[00:05] <broder> hmm...this run is much better, but disk utilization is still really high after ureadahead finishes
[00:46] <slangasek> if I look at broder's bootchart, will my X server crash again
[00:48] <RAOF> slangasek: Mine didn't, so I don't think yours will :)
[00:48] <broder> did...one of my other bootcharts crash your X server?
[00:48] <RAOF> Firefox loves itself some huge X pixmaps.
[00:49] <slangasek> broder: you have non-trivial I/O from fsck and update-motd-updates-available; is this reproducible across boots, or did you just get lucky with this bootchart that everything needed to be done at boot?
[00:49] <slangasek> broder: no, someone else's bootchart did
[00:49] <slangasek> it's an X server bug
[00:49] <slangasek> (not a firefox bug at all)
[00:49] <broder> slangasek: that was my second reboot after installation, so it may have just been an immediately-post-install thing
[00:50] <slangasek> hmm
[00:50] <slangasek> I wouldn't expect a fsck on the second reboot after install
[00:50] <slangasek> do you have the matching /var/log/boot.log?
[00:51] <broder> i rebooted the machine one more time after the chart i sent to ubuntu-desktop. let me see what i can get off the disk
[00:51] <broder> (i will also have one more bootchart so we can see if the i/o was spurious)
[00:57] <broder> slangasek: http://paste.ubuntu.com/702524/ is the boot.log that's there, but i assume that's from boot #3, not #2
[00:58] <broder> slangasek: the round 3 bootcharts are also in http://web.mit.edu/broder/Public/livecd-bootchart/ - they seem to have very similar disk i/o from fsck and update-motd
[03:14] <jbicha> smspillaz: I managed to get my launcher hiding behind my desktop again
[03:15] <jbicha> Alt-F1 or the Meta key still makes it pop out, but the edge reveal doesn't work, nor does it show when I alt-tab to the Desktop
[03:19] <jbicha> Ctrl-Alt-arrow keys to a blank desktop works and if I unmaximize my windows I'm ok too, but I like maximized windows on a relatively small screen
[04:06] <slangasek> broder: which one of those is the round-three boot chart?
[04:07] <broder> slangasek: bah, round 3 is the middle one
[04:07] <broder> (it ends in -3, not that you can see that)
[04:08] <slangasek> broder: hrm, how is that comparable disk i/o?
[04:08] <broder> slangasek: i'm comparing http://web.mit.edu/broder/Public/livecd-bootchart/postinstall-oneiric-20111004-2.png and http://web.mit.edu/broder/Public/livecd-bootchart/postinstall-oneiric-20111004-3.png
[04:09] <broder> which is 2nd boot after installing and 3rd boot, respectively
[04:09] <slangasek> well now I'm confused
[04:09] <slangasek> because I swear this isn't the boot chart I looked at earlier :)
[04:09] <broder> yeah, the...filename ordering was not ideal
[04:09] <broder> err, right. sorry
[04:10] <broder> the first bootchart i sent to the channel was one that was broken, because way too much I/O was happening after ureadahead
[04:10] <slangasek> oh
[04:10] <broder> but i had already purged the drive i was using when i realized that
[04:10] <slangasek> hah, ok
[04:10] <broder> and started over
[04:10] <slangasek> I thought that's still what we were debugging :)
[04:10] <broder> yeah, sorry
[04:11] <broder> i wish i still had it, because it was kind of strange
[04:11] <broder> but i'm assuming something just went horribly wrong with ureadahead's collection mode
[04:11] <slangasek> ureadahead isn't going to collect on behalf of fsck :)
[04:11] <broder> well, it should for the minimum amount of reading fsck does on every boot
[04:11] <broder> since mountall always does an fsck, it just usually doesn't do much
[04:12] <slangasek> nor update-motd-updates-available, which a) doesn't happen every boot, b) does writing more than reading
[04:12] <slangasek> right, but that boot chart showed a *non*-minimal fsck
[04:13] <broder> are you sure it wasn't just a slow minimal fsck that was slow enough that bootchart caught the child in the act?
[04:14] <slangasek> yes, because unless you're turning your disk by hand, there's no reason a fsck of a clean disk would take > 1s
[04:14] <broder> not even bad i/o scheduling?
[04:15] <slangasek> bad i/o scheduling wouldn't have shown up as solid red attributed to fsck on the bootchart :)
[04:15] <slangasek> (bootchart knows *which* process is actively engaged in i/o and colors the process appropriately - if it's just waiting its turn, it wouldn't have been red)
[04:16] <broder> i think that's light red - there's a modprobe further down that's dark red you can compare it against
[04:17] <slangasek> right, when I said "red" I mean "red that I can meaningfully distinguish from the background" :)
[04:17]  * broder shrugs
[04:18] <broder> i'll take your word for it - my bootchart reading ability is pretty minimal
[04:18] <slangasek> http://web.mit.edu/broder/Public/postinstall-oneiric-20111004-2.png - red fsck, not a no-op
[04:20] <robert_ancell> bryceh, RAOF, is rootless X back in scope for 12.04?
[04:20] <RAOF> robert_ancell: As in - not running X as root?  Probably not.
[04:20] <robert_ancell> RAOF, what's blocking it?
[04:21] <RAOF> It's the continuing saga of input routing.
[04:21] <robert_ancell> RAOF, i.e. udev?
[04:22] <RAOF> No - the problem is that when user switching you have no way of knowing that the old X server *actually* gave up it's read access to the tty, and isn't sitting pretty, nabbing all your passwords.
[04:22] <RAOF> Bah.  Apostropheing.
[04:23] <robert_ancell> RAOF, this is where the central wayland compositor fits in?
[04:23] <RAOF> Right.
[04:23] <robert_ancell> :)
[04:23] <RAOF> It's the trusted input router.
[05:46] <didrocks> good morning
[05:47] <RAOF> Hey, ho, didrocks!
[05:49] <didrocks> Ho, hey, RAOF ;)
[05:55] <didrocks> RAOF: we should track bug #830949 at UDS to have a session on those configurations with dx
[05:55] <ubot2> Launchpad bug 830949 in xserver-xorg-video-intel "[Intel N10 Graphics] Plugging in external monitor to VGA port makes both displays corrupted with thick slanted lines " [Undecided,Invalid] https://launchpad.net/bugs/830949
[05:55] <didrocks> RAOF: (hint hint, blueprint, blueprint ;))
[05:56] <RAOF> didrocks: I don't think there's a whole lot to discuss - they know what they need to implement, it just needs to be implemented.
[05:57] <didrocks> RAOF: yeah, but there was already a "we should do that" last cycle, if we don't mark it down somewhere, it won't be done next cycle either. Maybe not a session, but being part of one (multi-monitor enhancement, maybe?)
[05:57] <RAOF> I guess it could be useful for workitem tracking.
[05:57] <didrocks> right, the discussion itself won't, just tracking and not forgetting :)
[05:57] <RAOF> Right.
[06:14] <dupondje> mmmm telepathy-mission-control 5.9.3 is out today
[06:25] <pitti> Good morning
[06:25] <pitti> robert_ancell: still here?
[06:25] <didrocks> guten morgen pitti
[06:25] <robert_ancell> pitti, yes, on phone, will talk when off
[06:26] <pitti> bonjour didrocks
[06:26] <pitti> robert_ancell: great, thanks
[06:27] <jbicha> pitti: we're not disabling the login sound for oneiric, right?
[06:27] <pitti> jbicha: too late now, I think (unfortunately)
[06:27] <RAOF> Morning pitti
[06:28] <pitti> hey RAOF
[06:28] <jbicha> pitti: this is one guide for users to disable it: http://maketecheasier.com/disable-login-sound-in-ubuntu-oneiric-quick-tips/2011/09/15
[06:28] <pitti> jbicha: it's quite easy to do in the audio control panel
[06:28] <jbicha> we could make it simpler by removing the NoDisplay line but seb128 didn't like that by default Startup Applications would have one choice
[06:29] <jbicha> well that affects all sound effects, I like the other ones :)
[06:30] <jbicha> we did have a login preferences window in 11.04 with a 1-click login sound on/off button
[06:31] <jbicha> too bad this wasn't proposed sooner, I've not been a fan of the login noise for a few releases now
[06:31] <pitti> jbicha: oh, right, gnome-session-properties
[06:31] <didrocks> jbicha: you like the other one? waow, we found one! :-)
[06:31] <pitti> jbicha: but it doesn't seem to be there
[06:32] <jbicha> pitti: we dropped Login Preferences when we switched to GDM 3 as we didn't feel like rewriting it
[06:32] <pitti> jbicha: ah, that one (gdmsetup); I thought the autostart one
[06:33] <didrocks> the gdmsetup one was only the login sound only for gdm
[06:33] <didrocks> the "tadada"
[06:37] <robert_ancell> pitti, hey, about the LANG thing?
[06:39] <rickspencer3> didrocks, I just closed mumble and it caused my workspace to get visually garbled
[06:39] <rickspencer3> would this be at all a useful bug for me to log?
[06:40] <rickspencer3> *sigh*
[06:40] <robert_ancell> pitti, better get in quick before I log off...
[06:40] <rickspencer3> it seems to be that my desktop image is now mapped to my browser window
[06:43] <didrocks> rickspencer3: yeah, this bug is known and fixed for compiz
[06:43] <rickspencer3> thanks didrocks
[06:43] <didrocks> rickspencer3: I didn't want to push it yesterday, because it's risky
[06:43] <didrocks> and we know what happens when there is a change in the stacking code :)
[06:43] <rickspencer3> didrocks, np
[06:43] <rickspencer3> indeed
[06:43] <didrocks> rickspencer3: so basically first SRU, with a day of staging
[06:43] <didrocks> a week*
[06:43] <pitti> robert_ancell: whoops, notify fail
[06:43] <pitti> robert_ancell: I just followed up to the MP and to the bug
[06:44] <pitti> robert_ancell: so, the updated MP is rather bad IMHO (aside from the fact that it has a security vulnerability)
[06:44] <pitti> robert_ancell: I still don't understand why lightdm has to set $LANG in the first place, and second, why setting $LANGUAGE instead wouldn't be enough
[06:44] <pitti> robert_ancell: I thought the original idea was "control center and gdm only set a language, not a complete locale; if you want to have full control over locales you need langauge-selector"
[06:46] <robert_ancell> pitti, so the case is "systems that don't only use .dmrc and don't have a language changer need to set the language".  They have a Language defined in ~/.dmrc, and LightDM maps that to a locale (by scanning locale -a) and sets LANG overriding the system default set in PAM.  If you do use accountsservice then it is overriden by the values set in ~/.profile
[06:47] <pitti> robert_ancell: but that doesn't seem to be cjwatson's bug
[06:47] <pitti> robert_ancell: cjwatson's .dmrc had complete locales
[06:47] <robert_ancell> pitti, the case that cjwatson was getting, is he had both en_GB and en_GB.utf8 as valid locales, and lightdm picked the former meaning he was using the wrong locale.  1.0.0 changed this to pick a locale with a codeset over one without
[06:47] <pitti> cjwatson's bug is that lightdm assigns $LANG to the "Language" property in accountsservic3e
[06:48] <pitti> robert_ancell: right, which made it just differently wrong
[06:48] <pitti> robert_ancell: but the Language property is not a locale
[06:48] <robert_ancell> pitti, I didn't touch base with him, but he must have had "en_GB" in the output of locale -a (I don't)
[06:48] <pitti> robert_ancell: yes, he did
[06:48] <robert_ancell> so in his case, en_GB was a valid locale
[06:48] <pitti> yes, but nowhere did it say that it was his selected locale
[06:48] <pitti> .dmrc said en_GB.UTF-8
[06:49] <pitti> and accountsservice doesn't have a locale
[06:49] <pitti> robert_ancell: I think the part of .dmrc is okay
[06:49] <pitti> which is for the "derivatives" case, AFAIUI
[06:49] <robert_ancell> pitti, now that is weird.  It should have contacted accountsservice and that should have returned "en_GB"
[06:49] <pitti> robert_ancell: right, it did
[06:50] <pitti> robert_ancell: but that is *not* a locale
[06:50] <pitti> it's a language name
[06:50] <pitti> German would have said "de"
[06:50] <robert_ancell> pitti, correct, and lightdm uses set_language to map that to a locale name
[06:50] <robert_ancell> pitti, not de_DE?
[06:50] <pitti> yes, and that's the very bug :)
[06:50] <pitti> robert_ancell: no, because it is not a locale :)
[06:51] <pitti> robert_ancell: English and Spanish are special because the country specific dialects are so far apart
[06:51] <pitti> so for English, Spanish, and Portugese we also need the country dialect
[06:51] <pitti> but all other languages are similar enough to only have one name
[06:51] <pitti> e. g. German in Austria or French in Belgium are so close that they don't warrant a separate set of translations
[06:51] <pitti> robert_ancell: and you cannot select a country in control-center, i. e. you can never pick a locale there
[06:52] <pitti> i. e. I can't say "I'm in Austria"
[06:52] <pitti> I can just say "I speak German"
[06:52] <pitti> and that concept is $LANGUAGE, while "I'm in Austria" is a locale
[06:52] <robert_ancell> pitti, so is accountsservice different to dmrc?
[06:53] <pitti> robert_ancell: I'm not entirely sure how different login managers handled .dmrc; it was a giant pile of underdefined mess
[06:53] <pitti> robert_ancell: but from what cjwatson had, it had a complete locale, i. e. for $LANG
[06:53] <pitti> and _also_ a language list in a different key
[06:53] <pitti> hang on, looking for his pastebin
[06:53] <pitti> robert_ancell: but at least accountsservice is defined to only be a language name, i. e. for $LANGUAGE (not for $LANG)
[06:54] <pitti> it would have helped immensely if $LANG was called $LOCALE, FWIW
[06:54] <robert_ancell> pitti, hang on, lets set up etherpad and write this out in a table
[06:54] <pitti> robert_ancell: http://paste.ubuntu.com/702159/ <- cjwatson's .dmrc
[06:54] <pitti> Language=en_GB.UTF-8
[06:54] <pitti> Langlist=en_GB:en
[06:54] <pitti> LCMess=en_GB.UTF-8
[06:55] <pitti> which maps to $LANG, $LANGUAGE, and $LC_MESSAGES in that order
[06:55] <pitti> robert_ancell: sure, want to create one, or want me to?
[06:55] <robert_ancell> pitti, but that's wrong isn't it - shouldn't Language be "en_GB"
[06:55] <robert_ancell> pitti, you create one, I'm still trying to find the link to it
[06:56] <pitti> robert_ancell: http://pad.ubuntu.com/locale-handling
[06:58] <didrocks> brb
[06:59] <jbicha> pitti: could you merge https://code.launchpad.net/~jbicha/ubuntu/oneiric/jockey/update-help-link/+merge/77782
[07:09] <pitti> robert_ancell: I'm not sure what kind of language configuration tool derivatives use
[07:09] <pitti> robert_ancell: we didn't use to have one before hardy either, we just had the installer, and then people edited .profile
[07:09] <robert_ancell> pitti, they used to rely on the greeter, but if the DM doesn't set LANG, then nothing does
[07:09] <pitti> robert_ancell: sure, /etc/default/locale does
[07:09] <robert_ancell> pitti, yes, nothing per user
[07:11] <robert_ancell> pitti, so it sounds like the concept of the MP is right - we should set LANG="Language" from .dmrc if dmrc is used, and do nothing if accountsservice is used?
[07:14] <pitti> robert_ancell: yes
[07:15] <pitti> robert_ancell: setting the language from .dmrc is kind of deprecated now, at least from teh GNOME side, right?
[07:16] <robert_ancell> pitti, and the current behaviour is safe I *think* in that it will at worst set LANG to the value in .profile anyway
[07:16] <pitti> robert_ancell: no, it's not safe, as it sets $LANG from accountsservice's Language property
[07:16] <robert_ancell> pitti, yes, the LANG stuff in lightdm is a compromise for legacy system
[07:16] <pitti> i. e. it derives a locale from a language name, which can't work
[07:17] <pitti> robert_ancell: the bug we must fix is setting $LANG from Language; presumably the ~/.dmrc stuff is alright (haven't tested it, though)
[07:17] <pitti> those are two orthogonal issues
[07:17] <robert_ancell> pitti, is there an example that it will break.  If a-s Language=de, lightdm will run locale -a, look for de.* and then de, if it exists, set LANG to that otherwise do nothing
[07:17] <pitti> robert_ancell: nonono
[07:18] <pitti> don't do all that locale -a stuff
[07:18] <pitti> it can't work
[07:18] <pitti> that is the very bug
[07:18] <robert_ancell> pitti, I'm just running through what it does right now
[07:18] <robert_ancell> pitti, if you've never set a language it will do nothing as "Language" from a-s will be blank
[07:18] <robert_ancell> if you have set a language with the GUI, then .profile will be updated anyway and will override whatever lightdm set
[07:19] <pitti> robert_ancell: that case needs to be checked, of course; i. e. dont set $LANGUAGE to that empty field, just leave it alone
[07:19] <robert_ancell> pitti, I think we're safe as is for release though?
[07:20] <pitti> robert_ancell: I think the only thing that we need to fix for the release is bug 864618
[07:20] <ubot2> Launchpad bug 864618 in lightdm "UTF-8 locale no longer set" [Critical,In progress] https://launchpad.net/bugs/864618
[07:20] <pitti> the .dmrc stuff is probably alright
[07:20] <pitti> robert_ancell: i. e. we need to drop all that `locale -a` parsing and setting $LANG from accountsservice
[07:21] <pitti> right now that's a rather grave bug
[07:21] <robert_ancell> pitti, cjwatson should have LANG=en_GB.UTF-8 in lightdm 1.0.1 right?  He had LANG=en_GB in 1.0.0
[07:21] <pitti> I'm not sure why https://code.launchpad.net/~gunnarhj/gnome-control-center/non-utf8-confusion/+merge/78140 exists and why we need it
[07:21] <pitti> robert_ancell: no, he doesn't
[07:22] <pitti> robert_ancell: you don't look for .UTF-8
[07:23] <pitti> robert_ancell: it'll find en_GB.iso88591 first
[07:23] <robert_ancell> pitti, but I do in 1.0.1!
[07:23] <robert_ancell> oh
[07:23] <pitti> robert_ancell: but even then, you only look at the locale name, which can be anything
[07:23] <pitti> really, we need to drop that code
[07:23] <pitti> looking for .UTF-8 is just equally wrong
[07:23] <robert_ancell> well, I'm ok to drop it, and SRU the proper fix in later
[07:24] <pitti> many locales are UTF-8 and don't have such a suffix, and then have another encoding with a suffix
[07:24] <pitti> hy_AM UTF-8
[07:24] <pitti> hy_AM.ARMSCII-8 ARMSCII-8
[07:24] <pitti> for example
[07:24] <robert_ancell> pitti, you just need to comment out set_language()
[07:24] <pitti> robert_ancell: right, that was the original MP, AFAIUI
[07:25] <pitti> robert_ancell: what does that do exactly?
[07:25] <robert_ancell> pitti, but it will break it for derivatives
[07:25] <pitti> robert_ancell: set $LANG from accountsservice? or also handle .dmrc?
[07:25] <pitti> we need to drop the first, and keep the second
[07:25]  * pitti RTFS
[07:25] <robert_ancell> pitti, a-s and dmrc are the same inside lightdm, so both fields are being treated as a language code
[07:26] <robert_ancell> the proper fix is to have a locale property and a language property on the user object, the former set if using dmrc and the latter if using a-s.  If user.locale != null then set LANG
[07:27] <pitti> and if user.language != null, then set LANGUAGE (but that's optional as long as we write ~/.profile, but should be done later on)
[07:27] <robert_ancell> pitti, I think we just need to match the old GDM behaviour of setting LANG only.  If you want proper language support then use ~/.profile or ~/.pam_environment
[07:28] <robert_ancell> this is just a sticking plaster for desktops that haven't sorted out their language configuration
[07:28] <pitti> robert_ancell: that's easier indeed
[07:28] <pitti> so the long-term simple design is that lightdm/gdm don't set anything, and just make sure to source in .profile and .pam_environment
[07:28] <pitti> right?
[07:28] <robert_ancell> pitti, can you make the change?  (do we still have time)
[07:28] <robert_ancell> yes
[07:28] <pitti> robert_ancell: yes, this is a critical bug milestoned for final, we make time for these :)
[07:29] <pitti> robert_ancell: NB that if I upload it, I need to wait some time for someone else to review it, and then have the same conversation all over again
[07:29] <pitti> but Gunnar's original MP was correct, I think I can just sponsor that without the second /tmp/foo stamp stuff
[07:29] <pitti> robert_ancell: so yes, can do that ^
[07:29] <robert_ancell> ok.  thanks pitti!
[07:30] <robert_ancell> pitti, I have to go, is there anything else you need me to do?
[07:31] <robert_ancell> pitti, also you probably want to disable user_set_language in greeter.c as that will set the language wrong
[07:32] <robert_ancell> (only occurs if you've enabled the legacy language selector (in the gtk greeter))
[07:32] <Sweetshark> Moin Desktoppers!
[07:32] <pitti> robert_ancell: checking
[07:33] <pitti> robert_ancell: hm, that's an unrelated bug, though?
[07:33] <pitti> hey Sweetshark
[07:33] <robert_ancell> pitti, yeah, it's just the opposite of the other bug though
[07:33] <pitti> robert_ancell: so user_set_language gets passed a locale instead of a langauge name?
[07:34] <Sweetshark> pitti: 3.4.3-3ubuntu1~natty1 would fix postgres for natty, but needs pg 9.1 it seems. what should we do?
[07:34] <robert_ancell> pitti, I can't remember off hand, but it passes the same value to a-d and .dmrc so it will be wrong in one
[07:34] <pitti> Sweetshark: uh, how so? does it have server-side extensions, or does it actually use new API from libpq 9.1?
[07:34] <robert_ancell> to a-s rather
[07:34] <pitti> robert_ancell: right
[07:35] <Sweetshark> pitti: it uses the new api
[07:35] <pitti> robert_ancell: I think that's SRUable, though, as it only breaks when writing, not already when starting the session?
[07:35] <robert_ancell> pitti, yes
[07:35] <pitti> Sweetshark: so, I guess we can leave it broken then
[07:36] <robert_ancell> pitti, anything else?
[07:37] <Sweetshark> pitti: ok
[07:38] <robert_ancell> pitti, I have to go, email me and I'll check tonight if it's urgent
[07:38] <pitti> robert_ancell: good to go now, I think
[07:39] <pitti> jasoncwarner_: nothing on my list for 1-1 today, just a ton of stuff to sort out for oneiric, so I wouldn't mind dropping it
[07:39] <jasoncwarner_> pitti: done
[07:47] <rodrigo_> morning
[07:53] <didrocks> hey rodrigo_
[07:53] <rodrigo_> hi didrocks
[08:05] <micahg> good morning seb128
[08:05] <seb128> hey
[08:06] <didrocks> salut seb128
[08:06] <seb128> lut didrocks
[08:06] <seb128> hey micahg
[08:06] <micahg> seb128: I was wondering, can we switch the Ubuntu Desktop team owner to the DMB so that the DMB doesn't have to be a member and therefore does not get extra upload rights?
[08:07] <seb128> micahg, what is the dbm? what does that mean? why the dmb needs to own the team?
[08:07] <micahg> seb128: the DMB is the Developer Membership Board which delegates to the desktop team the right to confer upload rights
[08:08] <seb128> hum, I though we were entitled to decide who join our team?
[08:08] <pitti> hey seb128
[08:08] <micahg> seb128: yes, you are, that's the delegation :)
[08:08] <seb128> micahg, there is probably a reason but I don't understand what issue you try to solve
[08:08] <seb128> hey pitti ;-)
[08:09] <micahg> seb128: right now the DMB members who are not core-dev can upload from the desktop packageset, I'm trying to prevent that
[08:09] <seb128> micahg, why does the DMB need to have rights on the desktop set?
[08:09] <pitti> micahg: as long as some of the team members are admins, that makes sense indeed
[08:09] <micahg> seb128: the most probably case would be for removals
[08:10] <micahg> seb128: otherwise, I don't see us touching it, we're already admins on the team though, so no increase in permissions for us
[08:10] <micahg> in fact, this reduces our permissions
[08:11] <micahg> seb128: pitti and you would remain admins to manage as usual
[08:12] <seb128> I've no strong opinion, it feels a bit weird to have the ownership moved to a team which has nothing to do with desktop but if pitti is ok feel free to do it
[08:12] <pitti> seb128: core-dev is owned by DMB as well
[08:12] <seb128> i.e I don't understand why DMB needs to have rights in a team well administrated but there is probably a reason ;-)
[08:12] <pitti> seb128: that's the point -- we want to _remove_ right :)
[08:13] <seb128> pitti, well they can ping any admin for that as long as we have active admins ;-)
[08:13] <micahg> owner = modify membership w/out being a member
[08:13] <pitti> seb128: no, the thing micah wants to remove is DMB members being able to upload desktop packages
[08:13] <seb128> well I don't see DMB has "owning" desktop
[08:14] <seb128> pitti, well, just unsubscribe DMB from the team?
[08:14] <pitti> seb128: they need to be able to add/remove members to it, as per delegation policy
[08:14] <micahg> seb128: should we make a separate team just for the upload rights?
[08:14] <seb128> but if that's the standard way it's done for other team just do it for desktop
[08:14] <seb128> it just feels weird ;-)
[08:14] <pitti> micahg: ubuntu-desktop is the team for the upload rights, by and large
[08:14] <seb128> micahg, that's what the team is about already, uploads
[08:15] <micahg> ok
[08:15] <seb128> we have desktop-bugs for bug triagers, and gnome3-team for new crack
[08:15] <pitti> seb128: i. e. drop DMB as a member, but make it an owner; that's the standard way other teams are set up these days
[08:15] <seb128> pitti, micahg: ok, please do it then ;-)
[08:15] <micahg> yeah, I'm just doing some cleanup that should've been done a while ago
[08:15] <pitti> seb128: can't, you are the owner
[08:15] <seb128> lol
[08:15] <seb128> oh, I see,  you are trying to kick me out :p
[08:16] <seb128> micahg, pitti: ok, I will change the owner
[08:16] <micahg> seb128: heh, you're still a full admin for the team
[08:17] <seb128> no worry, I think the misunderstanding comes from my lack of understand of what is the "owner" role exactly for a team
[08:17] <seb128> but I will trust you guys ;-)
[08:18] <micahg> yeah, the role names can be confusing
[08:21] <jbicha> are we going to use a brand new PPA for the GNOME 3.2 extra backport stuff?
[08:21] <rodrigo_> pitti, seb128: so, what do we do with the rejected Recommends: mesa-utils in control-center?
[08:21] <pitti> rodrigo_: I think just keep what we have at this point
[08:22] <seb128> rodrigo_, yeah, I blame it on bryceh and RAOF that glxinfo is in universe
[08:22] <rodrigo_> pitti, ok, I'll remove that part and re-upload
[08:22] <seb128> jbicha, I would recommend using gnome3-team?
[08:22] <ricotz> jbicha, just use the oneiric pocket of the current ppa?
[08:22] <seb128> rodrigo_, can you get Gunnar's fix as well?
[08:23] <seb128> rodrigo_, https://code.launchpad.net/~gunnarhj/gnome-control-center/non-utf8-confusion/+merge/78140
[08:23] <ricotz> hi everyone
[08:23] <seb128> hey ricotz
[08:23] <rodrigo_> seb128, sure
[08:23] <ricotz> hey seb128
[08:23] <seb128> rodrigo_, though I'm not sure the patch is right
[08:23] <rodrigo_> looking
[08:23] <jbicha> well there's 161 packages in the PPA, for Oneiric we might not use more than 10% of that number!
[08:24] <rodrigo_> seb128, hmm, doesn't look good to me neither
[08:24] <jbicha> I was thinking a new PPA would be simpler, without needing to filter by series
[08:25] <ricotz> jbicha, that is the purpose of filtering ;)
[08:25] <ricotz> i dont see a need for a new one here
[08:25] <jbicha> ricotz: yes but it's clutter that was just for Natty and we don't need that baggage
[08:26] <seb128> jbicha, well nobody will see the ppa content, user just add one serie to their source
[08:26] <jbicha> but I see it! ;-)
[08:27] <ricotz> filter it! ;)
[08:27] <seb128> well, we can clean the natty packages once oneiric is out
[08:28] <jbicha> oh ok, I like that solution :)
[08:28] <ricotz> seb128, mhh, i would wait for that a few months
[08:28] <rodrigo_> seb128, pitti: with Gunnar's patch in https://bugs.launchpad.net/ubuntu/+source/lightdm/+bug/864618 there's no need for the "fix" in g-c-c, right?
[08:28] <ubot2> Launchpad bug 864618 in gnome-control-center "sets $LANG to a language name" [Low,In progress]
[08:29] <rodrigo_> the fix in g-c-c is https://code.launchpad.net/~gunnarhj/gnome-control-center/non-utf8-confusion/+merge/78140
[08:29] <jbicha> ricotz: well we want to discourage users from attempting to do GNOME 3 in Natty though
[08:29] <rodrigo_> which looks wrong to me
[08:29] <seb128> rodrigo_, there is still I think
[08:29] <seb128> rodrigo_, but it's very low importance, let's not bother for oneiric
[08:29] <rodrigo_> seb128, ok, so I upload without this?
[08:29] <seb128> rodrigo_, it seems that the user account can pick an iso8859 encoding when the locale doesn't reflect that
[08:29] <seb128> rodrigo_, right
[08:30] <rodrigo_> ok
[08:30] <pitti> rodrigo_, seb128: hang on, on the phone
[08:30] <ricotz> jbicha, ok, but some people doesnt upgrade instantly here, and i would wait for some later revisions of 3.2.x to land before cleaning it
[08:31] <pitti> rodrigo_: I'm not sure what that "confusion" MP is all about
[08:31] <pitti> rodrigo_: probably better to ask Gunnar directly (in the MP or on IRC)
[08:31] <rodrigo_> yes
[08:31] <jbicha> ricotz: I just think that GNOME 3 in Oneiric is far better than it was in Natty though
[08:32] <seb128> pitti, it's what cjwatson mentioned on #ubuntu-devl yesterday, he had "english (iso-8859-1)" in the user admin combo where his locale is not an iso one
[08:32] <pitti> rodrigo_, seb128: FYI, that's where much of this confusion came from: accountsservice and control-center do NOT select a locale, or encoding, or anything; just a language
[08:32] <rodrigo_> yeah
[08:32] <rodrigo_> seems the lightdm part is ok now, right?
[08:32] <seb128> ricotz, jbicha: well, cleaning the natty ppa will not uninstall GNOME3 from natty boxes which installed it from the ppa, just prevent new users to installit
[08:32] <pitti> the concept of encodings, iso, locales all don't apply to control-center
[08:32] <pitti> rodrigo_: it's not ok yet, but at least that part is
[08:32] <rodrigo_> ok
[08:33] <pitti> rodrigo_: lightdm stopped setting $LANG from accountsservice's Language property
[08:33] <pitti> which was the main bug
[08:34]  * pitti bbl
[08:34] <ricotz> seb128, i know, still leaving it some weeks active doesnt hurt here
[08:35] <seb128> ricotz, right, it's not useful but doesn't hurt a lot, out of the fact that I start having enough getting gnome-icon-theme bugs from people who try the ppa and try to roll back using ppa-purge :p
[08:35] <ricotz> seb128, i am still relying on this ppa as build-dep for mine
[08:35] <jbicha> it hurts me!! :)
[08:36] <jbicha> ricotz: that is the purpose of copy packages :)
[08:36] <micahg> jbicha: no, not really, that just wastes space on the servers
[08:38] <ricotz> seb128, jbicha, please leave it too me to clean the natty pocket
[08:38] <seb128> ok
[08:38] <jbicha> ok with me too
[08:40] <ricotz> alright, i will wait some weeks then, to see if the download counts are dropping
[08:41] <dupondje> any idea when kenvandine comes in? :) he still needs to patch mission-control
[08:43] <seb128> dupondje, he's u.s based, it's like 3am for him
[08:43] <seb128> dupondje, so in a few hours
[08:43] <dupondje> heh ok :)
[08:44] <seb128> ricotz, there are still lot of users installing it? just curious ;-)
[08:44] <ricotz> seb128, yes, at least updating from it
[08:45] <jbicha> what are the numbers? I don't think I've seen them
[08:51] <ricotz> jbicha, it might not be that accurate but it is like 800 downloads per day
[09:36] <didrocks> pitti: we don't use the desktop cache anymore? Seems /usr/share/applications/desktop.fr_FR.utf8.cache is out of date for me
[09:37] <pitti> didrocks: right, we didn't get it ported to the new gnome-menus, and didn't find it to be important any more for boot speed
[09:38] <pitti> didrocks: as we don't actually use gnome-panel any more, etc.
[09:38] <didrocks> pitti: ok, thanks or confirming :)
[09:48] <seb128> didrocks, pitti: we should perhaps make gnome-menus clean those on upgrade?
[09:48] <pitti> seb128: yes, right
[09:49] <pitti> seb128: btw, thanks for getting the pad restored
[09:49] <pitti> seb128: I'll upload a gnome-menus with that cleanup
[09:50] <didrocks> pitti: seb128: there was no use of it for unity-place-applications?
[09:50] <didrocks> kamstrup: ^
[09:50] <pitti> didrocks: there probably was, but we just didn't get around to porting the patch
[09:50] <pitti> gnome-menus 3.2 got changed pretty thoroughly
[09:51] <kamstrup> pitti, didrocks: yes, u-l-a uses libgmenu
[09:52] <kamstrup> butit is started lazily by unity
[09:53] <pitti> didrocks, seb128: filed as bug 868219
[09:54] <ubot2> Launchpad bug 868219 in gnome-menus "Needs to remove old menu caches" [Medium,In progress] https://launchpad.net/bugs/868219
[09:54] <chrisccoulson> hi everyone
[09:54] <pitti> hey chrisccoulson
[09:54] <chrisccoulson> seb128, did you know that we can't switch users from the lock dialog?
[09:54] <chrisccoulson> hey pitti, how are you?
[09:54] <jasoncwarner_> pitti seb128 just got done talking to mark about some stuff and he mentioned that he was having suspend/resume problems...is there a general thing happening (I'm not seeing it)
[09:54] <chrisccoulson> seb128, i just noticed that here when jo logged in to my laptop
[09:55] <pitti> jasoncwarner_: hm, works fine here
[09:55] <chrisccoulson> i can't switch back to my own account after suspending ;)
[09:55] <pitti> I have suspended my oneiric laptop maybe 20 times over the long weekend
[09:55] <jasoncwarner_> pitti: he also mentioned intel wifi locking up his system, which should never happen.
[09:55] <seb128> pitti, thanks
[09:55] <pitti> not generally known, I think
[09:56] <seb128> pitti, want to clean the buggy conffile as well? ;-)
[09:56] <jasoncwarner_> chrisccoulson: say what now? you can't what after resuming from suspend?
[09:56] <seb128> chrisccoulson, hey
[09:56] <chrisccoulson> jasoncwarner_, i can't switch users after resuming from suspend, as the lock dialog has no switch-user button
[09:56] <chrisccoulson> i guess that depends on gdm?
[09:56] <seb128> chrisccoulson, no, I didn't know, seems like a lightdm issue :-(
[09:57] <jasoncwarner_> oh, but you can resume and log in to the suspended account...the issue is you can't switch accounts...
[09:57] <seb128> jasoncwarner_, suspend resume is not known to be broken no
[09:57] <jasoncwarner_> I think mark's issue is that it generally doesn't suspend or resume
[09:57] <chrisccoulson> jasoncwarner_, yeah, which makes it difficult if jo resumes after i was the last person to use it
[09:57] <chrisccoulson> as she can't access her session again
[09:57] <jasoncwarner_> chrisccoulson: we should fix that
[09:57] <seb128> chrisccoulson, the lock dialog has a switch user button there
[09:57] <seb128> but the dbus interfaces it tries to use seems to not work
[09:58] <chrisccoulson> hmmm, i wonder why i don't have that here
[10:00] <chrisccoulson> seb128, do you have gdm installed?
[10:00] <pitti> seb128: I don't have a "switch user" button
[10:01] <chrisccoulson> it depends on gdm
[10:01] <seb128> chrisccoulson, yes, gdm is installed
[10:01] <chrisccoulson> seb128, that's why ;)
[10:01] <seb128> why does it depends on gdm?
[10:01] <seb128> we got a /usr/lib/lightdm/lightdm/gdmflexiserver
[10:01] <chrisccoulson> seb128, http://git.gnome.org/browse/gnome-screensaver/tree/src/gs-lock-plug.c#n1020
[10:01] <seb128> lightdm adds /usr/lib/lightdm/lightdm to the path as well
[10:01] <chrisccoulson> it checks if it is in the path
[10:01] <chrisccoulson> oh
[10:02] <chrisccoulson> that doesn't work here then, as it's not in my path
[10:04] <chrisccoulson> anyway, need to disappear again now
[10:05] <jasoncwarner_> didrocks: another issue mark mentioned was he was starting up and unity failed to start (no details). Any ideas?
[10:05] <seb128> chrisccoulson, thanks for pointing it
[10:05] <chrisccoulson> no problem :)
[10:05] <seb128> jasoncwarner_, get us his .xsession-errors
[10:05] <seb128> chrisccoulson, mterry did an hack to add the path to PATH but it seems to not work
[10:05] <didrocks> jasoncwarner_: yeah, I heard about this one :-) I asked for checking if lightdm is the latest one, if there is a crash file as well
[10:06] <didrocks> jasoncwarner_: nothing worrying in the french forum (no crash at startup since the fixed lightdm with root .xsessione-errors)
[10:06] <jasoncwarner_> didrocks: care to ping him and see?
[10:07] <didrocks> jasoncwarner_: yeah, apparently he's away for now, isn't it?
[10:12] <jibel> didrocks, I still get bug 851345 even with latest lightdm and .xsession-errors is not owned by root anymore.
[10:12] <ubot2> Launchpad bug 851345 in lightdm "compiz crashes on i915 graphic module with GLX_EXT_texture_from_pixmap is missing" [High,Confirmed] https://launchpad.net/bugs/851345
[10:12] <jibel> in that case, only nautilus starts and no unity
[10:12] <didrocks> jibel: oh, I didn't get it anymore after 10 reboots there. Better to ping RAOF about it
[10:12] <pitti> seb128: I did some more cleanup in the gnome-menus bzr, but not necessary to upload that into oneiric, FYI
[10:12] <seb128> pitti, ok
[10:13] <didrocks> jibel: ok, maybe that's what sabdfl is experiencing, thanks for the notice
[10:13] <didrocks> RAOF: when you will be around tomorrow ^^
[10:13] <seb128> didrocks, jibel: it's rather a robert_ancell issue
[10:13] <seb128> it seems the ck session is not correct
[10:13] <jibel> I pinged them last night but I got it this morning again after upgrading to the latest lightdm
[10:13]  * didrocks reopens the bug again
[10:13] <seb128> didrocks, don't
[10:14] <pitti> didrocks: meh, unity-lens-applications diff is basically unreadable due to v3 source
[10:14] <pitti> didrocks: can we also switch this back to v1 in P, similar to unity?
[10:14] <didrocks> yeah, ck
[10:14] <didrocks> seb128: reopens to read :)
[10:14] <seb128> didrocks, oh ok ;-)
[10:14] <didrocks> pitti: yeah, I'll do that for every bzr merge-upstream package
[10:14] <pitti> thanks
[10:14] <didrocks> pitti: u-l-a change is one line
[10:15] <didrocks> pitti: unity one is +1/-1
[10:15] <jasoncwarner_> jibel and didrocks :( thanks...robert_ancell and RAOF  will need to take a look tomorrow!
[10:15] <pitti> ++    DesktopAppInfo.set_desktop_env ("Unity");
[10:15] <pitti> didrocks: ^ is it that one?
[10:16] <didrocks> pitti: yeah
[10:16] <pitti> didrocks: thanks
[10:16] <didrocks> pitti: replaces GNOME by Unity
[10:17] <didrocks> pitti: for the unity upload, it's http://bazaar.launchpad.net/~unity-team/unity/trunk/revision/1693
[10:18] <pitti> didrocks: thanks
[10:35] <seb128> pitti, can we still get a glib upload today?
[10:36] <pitti> seb128: I guess we can, but that sounds risky; what does it fix?
[10:36] <seb128> pitti, the postrm has
[10:36] <seb128> if [ -d /usr/share/glib-2.0/schemas ]; then
[10:36] <seb128>     # Purge the compiled schemas
[10:36] <seb128>     rm -f /usr/share/glib-2.0/schemas/gschemas.compiled
[10:36] <seb128>     rmdir -p --ignore-fail-on-non-empty /usr/share/glib-2.0/schemas
[10:36] <seb128> fi
[10:37] <seb128> pitti, that basically leads to a time where there is no gsettings cache
[10:37] <seb128> which leads to applications randomly segfaulting
[10:37] <seb128> we should probably just do that only on purge
[10:37] <pitti> seb128: on upgrade, are the schemas updated in the postinst?
[10:38] <seb128> pitti, yes
[10:38] <seb128> and with a trigger
[10:38] <pitti> seb128: doesn't seem to be an install issue to me, is it? i. e. would be no worse off with an SRU?
[10:38] <seb128> pitti, we can sru yes
[10:39] <pitti> seb128: but anyway, if you want to work on this today, there's still time
[10:39] <seb128> pitti, right, it's a "during glib upgrade there is a small timeframe where the schemas mmap is missing and it leads to weird bugs"
[10:39] <seb128> pitti, I'm fine doing a SRU if you prefer
[10:39] <pitti> but seems to be "medium" priority to me
[10:40] <seb128> let's SRU it
[10:40] <pitti> seb128: yeah, let's; takes the pressure out, and we don't gain anything there by rushing IMHO
[10:40] <seb128> right
[10:45] <Sweetshark> pitti: shouldnt https://launchpad.net/ubuntu/+source/libreoffice/1:3.3.4-0ubuntu1 be long through proposed by now? see also http://blog.documentfoundation.org/2011/10/05/the-document-foundation-publishes-details-of-libreoffice-3-4-3-security-fixes/
[10:45] <pitti> Sweetshark: nobody followed up to the bug yet with testing results
[10:47] <GunnarHj> pitti: Hi Martin, seems like bug 864618 was decently resolved.
[10:47] <ubot2> Launchpad bug 864618 in gnome-control-center "sets $LANG to a language name" [Low,In progress] https://launchpad.net/bugs/864618
[10:47] <GunnarHj> I noticed another similar lightdm problem, and wrote a new MP that would fix it in Oneiric:
[10:47] <GunnarHj> https://code.launchpad.net/~gunnarhj/lightdm/lang-chooser/+merge/78226
[10:47] <pitti> hey GunnarHj
[10:48] <pitti> GunnarHj: dropping the set_language() call was indeed a bit blunt, but Robert said that fixing it properly is more intrusive and SRU material; I filed a new bug about the .dmrc issue
[10:48] <pitti> GunnarHj: had an 1.5 h discussion with him this morning and we created that pad to collect info
[10:48] <pitti> GunnarHj: ugh, locales are hard :)
[10:49] <GunnarHj> pitti: Ok. Yes it's hard because of all the different ideas about it that's floating around. :)
[10:50] <pitti> GunnarHj: why do we need to turn a language into a locale?
[10:50] <pitti> (and calling a perl script during login is also a bit sad)
[10:50] <GunnarHj> pitti: Because of the way it's matched against locale -a items.
[10:50] <pitti> well, not any more
[10:51] <pitti> that stuff got disabled
[10:51] <Sweetshark> pitti: shall I? It would be quite stupid and nonsense since I created to package in the first place, though.
[10:51] <pitti> Sweetshark: as long as you use the actual proposed .debs and give them some exercise in a natty environment, that's fine
[10:51] <GunnarHj> pitti: No, no, that was another set_language() function.
[10:52] <GunnarHj> pitti: The solution in the MP is definately not optimal, but it was the thing I was able to come up with at this point.
[10:52] <pitti> GunnarHj: I thought this part of the bug is solved for good now
[10:53] <pitti> GunnarHj: what's missing is to read .dmrc and set $LANG from it, for backwards compat/derivatives?
[10:53] <GunnarHj> pitti: And it's not the Perl script, but a shell script.
[10:53] <GunnarHj> pitti: No, they are separate things.
[10:58] <Sweetshark> pitti: do you have any good idea for javahelper wrt porting 3.4.3 back to natty? it seems to break on symlinks pre-0.37 and we dont have that in natty. Can we backport javahelper too? It prolly easier and safer that finding/fixing all the stuff it does wrong ...
[10:59] <pitti> Sweetshark: that's for the PPA? sure, you can just put the new helper into that PPA, too
[10:59] <Sweetshark> k
[10:59] <pitti> please make sure that the build-dep on it is strict enough
[11:05] <Sweetshark> pitti: you mean pre-depends
[11:05] <Sweetshark> ?
[11:05] <pitti> whereever it is used, yes
[11:06] <Sweetshark> although
[11:06] <Sweetshark> build-deps is even earlier that pre-depends, so what exactly do you mean?
[11:07] <pitti> Sweetshark: both then
[11:07] <pitti> Sweetshark: if the package doesn't build with earier javahelper, bump the b-dep to >= 0.37
[11:07] <pitti> if the package doesn't install with an earlier one, bump the (pre-)depends
[11:08] <Sweetshark> pitti: javahelper <0.37 creates wrong deps on build, I am not aware of any bugs on installing.
[11:08] <pitti> Sweetshark: ah, then build-depends >= 0.37 seems fine
[11:13] <Sweetshark> pitti: tagged SRU as verification-done.
[11:13] <pitti> Sweetshark: oh, that was quick :)
[11:14] <Sweetshark> pitti: 3.3 has no sensible test suite :(
[11:25] <GunnarHj> pitti: Did I succeed in convincing you that https://code.launchpad.net/~gunnarhj/lightdm/lang-chooser/+merge/78226 deals with a separate issue?
[11:27] <pitti> GunnarHj: oh, right, you did; I just got confused initially because I was discussing bug 864618 all morning :)
[11:27] <ubot2> Launchpad bug 864618 in gnome-control-center "sets $LANG to a language name" [Low,In progress] https://launchpad.net/bugs/864618
[11:27] <GunnarHj> pitti: Ok, good. :)
[11:27] <pitti> GunnarHj: this should be a nice SRU, IMHO not critical for final
[11:28] <GunnarHj> pitti: Ok, as long as you'll accept is as an SRU, it's ok with me to wait.
[11:28] <pitti> so it can go through the usual "review by robert/land upstream" cycle first
[11:29] <GunnarHj> pitti: And it may give us time to find a better solution, too.
[11:36] <pitti> GunnarHj: btw, that pad wasn't really meant to be a long-term documentation place (wiki would be better for that)
[11:36] <pitti> GunnarHj: robert and I mostly used it to co-edit the current questions and resolve misunderstandings
[11:36] <Sweetshark> any gnome guru around? I need to know if there is a way to find out if gnome-session is currently shutting down the session (for bug 562027)
[11:36] <ubot2> Launchpad bug 562027 in libreoffice "[Upstream] [ooo-build] QuickStarter: unable to shutdown / reboot / logout when quickstarter is active" [Medium,Confirmed] https://launchpad.net/bugs/562027
[11:38] <seb128> Sweetshark, not sure to understand the question
[11:40] <seb128> Sweetshark, gnome-session has 2 interfaces for session handling, a dbus one and the old xsmp protocol from xorg
[11:41] <seb128> Sweetshark, seems like the xsmp code in lo is broken
[11:54] <kancerman> I'm getting several BAD SIG error lines ... is there a way to reacquire all the signing keys files ... ??
[11:57] <seb128> mvo, there?
[11:59] <seb128> mvo, unping
[12:01] <mvo> seb128: pong
[12:01] <mvo> seb128: unpong
[12:01] <seb128> ;-)
[12:02] <seb128> mvo, I had a weird software-properties-gtk source edition, I added ppa:ubuntu-desktop/ppa and got a source file with the deb line commented and the second line starting with "src" rather than "deb-src"
[12:02] <seb128> mvo, but that seems to work no, I wonder if that was a leftover file on disk that it didn't like or something
[12:02] <mvo> seb128: *weeh*
[12:03] <seb128> mvo, I will ping again if I get the issue again ;-)
[12:05]  * rodrigo_ lunch
[12:10] <GunnarHj> pitti: Thanks for the clarification as regards the pad. Then I'll concentrate on putting down some stuff on a wiki page.
[12:13] <Sweetshark> seb128: thx, getting closer to it now.
[12:22] <seb128> mterry, hey
[12:22] <seb128> mterry, do you have time for a lightdm bug today? ;-)
[12:22] <mterry> seb128, sure
[12:23] <seb128> mterry, the gdmflexiserver path hack still doesn't work...
[12:23] <seb128> mterry, that breaks user switching on the lock screen
[12:24] <mterry> seb128, guh
[12:24] <mterry> seb128, bug number?
[12:24] <seb128> I knew you were going to love that :p
[12:24] <seb128> mterry, it was IRC reported but I can file a bug if you want
[12:24] <seb128> let me check if there is one
[12:25] <mterry> seb128, so to reproduce, I lock my screen and try to switch users?
[12:26] <seb128> mterry, well basically you uninstall gdm or move its gdmflexiserver away, (restart gnome-screensaver?), then try to lock your screen
[12:26] <seb128> mterry, you should get no "switch user" button
[12:27] <seb128> mterry, the code is basically looking if gdmflexiserver is in the path, but it's not
[12:27] <mterry> seb128, this is without lightdm?
[12:28] <seb128> mterry, no, that's with lightdm
[12:28]  * mterry wonders why gdmflexiserver isn't in the path
[12:28] <seb128> mterry, because lightdm install it in /usr/lib/lightdm and your path hack seems to not work (my guess)
[12:28] <seb128> mterry, remember you added code to add /usr/lib/lightdm to the PATH
[12:28] <seb128> which is somewhat hackish ;-)
[12:29] <mterry> seb128, I remember.  But it was working...
[12:29] <seb128> well doing a "strings /proc/$(pidof gnome-session)/environ | grep PATH" shows it's not in the path
[12:29] <seb128> mterry, yeah, I blame it on robert_ancell he probably broke it :p
[12:30] <seb128> mterry, well robert_ancell already had a few other lightdm blockers on his light and is flying to the GNOME summit in the next days so I assume he will not have time to look at it
[12:30] <seb128> mterry, if you can that would be nice, otherwise we will SRU it later
[12:31] <mterry> seb128, will look
[12:31] <seb128> mterry, thanks
[13:03] <kenvandine> jasoncwarner_, still awake?
[13:06] <seb128> kenvandine, not likely, he said he was going to bed a few hours ago
[13:07] <kenvandine> i figured as much, i wanted him to run a debugging script for me... oh well :)
[13:07] <kenvandine> email it is!
[13:07] <kenvandine> good morning seb128!
[13:07] <seb128> ;-)
[13:07] <seb128> hey kenvandine, how are you?
[13:07] <kenvandine> tired... i need more coffee :)
[13:07] <kenvandine> and you?
[13:09] <mterry> seb128, gdmflexiserver fix uploaded.  I blame robert-ancell
[13:09] <mterry> (since he's not here :))
[13:10] <seb128> mterry, thanks ;-)
[13:10] <seb128> (hehe)
[13:10]  * kenvandine goes to make more coffee, bbiab
[13:23] <dupondje> kenvandine: your still checking mission-control?
[13:25] <kenvandine> dupondje, i decided against it, but i will try to backport the patches for an SRU
[13:25] <kenvandine> the diff of the new release was more than i felt comfortable uploading in hard freeze
[13:26] <dupondje> heh ok :)
[13:26] <dupondje> new version is out also btw since today
[13:26] <dupondje> but thats TO much diff prolly :)
[13:26] <kenvandine> yeah, which includes the second fix you pointed out
[13:26] <kenvandine> that wasn't in the previous release yet
[13:26] <kenvandine> i am going to try to back port both of those fixes for SRU
[13:27] <dupondje> and it has nice integration with NetworkManager also now
[13:27] <dupondje> wich is even cooler ;)
[13:27] <kenvandine> that will be for 12.04 then :)
[13:27] <dupondje> when does 12.04 pre-alpha come out ? :p
[13:30] <dupondje> and its not possible to only backport the fixes for the crashes ?
[13:31] <kenvandine> dupondje, that is what i plan to do
[13:31] <dupondje> but after release ?
[13:33] <didrocks> brb
[13:39] <kenvandine> dupondje, yeah, after release
[13:41] <ogra_> well, you can upload SRUs to the queue before release
[13:41] <ogra_> so they roll in as zero day updates
[13:42] <kenvandine> ogra_, yeah, i want to test it a bit though
[13:42] <ogra_> true :)
[13:55] <ogra_> mpt, seeing that you discuss the window crash dialogs, would it be possible to make them not steal the users focus ? its really annoying if something dies in the backrgound and pops into the middle of your sentence
[13:55] <ogra_> s/and pops/and the error pops/
[13:56] <mpt> ogra_, sure
[13:56]  * ogra_ runs into that quite often 
[13:56] <ogra_> probably because i use arm and have more crashes than others :)
[13:59] <seb128> the wm focus stealing prevention should avoid that
[13:59] <seb128> but yeah, "bug"
[14:00] <ogra_> yep
[14:08] <didrocks> seb128: a lot of people seems to have nautilus-open-terminal making nautilus crashing
[14:09] <didrocks> (there is no new release from what I see)
[14:09] <seb128> didrocks, there was a fix uploaded yesterday
[14:09] <seb128> didrocks, does it still happen or are those on lagging mirrors?
[14:09] <didrocks> seb128: maybe lagging mirrors, the upload yesterday was not only ubuntuone-nautilus?
[14:09] <seb128> didrocks, there was an ubuntuone-client-gnome issue as well, different bug, different upload ;-)
[14:10] <didrocks> seb128: ok, I would say lagging mirrors then, let's see if there are still complains tomorrow
[14:10] <seb128> didrocks, right
[14:10] <seb128> didrocks, it got uploaded a 9pm yesterday
[14:10] <seb128> usually it takes a day or two for reports to settle
[14:11] <didrocks> yeah
[14:11] <didrocks> ok, I'll keep it under the radar
[14:11] <dobey> we think people are still having issues with the g_value_unref issue, and Our patch will 'fix' their crashes
[14:13] <seb128> dobey, we being?
[14:13] <seb128> dobey, didn't you add the schemas? that should fix the issue
[14:14] <seb128> nautilus works fine with u1 for me since the update
[14:18]  * dobey is we
[14:19] <kenvandine> no crashes for me
[14:22] <seb128> it's a bit annoying that opening a guest session trigger keyring dialogs
[14:27] <dobey> seb128: we added the schemas, yes; but schemas being there is irrelevant if the cache gets corrupted or gsettings doesn't load them or whatever
[14:28] <ogra_> seb128, mpt  https://lists.ubuntu.com/archives/ubuntu-users/2011-October/252272.html ... /me thinks thats a good question, do we intend to have syslog selected by default ?
[14:29] <ogra_> now that the world only uses a single file for logging it probably makes sense to have tehh GUI point to it by default
[14:30] <mpt> ogra_, I didn't know that either System Log Viewer or syslog existed :-)
[14:30] <ogra_> heh, we install it by default
[14:30] <ogra_> or rather gnome does, we inherit
[14:31] <mpt> Why does one of its menu titles end with a colon?
[14:31] <mpt> "Filters:"
[14:31] <seb128> ogra_, seems a bug, not sure what you say about syslog, it should have a left pane and a collection of logs
[14:31] <ogra_> a bug ... perhaps ?
[14:31] <seb128> ogra_, bug #841085
[14:31] <ubot2> Launchpad bug 841085 in gnome-utils "Log File Viewer empty" [Undecided,Confirmed] https://launchpad.net/bugs/841085
[14:31] <seb128> ogra_, I will look at it
[14:31] <ogra_> oh, cool, there is always already a bug :=
[14:32]  * ogra_ sighs about his typing, why are german keyboards designed in a way that i always end up with hitler smileys when i typo 
[14:32] <kenvandine> ogra_, and seb128 always finds it
[14:32] <ogra_> kenvandine, yeah :)
[14:32] <kenvandine> haha
[14:32]  * kenvandine sucks at searching launchpad for bugs
[14:32]  * ogra_ too
[14:33] <seb128> dobey, "if the cache gets corrupted or gsettings doesn't load them or whatever" is not a normal day to day situation though
[14:33] <seb128> dobey, yes, if the install is corrupted you will get bugs
[14:33] <seb128> it's the same if some lib get corrupted on disk
[14:33] <walters> sorry for the user-ish question but i'm curious -  so i have a Lucid VM and I'm getting a bunch of packages prominently listed as NOT AUTHENTICATED in update-manager
[14:34] <seb128> walters, seems like you don't have the gpg key that sign the repository for those
[14:34] <dobey> seb128: yes, but it still shouldn't crash. and Our glib patch makes it not crash.
[14:34] <walters> seb128, is the gpg key not installed by default?
[14:34] <seb128> walters, what does "apt-key list" say?
[14:35] <seb128> walters, it should
[14:35] <seb128> walters, is that the normal archive or a ppa? how did you add it?
[14:35] <walters> seb128, http://pastebin.mozilla.org/1347119
[14:35] <walters> i haven't installed any PPAs
[14:35] <walters> i could probably revert to my zygote to double check this is reproducible
[14:36] <walters> (zygote = immediately after installing a VM i clone it and leave the original untouched)
[14:36] <seb128> mvo, ^
[14:36] <walters> it's a pretty scary dialog so
[14:37] <seb128> walters, well, try to redo a sudo apt-get update or refresh in update-manager
[14:37] <walters> incidentally i wrote (with a friend) one the original apt-get/gpg integration, so it's interesting to see it today =)
[14:37] <seb128> walters, could you copy the error you get or screenshot the dialog?
[14:37] <seb128> hehe
[14:38] <walters> http://pastebin.mozilla.org/1347120
[14:42] <walters> http://pastebin.mozilla.org/1347125 has a bit more info, i'm looking for logs
[14:45] <seb128> walters, what about sudo apt-get update?
[14:45] <walters> oh man
[14:45] <walters> sorry, i totally forgot about the difference
[14:45] <walters> =)
[14:46] <walters> indeed, apt-get upgrade is now happy - maybe update-manager is just working on stale data?
[14:47] <walters> anyways thanks, and sorry for the irc noise
[14:47] <seb128> walters, no worry
[14:47] <seb128> walters, well sometimes that happens here when I undo an index update
[14:47] <seb128> walters, could also happen if the mirror you use is having some issues or you get some connection issue
[14:48] <seb128> walters, mvo would know better
[14:48] <xclaesse> seb128, is there an easy way to import a debian package into a ppa ?
[14:48] <xclaesse> can it be automated, or does it need changes in the packaging?
[14:48] <seb128> walters, update-manager should be happy again if the command line, usually the index refresh fixes those transient bugs
[14:49] <seb128> xclaesse, you can dget the .dsc, sign it and upload to the ppa
[14:49] <seb128> xclaesse, i.e dget the_dsc_url; do a no change source build and upload
[14:50] <chrisccoulson> pitti - g'ah, bug 858683 exists for natty upgraders still, now that natty has the same firefox version as oneiric :/
[14:50] <ubot2> Launchpad bug 858683 in firefox "Lost Search engines on upgrade" [Critical,Fix released] https://launchpad.net/bugs/858683
[14:50] <xclaesse> seb128, ok let's try. thanx
[14:50] <seb128> xclaesse, yw
[14:53] <dobey> chrisccoulson: oh yay, another upgrade/migration bug :)
[14:55] <pitti> seb128: is there a successor for gnome-open (for URLs), or should programs just call xdg-open?
[14:55] <seb128> pitti, gvfs-open, but better to use xdg-open since it's cross desktop
[14:55] <pitti> seb128: thanks
[14:55] <seb128> pitti, xdg-open will call gvfs-open on GNOME
[14:55] <seb128> yw
[14:58] <flacoste> hi
[14:59] <flacoste> since my last oneiric update, i've got a very weird problem with skype
[14:59] <chrisccoulson> pitti - did you see my earlier comment? not sure what i should do, and i have to dash in a couple of minutes ;)
[14:59] <seb128> hey flacoste
[14:59] <flacoste> and i'm not really sure where to start looking
[14:59] <seb128> flacoste, define "very weird problem" to start? ;-)
[14:59] <flacoste> sure :-)
[14:59] <flacoste> sounds work with for me
[14:59] <flacoste> i'm hearing other people
[14:59] <pitti> chrisccoulson: sorry, I'm not familiar with that bug; why can't the oneiric fix be applied to the natty version as well?
[15:00] <flacoste> but they hear me like if i sounded as "max headroom"
[15:00] <flacoste> or a remix of electronic music
[15:00] <rickspencer3> chrisccoulson,  hey man
[15:00] <flacoste> mumble and empathy works fine
[15:00] <rickspencer3> oops, sorry to interupt
[15:00] <chrisccoulson> pitti - this is the "search engines lost on upgrade" issue due to a change i made in oneiric last week, but it's broken currently for natty -> oneiric upgraders still
[15:00]  * rickspencer3 waits for flacoste to finish
[15:01] <flacoste> rickspencer3: roger
[15:01] <chrisccoulson> i'm just asking your opinion on what i should do, seeing as i'm still on vacation and about to disappear again in a couple of minutes :)
[15:01] <pitti> chrisccoulson: could you prepare an SRU next week?
[15:01] <rickspencer3> chrisccoulson, who else has the knowledge to fix it?
[15:01] <rickspencer3> pitti, sru?
[15:01] <pitti> chrisccoulson: or are you still on vacation then?
[15:02] <rickspencer3> I think we should fix that before ship, surely
[15:02] <rickspencer3> that's seems like a glaring issue
[15:02] <pitti> rickspencer3: well, I suppose it needs to be fixed in oneiric, we can't retroactively fix it in natty
[15:02] <pitti> rickspencer3: but it's not an install problem, but an upgrade problem
[15:02] <rickspencer3> right
[15:02] <pitti> rickspencer3: so a 0-day SRU will work just as well
[15:02] <rickspencer3> so, natty users are not hitting, it right?
[15:02] <chrisccoulson> rickspencer3, i fixed it right now, i just need to upload it - http://bazaar.launchpad.net/~mozillateam/firefox/firefox.head/revision/876
[15:02] <rickspencer3> only oneiric users
[15:03] <rickspencer3> pitti, it just seems like such a glaring issue
[15:03] <rickspencer3> I think it would be a really bad wart to ship
[15:03] <seb128> flacoste, not sure somebody will be able to help you there, seems like a skype issue, not something we have control on
[15:03] <pitti> rickspencer3: it doesn't affect the install images
[15:03] <rickspencer3> oh
[15:03] <flacoste> seb128: well, it was broken by a new update
[15:03] <flacoste> seb128: skype wasn't updated
[15:03] <flacoste> i'd suspect pulseaudio or kernel
[15:03] <pitti> rickspencer3, chrisccoulson: but anyway, as you have a fix right now, fine for me to fix it for release; I just wanted to take the pressure out for fixing it
[15:03] <flacoste> or something else like that
[15:04] <seb128> flacoste, try booting an older kernel? if it's pulse check with TheMuso when he's online
[15:04] <chrisccoulson> pitti - yeah, although the fix is untested, and i don't really have time to test it right this moment
[15:04] <flacoste> seb128: ok, thanks for the pointer, i'll try that
[15:04] <pitti> chrisccoulson: so, let's do a 0-day SRU next week
[15:05] <pitti> upgraders won't notice if it's -updates or final
[15:05] <rickspencer3> pitti, so when users upgrade, on upgrade, they will only have Ask for a search engine?
[15:05] <rickspencer3> oh, I see
[15:05] <rickspencer3> nm
[15:05] <pitti> rickspencer3: no, they'll upgrade straight to the oneiric-updates version
[15:05] <rickspencer3> this is why I should never talk
[15:05] <kenvandine> hehe
[15:05] <pitti> the trick is to have it in -proposed next week, test it, and move it to -updates at the same time as release
[15:06] <pitti> that's the standard recipe for fixing upgrade-only bugs without jeopardizing the image builds
[15:06] <chrisccoulson> pitti - ok, that sounds fine
[15:06] <chrisccoulson> i'm annoyed with myself that i didn't think about the natty -> oneiric upgrade path ;)
[15:06] <rickspencer3> kenvandine, hey, I thought that the "Nautilus crashes when there is a U1 widget" bug was fixed
[15:06] <kenvandine> james_w, what is the best way to fix an outdated source package branch?
[15:06] <chrisccoulson> and even more annoyed that i discovered it whilst on vacation ;)
[15:06] <rickspencer3> hmmm, maybe I need to restart my session
[15:06] <kenvandine> rickspencer3, it is...
[15:07] <kenvandine> you do need to restart, but you also might not have gotten the update yet
[15:07] <james_w> kenvandine, depends on why it is outdated
[15:07] <james_w> kenvandine, which package is of interest here?
[15:07] <kenvandine> mirrors might not be caught up
[15:07] <kenvandine> james_w, telepathy-mission-control-5
[15:07] <kenvandine> it is way out dated, and i hate working on packages without bzr :-D
[15:08] <kenvandine> rickspencer3, make sure you have ubuntuone-client-gnome 2.0.1
[15:08] <james_w> kenvandine, hmm, not sure why that one is out of date
[15:09] <seb128> rodrigo_, the calendar bug, one other issue is that gedit is listed as a calendar application, when it's not ;-)
[15:09] <kenvandine> a merge-upstream blows up horribly
[15:09] <rodrigo_> seb128, yes, just saw it
[15:09] <rodrigo_> seb128, seems to be a bug in g_app_info, since it just has text/plain in its .desktop file
[15:10] <james_w> kenvandine, oneiric?
[15:10] <kenvandine> yes
[15:10] <kenvandine> merge conflicts all over the place
[15:10] <james_w> I don't see it as out of date: http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/oneiric/telepathy-mission-control-5/oneiric/changes/24?start_revid=24
[15:11] <james_w> 1:5.9.1-0ubuntu2
[15:11] <kenvandine> oh... wtf
[15:11] <kenvandine> maybe the alias is broken?
[15:11] <pitti> dobey: seems bug 817656 has reports that it's still happening even with the cherrypicked fix?
[15:11] <ubot2> Launchpad bug 817656 in couchdb "[FFE] Please merge couchdb 1.1.0-1 (main) from Debian unstable (main)" [High,In progress] https://launchpad.net/bugs/817656
[15:12] <kenvandine> james_w, well nm then...
[15:12] <kenvandine> it is fine... on tuesday it checked out a 5.8 version
[15:13] <james_w> kenvandine, ah, that's possible
[15:13] <kenvandine> actually 5.6, it was revision 8
[15:13] <kenvandine> ha... and a bzr pull now updates
[15:14] <kenvandine> anyway... i am happy now :)
[15:15] <rodrigo_> seb128, it seems to be a bug in gtk_app_chooser_button
[15:15]  * rodrigo_ looks
[15:16] <seb128> rodrigo_, or the calendar type is a subtype of text and it consider subtypes?
[15:17] <rodrigo_> maybe
[15:18] <GunnarHj> rodrigo_: Hello Rodrigo, I understand that the purpose of https://code.launchpad.net/~gunnarhj/gnome-control-center/non-utf8-confusion/+merge/78140 wasn't all that clearly explained. :(  I added a comment; let me know if you need further clarification.
[15:19] <rodrigo_> GunnarHj, ok, looking
[15:20] <rodrigo_> GunnarHj, I still don't think that fix is needed, given the fixes in lightdm, right?
[15:21] <GunnarHj> rodrigo_: No, they are completely unrelated.
[15:21] <rodrigo_> hmm, I misunderstood then, pitti?
[15:21] <pitti> rodrigo_: right, seems I originally misunderstood it as well
[15:22] <GunnarHj> rodrigo_: If you want to see it for yourself, you can do e.g. 'sudo locale-gen en_GB' and open User Accounts.
[15:24] <rodrigo_> so, we don't want to show the codeset if it's not UTF-8?
[15:25] <rodrigo_> if so, why not remove the check, rather than commenting out the call to get the values and then checking for values we never change?
[15:26] <GunnarHj> rodrigo_: It's my understanding that we shouldn't deal with non-utf8 locales from the UI at all.
[15:26] <rodrigo_> the check is: if (!is_utf8 && codeset_code) {
[15:26] <rodrigo_> GunnarHj, ok, then we need a different fix, which is to ignore non-utf8 locales when loading them, right?
[15:28] <rodrigo_> in fact, we already do -> add_locale (names[cnt].name, TRUE);
[15:28] <GunnarHj> rodrigo_: Well, in Ubuntu that's what we are doing already. I haven't checked out g-c-c that carefully, but yes, if you load non-utf8 locale names, it should be fixed. At least if I have understood it correctly - you'd better confirm this with somebody before going on.
[15:29] <rodrigo_> that TRUE means to only load utf8 locales
[15:31] <GunnarHj> rodrigo_: No, that's just an initial value of the variable. Then it checks if the locale name represents a non-utf8 locale, and if it does, changes it to FALSE. The problem exists if your 'locale -a' includes non-utf8 locales.
[15:32] <GunnarHj> rodrigo_: Why not install that British latin1 locale? ;-)
[15:32] <rodrigo_> GunnarHj, yes, we load non-UTF8 locales from /usr/share/gdm/locale.alias , although that's empty for me
[15:32] <rodrigo_> is that the case for the bug?
[15:33] <GunnarHj> rodrigo_: Now I don't quite follow you ... Exactly what do you mean by loading?
[15:33] <rodrigo_> we add them to the list
[15:34] <rodrigo_> to show to the user
[15:34] <GunnarHj> rodrigo_: Ok.
[15:34] <rodrigo_> oh, but:
[15:34] <rodrigo_> #ifdef WITH_INCOMPLETE_LOCALES
[15:34] <rodrigo_>         if (utf8_only) {
[15:34] <rodrigo_>                 if (locale->territory_code == NULL || locale->modifier) {
[15:34] <rodrigo_>                         g_debug ("Ignoring '%s' as a locale, since it lacks territory code or modifier", name);
[15:34] <rodrigo_>                         gdm_locale_free (locale);
[15:34] <rodrigo_>                         return FALSE;
[15:34] <rodrigo_>                 }
[15:34] <rodrigo_>         }
[15:34] <rodrigo_> #endif
[15:34] <rodrigo_> I guess that's a type, and should be #ifndef WITH...
[15:35] <GunnarHj> rodrigo_: I'm not enough of a C programmer to have an opinion, I'm afraid.
[15:35] <rodrigo_> checking...
[15:38] <GunnarHj> rodrigo_: For language-selector and the Ubuntu UI of AccountsService we don't load any non-utf8 locales.
[15:39] <GunnarHj> rodrigo_: Are you otherwise using gdm-languages.c to look for available locales?
[15:39] <rodrigo_> yes, that's what g-c-c does
[15:40] <GunnarHj> I see.
[15:45] <GunnarHj> rodrigo_: When calling add_locale(), collect_locales_from_archive() sets 'utf8_only' in add_locale() to TRUE. Doesn't that mean that possible non-utf8 locale names are skipped?
[15:45] <rodrigo_> yes
[15:46] <GunnarHj> rodrigo_: So then you don't need to worry about that side of it, do you?
[15:49] <GunnarHj> rodrigo_: The reason why the proposed fix is needed for Ubuntu is that it's not complete locale names that are passed to gdm_get_language_from_name(), but it is language denotations such as 'en_GB' or 'de'.
[15:51] <pitti> good night everyone!
[15:51] <GunnarHj> rodrigo_: So if there happens to be a locale with an alias that equals the language that is passed to it, it results in the wrong label.
[15:51] <GunnarHj> pitti: Good night!
[15:52] <rodrigo_> GunnarHj, so, we don't want to show the codeset for any locale at all?
[15:52] <GunnarHj> rodrigo_: Correct.
[15:53] <GunnarHj> rodrigo_: Since we don't even bother to load the non-utf8 locale names, showing codeset should never be applicable.
[15:56] <GunnarHj> rodrigo_: So commenting out that function call (or something else to the same effect) should be completely harmless.
[16:04] <dobey> pitti: the "book opening" error isn't in couchdb. it's an issue with newer e-d-s. i thought rodrigo_ fixed it though
[16:05] <rodrigo_> dobey, what? the same problem again?
[16:05] <rodrigo_> GunnarHj, I'm looking, will answer in a bit :)
[16:05] <dobey> rodrigo_: the weird "book already opening" was fixed in couchdb-glib right?
[16:06] <GunnarHj> rodrigo_: No hurry.
[16:07] <rodrigo_> GunnarHj, you're using en_GB?
[16:08] <rodrigo_> dobey, yes
[16:08] <GunnarHj> rodrigo_: Not really. I just downloaded it to understand myself what Colin was talking about. :)
[16:08] <GunnarHj> s/downloaded/installed/
[16:08] <dobey> rodrigo_: seems some people still say they get it. though it also seems like some of them may be in other backends
[16:09] <rodrigo_> ok, so the problem seems to be really the en_GB locale (and maybe others), which uses ISO-8859-1 codeset instead of UTF-8
[16:09] <GunnarHj> Yes.
[16:09] <rodrigo_> there is a en_GB.UTF-8 locale also, which is the one that should be used, right?
[16:09] <GunnarHj> Absolutely.
[16:10] <rodrigo_> so, why do we have en_GB and not en_GB.UTF-8?
[16:10] <GunnarHj> pls define "we have"
[16:10] <rodrigo_> with your patch, we'll show no codeset, but the locale will be en_GB when selected
[16:11] <rodrigo_> right?
[16:11] <rodrigo_> and that's wrong, isn't it?
[16:11] <GunnarHj> Well, no. en_GB will be passed to AccountsService's SetLanguage, which does the right thing.
[16:12] <rodrigo_> oh
[16:12] <rodrigo_> it sets it to *.UTF-8?
[16:12] <GunnarHj> Yep.
[16:12] <GunnarHj> Always UTF-8.
[16:12] <rodrigo_> ok
[16:13] <rodrigo_> but we need a better patch anyway, you are just commenting the call, but then doing 2 checks for the variables
[16:13] <rodrigo_> so all that should be removed, from the language_name_get_codeset_details call, to the out: line
[16:13] <rodrigo_> GunnarHj, I'll do it, don't worry
[16:14] <mdeslaur> seb128: Is bug 863038 being targeted to get fixed?
[16:14] <ubot2> Launchpad bug 863038 in gnome-screensaver "If turnoff the screen setting set to 'never' screen turns off instantly" [High,In progress] https://launchpad.net/bugs/863038
[16:14] <GunnarHj> rodrigo_: Sure, I have no problem if you delete a few lines instead, if you prefer that. Just thought that that kind of cleanup may be a separate exercise.
[16:17] <mterry> tedg, heyo.  indicator-appmenu question for you.  I'm trying to find where "entry->label" every get inserted into a container.  (for the IndicatorObjectEntry objects appmenu creates)
[16:20] <tedg> mterry, I believe that it's in unity-panel-service
[16:21] <mterry> tedg, I looked, but all I could see was it hiding/showing the object.  I didn't see a container add
[16:22] <tedg> mterry, Hmm, it might not...
[16:22] <tedg> mterry, indicator-applet does.
[16:25] <rodrigo_> GunnarHj, ok, so it's the other part of debian/patches/52_ubuntu_language_list_mods.patch that makes it show the codeset
[16:25] <rodrigo_> GunnarHj, the upstream version doesn't
[16:25] <rodrigo_> also, upstream version only shows languages for the language selection, not locales
[16:26] <rodrigo_> GunnarHj, can you please build the g-c-c package with that patch disabled and see what do we need from the patch please?
[16:28] <GunnarHj> rodrigo_: With patch 52 disabled?? What do you expect to learn from that?
[16:28] <rodrigo_> GunnarHj, that the patch makes the upstream code behave differently :) ?
[16:28] <rodrigo_> differently = not deal correctly with the non-utf8 case
[16:30] <rodrigo_> GunnarHj, so, given we did several changes in upstream, I'm asking you to test without that patch to see if we're missing something
[16:30] <rodrigo_> to remove the patch, if possible
[16:31] <rodrigo_> the list of langs I see with or without the patch are the same, so not sure it's needed at all
[16:32] <rodrigo_> also, upstream just shows generic language names, and only English (country) when adding new ones
[16:33] <GunnarHj> rodrigo_: Hmm... The patch was essential to make it work consistently with language-selector. But sure, I can build without it and take a look, if you like.
[16:34] <rodrigo_> yes, please
[16:35] <GunnarHj> But doing such changes now, one day before the release candidate, does not appeal to me. (But I'll check.)
[16:35] <rodrigo_> well, we can do it as a SRU
[16:35] <rodrigo_> it looks to me like we're trying to patch what we broke with the patch
[16:37] <GunnarHj> With my Ubuntu eyes we didn't break anything with the patch - we prevented it from breaking. :)
[16:38] <rodrigo_> GunnarHj, I'm running the vanilla upstream, and nothing is broken AFAICS
[16:38] <rodrigo_> and the list of languages is smaller on the initial choice
[16:39] <GunnarHj> Ok. Disconnecting now to build and check it out.
[16:45] <didrocks> ok, time for dancing!
[16:45] <didrocks> see you guys
[17:16] <seb128> mdeslaur, rodrigo_ is supposed to work on it
[17:16] <seb128> not sure how much progress he did
[17:16] <seb128> rodrigo_, how is the gnome-screensaver diming immediatly on "never" going?
[17:19] <jbicha> seb128: https://code.launchpad.net/~jbicha/ubuntu/oneiric/jockey/update-help-link/+merge/77782
[17:19] <seb128> jbicha, thanks!
[17:20] <seb128> pitti, ^
[17:20] <seb128> jbicha, (pitti is upstream)
[17:20] <jbicha> I asked him earlier today about it but I probably didn't speak up loud enough :)
[17:21] <seb128> jbicha, days tend to be busy with pings for pitti :p
[17:21] <seb128> jbicha, I will make sure with him it's getting merged tomorrow morning
[17:21] <ogra_> it works if you wear a pink shirt
[17:56] <pulb> hi,  I've found a bug in oneiric which is pretty critical for me but fixing should take minutes (I attached a fix). could someone responsible please have a look and predict if a fix will make it into oneiric? https://bugs.launchpad.net/ubuntu/+source/gconf2/+bug/864615
[17:56] <ubot2> Launchpad bug 864615 in gconf2 "Thumbnailers are missing in gconf" [Undecided,New]
[17:59] <dobey> pulb: i don't see any reason for that to go into oneiric. it isn't a fix
[17:59] <seb128> pulb, it's wrong
[17:59] <seb128> oneiric doesn't use gconf
[17:59] <seb128> nautilus uses gsettings
[17:59] <dobey> right, libgnome-desktop uses gsettings
[17:59] <pulb> seb128: i'm not talking about nautilus
[17:59] <seb128> gnome-vfs is also deprecated for years
[18:00] <seb128> well, gconf is deprecated
[18:00] <seb128> nautilus and other GNOME component use gsettings
[18:00] <dobey> as is static python bindings
[18:00] <pulb> seb128: so all gtk2 that make use of that api are deprecated as well=
[18:01] <pulb> dobey: thats just a python example
[18:01] <seb128> yes
[18:01] <pulb> seb128: sure..
[18:01] <seb128> you should port your code
[18:01] <dobey> pulb: and it is using deprecated code
[18:01] <pulb> guys, gtk3 bindings arent available for all languages yet..
[18:02] <kenvandine> it isn't just about gtk2 vs gtk3
[18:02] <pulb> and if this is your opinion, why do you ship gconf in the first place?
[18:03] <dobey> pulb: what do you mean by all languages?
[18:03] <dobey> pulb: there aren't going to be static python bindings
[18:03] <dobey> use gobject-introspection
[18:04] <pulb> dobey: i'm actually not using python
[18:04] <kenvandine> gtk2 apps can't use deprecated APIs forever
[18:04] <seb128> pulb, gconf still works as it used to work
[18:04] <jbicha> pulb: for one thing, you haven't pointed us to an app in the repositories that's broken, we can't take full responsibility for third party apps using deprecated code
[18:04] <kenvandine> even gtk2 apps need to be ported if they use the deprecated apis
[18:04] <seb128> pulb, it's just that thumbnailers migrated to gtk3 and gsettings
[18:04] <dobey> kenvandine: yes they can, otherwise they won't be gtk2 apps any more :)
[18:05] <flacoste> how can i change the application used to open PDF application?
[18:05] <flacoste> i changed it to okular, but that association was lost by a recent upgrade
[18:05] <dobey> flacoste: right-click on a pdf in nautilus and choose "open with" and select your app
[18:05] <kenvandine> dobey, i am not ready to say gtk2 is deprecated :)
[18:05] <flacoste> dobey: how can i change the default
[18:05] <flacoste> ?
[18:05] <pulb> but the only thumbnail api available in gtk2(-sharp) is broken now: http://developer.gnome.org/gnome-desktop/stable/GnomeDesktopThumbnailFactory.html
[18:06] <flacoste> i remember a "default application" settings, but can't find it anymore
[18:06] <dobey> flacoste: in the dialog that pops up doing what i said. choose "other"
[18:06] <pulb> there are now gtk3 bindings for c# by now for example
[18:06] <kenvandine> that is from libgnomeui though...
[18:06] <dobey> flacoste: that hasn't been around since gnome 1.x
[18:06] <dobey> err, well, not one that set a pdf viewer anyway
[18:06] <flacoste> dobey: well, i remember being able to change that in gnome 2 at least
[18:07] <dobey> pulb: the API isn't broken
[18:07] <dobey> pulb: there are just no thumbnailers available
[18:08] <pulb> dobey: and thats what the bugreport is about
[18:08] <dobey> flacoste: afaik, the only way to set the default handler for pdf files has been through nautilus, or by hand
[18:08] <pulb> all apps that make use of that api are broken in oneiric
[18:08] <dobey> pulb: it's not a bug
[18:08] <pulb> so?
[18:08] <kenvandine> pulb, what apps use it?
[18:08] <dobey> at least, it's not a bug in ubuntu afaict
[18:08] <flacoste> dobey: ah, thanks, i found a "Set as default" in nautilus
[18:09] <dobey> it is a bug in some app that apparently requires some other specific version of an app to be installed
[18:09] <flacoste> that was the pointer i needed
[18:09] <dobey> flacoste: no problem
[18:09] <pulb> kenvandine: e.g. https://launchpad.net/basenji
[18:10] <jbicha> pulb: oh, it's your app
[18:10] <kenvandine> that isn't oneiric though
[18:10] <kenvandine> not that it means it should be broken :)
[18:11] <kenvandine> we just have no way of knowing it is using the api
[18:11] <pulb> kenvandine: thats the problem i'm packaging it for ubuntu/debian
[18:11] <pulb> then i stumbled about that bug
[18:11] <dobey> how is that app broken?
[18:12] <pulb> packages in the daily ppa won't work in oneiric properly
[18:12] <dobey> specifics please.
[18:12] <pulb> no more previews for indexed media anymore
[18:13] <pulb> as seen here: http://www.shockshit.net/basenji/screenshots/basenji05-mainwindow_2.png
[18:15] <dobey> pulb: do you have libgnome2-common installed?
[18:16] <pulb> dobey: hmm have to check that
[18:17] <dobey> was your screenshot link a working, or non-working, example?
[18:20] <pulb> dobe libgnome2-common is installed
[18:21] <pulb> that sreenshot shows a  working example :-)
[18:22] <pulb> dobey: but pictures are actually still working, but movies, documents etc arent
[18:23] <dobey> right, because images are done inside gtk+ directly
[18:23] <dobey> wel, most images are
[18:23] <dobey> some aren't
[18:24] <dobey> however, it seems your problem is a blind reliance on functionality from specific versions of external applications
[18:24] <pulb> its a reliance of GtkThumbnailFactory
[18:24] <dobey> no, because that works fine
[18:24] <pulb> the responsible apps are actually installed
[18:25] <pulb> (e.g totem-thumbnailer)
[18:25] <pulb> dobey: not in gtk2
[18:25] <dobey> yes it does
[18:25] <dobey> works, and gives you the results you think you should get, are not necessarily the same
[18:26] <dobey> if you think totem 3.x should install settings for old gconf still, perhaps you should file upstream bug reports
[18:26] <dobey> as it seems that is where your problem lies
[18:27] <dobey> but i suspect the answer will be "port your app to the new stuff"
[18:27] <dobey> granted, the upstream fix is trivial
[18:27] <pulb> dobey: there are no gtk3 bindings yet
[18:28] <pulb> thats why even banshee still uses gtk2
[18:28] <dobey> not entirely true
[18:28] <pulb> ok, but its wip
[18:28] <pulb> not packaged
[18:28] <dobey> there is a gtk3 port of banshee
[18:29] <pulb> dobey: yes but the dependencies (gtk3-sharp) are not packaged yet
[18:30] <dobey> yes, but your bug against gconf2 isn't an issue with gconf, and your hack that makes it work for you, isn't a fix
[18:31] <dobey> the api you are using works fine. it just does not give the same results you got running on a gnome2 system because gnome3 changed things
[18:31] <pulb> dobey: it's not a hack, those are proper keys comming with natty
[18:31] <dobey> it's a hack. all you did was dump your gconf tree
[18:31] <pulb> its not gnome3 related, its a incomplete gnom2 stack install
[18:32] <pulb> the proper gconf tree
[18:32] <dobey> oneiric doesn't use gnome 2
[18:32] <dobey> your problem is you have totem 3.x instead of totem 2.x
[18:32] <pulb> it shippes the gnome2 api, side by side to gnome3
[18:33] <pulb> no, its working with totem 3
[18:33] <pulb> its just the missing mapping in gconf
[18:34] <dobey> no, totem 3 doesn't use gconf
[18:34] <dobey> because it is deprecated
[18:35] <seb128> dobey, well it doesn't need to use gconf
[18:35] <pulb> true, it doesnt, but GtkThumbnailFactory uses gconf to get the mapping from mimetype to totem 3
[18:35] <seb128> the api just look to gconf keys and call the commands in there
[18:35] <dobey> seb128: it could probably install the schema for the thumbnailers still though
[18:35] <seb128> right
[18:35] <seb128> pulb, well in any case it's not a bug in gconf
[18:35] <dobey> seb128: and it doesn't becuase gconf is dead as far as gnome is concerned
[18:36] <seb128> pulb, it's a request for each thumbnailer which switched to gsettings to ship back a gconf schemas with those keys
[18:36] <pulb> dobey: you cant say its dead yet
[18:36] <pulb> there are a lot apps relying on it, even gtk3 apps
[18:36] <dobey> pulb: upstream gnome does not use gconf any more.
[18:37] <seb128> pulb, well they stopped using gconf so they will not add back gconf schemas
[18:37] <seb128> we need a gconf-thumbnailer-schemas-compat source
[18:37] <pulb> ok, but that does not appy to all other apps out there
[18:37] <dobey> and totem doesn't have to provide those. mplayer could install a thumbnailer just as well as totem could
[18:38] <seb128> pulb, well it's an issue but we are in hard freeze and it's late now to fix it
[18:38] <pulb> i'm just saying that those thumbnailers should still be available to the gtk2 apu
[18:38] <pulb> api
[18:39] <kenvandine> which is why seb128's solution would work
[18:39] <pulb> because currently the gtk2 GThumbnailFactory is broken without the mapping
[18:39] <kenvandine> but we are in hard freeze now
[18:39] <dobey> pulb: it is NOT broken.
[18:39] <kenvandine> dobey, choice of words, it doesn't find any thumbnailers
[18:39] <pulb> kenvandine, seb128: ok, do you think there will be a fix after release?
[18:40] <dobey> kenvandine: well, it finds the ones that are registered with it
[18:40] <seb128> pulb, we need to think to what the proper way to fix that is
[18:40] <kenvandine> dobey, yes... agreed
[18:40] <dobey> kenvandine: the fact that nothing is registering with it, doesn't mean it is broken
[18:40] <dobey> kenvandine: it works the same as it did in lucid, if you didn't have any of those apps installed.
[18:40] <kenvandine> dobey, you and pulb are just looking at the same coin from opposite sides :)
[18:40] <seb128> dobey, right, well what he's saying is that i.e totem should still register in gconf for all applications using the deprecated thumbnailer api
[18:40] <pulb> dobey: let's call it... useless
[18:43] <seb128> hum
[18:43] <seb128> not sure how to solve that best
[18:43] <seb128> but I don't want to add gconf schemas back to application that stop using gconf
[18:43] <seb128> especially if we try to drop gconf from the default installation next cycle
[18:44] <dobey> seb128: SRU partial schema back in for thumbnailer bits only as a vendor patch?
[18:45] <seb128> dobey, still that would add a depends on gconf to those binaries since they need to register the schemas
[18:45] <pulb> seb128: oneirci+1 shouldn't be affected, because up 2 date langbindings should be available by then. its up to the developers to fix their apps then
[18:45] <dobey> seb128: do they need a Depends? just having the file should trigger the gconf on postinst right?
[18:45] <pulb> what about a gnome-2-legacy package that fixes those issues?
[18:46] <dobey> because it won't fix those issues
[18:46] <dobey> it really has to be done in each app that provides a thumbnailer
[18:46] <mpt> If Nautilus has a "Segmentation fault (core dumped)", where can I find that core?
[18:47] <dobey> mpt: in the .crash file for it in /var/crash ?
[18:55] <flacoste> any reason why the 'Spelling menu' now be disabled in thunderbird?
[19:02] <seb128> flacoste, what spelling menu?
[19:14] <flacoste> seb128: in the compose window, thre is a spelling icon allowing to select language
[19:14] <flacoste> it's now disabled
[19:14] <flacoste> but this was working before my last update
[19:32] <seb128> flacoste, could be an issue with the translations updates, not sure
[19:32] <seb128> chrisccoulson is out this week
[19:32] <seb128> tb didn't change but we got new langpacks yesterday
[19:32] <kenvandine> now i need to google pangolin
[19:34] <micahg> flacoste: wfm, there needs to be some text in teh context you're in
[19:36] <jbicha> kenvandine: but at least I didn't need a dictionary for the adjective like I did for Natty & Oneiric
[19:36] <kenvandine> true
[19:41] <flacoste> micahg, seb128: PEBKAC, it's just disabled when the cursor is in the subject, the menu works fine when i click in the body
[19:41] <flacoste> so i guess thunderbird doesn't allow you to customize the dictionary for the message subject
[19:41] <flacoste> actually, right-clicking allows me to change the language in the subject
[19:41] <flacoste> so a minor ui bug i guess
[19:43] <micahg> flacoste: sure, feel free to file it
[20:08] <GunnarHj> rodrigo_: Still there?
[20:36] <flacoste> how does one access the boot menu?
[20:36] <flacoste> esc doesn't seem to work anymore
[20:36] <flacoste> also, 'Restart' not actually working is a known bug right?
[20:37] <kenvandine> restart works for me
[20:37] <kenvandine> flacoste, you hold down the shift key
[20:37] <kenvandine> for grub
[20:37] <flacoste> shift!
[20:37] <flacoste> ok thanks
[20:37] <flacoste> i'll give this a try
[20:50] <flacoste> is it possible to see what packages were updated in my recent upgrade?
[20:50] <flacoste> TheMuso: around?
[20:51] <TheMuso> flacoste: Yes.
[20:53] <flacoste> TheMuso: seb128 suggested you might be able to help me, I've got a sound issue in skype after i upgraded a bunch of packages yesterday
[20:53] <flacoste> TheMuso: it was working fine before, and sounds work fine in empathy and mumble after the update
[20:54] <flacoste> TheMuso: for more amusement, Skype Test Call works fine for me, but everyone else hear me like i'm Max Headroom
[20:54] <flacoste> well, a garbled version of Max Headroom
[20:54] <flacoste> i'm running Ubuntu 64bits
[20:54] <TheMuso> flacoste: Ah, yes this is known. There is an issue with the way skype checks for the library version of pulse, and since switching to 1.0, upstream accidentally changed things such that it broke skype. There is a fix in upstrea git, its just a matter of pushing it through as an SRU, or if the release gods think it important, it could be pushed through prior to final release.
[20:54] <flacoste> TheMuso: thank you very much!
[20:55]  * flacoste is very happy to know that skype will be fixed relatively soon
[20:56] <flacoste> TheMuso: is there a bug i can subscribe to?
[20:56] <TheMuso> flacoste: There has been a bug filed about this issue, give me a bit to dig up the bug number./
[20:58] <TheMuso> flacoste: bug 865820
[20:58] <ubot2> Launchpad bug 865820 in pulseaudio "Mic only works when pulseaudio sound settings open with skype" [Undecided,Confirmed] https://launchpad.net/bugs/865820
[21:02] <flacoste> TheMuso: thanks
[21:02] <TheMuso> np
[21:04] <flacoste> there is even a work-around, even better!
[21:15] <flacoste> so my restart not working is bug 838792
[21:15] <ubot2> Launchpad bug 838792 in gnome-session ""Restart" logs out, even when there are no other people logged in" [Undecided,Confirmed] https://launchpad.net/bugs/838792
[21:40] <TheMuso> flacoste: Seems the patch I was referring to has already been uploaded. Please check that you have pulseaudio 1.0-0ubuntu2.
[21:41] <flacoste> TheMuso:
[21:41] <flacoste> pulseaudio:
[21:41] <flacoste>   Installed: 1:1.0-0ubuntu2
[21:41] <TheMuso> hrm ok then.
[21:41] <flacoste> TheMuso: so it doesn't look like it fixes the issue :-/
[21:41] <TheMuso> No.
[21:41] <flacoste> the same work-around works though
[21:41] <flacoste> if i open the sound settings, the other side hears me fine
[21:41] <flacoste> other it doesn't
[21:42] <flacoste> which is why the Skype Test Call was working
[21:42] <flacoste> i was always opening it to make sure that the input/output flags were fine :-)
[21:42] <TheMuso> Right.
[22:30]  * bryceh runs into bug #861527
[22:30] <ubot2> Launchpad bug 861527 in gnome-icon-theme-symbolic "package gnome-icon-theme-symbolic 3.1.4-0ubuntu1 failed to install/upgrade: trying to overwrite '/usr/share/icons/gnome/scalable/status/keyboard-brightness-symbolic.svg', which is also in package gnome-power-manager 3.1.90-0ubuntu1" [Undecided,Confirmed] https://launchpad.net/bugs/861527
[22:35] <bryceh> looks like it's trying to upgrade gnome-icon-theme-symbolic before gnome-power-manager.
[22:38] <bryceh> guess it needs a Breaks
[22:42] <RAOF> I think you mean either a Replaces or a Conflicts, don't you?  Breaks won't prevent it from being unpacked before the upgrade?
[22:42] <bryceh> RAOF, not sure, but from reading 7.3 at http://www.debian.org/doc/debian-policy/ch-relationships.html it sounds like Breaks would do it, no?
[22:43] <bryceh> "When one binary package declares that it breaks another, dpkg will refuse to allow the package which declares Breaks to be unpacked unless the broken package is deconfigured first, and it will refuse to allow the broken package to be reconfigured. "
[22:45] <RAOF> deconfigured doesn't mean its files have been removed from the system, though, right?
[22:46] <bryceh> oh, you're right - "If the breaking package also overwrites some files from the older package, it should use Replaces to ensure this goes smoothly."
[22:47] <bryceh> hmm, actually http://www.debian.org/doc/debian-policy/ch-relationships.html#s-replaces suggests both Replaces and Breaks
[22:48] <jasoncwarner_> hey RAOF , did you see the notes about the bug mark was having last night (in the backscroll)?
[22:48] <jasoncwarner_> robert_ancell: when do you leave?
[22:48] <robert_ancell> jasoncwarner_, 11:30 tomorrow
[22:48] <jasoncwarner_> ah, nice.
[22:49] <jasoncwarner_> seb128 mentioned he needed something from you for lightdm. were you and he able to connect?
[22:51] <robert_ancell> jasoncwarner_, I haven't got any email about anything
[22:52] <jasoncwarner_> didrocks were you and mark able to connect about his problem with Unity starting?
[22:52] <jasoncwarner_> robert_ancell: let me find the backscroll...one sec
[22:53] <RAOF> jasoncwarner_: In the backscroll of #ubuntu-x?  Yeah.  I'll take over from tjaalton if need be (although we do not seem to have got the requested debugging info yet).
[22:53] <jasoncwarner_> robert_ancell: it was the consolekit not starting properly. seb128 , you still online?
[22:54] <jasoncwarner_> Thanks, RAOF
[22:55] <TheMuso> jasoncwarner_: No he is not.
[22:56] <bryceh> RAOF, this look sane?  http://paste.ubuntu.com/703047/
[22:59] <RAOF> bryceh: Looks sane.
[23:01] <bryceh> thanks
[23:04] <RAOF> Oh, urgh.  What's happened to my indicators?  They disappear on ButtonRelease.
[23:06] <bryceh> my wife had a good question... why is the trash icon on the dock bar?
[23:06] <bryceh> you can't seem to drag files onto it (we couldn't anyway)
[23:07] <RAOF> I guess so that you can access it, but failure to be a DND target is a bit weird.
[23:08] <bryceh> yeah that's what we concluded.  But there's a trash icon in the nautilus window easily accessible (which does accept files dragged in), so seems superfluous
[23:08] <bryceh> she thought that it was how you deleted icons from the dock, which I guess would be kinda cool
[23:08] <bryceh> (but bad ui if you need to delete stuff so much that you need a trash can visible to clean it up!)
[23:08] <RAOF> :)
[23:09] <jasoncwarner_> robert_ancell: bug #851345 I think that was one seb128 was talking about.
[23:09] <ubot2> Launchpad bug 851345 in lightdm "compiz crashes on i915 graphic module with GLX_EXT_texture_from_pixmap is missing" [High,Confirmed] https://launchpad.net/bugs/851345
[23:13] <RAOF> Again with the ConsoleKit?
[23:14] <robert_ancell> RAOF, I *think* what may be happening is lightdm has crashed for some reason, and CK automatically removes the CK sessions when that happens, so there is a flow on effect
[23:15] <robert_ancell> jibel has dmesg showing lightdm was killed by SIGABRT, but there's no indication why that is
[23:15] <robert_ancell> I pretty sure the ~/.xsession-errors is a red herring
[23:16] <RAOF> This sounds like a job for gdb.
[23:16] <robert_ancell> I was playing around with killing X in strange ways to see if that caused lightdm to fail, but it seems pretty happy to handle that
[23:17] <RAOF> I wonder if my netbook can reproduce it.
[23:18] <popey> anyone know what package a keyboard layout issue would go against, in GNOME ubuntu 11.10?
[23:19] <bryceh> popey, depends on the type of issue, but xkeyboard-config has the layout data
[23:19] <popey> its a layout issue, none of the layouts actually match the one in front of me
[23:21] <bryceh> popey, https://bugs.launchpad.net/ubuntu/+source/xkeyboard-config
[23:21] <bryceh> popey, see bug #750469 for some directions on getting a layout added
[23:21] <popey> i tried ubuntu-bug xkeyboard-config and it says that doesnt exist
[23:21] <ubot2> Launchpad bug 750469 in xkeyboard-config "can't find my keyboard layout even in 11.04 !!" [Undecided,New] https://launchpad.net/bugs/750469
[23:21] <bryceh> popey, yeah that's the source package name
[23:22] <popey> well, here's the tricky thing
[23:22] <popey> the layout _is_ there
[23:22] <bryceh> try ubuntu-bug xkb-data
[23:22] <popey> but I have to press the key then space to get the character
[23:22] <popey> like " or ~
[23:23] <RAOF> Oh, you've got deadkeys enabled.
[23:23] <bryceh> yeah sounds like it
[23:23] <RAOF> "e results in ?, right?
[23:23] <popey> i get an accented e
[23:24] <popey> ë
[23:24] <RAOF> Yeah, you've got a keyboard layout with deadkeys selected.
[23:24] <bryceh> popey, in Keyboard Preferences > Layouts > Options check Compose key position
[23:24] <popey> none are ticked
[23:25] <bryceh> ok, then it's set in the layout itself
[23:25] <bryceh> popey, so far you've not mentioned what the layout in question actually is?
[23:25] <RAOF> What is the layout that ?Keyboard Layout? says you have?
[23:25] <popey> UK Apple
[23:26] <popey> So choosing English (UK, Macintosh International) gets me the right layout
[23:26] <popey> but with this deadkeys thing
[23:26] <popey> choosing English (UK, Macintosh) it gives me a wrong layout, but no deadkeys
[23:26] <bryceh> yeah 'International' == 'with deadkeys'
[23:26] <popey> is it possible to switch this deadkeys thing off for this layout?
[23:26] <popey> because itś kinda annoying
[23:27] <bryceh> popey, well, that's why I had you check the Layout Options
[23:27] <popey> sorry, should I have ticked something in there?
[23:27] <bryceh> popey, possibly it'll need another macintosh variant
[23:28] <bryceh> popey, no
[23:28] <RAOF> Yeah, sounds like we need a (UK, Macintosh International no dead keys) to go along with all our other no-dead-keys variant.
[23:28]  * bryceh nods
[23:28] <popey> i cant see how any uk english person would want this deadkeys on
[23:28] <popey> i can see a european (fr, de, se) person might for all the accents
[23:29] <RAOF> popey: How else will you spell caf? correctly? :)
[23:29] <popey> hah
[23:29] <RAOF> Or, rather, this sounds like what we *actually* need is for deadkeys to be moved out of the layout and into a toggleable option.
[23:29] <popey> bah, just killed compiz ⍨
[23:30] <popey> I'm guessing this wont happen for 11.10?
[23:30] <RAOF> There's probably an SRU for your compiz crash :/
[23:30] <popey> hah
[23:31] <popey> i kinda meant the keyboard of course ☺
[23:32] <bryceh> popey, I wouldn't say that... xkeyboard-config patches tend to be pretty straightforward and low risk, so no particular reason it shouldn't go into 11.10
[23:33] <RAOF> It'd be pretty easy to do that patch.
[23:33] <bryceh> we might need you to do a bit of text file editing
[23:33] <popey> thats fine
[23:33] <bryceh> the layouts are in /usr/share/X11/xkb
[23:33] <bryceh> trying to find where the macintosh international one is
[23:34] <popey> ./keymap/macintosh
[23:34] <popey> ?
[23:34] <bryceh> very warm
[23:35] <RAOF>  /usr/share/X11/xkb/symbols/us
[23:36] <popey> for uk?
[23:36] <popey> that would be symbols/gb
[23:37] <bryceh> bingo
[23:37] <RAOF> Yup.
[23:37] <popey> lots of dead_ stuff
[23:37] <bryceh> yeah hack on that until you got something you're happy with, then post that file to  a bug report, and ping me or raof
[23:38] <popey> winner
[23:38] <bryceh> we probably ought to run it by upstream before sru'ing it, but that usually goes fairly smoothly
[23:38] <popey> can i make a new layout? based on the existing one?
[23:38] <popey> copy and paste the segment
[23:38] <bryceh> yep
[23:38] <popey> cool, thanks for the help chaps!
[23:38] <bryceh> that should be fine
[23:38] <RAOF> You might want to check out ?English (international AltGr dead keys)? from us
[23:39] <bryceh> let's just doublecheck this work hasn't already been done upstream... one sec
[23:40] <popey> does xkb_symbols need to be unique?
[23:40] <popey> or just the name[group]='foo' bit?
[23:40] <bryceh> not sure offhand
[23:41] <popey> ok, will fiddle
[23:41] <popey> and does x need restarting to take effect?
[23:41] <RAOF> I don't _think_ so.
[23:41] <bryceh> maybe refresh the xkb cache?
[23:41] <bryceh> might need to restart gnome-settings-daemon or something for it to show up in the UI.  not sure
[23:44] <bryceh> yep, looks like we're up to date with upstream on symbols/gb
[23:47] <popey> hmm. logged out and back in
[23:47] <popey> dont see it in the selection list
[23:48] <bryceh> yeah xkb cache probably
[23:48] <dingham> howdy
[23:49] <bryceh> xkb cache files are /var/lib/xkb/*.xkm
[23:50] <popey> gotcha
[23:50] <bryceh> RAOF, do we have a handy way to refresh the xkb cache, or just delete those files?
[23:51] <RAOF> Hm.  X *should* regenerate them when the files they cache change, but that would be on server startup.
[23:51] <dingham> Does anyone here have much experience with pulseaudio & multicast? In particular, why it might affect the packets & routig on my local network whilst it's broadcasting...
[23:52] <dingham> *routing
[23:52] <bryceh> popey, oh btw 'xkbcomp symbols/gb' would be handy for testing your layout
[23:53] <bryceh> might be enough to refresh the cache too, not sure
[23:58] <popey> deleted the xkm files, rebooted
[23:59] <popey> cant see the layout change in the UI
[23:59] <bryceh> this looks cool - http://simos.info/blog/archives/747