[02:47] pitti, currently removing libpam-systemd breaks gnome-shell since the upstart script is still run, is there anything we can do about that? [03:38] good morning [03:38] darkxst: you mean if you remove it, but not purge? [03:39] pitti, yes [03:39] darkxst: we could change the upstart job to check whether libpam-systemd is actually installed [03:39] its on staging now, but requires the manual purge step if people want to remove it [03:39] pitti: isn't it around 5:30am for you? [03:39] thumper: yes [03:39] geez [03:39] crazy [03:39] darkxst: so, it's a relatively simple fix, but I think at this point it needs to become an SRU [03:40] pitti: oh, and hai [03:40] darkxst: if you can file a bug for it, I'll upload a fix [03:40] pitti, ok, thanks [03:40] thumper: my wife gets up at 5, and now that it becomes summer I slowly adapt to getting up early, too :) [03:40] * thumper nods [03:43] pitti: is she sadistic? [03:43] lifeless: not in general, just an early bird :) [03:44] pitti, https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1171691 [03:44] Launchpad bug 1171691 in systemd (Ubuntu) "Removing libpam-systemd (without purge) breaks system" [Undecided,New] [03:47] darkxst: thanks === rsalveti_ is now known as rsalveti [04:49] g'day [05:39] darkxst: I uploaded a fix [05:43] pitti, great! thanks [07:34] pitti, suspend on lid closed is handled in logind? [07:35] darkxst: yes; gnome-settings-daemon's power plugin can set an inhibitor in some cases, but generally logind does that [07:35] hey desktopers [07:36] bonjour seb128 [07:36] lut pitti, ca va bie n? [07:36] salut pitti, seb128! [07:36] seb128: oui, merci ! et toi ? [07:36] bonjour didrocks [07:36] lut didrocks [07:36] pitti, je vais bien, merci ;-) [07:39] pitti, it should work? (beacuse it isnt) [07:40] darkxst: I haven't tested that yet; if it doesn't work, I guess that's something which again depends on systemd init, so we need a fallback [07:40] right === czajkows1i is now known as czajkowski [07:51] pitti, seems to use the same target as a normal suspend, but its not firing for some reason [07:58] perhaps its being inhibited for some reason? but I can't see anything logged ;( [08:00] darkxst: you can try from gdm/lightdm, or from a VT [08:02] pitti, nope doesnt work there either [08:02] so I guess it relies on systemd-init to send the actual event [08:02] bonjour mes amis! [08:03] Laney: bonjour Laney, comment vas-tu aujourd'hui ? [08:03] très bien merci, et tu? [08:04] je vais bien aussi [08:04] mais je souhaite le retour du soleil [08:05] salut Laney! [08:05] it will be back soon :) [08:05] * didrocks apprécie le soleil ici [08:06] la semaine prochaine nous allons être à California :) [08:06] pitti, possibly, g-s-d does see the lid close event, so I guess we deal with it there [08:06] darkxst: raring's g-s-d is built against CK, so it does do the lid handling itself, yes; I assumed you had your own 3.8 version built against logind? [08:06] pitti, with a dbus call to org.gnome.logind1.suspend() perhaps? [08:07] aujourd'hui j'ai soleil ici :-) [08:07] darkxst: no, raring's g-s-d calls upower [08:07] pitti, yes we are built against 3.8 [08:07] Laney, salut [08:07] hallo seb128! wie gehts? [08:07] Laney, gut, danke! und dir? [08:08] prima! die Sonne scheint :-) [08:09] ;-) [08:13] darkxst: do a dbus-monitor while you suspend and see what calls are taking place [08:14] assuming you've correctly enabled logind everywhere, that is [08:15] darkxst, hey [08:16] darkxst, re. libgweather, I looked at it a bit last week while I was trying to fix indicator-weather, it doesn't seem to use an online service for the locations but rely on the woeid to be in the db [08:16] darkxst, e.g https://git.gnome.org/browse/libgweather/commit/data/Locations.xml.in?id=349f453c9ac8c8b464b631b6c53e1de30932bc77 [08:17] but there is only 1 location tagged for example so far [08:17] so that's not really useful [08:17] seb128, oh I see, gnome-weather actually uses the yr.no provider [08:18] although I though it also used yahoo if available [08:18] Laney, logind is enabled and working fine, I am not seeing any relevent dbus calls when closing lid [08:18] well, using the service is easy, indicator-weather was doing that and can keep doing it [08:19] it only requires somebody to own the project and ask for an appid key [08:26] seb128: jtv is about if you need to ask why a POT isn't importing, sorry catching up on reading mails just saw you comment on on on coe-doc [08:26] *core [08:26] Laney, when I use the actual suspend key there is a call to "prepareToSleep" on the logind1 service [08:26] darkxst: that's logind telling you it's about to sleep [08:27] if it's prepareToSleep "true" that is [08:27] "false" is when it's just finished sleeping (failed or succeeded) [08:27] czajkowski, hey, thanks, I was going to ask dpm [08:27] usually I just fix the inport selection by sorting out the name if it's not there [08:27] darkxst: you should be able to see whatever it is calling the Suspend method on logind if that's what's happening [08:27] but I cant see those pages to help where as jtv will know in a second whats up [08:29] seb128, on a call, will look into it in a minute. Do you have more context, or is it on the scrollback comments? [08:29] Laney, nope, all I see is that and "DelayInhibited" [08:29] Laney, and when actually closing the lid, I see nothing [08:29] dpm, hey, no hurry [08:30] darkxst: What does "dbus-send --system --dest=org.freedesktop.login1 --type=method_call /org/freedesktop/login1 --print-reply org.freedesktop.login1.Manager.CanSuspend" give you? [08:30] dpm, https://code.launchpad.net/~seb128/ubuntu-docs/updated-translations/+merge/160106/comments/352944 [08:30] Laney, that would be 'true' [08:30] dpm, those have a label "No import target selected yet.", dunno why [08:31] should be 'yes' 'no' or 'challenge' [08:31] and calling org.freedesktop.login1.Suspend works [08:31] right, so the problem must be in gsd somewhere then [08:31] Laney, yes sorry the answer is 'yes' [08:32] there doesnt seem to be anything in gsd that fires the event [08:32] so I guess its back to pitti's theory that the event comes from systemd init [08:33] seb128: the templates aren't set I think [08:53] pitti: petite question sur les WI! How much work would it be to have the WI tracker to support [nameA,nameB,nameC] blablabla: TODO [08:54] pitti: this would duplicate the tracker status per person [08:54] having just one status [08:54] I think there is the blueprint bot parser as well as the launchpad side to change? [08:54] didrocks: I don't know about the LP side (for https://launchpad.net/people/+me/+upcomingwork), the parsing is all done in LP now [08:55] pitti: oh? no more cron then? [08:55] didrocks: as for the DB side on status.u.c., it's a fundamental change as you would have to move from "a WI is assigned to a person" to "a set of persons" [08:55] didrocks: and rewrite the whole thing with that in mind [08:55] pitti: well, internally, it can be duplicated in the db [08:55] so frankly, it's better to create a team and assign it to that [08:56] didrocks: that would lead to wrong counts, though [08:56] ah? [08:56] as that would count as three WIs [08:56] whereas the intended meaning is certainly "this or that or that person should do it" [08:57] pitti: it's more "this is a topic, and those people are working on it" [08:57] well, topic == BP, not WI [08:57] pitti: right, but a weekly BP seems overkill, especially when you will have few WI [08:57] didrocks: so taking a step back, what is your interpretation of that syntax? [08:58] pitti: nameA, nameB and nameC are working together on that task [08:58] pitti: this would avoid having tons of WI like on https://blueprints.launchpad.net/ubuntu/+spec/client-1303-delivering-touch-apps-to-raring [08:58] pitti: those WI seems separated, but they are not [08:59] and all is interleaved [08:59] like transitionning to HUD2 is impacting a lot of components [08:59] with "needs merge A, B and C" [08:59] I think a WI for each is overkill [08:59] especially that we maybe need to merge D and E, so need to add a WI for the purpose of adding one [08:59] where fixing it will be few minutes [09:00] what's an example of duplication on that BP? [09:00] well, you don't have to create WIs for a three-minute thing [09:01] you can summarize them like "transition all ubuntu packages to libfoo2" [09:02] didrocks: so "working together" means "any of the persons can do it" (that would be a team), or "this one WI is really three WIs for each assigned person" (that would be syntactic sugar for just expanding them)? [09:04] pitti: I think it would mean "those 3 persons are working together to get that WI done, but we don't know/care about the great each 3 minutes individual subitems to do it" [09:04] pitti: duplicating the same items 3 times for each person won't give that intent of "working together to get X done" [09:04] didrocks: ok, so you mean the first interpretation [09:05] yeah [09:05] but it's not really a team as the ones we have per say [09:05] if it's such a trivial thing, why bother splitting it amongst n people? [09:05] can involve some people from our team, PS, foundation [09:05] pitti: the task itself is not trivial, it's made of trivial things [09:05] but the total amount is a big task [09:05] and we need tracking that one [09:06] for instance [09:06] so, I guess LP could make up teams on the fly for such groups [09:06] but TBH I don't see that happening [09:06] I won't put "touch into S" as such [09:06] but "transitionning to HUD2" is a perfect target [09:07] I would just leave the items unassigned, and everyone grabs a bunch when they have time; or pre-assign them and allow people to reassign [09:07] that way you also always have someone who is accountable [09:08] pitti: hum, this doesn't really give the intent though of some people working on parallel on that task [09:09] pitti: but if it was translated to "tihs one WI is separated in 3 WI on the db", I wouldn't really mind [09:09] didrocks: but that is the second interpretation, where you just said that's the one you don't want? [09:09] (first = "or", second = "multiply") [09:10] pitti: as long as it's not what is shown in the blueprint and there is only one status, I'm fine with it to be translated as 3 separates things [09:10] pitti: it's not or, it's "and" [09:10] "those 3 persons are working together to get that WI done" [09:10] that means each individual person has to have a separate status and flip it to DONE [09:10] which is essentially a multiplication, yes [09:11] but how would you do that, if there is only one line to change? [09:11] pitti: yeah, but that it doesn't represent the exact same intent ;) [09:11] if you have [seb128,didrocks,Laney]: foo, and Laney does his part, but you didn't yet, what would the status be? [09:11] pitti: well, those 3 people can only consider to have their work DONE once the whole item is DONE [09:11] basically [09:11] no, for those WI, there is not really "that's my part and I don't care of yours" [09:11] it's the exact opposite intent I hope we can get ;) [09:12] liek seb128, didrocks and Laney are working together to get foo done [09:12] even if seb128 and Laney are pestering didrocks to do his part he's slacking on :p [09:12] ok, so everyone does their part, or it doesn't matter who in particular does that WI? [09:12] pitti: right [09:12] right what? :-) (either-or) [09:13] sorry, but I don't think I understand this yet [09:13] pitti: everyone participates to this WI [09:13] pitti: but there is not real "delimited part" [09:13] didrocks: so when is that WI done? [09:13] i. e. would [seb128,didrocks,Laney]: foo expand to [09:13] once "foo" is DONE? :) [09:14] [seb128] do seb's part of foo: [09:14] [Laney] do Iain's part of foo: [09:14] etc.? [09:14] (that would be multiplication) [09:14] in some way… the idea is that if didrocks can't work on foo today, seb can cover him :) [09:14] or whoever of seb128, Laney gets to it first, flips it to DONE? [09:15] didrocks: well, if that's the idea, what's wrong with [09:15] [didrocks] foo [09:15] pitti: so idea is that the task will have at least, even a small part done by everyone of them [09:15] and Laney steals the WI if he has time, or the tech lead balances? [09:16] xnox: hi. What would be the best way to do an archive-index run for data found in a PPA? [09:17] pitti: right [09:17] jamesh: last time, I had to rsync the ppa, setup apt-archive to generate Contents file, then point archive indexer at it (hack it up a little) and then it generates an index. [09:18] pitti: the hidden goal is to have Laney doing everything TBH, but shhhh :p [09:18] jamesh: I've done it once, and maybe can setup up again, but more permament on like people.c.c. Which PPA are you after? [09:18] xnox: The unity experimental one [09:18] * Laney has been ignoring all of these hilights for a good reason ... [09:19] jamesh: there a few, can you point which one you want? =) [09:19] xnox: http://ppa.launchpad.net/ubuntu-unity/experimental-prevalidation/ubuntu [09:19] Laney: don't worry, that won't change you will have everything to do :p [09:20] jamesh: ack. I can't promise to do it today, but hopefully soonish (in the middle of raring release). When do you need it by? i should have some older packages with index somewhere, if this is for software-center testing. [09:22] xnox: if you can point me in the right direction, I could do a scan myself. [09:24] I've got the archive-index code here. I hit the "missing Contents" file problem, so it is mainly "download the PPA" and "build a Contents" file that I need [09:25] jamesh: right. So one can use recursive wget on http://ppa.launchpad.net/ubuntu-unity/experimental-prevalidation/ubuntu/ [09:26] seb128, ok, looking at it now... [09:26] jamesh: and then write an apt-ftparchive config file to generate Contents file. http://manpages.ubuntu.com/manpages/lucid/man1/apt-ftparchive.1.html let me give slightly better reference. [09:27] jamesh: something like http://debian.scribus.net/debian/apt-ftparchive.conf [09:28] jamesh: but twiddle it for your own paths. Please note that regexpes in archive scanner are very narrow, and expect to have copy-paste comments from the top of the file. [09:28] just like on archive.ubuntu.com [09:29] thank you [09:42] seb128, about the issue we discussed yesterday, do you know of any way to reproduce it? [09:42] in any application [09:42] mhr3, no, it happens on my box once every second week... === seif is now known as Guest93284 === ara is now known as Guest91539 === seif is now known as Guest16675 [10:55] good morning! === MacSlow is now known as MacSlow|lunch [11:37] cyphermox, hey, how are you? [11:38] seb128: hey. good, you? [11:38] I'm good thanks === MacSlow|lunch is now known as MacSlow === highvolt1ge is now known as highvoltage === highvolt1ge is now known as highvoltage [12:57] mvo: ping [13:03] pstolowski: pong [13:13] mvo: hey! [13:14] didrocks, i figured out how to version QML plugins https://code.launchpad.net/~ken-vandine/libhud-qt/versioned/+merge/160363 [13:15] kenvandine: \o/ [13:15] didrocks, what do you think of qtdeclarative5-hud1-plugin for a versioned package name? [13:15] * didrocks is eager to see the diff :) [13:15] kenvandine, good morning [13:15] mvo: I'm looking at a long-standing issue in applications lens, where we don't have icons for some installable stuff in more-suggestions (until you run software-center) [13:15] good morning seb128 [13:15] kenvandine: hud1 is the soname, or rather, the import name? [13:15] import version* [13:15] yeah [13:15] pstolowski: this is about on-demand-downloding of icons? [13:15] mvo: in general, yes [13:15] we could do hud1.0 there too [13:16] kenvandine: it should match the "import x" [13:16] so if it's 1.0 I think 1.0 would make sense :) [13:16] mvo: app-lens looks for icons for uninstalled stuff in ~/.cache/software-center/icons and in /usr/share/app-install/icons [13:16] kenvandine: btw, why -plugin? [13:16] so you prefer qtdeclarative5-hud1.0-plugin [13:16] well, that's the convention we've been using already [13:16] i would prefer we drop that :) [13:17] qtdeclarative5-hud1.0 [13:17] mvo: what exactly lands in app-install-data package? (and why not all the icons, just some ;))? [13:17] kenvandine: let's use qtdeclarative5-hud1.0 [13:17] great [13:17] kenvandine: do you mind adding Replaces/Break/conflicts? [13:17] kenvandine: for the transition [13:17] we shouldn't need to [13:17] since it is parallel installable [13:17] with the 0.1 version [13:18] and with the version in the ppa right now? [13:18] (the next ppaà [13:18] right [13:18] you rename some file [13:18] different path [13:18] but it's not parallely installable [13:18] usr/lib/*/qt5/qml ? [13:19] the file name are changing? [13:19] usr/lib/*/qt5/qml /Ubuntu/Hud.1/ [13:19] usr/lib/*/qt5/qml/Ubuntu/Hud.1/ [13:19] ah ;) [13:19] i saw that is how qtquick.2 was named [13:19] it has to be a single directory though, we can't do Ubuntu/HUD/1.0/ [13:20] maybe i should make it Ubuntu/HUD.1.0 [13:20] hum [13:20] what's the "Ubuntu" directory means? [13:20] namespace [13:20] import Ubuntu.HUD 1.0 [13:20] ok, so we can have no namespace [13:20] or one namespace [13:20] but not more? [13:21] so no import Ubuntu.Awesome.HUD 1.0 [13:21] for instance? [13:21] i think we can [13:21] pstolowski: the icons from software-center.ubuntu.com are downloaded on demand [13:21] kenvandine: ah, so only no version in namespace? [13:21] right [13:21] that didn't work [13:21] i couldn't find docs on why though :) [13:21] it was trial and error :) [13:22] and it's not possible to ship different qml file and having qml importing the HUD.qml, HUD1.0.qml? [13:22] pstolowski: they are more dynamic, I need to look if the data-provider downloads them currently, I vaguely remember it does, but I may be wrong [13:22] didrocks, it's a .so and no [13:22] there is a qmldir file in that directory [13:22] with the module name and plugin name [13:22] i tried versioning the .so [13:22] but that doesn't work [13:22] it expects it to be unversioned [13:23] but there is prior art for versioning the import dir [13:23] kenvandine: and how does it do the mapping that Ubuntu.HUD 1.0 is Ubuntu/HUD.1.0/ for instance? [13:23] kenvandine: there is no mapping? [13:23] it's just happens that the dir name is having a .so file which exposes 1.0? [13:23] i think it is in the plugin [13:23] the dir name doesn't identify that [13:24] i knows the difference [13:24] ok, it's weird that versionning the .so doesn't work [13:24] it doesn't find the plugin at all [13:24] when it's versioned [13:24] kenvandine: it expects a hud.so file? [13:24] basically [13:24] libhud-qml.so [13:24] ok, so lib-qml.so [13:25] and there can only be a single qmldir file in the directory [13:25] yeah [13:25] kenvandine: what's the qmldir is for in it? [13:25] well, the name of the file is in the qmldir file [13:25] kenvandine: sorry for my noobs questions :) [13:25] module Ubuntu.HUD [13:25] plugin hud-qml [13:25] is the contents [13:26] so plugin hud-qml == libhud-qml.so [13:26] and I guess no way to put version in it? [13:26] nope [13:26] that would have been cool [13:26] like [13:26] yeah [13:26] module Ubuntu.HUD\nplugin hud-qml1.0 1.0 [13:26] i was hoping for something like plugin hud-qml1 version 1.0 [13:27] plugin hud-qml1.1 1.1 [13:27] yeah [13:27] it doesn't do that :) [13:27] to define a mapping version <-> so filename [13:27] yep [13:27] ok, I think you came with the right decision then [13:27] dirname == soname [13:27] yeah [13:27] they did that with qtquick.2 already too [13:28] hum [13:28] but the import has to change? [13:28] no [13:28] as we change the namespace? [13:28] what magic? ;) [13:29] ah [13:29] module foo.bar [13:29] /usr/lib/x86_64-linux-gnu/qt5/qml/QtQuick.2/ [13:29] for example [13:29] yeah [13:29] exactly [13:29] so, I don't know why it doesn't take [13:29] that module line identifies that [13:29] ubuntu/HUD/1.0/ [13:29] seems to be a side effect, not a namespace issue :) [13:29] dunno... that was what i assumed :) [13:30] kenvandine: However, I do prefer your dir-soname [13:30] can you change to map it? [13:30] as well as the package name? [13:30] and maybe just a replace (no conflict) for the transition? [13:31] so you mean changing it to Ubuntu/HUD.1.0 ? [13:31] i wonder if HUD-1.0 would work [13:31] probably [13:31] or [13:31] HUD1.0 [13:31] kenvandine: yeah, the last one is the more readable [13:32] i'll give it a swing :) [13:32] great! [13:32] kenvandine: do you mind coordinating with sil2100? I think we need to change the dep in a lot of places :) [13:32] * sil2100 turns into an ear [13:32] kenvandine: sil2100 is doing the transition to the new HUD today, but I guess, you won't be enough of being two if you have time :) [13:33] ;) [13:33] sil2100: the HUD runtime package is going to be renamed [13:33] sil2100: so need to change that in the MP for new hud as well autopilot runs :) [13:33] (list of packages to install and so on…) [13:33] didrocks, kenvandine: oh, is there a MR for the runtime package rename already? [13:34] for the qml package [13:34] i am tweaking it one more time now [13:34] https://code.launchpad.net/~ken-vandine/libhud-qt/versioned/+merge/160363 [13:34] sil2100, ^^ [13:35] kenvandine: thanks! [13:35] didrocks, it doesn't work with a dash in that dir name... grrr [13:36] oh... i didn't want that either [13:36] kenvandine: ahah, lost in the discussion? :) [13:36] kenvandine: no dash in the dir, no -plugin in the package name! :-) [13:36] (but a beer, please! ;)) === m_conley_away is now known as m_conley [13:38] ok, HUD1.0 doesn't work either [13:39] i wish i understood why... [13:39] HUD.1.0? :/ [13:39] trying that now :) [13:40] ok, that works ;) [13:41] actually i do need a R/B/C for libhud-qt -> libhud-qt1 transition [13:41] that does conflict in the ppa [13:42] ok ;) [13:45] i do feel much better knowing we can version these qml packages [13:45] that's been stressing me out! [13:46] hey guys [13:46] hey rickspencer3 [13:46] seb128, so, one of my computers isn't loading x this morning :/ [13:47] I seem to recall that there was something about telling plymouth to sleep or something to work around something racey [13:47] sound familiar at all [13:47] ? [13:47] rickspencer3: just enable -proposed [13:47] and steal the lightdm upload [13:48] what lightdm upload in proposed? [13:48] https://launchpad.net/ubuntu/raring/+queue?queue_state=1 [13:48] that's a queue, and it's plymouth only [13:49] kenvandine: yeah, nice work! :) [13:49] (tjaalton asked for lightdm to be rejected earlier) [13:49] ah [13:49] and advising people to enable proposed isn't really good [13:49] Laney, I suppose he wanted me to test it [13:50] well I guess it depends, i dont know what is needed atm, tjaalton might :-) [13:50] mlankhorst, wasn't there some way that I could add a sleep to plymouth to work around it? [13:50] didrocks, thanks, i pushed my changes [13:50] can you give it a review? [13:50] if you want the hack, just add sleep to lightdm.conf before the exec [13:50] sleep 3 or so [13:50] kenvandine: sure, doing [13:50] sil2100, the new package is qtdeclarative5-hud1.0 [13:51] mlankhorst, can you tell me more specifically where that file is? [13:51] /etc/init/lightdm.conf [13:51] upstart job [13:52] mlankhorst, it already says "sleep 5" in the post-start script [13:52]