[00:02] robert_ancell, btw I did add small changes to version to let specify a different regexp or url for some sources [00:02] seb128, sweet [00:02] robert_ancell, and I did add a lucid table [00:02] so we can clean a bit the list [00:02] seb128, where is the offical url now? I notice pittis one no longer works [00:03] cf topic? :-) [00:03] ha! [00:04] hehe [00:04] anyway enough work for today [00:04] good night everybody! [00:05] night! [00:51] vish - i understand what is going on with this gpm blanking issue now [01:22] hey, from python how do you detect if a program is available in the path (i.e. equivalent of `which dkms`) [01:27] bryceh: I believe there is no standard library function [01:28] ok, I'll just check for path existence [01:28] bryceh: you can subprocess.Popen it and catch OSError and check for errno.ENOENT, which is likely to be that the command isn't available [01:29] I think checking existence of /var/lib/dkms is probably cleaner [01:38] vish - i fixed the blanking bug :) === kermiac is now known as kermiac_ [03:43] chrisccoulson, I can confirm 515155 [03:43] bug 515155 [03:43] Launchpad bug 515155 in gnome-screensaver "unlocking screensaver hangs when ecryptfs is unmounted" [Low,Incomplete] https://launchpad.net/bugs/515155 [03:44] kklimonda, thanks [03:45] chrisccoulson, do you want stacktrace? [03:45] it doesn't crash but I've connected to the process itself [03:45] i'm not going to look at that tonight though. i've spent all evening debugging a gpm race and i'm going to bed now :) [03:45] sure [03:45] yeah, a backtrace would be useful [03:45] i suspect that it's not really a gnome-screensaver issue [03:46] anyway, i can look at that tomorrow sometime [03:46] if so I'll have to install more debug symbols :) [03:47] heh ;) [03:47] anyway, me -> sleep [03:47] 'night! [03:47] good night === JamieBen1ett is now known as JamieBennett === kermiac_ is now known as kermiac [06:40] Good morning [06:50] didrocks: bonjour === kermiac is now known as kermiac_ [07:13] good morning pitti [07:13] http://people.canonical.com/~pitti/bootcharts/daniel-lucid-20100212-1.png === kenvandine_ is now known as kenvandine [07:14] pitti, nice! [07:18] * kenvandine just uploaded (to my gnomex ppa) the first pass at merging the f-spot patch adding editing to view mode [07:24] pitti: morning! one question: if I want to update the humanity-icon-theme which package should I start from? In the meantime it is quite unclear to me what is in lucid and where that is in LP [07:25] kenvandine: ooh! [07:25] kenvandine: as a backport, or trunk snapshot? [07:25] backport [07:25] pitti: morning! one question: if I want to update the humanity-icon-theme which package should I start from? In the meantime it is quite unclear to me what is in lucid and where that is in LP [07:26] kwwii: Vcs-Bzr: says lp:~ubuntu-art-pkg/humanity/release [07:26] kwwii: that ought to be the authoritative place [07:26] there are a couple of crashers, but shouldn't be hard to work out [07:26] and we should make edit mode more discoverable [07:26] details :) [07:26] kwwii: lucid has 0.4.1ubuntu5, which is in that bzr as well, so it seems fine [07:27] even better: http://people.canonical.com/~pitti/bootcharts/daniel-lucid-20100212-1-no-rsyslog-dd.png [07:27] pitti: ok, cool...things are better than I thought. upstream as two repo's for it and someone has a ppa with new packages [07:27] (dropping rsyslog's dd process, which is on the plan for lucid) [07:27] pitti, that is pretty damn impressive [07:27] pitti: thanks for the info, I'll merge things, review them and then get a package ready for upload [07:28] kenvandine: 11.5 s is quite good already \o/ [07:28] kenvandine: and we have didrocks' bg caching in the pipe still [07:29] * kenvandine drools [07:30] * kenvandine should go get a little sleep before the little ones wake me :) [07:30] good night folks! [07:31] kenvandine: sleep well! [07:31] kwwii: thanks [07:35] I need to run out for ~ 2 hours [07:50] I'm having a bit of trouble registering acpi events. [07:51] I've created an event file, and an action script. the acpi events are reported from acpi_listen, however, the script doesn't execute when the event occurs [07:51] when I attempt to start, stop, or debug acpid i get the message: acpid: can't open /proc/acpi/event: Device or resource busy [08:13] good morning pitti [08:14] pitti: I think I need some sponsoring for gnome-desktop, just grab the branch and it should be ok :) [08:15] morning [08:24] hey huats [08:25] hello didrocks [08:25] don't try to talk to me here :) [08:25] :p === kermiac_ is now known as kermiac [09:06] good morning everyone [09:07] hey chrisccoulson, ready for your week-end? ;) [09:07] hey didrocks, yeah, i'm definately ready for the weekend. it was quite a late night again last night [09:07] are you ready for the weekend? [09:09] chrisccoulson: tons of stuff to finish first, but yeah, it'll be again a FLOSS week-end :) [09:10] ooh, bbiab - breakfast has arrived :) [09:11] enjoy ;) [09:12] heya [09:14] salut baptistemm [09:15] salut didrocks [09:29] well, i could sleep after that :) [09:34] heh [09:34] ;) [09:34] short Friday, isn't? [09:35] wake up, go to work, breakfast and sleep :) [09:35] kklimonda: your gnome-screensaver backtrace misses quite a few symbols. also, was that the only thread running? (i expected to see the PAM stuff in another thread there) [09:35] didrocks - yeah, short friday :) [09:35] then home to do some hacking ;) [09:36] i think i'm going to go the whole day without looking at anything gpm related today [09:36] some kind of "gpm day" :) [09:36] or gpm-free day [09:36] rather :) [09:37] gpm-free day today ;) [09:37] heh [09:37] after spending all last night investigating an issue with blanking [09:38] asac: so, netbook-launcher-efl works well with netbook-launcher installed. Some packaging tweaks to remove the conflicts/replaces against netbook-launcher (not sure why there is a replaces btw as there is no common file) and so, creating a wrapper. That should be doable [09:40] hey seb128 [09:40] hey chrisccoulson [09:40] how are you today? [09:40] good! [09:40] I overslept again ;-) [09:40] salut seb128, on se couche toujours aussi tardivement ? :) [09:40] you? [09:40] seb128 - i overslept today too ;) [09:40] didrocks, lut, ouais, mais on dors le matin pour compenser! [09:41] i had quite a late one again last night [09:41] seb128: héhé. Ça t'a permi d'uploader seahorse-plugins vu qu'il a poussé sa branche à 22h30 :) [09:41] didrocks, oui [09:42] chrisccoulson, so you didn't manage to get your early night again? ;-) [09:43] gir1.0-webkit-1.0 is currently build from two packages (webkit and gir-repository) which makes currently the build of gir-repository "Fail to Upload". Does somebody know which of those two packages should build it? [09:44] geser, ignore that, I've it on my list [09:44] ok [09:44] geser, there is no difference between current gir and the new one [09:44] so we don't really need it [09:44] debian will drop webkit from gir in the next upload [09:44] I pinged one of the debian guys about it some days ago [09:45] so next upload will work [09:49] seb128 - no, i didn't manage an early night this time. i've had 1 early night this week, so that's ok ;) [09:53] re [09:54] didrocks: yep, will do [09:54] pitti: thanks a lot ;) [09:55] hey pitti [09:56] seb128: bonjour! [09:56] seb128: not sure whether you saw the chart from this morning: http://people.canonical.com/~pitti/bootcharts/daniel-lucid-20100212-1-no-rsyslog-dd.png [09:57] pitti, no [09:57] this is with the rsyslog/dd thing fixed locally [09:57] pitti, I'm doing indicator-sound sponsoring [09:57] waouh, 11.5 seconds? [09:58] seb128: :) [09:58] seb128: i-s> merci [09:58] * pitti tries some more shuffling [09:59] the bar seems misleading [09:59] actual end of cpu use is around 13s [09:59] you don't have g-s-d background tweaked there right? [10:00] seb128: no, not yet; I just sponsored that [10:00] will do another chart once that's built [10:00] seb128: yeah, seems to do some lazy init later on [10:01] but that's pretty much what we do with gnome-screensaver, e-d-s, and the like [10:01] hey pitti [10:02] hey chrisccoulson, how are you? [10:02] i'm good thanks, how are you? [10:03] I'm good, thanks! [10:08] pitti, I will need you soon ;-) [10:08] pitti, to NEW review indicator-sound [10:09] chrisccoulson: is it still on your plan to speed up the xrandr plugin? or want me to look into that? [10:09] seb128: standing by [10:09] pitti, if you want to clean a wi too both didrocks and kenvandine have a gwibber mir listed [10:09] pitti - i can still look at that to see if there is anything we can do [10:09] pitti, one can go [10:11] seb128: hm, I see it on https://blueprints.edge.launchpad.net/ubuntu/+spec/desktop-lucid-une-applications, but none on https://blueprints.edge.launchpad.net/ubuntu/+spec/desktop-lucid-social-from-the-start [10:11] seb128: or anywhere at http://people.canonical.com/~pitti/workitems/canonical-desktop-team-lucid-alpha-3.html#ken-vandine ? [10:11] pitti, ok, I pinged kenvandine about that yesterday he must have cleaned it already [10:12] I'm not subscribed to this spec [10:12] so I didn't get the diff to tell [10:12] ah, good [10:12] I remember that diff flying by [10:14] * pitti hmmms at http://people.canonical.com/~pitti/bootcharts/daniel-lucid-20100212-1-mutter-phase-init.png [10:14] that could be something to consider (start mutter early on) [10:15] this is now _real_ 11.8 s (no lazy init) [10:15] and the CPU block looks almost perfect [10:15] no pause in between any more [10:16] it's funny how sensitive all of this is [10:17] http://people.canonical.com/~pitti/bootcharts/daniel-lucid-20100212-1-mutter-phase-wm.png is the "middle ground" between "application" and "init" phase, and it's much worse [10:17] what did you change? [10:17] seb128: move mutter to autostart-phase=init [10:17] so that it starts with g-s-d [10:17] ah right [10:17] I don't think it's using a theme or so [10:20] pitti: there is still some "not 100% busy cycle" around gdm start time, weird [10:20] that's a needle eye indeed [10:21] one CPU is fully busy, the other is free [10:21] so there's a theoretical saving of .5 s there [10:21] hum, sounds great if we can launch something in between :) [10:22] it's the I/O block of X.org and Xsession.d/ [10:22] pitti: how do you see that? (and also that only one CPU is fully busy), on the bootchart? [10:22] but I think we should deal with that once the kernel fix lands, and the loadkeys.sh block goes away [10:23] because I bet it'll look differently again then :) [10:23] didrocks: it has two CPUs, thus if the CPU chart is at 50%, it means that one CPU is idle [10:23] gnome-desktop is built, installed locally [10:23] * pitti bootcharts [10:24] didrocks: I just put a ruler on the point when it drops CPU usage, follow downwards, and see what happens at that time [10:24] pitti: ok, just from assumptions so, not from something strictely written in the bootchart :) [10:24] ok [10:25] didrocks: well, it does enumerate the CPUs in the headers [10:25] ~/.cache/wallpaper/stretched_1024_600__usr_share_backgrounds_warty-final-ubuntu.png [10:25] \o/ [10:25] * pitti hugs didrocks [10:25] * didrocks hugs pitti back ;) [10:27] pitti, should we use a -1 or -0ubuntu1 for dx indicator uploads? [10:28] seb128: I don't mind so much, but I'm slightly biased towards -0ubuntu1 [10:28] just for consistency [10:28] * seb128 is annoyed at debuild complaining about ubuntu revision but no ubuntu maintainer [10:28] which doesn't let me build it [10:28] I can tweak DEBEMAIL to build but still annoying [10:28] DEBEMAIL= bzr bd -S ? [10:29] right, what I just said [10:29] but oh well, why not set it to ubuntu-desktop@ or u-d-discuss@ [10:29] we should not have to pay for being upstream [10:29] seb128: but it just checks for any @ubuntu.com address [10:29] *shrug* [10:29] what is in the Maintainer: field right now? [10:29] the email is a canonical one [10:29] ah [10:29] Maintainer: Conor Curran [10:31] I saw some packages using the Original Maintainer when we are upstream and ubuntu-desktop@ as maintainer (a bit weird, but well…) [10:32] yeah, I did tha [10:32] yeah, I did that [10:32] didrocks: http://people.canonical.com/~pitti/bootcharts/daniel-lucid-20100212-2-mutter-phase-init.png [10:32] that's with your caching patch [10:33] it's not faster in total, but we have some free CPU cycles left [10:33] so I'll shuffle the bits some more to get a compact chart again [10:33] pitti: right, I'm a bit disappointed that the time is not used for anything else :) [10:33] pitti: moving some bits to start earlier? [10:33] or later, I don't know [10:34] pitti: you have the collector running, isn't it? [10:35] oh, hang on, this was the wrong one [10:35] it started mutter in phase=windowmanager, which is bad [10:35] didrocks: it's always running apparently [10:35] that's ureadahead [10:35] but ureadahead isn't showed in the graph? [10:35] pitti, what do we need to do to be able to push to lp:ubuntu/source? [10:35] oh sorry, seeing it now :) [10:35] pitti, we need to upload the source first right? [10:36] seb128: the package must be in any PPA, or in NEW [10:36] seb128: right, get it into NEW, push it to ~ubuntu-desktop/ubuntu/lucid/pkgname/lucid, and afterwards ping james_w to set that as the official branch [10:36] it's currently a bit complicated for new packages [10:37] pitti, why not just ~ubuntu-desktop/ubuntu/source? [10:37] if that works [10:37] I thought you had to specify the full 5-component path [10:38] we are there yet for the easy to use ;-) [10:38] james_w, help, I'm being confused :p [10:40] pitti, indicator-sound your way btw (just uploaded should hit NEW) [10:47] pitti: came across a weird issue last night. I did an upgrade from karmic to lucid and somehow it removed my broadcom sta driver, problem is I can't get it back on. Jockey log say /sys/module/wl/drivers does not exist, cannot rebind wl driver [10:47] davmor2: could it be bug 506816 ? [10:47] Launchpad bug 506816 in dkms "wl missing after Karmic -> Lucid upgrade" [High,Confirmed] https://launchpad.net/bugs/506816 [10:48] * davmor2 reads [10:50] pitti: yeap [10:52] pitti: any more info you need for it? [10:52] I don't think so, it's pretty well triaged now, and easy to reproduce [10:53] okay cool [10:53] Qapla! [10:53] seb128, didrocks: http://people.canonical.com/~pitti/bootcharts/daniel-lucid-20100212-2-mutter-phase-init_maximus-phase-app.png [10:53] purrrrfect [10:54] what did you shuffle? [10:54] mutter -> phase init, maximus -> phase app (instead of wm) [10:54] niiiiiiiiiiiccccceeeeee! [10:54] so that maximus can spread out into the empty CPU ranges instead of blocking startup [10:54] 11 seconds isn't long enough to go and make a coffee though ;) [10:54] cool [10:54] hehe [10:54] * pitti ^5s seb128, didrocks, and chrisccoulson [10:54] sweet ;) [10:55] and Keybuk, too! [10:55] I'll reboot two or three times to see how it reproduces [10:55] I've a friend who told me that the computer work he's using under win takes 7 minutes to start and finish activity [10:55] * seb128 hugs pitti [10:55] minutes?? [10:55] yeah [10:55] that's with a million virus scanners surely? [10:55] seb128 - i've found that too with most company PC's [10:55] they put all sorts of rubbish on [10:56] hence the comment about not being long enough to go and make a coffee ;) [10:56] seb128: it was 4 at my previous work. Time to go to bathroom, say hello to co-worker ;) [10:56] pitti, likely yes [10:56] wow, in 2 weeks time i'll not have to use a windows machine again [10:56] pitti: I got 10.2s over the weekend with some fiddling [10:57] chrisccoulson, feels good ;-) [10:57] as soon as I can get reliably dailies, I'm going to retest and if it still works, upload :D [10:57] chrisccoulson: you'll see, it's very good. I'm addicted now ;) [10:57] seb128 - oh yes ;) [10:57] Keybuk, plumber fiddling or desktop? [10:57] (removes almost a second from the kernel-X bit) [10:57] seb128: plumbing fiddling [10:57] cool [10:57] Keybuk: rock, and that's not even with the recent bg caching and the autostart shuffling [10:57] indeed [10:57] so with the background cache + pitti tweaks we are on target [10:57] \o/ [10:58] Keybuk, btw what is happening with gdm to desktop? [10:58] you know, as soon as we hit 10s, I'm going to send an e-mail telling everyone to stop what they're doing, the release is DONE! :p [10:58] will we get plymouth covering background during loading or something? [10:58] seb128: not sure what you mean? [10:58] yeah we should [10:58] right now you can see the gnome-panel being drawned etc [10:58] there's some plymouth vt bugs with fixes pending [10:58] oh, that's DX's stuff [10:58] you'd have to ask Cody [10:59] * Keybuk will biab - got a hair cut [10:59] meh, between two identical boots there's so much noise these days [10:59] i really should get a bootchart from my laptop [10:59] ("being drawn") [10:59] it definately isn't 11 seconds though ;) [10:59] Keybuk, ok thanks [11:00] chrisccoulson, ssd? [11:00] seb128 - no, it's not ssd unfortunately [11:00] i might get a mini though [11:00] Keybuk: your daily charts are still killed by the broken ubiquity automation, I figure? [11:01] darn, the next boot takes 11.42 instead of 11.15 [11:01] but either way, we've come a loong ay [11:01] way, even [11:01] yeah [11:02] but as long as we get one solid CPU block without pauses, we're good [11:02] didrocks: oh, we still need the ubiquity hook for the bg cache, right? [11:02] pitti: right, ubiquity and update-manager one [11:02] pitti: that's why I didn't close the WI :) [11:03] didrocks: it won't be in u-m, but in the gnome-desktop postinst, I figure [11:03] we need a 5-line mini C program to refresh the background and trigger the caching, right? [11:03] exactly [11:03] didrocks: I'll do those WIs explicitly [11:04] pitti: ok. I'll complete them on Monday morning (I want to finish triaging and selecting bugs for UNE today) [11:05] didrocks: done [11:05] seb128, pitti: yeah, it has to be the full 5-part path for the first push: lp:~ubuntu-desktop/ubuntu/lucid/package/lucid [11:05] didrocks: that's fine, it's not so urgent [11:05] pitti: thanks :) [11:05] james_w, why? [11:05] james_w, I used ~ubuntu-desktop/source/ubuntu that doesn't work? [11:05] seb128: that's what LP requires [11:06] bah, sucker [11:06] seb128: that does work [11:06] (not you) [11:06] james_w, thanks ;-) [11:06] that's just not what you said earlier :-) [11:06] james_w, can you get ~ubuntu-desktop/indicator-sound/ubuntu [11:06] to be lp:ubuntu/indicator-sound? [11:06] yes [11:06] thanks [11:06] james_w, ups, did I switch ubuntu and source again in the url? [11:07] I tend to do that often ;-) [11:07] yeah :-) [11:08] james_w: hey o/ btw, does the new plugin version/bzr are now available into lucid for merge-upstream command? [11:08] james_w, sorry about that and thanks for looking into it ;-) [11:08] james_w: (btw, I've converted two new branches to that workflow and it really rocks :-)) [11:08] didrocks: btw, I'm using lp:ubuntu/maxiums now (it doesn't have a Vcs-Bzr, and I didn't find a packaging branch) [11:09] pitti: I've done that too IIRC, maybe the changes have been removed as it was in the time I encountered that issue [11:09] seb128: done [11:10] didrocks: nice [11:10] didrocks: no, I haven't done an upload yet [11:10] james_w, thanks! [11:10] james_w: ok thanks, I keep doing from the local checkout so ;) [11:11] * seb128 gently nudge pitti about NEW === kermiac is now known as kermiac_ [11:11] pitti, I would like to get that new indicator landing before weekend if possible ;-) [11:11] james_w: oh, the only annoying issue is that for the first merge, I have always to bzr revert debian/ [11:11] in case things break so somebody is still around working to fix it [11:11] didrocks: because all the dx branches use to have packaging in them? [11:12] seb128: doing, sorry [11:12] james_w: I was thinking that too, but I'm sure not the last one [11:13] didrocks: if you find a case where they didn't then please file a bug [11:13] pitti, np, thanks [11:13] james_w: ok, will do :) [11:20] seb128: looks good except the wrong license in debian/copyright; can you please reupload with dropping the "or any later version"? [11:20] pitti, ok [11:21] sorry I overlooked that one [11:24] pitti, done, bzr updated and source reuploaded [11:29] seb128: wrt. the bzr branch, I still don't know how to get pristine-tar manually [11:29] the auto-imports have those [11:29] seb128, kenvandine: so indicator-sound will replace gnome-volume-control in lucid already? [11:29] (source NEWed into main, FTR) [11:30] pitti, yes, when you NEW the indicator I will comment gnome-media [11:30] pitti, ok thanks [11:30] doing the gnome-media autostart change now [11:32] gnome-volume-control uses a fair chunk of CPU; let's see how i-s will behave [11:32] yeah [11:32] I'm out of benchmarks land for now [11:32] I don't want to reboot my work box all the time [11:33] and cpu there is fast anyway so it doesn't behave the same way [11:35] lunch time, bbl [11:36] seb128: I'll benchmark it, don't worry [11:36] enjoy lunch! [11:43] seb128: binary-NEWed, too [11:46] fun, bug 330766 haunts me, but on NFSv4 [11:46] Launchpad bug 330766 in pulseaudio "pulseaudio hangs, prevents login, home as ntfs" [Medium,Fix released] https://launchpad.net/bugs/330766 === MacSlow is now known as MacSlow|pizza [11:46] doesn't hang, but fails to start [11:48] ok, saw why [11:50] didrocks: coool. thanks [11:51] didrocks: do you need anything from us? [11:51] or will you just try to do the fallback? [11:52] * asac wonder if we have enough space for both launcher on UNE image [11:57] asac: I didn't look at the space. It brings a lot of depends [11:59] asac: I think it's ok, I'll change the packaging for netbook-launcher-elf and try to run the fallback in that case [12:06] didrocks: rock. [12:07] didrocks: even if general UNE doesnt ship it that way we probably want to do that in our armel images [12:08] asac: I think ship it directly will depend on size/speed of startup of the wrapper [12:08] kenvandine, seb128: your libindicate patch introduced *another* weird crash which has been reported upstream and waste my time again... see https://bugzilla.gnome.org/show_bug.cgi?id=609417#c9 [12:08] Gnome bug 609417 in General "Crash when I double click in notification icon (panel)" [Critical,Resolved: notabug] [12:08] asac: but as you can now install UNE in a general desktop, that can be interested [12:08] RIGHT [12:08] cassidy, the tone is not really required or useful there [12:08] oops [12:08] * asac fights with his keyboard being kind of broken [12:08] seb128, I'm a bit tired of those tbh [12:09] cassidy, what do you suggest doing? [12:09] especially as the patch here is obviously wrong [12:09] be more careful before pushing patches? [12:09] you really think people add bugs for the fun? [12:11] I don't. But it's pretty frustrating for me to see that's the same patch adding bugs again and again [12:12] what about getting the change upstream so it's maintained properly there? [12:13] what about getting the change upstream so it's maintained properly there? [12:13] ups [12:13] cassidy, is there anything which was blocking that? [12:13] I've not been following reviews and issues [12:13] but I'm happy to help getting it in shape if work is still needed [12:14] the plan was to re-implement the libindicate integration by using the proper Telepathy API; but this has been delayed to lucid+1 [12:14] what can we do meanwhile to make things better? [12:14] (partially because Empathy doesn't implement approvers either, so that's our fault too) [12:15] I can understand it's annoying sometime for you [12:15] as said, be more careful before pushing your patches. This hunk is obviously wrong, a simple review of the patch would have spot the issue [12:15] on an another side it's not trivial to jungle with hundred of component for which you don't know the codebase well as a distributor either [12:16] especially when you are under pressure to get a stack ported in a day for a transition [12:16] (which is what happened for this one) [12:16] :\ [12:16] cassidy, I will have to look to the code but the libindicator api added timestamp now [12:17]