[00:03] vuntz: re [00:04] vuntz: around? [00:05] seb128: yes. Tell me it crashes :-) [00:05] should I lie? ;-) [00:05] it doesn't crash [00:05] but it displays no weather now [00:06] I think he doesn't detect my current location [00:06] s/he/it [00:06] hrm [00:06] it gives me a "set" button next to each location in the list [00:06] isn't this expected? [00:06] no [00:06] I had one location which has a house icon before [00:06] ah [00:06] and some weather [00:07] that's sad :-) [00:07] I'm homeless now due to you! [00:07] I want my home back ;-) [00:08] seb128: can you go in applets/clock/ and compile system-timezone.c with the instructions in the header? [00:08] then run it [00:09] does't work, hum [00:09] where should SYSTEM_ZONEINFODIR etc be defined? [00:10] in system-timezone.h [00:10] system-timezone.c:348: erreur: «SYSTEM_TIMEZONE_ERROR» undeclared (first use in this function) [00:10] etc [00:11] hum, weird [00:12] ah [00:12] move the include for system-timezone.h [00:12] it's in the #ifdef while it shouldn't [00:13] ah, indeed [00:13] $ ./system-timezone [00:13] Europe/Paris [00:13] which is correct [00:14] hum [00:14] interesting [00:14] my locations use the luxembourg timezone [00:14] let me try the hardy version ;-) [00:14] ah [00:15] I know [00:15] you used the old clock applet [00:15] which didn't modify /etc/timezone [00:15] and why did gnome-session stopped replying [00:15] so you have /etc/localtime which is set to Luxembourg but /etc/timezone set to Paris [00:15] ah, likely [00:15] stop replying? [00:16] gnome-session-remove --list just hangs [00:16] idem when trying to remove gnome-panel or nautilus [00:16] anyway no big deal there [00:17] ok, the set button updated /etc/timezone correctly [00:17] and I have some weather now [00:17] still no home icon though [00:17] weird, I have a home icon here [00:17] ah [00:17] add your location again [00:17] that's because I had metz and luxembourg using the luxembourg timezone [00:17] it's probably set to luxembourg too in gconf ;-) [00:18] after removing luxembourg it's working [00:18] good [00:18] I've a home again ;-) [00:18] bouhouh [00:19] adding locations doesn't work [00:19] err, I didn't touch this [00:20] well, I do pick a city, it selects the right timezone, I ack the dialog and the list doesn't change [00:20] it works here [00:20] weird [00:22] nop, restarting the panel makes no difference [00:23] I'll just pretend it's your fault ;-) [00:26] vuntz: you broke it! [00:26] vuntz: I downgraded to the hardy version which works fine to confirm and upgraded again and still the same issue [00:27] nothing printed on the command line though [00:28] seb128: the gconf key is not changed? [00:28] what gconf key? [00:29] vuntz: /apps/panel/applets/clock_screen0/prefs/cities has only one location [00:29] so it's not saved [00:29] hrm [00:29] do you want me to gdb break on some function? [00:30] seb128: yep [00:30] seb128: run_prefs_edit_save() [00:31] it's possible that info is NULL and that you just return [00:33] vuntz: Breakpoint 1, run_prefs_edit_save (button=0x83dd440, cd=0x8394300) at clock.c:3042 [00:34] vuntz: [00:34] (gdb) p info [00:34] $2 = (ClockZoneInfo *) 0x0 [00:34] vuntz: indeed [00:35] hmm, it might not be initialized yet? Can you do some ,? [00:35] vuntz: [00:35] 3064 if (!info) { [00:35] (gdb) [00:35] 3065 edit_hide (NULL, cd); [00:35] (gdb) p info [00:35] $2 = (ClockZoneInfo *) 0x0 [00:35] (gdb) n [00:35] 3106 } [00:35] (gdb) [00:35] 0xb771da4f in g_cclosure_marshal_VOID__VOID () from /usr/lib/libgobject-2.0.so.0 [00:35] ok [00:35] so, that's bad [00:36] (gdb) p *zones [00:36] $4 = {g_object = {g_type_instance = {g_class = 0x83f6b20}, ref_count = 1, qdata = 0x0}} [00:36] (gdb) p *timezone_l10n [00:36] $5 = 69 'E' [00:37] E? [00:37] heh [00:37] vuntz: sorry [00:38] (gdb) p timezone_l10n [00:38] $6 = (gchar *) 0x844a2e8 "Europe/Paris" [00:38] vuntz: that's already a pointer ;-) [00:39] hmm [00:42] wait [00:42] it's fine [00:43] is it? [00:43] timezone_l10n is a string, so we're good [00:44] what's not good is that you don't get a zone with Europe/Paris [00:44] still info == 0x0 [00:44] right [00:48] vuntz: the hardy version returns correctly a non null value there [00:48] I just verified [00:50] seb128: can you patch with http://pastebin.com/d66179342 [00:51] you should get the name of all timezones when you click ok [00:54] vuntz: no, seems to not be called [00:55] Breakpoint 1, clock_zonetable_get_l10n_zone (this=0x8314228, l10n_name=0x86ecfd0 "Europe/Paris") [00:55] at clock-zonetable.c:370 [00:55] 370 { [00:55] (gdb) n [00:55] 372 } [00:55] (gdb) n [00:55] 373 [00:55] (gdb) n [00:55] run_prefs_edit_save (button=0x8386b40, cd=0x8393c88) at clock.c:3061 [00:56] eerrrrr [00:56] maybe the hash table is empty? [00:59] vuntz: likely, I see no other reason to have nothing printed there [00:59] vuntz: (gdb) p *zones [00:59] $3 = {g_object = {g_type_instance = {g_class = 0x83f7878}, ref_count = 1, qdata = 0x0}} [00:59] that's from before [01:00] is qdata supposed to be the datas? [01:01] no idea [01:01] never looked at how the data is structured in a gobject [01:03] seb128: you didn't get any output at all? [01:03] no [01:03] I copied the gdb log some lines ago [01:03] you should at least get some when clicking on the add button [01:04] (before choosing a location) [01:04] no [01:05] weird [01:05] and the combo is not empty? [01:06] ah [01:06] the timezone combo is empty indeed [01:06] I didn't notice because selecting a city make the right timezone be written in the next entry [01:06] so I didn't click on the combo before [01:07] ah, interesting [01:07] can you put your /usr/share/zoneinfo/zone.tab somewhere? [01:08] vuntz: http://people.ubuntu.com/~seb128/zone.tab [01:10] looks like what is expected [01:10] seb128: can you br on clock_zonetable_load_zonetab() and look at what's going on there? [01:11] something will probably fail in clock_zonetable_parse_info_line() [01:14] vuntz: never called? [01:14] seb128: only at startup [01:15] gdb doesn't stop on it [01:15] I tried removing the static no difference [01:15] really? [01:15] gra [01:15] sorry, I was running the wrong version [01:17] it's working now?! [01:18] seb128: I blame you [01:18] ok, interesting [01:18] vuntz: it's working when building using -O0 apparently [01:18] but not when using -O2 [01:19] mmh [01:19] init issue? [01:19] br clock_zonetable_constructor [01:19] I'm seeing a small buglet there, but it shouldn't hurt you in this case [01:20] ok, this function is called [01:20] what do you want to know there? [01:20] Breakpoint 1, clock_zonetable_constructor (type=138542584, n_construct_properties=0, construct_properties=0x0) [01:20] at clock-zonetable.c:91 [01:20] 91 if (obj) [01:20] (gdb) n [01:20] 94 obj = G_OBJECT_CLASS (clock_zonetable_parent_class)->constructor ( [01:20] (gdb) [01:20] 99 clock_zonetable_load_zonetab (CLOCK_ZONETABLE (obj)); [01:20] (gdb) [01:20] 267 ClockZoneTablePrivate *priv = PRIVATE (this); [01:20] (gdb) [01:20] 273 priv->table = g_hash_table_new_full (g_str_hash, g_str_equal, [01:20] (gdb) n [01:21] 267 ClockZoneTablePrivate *priv = PRIVATE (this); [01:21] (gdb) [01:21] 273 priv->table = g_hash_table_new_full (g_str_hash, g_str_equal, [01:21] (gdb) [01:21] [Thread 0xb5bfdb90 (LWP 20643) exited] [01:21] 0xb7eff410 in __kernel_vsyscall () [01:22] seb128: you don't enter the while loop in clock_zonetable_load_zonetab? [01:22] no [01:23] I don't understand what's going on [01:23] it's returning from clock_zonetable_load_zonetab() after the first g_hash_table_new_full() ??? [01:23] how is it possible? [01:30] vuntz: ah, better now [01:30] what did you change? [01:30] vuntz: it does loop on g_io_channel loop [01:31] vuntz: rebuild using -01 and -ggdb [01:31] ah [01:31] still get the empty combo list [01:31] good [01:31] can you look at what's going on in clock_zonetable_parse_info_line()? [01:33] looking [01:39] vuntz: not finding anything wrong [01:39] I think I'll continue tomorrow [01:39] it iters over the locations correctly [01:40] and the parsing seems to be alright [01:40] weird [01:40] the night will help [01:40] maybe it breaks on one location or something [01:48] vuntz: ahhh [01:50] vuntz: I found it! [01:51] seb128: ah? [01:52] vuntz: http://paste.ubuntu.com/6684/ [01:52] vuntz: that fixes it ;-) [01:53] heh [01:53] stupid bug [01:53] indeed [01:54] thanks [01:54] vuntz: you are welcome ;-) [01:54] ok, so we can continue tomorrow if you have other things to try [01:55] yep [01:55] I'll get some sleep now I think ;-) [01:55] bon repos :-) [01:55] vuntz: bonne nuit ;-) [02:14] kwwii: uploading u-wallpapers === asac_ is now known as asac [06:49] good morning [07:59] bonjour [09:15] seb128: do you happen to know if/why apps/metacity/global_keybindings/switch_windows_backward is not set anymore in current hardy? [09:15] good monring btw :) [09:15] hey mvo [09:15] mvo: only synaptic remains to upload for the launchpad-integration soname change ;-) [09:17] (looking for your questions) [09:18] *cough* [09:19] ok, I will do that this morning [09:19] thanks [09:19] I had to write my activity report ... [09:19] I would have done it already if it was not in bzr [09:19] where is yours ;)? [09:19] doh, still have to do that [09:19] I helped vuntz debugging a gnome-panel issue until 3am, short night again [09:20] I'm not really on speed yet ;-) [09:20] just teasing :P [09:20] mvo: seems that this gconf key was not set in gutsy either [09:21] seb128: can we fix it please? I'm happy to do it [09:21] how fix? [09:21] set one you mean? [09:22] yeah, set it to sift-alt-tab (as it was before) [09:22] before being? [09:22] gutsy had no such keybinding [09:24] give me a sec I can check [09:24] but I see no rationale why it went away, its quite common, I think it works like this on win too [09:24] and its reasonable obscure that people do not hit it accidentally :) [09:24] mvo: it didn't went away, I'm checking svn it has always been disabled [09:24] oh? [09:24] hmmmmmm [09:29] Amaranth: you mentioned that git has a fix for the panel-sometimes-hidden bug in git, do you happen to know if its in compiz or in the plugins? [09:30] * mvo twiddles his thumbs while waiting for seb [09:30] mvo: http://cgit.compiz-fusion.org/fusion/plugins/expo/diff/?id=503e4eb7ac18fee1d6137674ceb0da65c1323d21&id2=176bdee796b9824509a4d80dbc1778b5598f5628 [09:30] mvo: it's a bug in expo's code to hide panels when in expo mode [09:31] it thinks it is in expo mode on start [09:31] Amaranth: great, thanks! [09:34] mvo: hey we almost know what causes weird shadows and such on geforce 8xxx cards [09:34] oooohhhh? [09:34] tell me more! [09:34] it is something lrm does to the nvidia driver, installing from .run fixes the problem [09:35] originally we thought it was dh_strip [09:35] but that doesn't seem to be the complete fix [09:35] but the files nvidia-glx-new installs do not have the same md5sum as the ones from nvidia's installer so there is obviously something wrong [09:35] woah, we need to milestone that one [09:36] i did [09:36] I have a 8300 but don't see the bug (I guess my card is too low level) [09:36] re [09:36] bug 214836 [09:36] Launchpad bug 214836 in linux-restricted-modules-2.6.24 "lrm incorrectly installs nvidia-glx-new" [High,Triaged] https://launchpad.net/bugs/214836 [09:36] seb! [09:36] seb128: ! [09:36] mvo: ! [09:37] mvo: so? ;-) [09:37] seb128: :) [09:37] seb128: dapper has a owrking alt-shift-tab [09:37] maybe by accident, I'm checking that now [09:37] mvo: the schemas was not set though [09:37] yes, I see the same here [09:37] maybe it used to code the action when no other value was set or something [09:39] seb128: same for gutsy [09:40] mvo: ask to marnanel when he's around ;-) [09:49] morning everyone [09:49] lut seb128 mvo [09:49] morning people [09:50] 'lu huats [09:50] hey Nafallo [09:50] Nafallo: evince issue fixed? [09:51] seb128: not sure. wrong laptop at this job, [09:51] mvo: so i'm guessing sabdfl doesn't like his shift-alt-tab being broken? noticed he jacked the priority up on that bug :P [09:52] Amaranth: indeed, that one will get fixed very soon [09:52] i've never lost that particular config while upgrading and i don't have the ability to do upgrade tests in anything other than vmware which isn't useful for testing compiz [09:53] apparently you can make vmware do opengl, I haven't really explored it yet, but it seems to be possible [09:54] and we need to blacklist the rs350 (either that, or the card is broken, both is possible) [09:54] you can't make it do aiglx :P [09:55] and only workstation does opengl, server has 'experimental' support which basically means broken [09:55] at least last time i checked [09:56] aha, you have server? [09:57] workstation is expensive :P [10:01] seb128: not reproducable on the vaio at least. [10:02] ok, thanks for testing [10:47] seb128: do you know about file corruption when copying a file from a gvfs-gphoto2 location? :) [10:48] slomo_: no, we don't build gvfsd-gphoto2 in hardy ;-) [10:48] i know, but you did for some time :) [10:48] right [10:48] I don't think many users tried that [10:48] or we would have received extra bugs about opening photos not working ;-) [10:49] ok :) i'm just wondering whether http://bugzilla.gnome.org/show_bug.cgi?id=526794 is "my" bug (i.e. in the gstreamer gio plugin) or just the gphoto2 backend being broken :) [10:49] Gnome bug 526794 in gst-plugins-base "[giosrc] totem doesn't work with some gvfs backends" [Normal,Unconfirmed] [10:50] no idea about this one [11:36] seb128: happy with the clock applet? [11:36] seb128: I have another small tarball for you to test before I release [11:37] vuntz: seems to be alright, I'll go for lunch soon but I will give it a try after lunch [11:37] seb128: http://www.gnome.org/~vuntz/tmp/gnome-panel-2.22.1.2.tar.bz2 [11:37] seb128: try to measure how much time it takes to open the calendar popup before & after this tarball [11:38] ok [11:40] seb128: are you having the libglib debug symbols not useful bug in ubuntu too? [11:40] I think yes [11:40] but the dbgsym works correctly [11:41] not sure why [11:41] interesting [11:41] doesn't make much sense to me :) [11:41] i want my debug symbols back :) [11:49] vuntz: seems to work correctly, adding locations, removing those, changing the local timezone using the set button, etc [11:50] vuntz: I'm away for lunch but I can give it extra testing after that if you want [11:50] vuntz: btw would be nice to copy the location entry text when clicking on the find button ;-) [11:51] vuntz: so user who type something and click on find don't have to enter the same text again [11:51] vuntz: I did that the first time I used the applet :-) [11:51] I did too [11:51] there are too much dialog btw [12:07] crevette, seb128: done :-) [12:08] mici [13:27] * seb128 hugs vuntz [15:01] vuntz: the clock applet opening is blasting fast now, you rock ;-) [15:42] * mpt accidentally adds six copies of Deskbar to his panel [15:42] What is the point of that? [15:43] What is the point of having more than one copy of an applet in the panel? [15:43] mpt: some people like to have 4 cpufreq applets, eg [15:43] for the weather one it is handy [15:43] pre-intlclock anyway [15:57] asac: sorry, I subscribed you to bug 214192 by mistake... I wanted to subscribe you to bug 215016 instead [15:57] Launchpad bug 214192 in liferea "liferea-bin crashed with SIGSEGV in _nss_wins_gethostbyname_r()" [Medium,Incomplete] https://launchpad.net/bugs/214192 [15:57] Launchpad bug 215016 in liferea "Liferea (Hardy) crashes on startup" [Medium,New] https://launchpad.net/bugs/215016 === cody-somerville_ is now known as cody-somerville [16:01] pochu: since when? [16:01] asac: sorry, since when what? [16:02] pochu: does it crash for everyone? i remember that it started yesterady or so [16:02] * asac starting liferea [16:03] ok not a problem here. [16:03] asac: ah. It doesn't crash for me, but I haven't been updating my system very regularly these days... [16:03] he said his system is up-to-date [16:03] pochu: : /usr/lib/python2.5/lib-dynload/time.so: undefined symbol: PyExc_ValueError [16:04] yes [16:04] for me it looks more like thats his problem [16:04] I get that too, but it doesn't crashes [16:04] and if you unset LD_LIBRARY_PATH you won't get that error [16:04] really? i only get: [16:04] $ liferea [16:04] Registering '@mozilla.org/module-loader/python;1' (libpyloader.so) [16:04] Registering '@mozilla.org/network/protocol/about;1?what=python' (pyabout.py) [16:04] (or if you launch liferea-bin directly) [16:04] thats all [16:05] hmm, I don't get it anymore... [16:05] you set LD_LIBRARY_PATH? thats definitly not good [16:05] bug 211479 [16:05] Launchpad bug 211479 in liferea "Hardy: liferea gives errors upon launch" [Undecided,Confirmed] https://launchpad.net/bugs/211479 [16:05] if liferea script still set that we should fix it [16:05] the script sets it to LD_LIBRARY_PATH=/usr/lib/xulrunner-1.9b5 [16:06] (not hardcoded though) [16:07] pochu: how does it detect that? [16:07] * asac looking [16:08] The ImportError for time.so is known and old, definitely not a crasher though. [16:08] oh right. thats a preventive measure because of sqlite bustage [16:08] right [16:08] but I don't get it anymore... dunno why [16:09] Amaranth: if you have a moment, could you please check out https://bugs.edge.launchpad.net/ubuntu/+source/compiz/+bug/206866 ? [16:09] Launchpad bug 206866 in compiz "compiz.real crashed with SIGSEGV in updateWindowAttributes()" [Medium,Triaged] [16:09] pochu: maybe the reporter still has a xulrunner-1.9b4 directory that has libsqlite3.so in it (unlikely) ? [16:09] he's reported it today, so probably not... I'll ask him to make sure anyway [16:10] thanks ... attaching strace -f -eopen can't hurt either [16:11] pochu, ImportErrors occur only once, just after a xul upgrade [16:11] fta: ah, ok [16:13] asac: ok, done [16:13] thanks [16:13] i'm no python expert so i haven't investigated much into that [16:15] pochu: i asked for eventual a11y tools [16:15] because it crashes in acessibiliy [16:31] seb128: could you sponsor bug #205749 [16:32] the .rules move was already reviewed by pitti [16:32] ubotu failed to get the data :( [16:32] https://bugs.edge.launchpad.net/ubuntu/+source/libmtp/+bug/205749 [16:32] Launchpad bug 205749 in libmtp "libmtp package doesnt install libmtp.fdi" [Medium,Confirmed] [16:34] i've subscribed pitti but i believe he is away [16:42] xhaker: subscribe ubuntu-main-sponsors [16:42] xhaker: I'll have a look [16:43] seb128: Could you sponsor a GPM while you're looking that way? :) [16:43] ted1: alright, in your ppa? [16:44]