete so you miss translations on disk [06:52] ah, I see [06:52] I tested by exporting the .mo from launchpad and copying it manually [06:52] seb128, do you know if pitti is back? It'd be good to release a langpack this week [06:52] works here, I'm just asking to make sure somebody test on a device non hacked locally [06:53] dpm, those are cron-ed and updated weekly afaik? [06:53] dpm: he's at plumbers [06:53] dpm, he's ... what larsu said [06:53] he's around on IRC at times [06:53] just ask and he's likely to get back to you during the day [06:53] ok [06:54] pitti, if you happen to be around, would it be possible to arrange a touch langpacks release this week? Or give me some info on how someone could do that while you're away? [07:01] dpm, those happen on friday automatically no? [07:02] seb128, the exports should happen automatically, but I'm not sure if pitti has automated the phone langpack package builds or if he triggers them manually [07:03] oh ok === bigon_ is now known as bigon === fredp` is now known as fredp [07:33] dpm: well, I'm not really "away", I just don't pay attention to IRC; email works much better this week [07:34] pitti, o/ [07:34] dpm: we have automatic touch langpacks every week [07:34] dpm: for RTM they happen on Thursday evenings [07:36] pitti, cool, thanks. Are the exports from LP full exports only? I remember you requested that to wgrant, but I can still see deltas at https://translations.launchpad.net/ubuntu-rtm/14.09/+language-packs - i.e. do we need to manually request full exports before every Thursday? [07:37] dpm: they are not, I currently need to request them manually; I added a calendar event for that now [07:37] ok, gotcha [07:51] morning all [07:52] hey willcooke [07:52] how are you? [07:52] hey pitti, how is the conf? [07:53] hey seb128 - pretty good, I think my slides are "done" [07:53] \o/ [07:53] seb128: quite nice so far, thanks; I have desrt sitting next to me now, he made it :) [07:53] hey desrt - you managed to get in! ;) [07:54] pitti, are you using windows? [07:55] willcooke: into germany? :P [07:55] willcooke: (morning) [07:55] seb128: yes, every day -- that's how I can see what's outside and let fresh air in [07:55] larsu, ha :D [07:55] * larsu high-fives pitti [07:56] pitti, :-) [07:56] pitti, from another channel, and the guy in the audience beside me is running windows [07:56] ;-) [07:56] seb128: if you mean the MS product, I believe I still have some Win7 image collecting virtual dust on my USB HDD [07:56] willcooke, well done on the slides! [07:57] pitti, yeah, I was just wondering because of that quote ^ [07:57] seb128, olli has checked them, but I've got to get past rickspencer3 still :) [07:57] I'm sure it's going to be fine! === tiheum is now known as tiheum_ [08:05] hallo! [08:06] hey Laney [08:07] hey Laney, pitti, seb128 [08:07] hey pitti, how's DUS? [08:07] hey darkxst [08:07] what up darkxst [08:07] hey Laney [08:07] hey darkxst, how are you? [08:07] how are you? [08:07] happyaron: yeah, I usually read emails so no need to follow up with a ping :-) [08:07] I'm good, but I don't like painting! [08:07] Laney, I published the u-s-d update, hope it's ok [08:07] it's in unapproved still though [08:07] YOU DID WHAT!!!!!!! [08:07] ;-) [08:08] that button is quite tempting eh [08:08] indeed [08:08] shame that you set up that landing [08:08] i'm sure somebody will review it [08:08] if I did it you could be waving it through unapproved [08:08] it's like 25 lines instead of 999999 that the other one was [08:08] seb128, or Laney can you look at bug 1381297, would be good to get that through before final freeze [08:08] bug 1381297 in gnome-session (Ubuntu) "gnome-classic sessions broken with upstart user sessions" [Undecided,Confirmed] https://launchpad.net/bugs/1381297 [08:19] willcooke: why would i have had trouble getting in? :) [08:19] desrt: ceta [08:19] desrt, could flight to Germany and handle Frankfurt airport while sleeping [08:19] -, [08:20] it's true [08:20] seb128: I think he needs to be awake in order to make himself stay home. His default state is to get into a plane to FRA [08:20] minus the crazy guy attacking me [08:21] :) [08:21] random thought of the day: man, GOptionContext is bad [08:21] larsu: it's not _that_ bad [08:21] desrt: it's pretty bad [08:22] faced with the choice of killing it off and rewriting it or giving it new life, i elected new life [08:22] by putting it on gapplication? [08:22] ya [08:22] it's quite workable for the sort of things that most apps want to do [08:22] it doesn't do most of the stuff modern command line tools need [08:22] bad API, sure.... but we can take care of that with the gapp wrapping we do [08:22] like positional arguments [08:22] positional args [08:22] or command-line structure [08:22] ? [08:23] desrt: "command-like". Think git [08:23] i added support for that recently [08:23] really? How does it work? [08:24] hm. maybe it never landed?! [08:24] also, aliases would be nice. i.e., --no-color ==> --color=never, etc [08:24] because now you need to keep all kinds of booleans around which might conflict [08:24] and you never find out which one was specified last [08:24] man, this really is a bad api [08:24] desrt: I know, patches welcome [08:24] * larsu stops ranting already [08:25] i added a 'bsd mode' to goption [08:25] aka posixly-correct mode [08:26] how does that differ? [08:26] if you do this: [08:26] ./mycmd -x foo -y [08:27] then it stops at "foo" and treats the rest as filenames [08:27] including the -y [08:27] which can be repurposed for git-style commands by treating the filename stuff as 'the rest' and invoking a second time with the options of the subcommand [08:27] seb128, Laney: gnome-terminal in bzr is working fine for me [08:27] https://bugzilla.gnome.org/show_bug.cgi?id=723160 [08:27] Gnome bug 723160 in general "GOption: add strict posix mode" [Normal,New] [08:28] yup -- never got committed [08:28] Laney: I documented dropping these two patches, forward-ported the LP: # one, and added the patch for the cwd issue [08:28] pitti: cool, looking forward to V [08:28] desrt: right, but that still leaves too much work for the application [08:28] pitti, k [08:28] larsu: i disagree [08:28] Laney: NB I kept 3.12.2 for now, as 3.14 needs a merge of vte (0.3.8) [08:28] pitti, thanks for the work [08:28] you need to pick the subcommand first before you know what you want to do anyway [08:28] * Laney nods [08:28] larsu: reviews welcome on that bug.... [08:28] seb128: yay for the last item in my ~/.gconf gone except for chrome :) [08:29] but aside from merging vte, updating g-t from 3.12 to 3.14 should be reasonably easy [08:29] did the profile migration work properly? [08:29] desrt: why do you need a posixly correct? [08:29] Laney: yeah; I maded changes in 3.6 (background color, font, etc.), dist-upgraded, restarted g-t, and everywhing was correct [08:30] great [08:30] larsu: consider git for example... it has arguments to the global thing (--git-dir for example) and separately arguments to subcommands like git clean -xdf or whatever [08:30] so consider: [08:30] git --git-dir=foo clean -xdf [08:30] you want to be able to parse the --git-dir=foo and leave the "clean -xdf" unmodified [08:30] once you see "clean" you can then parse for the subcommand and get the -xdf handled [08:31] if you don't do the posixally-correct parsing then the -xdf will get sucked up in the first pass [08:32] fair enough (even though it seems like quite an edge case) [08:32] I'll review your patch [08:32] it's not an edge case: it's precisely the feature that you were complaining about being missing :p [08:32] it also happens to be the expected behaviour on BSD [08:33] (or if you set POSIXALLY_CORRECT on GNU) [08:34] desrt: it's not what I was complaining about. It enables a feature (global and command-specific options) for a feature I was complaining about (no proper support for commands) [08:34] i'm not sure how you could possibly have a better API for that [08:35] i don't want to feed the global information about every possibly subcommand into the parser at the start [08:35] i want to first discover the subcommand and then do the options only for that subcommand [08:35] by its natures that involves two passes... [08:35] *nature [08:35] but that messes up the help message goptioncontext spits out [08:36] how? [08:36] xxx --help : shows help for toplevels [08:36] xxx foo --help: shows help for 'foo' subcommand [08:38] desrt: of course, but you can't let it generate the xxx --help, because it doesn't label things right [08:38] desrt: which is why gdbus prints it's own message there [08:38] I'm not saying doing these kinds of things is impossible, just that it is hard [08:38] so let's get this patch lined up and rewrite one of our subcommand tools [08:38] and see what issues pop up [08:38] * larsu is up for that [08:39] desrt: in fact, I want a new gdbus, for a couple of reasons: [08:39] commandline parsing is just so very boring :p [08:39] (1) it's UI is really bad (--dest?) [08:39] (2) it's output could be pretty-printed and syntax highlighted [08:39] we should invent some quasi-uri scheme [08:39] (3) monitor doesn't monitor anything but signals [08:40] session://ca.desrt.ca/ca/desrt/dconf or so [08:40] I thought about that, but dismissed for another thing to learn [08:40] ya... monitor is bad [08:40] we should fix that [08:40] when i find myself using dbus-monitor instead you know there is a problem [08:40] I'd much prefer: gdbus call [-e] ca.desrt.dconf /ca/desrt/dconf ca.desrt.dconf.SomeMethod [08:41] desrt: ya... [08:41] anyway.... feel free to work on this area with my blessings :) [08:42] i don't expect to have a lot of time for it unfortunately [08:42] thanks, same here [08:42] but if we could get that posixly-correct patch in at least i'd be happy about that [08:42] we could do a session on the plane [08:42] we should try to get seats together [08:42] i can sleep on your shoulder [08:43] sit next to you? nah [08:43] :P [08:43] pitti is on our flight too, eh? [08:43] is he? Nice! [08:43] too bad they don't have those cool little rooms like on the train [08:43] they do [08:43] too bad they don't have them in 2nd class :p [08:43] just not on every plane, and they're something like $15,000 [08:44] * desrt had a private room right at the front of the train with a view through the driver's front window yesterday [08:44] they give you free candy when you're in 1st [08:44] meanwhile, we have the "we should get rid of drivers" discussion again [08:45] in light of all the recent strikes [08:45] oh. those kinds of drivers :) [08:45] strikes in situations like this are hilarious [08:45] situations like this? [08:45] when someone's job could be replaced by a machine [08:45] and they strike [08:45] but, moar money! [09:08] darkxst: where again did you get the idle monitor from? Current g-s-d seems to get one called gnome-idle-monitor (instead of gsd-) from libgnome-desktop [09:30] larsu: It comes from mutter (3.10, I believe) [09:31] that's what does the idle monitoring for shell, and gnome-desktop calls into that using dbus [09:34] Laney: weird. thanks [09:34] I think it's part of de-Xification [09:35] maybe we should have that in compiz as well then? [09:35] oh wait... [09:53] larsu, yes as Laney said, idle-monitor was moved into mutter and abstracted so gnome-desktop doesn't have any deps on X [09:55] then mutter has various backends for X, wayland etc [09:59] darkxst: got it. The idle monitor code in mutter seems to have changed a lot [09:59] darkxst: I was about to backport my changes there, but I think that doesn't make much sense [10:02] larsu, right, and its a bit tangled up in other libs as well, but maybe the monitors ref could be useful [10:03] g-s-d still calls via gome-desktop so the gsd-cursor snippet probably doesnt [10:03] right, I just noticed that as well [10:03] (maybe?) [10:05] darkxst: did you check your gnome-session change? [10:05] I don't think that shell is right [10:10] anyway lemme fix and upload [10:12] Laney, yes I tested but seems I pushed an older branch [10:16] Laney, https://code.launchpad.net/~darkxst/gnome-session/lp1381297 is what I tested [11:16] blerg [11:16] evince bumped the required gtk from 3.12 to 3.14 between .0 and .1 === MacSlow is now known as MacSlow|lunch [12:04] Laney, well, they already required 3.13, they just forgot to update the configure [12:04] see the revert of gesture use patch I added [12:04] blerg! [12:05] fixed it to use GTK_CHECK_VERSION [12:11] what's the point? [12:12] we build our evince with a known gtk, when we update gtk we drop the patch [12:12] or do you want upstream to have the ifdef cases and lower the requirement? [12:12] not for the gesture stuff, the new patch in .1 [12:12] it's trivial to do [12:16] k [12:16] still what's the point? they don't support GTK 3.12 due to the gesture api use which is > 3.13 [12:18] it's not for forwarding [12:21] hmmm, my firefox on trusty just killed itself after hanging on a flash player fail (vs. youtube, which should be somewhat sane). After coming back, it lost all passwords ... [12:24] * Laney goes to the cafe for lunch, biab [12:26] Laney, k, enjoy lunch! [12:27] Sweetsha1k, instance still running locking your profile? [12:27] or talk to chrisccoulson, he might be able to help you [12:27] do you room with him next week? ;-) [12:33] seb128: no, Ill room with the guy who will say: *shrugs* just use chromium ;) [12:35] seb128: also: no, can be a locked profile. E.g. my cockies, extensions etc. are all still there. just the passwords are gone === Sweetsha1k is now known as Sweeshark === Sweeshark is now known as Sweetshark [12:35] weird [12:36] there might still be a file on disk that you might be able to copy over [12:36] dpm, hey, string translation question for you ;-) [12:37] dpm, do you know if there is a standard way to deal with several singular/plural forms in one string? [12:38] dpm, like "copy %d folder(s) and %d directory/ies" === MacSlow|lunch is now known as MacSlow [12:55] * dpm thinks about it [12:56] don't we have ngettext for these kinds of strings? [12:57] seb128, hm , I'm afraid I don't know, I've never come across that, and I can't think of the best way to handle those. I think perhaps rephrase it so that it can be done in 2 sentences? [12:57] dpm, I think we need code cases [12:57] larsu, no, ngettext is for one argument [12:57] larsu, e.g "copy %d folder" "copy %d folders" [12:57] singular/plural [12:58] but if you have 2 arguments I don't think it can mix 1 being singular and 2 being plural [12:58] and all possible combinaison [12:58] copy 3 folders and 1 file [12:58] copy 3 folders and 2 files [12:58] copy 1 folder and 6 files [12:58] etc [12:59] you could split the string [12:59] split how? [12:59] but there's probably some weird languge for which that doesn't work [12:59] or you mean change the design, reword? [12:59] seb128: no, have two translatable string that you concatenate [12:59] like "copy %d folder" + "and %d files" [13:00] yeah [13:00] hehe, ya [13:00] (had the same example typed) [13:00] what if you need to change the order or words in some locales? [13:00] I don't think that works [13:00] yeah this is what I meant - it probably won't work for some language [13:02] I can do a complex [13:02] the alternative is to have 9 translatable strings :-/ [13:02] or just do what you proposed [13:02] if n_files = 1 & n_folder > 1 string = [13:02] n_files > 1 & n_folder > 1, string = [13:02] etc [13:02] seb128: all 9 cases? [13:03] poor translators [13:03] I don't count 9 [13:03] but 3 [13:03] files = plural [13:03] folder = plural [13:03] both = plural [13:03] or none [13:03] 0,0 - 0,1 - 0,n - 1,0 - 1,1 - 1,n - n,0 - n,1 - n,n [13:04] 0,1 and 0,n are one case [13:04] it's "copy %d files" in ngettext [13:04] and 0,0 doesn't exist [13:04] it's a non copy [13:05] but yeah, then you still get quite some cases :-/ [13:05] that can't be, that's even wrong in english: "copy 1 files" [13:05] no, ngettext is [13:05] but yeah, cross out (0,0) :) [13:05] "copy %d file", "copy %d files", n_files [13:05] one string with singular/plural forms [13:05] but you can indeed count that as 2 strings [13:06] right, but you also have the cases where you only copy directories or only files, no? [13:06] yeah [13:06] ah, now I understand [13:06] sorry [13:06] to be fair that code already sucks [13:07] http://bazaar.launchpad.net/~phablet-team/gallery-app/trunk/view/head:/rc/qml/EventsOverview.qml#L67 [13:07] it already has the 9 (or rather 8) strings [13:07] uh oh [13:07] it's just not using ngettext in the plural cases [13:08] so I guess I can just [13:08] -i18n.tr("Delete %1 photos") [13:08] +i18n.tr("Delete %1 photo", "Delete %1 photo", organicEventView.selection.selectedPhotosCount) [13:08] ups [13:08] the second arg misses a "s" [13:08] I guess, yes [13:09] larsu, dpm: thanks for the discussion [13:09] that feels suboptimal but I guess it's the only option [13:09] out of changing the wording [13:09] seb128, np, sorry I couldn't help more [13:09] nw! === dobey_ is now known as dobey [13:36] Sweetshark, so libixion -- how do you want to proceed? I like split packages in general, but not splitting it for 14.10 seemed like an option [13:39] mterry: personally, I would go for libreoffice shipping its own bundled versions and not doing any releases of liborcus/libixion outside of that at all. At least not as long there is no clearcut request for that. [13:39] Sweetshark, does anyone use liborcus today? [13:40] mterry: I am not aware of any users of liborcus besides libreoffice. [13:41] mterry: did you see https://lists.ubuntu.com/archives/ubuntu-devel/2014-October/038510.html ? [13:42] Sweetshark, yes. The libixion part was interesting [13:42] Sweetshark, it sounds like it's easiest just to punt this for now. But I like when we can just piggyback on Debian [13:43] mterry: yes. [13:45] Sweetshark, commented in MIR bug [13:47] mterry: thanks! [13:54] Please, backport these patch: https://github.com/GNOME/gtk/commit/bde4e863579f05368661347bfaf55e0480e4555c [13:54] https://github.com/GNOME/gtk/commit/60cd7076cf885cbc421babfba44c9081ded3ad78 [13:54] they are useful for have a '.csd' class on the windows CSDed [13:54] l3on, can you open a bug about that and subscribe ubuntu-sponsors? [13:54] seb128, sure [13:55] thanks [13:57] seb128, done https://bugs.launchpad.net/ubuntu/+source/gtk+3.0/+bug/1381546 [13:57] Ubuntu bug 1381546 in gtk+3.0 (Ubuntu) "Please backport patches to add '.csd' class to CSDed windows" [Undecided,New] [13:58] l3on, thanks [13:58] larsu, ^ want to review that? [13:58] it's for next cycle anyway [13:58] so no need to be today [13:58] it might not be needed if we update gtk early in the cycle [13:59] I think it should be better have them now... they are really necessary to distinguish windows CSDed in themes.. [14:00] well, we don't have proper csd support in unity in utopic [14:00] so I doubt theming makes a difference [14:00] that's work for next cycle [14:00] no? [14:00] what about non-unity desktops? [14:00] like? [14:00] gnome ? [14:01] they are on 3.12 and I don't think their theme is making use of those classes? [14:02] well I need those patch for my ubuntu-themes branch which adds support to CSD windows.. but without that patch I can't recognize which windows using GTKHEader bar have been patched in Ubuntu to use Unity borders (like Nautilus) .. that's all. [14:04] k [14:04] well I don't say it's not useful [14:04] it's just that it's late in the cycle for such changes [14:04] better to SRU that [14:04] it's not that important for the iso itself [14:04] imho [14:06] seb128: will do [14:06] larsu, thanks [14:16] dpm, larsu, in fact the several cases doesn't work either :/ [14:17] why not? [14:17] you can't do "delete %1 photos and %2 videos" in cases [14:17] some locales have different cases for n=5 or =10 or between 10 and 20 [14:18] you can't have a combinaison of photos = 5, video = 2 [14:18] and ngettext supports those? [14:18] well, if you see what I mean [14:18] yes [14:18] but for 1 argument [14:18] right [14:18] you do "copy %d files" [14:18] and it spill the cases for your locales [14:18] so basically this is impossible because of combinatorical explosion [14:18] like [1] [2] [5] [others] [14:18] maybe concatenating the strings wasn't the worst idea then? [14:18] but it doesn't do it on substrings, or combinaisons for 2 words [14:19] yeah, I don't see a better way around [14:22] we could make mpt eat his hat and do something like: Do you want to delete the following things? [14:22] * %d files [14:22] * %d folders [14:23] I was just looking at the relevant part of the gettext manual an hour ago [14:23] I don't think gettext can do it [14:23] concatenating is a bad idea [14:23] mpt: I was joking and if I weren't, I hope you have a hat made out of chocolate [14:23] or something equally tasty [14:24] Laney: why? Are there languages that switch the order of terms in "X and Y"? [14:24] larsu, https://www.gnu.org/software/gettext/manual/gettext.html#Special-cases [14:25] wait, wrong section [14:25] I don't know all languages in the world, sorry [14:26] Laney: you seem to know that concatenating is a bad idea, though [14:26] Laney, contenating sucks but it's the only option [14:26] that might work [14:27] Laney, we can also give translator comment for both halfs "that's the first/second part of that sentence" [14:27] and let them trick the 2 subparts [14:27] what if you have to put something in the first part depending on the second one? [14:28] you are screwed [14:28] but I think it's less likely than have one sentence with %1 photos and %2 videos [14:28] and %1 and %2 being e.g 2 and 5 which are different forms in russian [14:29] have->having [14:29] Laney, if you have a better suggestion I'm open to it though, just can't think of any [14:29] (out of changing the design) [14:29] (which is not up to me) [14:31] i18n is a bitch [14:32] seb128: I'm fine taking those csd patches. They're very unintrusive [14:32] larsu, thanks, can you write that on the bug? [14:32] l3on, ^ [14:32] seb128: it's a tad late for this cycle though, no? [14:32] larsu, I was suggesting to SRU those [14:32]