[00:48] for Bug 62230 ... how is the video driver autodetected? i'd like to force certain problematic cards to go to vesa. this bug has been here for at laest the past three releases [00:48] Launchpad bug 62230 in xserver-xorg-video-nv "Corrupt graphics on boot with 7800GT/nv" [Medium,Triaged] https://launchpad.net/bugs/62230 [00:49] safe graphics mode works sure but when you alternate install you get a frozen system [00:49] actually even when you regular install you do [01:26] tjaalton: well, I do see /two/ new binaries added by vdr... :) [01:28] He must want you to leave the other one ;-) [01:31] Hmm [01:31] It seems that when gnome-keyring acts as the SSH agent, it can't cope with loading an RSA1 key === evalles_ is now known as effie_jayx [03:09] jdong: already back, it looks like [03:10] Hobbsee: magical [03:10] Hobbsee: err, on gutsy-backports? [03:28] jdong: oh, meh. [03:29] heya Hobbsee :) [03:30] heya [03:31] jdong: i'll actually have to use the LP UI for that :( [03:32] Hobbsee: aww so it'll take 10 minutes of your time waiting for LP to load? [03:34] jdong: given back [03:34] whee [03:34] and if it FTBFSes again, can I get an Ubuntu brown paper bag? [03:35] Flaming? On a doorstep? [03:40] anteater 1.0 beta is out! :) [03:41] ScottK: :) [03:41] I knew someone would get it. I should've figured you'd be the someone. [03:41] heh [03:46] jdong: why does it die? [03:47] Hobbsee: it should build-dep on bzrtools >= 1.3 [03:47] Hobbsee: but Debian didn't do it, I didn't want to diverge to do it, a rebuild would fix it at worst.. [03:48] Hobbsee: this particular FTBFS results when the archive has new bzr, old bzrtools, and tries to build new bzr-builddeb [03:48] hmm [03:48] that leaves new bzr + old bzrtools in uninstallable state [03:48] so you'll need to rebuild bzrtools too, then. [03:48] no, bzrtools built fine [03:49] it just built a matter of MINUTES after bzr-builddeb tried to build [03:49] hang on, i thought you said it all died again. [03:49] Hobbsee: no I said if! [03:49] hypothetical :) [03:49] oh, right. [03:49] i'm confused, apparently [03:49] * Hobbsee has been dealing with far too much other shit, apparently. [03:51] * jdong gives Hobbsee a hug [03:51] * Hobbsee hugs jdong back. And you're unfortunate enough to know what a great section of it is about, too. [03:53] :( yeah [03:55] Hm. SIGXCPU strikes again. And it's not banshee-trunk-specific, I've just had rhythmbox do it. [03:55] Maybe it only occurs on battery - I couldn't reproduce yesterday afternoon, and I think the difference was I plugged the lappy in. [03:56] RAOF: isn't SIGXCPU used by mono, specifically, the Boehm GC, to signal collection? [03:56] could there be some bug at that level? [03:57] jdong: Except that neither pulse nor rhythmbox are mono apps? [03:57] RAOF: oh sorry. you mentioned banshee [03:57] -ECONTEXT [03:57] Hm. Rhythmbox _really_ doesn't like pulse dying on it. 1.1GB of memory is not a nice mem usage ;) [03:59] jdong: I think you're right about mono's usage of XCPU. At least, debugging banshee suggests that you make gdb pass XCPU through. [04:02] is there a good diet softwares? like for a diabetes guy or a healthy living diet software for person who has heart d eases? [04:04] not the right place to ask.... Please use #ubuntu [04:05] he just asked in at least channels [04:10] pleaseandthankyo: While this isn't the place to ask, the only diet software I know for Linux is "Shrinkingman" -- but I don't believe it is maintained anymore. [04:11] pleaseandthankyo: http://debain.org/software/shrinkingman/ === elkbuntu_ is now known as elkbuntu === golddragon24 is now known as golddragon24_ === golddragon24_ is now known as golddragon24 [05:38] For those playing at home, pulse being killed with SIGXCPU? Not a on-battery problem. [05:50] meh. dapper-backports upload of git-core is blocked on cvsps being in universe in dapper. [05:50] doh [06:01] LaserJock: lovely [06:01] err lamont [06:01] lamont: do you have closer connections with the soyuz folks than us? [06:01] please I beg of you or someone, do something about it [06:01] I've simply fallen upon deaf ears to fix this several-month-old regression [06:01] does it have a bug? [06:01] number that is [06:03] jdong: because a bug number would be a major step towards escalating the fix.. esp if said bug was several months old... [06:03] lamont: yes it does [06:04] https://bugs.edge.launchpad.net/soyuz/+bug/198936 [06:04] Launchpad bug 198936 in soyuz "Regression: Backports should ignore main-universe split" [Undecided,New] [06:04] note the lovely lack of any activity or acknowledgement it exists [06:05] jdong: I have a couple of things ahead of it (not just pleading for the sync of bind9.4.2-10 (after I upload it...)). I'll poke someone or 3 about it this week [06:05] lamont: thank you so much [06:05] jdong: I mean, nfc if it'll have any more effect that your efforts... === asac_ is now known as asac [06:20] hrm... so do I mark a bug 'fix released' when I upload to debian, or when it's synced into hardy? [06:21] When it's sync'd, surely. You could mark it as 'fix committed' when you upload to Debian, probably :) [06:22] nah, it get's "fix committed" when the fix is, uh, committed. (to the vcs) [06:23] lamont: well, I suppose the LP ideal would be that it would be Fix Released on a Debian task and probably Fix Committed on an Ubuntu task [06:25] lamont: you put the LP: line in the Debian changelog, and let the sync close the bug ;) [06:25] slangasek: did that. :) [06:25] then let LP take care of the rest? [06:26] slangasek: yeah - that was my plan.. the only thing it doesn't take care of is figuring out how to trigger the sync. :-) [06:26] but I'll worry about that after dinstall tomorrow [06:26] ah, heh [06:27] the upload is 2 ubuntu bugs, zar0 debian bugs. \o/ [06:27] slangasek: do you think you could push flashplugin-nonfree from the gutsy unapproved queue? [06:32] LaserJock: can you tag the archive admin on duty, please? :) [06:33] I guess [06:34] I was just trying to be as quick as possible [06:34] but I can sub ubuntu-archive [06:36] actually I won't, that'd be mean [06:36] I'll poke in the morning if it's not done [06:36] night all [06:37] * lamont -> sleep [06:38] sleep is banned [06:42] slangasek: oh right, vdr-plugin-pictures.. totally harmless, it seems :) [07:11] good morning [07:14] Good morning [07:15] morning === kagou_ is now known as kagou [07:22] Good morning [07:31] is there supposed to be a konqueror-kde4-dbgsym package? i don't see one. ditto for konqueror-nsplugins-kde4-dbgsym [07:44] 5~/c [07:44] ugh [07:45] :) [07:46] kagou: bug #208419> did you ever install libpam-smbpass? [07:46] Launchpad bug 208419 in pam "Integrate samba password in PAM" [Wishlist,In progress] https://launchpad.net/bugs/208419 [07:46] ah, apparently not [07:46] slangasek, i finally installed libpam-smbpass [07:46] :) [07:46] did it ever work? [07:47] kagou: right - the pam change makes it optional, so that PAM itself continues to work with or without libpam-smbpass installed [07:47] but now i'm having problems with samba share... [07:47] so i'm investigating [07:48] kagou: and making libpam-smbpass be installed by default is an additional change to make in support of this, but I want to be darned sure the PAM changes are right first [07:48] indeed [07:50] mmm apparently, since i'v installed libpam-smbpasswd i can not enter in my PC from a windows client. An authentification is needed [07:51] what do you mean by "enter"? [07:53] BalaamsMiracle: I'm planning to import translations for ubuntu-docs in the next 2 days, time permitting [07:53] from windows, i browse PC, and when I enter (double-click) in my PC name (to see shared folders), now an athentification is required [07:54] i purge my system and recheck all [07:57] slangasek: is it going to be ok if the final ubuntu-docs upload with imported translations is ready by this weekend (i.e. a couple of days later than the freeze)? [07:57] mdke: yes [07:57] slangasek: thanks [08:06] slangasek, i suspect patched pam packages to interfer with samba server [08:07] i'm installing a fresh daily image (to ensure that all is clean). [08:07] brb [08:07] kagou: with or without libpam-smbpass installed? [08:08] without [08:09] I'd want to see the share config [08:09] slangasek, it's the default, installed by samba package [08:09] the samba package doesn't ship with any shares enabled by default... [08:10] indeed, i making share using usershare system (nautilus-share). This way smb.conf is not modified [08:10] ok, and is it set to allow anonymous access? [08:11] what ? smb.conf or usershare folders ? [08:11] the usershare folder [08:12] no, set to authentification [08:12] so where's the bug? [08:13] slangasek, i can not enter in my PC (from windows samba browser) to see shares (printer/filders) [08:14] sorry i must go for 10min [08:14] brb [08:14] ok, I don't see why you expect this to work currently. You've said that libpam-smbpass wasn't installed, so no Samba password entry was set up... so why would you expect it to work yet? [08:44] seb128, lut. do you have problem on launching gdmsetup. Here it tkae long time to launch, an harddrive have a lot of activity [08:45] kagou: hi, no, but we have several report about it being slow on the first run after installation, the harddrive activity is interesting, could you strace it? [08:47] slangasek, to resume. On a fresh/clean install/with samba/a folder shared for guest with nautilus-share (usershare). Browsing samba from a windows XP client is done like this : Enter workgroup/see PC/enter PC/see shares/ enter shares. Now that i'v played with pam pached/libpam-smbpass, ica n do browse/entre workgroup/see PC/CAN NOT entre PC [08:47] seb128, sure, i'm finishing a fresh install [08:48] seb128, a "strace gdmsetup > info" is enough for you ? [08:48] kagou: erm. if the folder(s) that you're sharing require authentication, and you're getting a prompt to authenticate....what, exactly, is the problem? [08:48] kagou: you appear to think that you should not have to authenticate, even though you've set it up so it requires authentication. [08:48] Hobbsee, i'm getting a prompt to enter in the PC, not in the share [08:49] this is the problem, as i can't see pdf/folder shares [08:49] this is the problem, as i can't see printers/folder shares [08:49] you're trying to connect to a windows box, i assume? [08:50] from a windows box, to a Samba server [08:50] from a windows XP client [08:50] oh right [08:51] kagou: why would you expect to have to have to put in your password at the physical samba share, then? [08:51] kagou: that sounds perfectly normal to me that you would be asked to authenticate when connecting to the server [08:53] having to bash down the server room door, so you can type in a password, so you can pull files from the server room back to your windows machine seems...backwards. [08:54] ...? [08:55] which appears to be what kagou is saying? [08:55] * Hobbsee gives up [08:55] he just wants to see what shares are broadcast [08:55] instead of actually getting into a share [08:55] not sure if you need to auth to do that [08:55] I believe he's saying that he gets a password prompt trying to browse the host from windows [08:55] which is normal [08:55] well, yes... [08:55] slangasek, no it's not normal [08:55] ah, right. so that's what he's saying. [08:56] er, but it is... [08:56] :p [08:56] * mvo waves to xtknight [08:56] sorry but my english is poor and sometimes i'v difficulty to explain well [08:56] mvo, oh yeah i had another bug to talk to you about [08:56] Bug 212542 [08:56] Launchpad bug 212542 in compiz-fusion-plugins-main "Drop type=Utility from 01-animation-defaults.patch" [Low,Triaged] https://launchpad.net/bugs/212542 [08:57] get my earlier messages? [08:58] kagou: when you try to connect to an SMB server, the server will request credentials from the client even if all you're doing is browsing it - and different shares may be visible depending on the credentials used [08:58] so yes, you will get a password prompt when browsing the host, unless your Windows client already knows credentials to use [09:00] slangasek, since today, connecting SMB server do not requier password prompt, and so i can see shares (printers/shared folders) [09:01] head-->desk. [09:01] (mine) [09:03] xtknight: yeah, that one is scheduled for today [09:05] mvo, oh ok i didn't even bother to check the page. looks like you've got that taken care of, then [09:05] i'm off to get sleep, laer [09:05] later [09:06] kagou: by "since today", do you mean "before today"? [09:06] thanks xtknight [09:06] mvo, thanks for dealing w/ the bugs, your job seems heftier than mine [09:08] slangasek, indeed [09:09] kagou: what's the value of 'map to guest' in /etc/samba/smb.conf? [09:11] slangasek, i can't say this now. Let me finish install :) [09:20] asac, if i dont use dhcp, can it be that NM doesnt show the network at all ? [09:20] (the wlan i mean) [09:21] ogra_cmpc: yes, when you use static ip, NM will pretend as if it doe snot know what the status is. :-) [09:22] ogra_cmpc: if you have a config in interfaces, NM will not manage that device [09:22] asac, no, thats different [09:22] ogra_cmpc: it will instead assume that you are online (so apps don't come up in offline mode) [09:22] asac, RichEd doesnt have a dhcp server in his network ... seems he doesnt see his wlan on the cdlassmate [09:23] ogra_cmpc: if the interface is managed, network manager cannot know about dhcp or not dhcp before it actually associates. so its unrelated for sure [09:23] ogra_cmpc: maybe he has a hidden SSID [09:23] no, he said he doesnt [09:23] sudo iwlist scan [09:23] (was my first question as well :) ) [09:24] slangasek, i'v a meeting. I'll be back in 1 hour [09:24] ogra_cmpc: if manually scanning doesn't show the network its a driver or AP issue [09:25] but since the cmpc works in general, i'd guess that its an AP issue for him [09:25] asac, i would guess APv since i havent seen the cmpc break on wlan at all with hardy yet [09:25] ogra_cmpc: maybe he uses a standard not supported by the cmpc driver? [09:25] like abg [09:25] and had no reports from others either [09:27] asac, abg ? wtf ? [09:28] if he does, he doesnt know i guess :) [09:29] ogra_cmpc: my router provides options to select modus "g or b" [09:30] ogra_cmpc: maybe some chipsets do not yet support one of abg [09:30] thats the only idea i currently have [09:30] i dumped all of that on him ... lets see what comes out === davmor2 is now known as davmor2_away === ubiq_ is now known as lucid [10:23] slangasek, still here ? [10:23] kagou: only barely [10:23] slangasek, with a fresh clean install+samba, browsing from a windows client let me see shares on my ubuntu samba server without password [10:24] i see "PDF" + "Printers..." [10:24] kagou: this is with 'map to guest = bad user' in /etc/samba/smb.conf? [10:24] yes, this is the default in smb.conf [10:25] now i will install patched pam + libpam and try if this is changing something [10:25] so the username you've used to log in to the Windows client is one that doesn't exist on the server, right? [10:25] indeed [10:26] ok; in that case, this is the expected behavior. Yes, please let me know if something changes just by installing the new pam package [10:26] slangasek, is there files to save before installing patched pam and libpam-smbpass ? [10:27] the only files that will be changed are /etc/pam.d/common-auth and /etc/pam.d/common-password [10:27] please install *just* the patched pam first [10:27] and test [10:27] thanks [10:27] ok [10:27] and then install libpam-smbpass, and test again [10:28] 3 packages to upgrade... [10:28] ->reboot [10:29] I have a more complicated PAM configuration on my own machine rather than the proposed default, but I believe that the tests I've done with the local samba are consistent and correct [10:31] slangasek, works the same [10:31] calc: pitti: Does anyone of you mind if I reopen the bug 205923? [10:31] Launchpad bug 205923 in openoffice.org "openoffice in hardy doesn't build on powerpc" [Critical,Fix released] https://launchpad.net/bugs/205923 [10:31] i install libpam-smbpass [10:32] reboot [10:34] slytherin: how does that differ from bug #214042? [10:34] Launchpad bug 214042 in openoffice.org "[Ubuntu] [hardy] OOo FTBFS on powerpc due to timeout" [Critical,In progress] https://launchpad.net/bugs/214042 [10:36] slangasek: I didn't log any of those two. The timeout ones seems to be logged yesterday only. [10:37] slangasek, seems ok too. [10:37] slangasek: and timeout is the reason I was going to reopen first one. But anyway since calc himself logged another bug and looking into it, I will not intervene. :-) [10:37] i add a usershare for guest now [10:38] mdke: Thanks for the info. But is there an official schedule after which the translations for all documentation are imported? Like once a month, one every 3 months, once a release? [10:40] slangasek, damn seems to work now [10:40] slangasek, i wil add new user, use it to share a folder with password [10:42] anyone here in the Ubuntu German translators team? I need somebody to approve the typo correction (Standardspache -> Standardsprache) in https://translations.launchpad.net/ubuntu/hardy/+source/debian-installer/+pots/debian-installer/de/61/+translate for bug 195075 [10:42] Launchpad bug 195075 in ubiquity "Typo in german translation of installer" [Low,Fix committed] https://launchpad.net/bugs/195075 [10:43] argll, windows have password+user in cache i must find how to empty this [10:43] i think i've applied twice for the german translators team but i havent been approved yet.. i guess i need to bug them in their irc channel [10:45] the Launchpad translations team tell me that the translator teams are really supposed to be translation moderator teams, and that casual translators should just use suggestions [10:45] the UI isn't clear at the moment [10:46] slangasek, i found my bug ! [10:46] ah. ok. [10:46] cjwatson: how are "suggestions" then merged? by the moderator teams? [10:46] I'm told that that's the idea [10:46] slangasek, all seems to work great, before i change my password for the first user. After that i can not more see shares [10:46] just out of curiosity, does anyone know why installing xfonts-* packages wants to have /lib/modules/$kvers/modules.dep in place ? [10:46] I suspect that a number of translator teams don't actually work this way in practice at the moment, though [10:47] FATAL: Could not load /lib/modules/2.6.24-8-generic/modules.dep: No such file or directory [10:47] cjwatson: well, i was going to translate a few of "my" apps in rosetta to my native language.. i guess i'll resort to vim then and let rosetta handle other languages [10:47] TheMuso: looks like the launchpad-integration stuff is broken - conflicting file bewteen liblaunchpad-integration0 and liblaunchpad-integration1. Please fix. [10:48] laga: don't let me discourage you [10:48] if you're doing serious chunks of translation, it's probably more effective to be in the moderator team anyway [10:48] TheMuso: wait, never mind. mirror is out of date. [10:49] cjwatson: no worries, it's just a different choice of UI for me, the translation will end up in the repos anyways. [10:49] true [10:51] ARGH [10:51] LAGA ! [10:51] laga, # do not start ltsp-client-core [10:51] chroot $ROOT update-rc.d -f ltsp-client-core remove || true [10:51] i guess i broke something. [10:52] laga, could you move that and the following update-rc.d commands inside the if statement ? [10:52] ogra_cmpc: yes. sorry :/ [10:52] so it doesnt get executed on normal ltsp ... its somewhat fatal to not have ltsp-client-core starting :) [10:53] slangasek, bring back the "old" password, and all is ok [10:53] kagou: what "first user"? At this point, I don't think this has anything at all to do with the pam changes; either your expectations for Samba are at odds with the intended behavior, or there's a bug in Samba, either way it doesn't correlate with the PAM change [10:53] sorry, i should have spotted that the fi sits to far up [10:53] first user ->user created on hardy installation [10:54] ogra_cmpc: is that bug 214481 ? [10:54] Launchpad bug 214481 in ltsp "[hardy] ltsp-build-client builds incomplete filesystem" [Undecided,New] https://launchpad.net/bugs/214481 [10:54] ogra_cmpc: just a debdiff this time? easier for you :) [10:54] laga, sure [10:54] stgraber, right, looks like it [10:55] laga, ^^^ [10:55] laga, ogra_cmpc: could you not use update-rc.d -f, which will be completely undone by any upgrades to the ltsp-client-core package? :) [10:55] slangasek, should I report this bug to samba [10:55] kagou: if you think there's a bug, yes. [10:55] slangasek: what do you suggest? [10:56] kagou: in the meantime, I'm off to bed [10:56] slangasek: deleting the init script? [10:56] slangasek, well, thst the mythbuntu plugin doing that, i.d actually appreciate to have ltsp-client-core starting on ltsp :) but i'm sure laga will take it into account [10:56] laga: I don't think there's really a proper way to do that to another package's init script [10:57] slangasek: echo "exit 0" >> /etc/init.d/ltsp-client-core ;) [10:57] slangasek, thank you for your time, i will remove pam patched and libpam-smb to test if this bug i sstill here. cya [10:57] laga, move it to K [10:57] laga: the system is designed so that you must get the other package to cooperate, rather than non-consensually fiddling with its bits [10:58] cjwatson: that's planned for +1 [10:58] upcdate-rc.d of the package will do nothing if a K link exists, just dont remove it [10:59] ogra_cmpc: so i just remove the S links. i guess for +1, i should write a blacklist function similar to the whitelist one [10:59] If any files /etc/rcrunlevel.d/[SK]??name already exist then update- [10:59] rc.d does nothing. The program was written this way so that it will [10:59] never change an existing configuration, which may have been customized [10:59] by the system administrator. The program will only install links if [10:59] none are present, i.e., if it appears that the service has never been [10:59] installed before. [10:59] from man ... === davmor2_away is now known as davmor2 [11:15] laga, oh, btw,why did you disable nbd-client ? it wont do any harm (will just fail silently if there is no configuration) [11:15] ogra_cmpc: i'm not worried about starting it, i'm worried about it getting killed when the box is shut down [11:16] ogra_cmpc: whatfor would libflashsupport require SSL? [11:16] ah, right, you should move it [11:17] hum, is it "for i in $@; do" or "for i in "$@"; do" - i wonder if i these quotation marks will affect semantics [11:18] asac, hmm, i knew that once ... its so long ago that i had to do with actual code there, let me look [11:19] laga: use "$@" or just for i; do ... ; done [11:19] communication with sound server? over ssl? [11:19] Mithrandir: thanks [11:19] ogra_cmpc: please fix it ;) [11:19] asac, whats wrong with it ? [11:19] ogra_cmpc: from the strace it looks like it invalidates file descriptors and finally crashes firefox/youtube [11:20] ogra_cmpc: go to youtube ... select a video with sound ... and push backwards and forwards ... wait till sounds start. repeat [11:20] crash! [11:20] asac, aha [11:20] laga: sh -c 'set -- "foo bar" "baz x"; printf "%s\n" "$@"' and sh -c 'set -- "foo bar" "baz x"; printf "%s\n" $@' [11:20] asac, [11:20] * built with tls support patch from Daniel T. Chen [11:20] https://code.edge.launchpad.net/~crimsun/libflashsupport-pulse/devel [11:20] ogra_cmpc: yeah i disabled that already [11:20] by default it uses OPENSSL [11:21] laga: And yeah, as Mithrandir said, plain “for i” is sufficient. [11:21] asac, originally it used libssl [11:21] that patch switches to gnutls [11:21] yeah ... but that didn'thelp [11:21] well, then its not ssl related i'd say [11:21] i even switched from PULSEAUDIO to ALSA ... in that way the crashes are even more frequent [11:21] yeah, flashsupport shouldnt be hooked between flash and alsa [11:21] ion_: thanks [11:22] ogra_cmpc: well. it was just to evaluate [11:22] its really only to work around pulse not closing its handles [11:22] err flash [11:22] not pulse [11:22] ogra_cmpc: and see where the problem might be. so given that neither SSL nor the sound output is responsible i guess that its really in the core of flashsupport [11:22] they have a worker thread for instance. [11:23] maybe that tries to access filedescriptors after firefox already closed the page or something [11:23] did you check the urls listed in the package description ? [11:23] ogra_cmpc: i don't see that purpose in the code [11:23] ogra_cmpc: for updates? Luke did. there is no new code [11:23] but probably info we dont know yet :) [11:24] * ogra_cmpc looks [11:25] ogra_cmpc: #225 Deadlock/crash in flashsupport closing stream [11:25] maybe that ticket [11:26] hmm, the only git checkin is a version change to support pulse 0.9.5 [11:26] laga: as a general rule, double-quote all parameter expansions in shell unless you know that you have an explicit reason not to do so [11:27] ogra_cmpc: yeah. thats what luke told me [11:27] damn. now gdb crashes if you run firefox from within [11:28] doko: ^^ [11:30] so that looks like there a partial fix already [11:31] 6 lines is not that much, but seems there is more to fix ... [11:33] ogra_cmpc: where? [11:33] in the ticket? [11:33] yeah [11:33] oh in the mids of the long pastes there is indeed something [11:33] * asac reading [11:34] let me try that [11:34] http://paste.ubuntu.com/6658/ [11:34] but apparently that doesnt fix it completely yet [11:37] hmm [11:38] works ? [11:39] ogra_cmpc: no, but i had a hacked version trying a clean one now [11:39] laga: (cases where you would generally not want to double-quote parameter expansions are when they're the operand to 'case', or when you want to deliberately word-split something) [11:39] ogra_cmpc: how is that supposed to fix a dead lock? [11:40] cjwatson: i once read a good explanation of all that, but i forgot half of it, hence my question :) thanks for the explanation [11:41] laga: http://www.greenend.org.uk/rjk/2001/04/shell.html is a good article by a friend of mine on the subject [11:42] ogra_cmpc: now it crashes firefox when playing the video for first time [11:42] cjwatson: thanks, bookmarked [11:44] asac, as i understand it he tries to kill the thread before it can lock up [11:45] asac: architecture specific? [11:49] asac, http://www.pulseaudio.org/ticket/225#comment:4 is the relevant info here i think [11:50] doko: i don't think so. [11:51] doko: can you try to start gdb /usr/lib/firefox-3.0b5/firefox [11:51] run [11:51] with dbgsym for xulrunner-1.9 and firefox-3.0 installed? [11:51] maybe flashplugin-nonfree and libflashsupport-dbgsym as well [11:52] ogra_cmpc: hmm. we could use a timed mutex then in shutdown [11:54] i wonder if there is simply siomething missing in FPX_SoundOutput_Close [11:54] where the heck is p DEFINED [11:54] oops [11:55] ogra_cmpc: pulse doesn't provide such a timeout lock [11:55] ogra_cmpc: unlikely that p is not allocated :) [11:55] no, i mean not freed correctly [11:56] ogra_cmpc: zeroing is done in line 943 [11:56] FPX_SoundOutput_Close frees p->mainloop, ->context and ->stream [11:56] and alloc in the line before [11:58] ogra_cmpc: i think its created in init and then only kept as user_data in the callback [11:58] ogra_cmpc: pa_context_set_state_callback(p->context, context_state_cb, p); [11:59] hmm, what about p->first ? [12:00] welll and a few more callbacks get that [12:00] so maybe we really get a call with that as data after free === smarter_ is now known as smarter [12:00] for instance _write_callback might be a candidate [12:01] maybe we can cancel those explicitly on free? [12:01] write_data breaks oif p->first is 1 and then sets it to 0 ... but it doesnt get freed [12:01] at least not explicitly anywhere [12:03]