=== asac_ is now known as asac [01:35] jasoncwarner_, it's built on amd64 btw [01:36] i've just had to do another upload though - https://launchpad.net/ubuntu/+source/thunderbird/7.0~b1+build1+nobinonly-0ubuntu2 [01:36] i noticed some broken icons ;) [01:36] chrisccoulson: :) [01:36] I'm on i386 on this machine [01:36] ah, that's still building anyway [01:36] that one takes longer because it has to build all the translations too [01:37] i need to figure out how to make the toolbar draggable now [01:38] it's kinda weird having the toolbar colour match the window border, but not being able to drag the window with it [02:19] quality technology. we can make marvelous things and (relatively) bug-free programs, but water distribution stations in cities still end up bringing up E-coli, boil-your-water warnings. :'( [02:20] cyphermox: it's just different kinds of bugs :-P [02:22] yes, isn't technology great? [02:26] broder: please mark life-threatening bugs "Critical" :D [02:26] RAOF: still around? [02:28] finishing up another xapian-core build test with the CJK patch, but it's already running on my system, and doesn't seem to be causing any issues. I'll send a merge request in a minute [02:41] TheMuso, is there an a11y indicator? [02:49] TheMuso: do you have time to take a look at bug 829370 [02:49] Launchpad bug 829370 in at-spi2-core "at-spi-bus-launcher crashed with SIGSEGV in XFlush()" [High,Triaged] https://launchpad.net/bugs/829370 [03:43] cyphermox: Back from lunch now; what's up? [04:11] Good morning [04:12] dobey: I am now [04:42] chrisccoulson: still now i386 thunderbird theme ;) ... taking a long time to build! [06:02] * pitti is off for an hour for running [06:15] chrisccoulson: getting new thunderbird and theme now...I'll let you know later today! [06:25] chrisccoulson: tbird disabled toolbar btns aren't really distinguisable from active ones [06:28] lightdm does not start for me any more. Anyone knows how to get it working again? [06:29] replace it with gdm... ;-) [06:51] tkamppeter ...what is the issue? I'm sure robert_ancell would be willing to take a look [07:05] jasoncwarner_, what happens is that lightdm tries to start repeatedly showing by the screen getting shortly white and returning to the text screen. [07:05] robert_ancell: Hi Robert, how about fixing that MP today? :) [07:05] https://code.launchpad.net/~gunnarhj/lightdm/langfix-oneiric/+merge/70969 [07:06] jasoncwarner_, no errors in /var/log/xorg.0.log nor segfaults in /var/log/syslog. [07:06] robert_ancell, ^^ [07:12] tkamppeter, what does /var/log/lightdm/lightdm.log say? [07:13] GunnarHj, so, what sets these variables incorrectly? [07:16] robert_ancell: It's accountsservice that writes to ~/.profile. But they are not set incorrectly; the purpose with the proposed code is to avoid a side effect with setting LC_MESSAGES etc. explicitly. [07:19] robert_ancell: There is no obvious package for it IMO. What speaks for lightdm is that, after all, it's lightdm that actually populates the env. variables by sourcing ~/.profile. [07:24] GunnarHj, so I'm still confused. If a user sets LC_MESSAGES to LANG this causes a problem? [07:24] And we want to catch if this is done inside a user script and undo it? So it applies to all login systems (lightdm, gdm, getty)? [07:29] robert_ancell: No, the problem that showed up is that if e.g. LC_MESSAGES is explicitly set, and you ssh into another machine without that locale, the environment is messed up. It's a special case, and yes, it applies to all login systems. Hmm... Maybe the latter speaks for moving the code, after all, to a place that is always sourced... [07:30] GunnarHj, yes, so Xsession.d sounds like the right location (although that only works for graphical logins). x11-common contains a lot of these [07:30] RAOF, does x11-common sound like the right place for something like this ^^^ ? [07:31] GunnarHj, so ssh handles if LANG is set but not LC_MESSAGES? [07:31] robert_ancell: It's not entirely clear to me what "this" is? There's some environment which, if you ssh to a machine on which that environment is unsuitable, is unsuitable? [07:33] RAOF, given there was a requirement to unset LC_MESSAGES if it equalled LANG for all logins, would the right place be for this script to be in Xsession.d in the x11-common package? [07:34] There is more questions, which are: "What is breaking in this case, what set the variables like this and should ssh break on this?" [07:34] I mean "What is breaking in this case, what set the variables like this and should we be correcting for this problem?" [07:35] robert_ancell: That sounds like something that may well want to be in x11-common. [07:36] robert_ancell, RAOF: I'm not able to answer all those follow-up questions. Maybe pitti, who was involved in the decision to take this step, is able to add something. [07:36] GunnarHj, ok, so it sounds like the merge should be moved to x11-common, but before we do that we need to know if the fix should be done. [07:36] robert_ancell, http://paste.ubuntu.com/674350/ [07:37] pitti knows all :) [07:37] sorry, I didn't follow -- this is about not setting LC_MESSAGES and friends when it's equal to LANG? [07:37] yes, that would be good to have somewhere [07:37] pitti: Yes. [07:37] arguably l-s shoudln't write it like that into ~/.profile in the first place [07:37] tkamppeter, unity-greeter crashed, do you have an apport bug? [07:38] once it does, we'd need to fix all login managers or x11-common or even bash itself (VT logins!) to un-set them any more, which is unpractical [07:38] GunnarHj: could we change l-s to not write them unless they are not equal to LANG? [07:38] and if you re-write .profile, clean up the redundant LC_* stuff? [07:40] pitti: Well, basically you don't know at the time when you write to ~/.profile what value LANG will have at next login. ;-) [07:40] morning all! [07:40] GunnarHj: why not? it was removed from gdm, and lightdm doesn't even touch it [07:41] hey Sweetshark, guten Morgen! [07:42] pitti: I mean that the user can change LANG after he sets LC_MESSAGES etc. Which shouldn't happen very often, I suppose. ;-) [07:43] * Sweetshark is torturing kakadus still ... [07:43] GunnarHj: but that's precisely the case which we want to fix here [07:43] GunnarHj: i. e. if you run a program with e. g. LANG=fr_FR.UTF-8 ..., I actually expect that program to be in French [07:43] these days I need to set LC_MESSAGES= LC_CTYPE= LANGUAGE= LANG=fr_FR.UTF-8 [07:44] and this also causes ssh logins to act up, as on most servers your ~/.profile just sets $LANG [07:44] pitti, oh, that is the problem [07:44] pitti, GunnarHj, so l-s should only set LC_MESSAGES if the user has explicitly asked for a different one right? [07:44] i.e. different to LANG [07:44] yes [07:45] but currently it sets all four variables? [07:45] in the past it was necessary to set them all, as gdm changed them dynamically [07:45] ah, and now that problem no longer exists [07:47] GunnarHj, so, it sounds like the fix should be in l-s - do you agree? [07:47] pitti, robert_ancell: I suppose that the test against LANG can be done at the time when the user sets a new language, i.e. when ~/.profile is written to. [07:48] GunnarHj: right; it couldn't in the past, as gdm changed it, but now it should be possible, right? [07:48] GunnarHj, yes. There will be some existing users with the problem, they need to go to the language selector to fix that === abhinav_ is now known as abhinav- [07:50] glatzor: good morning! just fyi https://bugs.launchpad.net/software-center-agent/+bug/833053 talks about restricting the license key stuff to /opt/pkgname, so we should be fine [07:50] Ubuntu bug 833053 in software-center-agent "license_key_path should not (yet) support per-user location" [Undecided,New] [07:51] robert_ancell, pitti: I'm beginning to hesitate. If we wouldn't write to ~/.profile (or somewhere else), then if and when the user changes the locale for regional formats, the language is changed as well unsolicitedly. [07:52] GunnarHj: but in that case we actually should write LANGUAGE/LC_MESSAGES/LC_CTYPE into ~/.profile [07:53] pitti: I forgot that one, we still need a MIR for lo-menubar for now as I add it as a recommend to the libreoffice metapackage. [07:53] Sweetshark: ah, jasoncwarner_ already asked about that [07:53] Sweetshark: shouldn't that also be in the default install? [07:53] morning [07:54] pitti: That's what we are doing currently. [07:54] pitti: yes [07:55] GunnarHj: ah, then everything seems to be okay? [07:55] indeed, it just sets $LANG in my ~/.profile now [07:55] pitti: Going forward, I can think of a solution where this is completely handled at one place, probably accountsservice. [07:55] pitti: as of now, lets just use the existenting lo-menubar package, while I am already building it along with the libreoffice source package I dont package it there yet. [07:55] GunnarHj: I thought accountsservice already writes ~/.profile in SetLanguage now? [07:55] Sweetshark: sounds fine [07:56] Sweetshark: and from 12.04 on it will be built into libo? [07:56] mvo, hm. not really. but you still want to allow a random location under /opt/pkgname? [07:56] pitti: Yes it does. But it always do so, irrespective of the value of LANG. [07:57] mvo, I would like to see a trusted source for dropping files, e.g. the package control field in the prototype [07:57] GunnarHj: I thought if you set a different language than regional format, it would write all four? if not, how does that work then? [07:58] pitti: well, I could still do it post-beta (this is only changing packaging, the code is the same). [07:58] Sweetshark: that sounds even better then [07:58] pitti: The problem here, I think, is that ~/.profile is a static file. That's why we currently need to do that test at the login time. [07:58] Sweetshark: so for beta we pull in lo-menubar, and handle the MIR/FFE for it; and the next upload would then just move it into libo itself [07:59] mvo, we don't only have to think about people dropping scripts/exectuable but also configuration files [07:59] Sweetshark: I guess we'll need another pre-beta upload for the armel FTBFS anyway, so this upload could also add teh Recommends: lo-menubar? [07:59] pitti: Yes, of course it works as you just said. But it writes all four always. [07:59] GunnarHj: I just started l-s and wiggled the language, and just got $LANG; weren't you saying above that this is intended? [08:00] mvo, think of /opt/pkgnmae/conf.d/ - especially if the the /opt location/app thing is opened more widely in the future [08:00] pitti: thats how it looks in my current checkout ;) [08:00] pitti: Yes, that scenario does not sound right. Your ~/.profile should be changed each time you change language. [08:01] Sweetshark: nice; please file the MIR/FFE bugs then [08:04] glatzor: right, that is fair enough, I will add info on that on the bugreport, I think the packagerecord is the best source indeed [08:04] pitti: Please note that accountsservice was recently updated with all those scripts, so most people using Oneiric does not have LANGUAGE/LC_MESSAGES/LC_CTYPE/LC_COLLATE set in their ~/.profile files. [08:05] GunnarHj: I don't understand that -- it just moved from l-s to accountsservice, that shoudln't have any effect on ~/.profile? [08:05] mvo, i am currently writting a short comment :) [08:06] glatzor: \o/ [08:07] pitti: No, but developers don't change their language settings that often. ;-) [08:07] pitti: But you are right, accountsservice does what l-s (and gdm) did previously. [08:10] nice, the retracer caught up with everything [08:11] jasoncwarner_: FYI, I just got the alternates into their size limits [08:11] they will shrink by another 5 MB or so over the weekend (with the langpack refresh) [08:11] pitti: This is what I think right now: We put that code that unsets LC_MESSAGES etc. in a directory that is typically sourced by DMs. That should cover most cases - after all, the environment problem isn't exactly a disaster... Then, during the next cycle, we add enough fields to accountsservice to store all the settings there, and drop ~/.profile. Accountsservice will be able to do 'the LANG test' dynamically and serve an [08:11] y DM with the right info for setting the environment. [08:12] GunnarHj: I thuoght accountsservice's SetLanguage() call already does that test? [08:12] it should certainly do [08:13] ah, I just tried that again [08:13] so, I had nothing in ~/.profile, then changed region to German/Switzerland and then back to German/Germany [08:13] then I only had $LANG in my ~/.profile, which is right [08:13] pitti: SetLanguage does what should be done with respect to storing the info. Now I'm talking about getting the info at login in order to actually populate the variables. [08:14] then I switched the langauge to En/UK, and back to German, and now I have all four [08:14] the last case is the bug [08:14] GunnarHj: storing the info and populating the variables for the session is one and the same no -- ~/.profile [08:16] pitti: We should be able to do it without ~/.profile, but that's not how we currently do it. [08:17] pitti: Accountsservice does not have fields for storing both LANGUAGE and the locale name that goes to LC_MESSAGES etc. [08:18] GunnarHj: I thought accountsservice would store it into ~/.profile now? [08:18] pitti: But what I mean is that during next circle we should change that. [08:18] is today a good day for upgrading to oneiric? [08:19] hey [08:19] hello! [08:19] pitti: Yes, now a-s writes to ~/.profile. [08:19] hey pitti, asac [08:19] hello seb128 ! [08:19] bonjour seb128 [08:19] hey asac [08:19] yay pitti ! [08:19] GunnarHj: right, and I think at this time we should check whether or not to write LC_* and LANGUAGE [08:20] seb128: FYI, retracer has caught up with all bugs now [08:20] asac, it's not a bad day but it's also the beta freeze and ui freeze day so it's likely to be a bit busy and bumpy [08:20] pitti, \o/ [08:20] seb128: including lucid/maverick/etc. [08:20] pitti, that was short [08:20] so my thinkpad broke today ... no fan anymore [08:20] and this netbook isnt good in natty ... hoped its better in oneiric [08:20] guess i will just upgrade [08:21] pitti: I'll get back about this later. IRC has its limitations. ;-) [08:22] asac, there is no known breakage so you can as well give it a try ;-) [08:23] no known breakage in oneiric? *cough* [08:32] pitti: hmm, there is already a mir: https://bugs.launchpad.net/ubuntu/+source/lo-menubar/+bug/725250 [08:32] Ubuntu bug 725250 in lo-menubar "[MIR] lo-menubar" [Undecided,Incomplete] [08:32] Sweetshark: ah, seems it's blocked by bug 739184 then? [08:32] Launchpad bug 739184 in lo-menubar "Keyboard shortcuts to menus do not work." [Undecided,Confirmed] https://launchpad.net/bugs/739184 [08:35] Sweetshark: that might very well be fixed with the current version, of course [08:35] robert_ancell, apport did not get triggered [08:35] so, at what time is the UI freeze? [08:35] robert_ancell, also no segfault in syslog [08:35] 2100 UTC, AFAIR [08:36] pitti, define breakage? ;-) I didn't say "no bug" ;-) [08:36] seb128: ah, ok [08:36] seb128, one quick question, if the unity-greeter branch is debian/ only why does it have all the other files in it? [08:36] pitti, it's just that today is not especially bad to upgrade [08:37] pitti, ok cool, plenty of time then :) [08:37] tkamppeter, ok, you'll need to fix apport so it can generate one [08:37] robert_ancell, hey, hum, like? [08:37] robert_ancell, http://bazaar.launchpad.net/~ubuntu-desktop/unity-greeter/ubuntu/files [08:38] robert_ancell, I think I don't understand the question [08:38] robert_ancell, pitti, apport is reporting a lot of crashes for me (and yesterday I passed on 2 crashes, one on Libreoffice. [08:39] pitti, oh, since lightdm catches the segfault does apport miss it? any way around that? [08:39] robert_ancell: why does it catch segfaults? [08:39] Are Ubuntu fonts included in Oneiric or I still have to use the private ppa? [08:39] seb128, oh, debcheckout has changed behaviour [08:39] robert_ancell: if you want to do cleanup when it happens, then you can catch it, clean up, and re-raise it [08:40] BigWhale: it has been in the default install for several releases already [08:40] rodrigo_, hey... the example for gir ebook that I had now works. But I was extra busy so I just gave it a quick test. :) [08:40] pitti, really? then I have to throw out the ppa :> [08:40] robert_ancell, I am not sure whether lightdm actually segfaulted, as syslog does not report a segfault. [08:40] BigWhale, cool! [08:41] tkamppeter, Launching process 5248: LIGHTDM_FROM_SERVER_FD=13 DISPLAY=:0 HOME=/var/lib/lightdm GNOME_KEYRING_CONTROL=/tmp/keyring-WzmCzJ XDG_SESSION_COOKIE=a1517f691661ea5ee1aaf23b4b019b21-1314253506.490179-2123213501 GDMSESSION=unity-greeter LANG=en_US.UTF-8 LIGHTDM_TO_SERVER_FD=16 USERNAME=lightdm GNOME_KEYRING_PID=5246 DESKTOP_SESSION=unity-greeter XAUTHORITY=/var/lib/lightdm/.Xauthority USER=lightdm PATH=/usr/ [08:41] lib/lightdm/lightdm:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games SHELL=/bin/false /usr/sbin/unity-greeter [08:41] Process 5248 terminated with signal 11 [08:41] pitti, only the parent process catches it, but that shouldn't bother apport right? [08:42] tkamppeter, have you set to "ignore this problem" with unity-greeter? [08:42] pitti: nope, nobody has touched the bzr lo-menubar branch since I merged it in libreofffice and commited the adjustments back: https://code.launchpad.net/~lo-menubar-team/lo-menubar/trunk [08:42] robert_ancell: as long as any process actually raises a SIGSEGV, apport will see it [08:42] tkamppeter, so it looks like something is wrong on your system [08:42] checking /var/log/apport migh thelp [08:43] or /var/crash if it doesn't already have one [08:43] mvo: I'm a bit puzzled by bug 828759 [08:43] Launchpad bug 828759 in ubuntu-meta "package ubuntu-desktop 1.240 failed to install/upgrade: ErrorMessage: dependency problems - leaving unconfigured" [Critical,Triaged] https://launchpad.net/bugs/828759 [08:43] mvo: I have booted a natty desktop, and try a dist-upgrade to oneiric now [08:43] mvo: when I do apt-get dist-upgrade, it wants to remove ubuntu-desktop, and keep the old at-spi [08:44] mvo: but when I do "sudo apt-get install ubuntu-desktop", it correctly upgrades it, removes the old at-spi, and pulls in the new at-spi2-core [08:44] TheMuso: ^ FYI [08:44] mvo: is there a way how I can debug why it wants to remove ubuntu-desktop instead of dropping the old packages? [08:45] pitti: hmm, if we do not MIR lo-menubar, and I package it from the libreoffice-source package, I kinda sneak lo-menubar in main, dont I? [08:45] Sweetshark: yes, but that's not such a big deal if it's upstream now [08:45] that happens all the time with new usptream stuff [08:45] pitti: -o Debug::pkgProblemResolver=true -o Debug::pkgDepCache::AutoInstall=1 [08:46] Sweetshark: but it would still require an FFE, and we wouldn't accept it until this bug gets fixed [08:46] pitti: the fact that ubuntu-desktop is installed when you ask for it explictely is expected, apt assigns it a very high value in this case, the interessting question is why at-spi is kept, probably a bunch of reverse depends or something like that [08:47] hiya [08:48] mvo: http://paste.ubuntu.com/674379/ [08:48] pitti: well, ok. I will find absolutely no time for that and at upstream nobody would care. So DX would have to work on it upstream ... [08:48] Investigating (0) at-spi2-core [ amd64 ] < none -> 2.1.5-0ubuntu1 > ( misc ) [08:48] Broken at-spi2-core:amd64 Depends on libatspi2.0-0 [ amd64 ] < none -> 2.1.5-0ubuntu1 > ( misc ) (>= 2.1.1) [08:48] Considering libatspi2.0-0:amd64 2 as a solution to at-spi2-core:amd64 0 [08:48] Holding Back at-spi2-core:amd64 rather than change libatspi2.0-0:amd64 [08:48] mvo: ^ that could be it? [08:48] the default photo viewer is set to eog; shouldn't it be eog? [08:49] Sweetshark: ok, let's keep it for next cycle then [08:49] hey Kaleo [08:49] Kaleo: eog == eog ? [08:49] pitti: eye of gnome [08:49] :) [08:49] Kaleo: yes, I know that [08:49] Kaleo | the default photo viewer is set to eog; shouldn't it be eog? [08:49] sorry, I re-read myself [08:49] eog == eog [08:49] shotwell [08:49] yes, it should :) [08:49] no [08:50] :) [08:50] shotwell always imports stuff [08:50] and you don't want to do that for every picture you look at [08:50] many pngs/jpgs are not even photos [08:51] TheMuso: do at-spi and at-spi2-core really conflict package/functionality-wise? [08:51] TheMuso: it seems that conflict is causing bug 828759 [08:51] Launchpad bug 828759 in ubuntu-meta "package ubuntu-desktop 1.240 failed to install/upgrade: ErrorMessage: dependency problems - leaving unconfigured" [Critical,Triaged] https://launchpad.net/bugs/828759 [08:52] Kaleo, sorry just saw your ping on ayatana [08:52] Kaleo, no it shouldn't [08:52] Kaleo, well the reason is that shotwell only handle a few photo format, not gif, bmp, etc [08:53] Kaleo, they defined themself as a photo viewer not an image viewer [08:53] cyphermox, RAOF: did anyone worked on xapian during the european night? [08:54] pitti: Ok, so for 3.4.2-2ubuntu1 I will: a) have armel fixed b) will _not_ recommend lo-menubar from libreoffice c) but prepare everything for lo-menubar being packed from LO with be just changing one switch ... [08:54] good morning everyone [08:55] chrisccoulson: Morning, you nightworker! [08:55] hey chrisccoulson, how are you? [08:55] heh [08:55] i'm good thanks, a bit tired though [08:55] late night ;) [08:55] tb kept you up late? ;-) [08:55] yeah [08:56] getting that theme in ;) [08:56] TheMuso: I updated bug 828759, would appreciate feedback from you about that [08:56] Launchpad bug 828759 in ubuntu-meta "package ubuntu-desktop 1.240 failed to install/upgrade: ErrorMessage: dependency problems - leaving unconfigured" [Critical,Triaged] https://launchpad.net/bugs/828759 [08:56] Sweetshark: sounds good [08:57] ok, time to start on unity updates [08:57] is somebody having a sponsoring slot there? [08:57] https://code.launchpad.net/~mathieu-tl/ubuntu/oneiric/xapian-core/lp833172/+merge/72821 [08:57] seb128, pitti: I see; so do you know any way for the user to define their photo management app [08:57] ? [08:57] that's the xapian cjk update dx needs for cjk support in unity [08:58] Anyone know who runs ubot2? [08:58] Kaleo, no [08:58] seb128: ok, thank you [08:59] mpt, I think dholbach knows [08:59] when the bot went to miss he told me who to ping [08:59] but I don't remember who that was now [08:59] thanks seb128 [08:59] yw [09:02] pitti: its line 404 (how fitting!) - Considering libatspi1.0-0:amd64 10 as a solution to libatspi2.0-0:amd64 2 - so libatspi1.0-0 has a score of 10 and libatspi2.0-0 of just 2 so apt things that the former is more important [09:03] pitti: probably due to reverse depends or something like this, I can check that in detail, but that raises the question, do they need to conflicts at all? aren't both libraries? [09:04] mvo, you know a bit about xapian? [09:04] seb128: depends on what you want to know [09:04] mvo, could you test build and run https://code.launchpad.net/~mathieu-tl/ubuntu/oneiric/xapian-core/lp833172/+merge/72821 just to double check it doesn't break s-c? [09:05] mvo, and upload if it works? [09:05] mvo, it's what dx needs for cjk, the patch is a backport from upstream stable svn and has been tested by dx and cyphermox [09:05] mvo, so it "should" be ok, I would still appreciate if you could test run it, I don't know much how to test xapian [09:11] pitti, robert_ancell has already left, but I can say that I did not set "Ignore this problem" on anything. [09:11] tkamppeter: do you have a /var/crash/ file which sounds like lightdm? [09:13] pitti, I have [09:13] -rw------- 1 lightdm lightdm 2023993 2011-08-24 23:42 /var/crash/_usr_sbin_unity-greeter.103.crash [09:13] there you go [09:13] pitti, how do I trigger the apport process manually? [09:13] sudo apport-bug /var/crash/_usr_sbin_unity-greeter.103.crash [09:13] to report it [09:14] tkamppeter: it'll only get popped up as "system problem blabla", probably you clicked it away [09:14] unfortunately update-notifier is really annoying with these [09:14] pitti, that is possible, these dialogs should tell which program crashed. [09:15] pitti, I got very many of them. [09:15] Excuse me. Have anyone know about why packages language-support-... are disappeared at Oneiric? [09:15] jincreator: yes, they are obsolete; language-selector installs the actual support packages dynamically now [09:17] Then language-selector works at installing ubuntu, too? [09:17] pitti, bug 833619 [09:17] Launchpad bug 833619 in unity-greeter "unity-greeter crashed with SIGSEGV in FT_Set_Transform()" [High,New] https://launchpad.net/bugs/833619 [09:18] tkamppeter: retrace should come in a few minutes [09:24] and there it is [09:25] seb128: sure, hold on a wee bit, is that the one that was in the desktop PPA? [09:34] seb128: do you know if indicator-applet is supposed to work on in current oneiric? [09:34] * pitti currently looking at bug 829491 [09:34] Launchpad bug 829491 in indicator-applet "indicator-applet version 0.4.12-0ubuntu1 failed to build in oneiric" [High,Confirmed] https://launchpad.net/bugs/829491 [09:35] dbarth: is DX still interested in maintaining indicator-applet and porting it to GTK3, or should we just drop it? [09:35] mvo, it's a new revision of it [09:35] dbarth: (now that we have unity 2d in the system, and actually want to provide a rather clean GNOME experience when people run gnome-panel or shell) [09:35] pitti, ted is working on it, he got some gtk3 style issue I think [09:35] pitti, it just came after default installation work [09:35] pitti, he mentioned it yesterday still so I think it's going to be fixed but after the beta freeze [09:35] seb128: ah, I wonder if it's actually good time spent, or whether it should just be dropped [09:35] pitti, "default install first" [09:36] pitti, it's "do nothing, it will get an update, but after the main install issues are settled" [09:36] seb128: ok [09:36] seb128: ok, thanks; still wondering if that's well spent resources, though [09:37] pitti, it's a few hours work and there are still people out there who use gnome-panel and a GNOME2 style desktop and want indicators [09:37] pitti, well, "dx call" I would say, if they don't fix it we drop it [09:37] *nod* [09:37] I just don't think it's worth pursing the idea [09:37] GNOME has pretty much said that they don't want them, and have shell now [09:38] pitti, we got quite some Ubuntu users who still want to run gnome-panel and that said they miss their indicators [09:38] if it's easy to please them I think we should just fix it [09:39] not a priority for sure though, just a "would be nice" [09:39] pitti: how high would you rate the interest of porting it? [09:39] pitti: ie, worth a freeze exception? [09:39] dbarth: doesn't need an exception [09:40] dbarth: it's FTBFS and broken righ tnow [09:40] dbarth: it's "port or drop" [09:40] right [09:40] same for indicator-network [09:40] i'll see with ted how long that would take; it seems reasonable [09:41] dbarth: merci Monsieur! [09:41] dbarth: right, in C the gtk 2 -> 3 port is often rather trivial [09:41] pitti, cyphermox said he was working on that indicator ftbfs yesterday [09:42] nice [09:42] * pitti assigns to him then [09:45] can someone sponsor libindicate for me before the freeze? 832222 [09:45] bug 832222 [09:45] Launchpad bug 832222 in libindicate "Use cli-common-dev packaging helpers" [Undecided,Confirmed] https://launchpad.net/bugs/832222 [09:46] Laney: doing [09:46] ty [09:47] pitti: i've added an upstream task for ted, if you want to assign to cyphermox [09:48] seems it got lost in a recent refactoring [09:52] chrisccoulson_: \o/ (the addressbook bug :) [10:04] Guys from lightdm I can't select unity 2d after yesterdays updates [10:04] how so? [10:05] be happy you can use the greeter at all :P [10:05] ogra_, bug number? [10:05] seb128: only shows ubuntu from the drop down I think let me log out and check the exact wording [10:05] davmor2, is ubuntu-2d installed? [10:06] unity-2d [10:06] not ubuntu-2d [10:06] seb128, none :P ... i still wait for the next upstream ... i know it still has issues, waiting 5-8sec for a dot to appear on the passwd input might still be normal at this point of development :) [10:06] seb128: It only lists Ubuntu, I'll log back in now and check on unity-2d [10:06] ogra_, is that unity greeter on armel? [10:06] it works fine here, it just displays veeeery slooow [10:07] yep, unity-greeter on armel, up to date yesterday [10:07] well I don't think it's that of an issue on any decent video driver [10:07] where decent is intel, nvidia or ati [10:07] (my login is as fast as in gdm, the grereter just cant cope with updating the screen) [10:08] that's why it didn't get worked [10:08] we dont have any video drivers on armel [10:08] right, which is why only you see the issue :p [10:08] seb128: rc unity-2d 4.0.0-0ubuntu1 I'm assuming that means it got removed [10:08] and we also have tons of users with vesa [10:08] you should maybe default to the gtk greeter on armel [10:08] davmor2, it does [10:08] ogra_, use the gtk greeter? [10:08] you cant expect that people have GL supported HW for the greeter [10:08] no and we don't [10:08] you say yourself it runs [10:09] it's using cairo, but that's slow on vesa [10:09] seb128, i want to use ubuntu on my moms PC without having to tweak it, that was possible for the last 6 years, even if she doesnt have super powerful HW [10:09] I'm not sure how much you can do nice graphics using cairo in a vesa driver [10:09] i expect that a replacement doesnt have such massive regressions really [10:09] ogra_, it's still possible, use the gtk greeter? [10:10] thats preinstalled and autoselected on HW that cant cope ? [10:10] no [10:10] it's hard to define "visually slow" [10:10] sorry, but making the default login process depend on high end HW seems like a massive regression and a release blocker to me [10:11] pitti: thanks [10:11] ogra_, come on, it's not an high end hw [10:11] seb128, typing a char in the password field and waiting for 3-5 sec for it to appre is what i define as slow [10:11] ogra_, the unity greeter runs smoothly on my old atom 10v [10:11] which I would call an high end config [10:11] it's a nb with an atom cpu and an intel 945 video [10:11] or klicking on "other" to sdelect another uiser and watchiong a slideshow start (about 10 sec for the scrolling) is what i define as slow [10:12] seb128, it has GL capable HW [10:12] i would say 1/3 of our users doesnt [10:12] it has nothing to do with gl [10:12] it has to do with acceleration one way or the other [10:12] we can't lower down our standard to "should run fast on vesa driver" [10:12] imho we cant release like that [10:13] arm uses fbdev, not vesa (since there is no video bios) [10:13] and i wouldnt expect us to given there is still time [10:13] let's see [10:13] seb128, but talking to you it sounds like that issue is completely ignored [10:13] it's not an issue on intel or amd config [10:13] I think armel should default to the gtk greeter if armel hardware can't display cairo graphics [10:13] sure, it wont be an issue in nvidia either [10:14] we cant ship accelerated drivers [10:14] sorry x86 or amd64 [10:14] that doesnt mean that users dont use accelerated drivers [10:14] when I say intel or amd [10:14] yeah, got that [10:15] it's just hard to do shiny graphics on a smooth way on almost not working drivers [10:15] that we default to something unusable in a place where its simply not appropriate is a big failure imho [10:15] its not hard to do shiny graphics [10:15] well we can use the gtk greeter by default [10:15] nor is it hard to do thzem in a smooth way [10:15] you will have a 10years old looking and fast greeter [10:16] it's just undermine all the modern config users for some who have buggy drivers [10:16] if you pick supporting libs that require certain HW then, thats indeed the outcome [10:16] doesn't seem a smart move [10:16] sorry, but i think its the other way round [10:16] seb128: is it a feature that it got removed or a flaw in the upgrade procedure? I'm assuming the latter [10:16] davmor2, it's probably you using dist-upgrade without reading what it does ;-) [10:16] davmor2, you should use upgrade [10:17] i.e don't do partial upgrades if you don't understand what it wants to do [10:17] it usually leads to such issues [10:17] seb128: Good chance :) [10:19] ogra_, well, what about you write to the list to suggest we default to the gtk greeter then? [10:19] ogra_, that's not a discussion we will solve on IRC [10:20] but I don't think we should level down on the slowest config you can find and say that things should run fast on that, you can't have high quality graphics on crappy hardware [10:21] i do [10:21] at least i did the last releases [10:22] why not do the same as with unity, and fall back to an "easier-on-the-hw" config when the hw is not up to it [10:22] tjaalton, thats what i'm saying [10:22] tjaalton, because it's hard to define "slow to render" [10:22] tjaalton, the greeter is pure gtk and cairo and run on any config, it's not using gl or anything [10:22] well, our session script can apprently determine it [10:22] tjaalton, it seems just that cairo is slow on fb [10:23] ogra_, no, our session script checks for gl [10:23] ogra_, because unity uses gl [10:23] sure [10:23] ogra_, the unity greeter has been made on purpose using nothing that requires 3d or gl [10:23] seb128: hard, but not necessarily too hard :) [10:23] but if you see there is no GL at all thats a pretty good indicator [10:23] ogra_, it's pure gtk and cairo [10:23] ogra_, well if we do that we could as well do a gl greeter and a fallback to the gtk one then [10:24] I'll see what my sis laptop likes of the new greeter [10:24] it makes no sense to make efforts to not use gl to default to not use the greeter on non gl configs anyway [10:27] well, you can look for vesa or xfbdev too [10:27] Laney, ping [10:27] hi [10:28] rodrigo_: hi there :) [10:28] Laney, seems you didn't push your last changes in tomboy package to the ubuntu-desktop branch [10:28] seb128, other apps use cairo too here and arent slow, so i'm willing to blame the development level of the code (which is why i didnt file a bug yet) [10:28] hola Chipaca, long time! :-D [10:28] yeah i forgot [10:28] did the merge in git, much easier [10:28] Laney, also, Chipaca wanted to chat with you about the libsyncdaemon dep [10:29] Laney, in gnome git? [10:29] Laney: hi there. Question about that, it now depends on libsyncdaemon, for what seems to be merely getting credentials? [10:29] ogra_, well there will be some optimization, but cairo is slow, it's one of the reason gtk wasn't picked by then for 2d launchers [10:29] no, debian git [10:29] Chipaca: I don't know anything about that, rodrigo_ did that patch [10:29] Laney, ah [10:29] Laney: ah. [10:29] * Chipaca *looks* at rodrigo_ [10:29] * ogra_ wants gnomecanvas back :P [10:29] Chipaca, yeah, now that I remember, it's always depended on libsyncdaemon, to talk to u-sso-client [10:29] ogra_, other apps use cairo but don't render a screen with effects like the greeter tries to do [10:30] err.. I enabled multiarch support and now apt wants to download a bunch i386 packages when I try to install nonfree flash plugin [10:30] rodrigo_: not in natty it didn't [10:30] Chipaca, hmm [10:30] * rodrigo_ looks at the bzr log [10:30] seb128, yeah, we probably could make the effects switchable :) [10:30] rodrigo_: in natty i can remove libsyncdaemon without it uninstalling tomboy :) [10:30] like in metacity [10:30] with libc and everything [10:31] jbicha: you need sponsoring for gnome-shell? [10:32] bigon: yes but I was going to try to remove some extra dependencies [10:32] rodrigo_: ah, but in natty, *if* you removed libsyncdaemon, and changed sync in tomboy to u1, tomboy crashed [10:32] there was a bug in tomboy in natty [10:32] it didn't have the right dependencies [10:33] looks like it, yes [10:33] jbicha: tell me when you need me to upload it [10:34] Chipaca: it is fixed in -proposed [10:34] Chipaca, Laney: [10:34] revno: 78 [10:34] tags: 1.6.1-0ubuntu5 [10:34] author: Iain Lane [10:34] committer: Bazaar Package Importer [10:34] branch nick: oneiric [10:34] timestamp: Sun 2011-06-26 14:56:53 +0100 [10:34] message: [10:34] * Rebuild against CLR 4.0 [10:34] * 01_dllmaps.patch: Fix glib DLLMap to resolve correctly and not bail with [10:34] stricter dh_clideps [10:34] * Add Build-Dep on libsyncdaemon-1.0-dev so that the DLLMap can be resolved [10:34] https://bugs.launchpad.net/ubuntu/+source/tomboy/+bug/667736 [10:34] Ubuntu bug 667736 in tomboy "tomboy help content not available" [Medium,Fix committed] [10:34] please verify it there [10:34] then it can be sent to -updates [10:35] Laney: rodrigo_: ok, thanks. But I need to break that dependency, now :( [10:35] i probably uploaded a broken changesfile without Launcpad-Bugs-Fixed, because I built it on Debian [10:35] tomboy will be on the CD in oneiric, yes? [10:35] fyi [10:35] yes [10:35] right, so I need to fix tomboy to not get creds that way [10:36] *sigh* I didn't want to learn C# :) [10:36] Chipaca, iirc, the patch loads the lib dynamically, so not sure why the build dep is needed, I guess some C# DDLmap problem [10:36] * rodrigo_ looks at the patch [10:37] it's not a problem [10:37] if you use the library then you have to depend on it [10:37] rodrigo_: it's a dependency becaues if you choose u1 from the sync menu and libsyncdaemon isn't installed, tomboy crashes [10:37] Laney, but it's using DllImport's [10:37] ah ok [10:37] but, it's not needed is my point [10:37] it's only being used to talk to sso [10:37] Chipaca, you'll still use u-sso-client? [10:38] which anybody can do over dbus [10:38] right, so you have to find out where the library is used and then port that call to not need it [10:38] ricotz: https://launchpadlibrarian.net/78035997/buildlog_ubuntu-oneiric-armel.gnome-shell_3.1.3-0ubuntu4_FAILEDTOBUILD.txt.gz :/ [10:38] Laney: exactly [10:38] right, so maybe you can just replace the class SyncdaemonAuthentication in the patch to use DBus [10:38] rodrigo_: right [10:38] rodrigo_: so i need to learn c#, and how to do async dbus calls in it :) [10:38] you should find dbus-sharp nice to use [10:38] :-) [10:38] * Chipaca reads up [10:39] yeah, dbus-sharp should be quite straightforward [10:39] Chipaca, if you need help, let me know [10:39] rodrigo_: have no fear, I will :) [10:39] there's that thing I pinged rodrigo_ about yesterday too, that ought to be fixed [10:39] :) [10:39] both need help, and let you know [10:39] Laney: what's that thing? [10:39] Laney, oh, what was the ping about? [10:39] it tells you to press Save after you get credentials [10:40] but a) save is always insensitive and b) that doesn't appear to be necessary anyway (it got the credentials) [10:40] get them from U1, that is [10:40] that's a leftover from when we used a different way of getting credentials, i guess? [10:40] i suppose [10:40] yes, probably, haven't really checked with the latest tomboy uploads [10:42] Laney: is dbus-sharp and libdbus1.0-cil the same thing? [10:42] dbus-sharp is the source package and name of the project [10:42] ok [10:42] the second one is the name of the library package [10:42] yep, just saw it :) [10:43] so it's being used already -- don't need to worry about the init sequence [10:48] bigon, hmm, i see [10:50] pitti - how does apport decide which fields in a bug report appear inline or are attached as a file? (or is that launchpad?). I'm just wondering why "Plugins" appears inline in bug 833591, whereas "Themes" is attached as a text file [10:50] Launchpad bug 833591 in thunderbird "Addresbook entries cannot be edited" [High,Triaged] https://launchpad.net/bugs/833591 [10:51] chrisccoulson_: if it's only text and <= 5 lines, it's inline [10:51] otherwise attachment [10:51] chrisccoulson_: no, it's apport who decides it [10:52] pitti - ah, ok. how come this ended up as an attachment then? https://launchpadlibrarian.net/78028753/Themes.txt [10:52] I'm not sure [10:52] chrisccoulson_: oh, there's another possibility -- if a hook only specifies a file reference, it's always an attachment, I think [10:53] * Sweetshark watches the arm builder and wonder what the airspeed of an unladen kakadu is ... [10:54] chrisccoulson_: or there's a hidden character in it which is < 32 and not LF [10:54] I actually prefer them attached. The Plugins field is quite hard to read in that bug report, as it gets line wrapping in all the wrong places [10:54] hm, hd doesn't see any weird chars [10:55] chrisccoulson_: so, I don't know off-hand, need to debug [10:55] chrisccoulson_: does that always happen, or only on some bugs? [10:55] pitti - thanks. i'll take a look at that in a bit [10:55] not sure, i've only just started getting bug reports with this new hook, as i only uploaded it yesterday [10:55] brb, lunchtime [10:56] Sweetshark, doesnt that depend on if it is an asian or tasmanian kakadu ? [11:13] bigon, this will need a patch to conditionally use GLES2 api on armel [11:18] hmmm, bug 830110 is annoying [11:18] Launchpad bug 830110 in compiz "Horrifically bright "Aero Snap" color (grid plugin?)" [Undecided,Confirmed] https://launchpad.net/bugs/830110 [11:18] i feel my retina getting more and more burned everytime i see it ;) [11:23] ogra_: yes ... no ... I dont know .... aaaargh [11:24] *giggle* === abhinav_ is now known as abhinav- [11:51] jbicha, do you have your gnome-shell package ready? [11:51] jbicha, hi [11:53] ricotz, pad.ubuntu-uk.org/oneiric-desktop [11:53] ups [11:53] ricotz, https://code.launchpad.net/~jbicha/ubuntu/oneiric/gnome-shell/gnome-shell-3.1.4/+merge/72796 [11:53] it's in the sponsor queue [11:53] seb128, hello [11:53] i know, this isnt usabel [11:54] i talked to him last night, and i was hoping he could use a bug report with a debian-only debdiff [11:54] this merge contains conflicts too [11:54] ook [11:55] would be nice to comment on the merge request when you review it and have issue [11:55] would it only to avoid to others to go through the same review than you [11:55] seb128, that is why i talked to him ;) and was hoping he already did it [11:56] seb128, you have seen the gdm merge? [11:56] yes [11:56] see my comment on the etherpad [11:57] I'm not fan of adding new binaries over debian especially when nothing use those [11:57] yeah saw that ;) [11:57] right, only gdm itself linking against it [11:58] ricotz: yes, http://paste.ubuntu.com/674476/ [11:58] ricotz: sorry for the delay [11:58] jbicha, hi [11:58] dont add gir1.2-soup-2.4 to build-deps, move it to the gnome-shell package [11:59] we don't hjave ambiance and radiance GTK3 themes, right? [11:59] rodrigo_, what do you think oneiric is using? [11:59] jbicha, could you add a build-dep on "mesa-common-dev [armel]" [12:00] seb128, they don't show up on gnome-tweak-tool, and the package only has gtk2.0 dirs [12:01] rodrigo_, light-themes in oneiric has gtk3 version [12:01] hmm [12:01] maybe I'm not up to date, checking [12:01] it has those for months [12:01] since before the distro rally [12:02] lunch, bbl [12:02] jbicha, did you check if the gir1.2-* deps of the gnome-shell package your dropped are autogenerated again? [12:02] seb128, can you dpkg -L light-themes and pastebin it? [12:02] ricotz: yes, except that they lose their version numbers, is that important? [12:03] ah, it's kept back, so installing light-themes now [12:03] jbicha, they are not versioned [12:05] jbicha, dont mix up build-deps and deps [12:05] good morning! [12:05] ricotz: for instance gir1.2-freedesktop (>= 0.10.6) loses the minimum version when autogenerated [12:06] seb128: xapian-core: did the patch, built, tested, and I can search with chinese or korean characters (haven´t tried japanese, but hey) and I get search results [12:06] jbicha, this version is force by the libgirepository1.0-dev, so you need to bump that anyway on 1.29.15 [12:06] the issue is, yesterday I was still able to search with such glyphs even after downgrading xapian, so I don´t know for sure that the patch works [12:07] so if there´s something clearer that should be tested, I can, my PC is all localized in chinese right now :) [12:12] ricotz: ok, I repushed http://paste.ubuntu.com/674480/ [12:14] * rodrigo_ lunch [12:15] jbicha, dont remove libglib2.0-dev ;) [12:16] ricotz: libgtk-3-dev depends on it [12:16] and it built like that in my pbuilder [12:16] jbicha, yes, so does gnome-shell [12:17] and the min-versions can be different [12:19] jbicha, why not dropping build-dep on gir1.2-polkit [12:25] ricotz: ok, done [12:28] jbicha, thanks [12:29] jbicha, try to compare the needed versions with the ones in configure.ac, i think libgnome-bluetooth-dev 3.1.0 and libnm-glib 0.8.999 [12:31] jbicha, also keep libedataserver1.2-dev (>= 3.0.0) which is explictly listed there [12:32] ricotz: jbicha: moo? [12:32] cyphermox, moo? [12:32] um, moo [12:32] yeah, what needs libgnome-bluetooth-dev and libnm-glib? [12:32] gnome-shell [12:32] ah, fun === MacSlow is now known as MacSlow|lunch [12:33] jbicha, ok, get it sponsored :P [12:33] jbicha, g2g, good luck [12:33] ricotz: I forgot to look closely at configure.ac, but libedataserverui-dev depends on libedataserver1.2-dev [12:34] ricotz: thanks [12:36] re [12:36] rodrigo_, sorry I was at lunch, why was it kept back? did the update fix it? [12:37] cyphermox, hey, thanks for the work! I got mvo to confirm that xapian was still working for him as well and to upload, it's fine if it doesn't work as it should (well it's an issue but not ours), I just didn't want to break what standard locales use [12:38] cyphermox, just got a bit confused in the morning because there was no update on the bug nor email in my inbox, it took me a bit to notice the sponsoring request [12:38] of course [12:38] sorry ;) [12:38] hey cyphermox [12:38] hey pitti [12:38] no worry [12:38] cyphermox: do yo plan to update nm-pptp to 0.9, too? [12:38] yes [12:38] trying to get that done now [12:39] (indicator-network can wait a bit) [12:39] cyphermox: while you are at it, could you please do bug 829478 as well? [12:39] Launchpad bug 829478 in network-manager-pptp "network-manager-pptp version 0.8.999-1ubuntu1 failed to build in oneiric" [High,Confirmed] https://launchpad.net/bugs/829478 [12:39] cyphermox: it's kind of implied to get it buildable, of course, but the chagnelog can then close the bug [12:39] yes, will be fixed with the update to 0.9 [12:40] cyphermox: cool, thanks [12:40] I fixed a few other desktopish FTBFS, trying to get through the 9 desktop ones of the recent test rebuild [12:41] yeah, I wanted to do more last night but xapian took a while [12:41] fwiw, I'll answer soon on the thread but I like the maintenance team idea [12:42] thanks for landing the CJK xapian fix [12:42] hooray for CJK support at last! [12:42] cyphermox: ah, looking forward to a few more opinions [12:43] I'm particularly interested in "one day a week" vs. "stable squad" [12:43] chrisccoulson_, seems like they landed a fix for your compiz grid color issue [12:43] pitti: dunno, one day a week sounds sane [12:44] I much favor regular shared slots [12:44] pitti: why did you vote 'resubmit' on your own merge proposals? there's no need to resubmit them for new revisions to fix small needs fixing issues; and that's not how they get resubmitted anyway. :) [12:44] which reminds me I should add myself to some spot for sponsoring [12:44] well I mean have a day a week and rotate between team members rather than having a team focussed on that [12:45] seb128, oh, nice [12:45] dobey: hm, I interepreted "resubmit" as the flag "I addressed the things the needsfixing comments raised" [12:45] seb128: you mean one day a week of nbs/ftbfs/stability checks, or one day per member? [12:45] dobey: so that's not how it's supposed to be used? [12:45] dobey: I. e. "I fixed stuff, and now I ask you to look at the MP again" [12:45] pitti: no. that resubmit vote is "i think you should resubmit this as a new proposal" [12:46] cyphermox, well I mean I prefer having something similar to what we do for sponsoring, a rotation or regular schedule of maintenance slots [12:46] dobey: is that what's holding tarmac back to actually merge it? [12:46] just adding the comment "I fixed that." is enough to send mail out to the reviewers. voting on one's own proposal really doesn't make any sense, and i'm not entirely sure whey LP allows it :) [12:46] pitti: no [12:46] rather than having a few people full time on it for a while [12:47] yeah, well, people on it full time for a while, I can see some getting tired / bored from it [12:47] dobey: ok, thanks for the explanation; some parts of LP are still a mystery to me [12:47] like some people like to alternate their activities, and having a few people full time give the message to others that they don't need to care because there is a team assigned to it [12:47] seb128: sounds fine to me, too; and with our current size, it'll amount to about three people each day anyway [12:47] which is the same size as we had in mind for the squad [12:48] pitti: you're not part of the "official" reviewer team, so tarmac just ignores your votes anyway. if one of those branches is approved, and it's not landing, likely it doesn't have the commit message set on the merge proposal [12:48] ah [12:48] dobey: right, it doesn't have one; I'm not sure how to create one [12:48] ah, "Set commit message" [12:48] dobey: want me to set one, or do you developers prefer to do that? [12:49] pitti: you can, just make sure it's a good commit message :) [12:49] ah, I was thinking just alternating one member each day of the week to look after these [12:49] dobey: "Fix stuff", what else? [12:50] pitti: "Allow using either GI or static bindings" maybe? "fix stuff" is NOT a good commit message :) [12:50] pitti: if i put "fix stuff" in an upload to ubuntu, you'd bitch at me :) [12:51] seb128: fwiw, gtkhtml is pretty much ready, and I put gbrainy next to my name because I already have much of the mono stack installed, so it should be easy to build/test [12:51] dobey: set [12:51] cyphermox, thanks! [12:51] dobey: (you didn't take that serious, did you) [12:51] oh man, gtkhtml. them was the days [12:51] cyphermox, do you have any clue if we should go for the new libgdata btw? [12:51] dobey: does it look ok to you at https://code.launchpad.net/~pitti/ubuntuone-client/gi-fixes/+merge/72447 ? [12:51] seb128: do we need it for something? [12:52] dobey: I don't know whether you have a particular convention wrt. formatting, or single header line etc. === MacSlow is now known as MacSlow|lunch [12:52] cyphermox, not that I know of, it's 0.9 to 0.10 so I was wondering if it's unstable to stable serie [12:52] cyphermox, it's better to be a stable serie if that's the case [12:52] cyphermox, but otherwise no need to update [12:53] Laney: ping, got a minute or 10? [12:53] it could be, but Id check carefully. if it drops auth mechanisms in favor of oauth this will mean breaking things like totem again (saying this because the old auth mechanism is deprecated by google) [12:53] pitti: well, we have some disagreement on that ourselves; but we don't generally list changed files and what changes are in them directly. a general statement or set of statements, describing hte set of changes makes more sense i think [12:54] dobey: ah, that's what I mean; in my stuff I prefer listing the files [12:54] dobey: ok, reworking then, before tarmac gets to it [12:54] dobey: done [12:54] bigon: ok, I'm ready for sponsoring [12:54] seb128: it looks rather a lot like bugfix, checking quickly [12:55] oh, gtk2 indicators should still work, right? [12:56] cyphermox, they should yes [12:56] I've had the initial fix to the indicator-network ftbfs in a merge request for a little while now, so if it's not absolutely necessary to port it to gtk3 then it probably works [12:57] oh [12:57] I'll give it a shot later [12:57] that's a system indicator right? [12:57] yeah [12:57] you can't mix gtk2 and gtk3 ones [12:57] d'oh [12:57] so you need to go to gtk3 if you want it loaded by unity [12:57] what I feared [12:57] well I wouldn't bother much [12:57] well, it's in progress, but I hit a snag in vala code [12:57] it's not our job to fix it [12:58] it's only an universe non maintained indicator [12:58] sure it would be nice to fix but it's lower priority than other desktop work [12:58] if dx wants it they should fix it [12:58] how can I prevent Evolution from re-creating the ubuntu one couchdb address book? [12:58] cassidy, uninstall evolution-couchdb? [13:00] seb128, that did it, thanks! [13:00] cassidy, yw [13:01] cassidy, you can check with rodrigo_ when he's back from lunch on how to turn it off for one user when evolution-couchdb if that's installed if that's possible, he probably knows === m_conley_away is now known as m_conley [13:02] seb128, that's fine I don't use it so it's easier for me to just remove the pkg [13:03] cassidy, ok === hggdh_ is now known as hggdh [13:22] gosh i can't understand this unity alt+tab :-( [13:24] pedro_: unity has clearly stolen gnome-shell's AltTabConfusion+ technology [13:24] in all fairness, though, we just stole it from macos [13:24] pedro_: you mean because it pops up behind the nautilus background window? :) [13:25] cause i kind of need extra fingers to be able to switch between apps with 2 or more windows [13:26] oh, did it change today? [13:26] pitti: https://launchpad.net/ubuntu/+source/network-manager-pptp/0.9.0-0ubuntu1 [13:26] desrt, the gnome-shell one works better IMHO at least i don't feel lost when i'm using it [13:26] (just pointing out it's done) [13:27] pedro_: you're smarter than me, then :) [13:27] * desrt misses his good old-fasioned application-unaware alt+tab [13:27] i think i'll just buy an EPOC and use it to switch windows [13:28] cyphermox: yay [13:31] i find the behaviour of the alt+tab switcher a bit confusing, especially with it showing everything on all workspaces. [13:32] but i'm so glad the thumbnails have gone [13:32] those were useless [13:32] chrisccoulson_: that's because you have one browser window with a million tabs :) [13:32] thumbnails are quite useful for me [13:32] heh [13:33] especially with lots of emacs/terminal/firefox windows [13:33] chrisccoulson_: is the 14_no_fade_on_user_switch.patch patch in gnome-screensaver still necessary? [13:33] i could never tell my e-mail client, browser and text editor apart with the thumbnails [13:33] mdeslaur, i'm not sure. i was wondering that last week [13:33] i can't even remember the problem it fixed ;) [13:33] well, alt+tab in metacity is nice [13:34] chrisccoulson_: yeah, the original bug seems to indicate it may have been fixed with a -nvidia update, etc. [13:34] mdeslaur, lol [13:34] nvidia? [13:34] ;) [13:34] i surprisingly still have unity running on my one laptop [13:35] dobey, nessita: what is left for me to do on https://code.launchpad.net/~pitti/ubuntuone-control-panel/gi-fixes/+merge/72449 ? [13:35] chrisccoulson_: mind if I disable the patch to get the fade hotness back? [13:35] mdeslaur, i don't mind. if you think it's worth it ;) [13:35] do you have an nvidia card to test it on? [13:35] pitti: was there a failing test, I think. let me check! [13:35] chrisccoulson_: it makes me feel warm and fuzzy when my screen fades :) [13:36] chrisccoulson_: actually, I do...I'll see if I can reproduce it with oneiric. [13:36] pitti: oh, i guess i'll have to make releases of everything for these changes. i presume it's ok to get all these in now, even though we're in beta freeze? [13:36] dobey: we are not in beta freeze yet [13:37] dobey: 2100 UTC [13:37] pitti: oh. *all* freezes changed to 2100? [13:37] mdeslaur, Guinness makes me feel warm and fuzzy [13:37] I'm not sure, but the recent ones were [13:37] pitti: right, there is a failing test. Let me give you a hint how to fix [13:37] pitti: ok [13:37] chrisccoulson_: so does that hot sun out there [13:38] pitti - you have sun? [13:38] chrisccoulson_: you don't? we have nothing else since saturday [13:38] pitti - we've had a bit of sun, and it was sunny this morning [13:38] but it's just got so dark here that i might have to switch the light on [13:38] I use a very advanced technology to keep my brains within specified temperature parameters [13:38] ... called "eating lots of ice cream" [13:38] heh :-) [13:39] i'd love some ice cream, but i'm trying to be disciplined with my food intake ;) [13:40] pitti: error, solution and needs fixing request added to https://code.launchpad.net/~pitti/ubuntuone-control-panel/gi-fixes/+merge/72449 [13:40] nessita: cheers [13:46] chrisccoulson_, mdeslaur: upload, if that's broken somebody will complain and we can put it back [13:47] seb128: ok, will do, thanks [13:49] cassidy, evolution-couchdb always create the source if it's not available [13:49] maybe we should make it be more clever and only show if you have an u1 account setup [13:50] seb128, any reason to not have gobject-introspection-doc package? [13:50] afaik they have some gtk-doc [13:52] xclaesse, that's a question for pitti [13:52] pitti, ^ [13:55] seb128, I assigned bug 831812 to the desktop team, not sure where the bug lies. It makes ubiquity crash on launch. [13:55] Launchpad bug 831812 in gtk+3.0 "ubiquity fails to start: Gtk:ERROR:/build/buildd/gtk+3.0-3.1.12/./gtk/gtkcssprovider.c:1275:gtk_css_scanner_new: assertion failed: (data[length] == 0)" [Critical,Confirmed] https://launchpad.net/bugs/831812 [13:56] jibel, it's not likely gtk itself no [13:56] seems like they have some buggy css [13:57] xclaesse: libgirepository1.0-doc ? [13:57] seb128, pitti: bah it is libgirepository1.0-doc [13:57] that's for the API [13:57] hehe, found at the same time :) [13:58] thanks [14:03] ev, seb128 about 831812 buggy css in ubiquity or gtk+3 bug ? [14:04] jibel, ev: well, I will have to check but gtk didn't change for a while and that seems a css parsing issue [14:04] so without looking at it I would first check if ubiquity is shipping a custom css and if it's a valid one [14:08] I don't think segfaulting is appropriate, even if it's invalid CSS [14:09] http://paste.ubuntu.com/674535 is the CSS in question [14:10] and the fact that it works if you run it enough times leads me to believe it's not a matter of invalid CSS [14:11] ok, maybe not [14:11] well as said I've no read it [14:11] not [14:12] I've only just jumped in here, so I'm merely stating my take on it. I don't mean for it to come across as confrontational. [14:13] pitti: hrmm. apt-cache policy ubuntuone-installer does not show it as being in the repository? [14:14] Laney: when you have a minute or three, I'm stuck with dbus-sharp ... :) [14:14] dobey: might still be in binNEW? [14:14] pitti: ah, it looks like it still is, yeah :-/ [14:15] ev, oh you didn't, it's just that things are busy as well there so I didn't have time to look at it ;-) [14:15] understood :) [14:15] ev, it's just that gtk didn't change for a bit and no other application hit that issue [14:15] ev, so either ubiquity is doing special things, or is unlucky or has a bug [14:15] will try to look at it once I'm done with the unity updates [14:16] okay, thanks [14:19] Chipaca: not really, try #tomboy if you need help [14:19] or #mono [14:20] Laney: ok, thanks [14:20] tedg: poke [14:20] desrt, howdy [14:20] tedg: can you do an upload for me to the dx team ppa? [14:21] basically, (1) track the newer glib release (2) new patch from me [14:21] otherwise same story as last time [14:22] desrt, Sure [14:23] i'll mail the patch [14:32] does somebody want to do some review and sponsoring for dx? [14:32] http://bazaar.launchpad.net/~unity-team/unity-lens-applications/trunk/revision/222 [14:32] https://code.launchpad.net/~dbarth/zeitgeist-extensions/fts-cjk-support/+merge/72903 [14:32] [14:33] those are dbarth's changes to turn on cjk xapian for unity [14:35] ugh - lucid really has pkg-config broken out of the box? gnome-doc-utils ships xml2po.pc which deps on libxml2.pc, but the gnome-doc-utils package doesn't dep on libxml2-dev [14:36] walters, that would be having g-d-u depends broken, which is different of having pkg-config broken [14:36] i.e those depends are not autogenerated, they are manually defined [14:37] seb128, it's true pkg-config --list-all doesn't blow up, but my jhbuild code is calling --modversion on each one [14:37] hm, this may actually be a gnome-doc-utils bug; it's probably wrong for xml2po.pc to require libxml2.pc since it only needs runtime, not the headers [14:38] (and gnome-doc-utils appears to be installed by default due to ubuntu-docs) [14:40] seb128: for the application lens, will that mean you can actually send ibus-generated characters in the search bar? [14:40] dbarth, ^ [14:41] cyphermox, that's the goal but I'm not sure if that still needs an unity side change as well or not [14:42] ok, new theme selector (in appearance panel in g-c-c) uploaded, will apreciate some testing [14:42] seb128: well, didn't work yesterday, it either didn't take any characters or just got the scancodes/keycodes, whatever [14:42] cyphermox: no, not yet [14:42] rodrigo_, great work, thanks [14:43] cyphermox: that's a separate bug, the upper half side [14:43] we have a theme selector again? [14:43] dbarth: oh ok, still planned to land soon? [14:43] cyphermox: the cjk changes you're seeing here are a continuation of the libxapian changes, the bottom half part of the whole CJK story [14:43] cyphermox: should be in a ppa by the end of this week, the bulk of the work is now donw [14:43] the upper half missing makes the lower half hard to verify ;) [14:44] cool [14:44] seb128, hm, though one could also argue that there should be a gnome-doc-utils-dev package which has the xml2po.pc [14:44] cyphermox: you can verify that with unity-2d [14:44] cyphermox: which already does support ibus [14:44] ohh [14:44] ehh [14:44] thanks for the hint ;) [14:44] nw ;) [14:45] anyway, we'll see how it goes [14:45] walters, indeed, would be the "normal" debian way [14:46] seb128, hm, fedora has this same bug actually. blah. I think i'll fix gnome-doc-utils [14:47] though there is this general tension that some people want to move pkg-config files to be useful for runtime detection and not just devel packages [14:52] walters, that's not what .pc are made for though, it's the same sort of issue that python checking for python bindings in their configure because they are needed a runtime [14:52] right [14:53] it doesn't solve the issue since those don't convert in real depends but they do create build time issues or unrequired build-depends for distributors [14:54] alright i did https://bugzilla.gnome.org/show_bug.cgi?id=657339 and will move on [14:54] Gnome bug 657339 in build utils "Fix xml2po.pc to not Require libxml.pc" [Normal,Unconfirmed] [14:54] it's just annoying to work around in jhbuild =/ [15:02] wow, it really is dark here now [15:02] what happened to the daylight? [15:05] chrisccoulson_, Hah, America takes one more thing from the old world! [15:05] lol [15:06] chrisccoulson_, you get compiz grid sun, no need of extra light ;-) [15:07] heh :) === dpm_ is now known as dpm === tkamppeter_ is now known as tkamppeter [15:17] pitti: status update for LO: it builds on arm past way the last buildbreaker, succeeds on intel, havent tested yet on ppc (last time I tried, apt-get 404'ed on installing deps) [15:18] oh, *fiddlesticks*. looks like dbus-sharp 0.7 has a bug wrt receiving signals [15:19] pitti: Do I need a FFE at all since we are not introducing any new features (3.4.2 is a bugfix release)? If so can I reuse issue 828724? [15:25] Sweetshark: no, that's fine [15:25] i. e. no FFE necessary [15:28] pitti: k, preparing an upload === htorque_ is now known as htorque [15:39] who do I need to talk with about a patch to dbus-sharp? [15:39] (I've got a patch, bug is fixed upstream in git) [15:51] Chipaca: official way is to open a bug for it, point to the patch, and subscribe ubuntu-sponsors [15:51] pitti: ok [15:51] Chipaca: I guess everyone is in pre-beta-freeze rush (including me), so feedback might be a little thin right now [15:51] yeap [15:51] Chipaca: but we'll get it sponsored quickly [15:52] nessita: ah, I totally didn't notice the two new pylint warnings amongst the other 3000 it already had, sorry; will fix them [15:53] pitti: thanks! (you should be getting 3000 because you may not be using our patched pylint version) [15:53] (which makes sense) [15:53] this is running u1lint [15:53] from the packaged u1dev-tools [16:08] nessita: is there a trick to run u1lint on just one file? it takes ages otherwise [16:08] * pitti still fighting with the #pylint magic markers [16:08] pitti: no with u1lint, no :-/ NO worries, I can run it myself and tell you if you need to fix something [16:09] nessita: ah, that's fine then; I just wait :) [16:09] I hope I have it nailed now [16:10] yay [16:10] nessita: pushed, shoud both be fixed now [16:10] great! [16:10] test suite runs cleanly now with pygobject 2.28 [16:10] there are a couple of inevitable failures with 2.90 [16:11] (but the branch fixes a lot of the regressions, the remaining ones are specific to the test suite) [16:11] bbl [16:11] ok, need to leave for a bit, back for TB meeting === htorque_ is now known as htorque [16:13] Desktop is rather crashy and unstable. Have sent my whole /var/crash collection now. There happen between 6 and 10 segfaults after login. [16:13] yeah, but most of them are in lightdm [16:13] for me, anyway [16:14] it's mterry's dault! [16:14] pitti: Yes, one in unity-greeter, two in gnome-settings-daemon. [16:15] joke aside it's mostly noise [16:15] would be nice gvfs etc was not running in the greeter to start [16:15] well, the g-s-d one isn't noise, but the rest is, yes [16:15] pitti, and in addition, compiz seems to be a complete mess: when moving a window, the window is drawn on its new position but not removed from its old position [16:15] pitti, so it leaves traces. [16:15] oh, I don't have that here [16:15] pitti, do you have a bug number of the g-s-d issue? [16:16] I've declared retracer emails backrupts while they were catchup yesterday [16:16] pitti, and when answering a crash dialog, the dialog closes but the window's shadow stays. [16:16] for the theme bug? no [16:16] tkamppeter: that sounds like a bug that has been fixed in compiz two days ago, hmm [16:18] pitti, I have updated today, and it pulled compiz and I logged in afterwards ... [16:18] tkamppeter, does it happen in a guest session? [16:20] logged in again and five new crashes in /var/crash, and desktop does not show indicator upplets, and Unity launcher bar at the left does not open. [16:20] tkamppeter, do you have unity and compiz-gnome installed? [16:21] seb128, will try ... [16:21] seb128, yes, both. Do I have to uninstall one of them? [16:22] tkamppeter, no, I was just checking you have those [16:22] tkamppeter, can you pastebin your .xsession-errors? [16:22] brb session restart [16:23] ok, really off now, bbl [16:24] seb128, pitti, http://paste.ubuntu.com/674608/ [16:24] seb128, pitti, http://paste.ubuntu.com/674608/ [16:26] tkamppeter, ldd /usr/lib/compiz/libunityshell.so [16:26] can you pastebin that? [16:26] unity is not loading for you [16:27] seb128, and in addition, network-manager has also a problem. It cannot access to my WLAN. It tries, after a timeout it shows the password dialog for wireless (I have checked network name and password are correct) click Connect, tries again and after a timeout again the password dialog ... [16:28] seb128, ldd: /usr/lib/compiz/libunityshell.so: No such file or directory [16:28] tkamppeter, dpkg -l | grep unity [16:28] seems like you don't have unity installed [16:28] dpkg -S libunityshell [16:28] dpkg-query: no path found matching pattern *libunityshell*. [16:29] seb128, http://paste.ubuntu.com/674611/ [16:30] tkamppeter, unity is not installed [16:30] install it, it will work better [16:31] seb128, http://paste.ubuntu.com/674614/ [16:31] tkamppeter, yeah, you just picked the wrong time, the new unity is building, try again in 15 minutes [16:32] tkamppeter, or get the debs on https://launchpad.net/ubuntu/+source/unity/4.10.0-0ubuntu1 [16:36] DBO, hi :), is http://bazaar.launchpad.net/~unity-team/bamf/trunk/revision/409 suppose to fix this https://bugs.launchpad.net/bamf/+bug/826073 ? [16:36] Ubuntu bug 826073 in bamf "Segfault when apps quit that have been launched before Plank did." [Undecided,New] [16:36] seb128, trying to install ... [16:38] seb128, problem is unity-common, it is not under "Built files" on https://launchpad.net/ubuntu/+source/unity/4.10.0-0ubuntu1/+build/2744623 [16:38] tkamppeter, it's in the i386 build [16:38] ricotz, yeah [16:38] tkamppeter, the arch all binaries are built on i386 [16:38] DBO, nice [16:40] seb128, found it. [16:42] seb128, unity is now installed and I am logging in again. [16:42] is it working? [16:44] seb128, great, only one crash after login! [16:44] ok, that's better [16:44] pitti: upload to chinstrap is running, please sponsor [16:44] did you get your launcher back and the refresh issues fixed? [16:45] seb128, launcher is there, rest I am trying by reporting the crash on the machine's own desktop (and not via SSH). [16:50] seb128, from the total of 6 crash reports (5 from last session, 1 from this session, I got 3 to the browser until I lost access due to some desktop problem which I did not succeed to identify ... [16:51] seb128, I cannot move windows any more, nor can I click buttons on crash report dialogs or in firefox. [16:51] is the dash open? [16:51] i.e super key [16:52] hmm, since pitti is afk is anyone else volunteering to sponsor libreoffice-3.4.2-2ubuntu1 and libreoffice-l10n-3.4.2-2ubuntu1 pre-beta? (both can be found on chinstrap) [16:52] seb128, I do not know, how to determine this? [16:53] try pressing super? [16:53] ouch, colord added quite a bit to the CD :/ [16:53] seb128, Alt key? [16:53] super key [16:53] the microsoft logo key [16:54] Sweetshark, nice :) [16:55] seb128, Left-Windows gives me a mask with a search input line near the top and big icons for common tasks in the center of the screen. The desktop is blurred. [16:56] tkamppeter, ok, that's the unity dash, you can press it again to close it [16:56] if click still doesn't work I don't know why [16:56] you can ask on #ayatana [16:56] chrisccoulson_, how much is the bit? that was supposed to be a bit less than 1mb [16:57] seb128, it seems that window decorations are still visible but not sensitive. [16:57] seb128, well, it pulls in icc-profiles-free, which is 1.5MB alone [16:57] tkamppeter, dunno then [16:58] Anyone here who knows about window decorations? [16:58] chrisccoulson_, would it be of any use without those profiles? [16:58] tkamppeter, try #ayatana [16:58] seb128, i'm not sure [16:58] chrisccoulson_, we should perhaps lower that to a suggest? [16:59] seb128, yeah, possibly. i'm probably not the right person to make that decision though [16:59] i don't know enough about it ;) [16:59] tkamppeter, RAOF: ^ [17:01] seb128, icc-profiles-free is part of the new color management stack. it is a Recommends: in colord and provides the basic color profiles, which are needed to make the color management working at all. [17:01] tkamppeter, it's taking 1.5mb of CD space [17:01] can we lower that? [17:01] tkamppeter, what happens if the profiles are not there? [17:01] seb128, I think without the profiles color management simply will not work. [17:02] what devices do those profiles cover? [17:03] seb128, I think there are once the standard profiles, like sRGB and AdobeRGB, and perhaps some devices but I do not know exactly. Perhpas RAOF knows more. [17:03] ok [17:04] seb128, I am also not a CM expert, I have simply taken the packages which the color management people recommended. [17:05] seb128, what is this #ayatana channel? Is it about X, compiz, ... [17:05] tkamppeter, #ayatana is the dx team channel, unity, compiz, etc [17:07] seb128, thanks. [17:07] yw === abhinav_ is now known as abhinav- [17:12] ricotz: yes, I work 24 hours a day and if thats not enough I do work during the night. [17:15] Sweetshark, hehe ;) -- looking forward to 3.4.3 :P [17:16] the greatest problem of this thing is the buildtime :\ [17:18] seb128, do you think you can fix up the seeds for the u1 packages? [17:19] seb128, bug 833824 [17:19] Launchpad bug 833824 in ubuntu-meta "Remove Ubuntu One packages from Oneiric CD" [Undecided,New] https://launchpad.net/bugs/833824 [17:19] kenvandine, why me? ;-) [17:19] you should have access as well [17:19] seb128, no reason... :) [17:19] i do... just over loaded atm [17:20] if you have time, it would be great :) [17:20] ok, same here [17:20] ok [17:20] no worries [17:20] but I will sort it with pitti tomorrow [17:20] want it to be done before beta freeze [17:20] kenvandine, did you discuss the transitionnal binary thing? [17:20] ok [17:20] kenvandine, well the freeze is mostly an ui freeze, we can still get bug fixes in tomorrow [17:20] yes... he said just to make it easier if they ever bring it back [17:20] kenvandine, I expect we will clean up ftbfs and installability issues tomorrow [17:21] ok [17:21] i think i prefer adding the conflicts, breaks stuff instead [17:22] seb128, should i make rb conflict/break with rb-u1-music? or make rb-u1-music conflict/break rb? [17:22] the former would be easier, because right now i think the plugin is FTBFS :) [17:23] kenvandine, rb conflicts with it [17:24] so make it conflict and break right? [17:24] oh [17:24] it already does [17:25] i wonder why ted had upgrade problems... [17:25] oh, i guess an ordering issue... since it only breaks... doesn't conflict or replace [17:29] #ayatana is dead, anyone here who can help on insensitive window decorations? [17:31] tkamppeter, sorry... i think everyone is scrambling for UIF and beta freeze [17:33] tkamppeter, try restarting compiz [17:33] compiz --replace [17:43] seb128, tkamppeter - the color management control center panel seems to depend on gnome-color-manager too, which isn't installed by default [17:43] and that's another 2.7MB ;) [17:43] chrisccoulson_, what is it needed for? [17:43] chrisccoulson_, not sure ";)" is appropriate... ;-) [17:44] seb128, gcm-viewer, which seems to be needed for the "View details" button in the control center panel [17:44] seb128, "color-cc-panel-WARNING **: failed to run calibrate: Failed to execute child process "/usr/bin/gcm-viewer" (No such file or directory)" [17:44] and that takes almost 3mb? [17:44] we can probably patch that out [17:45] seb128, yeah, it's around 2.7MB [17:45] and it's in universe [17:45] I'm wondering how much of those calibration are useful without ui === _LibertyZero is now known as LibertyZero [17:46] can sponsor LibO, but Sweetshark seems offline now [17:47] hah, and installing gnome-color-manager pulls in argyll [17:47] chrisccoulson_: 2.7 MB out of the blue? [17:47] meh, this is really not the kind of change we can make at this point [17:47] pitti: ~3.25 hours left, right? :) [17:47] dobey: yep [17:47] 2.7MB + 14MB of argyll ;) [17:47] pitti, chrisccoulson_: well it's only a button calling an ui, we could just hide the button [17:48] not sure how much of that is actually needed to make the control center panel work though [17:48] pitti: ok, i'll make some releases real quick with your gi/static fixes [17:48] chrisccoulson_: nope, argyll was determined for universe, it'll be installed on deman [17:48] d [17:49] oh, after installing gnome-color-manager, the control-center button still doesn't work [17:49] and i get a different error ;) [17:49] well let's see what comes out from beta testing [17:50] we can roll those out if needed [17:50] or split the color.so in a new binary [17:50] Sweetshark: hm, this has changelog target UNRELEASED [17:50] which depends on the universe cracks [17:50] so those who want to install it can [17:50] Sweetshark: I'm uploading -l10n now, that seems fine [17:51] Sweetshark: can you regenerate the .dsc, debian.tar.gz and sources.changes for libo? (no reupload of orig tarballs necessary) [17:52] nessita: nice, so all the ubuntuone parts should be good now for the new pygobject [17:52] nessita: thanks! [17:52] dobey: cheers [17:52] pitti: thank you! but, you mentioned that desktopcouch needed a fix as well, no? [17:53] nessita: I'm not sure yet [17:53] but I'll keep it on the watch list [17:53] thanks, let me know [17:53] pitti, nessita: desktopcouch fix? [17:54] oh for gi? [17:56] yes, possibly [17:57] desktopcouch uses only static glib bindings [17:57] so should be fine [17:57] modulo the other problems it has of course :) [17:58] at least I get the very same "couchdb.http.Unauthorized: ('unauthorized', 'Authentication required.')" exception with pygobject 2-28 [17:58] so that exception doesn't seem to be a regression in pygobject [17:58] but that makes neither the backend nor gtk UI crash [17:59] and when the GUI finishes, I get a "Fatal Python error: PyEval_RestoreThread: NULL tstate" [17:59] that also happens with both the old and the new pygobject [17:59] and is a symptom of mixing static and GI bindings [17:59] so I think it doesn't block the new pygobject, but will have to be fixed anyway; I'll keep that on my list [18:01] it needs to be fixed, yeah; and i guess converted to dh_python2 would be nice [18:01] pitti: i think those errors are more a result of new couchdb in oneiric or something :( [18:02] the "auth required" one, yes [18:02] the PyEval one, I'm not sure [18:02] dobey: I wonder if I should already uplaod the new pygobject, or wait for the u1 releases? === davidcalle_ is now known as davidcalle [18:02] pitti - the couchdb error has been present for me in oneiric for the entire cycle so far [18:03] it's also the cause of bug 825280 :/ [18:03] Launchpad bug 825280 in thunderbird "Repeated Ubuntu One address book errors" [Undecided,Confirmed] https://launchpad.net/bugs/825280 [18:04] pitti: you can upload i think [18:12] hah, that was silly of me. i dist-upgraded earlier and clearly wasn't paying any attention [18:12] and now compiz has just crashed, i realize that i no longer have unity installed ;) [18:17] chrisccoulson_, than you have the same mess that I had ... [18:20] i'm back to normal now, except the menubar doesn't work on my secondary screen anymore [18:24] Sweetshark: AYT? [18:24] tedg_: hey. did you do that upload yet? [18:24] >:| [18:24] tedg: or here? :) [18:25] tedg_: !!! [18:49] pitti, how much do we care about powerpc? [18:50] neither firefox or thunderbird currently build on it [18:50] target of opportunity [18:50] ah, ok. thanks [18:50] i wasn't sure if i should be putting effort in to fixing that before beta [18:50] but i'll look at it if i get a chance anyway [18:51] chrisccoulson_, by the time you had the issue unity was published for 2 hours, stop using an outdated mirror ;-) [18:52] seb128, oh, the issue was that i upgraded earlier on, but didn't realize i'd uninstalled unity until compiz crashed a couple of hours later ;) [18:52] chrisccoulson_, no dist-upgrade for you! [18:52] chrisccoulson_, you should join the "upgrade" club [18:52] ;-) [18:53] heh [18:54] pitti, there is one bad thing about Apport: For each crash which happens I get a window "System program has crashed. Report, Cancel", not telling *which* program has crashed. If I click on "Report" I get asked for my password and after that the usual Apport interface comes up. With the current messy desktop I have to click "Report" several times and everytime enter the password. This is annoying and demotivates to report all crashes. [18:55] tkamppeter: I know, that's a tricky thing, as we can't make the system reports user readable [18:55] pitti, this was also the reason why I did not see the crash of unity-greeter. [18:56] pitti, can one not at least extract the package names for normal users? [18:56] pitti: can't for "security" or can't because the user doesn't have permissions to open the file? [18:56] we can reconstruct the program name at least [18:57] pitti: right, i was just going to suggest that, by parsing the filename, since the user can see that [18:57] dobey: we don't want mere mortal users to look at core dumps of e. g. postgresql, apache, shadow, or root daemons [18:57] the file name encodes the program name, so we can use that [18:57] right [18:57] so that is doable :) [19:02] pitti, well you could ask the password one and keep the credential for the serie of bugs queued [19:02] once [19:02] also, update-notifier shows way too many windows [19:03] it seems it reports the same .crash file over and over again [19:03] there's another bug there [19:03] yeah, ideally they would add in one ui [19:03] pitti, well, it seems every time you get an new bug it reprompt for all the bugs you didn't report yet [19:03] right, the system ones [19:03] users one as well [19:04] seems the "mark as seen" stuff isn't working there [19:04] is there a mark as seen? [19:05] yes, in apport/fileutils [19:05] well I think we should do something similar to update-manager [19:05] it's by and large just looking at atime != mtime [19:05] open one ui with the list of unreported bugs unfocussed [19:05] rather than prompt for each bug [19:05] yeah [19:07] is there a standard way to tell dpkg-buildpackage "redo the debs"? [19:07] like "debuild binary" works for most GNOME sources [19:07] it doesn't clean, just start again at make and what follows [19:07] I usually use debuild -us -uc -b -nc [19:07] but that doesn't work on unity, annoying! [19:07] -nc -> "no clean" [19:08] it's so incredibly useful, I call this "dreb" in an alias :) [19:09] doesn't work in unity either [19:09] yours is probably similar to "debuild binary" [19:09] which is what I use usually [19:09] unity is using cmake though, I guess it's another cmake stupidity (hate cmake) [19:15] https://code.launchpad.net/~dobey/ubuntu/oneiric/ubuntu-sso-client/release-133/+merge/72943 [19:15] pitti: ^ one down [19:17] dobey: can you upload yourself, or want me to sponsor? [19:18] pitti: if i could upload, i would have done so instead of making a merge proposal :) [19:18] i don't have perms for ubuntu-sso-client, or ubuntuone-control-panel, so they will need sponsored [19:19] dobey: merged/pushed/uploaded, thanks! [19:19] need to get that ubuntuone package set setup [19:21] bah, it's annoying that packaging tools can figure a way to include binaries update in the diff they generate [19:22] pitti: oh; do you know if twisted has been fixed for this already? it has reactors that use the glib/gtk main loops and might cause some issues? [19:22] unity trunk got some icon updates [19:22] it's easy to bzr merge the revision [19:22] dobey: that's indeed causing trouble with the test suite, when using GI in them; that's why I didn't port the tests [19:22] but the packaging tools don't manage to make a diff of it [19:40] pitti: https://code.launchpad.net/~dobey/ubuntu/oneiric/ubuntuone-control-panel/release-113/+merge/72945 is ready now :) [19:40] pitti: and ubuntuone-client, i will upload directly since i can [19:40] dobey: oh, doesn't -client need to go first? [19:40] ah [19:41] dobey, pitti: if one of you upload that one can you check why its desktop doesn't use dh_translations or equivalent [19:41] i.e it doesn't have its translations stripped not a gettext domain [19:41] (I built a list of those today and it's on it) [19:41] cyphermox, ^ nm-applet is on that list as well [19:41] seb128: control-panel is? [19:42] seb128: can we figure this out after beta? [19:42] dobey, [19:42] $ grep Gettext ubuntuone-control-panel-gtk.desktop [19:42] $ [19:42] seb128: ok [19:42] # langpack.mk is relevant on Ubuntu only, not Debian; it does not matter if it's missing [19:42] -include /usr/share/cdbs/1/rules/langpack.mk [19:42] pitti, yeah for sure, I though it would just be adding one line to the rules and could go in the same upload [19:42] seb128: ^ that would be it [19:42] it doesn't exist any more [19:43] I'll add dh_translations [19:43] thanks [19:43] pitti, sorry fo adding work on the beta day [19:43] I will come tomorrow with the other ones [19:43] np [19:43] pitti: should i do it? [19:43] dobey: I'm on it [19:43] pitti: i can make it use pure dh real quick [19:43] i thought it already did [19:44] but alas it doesn't [19:44] oh well [19:44] dobey: for u-client? [19:44] I'm currently test-building control-panel with the fix [19:44] pitti: for ubuntuone-control-panel [19:45] what is the fix? [19:45] gimme a minute to test [19:45] ok [19:48] back (just saying good night to my wife) [19:48] mterry, "funny bug" there [19:48] * mterry hopes this is the setup to a joke [19:49] mterry, it happened twice this week so I think there is a bug (I'm running the current lightdm with your fixes from yesterday) [19:49] dobey: http://paste.ubuntu.com/674762/ usually [19:49] mterry, got a "lightdm" locked screen over my user session when coming back from dinner [19:49] but it seems that doesn't work with -gtk [19:49] mterry, I think I closed the lid when going to dinner [19:50] seb128, interesting. And you couldn't get back in because lightdm doesn't have a password? [19:51] mterry, indeed, I get the locking screen with no input password [19:51] it was animating out and back on screen at every key press [19:52] seb128: ah, for control-panel, did that ever work? [19:52] seb128, seems bad [19:52] seb128: I don't think it's clever enough to figure out the domain from the build system [19:52] dobey: ^ FYI [19:52] at this point, I punt this to post-beta [19:52] pitti, oh, I don't know, I just grepped in the installed desktop to figure which ones don't have a domain [19:52] pitti, yeah, don't bother, I've a list I will open bugs tomorrow [19:53] pitti, "post-beta", means we can't fix those tomorrow? [19:53] seb128: we can [19:53] I was hopping tomorrow would be a good day for remaining ftbfses and a few of those [19:53] I mean "post the crazyness that is going on right now" [19:53] pitti, ok, great [19:54] damn... my gwibber upload keeps stalling! [19:54] dobey: merged/pushed/uploaded [19:54] * kenvandine hates his isp right now [19:54] go, mterry, go! [19:55] pitti, :) [19:55] * mterry notes that now, simply logging in will not cause crash reports [19:55] mterry: wow [19:55] mterry, awesome! [19:55] this morning I still got about 5 or so [19:56] mterry: did the greeter stop spawning an insane number of processes now? [19:56] pitti, you mean spawning and leaving? [19:56] evo-rss ftbfs/nbs/uninstallability almost done fixing :) [19:56] pitti, it might still spawn a few [19:56] mterry: yes, that [19:56] they just don't crash [19:56] mterry: also, it spawns a lot of stuff that it shouldn't in the first place [19:56] pitti, but they all shut down when logging in now, yes [19:56] mterry, thanks for backporting those fixes ;-) [19:56] cool [19:57] pitti: ok, cheers! [19:57] pitti, really, the extra spawned stuff it shouldn't are all indicator bugs that should be doing less when INDICATOR_GREETER_MODE is on. simply a matter of hunting each down [19:58] Ack! lightdm process are sticking around again [19:58] * mterry bets the patch got dropped from packaging; it's not upstream yet [20:00] chrisccoulson_, how busy are you? ;-) [20:03] chrisccoulson_, ok, randomly asking in case somebody is interesting in making xdg-open not try to use gnome-open when not installed [20:03] is there anyone interested by working on that? [20:04] * jdstrand looks curiously at his name being displayed as '[Invalid UTF-8]' with the new indicator updates... [20:04] jdstrand, already fixed :) [20:04] jdstrand, should be fixed in .1 [20:04] uploaded about 30m ago [20:04] :) [20:05] hey, since people are listening to me: what happened to the applications menu that was in the launcher? [20:05] jdstrand, the icon starting the place (lens)? [20:05] hi seb128 [20:05] sorry, just popped away to do some exercise! [20:06] seb128: yeah-- I had two: 'Applications' and 'Files & Folders' [20:06] don't laugh ;) [20:06] lol [20:06] jdstrand, in the dash now [20:06] look at the bottom of the dash [20:06] chrisccoulson_, I read "to drink some guinness" don't worry ;-) [20:06] heh :) [20:06] i wish! [20:06] kenvandine: ah-- see, before today's update, that stuff wasn't working for me :) [20:07] so I thought they just disappeared [20:07] jdstrand, you will get back keybindigns for direct access as well [20:07] jdstrand, you are on a small screen? [20:07] actually no [20:07] well anyway if it's fixed great ;-) [20:07] pretty big screen (external monitor) [20:07] they were out of the screen on my nb which has 1024x600 [20:07] well dunno then but it's fixed it's not worth bothering ;-) [20:07] no [20:08] it was that I would click them and nothing happen. stuff is happening now, so that is good [20:08] :) [20:08] great ;-) [20:08] ok [20:09] so does anyone know an easy way to include binary updates in a package revision? [20:09] oh yeah, they didn't start loading data until you typed in the search [20:09] that was annoying :) [20:09] easy mean "not using uuencode or equivalent" [20:09] jdstrand, oh right, what kenvandine said, you needed to type something in the text entry to have then getting content [20:09] seb128: format 3.0-quilt should work? [20:09] pitti, no it doesn't [20:09] seb128: put it into debian/local/ or so [20:10] pitti, does it work for out of the debian dir files? why isn't it doing it for me? ;-) [20:10] seb128: no, just for debian/ [20:10] :-( [20:10] you can't put binary data into patches/ [20:10] and v3 only has debian/ [20:10] seb128, what's using xdg-open btw? [20:11] pitti, it's not patch, I want to bzr merge lp:unity and they had an icon update [20:11] seb128: put them into debian/local/ and cp them to the source tree in rules? [20:11] seb128: make a new orig.tar.gz? [20:11] pitti, yeah, can do, it just sucks :p [20:11] pitti, I will do that I guess [20:11] seb128: unity_4.8.2+bzr123.orig.tar.bz2 ? [20:12] pitti, yeah, that I can do, I was just wondering if there was an easier way [20:12] seb128: no, I'm afraid not; v1; uuencode, v3: debian/, that's it [20:12] s/;/:/ [20:13] pitti, thanks for confirming, that's what I though [20:13] seb128: oh, with v3 there's also another way [20:13] seb128: you can have multiple orig tarballs [20:13] this is why i'm glad the new thunderbird icons are all svg's, else i would have had your problem yesterday ;) [20:13] for lots of files [20:13] but I think they will be unpacked into subdirs, so that doesn't help too much (then debian/ would work, too) [20:14] I will redo a tarball I think [20:14] easier [20:14] chrisccoulson_, btw you don't reply, how busy are you? ;-) [20:14] seb128, i can probably look at xdg-open ;) [20:15] pitti: and ubuntuone-client uploaded now :) [20:15] dobey: yay you [20:15] dobey: too bad that ubiquity spoiled the fun :/ [20:15] pitti: well at least ubuntuone will mostly work [20:15] seb128, you could put them in place where they belong in the source tree, look how plymouth does that [20:16] seb128, oh, do you have gvfs-bin installed? [20:16] seb128, with debian/source/include-binaries [20:16] chrisccoulson_, thanks [20:17] chrisccoulson_, no [20:17] seb128, it actually tries to use gvfs-open, and then falls back to gnome-open [20:17] ah [20:17] chrisccoulson_, it's trying gvs-open first and then gnome-open? [20:17] we need to seed gvfs-bin [20:17] yeah [20:17] we should make xdg-utils recommends gvfs-bin? [20:17] seb128, i think it suggests it. i'm not sure how a recommends will affect other DE's though [20:18] yeah, it just suggests it currently [20:18] XFCE is using gvfs these days, but KDE certainly doesn't [20:18] bah [20:19] let's get gvfs recommends gvfs-bin? [20:19] yeah, i guess that would work too [20:19] or xdg-utils recommends gvfs-bin | kde-equivalent [20:19] gvfs recommends: gvfs-bin sounds less intrusive at this point? [20:20] works for me [20:20] yeah, sounds good [20:20] that will fix the unity trash icon doing nothing on the default install [20:20] seem they use xdg-open [20:21] really? [20:21] it's likely going to fix other issues as well [20:21] and they also look in the wrong place for the default mail client ;) [20:21] chrisccoulson_, yeah, that's where I hit the issue [20:21] the last time i started a guest session, the dash showed evolution as default, but the message menu showed thunderbird [20:21] debugging on my test machine [20:21] is there a bug for that already? [20:21] xsession-errors has quite some" xdg-open: gnome-open: not found" [20:22] can we stop using that? [20:22] chrisccoulson_, bug #804397 [20:22] Launchpad bug 804397 in xdg-utils "xdg-open depends on gnome-open, but it's not on the CD" [High,Confirmed] https://launchpad.net/bugs/804397 [20:22] surely there's no need to use that [20:22] we probably can but that's not going to happen tonight [20:22] we should still fix xdg-open anyway [20:22] i'm going to end up writing a patch to remove all use of xdg-open! [20:22] yeah, it should still work :) [20:22] so getting gvfs-open on the CD is right [20:23] ;-) [20:24] good night everyone! [20:24] 'night pitti [20:25] I'm not going to solve the ubiquity crash in 30 mins [20:25] *frustrated* [20:25] pitti, oh, and stop working 19 hours a day! :p [20:25] yeah, that'd help, too.. [20:25] * pitti waves, good luck with the remaining 30 mins :) [20:25] hehe [20:25] pitti, thanks, have a good night [20:26] good night pitti [20:30] pitti: Just wanted to check, considering that UIFreeze/BetaFreeze happens in a few minutes. I think it's important that https://code.launchpad.net/~gunnarhj/gnome-control-center/langfix/+merge/72792 is included in b1. Is it ok that rodrigo_ does it tomorrow? [20:32] pitti: Never mind, missed that you waved good night. [20:34] GunnarHj, yeah, tomorrow should be ok [20:34] GunnarHj, that seems closer from a bug fix than from an ui redesign anyway [20:34] seb128: Thanks! Agree. Was about to ask you, but you came first. :) [20:34] ;-) [20:35] brb, session restart [20:38] hah - http://blog.mozilla.com/metrics/2011/08/25/ [20:38] i wish we had a similar way to gather hard data like that ;) [20:38] but i bet you all turn off test pilot! [20:38] :P [20:39] chrisccoulson_, btw can we get http:// back in the urlbar ? ;-) [20:39] seb128, no ;) [20:40] dunno who decided to drop it but that's a stupid move [20:40] i think there is a pref actually [20:40] why do you miss it? [20:40] because I keep copying urls to emails and bug report that don't get highlighted [20:40] chrisccoulson_: watch them remove ctrl-f by using that metric as justification [20:40] nor working on click [20:40] seb128, you mean copying URL's from the addressbar? [20:40] that should actually include the protocol when you copy [20:40] if it doesn't, then it's a bug [20:41] that's definitely the intended behaviour [20:41] (ie, protocol is hidden, but it's copied if you copy the url from the addressbar) [20:41] chrisccoulson_, ok, I got a few cases when it didn't work [20:41] yeah, it works here [20:41] were you copying partial url's? [20:42] oh, even that works here [20:42] no, that works as well [20:42] dunno, I will tell you if that happens again [20:43] chrisccoulson_, oh, I know [20:43] chrisccoulson_, I do type a few words, i.e "oneir", key down in the combo and ctrl-C [20:43] that doesn't work [20:43] like I wanted to copy the etherpad url without opening it [20:43] it's a fail === m_conley is now known as m_conley_away [20:45] ah, ok [20:45] i'm not sure how it would handle that, unless you type the protocol too [20:48] hmm, I "think" I found a bug, unless I just changed name and am now called "[Invalid UTF-8]" ;) [20:48] DBO, somehow i still got this crash, i dont really know how to get into this state -- i looked a bit around and changed this http://paste.debian.net/plain/127413 [20:48] indicator-session doesn't seem to like utf-8 ;) checking if system is up to date [20:49] ricotz, and that makes a difference? [20:49] stgraber, it is fixed [20:50] in 0.3.3.1 [20:50] DBO, cant say since it is hard to reproduce :\ [20:50] ricotz, I really doubt that makes a difference [20:50] kenvandine: ok, still on 0.3.3 here. thanks! [20:50] chrisccoulson, that sucks :p [20:50] chrisccoulson, I'm switching to chrome! [20:51] DBO, the crash is happening for me on freeing priv->path (bamf-view.c:549) [20:51] seb128, hah, chrome hides it too :P [20:51] ricotz looking in a second [20:52] chrisccoulson, ok, I'm switching to ie! [20:52] ;-) [20:52] lol [20:52] wait [20:52] I'm sure epiphany doesn't do it ;-) [20:56] ricotz, somehow priv->path is getting free'd externally... [20:57] though Im sure you figured that out... [20:57] sorry still playing catch up :) [20:57] DBO, probably, so i was thinking gdbus is crashing or messing with it somehow [20:57] seb128, ok... so since rb already has a breaks for the music store... i should add a conflcts to make sure that upgraders definitely get it removed before rb gets updated? [20:58] kenvandine, no, one should be enough [20:59] this was added before ted had the failed upgrade [20:59] ricotz, I guess thats possible, let me know if you think it's fixed [20:59] maybe he had some weird combination or something [20:59] DBO, if that is the case there are other places with dbus related methodes using the priv->... strings [21:00] thats kinda why I dont think this is the case [21:00] you know what this could be [21:00] a screwed up slice allocator [21:03] cyphermox: thanks for the pidgin upload before beta :) [21:05] DBO, would be easy to confirm if i know what triggers it :( [21:05] ricotz, heisenbug [21:05] DBO, might be caused by restarting the windowmanager [21:05] nooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo [21:05] dont do that? [21:05] and other things [21:06] things you shouldn't do: [21:06] anything that makes my code crash [21:06] the end [21:07] this isnt a intentional restart, sometimes gnome-shell wants to crash ;) [21:07] DBO, but this could mess around with bamfdaemon, i guess [21:08] do you only see this with GNOME Shell? [21:08] because it does silly things [21:09] DBO, i think the elementary guys seeing it too [21:09] with what WM? [21:09] so not this strictly related to g-s [21:09] I am more thinking mutter [21:09] i think metacity [21:10] gnome-shell is mutter, there is no separate process anymore [21:11] i mean it isnt a plugin anymore [21:15] DBO, i think another trigger of this crash is the bamf-trigger of dpkg -- bamf index update [21:15] O_o [21:16] but i dont have a trace of it [21:16] it is really a heisenbug [21:22] seb128, want to sponsor a glib upload (for bug 829778)? [21:22] Launchpad bug 829778 in glib2.0 "Breaks everything because of an API break in glib" [Critical,Triaged] https://launchpad.net/bugs/829778 [21:23] kenvandine, ^ [21:23] can you do? [21:23] chrisccoulson, I'm trying to get an unity bug fix upload out before skaet lock the archive :p [21:24] heh, sure, no problem :) [21:24] i've pushed it to bzr anyway [21:25] chrisccoulson, will do later otherwise, you real need to apply for upload rights btw ;-) [21:26] seb128, won't i end up with more work to do if i do that? ;) [21:26] no, less seeking for a sponsor rather ;-) [21:27] seb128, offset by others asking you to sponsor stuff ;-) [21:27] seb128, i can try [21:27] bah, I'm not like that :p [21:57] 'night [22:05] tkamppeter, seb128: Re: the standard profiles - we'd want at least the sRGB and AdobeRGB standard profiles. I think the rest could be dropped from the CD if push came to shove. [22:10] does anyone here know how to get autologin to work after an X crash? i can't seem to make gdm do such a thing [22:22] achiang, you have X crashing so much? [22:23] bryceh: no, but i want to know how to do it. :) [22:23] achiang, ah good :-) [22:24] achiang, afaik there isn't a configuration setting for that; I suspect you'd need to dig into the code for where it detects that X has crashed and add a check for if autologin is set [22:25] bryceh: ah, ok. i can try that. after a bit of googling, it seems kdm has an option to re-autologin after an X crash [22:25] but i couldn't find similar for gdm [22:26] I suspect it's not implemented; it's been a while since I dug through gdm internals though [22:28] * achiang starts digging around === nessita1 is now known as nessita