[00:30] how do you get apport to force filing a bug report even if you don't have all the latest bits installed? [00:31] nm, found it (rtfm bryce) [00:42] hm, don't work [00:48] pitti, APPORT_IGNORE_OBSOLETE_PACKAGES=true apport-cli -c /tmp/xserver-xorg-video-intel.2010-02-17_22\:31\:42.401787.crash does not let me file the bug [00:52] robert_ancell, isn't simple-scan already in Lucid? why would you need a ffe? [00:52] hi rickspencer3 [00:53] hiya bryceh [00:53] bryceh, seems like a plan is coming together for xorg + kernel bits? [00:54] yeah, we still need to get apw's view on it, but I think we got a solid plan finally [00:54] bryceh, I think that's because you already got the crash report ready for upload [00:55] bryceh, the variable works if you didn't do the "report the bug" yet [00:55] anywho ... [00:55] seb128, basically I have the .crash file and want to post it to launchpad so I can see what info it's including [00:55] ie if the crash file doesn't have the stacktrace processed etc yet [00:55] oh [00:55] ie before the bouncing bar work [00:56] so basically get your crash and run that command without using the ui auto opened [00:56] rickspencer3, new versions of softwares which are not purely bug fix updates need exceptions [00:57] seb128, I'm trying to do this remotely because the machine's keyboard isn't working [00:57] Is there a way to find out where in my source these warnings are coming from? http://paste.ubuntu.com/379439/ [00:57] bryceh, well, if you get the crash report and run the command directly without doing anything else with it before that should work [00:57] eh, I'll just upgrade the box to latest and reproduce the report [00:58] heh [00:58] bryce@blumonc:~$ ls /var/crash [00:58] _usr_share_apport_apport.0.crash xserver-xorg-video-intel.2010-02-17_22:37:01.055233.crash [00:59] you crashed apport? ;-) [00:59] guess so! [00:59] rickspencer3, ffe requires "a new upstream version of a package" to get an exception. I guess it depends how much you class Simple Scan as upstream [01:00] robert_ancell, non-bug-fix-only-version [01:00] robert_ancell, bug fix updates are fine === kenvandine_ is now known as kenvandine === ubott2 is now known as ubottu [05:39] Good morning [05:39] robbiew_: libgudev is built by udev source [05:40] bryceh: hm, perhaps it's not an obsolete package, but a non-native one? [06:25] hey pitti, good morning [06:26] hey kenvandine [06:26] i shaved more off of gwibber :) [06:26] 200K [06:26] yay diets [06:26] removed icons we don't use [06:26] and split out the extra themes to a different package [06:26] no need to ship 9 themes :) [06:26] kenvandine: nobody uploaded your egenix splits yet? [06:26] nope [06:26] seb128 reviewed it [06:27] but i guess not uploaded [06:27] bug 523969 if you want to look at it :) [06:27] Launchpad bug 523969 in egenix-mx-base "Split docs out into separate packages" [Undecided,New] https://launchpad.net/bugs/523969 [06:28] kenvandine: opening a tab, will do in a bit [06:29] thx! [06:30] kenvandine: can you forward the debdiff to Debian in the meantime, please? [06:30] sure [06:31] kenvandine: oh, both mtools and datetime are the same source? convenient [06:31] yup [06:31] urgh, yay for not using standard debhelper [06:31] kenvandine: thanks [06:34] pitti, yeah... it was an interesting package to mess with :) [06:34] but i just followed what they already did [06:34] movefiles.sh is an interesting approach [06:34] :) [06:36] kenvandine: right, that "urgh" wasn't directed at you [06:37] understand [06:37] i had a similar reaction when i first downloaded it :) [06:46] good morning [06:51] baptistemm: hm, I'm not sure what you mean -- why would any other process influence how bluez evaluates its options? [07:10] pitti, I don't know, The reporter sent a mail to the bluez mailing list and the developper said it was working for him, and asked if any component in ubuntu would influence the power preferences [07:55] didrocks: what I feared: http://cdimage.ubuntu.com/ubuntu-netbook/daily-live/current/ [07:55] didrocks: I think we need to cut some OO.o parts [08:15] Guten Tag pitti, wie geht's? [08:16] hey didrocks, bonjour! [08:16] I'm a bit tired, but great [08:16] I got up at 6, the thought of my xorg blunder wouldn't let me sleep any longer :) [08:16] (all fixed now) [08:16] Xsession.d/* is now really streamlined [08:16] ohoh, you didn't dream about it at least? :) [08:16] (and it only took 4 package uploads..) [08:16] didrocks: no, only took me an hour to get to sleep :) [08:17] rah… [08:17] but that's fine, TGIF :) [08:17] how was your concert? [08:17] oh, it was actually an iceskating performance [08:17] "Holiday on Ice" [08:17] artistic ice skating [08:18] sweet, I know it even if I never attended to their show [08:18] great show, great costumes, nice music, and gorgeous light effects [08:18] didrocks: how are you? survived the release day yesterday? [08:18] I can imagine. Glad you like it :) [08:18] well, they release late. [08:19] and as I'm still ill, it was not a very good experience :) [08:19] but well, new release ready [08:19] urgh, so you are in the middle of the cold? [08:19] didrocks: take it light then [08:19] exactly. I'll try to not make something too brain-damaged today :) [08:20] so, first fix this CD size [08:20] pitti: the OVERSIZE file is always empty? It's just an indicator that the CD is oversized? [08:20] didrocks: exactly [08:21] ok, so, in that can, most of the time you look at the seed and look at what can be moved away, right? [08:21] right [08:21] (I see OOo draw is a good candidate) [08:21] didrocks: it's also helpful to have a rather clean chroot/pbuilder login [08:21] and compare "apt-get install" download sizes [08:21] to see how much dropping a particular thing will buy [08:22] * pitti goes back to *nnnng* speed ups [08:22] pitti: hum, is it possible to install from the seed all softwares and try autoremove in the pbuilder , [08:22] or you just try some kind of heuristic to see what overlap or not? [08:23] didrocks: no, just compare "install oo.o-writer oo.o-calc oo.o-draw" against "install oo.o-writer oo.o-calc" [08:23] ok, trying that :) [08:23] thanks [08:23] (no need to actually install, of course) [08:24] right, just have the space took on disk :) [08:24] didrocks: no, not on disk [08:25] didrocks: look at the download size (.deb size) [08:25] didrocks: squashfs is compressed [08:25] which is pretty similar to the .deb size [08:25] didrocks: for "normal" .debs it's almost identical [08:25] pitti: oh ok, I didn't know it was corresponding approx to an ar archive [08:25] for OO.o the .deb size is a little smaller, since OO.o uses lzma [08:25] noted [08:39] hey there [08:40] hey seb128 [08:40] seb128: how is your cold? [08:40] lut didrocks [08:40] still there but it's not a really annoying one this time [08:40] no blocked nose or anything [08:41] you? [08:41] ok, a light one so :) You'll be able to enjoy your day off [08:41] yeah [08:41] hum, not very good I'm afraid. I think the week-end will be a slackering week-end in my bed :) [08:41] still working a bit this morning [08:41] then I will be off for lunch and enjoy my afternoon [08:41] * seb128 hugs didrocks [08:41] * didrocks hugs seb128 [08:48] urgh, we still install firefox-3.5 in addition to firefox-3.6 on netbook (it's only 0.1MB of additional download, but still). I have to fix also webfav to work with 3.6 [08:55] didrocks: shouldn't we just seed "firefox"? [08:55] hey seb128, good morning [08:55] seb128: wern't you supposed to have a holiday today? [08:56] pitti: for me, right, I just want to see why there is this "germinate workaround" before removing it [08:56] argh, I have gconf-sanity-check back; I thought we disabled that? [09:02] pitti, as said before I'm around a bit in the morning then away for lunch and afternoon [09:02] will likely going to a spa or something this afternoon [09:02] I still want to deal with some aftermatch of yesterday's update today [09:02] after the "all indicator message app are crashing" from yesterday [09:03] and gtk csd makes apps slugish and use cpu [09:03] pitti, btw I uploaded libgnome-keyring to NEW if you can have a look [09:03] pitti, I added the COPYING.GPL and upstream fixed that to git yesterday [09:03] ah, thanks [09:03] I guess it's ok until next upload? [09:04] the code is basically a split of what was in gnome-keyring before I don't think it needs mir review [09:04] right, no MIR necessary, just a new upstream version [09:09] good morning everyone [09:09] hey chrisccoulson, good morning [09:09] hey chrisccoulson [09:09] how are you? [09:09] chrisccoulson: btw, I'm taking another look at your 56gconfd-helper and "26_preload_sources.patch gconf" [09:09] pitti: after having look at the entire thing, I'm a bit concerning for the nebook image. I can only see this candicate: firefox 3.5 (wanting for lool to answer an only 0.1MB) and openoffice.org-help-en-us, and maybe computer-janitor-gtk?). openoffice.org-help-en-us seems to take approx 5MB, so that could work and enable us to add the music store plugin [09:09] hey chrisccoulson [09:10] chrisccoulson: seems it might get late honours :) and we might use it after all [09:10] chrisccoulson: after the latest optimizations, it now nicely squeezes into the free CPU slot before mutter starts [09:10] that sounds good then! [09:11] that's not a good start to the day - i'm out of coffee again! [09:11] oh? [09:11] didrocks: sorry, what's with firefox 3.5? [09:11] i should have got some on the way to work ;) [09:11] didrocks: Oh the seed thing, just saw it [09:11] This was a subtle one, let me recall [09:11] chrisccoulson: none in the office any more? [09:11] lool: right :) [09:12] pitti - there is some other coffee in the office, but everybody else drinks instant coffee [09:13] chrisccoulson, you don't have a starbucks nearby? ;-) [09:13] chrisccoulson: http://people.canonical.com/~pitti/bootcharts/daniel-lucid-20100219-1.png is normal lucid [09:13] chrisccoulson: http://people.canonical.com/~pitti/bootcharts/daniel-lucid-20100219-1-earlygconf.png is with your two gconf changes [09:13] seb128 - i suppose i could drive in to redditch and get a starbucks ;) [09:13] chrisccoulson: look how the CPU usage nicely increases right after the last red bar (desktop startup) [09:14] pitti, early gconf uses cpu for some 0.5s extra though [09:14] seb128: right, but nothing else is in that slot [09:15] so we might just as well start it when we have a bottleneck (gnome-session and nothing else) anyway [09:15] instead of starting it later where we already have enough waiting processes? [09:15] seb128: or do you mean it uses 0.5 s _more_ when you start it early? [09:16] didrocks: Right, so in theory this could be considered a germinate bug, but abrowser gets considered despite firefox being listed earlier (which will eventually pull firefox-3.5 in karmic); it was enough to list firefox-3.5 anywhere in the seed to get this fixed, so that's what I did [09:16] didrocks: Sadly, we need to bump this workaround with each firefox update [09:16] pitti, the earlygconf is 11.5s the other one 11s [09:16] yeah, i just noticed that. how accurate are those markers? [09:16] seb128, chrisccoulson: anyway, I'm currently experimenting with something else still; I'm not firm yet that the gconf change is optimal [09:16] didrocks: I'm not sure whether cjwatson had particular plans to address this, perhaps we will just have to live with the bug [09:16] lool: ok, so I'll just list firefox here [09:16] chrisccoulson: not very [09:17] the end marker varies some plusminus a second [09:17] didrocks: blacklists would make it easier to manage, but are usually a last resort kind of thing, and not meant for this [09:17] pitti, is that still worth spending efforts on boot speed now? [09:17] didrocks: You should list firefox-3.6 [09:17] depending on where it decides that the CPU is "empty" enough [09:17] seb128: we are still > 10 s.. [09:17] didrocks: firefox is already listed earlier in the netbook-remix karmic seed [09:17] pitti, we are down to 10.nn seconds and the remaining part to optimize doesn't seem really GNOME now [09:17] lool: understood, updating now. Thanks :) [09:17] pitti, did we say 10.00s? [09:17] (I checked karmic; let me check the lucid UNE see) [09:17] pitti, or 10s? [09:17] well, desktop is 5 s still [09:18] because 10.8s is 10s to me [09:18] lool: it's the same, we still have the firefox-3.5. I just have to replace by firefox as it's firefox 3.6 [09:18] Oh right we don't have a firefox-3.6 [09:18] didrocks: firefox should be already listed [09:19] pitti, is notify-osd dbus spawned when something open a notify bubble? [09:19] seb128: yes (dbus activation) [09:19] didrocks: it's already there on line 94 [09:19] lool: right, it's already listed, but abrowser depends: firefox [09:19] pitti, I'm wondering what sends a bubble during that boot [09:19] pitti, and if that could be fixed to not [09:19] seb128: hm, good point; nothing is [09:19] didrocks: Hmm so? [09:19] seb128: perhaps something querying for capabilities [09:19] pitti, well nm does the "you can connect by clocking on the notifyicon" [09:20] clicking [09:20] or do you autoconnect to something? [09:20] lool: no more need for the workaround, right? as abrowser will be considered and pull firefox, no? [09:20] seb128 / pitti - nm-applet [09:20] didrocks: firefox is explicitly seeded [09:20] seb128: nm-applet starts later than notify-osd [09:20] that shows a notification when it connects to wireless [09:20] so that's not it [09:20] hmmm [09:20] didrocks: We want to list firefox in the seed because we want to ship firefox, not abrowser [09:21] didrocks: the goal of seeding firefox-3.5 was for the webfav deps to be already satisfied by the seed when considered [09:21] the things before are: gconfd, mutter, g-s-d [09:21] didrocks: But that's not needed anymore since we seed "firefox" and that satisfies the webfav deps [09:21] pitti - i suspect that maybe my osd patch to g-s-d queries the capabilities [09:21] g-s-d maybe does query for notify capability? [09:21] i can look at that this afternoon and fix it if it's dong that [09:22] chrisccoulson: it might defer that until it has a notification to show? [09:22] lool: hence my "no more workaround needed" as webfav deps are already satisfied [09:22] didrocks: it used to be too indirect that we had to have a workaround (we wanted to avoid mentionning the firefox version in the seeded packages), so we added the workaround; it's not indirect anymore, we don't need the workaround [09:22] pitti, can you try a boot without g-s-d to see if notify-osd is still there? [09:22] seb128: good catch! [09:22] pitti, ;-) [09:22] didrocks: Exactly, but do keep firefox itself seeded on line 94 [09:22] seb128: sure [09:22] Just drop the Germinate workarounds sections with 3.5 [09:22] *section [09:22] lool: oh sure, I didn't mean that :) I have also to bump webfav dep [09:23] didrocks: Hmm the webfav deps are broken [09:23] lool: webfav doesn't currently work with firefox 3.6, but I'll fix that [09:23] didrocks: These might just need a rebuild IIRC [09:23] Ok [09:24] seb128: oops, I think it didn't like that; it's spinning on a black bg forever now [09:24] weird [09:24] lool: thanks :) [09:24] chrisccoulson: is that patch in a plugin? or in g-s-d itself? [09:25] pitti - it's in the media-keys plugin [09:25] pitti, did you move the autostart away or did you do your chmod trick? [09:25] bbiab, i have my exit interview with HR now [09:25] chrisccoulson, good luck! [09:25] seb128: chmod [09:25] disabled the plugin now [09:25] chrisccoulson, does g-u-s need sponsoring btw? [09:26] pitti, that might not work, gnome-session might be waiting on g-s-d to be done with registration to go to next steps [09:26] pitti, I usually move the autostart away which makes it not try at all [09:26] anyway turning off media-key seems better [09:27] nope, still starts up [09:28] I'll track that down [09:29] pitti, can you look at libgnome-keyring now? [09:30] pitti, I would like to upload the gnome-keyring updates which block on the lib to be accepted first before going for the afternoon [09:30] oops, sorry; doing [09:30] pitti, danke [09:31] so, no objection about removing openoffice.org-help-en-us from the netbook seed? I don't see any other opportunity of removal apart from computer-janitor-gtk [09:32] uh [09:32] didrocks: dropping presentation and draw is not enough? [09:33] and -base? [09:33] seb128: looks fine, accepted; thanks [09:33] pitti: presentation can be dropped. draw isn't listed, so it should be mixed with something else [09:33] pitti, danke [09:33] pitti, to main? [09:33] yes [09:33] excellent [09:33] pitti: but one of the point of providing OOo again was presentation… [09:33] * seb128 hugs pitti [09:34] didrocks: hm, ok; perhaps we could drop calc then [09:34] didrocks: (fair point, yes) [09:34] pitti: let me check [09:34] didrocks: and we don't even have mono back yet, do we? [09:35] seb128: so I dropped all the g-s-d plugins which link against libnotify, and now it's started later [09:36] pitti: 5 MB for calc apparently [09:36] pitti: no, no mono [09:36] pitti: and I think we'll be asked to bring the music store plugin [09:36] seb128: I'll just start a dbus-monitor and see what else is talking to it [09:36] (just hope it's few KB) [09:37] pitti, I guess every program using libnotify queries for capabilities at init [09:39] RAOF: Heya, you seem to be taking care of libdrm these days, any plans for merging .18 from Debian? it fixes a nasty random crash [09:39] seb128: right, it's in notify_init() [09:39] lool: I believe tjaalton is on it. [09:45] pitti: when you have some time: lp:~didrocks/+junk/netbook.lucid. simple-scan can be the next opportunity if we really want gwibber in :/ [09:46] didrocks: well, if OO.o is too big, we have to go with abiword and gnumeric, and then have space for gwibber, tomboy, mono [09:46] didrocks: dropping simple-scan is fair enough (although it is quite small); it's not a netbook-y app [09:46] pitti: right, I'm still looking for an presentation software which can open odp as this is typically a netbook usage (making a conference) [09:49] RAOF: thanks [09:54] RAOF: sure, I'll add patches to revert the two changes you mentioned [09:55] tjaalton: Thanks. You weren't already on it? I'll prepare a nouveau DDX upload that'll build againt it. [09:56] hey seb128 - you can sponsor g-u-s if you like. it's in a workable state at the moment anyway [09:56] i'll do some more work on that over the weekend [09:57] RAOF: I was about to ask you where we stand :) [09:59] chrisccoulson, ok thanks [09:59] chrisccoulson, how went your hr interview? [09:59] seb128 - yeah, it was ok. but i just wanted to finish and have my breakfast ;) [09:59] hehe [10:00] they didn't offer you some coffee? ;-) [10:06] hmm.. looks like firefox is yet another victim of the gtk-csd [10:07] same problem as with nautilus - when I rearrange tabs or drag selected text I see gray rectangle.. [10:08] issues were expected [10:08] opens bug if there is none and let us know [10:08] I will tag and assign them correctly [10:09] it's still before alpha3 [10:09] we have time for bugfixing before betas and lucid [10:10] yeah, opening it right now - LP was misbehaving once again :) [10:10] seb128: i was about to say "i didnt understand" what you meant in Bug #524301 , only now i noticed rhythmbox cpu usage :D.. duping my bug.. thanks :) [10:10] Launchpad bug 524301 in evolution "Evolution slow when switching mails [delayed formatting]" [Low,Invalid] https://launchpad.net/bugs/524301 [10:12] vish, sorry about the dup undup got some launchpad issues [10:12] vish, but yeah, I think all the "$software uses cpu" are the same gtk issue [10:12] np.. thanks :) [10:13] seb128: bug 524329 - should I tag next bugs I find as gtk-csd or just paste them here leave triaging for you guys? [10:13] Launchpad bug 524329 in gtk+2.0 "gtk csd change breaks firefox dnd text and tabs" [Undecided,New] https://launchpad.net/bugs/524329 [10:13] kklimonda, feel free to tag those gtk-csd [10:23] argh, I still had nautilus disabled on the mini [10:24] * pitti does a reinstall with today's image to get a clean slate again [10:34] didrocks: any chance to get the ubiquity icon into the favourites for alpha-3? [10:35] pitti: oh sure, I was thinking of making iso testing next week, but I still can have a look now [10:36] didrocks: no need to let everything drop, just asking whether that's on your radar [10:36] didrocks: we can certainly survive another alpha if it's hard for some reason [10:37] pitti: I hope it'll not be so complicated. I have a little idea why we don't have it anymore, but I need to check that [10:37] pitti: today, I'm triaging and ordering bugs which should be fixed and those two (with webfav not compatible with FF 3.6) are on my first priorities [10:38] didrocks: sounds fine [10:38] (the third being "no clock applet by defaut) [10:38] default* [10:46] seb128: another Bug #524304 probably gtk-csd related ? you want all to be duped to the meta bug or want to track the different issues? [10:46] Launchpad bug 524304 in gtk+2.0 "High CPU usage with murrine progressbars" [Undecided,New] https://launchpad.net/bugs/524304 [10:47] reassign to gtk for now and keep those separated [10:47] ok.. thanks.. [10:47] those might be apps issues too or theme issues [10:48] brb [11:00] pitti, did you measure what impact gnome-user-share has? [11:00] on boot speed [11:00] seb128: I didn't; I thought it was disabled by default? [11:00] just reinstalling [11:00] pitti, well it exit immediatly since no option is activated by default [11:00] but you were concerned by the loading cost [11:00] I thought it had an AutostartCondition? [11:01] chrisccoulson's update add a delay to the autostart but I'm wondering if that's really required [11:01] pitti, no, as said the other day I didn't do that because the server check for what is enable and exit immediatly [11:01] pitti, autostart can check one key not combinaison, and since you can use it for webdav and,or bluetooth [11:01] that still requires it to link, start up, and connect to gconf in vain? [11:02] well that's what you said the other day [11:02] but I thought the final plan was to run it on demand [11:02] that's why I'm asking ;-) [11:02] seb128 / pitti - if we want to use an autostartcondition, i could add a new merged gconf key to trigger the autostart [11:02] i plan to do some work on this over the weekend anyway [11:02] wah, fresh install, and it fails immediately [11:02] * pitti purges plymouth again and sighs [11:02] well I'm wondering if that makes sense, I would expect loading cost to be marginal [11:03] but since we are after 0.1 seconds [11:03] pitti - is it worth me reporting plymouth bugs, or is it known to be broken? [11:03] plymouth gets worse every time i try and install it... [11:03] pitti, anyway it has a delay=15s with the upload I just sponsored from chrisccoulson [11:03] pitti, so it's not a boot speed concern [11:03] seb128 - thanks for sponsoring :) [11:03] seb128: I just spent two hours to eliminate 4 greps and two "which"es from Xsession.d .. :) [11:03] seb128: ok, that should be fine [11:03] pitti, it just seems ridiculous to delay things this way to win 0.1s out of benchmark effect [11:04] but yeah, we are fighting benchmarks at this point [11:04] well, shouldn't it start on demand? [11:04] not working on improving any user perspective [11:04] well, needlessly starting a program is not just benchmark fighting [11:04] pitti, well, if you can detect that something try to push things over bluetooth that would be nice [11:05] I'm not sure that's the sort of things you get events for though [11:05] I would expect you need the service to listen for those [11:05] doesn't bluez daemon already listen to those? [11:05] we have a bluetooth applet and a system bluez daemon [11:06] adding a third one for functionality that isn't necessary for most people seems to be a waste to me [11:06] I don't have any bluez daemon running there [11:06] and the gnome-bluetooth applet is not a server [11:06] seb128: e. g. the bluetooth applet already notices new devices like phones [11:06] it's an ui to configure bluetooth [11:06] it could start g-u-s once it pairs with a pone [11:06] phone [11:06] right [11:07] pitti - i could look at doing that. it would need to handle the case though where a user allows non-paired phones to send files though [11:07] well right now as said the daemon start, check if anything needs it and exit [11:07] which is possible in the UI [11:07] which I don't think it's anything real world users would notice [11:07] seb128: right, but what's the point in that? it's overhead for zero benefit [11:07] if it's not running, then it won't be there when actually needed either [11:07] well it doesn't exit if you need it [11:08] like you do webdav sharing [11:08] or bluetooth sharing [11:08] we don't want to say "you must have your phone turned on and paired when starting your session" [11:08] right for webdav [11:08] well g-u-s shares things over bluetooth too [11:08] you can browse your computer from an another computer too [11:08] using nautilus [11:09] that again requires pairing first [11:09] depends of what option you selected [11:09] as chrisccoulson's said there is a "share without pairing" option [11:09] that would be a good autostart condition then :) [11:09] and the rest on demand, triggered on pairing [11:09] well we just start having or conditions [11:10] right [11:10] we have zillions of things which would be nice to have but limited manpower though ;-) [11:10] seb128: how come you don't have bluetoothd running, BTW? did you disable it for testing? [11:10] anyway delay=15s, load a binary which check keys and exit should not be end of the world [11:10]