[01:22] <YokoZar> slangasek: Are FTBFS bugs going to hold up the release?  (eg https://bugs.launchpad.net/ubuntu/+source/fltk1.1/+bug/445560)  -- I need those fixed before I can update ia32-libs, which really needs a push before release.
[01:49] <slangasek> YokoZar: it's supposed to be fixed before release, but we don't really have any traction on that bug yet AFAIK - why does ia32-libs need updated?
[02:03] <LaserJock> slangasek: would uploading http://launchpadlibrarian.net/33791667/moodle_1.9.4.dfsg-0ubuntu4.debdiff be OK?
[02:04] <LaserJock> it seems to fix the bug in Karmic anyway, I was going to test Jaunty->Karmic upgrades before uploading though
[06:29] <slangasek> : why is dropping puppet's recommends: on libaugeas-ruby the right thing to do?
[06:29] <slangasek> zul: why is dropping puppet's recommends: on libaugeas-ruby the right thing to do?
[06:40] <slangasek> superm1: is the patch for bug #385658 really tested to DTRT on unsupported cards? (cf. bug #413439)
[06:41] <slangasek> superm1: in particular, I notice that the code now adds an entry to NVChipsets,NVPciChipsets even if walking the list explicitly does *not* find a matching chip
[06:46] <slangasek> superm1: oh, n/m, I guess you must have tested it on your own hardware as the bug submitter :)  still, bug #413439 is still open
[09:40] <AnAnt> Hello, I have a question regarding the latest change in gdm (setting the default theme using an XML file), will this step help to ease branding ?
[10:45] <Laibsch> ArneGoetje: Any idea on the issue of CJKV characters again not displaying properly in Flash?
[10:47] <Laibsch> bug 207198
[10:47] <Laibsch> Who could we ask?
[10:48] <Laibsch> Unfortunately, flash is becoming more and more important and difficult to avoid
[13:34] <YokoZar> slangasek: ia32-libs needs updating because the libraries it contains are over a month old and out of sync with the real 32-bit libs (so any bugs fixed in them remain).  There's also a few bugs that I've patched in the ia32-libs package itself (mostly missing symlinks for apps to not get confused when they're on 64 bit)
[15:03] <bdrung_> who can close this blueprint? https://blueprints.launchpad.net/ubuntu/+spec/eclipse
[15:51] <ArneGoetje> Laibsch: hmm... was it working before? And if yes, what changed?
[15:56] <Laibsch> I think it was working before
[15:56] <Laibsch> I don't remember specifically changing anything as far as configuration
[15:56] <Laibsch> Just the normal upgrade routine
[15:57] <Laibsch> But this stretches over such a long period of time that I cannot absolutely guarantee that there were no changes to the configuration
[15:57] <Laibsch> All I can say is what I (Rolf) wrote in that ticket, IOW currently flash and CJKV don't mix for me, no matter what I do
[16:01] <ArneGoetje> Laibsch: well, we haven't changed anything in fontconfig or language-selector in that regard. the only package which was updated, was firefox
[16:01] <Laibsch> does CJKV work for you in FF?
[16:01] <Laibsch> and flash
[16:02] <ArneGoetje> Laibsch: let me check
[16:02] <Laibsch> In general CJKV works for me in FF, of course, but not in flash
[16:02] <Laibsch> Try the links I gave (jal.co.jp)
[16:02] <Laibsch> some things work, some don't
[16:04] <ArneGoetje> Laibsch: I tried the flash map of JAL, which you reported as not working. Works for me. But I would need to confirm on a fresh install in a VM...
[16:05] <Laibsch> thanks
[16:05] <Laibsch> interesting
[16:05] <Laibsch> I wonder what I may have done to break things
[16:06] <ArneGoetje> Laibsch: I'm starting up a VM with the Live CD... wait a few minutes, I will try to confirm.
[16:07] <Laibsch> If this is OffT, let's move this to a private chat?
[16:08] <Laibsch> ArneGoetje: Just to be sure, the innosked link works for you?
[16:13] <ArneGoetje> Laibsch: yes, that one works
[16:22] <ArneGoetje> Laibsch: just tried the same page in the daily Live CD from yesterday. Works without problems
[16:22] <ArneGoetje> Laibsch: I just installed the Adobe Flashplugin from software center
[16:26] <Laibsch> I'll see if I can try a different user and a live USB boot tomorrow
[16:26] <ArneGoetje> Laibsch: which locale did you use for testing?
[16:26] <Laibsch> I'll let you know the results
[16:26] <Laibsch> Which LC_* setting do you want to know?
[16:26] <Laibsch> I don't think that it had any relation to the problem, though
[16:27] <Laibsch> My locale is usually UTF8
[16:27] <ArneGoetje> Laibsch: LANG and LANGUAGE, maybe LC_CTYPE
[16:27] <Laibsch> some setting de_DE
[16:27] <Laibsch> others en_US
[16:27] <ArneGoetje> Laibsch: I tried en_US and it worked... will try other locales as well
[16:28] <Laibsch> LANG=de_DE.UTF-8, LANGUAGE=de_DE.UTF-8, LC_CTYPE="de_DE.UTF-8"
[16:29] <Laibsch> I think that is probably what FF is running at
[16:29] <ArneGoetje> Laibsch: just tried with de_DE: also works fine.
[16:30] <ArneGoetje> Laibsch: will try with CJK font settings
[16:30] <Laibsch> yes
[16:31] <Laibsch> I don't think it had any relation to the locale
[16:31] <Laibsch> I may have broken something in fontconfig
[16:31] <Laibsch> and it may be something totally different altogether
[16:31] <Laibsch> But judging from the reports in that bug ticket, I'm not the only one who currently has problems
[16:34] <ArneGoetje> Laibsch: there we go! using fontconfig-voodoo with ja_JP displays fine... but zh_TW has the empty boxes...
[16:45] <Laibsch> OK
[16:45] <Laibsch> I'm glad you can at least reproduce sometimes
[16:45] <Laibsch> That helps at least ;-)
[16:50] <ArneGoetje> Laibsch: yeah... I'm just trying now to find out what's the problem with the fontconfig snippets
[16:50] <Laibsch> cool
[16:50] <Laibsch> let me know if there is anything I can do to help with my limited knowledge
[17:01] <ArneGoetje> Laibsch: ok, found the problem... and I'm afraid it's not something easy to fix. :(
[17:02] <ArneGoetje> Laibsch: seems Adobe Flash can only use one font and doesn't fall back to another font when glyphs are missing.
[17:03] <ArneGoetje> Laibsch: For ja-JP we prefer Japanese fonts, even for ASCII, so there it's not a problem a Japanese font is used for the whole system
[17:04] <ArneGoetje> Laibsch: but for zh-* we prefer DejaVu and then the Chinese fonts, because Latin glyphs in DejaVu look better. apparently this doesn't work with Flash.
[17:04] <ArneGoetje> Laibsch: when I remove the DejaVu stance in the fontconfig file, a Chinese font is prefered, even for Latin glyphs. Then Flash displays the CJK characters fine.
[17:06] <ArneGoetje> Laibsch: it's just funny that this only happens if we prefer the specific ordering of DejaVu, followed by a CJK font... if we don't specify this and fontconfig picks a font by itself, flash also displays fine.
[17:07] <ArneGoetje> Laibsch: (i.e. we remove the 69-language-selector-* files)
[17:16] <Laibsch> OK
[17:16] <ArneGoetje> Laibsch, freeflying: will test more for bug #207198 tomorrow, need to sleep now
[17:16] <Laibsch> So what's the suggested solution and its drawbacks?
[17:16] <Laibsch> ArneGoetje: Sure, I was about to hit the sack myself much earlier
[17:17] <ArneGoetje> Laibsch: don't know yet. Let me do some more testing tomorrow.
[17:17] <Laibsch> I hope I'll get to it soon
[17:17] <Laibsch> Ok
[17:17] <Laibsch> If I'm not in this channel, please ping me
[17:17] <Laibsch> I should be online
[17:17] <Laibsch> good night
[17:17] <Laibsch> and thank you for your work
[17:17] <ArneGoetje> Laibsch: ok, good night
[17:18] <ArneGoetje> Laibsch: np
[19:13] <slangasek> YokoZar: what do you mean, missing symlinks?  ia32-libs can't use any paths that are used by the native 64-bit packages
[19:25] <LLStarks> hi.
[19:34] <siretart> slangasek: OK to upload http://launchpadlibrarian.net/33910661/g-s-s.diff, cf. lp bug #428884? should fix screensaver issues with xine and vlc (and any app that uses gnome-screensaver-command --poke)
[19:34] <siretart> patch against gnome-screensaver, not vlc's fault
[19:35] <LLStarks> hey reinhard, do you know if anything is being done to make sure that usplash isn't a static image for the final release?
[19:36] <superm1> siretart, by chance would you suspect that to help bug 453377 too?
[19:36] <siretart> LLStarks: sorry, I don't really follow usplash development
[19:36] <siretart> superm1: looking
[19:37] <LLStarks> who should i ask about it?
[19:37] <superm1> siretart, that patch, gs_monitor_simulate_user_activity, is that the function called when --poke is used?
[19:37] <siretart> I'd guess Keybuk or cjwatson, but check usplash's changelog yourself
[19:38] <siretart> superm1: AFAIU the code there, yes.
[19:38] <Keybuk> which code?
[19:39] <Keybuk> aiui, the current usplash theme is the final one
[19:39] <slangasek> siretart: we should certainly fix that bug; I can't comment on whether the code looks reasonable, I don't know what it's doing
[19:39] <siretart> superm1: yes, I'm actually fairly confident that the patch should fix that
[19:39] <slangasek> Keybuk: thread interleaving, he was responding to LLStarks' query
[19:40] <superm1> siretart, yeah, i just saw that do_poke calls into simulate activity, so it looks like that is the same.  i'll mark duplicate accordingly
[19:40] <siretart> slangasek: I'm not a g-s-s guru myself, but the patch is taken from upstream, and after reading the upstream comment on the patch, I believe we should include it
[19:40] <siretart> plus: we have an independent user who confirms the patch works
[19:40] <Keybuk> LLStarks: the current usplash theme is the final one I believe
[19:40] <slangasek> siretart: then yes, please upload it; you don't have to get prior approval before uploading
[19:41] <Keybuk> it's not entirely static, the logo does pulsate etc. in certain situations, and it does fade
[19:41] <slangasek> siretart: fwiw, I don't count "one user says it works" for much for changes like this :)
[19:41] <sladen> Keybuk: so, just to confirm, the _intended_ final theme is a static one (eg. no progress indication, no throbber)
[19:41] <LLStarks> i can't get it to pulse even on a fresh daily.
[19:41] <siretart> slangasek: ah, I see
[19:41] <Keybuk> sladen: right
[19:41] <slangasek> siretart: so the vlc task there should be closed as invalid?
[19:41] <Keybuk> LLStarks: the pulsing is quite subtle, it shows up for me on this laptop, but on my other one it's hard to see
[19:42] <superm1> LLStarks, it only pulses when using casper (eg booting the daily, not on the post install)
[19:42] <sladen> Keybuk: okay, so what is this "logo does pulsate etc. in certain situations, and it does fade" you speak off
[19:42] <LLStarks> if i was new to linux and had an old machine or non-ssd hard drive, i'd think my computer locked up.
[19:42] <siretart> slangasek: yes, I wanted to confirm first that it's really not vlc's fault
[19:43] <LLStarks> i have a 3 year old laptop and it spends about 20 seconds at usplash.
[19:43] <sladen> Keybuk: where should I be looking?---if I know that there should be some (a fortnight ago it faded on shutdown) and I can report bugs and it not doing so
[19:44] <sladen> Keybuk: do you have a video of what's intended
[19:45] <LLStarks> sladen, there is fade in and fade out on my machine, but no pulsing whatsoever. it would be awesome if that wasn't restricted to livecd/liveusb.
[19:45] <Keybuk> sladen: I don't understand? are you saying it does nothing for you?
[19:46]  * sladen reboots to confirm his presumped madness.  bbiab
[19:47] <LLStarks> for me, the shutdown usplash almost never runs.
[20:01] <sladen> Keybuk: right, fade on shutdown worked.  Bootup takes 2:20 from Grub finishing.  4 seconds textmode, 2 seconds cleared textmode (cursor blinking), 20 seconds usplash (no action, no throbb, no nothing), 9 seconds text mode "calling settle", 5 seconds blank text mode (again), 4 seconds blank/black graphics mode, 1 second busy cursor, 1 second nothing, 13 seconds Xsplash, GDM ... 12 seconds Xsplash, 13 seconds semi-desktop, 8 seconds Panels, 37 seonds Panel
[20:02] <Keybuk> sladen: bootchart
[20:07] <sladen> Keybuk: http://www.paul.sladen.org/ubuntu/karmic-20091018-2.png
[20:08] <chrisccoulson> siretart - that gnome-screensaver patch makes the assumption that the MIT-SCREEN-SAVER extension is always available, which might not always be the case
[20:08]  * sladen stars at the couchdb forkbomb "wtf"
[20:08] <chrisccoulson> if it's not, then gnome-screensaver will just crash instead
[20:09] <Keybuk> sladen: yeah you have the I/O problem of death
[20:09] <siretart> chrisccoulson: oh, can you perhaps improve the patch to check for that extension?
[20:09] <Keybuk> sladen: 4200 RPM hard drive?
[20:10] <slangasek> sladen: you should purge the couchdb package
[20:10] <slangasek> it's not installed by default
[20:11] <slangasek> (it was earlier in the cycle, then the package was split for this exact reason)
[20:12] <Keybuk> sladen: if you look at the chart, you're basically in I/O wait the entire time
[20:12] <Keybuk> (top graph mostly red)
[20:12] <Keybuk> but while there's always something on the I/O queue (bottom graph mostly red), you're achieving almost no throughput from your hard drive (green line almost flat)
[20:12] <Keybuk> if you want a challenge, figure out why!
[20:15] <chrisccoulson> siretart - i'll have a look if i get the chance, but the patch doesn't really look right anyway. gnome-screensaver currently gets session idle-ness from gnome-session, which uses the SYNC extension and IDLETIME counter. applications which want to inhibit the screensaver should really be the gnome-session inhibit API now, although it would be nice to get gnome-screensaver to work correctly too
[20:15] <chrisccoulson> anyway, dinner time for me
[20:15] <sladen> Keybuk: right, so the lack of I/O is the reason that usplash does not throb?
[20:15] <Keybuk> sladen: no...
[20:15] <Keybuk> nothing went PULSATE at it I guess
[20:16] <sladen> Keybuk: whilst I can make do with bootup taking twice as long as jaunty, I'd like it not to look rubbish
[20:16] <siretart> chrisccoulson: is there some migration documentation? The vlc developers seems pretty confused about this bug and I'd like to help them to understand the issue...
[20:16] <Keybuk> sladen: I'd rather we figured out why this has got so bad
[20:16] <Keybuk> than wasted time with pixels
[20:16] <siretart> slangasek: based on chrisccoulson comments, I please reject my gnome-screensaver upload :(
[20:17] <Keybuk> from my POV, screenfuls of random text that are only visible for a few seconds is better than something pretty that lasts an hour
[20:17] <Keybuk> sladen: you didn't confirm your drive, btw?
[20:17] <Keybuk> sladen: could you paste me your model number, I'm collecting some data on who's affected
[20:17] <sladen> Keybuk: no idea, but it's HDD, not flash
[20:18] <slangasek> siretart: ok, rejected
[20:18] <Keybuk> sladen: cat /sys/block/sda/device/model
[20:18] <elmo> hey, is there a wiki page on bootcharting?  my karmic boot is horrific, and I'd like to generate pretty pictures once I purge some of the non-desktop packages
[20:18] <sladen> Keybuk: ST980825AS
[20:19] <sladen> elmo: sudo apt-get install bootchart
[20:19] <Keybuk> sladen: interesting
[20:19] <elmo> sladen: that's it? :-P
[20:19] <sladen> elmo: then reboot and look in /var/log/bootchart/*
[20:19] <elmo> ok
[20:20] <slangasek> chrisccoulson: incidentally, after upgrading gnome-screensaver last night, this morning I came to a hung 'time expired' dialog... did this "who kills the child" change get tested to DTRT for running instances of gnome-screensaver?
[20:20] <Keybuk> sladen: that's 7200RPM with a 10.5ms seek
[20:20] <Keybuk> that's a *fast* drive
[20:20] <Keybuk> something's seriously fucked there
[20:20] <Keybuk> sladen: could you upload your dmesg somewhere
[20:22] <sladen> Keybuk: http://www.paul.sladen.org/ubuntu/karmic-20091018-2.dmesg.txt
[20:27] <Keybuk> sladen: one thing to try
[20:27] <Keybuk> boot with break=bottom
[20:27] <Keybuk> stick deadline into /sys/block/sda/queue/scheduler and then ^D to continue the boot
[20:27] <Keybuk> see if your chart looks better
[20:27] <Keybuk> sorry
[20:27] <Keybuk> break=top :)
[20:27] <Keybuk> (you may have to mount/unmount sys)
[20:29] <Keybuk> sladen: nothing in dmesg suggests a DMA problem or anything
[20:30] <Turl> hello
[20:33] <Turl> in a bootchart I just made I can see gnome-session taking a *lot* of time to launch things http://ubuntu-pics.de/bild/27825/laptop_karmic_20091018_1_uY9yMi.png
[20:33] <Turl> any ideas why?
[20:34] <Keybuk> Turl: uninstall couchdb
[20:34] <Turl> why Keybuk?
[20:36] <Keybuk> Turl: that's what's causing your problem and you don't need it
[20:38] <Turl> why is it included by default then?
[20:38] <Keybuk> it isn't
[20:39] <Keybuk> we did for a while during the development phase, but the package has now been split apart
[20:39] <Keybuk> apt-get remove --purge couchdb
[20:39] <Keybuk> should do it
[20:40] <Turl> mhm, I shouldn't have removed everything that had couchdb on its name ^^
[20:40] <Turl> reinstalling desktopcouch should reinstall it right?
[20:42] <Keybuk> right ;)
[20:42] <Turl> done :)
[20:43] <Turl> any other thing I should remove/add to gain more performance?
[20:43] <Turl> I'm using karmic since alpha 5/6 iirc
[20:44] <Keybuk> you have a bug we're trying to track down too
[20:44] <Keybuk> that's going to be my #1 task for all of next week
[20:44] <Turl> what bug Keybuk?
[20:44] <Turl> I can help with testing if you need :)
[20:45] <Keybuk> 432089
[20:45] <Keybuk> bug 432089
[20:46] <Turl> yep my disk is a bit slow
[20:47] <Turl> Keybuk: do you know if gnome power management bugs are high priority?
[20:48] <Turl> I reported one, https://bugs.launchpad.net/ubuntu/+source/gnome-power-manager/+bug/228932 but it is set to low
[20:48] <Turl> and IMO, it's high importance as it can cause data loss and system damage
[20:48] <Trewas> actually I just tested and sreadahead seems to improve boot speed, 80s with it and 120s without (5400rpm laptop disk)... however I only tested once without sreadahead
[20:49] <Keybuk> Turl: out of my remit
[20:49] <Keybuk> Trewas: yeah, it's that it doesn't *really* improve it, or *always* improve it
[20:49] <Keybuk> with sreadahead I'm looking for a particular shaped I/O graph that we're not achieveing
[20:49] <Keybuk> something in the kernel is defeating I/O performance
[20:50] <mdke> slangasek: just uploaded a new gnome-user-docs with new translations, in due course could I call on you to do your NEW dance for the new binaries?
[20:51] <Trewas> yeah bootchart shows that I/O is very slow and the processor is mostly waiting, and this is a slow atom
[20:53] <ion> su224300 < Turl> any other thing I should remove/add to gain more performance?
[20:53] <ion> Gnome for starters ;-)
[20:54] <Turl> ion: nah :P
[20:55] <LLStarks> keybuk, is there any chance that pulsing logo from the casper environment can be implemented for installed systems. also, do you find it acceptable that the usplash on an installed system shows no indicators of progress even if it takes 30 or more seconds for the usplash to be passed on a slower system?
[20:55] <Keybuk> LLStarks: I find it unacceptable that it takes 30s at all
[20:55] <Keybuk> I'd rather spend my time fixing that
[20:57] <LLStarks> http://img38.imageshack.us/img38/5394/kingfisherkarmic2009101.png
[20:57] <LLStarks> take a look at my bootchart.
[20:58] <LLStarks> 17 seconds of usplash.
[20:58] <LLStarks> 54000 rpm drive
[20:59] <slangasek> chrisccoulson: the 'time expired' hang is reproducible even after restarting gnome-screensaver; we've got a regression here
[20:59] <LLStarks> keybuk, i don't expect you get a system without a solid-state drive down to 10 seconds, but this is ridiculous.
[21:00] <Keybuk> LLStarks: you have the exact same I/O performance bug as sladen
[21:00] <Keybuk> you're getting almost no throughput on your drive
[21:01] <Turl> Keybuk: also, if fsck checks the disk, usplash stays there and there's no written indication that the disk is being checked
[21:01] <Keybuk> yes there is
[21:01] <Turl> so you wonder what's going on
[21:01] <Keybuk> grab the mountall package from the ubuntu-boot PPA
[21:01] <Keybuk> that'll be in the final release
[21:01] <Turl> Keybuk: the last time it checked my disk, there was none
[21:01] <Keybuk> you didn't have that installed
[21:01] <Turl> mhm, can you link me to the ppa?
[21:02] <Keybuk> yes.
[21:02] <Keybuk> PPA
[21:02] <Keybuk> err
[21:02] <Turl> nevermind, found it :)
[21:02] <Keybuk> http://launchpad.net/~ubuntu-boot/archive/++A
[21:03] <Keybuk> http://launchpad.net/~ubuntu-boot/archive/+ppa
[21:03] <Keybuk> even
[21:08] <LLStarks> keybuk, link to bug?
[21:08] <Keybuk> llstarks: bug 432089
[21:09] <LLStarks> what kind of boot times can i expect with the ppa package?
[21:10] <Keybuk> the PPA package isn't about boot times, it's about fsck progress notification
[21:10] <Keybuk> you're crossing threads
[21:11] <Turl> Keybuk: what happened to the graphical OS selector that karmic was suppoused to bring?
[21:11] <elmo> Keybuk: http://people.canonical.com/~james/tuna-karmic-20091018-1.png
[21:13] <Keybuk> Turl: unknown, not my dept
[21:13] <Keybuk> elmo: yup, same thing :-(
[21:13] <Keybuk> elmo: I wish I knew what causes it
[21:13]  * Turl reboots
[21:13] <LLStarks> bbiab
[21:15] <chrisccoulson> slangasek - i'll tty and reproduce that g-s-s hang now
[21:15] <slangasek> chrisccoulson: thanks
[21:15] <slangasek> mdke: yep, I'll be here for the NEWing when needed
[21:18] <chrisccoulson> slangasek - i can't recreate it here. i get the "timeout has expired" message, and then the dialog goes away
[21:18] <chrisccoulson> any chance you could get a backtrace of both processes after it appears to have hung?
[21:18] <slangasek> chrisccoulson: sure; it'll be a few minutes
[21:19] <chrisccoulson> no worries, thanks
[21:25] <mdke> slangasek: :)
[21:29] <LLStarks> keybuk. here's my bootchart with the newer mountall.
[21:29] <LLStarks> http://img32.imageshack.us/img32/5394/kingfisherkarmic2009101.png
[21:33] <joaopinto> LLStarks, <Keybuk> the PPA package isn't about boot times, it's about fsck progress notification
[21:33] <LLStarks> i know
[21:33] <LLStarks> but how would he know that my machine was affected?
[21:35] <ion> keybuk: progress[61] = '\0'; should be progress[60].
[21:37] <slangasek> chrisccoulson: bug #454959 with backtraces
[21:37] <chrisccoulson> thanks
[21:38] <slangasek> chrisccoulson: including a likely pointer to why it's not reproducible for you
[21:38] <joaopinto> LLStarks, becaue your symptom matches the bug already reported
[21:38] <joaopinto> because
[21:39] <LLStarks> okay
[21:40] <LLStarks> next topic. can we get this bug elevated?
[21:40] <LLStarks> https://bugs.edge.launchpad.net/firefox/+bug/438868
[21:41] <LLStarks> this is a very visible bug and i think asac needs help.
[21:50] <ion> keybuk: The PPA version crashed with an assertion in nih_free with ctx->destructor != NIH_ALLOC. I bet the progress[61] write happened to overwrite that pointer. :-)
[21:50] <chrisccoulson> slangasek - i found an issue with my patch
[21:51] <chrisccoulson> it seems that SIGTERM might be raised to quit the dialog after the main loop has already quit, which will probably make it hang like you describe
[21:51] <bdrung_> ScottK: i will let you know if we have something usable
[21:55] <slangasek> chrisccoulson: ah, ok
[21:55] <slangasek> chrisccoulson: I'm happy to test a patch
[21:56] <lifeless> slangasek: hi
[21:56] <lifeless> slangasek: are you aware of the bug with intel graphics sluggishnish + corruption ?
[21:56] <lifeless> slangasek: if so, is it a release blocker, if not, whom should I talk to to make it more widely known ?
[21:58] <sladen> Keybuk: still ~100 seconds with echo deadline > .../sda/queue/scheduler  do you want to see the chart
[21:59] <Turl> Keybuk: hi
[21:59] <Turl> Keybuk: that mountall left my system unbootable T_T
[22:01] <slangasek> lifeless: no, I've not heard of such a bug
[22:01] <Turl> Keybuk: usplash appeared, then it said sth like "/ waiting on /dev/..../by-uuid/...., /home waiting on /dev/...., /temp waiting on (null) press esc for a recovery console" and esc did nothing
[22:01] <slangasek> not counting the "Xorg eats 100% CPU because upstart stole its controlling tty" bug
[22:01] <Turl> rebooted, and the same thing happened
[22:01] <lifeless> slangasek: not that bug, no.
[22:01] <slangasek> lifeless: bug #?
[22:02] <Turl> so I reverted it to the karmic one and I'm using it now
[22:02] <lifeless> slangasek: digging, sec
[22:02] <Turl> slangasek: you published the mountall package on the ubuntu-boot ppa right?
[22:02] <slangasek> Turl: no
[22:03] <Turl> oh, it was Keybuk, sorry :P
[22:05] <lifeless> https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/421232
[22:06] <lifeless> https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/454984
[22:07] <lifeless> slangasek: ^
[22:07] <lifeless> the corruption bug I just filed because I saw there didn't seem to be one for it
[22:08] <slangasek> lifeless: I think it's pretty clear that you should open a new bug report for your performance issue, since you say you started seeing it on Thursday and the original submitter filed his report back in August
[22:08] <slangasek> lifeless: a list of what packages were upgraded for you on Thursday would be helpful, as well as information about what chip you actually have
[22:08] <lifeless> slangasek: that was thursday a week back
[22:08] <lifeless> slangasek: right after bryce announced mesa 3
[22:08] <slangasek> lifeless: a week back was still not August ;)
[22:08] <lifeless> slangasek: ack; filing new bug.
[22:09] <Turl> lifeless: my graphics performance on intel seems to have decreased recently, I don't know the exact date though
[22:09] <Turl> but compiz feels more sluggish
[22:09] <chrisccoulson> slangasek - would you mind testing http://paste.ubuntu.com/296366/ please?
[22:10] <chrisccoulson> the issue seems to be that when the dialog times out, it quits the main loop, and raises SIGTERM if there are still any PAM threads. the SIGTERM does nothing because i registered a handler which quits the main loop (which has already quit)
[22:10] <chrisccoulson> i think that's what causes your hang
[22:12] <dtchen> pitti: ok, i fixed your mute issue harder.
[22:12] <lifeless> slangasek: I've got another corruption bug a friend of mine is reporting; its suspend related but a regression.
[22:12] <lifeless> slangasek: [I'm taking advantage of your being online to talk about this :P]
[22:14] <slangasek> lifeless: yah, I can confirm bug #454984 here (done so in the bug report now)
[22:14] <slangasek> lifeless: suspend-related> new bug report?
[22:17] <lifeless> slangasek: intel card, when the user comes out of s2 gets corruption of a different sort
[22:18] <slangasek> chrisccoulson: something better than a pastebin for a patch, please?
[22:18] <slangasek> chrisccoulson: there's a bug open, you could upload it there :)
[22:19] <slangasek> lifeless: s2, not s3?
[22:20] <slangasek> (do we even have quirk handling for s2?)
[22:20] <chrisccoulson> slangasek - i've attached the patch to the bug report as well now
[22:24] <slangasek> chrisccoulson: building now
[22:28] <slangasek> and testing now
[22:28] <chrisccoulson> thanks
[22:29] <m3ga> lifeless asked me to log a bug about xserver-xorg-video-intel : https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/454999
[22:29] <slangasek> chrisccoulson: how long is the dialog timeout?
[22:30] <slangasek> oh, there it goes
[22:30] <slangasek> chrisccoulson: yep, confirmed fixed
[22:30] <chrisccoulson> excellent, thanks for testing
[22:30] <chrisccoulson> do you want to upload this once i've pushed it to bzr?
[22:30] <slangasek> (at least, with a sample size of 1)
[22:30] <slangasek> chrisccoulson: can do
[22:30] <chrisccoulson> cool, thanks
[22:32] <slangasek> m3ga: what does "s2disk" mean?  Are you using the standard hibernate functionality in GNOME, or does this refer to some other utility?
[22:32] <slangasek> m3ga: IIRC, s2disk is the name of a program provided by uswsusp?
[22:33] <chrisccoulson> slangasek - i've pushed it here: https://code.edge.launchpad.net/~ubuntu-desktop/gnome-screensaver/ubuntu
[22:34] <slangasek> chrisccoulson: got it
[22:34] <chrisccoulson> thanks
[22:34] <m3ga> slangasek: suspend to disk
[22:34] <slangasek> m3ga: yes, by what means?
[22:34] <m3ga> yes provided by uswsusp.
[22:34] <m3ga> running "sudo s2disk"
[22:34] <m3ga> in an xterm
[22:34] <slangasek> ok, I'm going to reassign this bug to uswsusp then
[22:35] <slangasek> unless you can reproduce this using pm-utils' hibernate mode?
[22:35] <slangasek> (sudo pm-hibernate)
[22:35] <m3ga> then i probably need to file a new bug about sluggish performance.
[22:36] <slangasek> chrisccoulson: uploaded
[22:36] <slangasek> m3ga: sluggish performance of what? pm-hibernate?
[22:36] <slangasek> (I think that bug's been filed a few dozen times before)
[22:36] <m3ga> x server has been sluggish in karmic for about 2 weeks
[22:36] <slangasek> ah
[22:37] <m3ga> i thought it was bug 421232 but lifeless seems to think it needs a new bug
[22:37] <slangasek> m3ga: can you please check whether this problem is reproducible with 'sudo pm-hibernate' instead of 'sudo s2disk'?  That will let us establish whether it should be treated as a driver bug, or if uswsusp is missing quirk handling that pm-utils knows about
[22:38] <m3ga> ok, i'll lose my xchat session but i'll be back
[22:38] <slangasek> m3ga: 421232 has a generic description and was reported months ago - it's far better to file a new bug report and let the triagers decide whether it's the same bug
[22:38] <m3ga> will do
[22:39] <lifeless> thanks slangasek
[22:41] <slangasek> sure
[22:47] <m3ga> slangasek: 'sudo pm-hibernate' causes similar screen corruption.
[22:47] <slangasek> m3ga: ok, thanks for confirming
[22:47] <m3ga> before it did it I restarted x server and it was fine.
[22:47] <m3ga> ie no corruption before pm-hibernate' corruption after
[22:48] <m3ga> slangasek: you want the sluggish perf bug report now or can that what til my evening (9am here).
[22:48] <m3ga> s/what/wait/
[22:50] <slangasek> m3ga: I'm not the person who's going to be triaging it anyway, I'm sure it can wait if you have other things to do
[22:50] <m3ga> i'll do it this evening. its monday morning and i'm at work :-)
[23:10] <slangasek> mdke: gnome-user-docs's build-dependency is on *both* iso-codes *and* isoquery; but it's only used for regenerating debian/control, which shouldn't happen at package build time, so can be ignored
[23:10] <slangasek> mdke: but if you're going to put it in build-depends, it should be both packages (which is what I tried to do in the previous upload, but then I regenerated debian/control right over myself)
[23:15] <slangasek> dtchen: how is this alsa-utils change supposed to fix a race condition?
[23:20] <dtchen> slangasek: apparently alsactl store was being called at odd times during reboot/shutdown, resulting in muted slevels
[23:21] <dtchen> ugh, lag
[23:22] <dtchen> slangasek: anyhow, the previous change in ubuntu5 was broken, so it's definitely a regression from 8.10 and 9.04
[23:22] <slangasek> dtchen: and how do you figure that this isn't also "at odd times"?
[23:23] <dtchen> slangasek: it can only be called once
[23:23] <slangasek> how is that different from the init script, which only shows up once in /etc/rc6.d?
[23:24] <dtchen> slangasek: asound.state was not being written to /var/lib/alsa
[23:24] <dtchen> slangasek: i can't reproduce it on my machines; maybe their disks are too slow
[23:27] <slangasek> so you don't actually have any evidence that there was a race
[23:27] <slangasek> or that this will fix it?
[23:27] <slangasek> sysvinit scripts are handled *serially* in Ubuntu, and upstart jobs are handled in *parallel*.  if anything, you've increased the chance of a race condition here
[23:28] <dtchen> i'm waiting for gmail to load so i can dig up the bug report
[23:28] <slangasek> ok
[23:29] <slangasek> I'm afk for a few hours, but if there's a bug report that explains how this is a race, please throw it at me while I'm out
[23:40] <dtchen> slangasek: https://bugs.edge.launchpad.net/ubuntu/+source/pulseaudio/+bug/352732/comments/63