=== tsimpson_ is now known as tsimpson [01:02] chrisccoulson: any update on the upload for bug 724324? [01:02] Launchpad bug 724324 in gnome-power-manager "FFE: Add support to g-p-m for controlling keyboard backlight via upower in natty" [Wishlist,Confirmed] https://launchpad.net/bugs/724324 [01:03] Amaranth, it's funny you should ask that. i'm just doing a test build of it now ;) [01:03] chrisccoulson: I just did a test build and verified the feature actually works :) [01:03] oh, ok. thanks :) [01:03] i'll just upload it then === smspillazzzz is now known as smspillaz [03:32] Anyone around to sponsor a partial fix for the mesa segfault that's plaguing software-centre? [03:35] RAOF, how risky is it? [03:35] i can sponsor it... but know very little about mesa :) [03:35] kenvandine: Not terribly risky. [03:35] Well… on IA32. [03:36] On amd64 it's got some shiny new assembler in the dispatch code, but that works for me and I see no reason why it won't work for everyone else, too :) [03:36] hehe.. famous last words [03:36] point me to it please :) [03:39] kenvandine: Do you have upload privs for mesa? [03:39] yup [03:39] * TheMuso can do it if kenvandine can't. [03:39] core-dev [03:39] ah ok forgot about that. :) [03:40] i avoid X as a rule though... so if you feel more comfortable :) [03:40] TheMuso, and actually it isn't midnight for you either :) [03:40] I have done it a few times, so if you don't feel comfortable... [03:40] kenvandine: THis is true. [03:40] :) [03:40] * kenvandine ducks that one [03:42] http://cooperteam.net/Packages [03:43] If you'd like to do some pre-upload testing I can push my builds there, too. [03:44] I'm happy to build it. What is the problem exactly? I am only running amd64. [03:45] The problem is bug #259219 - there's a minimal reproducing test-case on that bug. [03:45] Launchpad bug 259219 in mesa "Broken TLS support in libGL.so AKA: software-center crashed with SIGSEGV in __cxa_allocate_exception()" [High,In progress] https://launchpad.net/bugs/259219 [03:46] The mesa I've got there isn't a complete fix. Or, rather, it is a fix for that bug, but disables the optimised IA32 assembler because I haven't yet got that working. [03:47] But that bug's attracting duplicates at an annoying rate, so I think it's time to fix it with a performance regression now and keep working on making the IA32 assembler work with a little less time pressure. [03:47] Right, so does it affect amd64? [03:47] I'll trust your testing, I'll just review the patch. [03:51] It affects everything - it's using an incorrect TLS model, when libGL is dlopen'd libstdc++ doesn't get its TLS stuff initialised correctly and when something hits that codepath (stream IO and exception handling seem to hit it) it SEGVs. [03:53] The patch in the upload is only complete for amd64, though. It's *fixed* for IA32, but at the cost of a performance regression because it disables the optimised assembler. [03:55] (And there's a performance regression on amd64, too, but it's smaller and (possibly) mandatory; the proper TLS model needs to do more work ☺) [03:59] ok [04:57] RAOF: uploaded [04:58] Sweet, thanks. [05:00] np [05:26] I'm seeing the Gnome3 Team in Launchpad. Would this ppa be the best route to a native Gnome 3 experience under natty? [06:44] RAOF: Do the fglrx and nvidia libGL.so versions handle this correctly? Oh, they probably just don't use C++, right? [06:44] whoa late [07:05] Good morning [07:08] * bryceh_ waves [07:09] hey bryceh_, how are you? [07:10] pitti, good, you? [07:10] I'm great (except for aching muscles from the weekend :) ) [07:10] pitti, this evening we made a list of all the words my son can say now, and tallied 43 words! [07:11] "one word a day", heh? [07:12] we've been teaching him to say 'please' (he gets it as 'pleh'... close enough) [07:12] unfortunately I've created a monster [07:12] he and I feed the fish every night and I've been letting him help lately [07:12] so tonight we finished and he wanted to feed them again! "pleh! pleh! pleh!" [07:14] pitti, been out cycling or something? [07:14] bryceh_: no, helped a friend moving on Saturday, and on Sun we had an all-day Taekwondo camp [07:14] and last night the usual regular TKD training as well [07:15] pitti, or just sore from all the blog activity over the w-e? ;-) [07:15] I did go for some cycling, but not much [07:15] bryceh_: that's what makes my heart hurt, not my legs :( [07:15] pitti, yep too true [07:16] pitti, currently I'm mulling over bug 702090 which seems stuck [07:16] Launchpad bug 702090 in xserver-xorg-video-intel "i965gm GPU lockup if vesafb is left loaded (EIR: 0x00000010 PGTBL_ER: 0x00000100) - *ERROR* EIR stuck: 0x00000010, masking" [High,Confirmed] https://launchpad.net/bugs/702090 [07:17] pitti, this has tons of dupes and apw posted a kernel with a fix, and I've posted directions to test it but there's not been any response. It's very troubling [07:17] (no response except just scads more dupe bugs getting filed) [07:17] pitti, any advice? [07:25] bryceh_: oh, I forgot about the workaround kernel to test [07:29] Morning. [07:38] bryceh_: I have such a machine, but the lockups i was seeing stopped when mesa 7.10.1 was uploaded [07:38] used to hang when the screen blanked [07:41] hmm [07:41] not sure it's the same though [07:41] +bug [07:42] vesafb isn't loaded :) [07:42] tjaalton, yeah these particular hangs (pseudo-hangs) actually seem not to lock the gpu permanently most of the time, they just trip apport but -intel resets [07:42] it was able to revive the machine by killing X from a VT [07:42] *I [07:43] tjaalton, interesting, usually with gpu lockups that doesn't help [07:43] maybe it wasn't really locked up [07:43] of course I have no logs to prove either way anymore [07:44] dah, mystery bugs gotta love em [07:44] well, fixed by mesa anyway :) [07:47] heh speaking of freezes I notice my sandybridge box has frozen while I wasn't looking... intriguing [07:48] still waiting for sb boards to arrive to the shelves.. [07:49] got phoned by newegg that they have a replacement board for me... need to call them back [07:49] root@albagensia:/usr/share/apport# intel_gpu_dump [07:49] Couldn't map MMIO region: No such file or directory [07:49] that's odd [07:50] is the tool sb-aware? [07:50] far as I know. Hmm, nada in i915_error_state either [07:50] good morning [07:50] yes, it's sb aware; we get plenty of gpu lockup bug reports for sandybridge [07:51] didrocks, heya! [07:51] hey bryceh_! How are you? [07:51] didrocks, good, you? [07:51] bonjour didrocks [07:53] pitti: Guten morgen [07:53] bryceh_: very good, thanks :) [07:55] tjaalton, aha... just killing X and relogging in was enough to get it back. So maybe wasn't a gpu lockup at all? was running unity on it so maybe that was what froze? [07:55] bryceh_: yeah it might be that it just doesn't come back from the blanked state? [07:55] hmm actually, here it also just froze without blanking [07:56] wasn't actually blanked... the windows were displayed properly and the mouse cursor moved, just clicking or using keyboard didn't do anything [07:56] but killing X from the vt "fixed" it [07:56] fwiw, this is with the new X bits from the ubuntu-x-swat ppa [07:56] bryceh_: invisible windows maybe? [07:56] there's not much new [07:57] I should have checked if the clock updated [07:57] but doing a chvt didn't restore it, and just left the screen black [07:57] ok it was something else then [08:00] bryceh_: ugh, read the bug now [08:00] everyone <3's -intel [08:01] bryceh_: advice> I wish, but I'm afraid I don't know enough about these graphics drivers to be of help here :/ [08:01] bryceh_: so vesafb actually needs to get unloaded when intel is being loaded? [08:01] or is the problem that vesafb is getting loaded at all (instead of using direct DRM)? [08:02] I see vesafb on my dmesg, but it's not loaded anymore [08:02] pitti, well I was more looking for advice on how to stir people up to give feedback [08:02] tjaalton: ah, right [08:02] the issue itself is pretty clear - the vesafb module is loaded at boot in case no drm drivers are available, and then (supposed to be) unloaded in favor of the drm driver if found [08:03] curious that it only affects intel, though; does the radeon driver do the unloading properly? [08:03] however there's either a timing issue or something's holding onto vesafb too long (grub?) and so intel collides with it and that locks the gpu [08:03] but the bug with all its dupes should have a fair amount of subscribers now [08:04] pitti, intel is able to reset the gpu most of the time and keep on truckin' but it does trigger apport so we get a bug report (lots actually) [08:04] pitti, indeed [08:05] pitti, in fact we do see a few bugs here and there for -nouveau and -ati that seem to also be conflicts with vesafb [08:05] but at a *much* lower rate [08:05] my guess is something about how the -intel driver is catching the hang and resetting the gpu compounds the issue [08:06] or else -intel is just buggier than the other drivers [08:07] bryceh_: or radeon doesn't catch the hang via apport and thus we just don't hear about it (so much)? [08:08] pitti, possibly [08:13] Amaranth: Who knows what TLS model nvidia and fglrx use, or even *if* they use TLS. [08:13] Actually, I guess readelf would tell me. [08:15] Incidentally, is there a good GUI for gdb anywhere in the archives? Wandering through the disassembly would work much better with one, I feel. [08:16] RAOF: advertisement from my sun/orcle times: Netbeans might be a decent gdb frontend (as far as that is possible) [08:36] RAOF: I generally like nemiver quite a lot [08:36] RAOF: doesn't quite work with debugging X.org server, of course :) [08:38] Well, what I *really* want is the ability to keep the disassembly view up as I single-step through the instructions. [08:44] Hm, and register contents would be nice, too. [08:45] Looks like nemiver might be good. [08:52] RAOF: I haven't tried nemiver with assembly, but for normal C it's navigation, mouse-over values, variable view etc. are quite fine [08:52] and compared to eclipse & co it's ultra-lightweight :) [08:53] Does it handle gdb remoting? [08:53] I don't know [08:53] It's installing anyway. [08:54] RAOF: you can get a refund for it if you don't like it, anyway :) [08:54] * RAOF is slightly confused. Why doesn't ‘jmp *0x804f09c’ jump to 0x804f09c? [08:55] indirect addressing? [08:55] ... [08:55] Of course. [08:55] at least that's what * used to mean in assembly I used to look at a thousand years ago [08:55] i. e. "jump to the address that is stored at this pointer"? [08:56] Yeah, thanks. I misread that :) [08:56] Because it would clearly be too much to have indirect immediate addressing have the same syntax as indirect *register* addressing. [08:57] That would just be madness. [08:57] nonononono [08:57] RAOF: can you imagine by how much assembly programmer's salaries would drop if this stuff would suddenly get *logical*? [08:57] hey desktopers [08:58] RAOF: indirect register was something like add [ax], ..., wasn't it? [08:58] Hey seb! [08:58] You're thinking intel syntax. AT&T is (%eax) :) [08:59] RAOF: you know, my last endeavours with assembly was with Borland C++ 3.1 on DOS .. [08:59] no crazy AT&T syntax there [08:59] morning seb :-) [09:00] ahhh Borland [09:00] RAOF: actually no, it was for the Atmel 4414 class of microcontrollers; but their assembly is just awesome [09:00] regular syntax, RISC, tons of uniformly usable registers, etc. [09:01] i386 asm is crazy and annoying no matter which syntax is used. Me want SPARC for everyone. [09:01] x86-64 is much nicer. [09:01] And the ABI actually passes parameters in registers! [09:01] but i guess that battle is lost already. [09:02] RAOF: you mean like SPARC did for ages already? ;) [09:02] Sweetshark: Actually, if you'd *like* to touch some SPARC assembly, mesa could use some dispatch love. [09:02] Sweetshark: SPARC is the other architecture that has some hand-tuned assembly in the dispatch code, and I'm not planning to fix it :) [09:03] RAOF: I have not touched SPARC for some time and have no hardware for it. [09:04] who needs hardware when he has qemu-user [09:04] RAOF: Ill take a look at it once LO is perfect and there is nothing to be done there. [09:04] You'll get to play with SPARC assembly *and* learn about the TLS ABI! [09:04] Sweetshark: is that before or after the HURD release or the year of the Linux desktop? :-) [09:05] after the year of the Linux desktop, but before Duke Nukem Forever and HURD releases ... [09:05] ah, ok; always good to have a solid plan [09:05] Hey, Duke Nukem Forever has a release date! [09:06] RAOF: a release date is not a release [09:06] RAOF: but they can't release, they are still blocked on figuring out some mesa bugs on sparc [09:06] :) [09:06] * pitti STFUs and goes back to do something sensible [09:07] hehe [09:25] pitti, nice scour upload ;-) [09:26] :) [09:28] MORNING [09:28] ugh, sorry [09:29] morning [09:29] hey rodrigo_ [09:29] hi seb128 [09:29] pitti, reminding you about the meeting reminder [09:29] pitti, which jasoncwarner's email just reminded me about ;-) [09:30] seb128: heh, nice timing; I just finished fixing pygobject, was the next thing on my list [09:30] seb128: I think we need to extend our reminder chain long enough to span around the week, to never drop the ball [09:30] ;-) [09:30] hey rodrigo_ [09:33] hi pitti [10:14] Hello! Does anyone have an idea why mysql server is now showing as a login option in GDM greeter in Natty? I'm scratching my head if it's something that got lost in the last mysql merge, or something that has changed in gnome.. thanks :) [10:14] Daviey: !?! [10:15] Daviey: does /usr/share/xsessions/ have anythign mysql'ish for you? [10:15] Daviey: Is mysql run from upstart these days? [10:15] Daviey, as an user or as an option? [10:16] soren, believe so [10:16] one mo [10:16] Daviey: also, please check grep ^Session ~/.dmrc [10:16] Daviey: Using su to switch to mysql? [10:16] Daviey: Or does mysql do that on its own? [10:16] is that a session or an user? [10:16] pitti, grep for mysql returns null [10:16] ah, I guess he means user [10:16] "option" is not very obvious [10:16] Daviey: do you see it in the user list or in the sessino chooser in teh panel at the bottom? [10:16] Daviey, ck-history --frequent list it? [10:17] $ grep ^Session ~/.dmrc [10:17] Session=gnome [10:17] pitti, mysql has an invalid $home btw. [10:17] Daviey: so I guess you mean you see it as an user [10:17] /etc/passwd - mysql:x:117:128:MySQL Server,,,:/nonexistent:/bin/false [10:18] seb128, http://pb.daviey.com/6znJ/raw/ [10:18] weird, gdm has a patch to sort out users with id < 1000 [10:18] Daviey, you still didn't reply on where it's listed as an option, do you mean as an user? [10:18] pitti, I see it as a user, which i could possibly log in as [10:19] seb128, ^ ? [10:19] pitti, I think I've read someone saying that it doesn't apply this filtering to the ck-history users [10:19] gdm bug then [10:19] pitti, no, ck bug [10:19] Ah! [10:19] ck-history shouldn't list it [10:19] well bug on both I guess [10:19] Daviey: Is there any chance you've sudo'ed to mysql recently? [10:19] seb128: you mean bug 395281? [10:19] Launchpad bug 395281 in gdm "pam_ck_connector.so is called for non-login sessions" [Low,Confirmed] https://launchpad.net/bugs/395281 [10:20] pitti, yes... [10:20] Daviey: Or su'ed? [10:20] seb128, Hmm, if that was the case, wouldn't puppet also be a login user? [10:20] seb128: but 2 logins sounds very plausible for doing "su mysql" [10:20] seb128: it's not the daemon start [10:20] Daviey: Nova turns up on gdm for me, too, because we run it with su from upstart. su registers a pam session. [10:21] ah! [10:21] It's very annoying. [10:21] mysql upstart conf - http://pb.daviey.com/R3O0/raw/ [10:22] Daviey: That should be fine. [10:22] Daviey: No use of su in sight. [10:22] pitti, well in any case I didn't check but people mentioned gdm doesn't do the id filter on ck-history lists [10:22] pitti, we should maybe fix that... [10:22] right [10:23] seb128, But that doesn't explain why i don't see puppet as a login user in gdm greeter [10:23] surely if it was that, i'd see it aswell. [10:24] is it listed by ck-history --frequent? [10:24] Daviey, well as pitti said you maybe did su mysql two times on this box? [10:24] seb128, maybe as a maintainer script. :/ [10:25] I found it interesting that postgres is in /usr/share/gdm/gdm.schemas to not be shown. [10:26] bug #734910 [10:26] Launchpad bug 734910 in gdm "gdm greeter shows unwanted users" [Low,Incomplete] https://launchpad.net/bugs/734910 [10:26] is similar [10:26] (i've compared gdm.schemas on both maverick and natty, and they identical) [10:26] or bug #696038 [10:26] Launchpad bug 696038 in gdm "rabbitmq user appears in login list of users" [Undecided,New] https://launchpad.net/bugs/696038 [10:27] Daviey, well that list is a fixed list of users to ignore [10:27] but as pitti said before, all users with a system uid should be filtered out [10:27] Daviey: what am I looking for? [10:27] without having to code the list of existing users in the schemas [10:28] Mez, We were talking about the mysql user showing up. [10:31] and ? the outcome is? [10:31] Mez, Still not quite sure, but based on two bugs - looks like a possible regression in gdm. [10:32] the reason is likely that someone logged it as mysql, i.e su mysql on the box [10:32] there is a gdm bug there though as well, it seems the uid filtering doesn't work as it should on ck-history lists [10:32] we will likely use bug #696038 to track that issue [10:32] Launchpad bug 696038 in gdm "rabbitmq user appears in login list of users" [Low,New] https://launchpad.net/bugs/696038 [10:33] seb128: afaik, that doesn't work if the shell is set to /bin/false ? [10:33] (the su to the user) [10:33] Mez, there is also bug #395281 [10:33] Launchpad bug 395281 in gdm "pam_ck_connector.so is called for non-login sessions" [Low,Confirmed] https://launchpad.net/bugs/395281 [10:33] Mez, well there is different reason which could make the user be listed in the consolekit sessions [10:34] in any case we should filter users from those list out [10:34] seb128, grepping my installed packages maintainer scripts for "su mysql" [10:34] null return. [10:35] Daviey: and sudo ? [10:35] I would be quite suprised if a maintainer used sudo instead of su... :/ [10:35] empty set returned. [10:35] Daviey: you never know - some people are idiots :D [10:36] Daviey: Look in /var/log/auth, perhaps? [10:37] Daviey, it's not trivial to grep for that, you could have su -c or something or "su $USER" with user defined before etc [10:37] soren, http://pb.daviey.com/vWuL/raw/ [10:37] psst, Daviey, what pastebin are you using? [10:38] Daviey, Mez: in any case let's use bug #696038 for the fact that users with system uids are not filtered out as they should [10:38] Launchpad bug 696038 in gdm "rabbitmq user appears in login list of users" [Low,New] https://launchpad.net/bugs/696038 [10:38] having that fixing in gdm will fix your issue [10:38] it doesn't matter much how mysql got a login session counted [10:38] seb128, I grepped for " su ", that should be safe? [10:38] I thought ck-history would let you show the actual logins. Perhaps the times would provide a hint. [10:39] soren, ^ btw feel free to comment on that bug as well about your similar issue [10:39] soren, right, ck-history --log shows the list [10:40] seb128, Do you consider 734910 and 696038 (the two from above) to be dupes? [10:40] Daviey, yes [10:40] Daviey, well I will use 696038 [10:40] waiting on reply on the other bug but I will likely dup it [10:41] seb128, ok, i added my output to both incase. [10:41] * Daviey leaves it in your safe hands, if you want me for anything - feel free to ping. [10:42] Thanks for looking into it. [10:43] Daviey, thanks [10:44] Daviey, can you add you passwd lines for mysql and puppet to the bug? [10:44] it's a bit weird that mysql got listed but not puppet [10:44] that is what i thought! [10:46] I wonder if switching the UID's between puppet and mysql would make puppet show. [10:47] Daviey, what uid do they have? [10:48] are they listed in the same range? [10:48] hum, they are [10:49] I wonder if the broken user directory has something to do with it [10:49] mysql:x:117:128:MySQL Server,,,:/nonexistent:/bin/false [10:49] puppet:x:116:127:Puppet configuration management daemon,,,:/var/lib/puppet:/bin/false [10:50] well from there is needs code debugging I guess [10:50] Daviey, thanks, we will ping you back if we need extra details [10:50] seb128, thanks === Nafallo_ is now known as Nafallo [11:05] hi. [11:05] does anyone know about this bug: https://bugs.launchpad.net/bugs/735372 [11:05] ? [11:05] Launchpad bug 735372 in banshee "Rhythmbox not Banshee is set as default media player in Natty" [Undecided,New] === MacSlow is now known as MacSlow|lunch === MacSlow|lunch is now known as MacSlow [13:30] seb128: woudl you have some time to fix bug 703230 ? [13:30] Launchpad bug 703230 in pango1.0 ""rm: cannot remove `/usr/share/doc/libpango1.0-0': Is a directory" when updating to 1.28.3-4" [High,Triaged] https://launchpad.net/bugs/703230 [13:31] or rodrigo_ perhaps? ^ [13:31] pitti, can do [13:31] I guess it's also a nice little packaging learning exercise [13:31] pitti, did debian fix it now? I got some bts emails about it but didn't read the discussion [13:31] bite-size thing, but important for beta-1 upgrades [13:31] seb128: haven't checked yet [13:31] pitti, I will check [13:31] I've the emails there [13:31] seb128: merci [13:31] pitti, btw I was about to ping you for media-player-id [13:32] but seems you beat me to it ;-) [13:32] hehe [13:32] seb128: I'll sync it in half a day when it hit the Debian mirror [13:32] it's right to fix bug 735372 in libgnome2-common, yeah? [13:32] Launchpad bug 735372 in banshee "Rhythmbox not Banshee is set as default media player in Natty" [Undecided,New] https://launchpad.net/bugs/735372 [13:32] pitti, did you check if the icon thing being commented has side effects? [13:33] Laney, no, desktop-file-utils has the defaults.list in Ubuntu [13:33] seb128: it'll fix the handling of photos on mobile phones again, as a workaround to bug 481528 [13:33] Launchpad bug 481528 in gvfs "Volumes with photo MIME type do not have a camera icon" [Low,Triaged] https://launchpad.net/bugs/481528 [13:33] seb128: sorry, bug 657609 [13:33] Launchpad bug 657609 in gvfs "music players with UDISKS_PRESENTATION_ICON_NAME only get shown as photo device" [Undecided,Triaged] https://launchpad.net/bugs/657609 [13:33] seb128: ah ok, so that was just forgotten, thanks [13:33] pitti, but wasn't that added to fix wrong icon issues on some devices? [13:33] jbicha: ^^^^ [13:34] seb128: well, it's about nicer icons for some phones; I know it's a workaround, but it fixes much more than it breaks [13:34] pitti, ok, I just wanted to check it was not going to break another side [13:35] I don't remember the specifics but similar issues reversed ring a bell [13:35] seb128: I still have the optimistic hope that at some point I can get out of the "OMG RC bugs" rabbit hole, and find time to work on things like that.. [13:36] rodrigo_, btw I Cc-ed you on a gnome-media, g-c-c bug with a one liner patch, would be nice if you could commit it to git [13:37] seb128: isn't defaults.list different than the gconf default apps though? [13:37] jbicha, I didn't read the bug out of the title [13:38] jbicha, sorry in middle of other things [13:38] seb128: np [13:38] jbicha, the libgnome .gconf-defaults thing is what will be used when using the key on your keyboard [13:39] seems fine to update that one [13:39] didrocks: hm, I just tried booting the current desktop image on my wife's amd64/ATI machine, and alpha-3 back then on my dell mini 10; it still fails miserably and takes ages, I have to ctrl+alt+f1 and kill -9 -1, and re-login :/ [13:40] the bug is about gnome-default-applications-properties [13:40] but I guess that is from libgnome2 too [13:40] Laney, right, that uses the gconf key [13:40] pitti: so you have a crash fie or compiz just hangs? [13:40] Laney, jbicha: so yeah it's a one liner in libgnome's gconf-defaults [13:40] cool [13:41] didrocks: I also see apport after a while; I get nautilus, but nothing else (aside from the apport window) [13:41] but I guess desktop-file-utils wants doing too [13:41] didrocks: the long hang might just be gnome-session, of course [13:42] pitti: what do you see in .xsession-errors? [13:42] seb128: not trying to be difficult but it looks like Ubuntu has modified /usr/share/gconf/defaults and not libgnome directly [13:42] Laney, right, though there was still discussions on that, currently totem is the default handler there [13:43] jbicha, libgnome's debian directory has a libgnome2-common.gconf-defaults [13:43] I see rb for audio-cdda audio-dvd audio-player [13:43] didrocks: I don't know, I shut down again; no need to debug it now, just pointing out that we need to fix this by beta-1; three out of three (different) machines failing doesn't sound good, and it failed for Riddell in a3, too [13:43] Laney, right, those likely need an update, we are still not decided on what to do for mimetypes though which delayed the update [13:43] sure [13:43] pitti: yeah, but right now, there is no hang report from what I see, some crashes at starts that smspillaz have a look on AFAIK [13:44] * Laney has no idea what an audio DVD is [13:44] Laney, though it's not likely to make a difference since there is only one application handling those in the default install [13:44] didrocks: it probably just took long because gnome-session timed out on finding the WM? [13:44] Laney, it's only revelant when you have different options [13:45] seb128 & Laney : thanks for the pointers [13:45] jbicha, the defaults directory is what is used for changing defaults, it writes in a gconf directory used before the upstream values [13:45] jbicha, it's a way to overwrite values without changing the upstream schemas, so if you don't like the ubuntu customization you just need to edit the defaults or to clear it [13:46] pitti: the gnome-session timeout is only on the helper (which is the session tester) [13:47] seb128: what actually does consume this defaults.list? I thought nautilus handled mimetypes through desktop files [13:47] Laney, the defaults.list just tells nautilus or gio what to prefer when you have 2 options for a type [13:47] pitti: if it fails after 5 seconds, it fallbacks. With the current code, it should fallback to compiz (still trying to launch unity for now) [13:47] ah right ok [13:47] Laney, i.e several desktop claiming the same type [13:47] without that it would pick a random one [13:48] didrocks: does it actually work on your laptop(s)? [13:48] (the desktop live system, I mean) [13:48] pitti: I have two netbooks where it works and my laptop. I didn't try the desktop live system though [13:48] pitti: still having too much code right now to file in unity to give proper test :/ [13:49] didrocks: yeah, the second, third, etc. startup all work fine; only the first one in the live system [13:49] didrocks: is the test result cached somewhere? [13:49] pitti: there is a nux bug that I tagged priority on the startup performance [13:50] pitti: no, but I think that all libraries are cached with ureadahead, isn't it? [13:50] didrocks: they are, yes [13:50] but loading those shouldn't take 20 seconds.. [13:50] so I think nux is loading them faster [13:50] hum… there is no other cache [13:50] didrocks: I'm booting usb stick from today's image on the mini 10 [13:50] didrocks: I mean like caching the result in gconf or so [13:50] we don't [13:51] ok, thanks [13:52]