[00:09] rickspencer3: a thought just struck me: do we have a test and evaluation plan for desktop audio for Lucid? === bjf is now known as bjf-afk === robbiew is now known as robbiew_ [00:37] Hi! Anybody knows why /usr/bin/asoundconf is no longer included in alsa-utils package on Ubuntu 9.10? Is it installed by another package? === asac_ is now known as asac [00:42] dyek: it is no longer maintained. pitti wrote the original; I maintained it for several years. It is the wrong approach and encourages a mess of poor configuration files. [00:43] dyek: my opinion on the matter is fairly easy to search (and find) on the Web: if someone else wishes to pick it up for maintenance, s/he has my blessing, but be aware that it is the wrong approach. [00:44] dtchen: So, is the right approach is just have udev detects it as part of PulseAudio? [00:44] dyek: to answer your original question more completely: it was removed because of increased maintenance burden and because it is the wrong way to proceed with runtime device configuration in the context of PulseAudio. It is not installed by another package. [00:45] dyek: in the presence of PulseAudio, it is strictly unnecessary. If something doesn't work, one or more of linux, alsa-lib or pulseaudio needs to be fixed. [00:47] dtchen: PulseAudio works well on Ubuntu 9.10 in the beginning, but then I found it generates tons of this messages: sink-input.c: Failed to create sink input: sink is suspended. [00:47] dyek: for troubleshooting, please join me in #ubuntu-audio-help or #ubuntu-bugs (the former preferably). [00:48] Correction: ...some intermitten of the above message and tons of this message: alsa-util.c: Unable to set sw params: Too many open files [00:49] Now, PulseAudio can't detect my second card for some reason...speaker-test also doesn't work...which is why I'm looking for asoundconf, and whatever that could fix the issue. [00:50] Thanks for the response. I'll see what I can find out or get help elsewhere. [00:55] dyek: you should file a bug: ubuntu-bug alsa-base [00:55] dtchen: OK! Thanks! [00:55] dyek: the triagers will carry it forward === eeejay is now known as eeejay_away === eeejay_away is now known as eeejay === astechgeek is now known as techgeek [04:04] oh, that's crazy [04:04] udev and/or xorg is detecting the accelerometer in my macbook as a joystick, thus the mouse jumping back to the center of the screen [04:06] disabling the applesmc driver corrects this === bob__ is now known as Guest21788 === eeejay is now known as eeejay_away [08:00] Good morning [08:17] hey pitti [08:17] hey didrocks, bonjour [08:17] good morning tseliot [08:18] pitti: good morning to you :-) [08:27] good morning everyone [08:30] hey chrisccoulson [08:30] chrisccoulson: sorry for breaking screensaver [08:30] I'll get to it today [08:30] but I really want to keep the caching patch [08:31] hey pitti - that's ok :) [08:31] I compared karmic (with ureadahead) vs. lucid bootcharts on my system yesterday -- 89 vs. 49 seconds! [08:31] that's pretty impressive [08:32] look at http://people.canonical.com/~pitti/bootcharts/tick-karmic-20091119-1.png vs http://people.canonical.com/~pitti/bootcharts/tick-lucid-20091210.png [08:32] gnome-panel I/O went from some 25 seconds to some 5 :) [08:32] i still hope to gain another half a second from the changes i'm going to do in gnome-session / gconf / g-s-d [08:32] but that's a drop in the ocean compared to these numbers ;) [08:32] chrisccoulson: no, it's not [08:32] chrisccoulson: your "half a second" is on the Dell mini 10, I guess [08:33] mine is a Latitude D430 with an utterly crappy HD [08:33] so half a second on the mini will be like 4 seconds for me :) [08:33] if the dell mini would boot in 49 seconds, Scott would kill us on the spot :) [08:34] i'm hoping to eliminate the delay you see on your bootchart between gnome-session and g-s-d starting [08:34] s/eliminate/reduce [08:35] chrisccoulson: that's the gconf bit? [08:35] pitti - yeah, most of that is blocking on gconfd loading the defaults [08:35] good morning everybody [08:35] hey pitti chrisccoulson! [08:35] bonjour seb128 [08:35] good morning seb128 [08:35] speaking about login? [08:35] we are :) [08:35] seb128: was just telling chris: I did a new bootchart on my freshly installed alpha-1 yesterday (well, with my extra packages again) [08:36] pitti: actually I think Scott would make you suffer horribly as he killed you slowly and not on the spot at all ;) [08:36] seb128: 89 s (karmic+ureadahead) vs. 49 s (lucid alpha-2) [08:36] waouh [08:36] davmor2: yeah, it'd take at least 89 seconds of pain :) [08:36] pitti - your bootchart is interesting, as it's not using auto-login. you can see the GDM's gnome-session starts quicker than your users gnome-session [08:36] is that on d430? [08:36] seb128: look at http://people.canonical.com/~pitti/bootcharts/tick-karmic-20091119-1.png vs. http://people.canonical.com/~pitti/bootcharts/tick-lucid-20091210.png [08:36] where is the chart? [08:36] seb128: and see how the gnome-panel I/O reduced [08:36] thanks [08:36] and that's because the GDM's session is started without gconf [08:36] seb128: g-panel> from 25 s I/O to some 5 [08:36] weird [08:37] chrisccoulson: right, manual login [08:37] but I'd say I'm pretty constant in pressing enter and my password [08:37] when I bootchart, I hover over enter when gdm appears [08:37] I think the variance should be some 0.5 s [08:38] bah [08:41] login takes some 5 seconds on today's chart on my laptop... [08:41] or 6 rather [08:41] that's good [08:42] how did you manage that? [08:42] I don't know! [08:42] I didn't change anything out of upgrading [08:42] yesterday boot took 61 seconds [08:42] today's one takes 41 seconds [08:44] ureadahead was not working yesterday apparently [08:44] or it took 1 second where it takes over 25 seconds today [08:45] let me do a new chart just to see, brb [08:54] ok it's constant [08:55] http://people.canonical.com/~seb128/bootchart/seb128-laptop-lucid-20091210-5.png [08:56] I don't start nautilus on that box though [08:57] did anybody there tried to run bughugger on lucid? [08:57] seb128: wow, your ureadahead process takes crazily long [08:57] however, desktop startup is pretty good :) [08:58] right [08:59] gosh, your panel and compiz need next to no cpu/io [08:59] not true [08:59] its quite impressive how quickly everything starts after ureadahead finishes [08:59] if you look at the cpu graph around login it's at 100% [08:59] that doesn't really reflect on the chart colors for some reason though [09:10] chrisccoulson: where did you see the "[manager_select_theme_for_job] gs-manager.c:250 (22:27:34): Could not find information for theme: fiberlamp" line? [09:11] pitti, btw I had menu entries duplicated for things which got upgraded yesterday [09:11] until restart that's it [09:11] I think that's another bug due to your changes [09:11] I will open it if I find a way to trigger the issue [09:11] I did remove the cache before to test that gnome-screensaver bug though [09:11] so it might just be weird interaction in cases where you delete the cache [09:12] seb128: hm, a simple sudo touch /usr/share/applications/gedit.desktop doesn't reproduce it [09:12] seb128: thanks [09:16] pitti - i saw that when i ran "gnome-screensaver --debug --no-daemon", and let the screensaver activate [09:16] ah, I tried with gnome-screensaver-properties (which also has this bug) [09:16] chrisccoulson: thanks [09:17] I already found the corresponding code, I just wondered where you got it [09:17] i had a quick glance at the code last night, and i think the code path is triggered by not being able to fetch an Exec= value [09:17] ah, you found it already :) [09:17] I think I know what's going on [09:20] pitti, the dup bug is easy to get, remove the cache and install a package with a desktop entry, the entry will be duplicated [09:21] it's just a refresh issue though I think because it's fine after next start and the cache is correct [09:21] ah, ok [09:23] asac: any news on the root certificate addition to firefox? [09:27] TeTeT: everyone i asked said that its painful [09:27] to add that in a global way [09:27] asac: so, there's no recommended way to doing that? [09:28] reason is that mozilla always rejects contributions that would make it easier [09:28] because they dont want distros to mess around with the root cert list shipped [09:29] TeTeT: yes. there is no recommended way [09:29] asac: hmm, is there any way at all to do that? [09:29] there is. during build you can add more certs [09:30] but their is no way to do that from outside in a way that it wouldnt get wiped on upgrades [09:30] asac: so there can only be additional root certificates for Ubuntu, not for a single custom site [09:30] right [09:30] but even for ubuntu we are not allowed to [09:31] asac: because of redistribution rights? [09:31] imo we should file an upstream bug [09:31] yes trademark issues [09:31] let me check [09:31] asac: can you file the upstream bug, or want me to do that? [09:35] https://bugzilla.mozilla.org/buglist.cgi?query_format=advanced;version=Other%20Branch;component=Security;component=Security%3A%20CAPS;component=Security%3A%20PSM;component=Security%3A%20S%2FMIME;component=Security%3A%20UI;resolution=WONTFIX;resolution=DUPLICATE;resolution=---;product=Core [09:36] if there was a bug about it it should be in that list [09:37] i cant believe there was not bug about that ;) [09:42] asked again [09:42] pitti, how likely are we going to get respins today? ie how careful should we be with uploads? [09:43] seb128: I think the current images are final for a1 [09:43] asac: thanks [09:43] pitti, ok thanks [09:47] asac: in principle, could we fork ca-certificates for the customer and have them place their root cert there. When an update for ca-certificates would come, they'd have to merge and re package again [09:52] TeTeT: right. problem is that ca-certificates is not considered for mozilla [09:52] there was upstream discussion about that.... and we got turned down - basically because they want us to ship certs in sync with what they ship [09:53] asac: oh, where are the mozilla certs then? [09:53] TeTeT: they have built-in certs. directly put into a shared lib during build time [09:54] asac: what is /usr/share/ca-certificates/mozilla about then? [09:54] libnssckbi.so [09:55] good question. i think thats just a dir maintained/used by ca-certificates package [09:55] where the ca-certificate maintainer dumps from time to time all the built-in certs to [09:55] asac: I thought the lib is compiled via update-ca-certificates? [09:55] e.g. you can go and say: yeah, mozilla does a great job selecting ca-certs carefully ... i want to use those globally for other things likst openssl etc. [09:56] guess I misunderstood the role of update-ca-certificates [09:56] TeTeT: ca-certificates are shared among other ssl libs [09:56] but not for libnss [09:56] asac: ok [09:57] so it works for everything but mozilla/nss stuff [09:57] asac: got it [09:59] asac: there are two libnssckbi.so, one in xulrunner, one in nss. Which one does ff use? [10:04] TeTeT: i think the xulrunner one [10:04] everything else would be be a bug imo [10:06] asac: ok [10:23] seb128: there's a good chance that the fix for bug 494669 also fixes your issue [10:23] Launchpad bug 494669 in gnome-menus "gnome screensaver only works with blank screen in lucid" [Medium,In progress] https://launchpad.net/bugs/494669 [10:23] I messed up the file_ids [10:24] pitti, ok thanks [10:24] seb128: so, it's "purge gedit, drop cache, restart panel, install gedit" -> two gedit in the menu? [10:25] hm, no, the cache is only read at the start of gnome-panel [10:25] pitti, no, it's drop cache, dpkg -i gedit*.deb [10:25] seb128: so g-panel started with an existing and valid cache? [10:25] new cache is created and gedit is duped in menus [10:25] yes [10:28] hum, the mini didn't like today's update [10:28] gnome-panel keeps using cpu and has empty bars displayed [10:28] seb128: still can't reproduce [10:28] I'm afraid I need a more detailled recipe [10:28] pitti, using the current version or your fixed one? [10:28] current [10:28] ok, what I do is [10:28] * start desktop [10:29] * open menu, see one totem or gedit or whatever [10:29] * sudo rm /usr/share/applications/desktop.locale.... [10:29] where desktop.local is the cache [10:29] * open menu see they are still correct [10:29] ok, I'm up to here now [10:29] * sudo apt-get install --reinstall [10:30] or sudo dpkg -i *.deb for something [10:30] * open menu, notice that the launcher for that specific desktop is displayed twice [10:30] aah [10:30] ok, got it [10:30] good [10:30] sorry I was not being clear before ;-) [10:31] gnome-panel --replace -> one gedit; sudo apt-get install --reinstall gedit -> two entries [10:31] no need to fiddle with the cache (it's only read at panel startup anyway) [10:31] ok [10:31] nice, will test that with the fixed version [10:32] do you want me to open a bug about it? [10:32] seb128: give me 2 mins, it's at dh_gencontrol [10:33] seb128: ok, it's fixed [10:33] as I suspected, it also stumbled over the wrong file_id [10:33] it compares by file_id [10:34] * seb128 hugs pitti [10:34] good job [10:34] * pitti hugs seb128, well I broke it in the first place :) [10:34] but I really want to keep this patch [10:34] it makes such a difference here [10:34] yay, that also fixed software-center [10:34] \o/ [10:34] * pitti loves that [10:35] 3 bugs with one fix [10:35] good ;-) [10:42] well, it's three fixes actually, but still just one root cause [11:07] \o/ gnome-panel works again on the mini config [11:07] will teach me to enable ppas [11:07] it was due to gtk csd [11:27] meh, that third bit is so excruciatingly hard to fix :( [11:56] re [11:59] hum [11:59] why is locale.alias read several times? [12:00] * seb128 asks on #ubuntu-x [12:03] pitti rocks \o/ [12:04] same question a .ICEauthority [12:05] about [12:06] it's read 4 times in a row === MacSlow is now known as MacSlow|lunch [12:34] asac: I fanned out all the unassigned MIRs now, FYI [12:43] hmm ... /me feels outrun. thx [12:48] asac: I hope you aren't sad now :( [12:48] i am not ;) [12:49] i started to pick a few not-mobile ones for me last night ... but then never came back :/ [12:51] asac: so far I assigned the ARM ones to lool, since you already had a fair share [12:52] but I guess it would be better if you had taken the ARM related ones and lool the others [12:52] pitti: yeah. i didnt pick the ones from the mobile team. as i wasnt sure if we do that or not [12:52] since in the end it's your team who has to provide the support for those [12:52] usually i would suspect that server MIR member konws about server, but also is biased [12:52] ok [12:52] asac: of course there's somewhat of a conflict between "we want that" and "we are the best people to review that" [12:52] thats the other perspective [12:53] i wasnt sure which perspective we prefer [12:53] neither am I :) [12:53] it probably can't hurt to take non-related packages [12:53] but i agree that having me review mobile feels better in the long run. /me needs to stay unbiased ;) [12:53] and ask the team/tech lead for more input if in doubt [12:53] right [12:53] if that would work and doesnt start blocking teams thats better [12:53] ideally, the MIR report sohuld be complete enough to be useful for an "outsider" [12:53] i think the team lead will notice and bug the assignee anyway === MacSlow|lunch is now known as MacSlow [12:54] rather than just acking because its now needed [12:54] asac: ok, so let's just stay with the "gut feeling and common sense", and slightly prefer reviewing other team's packages? [12:54] lool: ^ any opinion? [12:54] yes [13:15] pitti: fine wiht me [13:15] Note: I'm helping with a booth at the arm symposium event, so will be mostly afk today === robbiew_ is now known as robbiew [14:49] /join #ubuntu-server [14:49] heh [14:49] hey rickspencer3 [14:49] hi pitti [14:50] space matters! :) [14:50] salut rickspencer3 ;-) [14:50] as a python programmer, you would think I know that [14:50] bonjour seb128 [14:51] how is everyone today? [14:51] * rickspencer3 has a full callendar for the next 3 hours [14:52] Hello, everyone! I have menu's missing in applications such as evolution, empathy, gnome-terminal and others. It appears the missing menu bar is only in gnome applications. Programs such as thunderbird and firefox are not affected. How do I get the menu bars back? [14:52] Zylogue, is this on Lucid? [14:53] rick, no. Actually Fedora, but I have seen several other questions similar to thin in ubuntu forums that did not have answers. I was hoping to find one here. [14:53] rickspencer3, okish, spending lot of time on nautilus speed without getting anything out of it [14:53] it's a bit frustrating [14:54] gtk csd breaks gnome-panel on my boot speed config btw... [14:54] Zylogue, oh, well this is a channel where we collaborate on creating the next version of ubuntu desktop, so we probably won't be able to help much [14:54] it uses cpu for ever and never display any content [14:54] oh? [14:54] kenvandine, ^ [14:54] hmmm [14:54] rickspencer3, somebody else commented on the bug having the same issue [14:54] let me talk to robbiew, I think the plan was to not let things in that wreck start up time [14:55] oh, it's not a start up issue, just a bug? [14:55] rickspencer3, oh no [14:55] rickspencer3, right, it just breaks gnome-panel [14:55] "just" ;-) [14:55] "just: [14:55] :) [14:55] seb128, ok [14:55] I've not be able to tell if it makes a difference on speed due to that though [14:56] seb128, maybe take a break from nautilus start time for a day? [14:56] rickspencer3, yes, I will do that once I think, thanks [14:56] rickspencer3: on the bright side, karmic boot time was 89 s for me; now it's at 49 :) [14:56] yes, good progress has been made already [14:57] and I have lots of faith in seb128, and I note it's only December 10th [14:57] ;) [14:57] thanks ;-) [14:57] seb128: shall we replace it with midnight commander, do you think? [14:57] well I think we will finish by not running nautilus nor compiz [14:57] much faster than naughtylus [14:58] but *shrug* let's see [14:58] all what we win is good for users anyway and as pitti pointed we are already much better than we did in karmic [15:01] rickspencer3, thanks. [15:01] seb128: no nautilus you think? [15:02] mvo, yes, no desktop [15:02] like run nautilus the first time somebody open a location [15:03] * mvo nods [15:03] I might be wrong though [15:04] but I'm adding markers to nautilus and stacing it for a day now [15:04] and lot of startup time is actually things we need to do [15:04] like fontconfig reading fonts [15:04] gtk reading icon caches [15:04] libx11 reading x11 cookies for session registration [15:05] etc etc [15:05] and we have lot of fonts, lot of icons, ... [15:08] I wonder if we can clean up the fontconfig stuff [15:09] seb128: it seems curious to me why only nautilus is interested in fontconfig? shouldn't it affect all apps? [15:10] it's not nautilus only [15:10] but I've been profiling that one [15:10] we have the fontconfig init cost at session login once anyway === bjf-afk is now known as bjf [15:32] kenvandine: hi, how comes empathy-dbg pkg doesn't exist in empathy debian/control file on lucid? [15:37] bigon, dunno but those are of not use with dbgsym anyway [15:37] they just pollute mirrors and the apt indexes [15:38] right but there is nothing in the changelog [15:54] seb128: is there anyplan to enable empathy with geoloc stuff? [15:54] no [15:54] not for lucid at least [15:54] ok [15:54] we discussed it at uds, desrt reviewed geoclue [15:54] it's buggy, crashy and not ready to be used [15:55] he tested with 2 gps, ip reverse, etc [15:55] and upstream says geoclue is not actively maintained [15:55] mmm ok [15:55] would be lot of effort for no real benefit [15:55] it's rather a gadget in empathy than something useful [15:56] we did consider it because we wanted to use it in the installer and clock applet though [15:58] seb128: and only enabling the map? [15:58] what does the map do if you don't use geoclue? [15:58] can it be installed without the geoclue libs, etc? [15:59] n900 publish localisation :p [15:59] only champlain [16:00] tseliot, hi [16:00] would be so nice if someone picked up the upstream ball there... [16:00] indeed [16:00] rickspencer3: hi [16:01] tseliot, call time? [16:01] we might have a look to work on that next cycle but we can't this cycle, too much to do with the coming lts and priority is not on new features [16:01] rickspencer3: yep [16:40] * pitti creates lucid retracer chroots [16:41] seb128: ^ FYI [16:41] pitti, thanks [16:41] I think we need to enable apport soon === eeejay_away is now known as eeejay === halfline_ is now known as halfline === mac_v_ is now known as mac_v [18:07] ok [18:10] good night everyone! [18:10] 'night pitti [18:10] * seb128 just finished going through a nautilus strace and taking notes [18:11] seb128 - did you find anything interesting? [18:11] well [18:11] it's rather a boring list of what is done [18:13] sec [18:15] http://people.canonical.com/~seb128/nautilus [18:15] has the notes [18:15] the first column is the time from the strace [18:16] I will have a look tomorrow to what we can cut [18:16] the early start doesn't seem to waste a lot [18:16] but there quite some time spent reading all the icon dirs [18:16] translations [18:16] the backgrounds [18:16] custom icons, bookmarks, etc [18:16] some things could perhaps be done better [18:17] like the bookmark list seems to be updated several times [18:18] strace is probably not the best tool to see where cpu is spent too [18:19] yeah, there seems to be lots and lots of things which use small amounts of time, which increases the effort needed to gain improvements [18:19] "12.216: use .X11-unix socket" - i take it that is part of drawing the background? [18:20] I started having too much of that and didn't investigate from the background part [18:20] I will do that later or tomorrow [18:20] but it's opening the background image several times [18:20] and doing some mmaping, socket writing, etc [18:20] dunno why all that though [18:23] urg, I'm late for sport need to run [18:23] see you later [18:23] bbl [18:43] pitti: bug 495161 ;) [18:43] Launchpad bug 495161 in cryptsetup "initramfs cryptroot hook bloats armel initrd by adding >13M of compressed modules" [Critical,Triaged] https://launchpad.net/bugs/495161 [18:45] pitti: you just did the merge i guess. anyone i should assign that to? or look on my own? etc. let me know [19:05] gah, it's compiz bug day [19:05] * Amaranth is flooded [19:06] pedro_: You have to give me whatever script you're using to flood my inbox with 50 mails a minute :) [19:07] Amaranth, sure, but it's just a silly script to close bug numbers, you need to pass the numbers to it anyways [19:07] oh, nevermind then [19:07] bryce_ has a much better script for that, it checks for bugs incomplete for > 30 days and closes them automatically [19:08] right, i just prefer to have an extra look in case i'm missing something [19:08] yeah mailbox recieving 1100 mails from ubuntu-x in an hour can attest to that :D [19:08] Amaranth, I'd be happy to extend that to run on compiz as well if you'd like (I'm slowly working on generalizing my scripts for more than just X.org) [19:08] heh [19:09] hehe [19:09] bryce_: I already run it with some modifications for compiz [19:09] Amaranth, aha kewl [19:09] is the archive frozen? [19:09] * Amaranth has a fix for panel applet shadows [19:09] open again now [19:10] they don't show their shadow when on top of another window but that's an applet bug as they have the same problem with metacity [19:10] but at least the shadow goes away [19:11] hmm, I never did learn how to sign a package where someone else's name is in the changelog [19:24] there, invest-applet users can stop yelling at me now [19:25] Amaranth - if you're building a source package with someone else in the changelog, you can run "debuild -S -sa -k [19:26] (if you haven't figured that out already) [19:26] I did, but only after I redid the changelog to have my name :) [19:27] i just have DEBSIGN_KEYID= set in ~/.devscripts [19:29] ooh, didn't know about that one [19:29] chrisccoulson: does the key need to be in quotes? [19:30] Amaranth - it doesn't need to be in quotes [19:30] i think that works anyway. i must have set it in the past for some reasons, and all packages I build are magically signed by me :) === eeejay is now known as eeejay_away [19:42] chrisccoulson: you seem to know to what /tmp/.X11-unix/X* are used [19:43] grr, another FTBFS on armel [19:43] still KDE stuff [19:44] * Amaranth starts pretending armel doesn't exist, get happier [19:44] s/get/gets/ [19:44] :) [19:47] kenvandine, how is the Dx release going? [19:53]