[00:29] hey guys, anyone interested in this bug? https://bugs.launchpad.net/ubuntu/+source/unity/+bug/801363 [00:29] Ubuntu bug 801363 in unity "Win+W won't show minimized windows" [Undecided,New] [00:30] seems like it is more of a unity problem/opinion === m_conley_away is now known as m_conley === asac_ is now known as asac === m_conley is now known as m_conley_away === m_conley_away is now known as m_conley === m_conley is now known as m_conley_away [06:22] a/c [07:18] Good morning [07:18] Morning pitti. [07:22] good morning pitti and TheMuso [07:22] hey chrisccoulson [07:22] Hey chrisccoulson. [07:42] good morning [07:42] bonjour didrocks [07:43] pitti: guten tag! Have you finally succeeded doing some gardening and bikeshedding despite the weather? [07:43] didrocks: some, it got sunny in the evening; but most of the day was pretty lazy admittedly :/ [07:44] pitti: well, when it's the weather forcing you, you can't be guilty :) and some rest is nice as well! [07:45] Morning didrocks. [07:46] good morning TheMuso [08:11] it's going to be an interesting week at the rally next week [08:14] hey chrisccoulson! [08:14] hi didrocks! === htorque_ is now known as htorque [08:43] hey [08:44] bonjour seb128 [08:44] pitti, guten tag, wie gehts? [08:52] nice, yesterday's iso is down by 10 to 11 mb [08:52] i.e almost in the budget [08:54] seb128: dou you think you could have a look at merging cheese from debian experimental? [08:55] salut seb128 [08:55] bigon, it's somewhat on my list but probably after the rally, it will need a collection of mirs [08:55] indeed [08:55] bigon, not speaking about the gst issue since we can only depends on good [08:55] yeah that's suxxs :/ [08:55] bigon, we have work items on the gnome3 specs to mir those or demote cheese [08:56] but since cheese will be used in empathy and maybe ubiquity we need to discuss what to do [08:56] well libcheese-gtk is now on optional dep for empathy [08:56] right [08:56] lut didrocks [08:57] bigon, well we could mir mx, the new clutter libs, etc but we would still be blocked on the gst issue [08:57] hey rodrigo_ [08:57] morning [08:57] bigon, do you need cheese for anything? [08:58] well for the empathy integration, but I can live without I guess :) [08:58] don't want to put 100 of pkg in telepathy ppa [08:58] argh @ gnome-shell FTBFS; I forgot that I had a locally hacked glib header [09:01] about gnome-shell I think that some changes should be merged from the gnome3-team ppa [09:02] (like using firefox) [09:02] it shouldn't use firefox [09:02] it should use the standalone libmozjs [09:02] seb128: iso size> cjwatson landed the .pyc cleaning, really great [09:02] the goal of that lib was to avoid breaking things on firefox updates ;-) [09:02] pitti, indeed [09:02] pitti, I'm waiting to see what today's build give with deja-jup etc [09:03] RAOF said that he can do some more dynamic linking of the gallium drivers, and some mono libs will still disappear [09:03] and dropping python 2.6 support should give us some extra MBs, too [09:03] seb128: well I mean adding the firefox icon instead of the epiphany one in the launcher [09:03] bigon, oh ok [09:04] bigon, the nice thing now is that we are on sync with Debian ;-) [09:04] :p but well not for long I guess [09:05] depends if somebody is wanting to maintaing gnome-shell in Ubuntu [09:05] there was no sign of that so far [09:06] * bigon leaves the room without making noises [09:07] lol [09:10] seb128: ok with me backporting http://git.gnome.org/browse/clutter/commit/?id=2b81d90dd7399c409cf23d7324ff909351dbefa3 to avoid having to patch gnome-shell? [09:10] or should we go to 1.7.12 right away? [09:10] 1.7.2, sorry [09:11] (it's fixed there) [09:12] pitti, I'm fine either way, if you want to update clutter go for it [09:12] pitti, shouldn't g-s build if you disable the deprecation errors? [09:12] well, "want to" is a bit too much (I don't care at all about it), but do we need to at some point anyway? [09:12] we do I guess [09:12] or should we stay with 1.6. (stable series?) for oneiric? [09:13] seb128: yes, but I'd need to patch the build system for that, and this woudl mean introducing an ubuntu delta [09:13] that's what I want to avoid [09:13] it builds fine in sid, and I'd rather fix the cause than the symptom [09:13] until they update their glib [09:13] pitti, but well pick whatever is easier for you [09:13] right, but it's still not a bug in g-s, it's a bug in clutter [09:13] so it'll hit other packages as well, not just g-s [09:14] pitti, I would just diff clutter 1.6 tarball with -1-6trunk and dump that in a patch [09:14] http://git.gnome.org/browse/clutter/log/?h=clutter-1.6 [09:14] the 4 commits [09:14] seb128: oh, right, they backported that [09:14] pitti, I don't want to decide on the new clutter today, we don't have enough visibility for that [09:14] right, doing that then [09:14] pitti, thanks [09:22] about gnome-shell I think that some changes should be merged from the gnome3-team ppa like using firefox [09:22] huh????! [09:22] firefox doesn't even provide a shared libmozjs anymore, unless you want to load the entire browser (ie, libxul.so), just to get jsapi.... [09:22] feel free to do that, but i *guarantee* that gnome-shell will break with every single firefox upload i do ;) [09:23] 10:03 < bigon> seb128: well I mean adding the firefox icon instead of the epiphany one in the launcher [09:23] (in addition to making gnome-shell take a decade to start) ;) [09:23] chrisccoulson: ^ [09:23] oh [09:23] :p [09:23] i misunderstood there then ;) [09:23] that sounds better [09:24] well I was only at my 1st coffee when I've posted that :p [09:24] heh [09:24] that was not really clear [09:24] i should have read the entire scrollback before commenting ;) [09:25] i saw your comment and had a "WTF?" moment ;) [09:25] :) [09:27] hey chrisccoulson, how are you? [09:27] chrisccoulson, you alternate late night and early morning days? ;-) [09:30] pitti, danke fur die clutter "fix"? ;-) [09:32] seb128: gern geschehen [09:32] g-s builds fine now [09:32] I'll retry the build when clutter is published [09:35] ok [09:45] hi seb128, i'm good thanks [09:45] how are you? [09:47] chrisccoulson, I'm fine thank you [10:00] ok, laptop works after a full upgrade, so ready to go to dublin :) [10:06] rodrigo_, there's still time for us to break it before dublin! [10:06] ;) [10:07] chrisccoulson, I won't upgrade anymore :) [10:07] specially if I see a firefox package :) [10:09] well, there are still some kept pack packages, so will install some of those, so yeah, something can break still :) [10:12] hey rodrigo_ [10:13] so, i wonder if people have started to notice yet that Synaptic hasn't actually dropped from the CD? ;) [10:13] hi again, seb128 :) [10:15] rodrigo_, oh, right, I said hello one hour earlier, it's friday, has been a busy week :p [10:15] chrisccoulson, do you wonder if they started to notice the reason is you and your firefox? ;-) [10:15] lol [10:16] i've saved everybody! \o/ [10:16] seb128, :) [10:16] seb128, don't worry, I like a lot when you tell me 'hi' :D [10:16] ;-) [10:17] rodrigo_, and it's a "hi" coming without bug or anything today! [10:17] just a friendly hello ;-) [10:17] seb128: you recently ran the lucid retracer manually, right? [10:18] pitti, yes, did I screw something? [10:18] seb128: the apport failure in bug 800984 looks like there were no configured apt sources for -updates/-security? [10:18] let's rephrase, I bet I did, "what did I break"? ;-) [10:19] seb128, yeah, those are the ones I like :D [10:19] just weird because usually they get added right from the start [10:19] pitti, that's a natty bug there, not a lucid one? [10:19] seb128: ah, indeed, sorry; misread [10:19] seb128: I'll have a look [10:19] pitti, thanks [10:20] pitti, but yes, I'm pretty sure we usually have -updates and -security [10:21] ok, last reboot hopefully [10:21] brb [10:22] seb128: eww -- it indeed only has maverick-{updates,proposed,security? [10:23] pitti, hum, I probably screwed the sed call in the sources.list, sorry :-( [10:24] I'll check the oneiric one as well while I'm at it [10:24] pitti, oh no, it's not screwed sed [10:24] pitti, it's right for the unstable cycle [10:24] right [10:24] pitti, we just didn't change it where natty-updates opened [10:24] when [10:25] we could keep both -1 and current sources for those in the sources.list to avoid those issues [10:25] I guess they would just error out during the unstable cycle or be empty [10:25] but that should break anything [10:26] "shouldnt" [10:27] yes, I'll add both [10:28] pitti, thanks [10:29] * didrocks cries at every bzr bd-do one Qt [10:29] on* [10:30] weirdly, if I just dpkg-source -x, or untar the tarball, it's very fast… [10:49] apropos very fast: the ubuntu wiki is not. it takes ~2 minutes to get a freaking login. [11:01] pitti - oh, you've done a new set of language packs for maverick? [11:02] ah yes [11:02] i need to update my firefox-stable PPA then ;) [11:04] hi smspillaz, would like to chat about some compiz default settings when you have a sec [11:18] chrisccoulson: dpm did, but yes [11:18] chrisccoulson: what did you change there? drop xpis and add the -locale dependencies? [11:19] pitti - yeah, for https://launchpad.net/~mozillateam/+archive/firefox-stable i repack them without the xpi's and add the recommends [11:26] oh, i'd best get them in whilst the PPA build queue is empty :) [11:33] Does LightDM support automatic logins? Or is that just broken in O? [11:35] njpatel, it does [11:35] njpatel, but you need to edit /etc/lightdm/lightdm.conf [11:35] seb128, aah [11:35] # default-user = Username of default user to log in as [11:35] # default-user-timeout = Delay before logging in in seconds (use 0 for automatic login) [11:35] njpatel, ^ those keys [11:37] hmm, where can I see the status of the uploaded packages for oneiric? [11:37] once accepted I mean [11:37] seb128: oh, we should probably fix that in casper, to work on the live system? [11:37] rodrigo_, gnome-desktop3 is in binary New due to the soname change if that's the question [11:37] seb128, also, do you know how to stop all the default locale warnings in the terminal? [11:37] pitti, done previous week [11:37] ah, great [11:38] * pitti rsyncs the current ones [11:38] njpatel, no, don't get that here [11:38] seb128, yes, that was the question [11:38] seb128, is it just for en_GB? ;) [11:38] pitti, [11:38] casper (1.269) oneiric; urgency=low [11:38] [ Colin Watson ] [11:38] * Add LightDM autologin support (LP: #797669). [11:38] seb128: thanks [11:38] seb128, so,where do you check that? [11:38] rodrigo_, launchpad, where else? ;-) [11:38] rodrigo_, https://launchpad.net/ubuntu/+source/gnome-desktop3/3.1.2-0ubuntu1 [11:38] it says "New" [11:39] rodrigo_, https://launchpad.net/ubuntu/oneiric/+queue is the queue [11:39] seb128, ok, was asking for a page showing the status of all source packages [11:40] rodrigo_, I'm not sure to understand the question, you want a list of 15000 sources with what info onit? [11:41] seb128, no, just a way to know what's been NEW'ed, etc, but I guess checking the package source I'm interested in is enough :) [11:41] rodrigo_, usually just check the queue, i.e the second URL I gave you [11:41] or the launchpad infos for the version you uploaded and are interested in [11:42] rodrigo_, do you need it to be Newed? [11:42] no, just that it didn't show up on the repos, so wanted a way to check why [11:42] but I'll build it locally for my g-s-d build [11:43] rodrigo_, you didn't test build it locally before uploading? ;-) [11:43] rodrigo_, I can new it but I was avoiding to throw a soname change in on friday before traveling [11:43] rodrigo_, I figured it could wait on monday [11:43] seb128, yes, I did, but on my desktop machine, not on the laptop, where I'm now :) [11:43] seb128, oh yes, it can wait [11:44] I won't upload g-s-d until a lot of testing, so it can wait [11:44] rodrigo_, well you can get the binaries from launchpad, the first URL I gave you [11:44] click on the architecture you want [11:44] it has links to the binaries [11:44] no need, just built it here :) [11:45] my question was just to check why it wasn't in the repos [11:45] so never mind :) [11:46] rodrigo_, new binaries go through New and need to be reviewed or approved so you can usually assume that if things are not in they either failed to build or are waiting in New ;-) [11:46] ok [11:46] https://launchpad.net/ubuntu/+source// has details about the status and is the place to check in any case [11:47] I've a firefox shortcut to "https://launchpad.net/ubuntu/+source/" [11:47] lpsource [11:47] that's handy ;-) [11:47] I've similar ones for bug lists etc [11:47] yeah, adding it [11:47] on bugzilla as well [11:52] hum, lunch time [11:52] bbl [11:53] seb128: indeed, autologin works fine now in kvm [11:53] I just don't get any default launchers in unity-2d, hmm [11:53] seb128: lpp (LaunchPad Package) here, next to lpb, bg… and "latest crack" (directing to daily live CD build) :) [11:53] pitti: oh? [11:54] didrocks: just try it [11:54] kvm -m 1024 -vga vmware -cdrom oneiric-desktop-amd64.iso [11:54] pitti: possible under some condition, let me check something [11:54] yeah, with fallbacking [11:54] I do get default launchers for a fresh user in the installed system (in -2d), though [11:54] righht, that's the auto-fallback [11:54] in my local testing that's with explicit choice in gdm [11:54] right, but you are choosing the session for the fresh user? [11:55] right [11:55] just one sec, confirming :) [11:55] pitti: so right, it's only applied on the selected session [11:55] pitti, didrocks empty launcher bug 797672. could you please confirm. [11:55] Launchpad bug 797672 in unity-2d "No default apps on launcher on a live session." [Undecided,New] https://launchpad.net/bugs/797672 [11:55] pitti: still the same issue with gconf path [11:55] jibel: cheers [11:56] pitti: the correct solution is to choose the right session before login, so implementing the detection at the dm level [11:56] which I didn't want to do with gdm, but I have a WI for that for lightdm [11:56] didrocks: OOI, what does that have to do with the default launchers? shouldn't they be in gconf or gsettings schemas? [11:56] it's one of my goal for the sprint, but I need robert next to me :) [11:57] heh, then doing this as a pair indeed makes sense next week [11:57] pitti: yeah, they can be in the common gconf set [11:57] but that doesn't fix everything [11:57] like the gconf option to tell metacity "remove decoration from maximized window" [11:58] oh, the session and gconf defaults are related through the default XDG path setting? [11:58] you only want that in unity-2d and the guys implemented that as a gconf key, not a dbus call to detect if unity-2d is running [11:58] pitti: there is a dedicated GCONF path [11:58] ok, now I underfstand [11:58] DEFAULTS_PATH [11:58] and MANDATORY_PATH [11:58] and as we are running gconf before gnome-session… [11:59] so yeah, all that detection should really be at the dm level to avoid that :) [11:59] didrocks, bug 797676 might be related too. [11:59] Launchpad bug 797676 in unity-2d "Launcher doesn't autohide" [High,Confirmed] https://launchpad.net/bugs/797676 [12:00] jibel: hum, I don't think so as it should be the default, with or without gconf key [12:00] jibel: so, if it's not, the code should, in case of no keys present, have that behavior === MacSlow is now known as MacSlow|lunch [12:33] pitti: does jockey have a function to get the current debian arch? Or shall I create one in both fglrx and nvidia's handlers (because of multi-arch)? [12:57] tseliot: not right now; but it's easy to find out with dpkg --print-architecture [12:58] tseliot: I think create it in the handlers for now [13:28] pitti: maybe I can add it in nvidia-common (as the handlers use it already) [13:35] pitti: I have a nifty little LO 3.3.3 release for natty here (actually there: https://launchpad.net/~bjoern-michaelsen/+archive/libreoffice-nattytest2/+sourcepub/1789243/+listing-archive-extra ) [13:40] hey Sweetshark [13:40] Sweetshark: oh, nice! [13:40] Sweetshark: some changes look a bit intrusive for an SRU, though? [13:41] pitti: I did some more changes (translations and icons updates) as 1:3.3.3-1ubuntu1~ppa2 (not uploaded yet). How should I upload that as ~ppa2 to the LO packaging archive or directly as 3.3.3-1ubuntu1, because we will SRU that (will we?) ... [13:42] Sweetshark: ah, it doesn't actually change multiarch bits? [13:42] the Debian changelog makes it sound like it, but your merge changelog says it's conditional? [13:43] the most intrustive changes are the ~30 critical upstream fixes between 3.3.2 and 3.3.3 ... [13:43] these seem fine [13:43] I was just concerned about introducing multiarch in an SRU for the libs [13:43] if the packages don't actually change that, it seems fine [13:43] lemme check a diff of debian/rules [13:44] Sweetshark: for the 30 upstream fixes I think we do want it as an SRU; WDYT? [13:47] pitti: yes, I have not investigated them all that close, but if somebody ports a patch back to 3.3 after 3.3.2, it is likely worth it. (and all those are under code review upstream) [13:50] pitti: is disabling (broken) sdbc-postgresql an issue? It is basically fubar in 3.3, so disabling it is just honest. [13:50] Sweetshark: yes; we can't remove a package from natty [13:50] Sweetshark: that's libreoffice-sdbc-postgresql, right? [13:51] Sweetshark: so removing it from debian/control is a no-go for natty, I'm afraid [13:51] Sweetshark: I'd just not touch it at all if it's broken [13:51] Sweetshark: for oneiric, you can make it an empty transitional package [13:51] (if it won't be fixed any more, ever) [13:52] Sweetshark: if you drop it from debian/control, we'd just introduce an archive inconsistency (NBS), and it wouldn't make the natty final version disappear [13:53] oh, I hope it will be fixed sometime, it is not broken beyond hope. [13:54] pitti: I can explicitly reenable it for natty, so we still explicitly build it on natty (even if broken). [13:56] Sweetshark: sounds better for the SRU [13:59] the multiarch changes seem to be not be about LOs install, but about finding multiarched deps, so that should be ok too. [14:01] hate the retracers [14:01] " from launchpadlib.errors import HTTPError" [14:01] who does that keep happening? fixing it... [14:02] oh great, rene silently disabled parallel builds with more than 2 jobs (likely because of https://bugs.launchpad.net/ubuntu/+source/make-dfsg/+bug/759704 ) [14:02] Ubuntu bug 759704 in make-dfsg "make crashed with SIGSEGV" [Undecided,Confirmed] [14:04] seb128: meh :/ [14:05] although it can still ftbfs with 2 jobs -- its just unlikely. [14:06] cyphermox, good morning [14:06] cyphermox, working on evolution too? === m_conley_away is now known as m_conley [14:14] pitti: btw for 3.4.X bug 759704 will be a sure build breaker for parallel builds -- does that qualify for importance 'high'? [14:14] Launchpad bug 759704 in make-dfsg "make crashed with SIGSEGV" [Undecided,Confirmed] https://launchpad.net/bugs/759704 [14:15] Sweetshark: oh, sure; but we actually need someone to work on this/a fix, too :) [14:17] pitti: well, there is a complete patch by me in the bug and actually a patched release in the LO ppa. [14:17] oh, did you subscribe sponsors? [14:18] ah, seems fixed upstream, nice [14:19] Sweetshark: (terminology nitpick: you attached a patch, not a debdiff -- there's no debian/changelog, debian/patches/* etc handling) [14:19] Sweetshark: want me to sponsor that for you? [14:19] pitti: I just did not to splash into somebody elses package (esp. since you can break a lot of stuff with tweaking GNU make). [14:21] pitti: please do. And there are changelog entries etc. in the ppa upload: https://launchpad.net/~libreoffice/+archive/ppa/+sourcepub/1733537/+listing-archive-extra === MacSlow|lunch is now known as MacSlow [14:24] pitti: I had that bug filed upstream too at debian http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=622644 including debdiff (no changelog changes there, as I did not want to mess with debians policies), but there nobody seems to really care. [14:24] Debian bug 622644 in make "parallel (-j2) make with $(eval) construct segfaults" [Normal,Open] [14:24] (my hope was it would have been tickled donw back to us by now ;) [14:27] Sweetshark: hmm, it has [14:27] https://launchpad.net/ubuntu/+source/make-dfsg/3.81-8.1ubuntu1 [14:27] I'm not sure why the bug wasn't auto-closed [14:27] Sweetshark: closed it now [14:29] * Sweetshark might not have noticed because of still sticking with natty quite often. [14:30] I need more tasty dogfooding. [14:31] heh [14:32] what that weird (debian) versioning (8.1) there, btw? [14:33] Sweetshark: -8.1 is an NMU in Debian, a Non-Maintainer Uload [14:33] Upload [14:33] it's not weird [14:33] kenvandine: gwibber on natty seems to of stopped syncing twitter since I signed up to twitter on oneiric last night :( Anything massively changed? [14:38] pitti: weird in the "havent seen that before" sense ;) [14:41] davmor2, not in natty... [14:41] davmor2, in a terminal, killall -9 gwibber-service && gwibber-service -d [14:41] let it sync, then post the log for me [14:41] ~/.cache/gwibber/gwibber.log === Guest49686 is now known as WelshDragon [14:47] ok, g-s-d 3.1.2 seems to work ok, so pushed to the ubuntu-desktop branch, if someone wants to test it [14:48] rodrigo_, does it conflicts with gpm? [14:49] seb128, no, gpm code is still in patch review, not yet in gsd code [14:49] but it needs to conflict with gcm, checking now with hughsie [14:49] ok [14:49] is gcm going to stay around? [14:50] does it have extra feature or is that merged in gcc? [14:50] all should be merged, checking [15:02] rodrigo_, seems fine from a review, I didn't try it [15:02] rodrigo_, but let's do the 3.1 landings next week as a team goal for the rally [15:03] seb128, running it for a few days myself before uploading [15:03] so yes, won't upload it for now [15:03] kenvandine: evo all done, [15:03] I had issues pushing last night though :/ [15:06] cyphermox, cool... point me to it and i'll take a look [15:06] i installed eds last night, which of course removed evo :) [15:07] yeah [15:07] you need gtkhtml 4.1.2 as well [15:07] evo: bzr branch lp:~mathieu-tl/evolution/3.1.2-0ubuntu1 [15:07] gtkhtml: lp:~mathieu-tl/gtkhtml/4.1.2-0ubuntu1 [15:10] cyphermox, ok, thx [15:11] time for a break and some errands, I'm away for ~2 hours, will read backlog when I'm back [15:11] see you later [15:11] later seb128 [15:11] cya, seb128 [15:11] or see you on sunday for those who called it a week before I'm back ;-) [15:12] bye seb128 [15:14] see you seb128 [15:19] doh [15:19] hrmm [15:19] bryceh: is https://blueprints.launchpad.net/ubuntu/+spec/desktop-maverick-rootless-x something that you actually want to pursue for oneiric, or should we just move it to the Q cycle (post-LTS)? [15:20] is there an equivalent to foo.gconf-defaults files for gsettings in packages in oneiric? [15:22] dobey: yes, e. g. /usr/share/glib-2.0/schemas/foopkg.gschema.override [15:22] pitti: so we install one of those files? cool [15:23] dobey: we have some, yes [15:23] great, thanks [15:25] any ideas on how i can get a stack trace (i.e. from gdb) of a crash in a gnome-panel applet, like the clock? [15:25] the complication of course being that it's started by the panel and crashes immediately, so no time to attach a gdb to it. [15:27] Shred00: does it keep crashing when you re-add it? [15:27] yup [15:28] Shred00: enable apport for one time, let it catch the crash, and then use apport-retrace -g /var/crash/gnome-panel.crash ? [15:28] Shred00: when it crashes the first time, run it inside gdb, then re-add it while it's running in gdb [15:28] well, i don't need to "re-add" it. i get an error dialog asking if i want to "reload" it [15:28] sudo service apport start force_start=1 [15:28] or what dobey says [15:28] for live debugging [15:28] Shred00: right, while that dialog is up, start up gdb with it, and when it's started running in gdb, hit "reload" [15:30] where's the hotel address for the dublin rally? [15:30] rodrigo_: https://wiki.canonical.com/UbuntuPlatform/Rally/Oneiric [15:31] i am both sad and glad i am not going to dublin :) [15:31] pitti, thanks! [15:31] dobey: i must not be understanding: [15:31] $ gdb /usr/lib/gnome-panel/clock-applet [15:31] ... [15:31] (gdb) [15:31] then i hit Reload in the dialog and the dialog disappears briefly as it starts a new one which immediately dies. [15:31] Shred00: you need to run the applet inside gdb [15:31] Shred00: not simply load it, it must be running [15:32] pitti, hmm, that page is empty [15:32] ahhhh. got it. [15:32] Shred00: iow, type "run" and press RET :) [15:32] rodrigo_: sorry, recent mass-renaming; https://wiki.canonical.com/UbuntuEngineering/Rally/Oneiric [15:32] damn. pacrunner_webkit again [15:33] i thought this got fixed in maveric [15:33] wtf is the clock applet using webkit for?! [15:33] loading weather info [15:33] let me rephrase [15:33] wtf is the clock applet using webkit for?! [15:33] :) [15:34] the clock loads weather info from the web. but to do so, it might need to use a proxy and to resolve the proxy configuration it might need to process a proxy "pac" specification, which is javascript, so the webkit JS engine is invoked. [15:34] hrmm, i guess it's probably not actually using it, but soup is using libproxy which has a backend that uses the js engine [15:34] dobey: you got it [15:34] this problem existed in maverick also, but was fixed, iirc. it seems to have leaked back into natty. [15:35] or maybe i just fixed it here myself. [15:36] which is not webkit, but still; https://twitter.com/#!/dohbee/status/81015987595513857 [15:37] because proxies provide filtering and redirection (and content substitution) abilities that simply changing an IP address cannot. [15:37] and the ability to authenticate [15:39] a proper network setup doesn't need proxies to do that. [15:39] anyway [15:42] pitti: hrmm, it's been a long time since i've had to install a file from debian/ to the system. and with multiple binary packages. do i put the "debian/foo /usr/share/foo/" bit inside foo.install, or in a generic "install" file, and just include it with /usr/share/foo/foo in the foo.install? [15:43] dobey: proxies also cache content for reduced bandwidth usage and lower latency for the user. [15:44] you can have a cache without an actual proxy thing without the annoyingly crap "proxy" part [15:44] err, that sentence was horrible [15:45] you can have a cache without the actual annoyingly crap "proxy" part [15:45] that's better :) [15:45] dobey: are you referring to a transparent proxy? [15:48] i'm referring to properly configuring a network so that users don't have to configure anything on their computers in every other application, to use the internet [15:50] dobey: i'm all ears (or eyes if you have something to point to, or fingers if you have something for me to google for) [15:50] dobey: debia/nfoo /usr/share/foo/ in foo.install is fine [15:50] dobey: I strongly recommend not to use debian/install; it'll fall on your feet if someone adds a second binary package, or changes the order in debian/control [15:50] Shred00: you're using a proxy on your *personal* network? [15:51] pitti: ok, yeah. we have like 8 binary packages already :) [15:51] cyphermox, looking at gtkhtml [15:51] dobey: our home network, yes, but could just as easily be a corporate network. [15:51] they didn't bump the api version, so it still builds a 4.0 .so [15:52] Shred00: well, there's a difference between "i mad a conscious choice to do this" and "the IT dept. at my job requires me to do this" [15:52] i am thinking your shlibs change should be libgtkhtml (>= 4.0) [15:52] they must not have removed any symbols, so we shouldn't break deps for things that might dep on it [15:53] dobey: sure. i am the IT dept. but you say there are better alternatives than a proxy server to provide (a) caching, (b) authentication and (c) content restriction, so i am asking what is it/are they? [15:54] kenvandine: ok [15:55] dobey, going to have the u1/sso fix uploaded soon? [15:55] kenvandine: yes, lots of fixes coming today [15:55] dobey, good... i am getting harassed constantly about u1 being busted :/ [15:55] kenvandine: do you know why indicator-*-gtk2 is on the CD images still? [15:55] and i kind of want to let my laptop sync before i travel [15:56] kenvandine: me too :) [15:56] pitti, yes... because we still need them [15:56] oh, indicator-session recommends: i-s-gtk2 [15:56] why's that? [15:56] it is kind of a hack [15:56] kenvandine: nightlies will get you synced [15:56] until unity can load the gtk3 indicators [15:56] kenvandine: oh, I thought unity would be gtk3 [15:56] not yet [15:56] almost :) [15:56] that is a goal for the rally [15:57] kenvandine: oh, it depends on both 2 and 3 [15:57] kenvandine: and you should be running nightlies of u1 :) [15:57] pitti: some part are [15:57] kenvandine: ah, so the recommends will drop once unity is fully ported over? [15:57] pitti, yes [15:57] ah, great; thanks for the heads-up! [15:57] np :) [15:57] pitti: njpatel didn't want to push unity panel in gtk3 until all indicators were ready [15:57] pitti: which wasn't the case for last release [15:57] it was better than changing the recommends in unity everytime i uploaded a ported indicatore [15:57] yeah, understood [15:58] i think indicator-sound is the only one left [15:58] kenvandine: si signore! [15:58] which ronoc just got merged [15:58] we'll upload that on monday [15:58] Shred00: i haven't gone about setting up such a configuration yet, so i don't have exct details. i have lots of bandwidth, so don't really need to. though once i set up vpn i may set up some stuff [16:05] dobey: i don't mean to be a dick, but you are hardly speaking from a point of experience in telling me to set up my network "properly". [16:31] mpt: having fun with math? :) [16:31] didrocks, my brain is melting [16:32] mpt: heh, congrats on your research on that part btw. It's nice to see a "sort by relevancy" dump and that you really explored the problem ;) [16:32] At least when we get to the "Recommended for you" algorithm, I know I will be *completely* out of my league and won't be tempted to figure it out myself [16:32] I'm quite scared about this algorithm TBH, a lot of research will be needed :) [16:33] let's try to put that as much as possible server side for easier update and refinement [16:34] Oh, yes, recommendations will definitely be server-side [16:35] If only because the amount of data representing everything that everyone (who has opted in) has installed will be both (a) semi-confidential and (b) large :-) [16:36] right, and (b) will force a lot of caching :) [16:42] kenvandine: shall I move the entire https://blueprints.launchpad.net/ubuntu/+spec/packageselection-desktop-n-geoclue to the p or q cycle then, and reopen the WIs? [16:42] kenvandine: or is that really meant to mean "DROP"? [16:43] i think drop [16:43] i think, at least for now, we have done all we are doing [16:43] ok [16:43] [magicfab] community participation to improve osm mapping: [16:43] is that still useful then? [16:59] pitti, that was more of a way to help people improve osm [16:59] not sure it is really relevant to the desktop [17:01] good night everyone! see you on Sunday evening! [17:01] good night pitti [17:01] see you soon :) [17:02] see you on Sunday pitti! [17:31] Hello. I have a potential bug and I would like to confirm that it should be filed against nautilus. It's miscalculating that /home is full because there are 2 samba shares mounted in /home through fstab [17:32] the /home partition is ~ 600GB and the 2 shares are totaling ~ 2TB. I get the messahe that the /home is full [17:32] if I unmount the drives then there's no message [17:35] I'm not sure against what I should file the bug: nautilus - linux - samba ? [17:48] cyphermox, evo and friends all uploaded [17:56] pitti, I do not have plans around it; RAOF might. Move it to Q [17:57] eod'ing, so see you all in Dublin on Monday (there were no flights on Sunday so arriving there monday morning) [17:59] amorphous, I'd go with nautilus [17:59] (it can always be moved) [18:08] kenvandine: thanks [18:09] mterry, also...I've forgot to mention: df says /home is 100% when the shares are mounted. So I guess nautilus reports what the kernel says. I'll file the bug against linux [18:09] amorphous, k [18:09] mterry, also when the drives are not mounted, df shows the reals %. [18:09] mterry, i just filed bug 801658 [18:09] Launchpad bug 801658 in duplicity "missing depends for ubuntuone-couch" [Undecided,New] https://launchpad.net/bugs/801658 [18:09] easy fix :) [18:11] kenvandine, that's not quite a depends on duplicity; better on deja-dup. But I thought it had a Recommends for it! [18:11] kenvandine, yeah, it does [18:11] ah,ok [18:11] weird [18:11] i just got dejadup installed today in a dist-upgrade [18:12] but no ubuntuone-couch [18:12] kenvandine, and it didn't pull in ubuntuone-couch? hmm.. maybe recommends are not pulled in for a dist-upgrade? That sounds familiar [18:13] maybe... that does sound familiar :/ [18:15] mterry, just confirmed in a VM... dist-upgrade brought in deja-dup without ubuntuone-couch [18:15] bummer [18:16] kenvandine, but a direct apt-get install does bring it in? [18:16] it would [18:16] weird [18:16] * kenvandine didn't test that :)( [18:16] I assume that's intentional behavior... [18:17] yeah, confirmed [18:17] a pane though... i guess you shouldn't display U1 as a target if the deps aren't there [18:18] * kenvandine goes to get some lunch, bbiab [18:18] dist-upgrade should, but it's a known issue it doesn't everytime [18:18] use --fix-policy regularly :) [18:26] kenvandine, we test whether 'ubuntuone' is importable, but not each piece of the ubuntuone puzzle. probably should [18:27] chrisccoulson, firefox and thunderbird are crashy for me right now. slow down my machine, then crash [18:29] mterry, it seems to be ok here, and i don't think anything has changed recently :/ [18:29] kenvandine, if you have a box showing the "don't install recommends" please keep it this way [18:29] got a trace? [18:29] kenvandine, so we can get mvo to debug it next week [18:32] chrisccoulson, hrm, not just those two apps now. I'm apt to blame compiz or unity [18:32] heh [18:32] seb128, so, i'm getting somewhere the indicator issue [18:32] with? [18:33] mterry, turn a11y off :p [18:33] :) [18:34] seb128, bug 799878 (and the empty messaging indicator) - the indicator is dropping all of the incoming ItemsPropertiesUpdated signals because they have the wrong signature :/ [18:34] Launchpad bug 799878 in dbusmenu/0.5 "Test suite 'glib-test-proxy' failes on GLib 2.29" [Critical,Confirmed] https://launchpad.net/bugs/799878 [18:34] but i'm not sure what changed there [18:34] chrisccoulson, check with dest or davidz on #gnome-hackers maybe? [18:34] chrisccoulson, or wait next week to check with him at the rally [18:35] seb128, is desrt going to be there? [18:35] I think so [18:35] cool, i can wait until next week :) [18:35] m_conley ^^ ;) [18:39] he will be there, right [18:40] hum, evolution is crashy there as well now [18:40] and qt creator! [18:48] chrisccoulson: hey - nice job narrowing that down! :) [18:48] m_conley, i still need to figure out exactly what is going on though ;) [19:33] just as a heads-up, we (chrome) are removing global bookmarks support [19:33] http://codereview.chromium.org/7252001/ [19:33] i'll file appropriate bugs [19:38] https://bugs.launchpad.net/dbusmenu/+bug/801699 [19:38] Ubuntu bug 801699 in dbusmenu "DBus menu is very slow when using large menu" [Undecided,New] [20:06] evmar, thanks. how big do the bookmarks menus need to be btw? i didn't hit any issues like this with firefox [20:06] but then, my bookmarks menu isn't really that big [20:14] oh, welcome tedg \o/ [20:14] did you see bug 801699? [20:14] Launchpad bug 801699 in dbusmenu "DBus menu is very slow when using large menu" [Undecided,New] https://launchpad.net/bugs/801699 [20:17] chrisccoulson: perhaps we're hitting some pathological behavior in the api? i see g_list_foreach calls in the call stack, maybe some algo is unintentionally O(n^2) [20:17] chrisccoulson, No, I didn't. Interesting. [20:17] chrisccoulson: do you have any pointer to how firefox integration works? (was it implemented by you or by ffox upstream?) [20:18] It is odd. As it should be only one foreach per level in the hierarchy. [20:18] evmar, Does the chrome bookmark menu have something like tag menus or something? [20:19] i don't think so [20:19] Hmm. [20:20] evmar, If you have the large menu running on a process could you run /usr/lib/libdbusmenu/dbusmenu-dumper on it and attach the JSON to the bug? [20:20] evmar, I can then put it as part of the dbusmenu test suite. [20:20] evmar, the firefox integration doesn't use GtkMenu (and thus bypasses appmenu-gtk and the GtkMenu parser), although i don't think that would make any difference here [20:20] evmar, it's sending favicons isn't it? [20:21] chrisccoulson, It wouldn't get the icons on the first pass, and that looks like where it's passing the structure. [20:23] mterry, kenvandine: can i get some sponsorship love for ubuntu-sso-client? :) https://code.launchpad.net/~dobey/ubuntu/oneiric/ubuntu-sso-client/release-1-3-1/+merge/65827 [20:26] chrisccoulson: yeah, favicons [20:26] chrisccoulson: back of the envelope calculation says we can be sending 100kb+ over dbus, not sure if that works [20:29] dobey, I'll look at it [20:30] evmar, We ask for the structure with labels in one pass, and then ask for the icons in groups of 100 to limit the impact. [20:31] evmar, So it shouldn't be 100kb at once. [20:31] evmar, tbh i'm not too sure what the physical limits of dbus are. i had a problem with icon data in thunderbird before though (although admittedly, it was because of an extension sending a 512x400 icon in a menu. but that resulted in thunderbird just being aborted after a couple of minutes of sending so much data that everything just gave up - bug 749440) [20:31] Launchpad bug 749440 in globalmenu-extension "Incompatibility with Thunderbird addon CuteButtons - Crystal SVG : makes TB crash when reply a mail" [Medium,Fix released] https://launchpad.net/bugs/749440 [20:32] chrisccoulson, The limits are set in /etc/dbus-1/session.conf [20:33] chrisccoulson, Though, they say it should be 1G, so it seems unlikely you hit those. [20:33] tedg, i guess i mean "limits at where things begin to suck", rather than actual physical limits [20:34] chrisccoulson, Heh, the "suck limits". We should coin that phrase. [20:34] :-) [20:34] "This is currently operating at 90% suck, so we need to start thinking about optimizing it" [20:43] see you on Sunday guys! [20:44] 'night didrocks [20:44] didrocks, have a good weekend [20:44] see ya' soon :) [20:44] tedg: you should try python on plane! :p [20:44] SNAKES ON A PLANE! [20:44] thanks, you too tedg, kenvandine! [20:44] ;-) [20:44] heh [20:48] dobey, uploaded [20:48] mterry: great, thanks! [20:52] mterry, dobey awesome... thx! [21:00] i love breaking ubuntu on a friday afternoon :) [21:00] * kenvandine hopes this dbusmenu upload doesn't cause too much havoc [21:01] kenvandine, if you didn't break things, what would you do during the rally? [21:01] kenvandine, esp. a friday before international travel :-) [21:01] drink beer? [21:01] bryceh, exactly :) [21:01] mterry, this upload is tricky too... i breaks apps that use libindicate and libunity via GIR [21:02] speaking of breaking Ubuntu on a friday afternoon before travelling, I need to upload an update to NM :D [21:02] so i added a breaks to make it hold back [21:02] cyphermox, ugh ;) [21:02] none of us will be able to connect to wifi when traveling :) [21:02] pfft, you should know how to use wpasupplicant by itself ;) [21:03] * kenvandine hides [21:03] * cyphermox says that because he still struggles with that black magic... though less than before [21:03] Hi all! [21:03] although i remember when dbusmenu made nm-applet toggle my interfaces over and over again [21:03] kenvandine: right, that was fun [21:03] i was not happy about that... while tedg was on some beach in hawaii [21:04] Is this the right place to inform about gdm 3 issues? [21:04] heh [21:04] ajvdvegt: bugs.launchpad.net/ubuntu/+source/gdm probably is the "right place" :) [21:04] ajvdvegt: depends, what kind of issue? also, bugs are great to inform developers of issues ;) [21:05] :) [21:05] consistent segfaults count as an issue, I suppose. [21:05] yup, the best kind [21:06] But as GDM suddenly stopped working for me, I suppse I'm not the only one seeing this. [21:06] kenvandine: speaking of nm-applet, do you see it in unity ? [21:06] I'll go file a bug report, tnx. [21:07] ajvdvegt: thanks, that will be useful especially if it comes from apport (e.g. with a backtrace and everything) [21:07] cyphermox, i do [21:07] kenvandine: hrm... so why am I missing it, the power button indicator thing, and the messaging menu? [21:08] is this related to the dbusmenu upload you're doing? [21:08] cyphermox, can't be... [21:08] but [21:08] do you have indicator-application-gtk2 installed? [21:08] and indicator-session-gtk2? [21:08] etc [21:09] ah, no. uninstalled, apparently. [21:09] ah [21:09] i wonder if this is another case of recommends being ignored on a dist-upgrade? [21:09] maybe. is that recommended in ubuntu-desktop? [21:10] no [21:10] indicator-session recommends indicator-session-gtk2 [21:10] ok [21:10] just until unity panel can load gtk3 indicators [21:10] but other people noticed them getting installed [21:10] i saw a lot of stuff get uninstalled [21:10] ok, I activated apport. I'm off crashing gdm now, thanks! [21:10] ajvdvegt, have fun :) [21:11] probably related to gnome-panel/libpanelapplet getting removed? [21:22] oh, dbusmenu broke ABI? [21:29] chrisccoulson, Yeah, we dropped an object no one was using. [21:29] chrisccoulson, Think of it less as "broke" and more as "refined" [21:33] s/no one/no one we care about/ :) [21:58] dobey, I'd be impressed if someone was using it. It didn't really do anything... I guess not as much impressed as confused :-) === m_conley is now known as m_conley_away [22:23] AND ubuntuone-client 1.7.0 uploaded to ubuntu [22:24] i hope the switch to dh_python2 didn't blow stuff up :)