[00:00] <asac> feels to me we should do it quickly if we want to do that at all too
[00:00] <asac> TheMuso: that is? yes, lets disable it?
[00:00] <TheMuso> asac: Yes
[00:00] <asac> i dont think daniel said we _should_ do it... just asking
[00:00] <asac> great
[00:00] <asac> TheMuso: do you need anything from me? or want me to ping you to not forget ... let me know!
[00:00] <rickspencer3> asac I suggest that we hold off on -intel for one day to:
[00:01] <rickspencer3> 1. give the x team some breathing room, so they can be well organized
[00:01] <asac> ok agreed. better not put too much speed
[00:01] <TheMuso> asac: Yeah I don't think Daniel would want this done, so its a toss up.
[00:01] <rickspencer3> 2. give us a day delta, in case mesa causes a problem, if we see it before -intel is uploaded, there won't be ambiguity
[00:01] <rickspencer3>  
[00:01] <asac> rickspencer3: to prevent mistakes etc.
[00:01] <rickspencer3> that said, this is not a mandate
[00:01] <rickspencer3> so, whatever the engineers think is best
[00:01] <asac> rickspencer3: yeah. though the 1 day window is probably not evnough.
[00:02] <asac> to distinguish reports
[00:02] <rickspencer3> asac it won't be, true
[00:02] <rickspencer3> unless the issue is severe
[00:02] <asac> but i think the argument is valid that we shouldnt hurry stuff in... better do accurate work
[00:02] <rickspencer3> or widespread
[00:02] <asac> redoing will cost more time
[00:02] <asac> rickspencer3: thanks for clarifying.
[00:02] <asac> just wanted to assure that the schedule is feasible ;)
[00:03] <rickspencer3> sure
[00:03] <asac> (even for -intel ... were regressions are double painful because we always told folks to buy intel if they want good linux support)
[00:03] <rickspencer3> I think the project is off and running now, so I will strive to stay out of the way and let the teams work ;)
[00:03] <asac> ack
[00:04]  * asac wonders if there are initial bits to get already
[00:05] <rickspencer3> asac, which bits are you referring to?
[00:05] <rickspencer3> mesa 7.6 is uploaded
[00:05] <rickspencer3> and -intel is sitting in one of the ppas
[00:06] <rickspencer3> I think tormond may have gotten it ready, but decided to upload in the morning after some rest, not 100% certain though
[00:08] <asac> mesa
[00:08] <asac> thats fine. just looking for mesa now ;)
[00:22] <bryce> asac, right, what Rick said.  :-)  I want to do some practice with reverting mesa before turning focus to -intel.
[00:23] <asac> kk thx
[00:26] <jono> kenvandine, thanks for sending the mail to Christian, I followed up with more details of the problem
[00:29] <kenvandine> jono thx
[01:35] <Amaranth> well crap
[01:36] <Amaranth> proper support for compiz with ati apparently requires KMS now
[01:36] <Amaranth> I guess that means I can ignore these bugs since it'll all work right in lucid :)
[01:37] <Amaranth> $ xbacklight
[01:37] <Amaranth> No outputs have backlight property
[01:37] <Amaranth> :(
[01:40] <johanbr> Amaranth, really? how did that come about?
[01:40] <Amaranth> johanbr: No idea
[01:40] <johanbr> did non-KMS stuff get ripped out of the xorg driver?
[01:40] <Amaranth> ati users are getting a max texture size of 2048 in glxinfo but actual support of only 1024
[01:41] <Amaranth> so compiz starts but large windows (including your desktop) are black
[01:41] <Amaranth> unless you turn on KMS
[01:42] <JanC> Amaranth: I saw in #ubuntu-x that they are thinking about pulling in some newer KMS drivers for ATI
[01:42] <JanC> well, s/thinking/working/
[01:42] <Amaranth> JanC: already done, afaik
[01:42] <Amaranth> I should probably be in that channel though...
[01:42] <JanC> well, and, more new ATI stuff
[01:43] <JanC> Amaranth: yeah, possibly done by now
[01:43]  * Amaranth is in too many channels
[01:43] <JanC> I saw them discussing stuff about ati and newer versions that are supposed to fix some issues etc.  ;)
[07:33] <pitti> Good morning
[08:02] <robert_ancell> dbus is driving me insane...
[08:03] <robert_ancell> pitti, have you ever used a dbus value of type 'ao' before?  I'm trying to work out how to handle it in the glib bindings
[08:03] <pitti> robert_ancell: hey
[08:04] <pitti> robert_ancell: only in Python, I think
[08:04] <robert_ancell> pitti, I know the dbus calls to make for the gdm guest session but I spend hours trying to bend C to actually accomplish them...
[08:06] <pitti> oh, for GetDisplays?
[08:07] <didrocks> hey pitti, good morning robert_ancell
[08:08] <pitti> hm, is 'o' actually a GObject, or just any GPtr which you need to know the type yourself?
[08:08] <pitti> gdm itself just seems to treat ao as a GPtrArray and stuffs strings into it
[08:09] <robert_ancell> pitti, I need to get the CK session to activate it, org.freedesktop.ConsoleKit.Manager.GetSessionsForUnixUser() returns 'ao' - a list of object paths (which are just strings afaik)
[08:09] <robert_ancell> didrocks, hey
[08:10] <robert_ancell> pitti, thanks, that sounds familiar
[08:11] <pitti> robert_ancell: right 'o' means 'object path', not 'any object' according to http://dbus.freedesktop.org/doc/dbus-specification.html#message-protocol-signatures
[08:21] <huats> morning everyone
[08:23] <didrocks> hey huats
[08:24] <huats> plop plop didrocks
[08:24] <huats> :)
[08:28] <seb128> good morning there
[08:30] <mvo> hey seb128
[08:30] <seb128> hello mvo
[08:30] <seb128> how are you?
[08:31] <huats> hey seb128
[08:31] <seb128> lut huats
[08:42] <seb128> pitti, hey, see bug #445846?
[08:44] <chrisccoulson> good morning everyone
[08:44] <seb128> hello chrisccoulson
[08:45] <chrisccoulson> hey seb128
[08:45] <chrisccoulson> how are you today?
[08:46] <seb128> good thanks, you?
[08:46] <chrisccoulson> yeah, not too bad. very tired though!
[08:46] <chrisccoulson> it was a late night last night
[08:51] <seb128> chrisccoulson, same here, yesterday I went to bed to a reasonable time but that didn't make up for the 2am hacking nights all the other days
[08:51] <seb128> chrisccoulson, did you manage to figure how to get the gdm theming working?
[08:53] <chrisccoulson> seb128 - i'm 70% of the way through the g-s-d change for it. the trickiest bit is propagating the gconf prefix string in to the xsettings plugin, but i've done that now (which was most of the work)
[08:53] <seb128> cool
[08:53] <chrisccoulson> tonight, i'll get the plugin to use the customizable location
[08:54] <chrisccoulson> but i think we'll need to make metacity use a different location in the GDM session too, as it seems we're setting a metacity theme as well
[08:54] <seb128> seems that start to me quite some work
[08:54] <pitti> seb128: I saw mdz's reply (have it in my mailbox), but didn't look at it yet
[08:54] <pitti> still processing mail
[08:55] <pitti> hey chrisccoulson
[08:55] <seb128> pitti, same here, I was just pointing the new bug, I did read that one before mdz's comment
[08:56] <chrisccoulson> seb128 - yeah, i think it's quite a lot of work too, but i'm not sure what option we have now. i think once i've done this change, i will start up a discussion on d-d-l to see if anyone can come up with some other suggestions for the next cycle, as this must be an issue that other distributions will face too
[08:56] <chrisccoulson> hey pitti
[08:57] <seb128> chrisccoulson, try on the gdm list rather
[08:57] <chrisccoulson> seb128 - ok, i can do that
[08:57]  * chrisccoulson subscribes to gdm list
[08:58] <seb128> chrisccoulson, thanks
[08:59] <seb128> I'm wondering if we should rather try to make the gconftool calls robust
[08:59] <seb128> ie sudo dbus-launch gconftool || true
[09:00] <seb128> and I think gvfs has a variable or something to not do fuse mounting
[09:00] <chrisccoulson> seb128 - possibly, but will that work around the issues with being prompted for a password?
[09:01] <seb128> chrisccoulson, sudo should not be asking a password or that's what I understood from pitti when he suggested to use sudo rather than su there
[09:01] <seb128> I'm not knowledgable about the pam stack
[09:01] <chrisccoulson> ah, ok. yeah, i'm not sure about that - I'm not knowledgable about that either
[09:02] <chrisccoulson> i suppose it could work, but the only other issue is that it will always overwrite user-defined preferences when the package is upgraded
[09:03] <chrisccoulson> although, i suppose you could check the existing key on upgrade and see if the user modified it
[09:05] <seb128> chrisccoulson, how so? the key is only written once on new installs
[09:06] <seb128> and on upgrades from the version before the gconftool calls
[09:06] <chrisccoulson> oh, ok. i didn't realise that ;)
[09:06] <seb128> we will get an issue if we want to change the default theme some day
[09:07] <mvo> hm, I seem to be unable to find the "give this bug a nickname" button in LP - where is it these days?
[09:07] <chrisccoulson> yeah, that's an issue 2
[09:07] <chrisccoulson> the new GDM is a lot of pain ;)
[09:14] <seb128> gvfs has a GVFS_DISABLE_FUSE variable
[09:17] <seb128> mvo, dunno, I never used that thing, maybe they dropped it because it's not used?
[09:19] <seb128> mvo, bug #416369
[09:19] <mvo> seb128: aha, thanks. *pff* I used it :)
[09:20] <mvo> I can't remember all the bugnumbers otherwise
[09:20] <mvo> oh well
[09:20] <seb128> mvo, use tags?
[09:20] <seb128> I could remember nicknames for bugs either
[09:20] <mvo> is that how you do it? what is your way to remember the gazillion of numbers :) ?
[09:22] <seb128> no, I'm good at using search
[09:22] <seb128> ie I sort by most duplicates for example for frequents bugs
[09:22] <seb128> or I milestone things I want to look at for a cycle and look at the milestoned list
[09:23] <mvo> what I don't like about the search is that it takes some seconds. not long, but long enough to make me anoyed. but I guess that is just me
[09:23] <seb128> ;-)
[09:24] <seb128> you aim for better effenciency than me apparently
[09:24] <seb128> efficiency
[09:24] <mvo> haha
[09:25] <mvo> I need to work on myself to become more calm
[09:27] <seb128> I often do find duplicates using my email client though and search in my bug box
[09:32] <chrisccoulson> seb128 - you can also use GIO_USE_VFS=local to disable gvfs entirely
[09:33] <seb128> right
[11:09]  * seb128 kicks firefox
[11:09] <seb128> asac, greasemonkey breaks every time there is a xulrunner change apparently
[11:10] <seb128> I've to clean the profile dir and download it again
[11:22] <joaopinto> does the new GDM supports themes in anyway ? is it just missing a configurator tool ?
[11:22] <pitti> asac: so should we wontfix the karmic task in bug 401823 then, if the problem isn't actually that bad?
[11:23] <seb128> joaopinto, not in the way the old one did
[11:24] <seb128> joaopinto, it's a gnome-session so it supports theme the same way than GNOME, gtk theme, icon theme
[11:24] <seb128> etc
[11:24] <asac> pitti: yes, invalidated karmic task now for firefox. not sure we want to remove the warning for gtk ... one concern was that it fills up xsessions-errors
[11:25] <joaopinto> seb128, tks, it's becoming a FAQ for Karmic
[11:25] <asac> pitti: hmm. odd i invalidated karmic and the normal task didnt come back
[11:25] <asac> oh thats ajax ... reloading
[11:25] <asac> no. that didnt help. no normal target
[11:26] <gnomefreak> seb128: is there a link for the gtk theme in wiki so i can add it to the bot?
[11:26] <pitti> asac: I fixed the tasks
[11:26] <seb128> tseliot, hey, could you look at bug #446199? that seems a candidate for a karmic milestoned bug
[11:26] <tgpraveen> asac: in ubuntu which spell checking engine does firefox 3.5 use ? is it hunspell?
[11:26] <asac> pitti: what is the trick?
[11:26] <pitti> asac: "invalid" will close the package task entirely; "wontfix" will revive the floating task
[11:26] <asac> ah ;)
[11:26] <asac> kk
[11:26] <tseliot> seb128: yes, sure
[11:26] <seb128> gnomefreak, not that I know about but I'm not a wiki user
[11:26] <pitti> fixed harder now
[11:26] <seb128> tseliot, thanks
[11:27] <asac> so i think we can live with the error output?
[11:27] <gnomefreak> seb128: ok thanks i couldnt find it so maybe ill get it right from gnome if i can
[11:27] <seb128> gnomefreak, I don't understand your question to be honest
[11:27] <seb128> gnomefreak, sudo -u gdm gconf-editor
[11:27] <asac> tgpraveen: yes. its hunspell
[11:27] <gnomefreak> ah thanks seb
[11:28] <seb128> gnomefreak, or sudo  -u gdm gnome-appearance-capplet
[11:28] <seb128> gnome-appearance-properties
[11:28] <seb128> rather
[11:28] <gnomefreak> seems gconf is broken
[11:29] <tgpraveen> asac: and since chrome uses hunspell too so a word if I add to chrome's dictionary then it should be known to firefox too right?
[11:29] <asac> tgpraveen: i think the hunspell dicts for chrome are a bit odd. what problem are you trying to solve?
[11:30] <gnomefreak> thats broke too, guessing its related to gconf but says its dbus
[11:31] <tgpraveen> asac: well basically to make it so that all apps on the desktop use the same dictionary. so am loking at which all apps use which dictionary?
[11:31] <tgpraveen> maybe every app should use hunspell
[11:31] <asac> tgpraveen: chrome is not even a distro package
[11:31] <tgpraveen> it seems to have a lot going for it.
[11:31] <asac> yes. hunspell is standard for most new stuff afaik
[11:31] <tgpraveen> asac: I know but still am looking at all the main/popluar apps including chrome,ff,pidgin,empathy
[11:32] <tgpraveen> anybody knows what pidgin uses?
[11:32] <seb128> gnomefreak, sudo -u gdm dbus-launch gnome-appearance-properties
[11:32] <seb128> gnomefreak, sudo -u gdm dbus-launch gnome-appearance-properties should be working...
[11:32] <gnomefreak> seb128: thanks
[11:32] <seb128> the dbus-launch is needed if there is no session bus for the gdm user
[11:33] <tgpraveen> got it it's aspell
[11:33] <tgpraveen> maybe a target for lucid to get all apps to use same dictionary. that would be awesome!
[11:33]  * tgpraveen looks for/files  abug
[11:34] <seb128> all using the same dict was a target for intrepid?
[11:34] <seb128> everything should be using enchant which uses aspell or something around those lines
[11:35] <tgpraveen> all mozila use hunspell
[11:35] <tgpraveen> chrome also uses hunspell. pidgin uses aspell
[11:35] <tgpraveen> *mozilla software
[11:36] <tgpraveen> openoffice.org uses hunspell
[11:36] <tgpraveen> so this issue still needs work to be done
[11:37] <seb128> https://wiki.ubuntu.com/ConsolidateSpellingLibs
[11:37] <seb128> gnome uses hunspell
[11:37] <tgpraveen> hmm about enchant "Enchant is a free software project developed as part of the AbiWord word processor with the aim of unifying access to the various existing spell-checker software."
[11:37] <seb128> right
[11:37] <seb128> which is what GNOME is using
[11:38] <tgpraveen> and enchant has backends for hunspell/aspell so it should work
[11:38] <tgpraveen> but it aint working for me. I added a word in pidgin and it didn't get added in ff
[11:38] <seb128> pidgin != gnome
[11:39] <seb128> dunno what those guys are using
[11:39] <seb128> seems they use gtkspell
[11:39] <seb128> which uses enchant
[11:39] <tgpraveen> yeah and since it has a backend for enchant too and
[11:39] <seb128> is your work listed in gedit for example?
[11:40] <tgpraveen> since as you say it was a target for intrepid shouldn't it work with pidgin
[11:40]  * tgpraveen tries with gedit
[11:40] <seb128> it should
[11:40] <seb128> if it doesn't there is a bug somewhere
[11:42] <tgpraveen> seb128: nope even after I add in gedit the word it aint recognized in ff 3, ff 3.5 , chrome
[11:42] <seb128> tgpraveen, I'm not asking about web browser but rather if it works between gnome applications
[11:42] <seb128> like gedit, pidgin
[11:43] <tgpraveen> um well it doesn't between gedit , pidgin
[11:43] <tgpraveen> but as you said pidgin not in gnome
[11:43] <seb128> right but they use gtkspell which uses enchant
[11:43] <tgpraveen> but still it should have worked if for nothing else then that intrepid targer
[11:43] <tgpraveen> yeah and it still doesn't work
[11:43] <seb128> well it works
[11:44] <tgpraveen> between which and which app?
[11:44] <seb128> but we never tried to do dictionnary changes I think
[11:44] <seb128> all apps use the same system dictionnaries
[11:44] <seb128> I'm not sure how user changes are supposed to work
[11:44] <seb128> try enchant -l example
[11:44] <seb128> see if it lists the same issues?
[11:46] <tgpraveen> enchant -l praveen ---> Error: Could not open the file "praveen" for reading.
[11:46] <Laney> looks to me like http://bugs.freedesktop.org/show_bug.cgi?id=21299 is not fixed
[11:46] <seb128> tgpraveen, do you have a praveen file?
[11:46] <tgpraveen> it gives the same thing for any word
[11:46] <Laney> anyone care to repro? Connect to the same MSN account in two instances, they fight for control
[11:47] <asac> ArneGoetje: was today the day where i should review the langpacks?
[11:47] <asac> what are the best instructions to install current karmic with a usb key?
[11:48] <tgpraveen> seb128: what exactly does a file mean here?
[11:48] <tgpraveen> -l     List only the misspellings.
[11:48] <seb128> tgpraveen, I don't know how to explain, it's a computer thing
[11:48] <tgpraveen> so oh ok
[11:48] <tgpraveen> I gottit
[11:48] <seb128> whatever logic info is written on a disk
[11:48] <tgpraveen> hehe yeah I got it
[11:49] <seb128> let me take a dictionary for you
[11:49] <seb128> ok, good
[11:49] <seb128> echo "some words" > example
[11:49] <seb128> enchant -d en -l example
[11:49] <seb128> if you want to use the en dictionary
[11:51] <tgpraveen> seb128: ok so I tried it
[11:51] <tgpraveen> and enchant still lists as a misspelling a word which I added
[11:51] <tgpraveen> to dictionary in gedit
[11:51] <tgpraveen> so something's broken
[11:52] <seb128> tgpraveen, grep word .config/enchant/*?
[11:54] <tgpraveen> um you want me to run the command in terminal "grep word .config/enchant/* "   ?
[11:54] <seb128> yes
[11:54] <seb128> where word is the word you added
[11:55] <chrisccoulson> pitti - screen locking doesnt work when suspending from System -> Shutdown -> Suspend because gnome-session uses "/apps/gnome-screensaver/lock_enabled" to determine whether to lock the screen or not, and that is false by default
[11:55] <tgpraveen> grep: .config/enchant/*: No such file or directory
[11:55] <chrisccoulson> this is all very depressing - every method of suspending does it's own thing ;)
[11:55] <seb128> tgpraveen, ~/.config...?
[11:55] <chrisccoulson> rather than being proxied through g-p-m like before
[11:55] <seb128> not sure where you are
[11:55] <tgpraveen> am on desktop
[11:56] <tgpraveen> praveen@praveen-desktop:~/Desktop$
[11:56] <seb128> well .config is in the user dir
[11:56] <seb128> so ~/.config
[11:57] <ArneGoetje> asac: yes, we are currently building new langpacks
[11:57] <tgpraveen> .config/enchant/en.dic:praveen
[11:57] <tgpraveen> .config/enchant/en.dic: praveen
[11:58] <tgpraveen> so its there in enchant? right
[11:58] <seb128> tgpraveen, ok, they are added to the dict correctly but enchant doesn't find those
[11:58] <seb128> seems a bug in enchant
[11:58] <seb128> since enchant -l has the same issue
[11:58] <seb128> I've to go for lunch but will look to that after lunch
[11:58] <tgpraveen> yeah
[11:58] <tgpraveen> k. gr8
[11:58] <tgpraveen> you want me to file a bug or anything
[11:58] <tgpraveen> though I don't knw how well I would file it
[11:59] <tgpraveen> and I would also add am on jaunty not on karmic.
[11:59] <seb128> just file a bug on enchant saying custom words are not being used
[11:59] <seb128> thanks
[12:02] <asac> ArneGoetje: ok eta? ~5h?
[12:11] <tgpraveen> seb128: done.  https://bugs.launchpad.net/ubuntu/+bug/446230
[12:31] <chrisccoulson> seb128 - bug 411083 is my fault ;)
[12:31] <chrisccoulson> there is an extra semicolon in the patch which probably triggers that
[12:31] <seb128> chrisccoulson, oh ok
[12:31] <chrisccoulson> which causes the migration to not be one-time-only
[12:31] <chrisccoulson> i'll fix that later
[12:31] <seb128> right, makes sense
[12:32] <seb128> it that only this one char change? do you want to do testing or should I just upload for you?
[12:32] <chrisccoulson> it's just a one character change (i've added a comment in the bug report)
[12:33] <chrisccoulson> feel free to upload it if you have some spare minutes:)
[12:34] <chrisccoulson> bbl, lunch time!
[13:01] <vuntz> chrisccoulson: heya. Didn't understand your comment about gnome-session using /apps/gnome-screensaver/lock_enabled
[13:02] <vuntz> chrisccoulson: is it supposed to use something else?
[13:02] <chrisccoulson> vuntz - that's different to what gnome-power-manager does by default, for deciding whether to lock the screen
[13:03] <chrisccoulson> it seems to be confusing for users, because they disable "Lock the screen when the screensaver is active" option in the screensaver preferences, but still expect the screen to lock when the suspend manually
[13:06] <chrisccoulson> bbl (again!), i have a meeting to go to now
[13:06]  * pitti sings "another one bites the dust" and uploads polkit-1 ported screen-resolution-extra
[13:06] <pitti> lunchtime now..
[13:06] <james_w> yay pitti
[13:07] <james_w> though there are a surprising number of complaints about polkit-gnome-authorizations going away
[13:10] <seb128> is dbus_g_proxy_new_for_name_owner() the right way to query if there is a provider for something on the bus?
[13:13] <james_w> I'm not sure
[13:14] <james_w> is it possible that the thing you are querying for is an activated service?
[13:14] <seb128> http://www.abisource.com/viewvc/enchant/trunk/src/zemberek/zemberek.cpp?revision=25545&view=markup&sortby=date&sortdir=down
[13:14] <seb128> see line 46
[13:15] <seb128> using that to see if zemberek-server is installed is not working
[13:15] <james_w> ah
[13:15] <seb128> using _owner seems to work
[13:15] <james_w> that just creates a proxy, doesn't do any DBus calls
[13:15] <seb128> but I want to check if that's right with somebody knowing dbus before uploading
[13:15] <seb128> well dbus_g_proxy_new_for_name_owner() seems to work
[13:15] <james_w> if you use _owner then (I think) it does a DBus call to find out the owner and creates a proxy for its unique name
[13:16] <seb128> "dbus_g_proxy_new_for_name_owner(): if you provide the well-known name "org.freedesktop.Database" dbus_g_proxy_new_for_name() remains bound to that name as it changes owner. dbus_g_proxy_new_for_name_owner() will fail if the name has no owner."
[13:16] <seb128> from the api
[13:16] <james_w> however, that means that if the owner changes you won't follow
[13:16] <james_w> yeah
[13:16] <james_w> what's the bug?
[13:16] <seb128> see the function, they don't use it, they free directly proxy
[13:16] <james_w> ah, I see it
[13:16] <seb128> it's just meant as a way to check if the provider is installed
[13:16] <james_w> zemberek_service_is_running
[13:16] <james_w> _owner is probably ok then
[13:17] <james_w> it's not going to be robust though
[13:17] <seb128> james_w, the bug is that it always returns true
[13:17] <seb128> which means everybody has a turkish dictionary listed everywhere in GNOME
[13:17] <james_w> because that could then return True, but when it goes to access it there is nothing there any more
[13:17] <seb128> james_w, robust against what?
[13:17] <seb128> oh
[13:17] <seb128> that will be good enough for karmic
[13:17] <james_w> but that's probably fairly unlikely
[13:18] <seb128> I've other bugs to track, I don't really care about that corner case
[13:18] <seb128> I just want turkish to not be wrongly listed everywhere
[13:18] <seb128> james_w, thanks
[13:18] <james_w> np
[13:20] <vuntz> seb128: can you make sure chrisccoulson files the bug upstream when he comes back? :-)
[13:20] <seb128> vuntz, yes, did you understand the issue?
[13:21] <vuntz> seb128: yep. The summary is "see gpm_control_get_lock_policy() in gpm-control.c"
[13:21] <vuntz> and I blame hughsie ;-)
[13:21] <seb128> ok
[13:21] <vuntz> seb128: thanks!
[13:21] <seb128> vuntz, you're welcome
[13:35] <aquarius> Is there a website planned to go along with the Software Center to help people thinking of moving to Ubuntu to see which apps are available, what they'd use as alternatives to their current apps, and generally to help answer the question "can I switch?"
[13:35] <Gewitterstern> good idea, aquarius
[13:37]  * aquarius takes that as a "no", then ;-)
[13:37] <Gewitterstern> aquarius, I am not from the staff
[13:38] <andreasn> aquarius: I've seen a couple of those, surely it would be possible to do copy+paste and modify the content of any of those and to link to apt:packagename or something like that
[13:38] <aquarius> It would be interesting to have; I'm just not sure how it would fit in with the current plans for Software Center
[13:38] <aquarius> this is an mpt question, I suppose. :)
[13:39] <aquarius> andreasn, yeah, there are a few around, but they're not Ubuntu-specific, and they're more about trying to mention every app than about guiding people to help them decide whether they can switch.
[13:40] <mvo> aquarius: we have support for keyword, so if you search for photoshop, it could show gimp, the whole keyword adding is a bit cumbersome at this point though (basicly I need to add them :)
[13:40] <Gewitterstern> thats true. also they are usually not uptodate
[13:40] <aquarius> mvo, that's just in the desktop app, though, yes?
[13:40] <mvo> but once that scales better with LP support etc then part of the problem will be gone
[13:40] <mvo> yeah, a website would be a seperate project
[13:41] <aquarius> mvo, this is to help people currently running Windows or OS X, so when I say to them "you seem to be annoyed with Apple, how about trying Ubuntu?" they can go to the website, type in the names of apps they use, and get a list of alternatives with decent review comments on whether they fulfil the role of replacing the app they're asking about
[13:42] <aquarius> mvo, where's the keyword database ("photoshop" -> "gimp") going to come from? :)
[13:49] <mvo> aquarius: currently its part of app-install-data-ubuntu
[13:50] <mvo> aquarius: it should be trivial to extend the xapian stuff to be a web app
[13:50] <mvo> it would need to be hostest somewhere etc
[13:51] <mvo> but the data to build this is all there (minus that its currently not that well fostered, but that will hopefully change once we get support from LP and the amazing contriburors)
[13:53] <aquarius> I have pinged a couple of people who are currently not Ubuntu users and have enquired whether such a website exists to see what exactly they would want such a website to do
[14:00]  * mac_v wonders when Amaranth will fix jumping windows ;p
[14:03] <joaopinto> aquarius, it is already possible to have software installation integrated with a web presentation using apturl
[14:04] <joaopinto> aquarius, check www.playdeb.net
[14:05] <aquarius> joaopinto, agreed. That's not quite what I'm looking for, though. Someone who already runs Ubuntu can use the Software Center. This is for people who are not yet running Ubuntu, but are thinking about doing so, and so they want to know "will I be able to switch to ubuntu? If I do, how will I carry on working like I do at the moment? which apps should I use?"
[14:06] <joaopinto> aquarius, you are talking about two different subjects, one thing is to present more information about apps, another thing is about migrating for another OS
[14:07] <aquarius> joaopinto, Sort of. There's already quite a lot of information about Ubuntu generally, why it's good, and so on. This is specifically to help people on another OS understand which software is available on Ubuntu, and which software they'd need to install to continue doing the work that they currently do, if they switched from Windows/OSX to Ubuntu.
[14:09] <joaopinto> aquarius, IMHO does is not related to software distribution in general, there is much more changing than the applications, so you are talking about a site devoted to  OS -> Ubuntu migration
[14:11] <aquarius> joaopinto, I'm specifically responding here to a few people I know who have said: "Ubuntu needs a good marketplace attached to it so people like me can see what software we can get for it." and "you need a 'can I switch?', 'will my apps work' type affair"
[14:15] <seb128> vuntz, is gnome-panel recent document not opening remote files correctly a known issue?
[14:15] <seb128> vuntz, ie will try to open a pdf over ssh with firefox where the .recently-used.xbel has the right mimetype and software
[14:15] <seb128> vuntz, on GNOME 2.28
[14:17] <vuntz> seb128: this looks suspiciously like an old bug you opened and which was a panel bug
[14:17] <seb128> mvo, bug #446268 seems for you, could you have a look?
[14:17] <vuntz> was not
[14:17] <seb128> vuntz, right, but this time the recently-used mimetype is correct
[14:17] <vuntz> seb128: but I can't test right now -- I'm leaving tomorrow and I have tons of things to finish before that :/
[14:18] <vuntz> seb128: so just open a bug if you're unsure
[14:18] <vuntz> seb128: and ping me in November :-)
[14:18] <seb128> mvo, some errors about communicating to aptdaemon
[14:18] <seb128> vuntz, coming to uds?
[14:18] <vuntz> seb128: don't know yet
[14:19] <seb128> ok
[14:21] <Amaranth> mac_v: they're fixed unless the app is doing weird things
[14:22] <mac_v> Amaranth: nope , not fixed here :( ... try with synaptic check the history child window , it jumps on close
[14:23] <Amaranth> Right, it moves up
[14:23] <mac_v> eog :about also does the same , several about windows too
[14:23] <Amaranth> But not every window does this
[14:23] <mac_v> yeah , not every window , but only some
[14:24] <Amaranth> The last time someone reporting a problem with windows jumping the bug was actually fixed in gdk, not compiz
[14:25] <mvo> seb128: that bug looks odd, I can see no failure inthe log
[14:26] <seb128> mvo, you see the debconf aptdaemon errors?
[14:29] <mvo> seb128: yes, now
[14:32] <fta> can't open attachments in evolution (karmic), it jumps to 100% cpu & 100% memory
[14:36] <fta> i can save it though
[14:39] <seb128> fta, wfm
[14:40] <seb128> I just tried on a pdf there
[14:42] <seb128> mvo, thanks for looking at the bug and reassigning ;-)
[14:44] <mpt> mac_v, hi, do you remember that the "You need to restart to finish installing updates" alert was reworded during Karmic? I'm trying to find the bug report
[14:45] <mvo> seb128: thanks for bringing it up - I'm doing triage all day, I so much don't like it :)
[14:49] <mac_v> mpt: that was fixed , the wording was changed... let me get the bug#
[14:50] <mpt> The closest I can find is bug 211616, but that was last year
[14:50] <mac_v> mpt: Bug #397324
[14:51] <mpt> thanks mac_v!
[14:51] <mac_v> np :)
[14:51] <mpt> Q: Why is Xorg consuming 99% of my CPU?
[14:52] <mpt> A: Because the other 1% is taken by tracker-indexer
[14:52] <mac_v> mpt: A: it hates you ;p
[15:02] <seb128> pitti, is there a way to get a bug out of the retracer signatures list?
[15:03] <pitti> seb128: you mean you have a signature and you want to map it to a bug?
[15:03] <seb128> pitti, see bug #429126, the retracing didn't work and the retracer is cleaning duplicates so we never get a stacktrace
[15:04] <pitti> oh, you want to delete the bug from the dup db
[15:04] <seb128> yes
[15:04] <pitti> seb128: dupdb-admin should be able to, hang on
[15:04] <seb128> or we will never get a stacktrace
[15:04] <pitti> %prog [options] dump
[15:04] <pitti> %prog [options] changeid <old ID> <new ID>
[15:04] <pitti> %prog [options] consolidate''')
[15:04] <pitti> oh, meh
[15:04] <pitti> seb128: so if it's utterly urgent, we need to use the sqlite command line tool or the firefox plugin
[15:05] <pitti> seb128: otherwise, let me just add a dupdb-admin delete command
[15:05] <seb128> would closing the bug work? or it will retrace new one and say it matches the close bug but still not retrace?
[15:05] <seb128> pitti, not urgent but that's not the first time we get a similar issue so I was rather wondering
[15:05] <pitti> seb128: if you close it, it will spit out warnings about "matches the stack trace in that other bug", but not mark as dup
[15:05] <seb128> we have enough bugs that I don't really care about this one
[15:06] <pitti> seb128: that is for "fix released" bugs anyway; let me check the code what it does for "invalid" bugs
[15:06] <seb128> pitti, right, but will it add the retracing in such cases?
[15:07] <mac_v> seb128: doubt regarding bzr  , when doing $bzr commit -m "message" --fixes lp :1234  , how do i use the "--fixes" option for 2 bugs?
[15:08] <seb128> mac_v, don't ask me I never used that option
[15:08] <mac_v> ;)
[15:08] <pitti> seb128: oh, even better
[15:08] <pitti> seb128: if you close the bug as invalid, the next time the retracer wants to touch it it will be removed from teh dup db
[15:09] <pitti> so seems we already had this case ages ago
[15:09] <pitti> and we don't really need dupdb-admin delete
[15:09] <seb128> pitti, ok thanks
[15:09] <pitti> seb128: (had to triple-check the code first)
[15:09] <seb128> pitti, thanks for checking
[15:09] <mac_v> pitti: any ideas for the --fixes option question ^?
[15:10] <pitti> mac_v: I read recent scrollback, I don't see the q?
[15:10] <mac_v>  doubt regarding bzr  , when doing $bzr commit -m "message" --fixes lp :1234  , how do i use the "--fixes" option for 2 bugs?
[15:13] <mac_v> pitti: ^
[15:14] <pitti> mac_v: should be --fixes lp:1 --fixes lp:2 ...
[15:14] <mac_v> pitti: hmm.. ok thanks :)
[15:20] <tedg> seb128: Were you able to figure anything out on the translations issue for indicator-messages?
[15:21] <pitti> btw, is it just me, or is the menu ordering in the indicator applet differently scrambled each time I start it?
[15:21] <seb128> tedg, seems to be what you said the other day, textdomain screwing translations when there is no translation for the domain
[15:21] <seb128> tedg, I was waiting today's language pack update to confirm it's fixed in french
[15:21] <seb128> tedg, if that's the case I would advice to do a call for translations to make sure all languages have a translation
[15:21] <tedg> seb128: Okay, I'll release that one last for kenvandine then, just to see if we need any changes.
[15:22] <tedg> seb128: Makes sense.
[15:22] <seb128> tedg, you don't, no need to wait
[15:22] <tedg> Okay
[15:22] <seb128> tedg, I did cp gedit.mo indicator-messages.mo and that fixed the bug
[15:22] <seb128> so I expect the next langpack round with the indicator translations will work too
[15:24] <Amaranth> grr
[15:24] <Amaranth> I'm not sure what part to even be mad at but something intel-related is making me angry :P
[15:25] <Amaranth> If I boot with nomodeset I get a fun strobe effect as it rapidly tries and fails to start X
[15:25] <Amaranth> So rapidly you can't login to a tty and see why
[15:26] <chrisccoulson> pitti - bug 440484?
[15:26] <pitti> chrisccoulson: that, too
[15:26] <pitti> might be just that indeed
[15:27] <chrisccoulson> there was also bug 430904, but that's fixed
[15:27] <seb128> chrisccoulson, wb, vuntz asked if you could open a GNOME bug about the locking issue you described before
[15:27] <chrisccoulson> seb128 - yeah, will do
[15:27] <seb128> chrisccoulson, thanks
[15:28] <chrisccoulson> the current behaviour seems to be confusing, and it is different to how it worked before when all the suspend/hibernate requests were proxied through g-p-m
[15:29] <vuntz> seb128: you didn't quote everything I said ;-) (but thanks for remembering, I already had forgotten)
[15:30] <vuntz> chrisccoulson: 14:21 < vuntz> seb128: yep. The summary is "see gpm_control_get_lock_policy() in gpm-control.c"
[15:30] <chrisccoulson> vuntz - in g-p-m?
[15:30] <vuntz> yep
[15:30] <chrisccoulson> yeah, i've seen that already
[15:31] <chrisccoulson> it has a key to choose whether to use it's own config, or gnome-screensaver config instead
[15:31] <chrisccoulson> and it defaults to it's own policy
[15:31] <chrisccoulson> vuntz - do you think gnome-session should do something similar?
[15:34] <vuntz> chrisccoulson: probably. And this is exactly why I'm unhappy about having to lock the screensaver in gnome-session, fwiw
[15:34] <vuntz> chrisccoulson: please cc hughsie in the bug you open, btw. Just so that he double-checks
[15:34] <chrisccoulson> vuntz - yeah, i agree with you as well, but it's the only solution so far:(
[15:35] <vuntz> chrisccoulson: should be in dk-power
[15:35] <chrisccoulson> yeah, i think so too. that way, other things can hook on to the suspend event too
[15:41] <rickspencer3> seb128, pedro_ could you take a look at:
[15:41] <rickspencer3> https://bugs.edge.launchpad.net/ubuntu/+source/xsane/+bug/446373
[15:41] <rickspencer3> this is a crasher in xsane that pgraner is hitting
[15:43] <seb128> pgraner, do you know how to get a stacktrace using gdb?
[15:44] <seb128> pgraner, can you install libgtk2.0-0-dbg, break on gdk_x_error and run with --sync and get one?
[15:45] <pgraner> seb128: sure will be a bit as I'm on calls for the next few hours
[15:46] <seb128> pgraner, no hurry just add it to the bug when you get it if I'm not around
[15:46] <pgraner> seb128: will do
[15:46] <seb128> thanks
[15:47] <mpt> mvo, robbiew: What do you think of pulling gnome-app-install off the CD now?
[15:48] <chrisccoulson> another xrandr crash!
[15:48] <robbiew> mpt: I'm fine with it
[15:50] <mvo> mpt: its already removed from the seeds
[15:50] <mpt> oh, cool
[15:50] <mvo> mpt: have you checked that its still on the cd?
[15:50] <seb128> chrisccoulson, ;-)
[15:50] <chrisccoulson> X_RRGetScreenSizeRange
[15:51] <seb128> chrisccoulson, you start seing xrandr crashes everywhere?
[15:51] <chrisccoulson> if i've decoded the error message correctly ;)
[15:51] <chrisccoulson> seb128 - only the ones in g-s-d
[15:52] <chrisccoulson> actually, minor_code=5 is X_RRGetScreenInfo
[15:52] <superm1> mvo, mpt robbiew it's a nautilus recommends so i'd bet it's still on the disk: http://people.canonical.com/~ubuntu-archive/germinate-output/ubuntu.karmic/desktop
[15:52] <mpt> mvo, one of my teammates showed me that "Add/Remove Applications" showed up as the first item in the "Installed Software" section, but I've just checked and that was a 9.04 image upgraded to Karmic last Friday
[15:53] <mpt> superm1, recommended by *nautilus*? why on earth? :-)
[15:55] <Amaranth> mpt: for the mime stuff
[15:56] <Amaranth> if you try to open a file you have no program for it open gnome-app-install to find one
[15:58] <mvo> mpt: so you want to remove it from the app-install data set too?
[16:00] <mpt> mvo, I don't think that's necessary (we show Synaptic in the Center, so why not Add/Remove?)
[16:00] <mvo> seb128: could you please remove g-a-i from the nautilus recommends (I can do it too, just let me know)
[16:00]  * mpt wonders what kind of weird file to download to see this "mime stuff" in action
[16:00] <mvo> mpt: ok
[16:00] <mvo> mpt: try a dia file
[16:00] <seb128> mvo, is software-store a drop in place for that
[16:00] <seb128> +	command = g_strconcat("gnome-app-install --mime-type=",
[16:00] <seb128> +			     qmimetype, " ", quri,
[16:00] <seb128> +			     (char*)0 /* NB NULL is wrong */);
[16:01] <mvo> seb128: no, we don't have that feature anymore
[16:01] <seb128> mvo, ie should we call software-store --mime-type
[16:01] <seb128> mvo, you ask me to drop a feature now?
[16:01] <mpt> "There is no application installed for AutoCAD image files"
[16:01] <mvo> seb128: that is how I understand mpt
[16:02] <mpt> I haven't seen this feature before, and it doesn't seem to be working in Karmic
[16:02]  * mpt tries with .dia
[16:03] <mvo> its not used a lot, it only offer suggestions for stuff in main
[16:03] <mvo> and we cover most of the relevant file types with the default install
[16:03] <seb128> mvo, I think debian has a patch variant calling synaptic
[16:03] <mvo> mpt: it was part of the list of regressions compared to g-a-i that I send out some weeks ago
[16:04] <mvo> seb128: oh, really? how do they do the mapping between mime->package with just synaptic?
[16:04] <rodrigo_> kenvandine, pitti: I need your nomination approval for https://bugs.edge.launchpad.net/ubuntu/+source/couchdb-glib/+bug/444576 , please :)
[16:04] <mpt> Double-clicking a .dia file does nothing at all
[16:04] <mvo> seb128: its not too hard to add for software-center, I just didn't had time to get it done
[16:05] <mvo> mpt: if you want to see what it looks like try "gnome-app-install --mime-type text/html"
[16:05] <mpt> mvo, I'm sorry, I didn't notice
[16:06] <seb128> mvo, they use the app-installs infos
[16:07] <mpt> mvo, not having used this feature, I'm not a good judge of how important it is. If it's important, keep Add/Remove on the CD. If it's not, kick it off, and I'll design it for the Center v2.
[16:08] <seb128> I think we can drop it
[16:08] <seb128> it was dropped in hardy and nobody ever noticed I think
[16:09] <mvo> seb128: ok with me, if its not trivial to use the debian approach then we can do that, as I said, most apps are covered
[16:09] <chrisccoulson> i didn't even know that feature existed
[16:10] <mvo> I guess it would be much more useful if there was a right click option in nautilus "find software that deals with that type" or something
[16:10] <mvo> because for most stuff there is a mime type handler already
[16:10] <mvo> a app installed I mean, so no reason for it to kick in
[16:10] <seb128> mvo, I would say it's too low detail to bother with it
[16:11] <seb128> mvo, you can probably fix higher profile issues in one hour of your time
[16:11] <seb128> so don't bother trying to look at the synaptic way
[16:11]  * mvo nods
[16:11] <seb128> mvo, I will do the nautilus change
[16:11] <mvo> liek bug triage
[16:11]  * mvo weeps
[16:11] <mvo> thanks seb128!
[16:11] <seb128> lol
[16:11] <seb128> mvo, I can assign you desktop bugs if you want... ;-)
[16:11] <mpt> thanks seb128, thanks mvo
[16:12] <mpt> and thanks robbiew
[16:12] <robbiew> ;)
[16:12] <mvo> the good thing is that we can re-introduce it in lucid and celebrate it as a new feature
[16:12] <mpt> And maybe do a better job of it, e.g. distinguishing between viewers and editors
[16:13] <mvo> more metadata!
[16:13] <chrisccoulson> mvo - do you not like bug triage? ;)
[16:13] <mvo> chrisccoulson: I hate it
[16:13] <mvo> with passion
[16:13] <chrisccoulson> heh:)
[16:13] <mvo> I wish there was someone/something doing it for me and just showing me what I consider relevant
[16:15]  * mvo stops ranting and goes back to work
[16:16]  * robbiew agrees with mvo's wish
[16:16] <mpt> Yeah, a dedicated QA person for package management would be nice :-)
[16:16] <chrisccoulson> perhaps there should be an update-manager bug day ;)
[16:16] <seb128> mvo, stop looking at pedro when you say that!
[16:16] <mpt> I added a note about searching (and launching) by filetype to the spec <https://wiki.ubuntu.com/SoftwareCenter?action=diff&rev2=220&rev1=219>
[16:17] <chrisccoulson> mvo - can i get you to look at some bugs?
[16:17]  * chrisccoulson runs
[16:19] <mpt> mvo, seb128: So is there anything we can do to remove g-a-i on Karmic upgrades, without making it actually Conflict with software-center?
[16:19] <mpt> (or vice versa)
[16:20] <seb128> mpt, I think update-manager will clean things which moved to universe but mvo should confirm that he knows better ;-)
[16:21] <mpt> e.g. how did nautilus-cd-burner get replaced?
[16:24] <Amaranth> mpt: brasero has Conflicts: bonfire (<= 0.4.4-1), nautilus-cd-burner
[16:24] <mpt> huh
[16:45] <pitti> seb128: is there a bug for teh schema locale default stripping? can you add a task to the existing one and assign it to me?
[16:46] <seb128> pitti, task on what component?
[16:46] <pitti> seb128: cdbs
[16:46] <pitti> rodrigo_: done, sorry for delay
[16:46] <seb128> pitti, done
[16:47] <pitti> thanks
[16:47] <mvo> seb128: we don't remove stuff automatically when it moved, that is risky, we don't know if the user actually likes the app in question :)
[16:48] <mvo> mpt: we can add a quriks handler, I don't think a conflict is right
[16:48] <mpt> mvo, me neither
[16:49] <seb128> mvo, ok ;-)
[16:50] <mvo> I add it to the forced obsoletes now
[16:51] <mvo> seb128: btw, after the triage of today I will switch u-m back to synaptic by default
[16:51] <seb128> mvo, oh?
[16:52] <mvo> I think its just too risky give the importance of u-m
[16:53] <seb128> mvo, right
[16:53] <mvo> especially the amount of policykit hangs worries me
[16:53] <mvo> the odd thing is that I got much less of those in software-center
[16:53] <seb128> mvo, you noticed quite some bugs or that's just being cautious?
[16:53] <mvo> either its much less used at this point or it excerises different code pathes
[16:53] <mvo> (or both ;)
[16:53] <seb128> mvo, much less used for sure
[16:53]  * mvo nods
[16:53] <seb128> everybody update daily using update-manager
[16:54] <seb128> I don't know who use software-store out of testing
[16:54] <seb128> I mean in the unstable distros users
[16:54] <seb128> those usually have want they want already installed
[16:54] <seb128> they don't install new things daily
[16:54] <mvo> yeah, its a great way to test aptdaemon and we got a lot of valuable feedback, but I'm just worried that there will be one app where it fails duing a post-release upgrade
[16:54]  * mvo nods
[16:54] <james_w> mvo: have a reference for the policykit hangs?
[16:55] <mvo> james_w: bug #445303
[16:55] <mvo> james_w: I'm not sure if its polkit or aptdaemon, could well be something aptdaemon too
[16:55] <james_w> that's not aptdaemon specific from what I have seen
[16:55] <mvo> oh?
[16:56] <chrisccoulson> i'm sure i've seen that somewhere other than aptdaemon too
[16:56] <mvo> hm, in this case I may recondier
[16:56] <mvo> its my biggest concern (also not the only one)
[16:56] <chrisccoulson> but aptdaemon does eventually crash after the dialog appears to have hung
[16:57] <mvo> well, it will timeout
[16:57] <mvo> not actually crash
[16:57] <chrisccoulson> yeah, that's what i thought
[16:57]  * mvo nods
[16:57] <james_w> bug 414877
[16:58] <mpt> ugh, I've seen that one
[16:58] <james_w> is that the same bug?
[16:58] <chrisccoulson> james_w - yes, i think so
[17:00] <james_w> bug 436413 as well?
[17:00] <mvo> can we milestone it please?
[17:00] <mvo> james_w: the frozen one might be, there was one issues where while the dialog was up the app would freeze, but that is fixed since
[17:00] <mvo> so the bugreport may conflate(?) two issues
[17:01] <james_w> ah, ok
[17:01] <james_w> shall I close the polkit task on the last one, dupe your bug to mine, and then milestone?
[17:01] <mvo> james_w: mine has some duplicates already, whatever is easier for you, I don't mind
[17:02] <mvo> I miss a "dup-this-bug-and-all-duplicates-to-new-bug" button in LP
[17:02] <james_w> there's lp-set-dup in ubuntu-qa-tools
[17:03] <mvo> oh sweet
[17:03] <james_w> can anyone reliably reproduce it?
[17:03] <mvo> I should talk to you more often :)
[17:03] <mvo> I don't know of anyone :/
[17:06] <chrisccoulson> james_w - i can't reliably reproduce it :(
[17:06] <chrisccoulson> sounds like we need some strategically placed printf's in polkit;)
[17:07] <james_w> g_warnings are good now :-)
[17:09]  * chrisccoulson wonders if https://wiki.ubuntu.com/DebuggingProgramCrash isn't clear enough when it says that "yelp" is just an example and you need to replace it with the appropriate package
[17:09] <chrisccoulson> someone just commented on a transmission bug, asking what to do now after installing yelp-dbgsym
[17:10] <kklimonda> heh, you've noticed that too? ;)
[17:10] <chrisccoulson> yeah ;)
[17:13] <james_w> hmmmmmm
[17:13] <james_w> I wonder if is the PAM_TEXT_INFO bug back again
[17:36] <james_w> not seeing anything obvious
[17:37] <james_w> there's probably a deadlock in communication with the helper or with PAM
[17:37] <james_w> I'm not seeing where though
[17:47] <seb128> mvo, so I changed the gnome-app-install recommends to a suggests and I let the patch, it deal fine with the case where gnome-app-install is not installed and some user might still want it
[17:53] <jono> kenvandine, so Christian seems to think Empathy isnt talking to pulse and i should
[17:53] <jono> is that the case?
[17:55] <kenvandine> jono, right
[17:55] <kenvandine> but for other people, it is
[17:55] <jono> kenvandine, can I test now
[17:55] <kenvandine> i don't know how to debug that
[17:56] <kenvandine> jono, not with me... give me a bit
[17:56] <jono> ahhh
[17:56] <jono> my gstreamer-properties were set to ALSA
[17:56] <jono> changed them to pulse, want to test
[17:56] <jono> it may fix it
[17:57] <kenvandine> ok
[17:57] <kenvandine> lets test real quick
[17:58] <jono> kenvandine, I can hear you
[17:58] <jono> one sec
[18:00] <chrisccoulson> seb128 - did you decide what you want to do with the GDM theming stuff in the end?
[18:00] <seb128> chrisccoulson, I want to raise the issue on the upstream list
[18:01] <chrisccoulson> before you make a decision?
[18:04] <seb128> chrisccoulson, well I'm still pondering it, what do you think?
[18:04] <seb128> I think for now I'm leaning toward the sudo dbus-launch || true
[18:04] <seb128> other changes seem to be complicated
[18:05] <chrisccoulson> yeah, if it works ok then that's probably the way to go
[18:05] <seb128> in the || true case if that breaks we will just have upstream theming
[18:05] <chrisccoulson> i can do the other changes, but they are quite complicated ;)
[18:05] <seb128> that + local gvfs should be quite stable
[18:05] <chrisccoulson> yeah, hopefully. i'll put the work i started to one side for now then
[18:05] <chrisccoulson> right, home time for me!
[18:06] <chrisccoulson> bbl
[18:06] <seb128> sorry about that
[18:06] <chrisccoulson> that's ok ;)
[18:06] <seb128> see you later!
[18:06] <seb128> time for sport here
[18:06] <chrisccoulson> have fun!
[18:06] <tgpraveen> kenvandine: jono did it work?
[18:06] <kenvandine> it did :)
[18:06] <kenvandine> sort of
[18:06] <kenvandine> he heard me
[18:06] <kenvandine> but connections are flaky now
[18:07] <kenvandine> he is getting UPnP timeouts
[18:07] <kenvandine> i think his router has some trouble with UPnP
[18:07] <kenvandine> not sure if we can do anything about that
[18:07] <tgpraveen> hmm something is better than nothing.
[18:07] <kenvandine> yeah
[18:07] <kenvandine> not crashing on audio anymore :)
[18:07] <tgpraveen> maybe empathy should be patched to tell user about UPnP.
[18:07] <kenvandine> yeah
[18:08] <kenvandine> we talked about that
[18:08] <kenvandine> not doable in time for karmic though :/
[18:08] <tgpraveen> also since he changed gstreamer settings, maybe that should be added in faq of empathy
[18:08] <kenvandine> but talked about adding a diagnostics button in the call dialog
[18:08] <kenvandine> so if you have problems, you can just click that
[18:08] <kenvandine> and it would run through some common problems
[18:08] <kenvandine> yeah
[18:09] <tgpraveen> also the gstreamer props are set to pulse by default in fresh karmic right?
[18:09] <tgpraveen> cool. the diagnostic making it in karmic?
[18:09] <kenvandine> yes
[18:09] <kenvandine> yes to pulse by default
[18:10] <kenvandine> no to diagnostics making it in time for karmic
[18:10] <kenvandine> it is way too late for that
[18:10] <tgpraveen> hmm ok. guess we should leave something for lucid too ;-)
[18:25] <jono> kenvandine, its working, on a video call now :)
[18:27] <kenvandine> woot
[18:27]  * Amaranth really needs his screen to not be so bright
[18:27] <kenvandine> jono, and if you got the update for empathy today... it shouldn't be crashing based on indicator events :)
[18:27] <Amaranth> writing code like this is painful :/
[18:39] <dobey> Amaranth: hey
[18:39] <Amaranth> dobey: uh oh :)
[18:39] <dobey> Amaranth: what was that command you suggested i try running for compiz?
[18:39] <Amaranth> CM_DRY=yes compiz
[18:39] <dobey> it seems to be beyond my scrollback
[18:39] <dobey> thanks
[18:40] <dobey> Amaranth: http://pastebin.ubuntu.com/288762/
[18:41] <Amaranth> ok, seems our blacklist isn't working since it can't find the VGA line in lspci
[18:41] <Amaranth> otherwise that looks fine
[18:42] <Amaranth> dobey: what happens if you just run `compiz`?
[18:42] <rodrigo_> pitti: thanks!
[18:43] <dobey> Amaranth: blinky blinky, compiz.real crashed, and back to metacity
[18:43] <dobey> *** glibc detected *** /usr/bin/compiz.real: double free or corruption (!prev): 0x08805c48 ***
[18:43] <Amaranth> fucking nvidia
[18:43]  * Amaranth throws things
[18:43] <dobey> i don't think nvidia puts double frees in compiz
[18:44] <Amaranth> dobey: I've got about 15 dupes of a crash 2-4 levels inside libGL.so that disagrees
[18:45] <dobey> well it worked fine before like alpha6, and i don't think nvidia drivers changed then?
[18:46] <Amaranth> dobey: The only thing I can think of is changing from one GLX call to a different one to prevent a memory leak in open source drivers
[18:46] <Amaranth> Perhaps no one ever used that call with nvidia before :P
[18:46] <Amaranth> dobey: But if you can bisect it to a change in compiz we can know for sure
[18:47] <dobey> well the thing you just mentioned sounds like it would be it, to me anyway :)
[18:47] <Amaranth> sure but we were doing it wrong before and are doing it right now
[18:48] <Amaranth> the way we were doing it before was not following the spec and causing memory leaks with DRI2 drivers
[18:48] <dobey> i really wish apport would split backtraces into individual frames and make a row per frame, rather than sticking them all in one row, that's 1243243 pixels tall
[18:48] <dobey> not following which spec?
[18:48] <Amaranth> bug 392850
[18:48] <Amaranth> dobey: glx
[18:49] <Amaranth> http://git.compiz.org/compiz/core/commit/?h=compiz-0.8&id=208b459b141b6d49f48796231146bb33d5dc1d26
[18:51] <dobey> the worst part is that it resizes my windows
[18:51] <dobey> and i have to use metacity
[18:53] <dobey> hrmm
[18:53] <dobey> i'd test, but not sure how
[18:53] <dobey> (i really don't want to build compiz from source)
[18:53] <dobey> guess i could shove it my ppa and see
[18:55] <Amaranth> hehe, you don't want to install gnome, kde, and all -dev packages for gnome, kde, and X? :)
[18:56] <Amaranth> If reverting that change fixes the crash I might be able to hack something together to do it only for nvidia
[19:10] <dobey> cgit really should have a "download reverse patch" button
[19:10] <Amaranth> heh
[19:10] <Amaranth> patch -R?
[19:10] <Amaranth> Although you don't want to do that either
[19:11] <Amaranth> If you revert the whole patch you'll change the API/ABI and screw everything up
[19:11] <Amaranth> dobey: you just want to put the calls to glXDestroyGLXPixmap back
[19:12] <Amaranth> although without a proper stack trace showing it crashes on that call...
[19:12] <Amaranth> but that's still the only thing I can think of that changed and would trigger a crash inside the nvidia driver
[19:18] <dobey> well i'll try and see
[19:53] <Amaranth> dobey: still waiting for the PPA to build?
[19:53] <dobey> Amaranth: well, busy doing actual work in the process :)
[19:54] <dobey> Amaranth: lp says "start in 28 minutes"
[19:54] <dobey> Amaranth: so i guess i'm still waiting, yeah :)
[19:54] <Amaranth> dobey: alright so we'll call it an hour :)
[20:33] <al-maisan> hello just did a "apt-get update", now I see the mouse pointer while at the gdm login mask but it becomes invisible as soon as I log on
[20:34] <al-maisan> playing with the touch pad does change the windows i.e the mouse pointer seems to be there but just invisible
[20:35] <al-maisan> an update for the following packages was installed: xserver-xorg-input-synaptics, xserver-xorg-video-intel
[20:35] <al-maisan> could I either of them case this behaviour?
[20:35] <al-maisan> s/I//
[20:35] <al-maisan> *cause
[20:39] <awe> TheMuso: I talked to rtg about the rtkit.  At this point, he's most comfortable with removing the PA runtime dependency on rtkit.  If you'd like, I can make the change, although I'll need someone to do the upload for me?
[20:40] <al-maisan> this is *very* funny after a few minutes into the session the pointer does become visible again
[20:56] <chrisccoulson> james_w - you there?
[22:19] <chrisccoulson> lol @ launchpad.net - "Featured projects - Drizzle" - i thought that the logo there was a local weather forecast!
[22:23] <chrisccoulson> seb128 - did you see we got a way to reproduce the issue with the session-properties capplet, where users were claiming it doesn't save changes?
[22:23] <seb128> chrisccoulson, no
[22:23] <seb128> one claims having the issue when not using a full path to the command
[22:23] <seb128> but I did try that and that worked for me
[22:23] <chrisccoulson> just by chance, another user noted that if you close the dialog too quickly, it doesn't save the changes
[22:23] <chrisccoulson> and sure enough, if you do that, it doesn't save any changes
[22:24] <chrisccoulson> it's wierd
[22:24] <seb128> oh
[22:24] <chrisccoulson> bug 444993
[22:24] <chrisccoulson> it seems to be the same issue that the other users were trying to describe, so i just dupe'd them all
[22:27] <seb128> thanks
[22:27] <seb128> you can still ask on the other bugs if they close it quickly
[22:28] <chrisccoulson> yeah, i left a comment on them
[22:33] <TheMuso`> awe: Daniel and I have a pulse upload pending, so I can take care of it.
[22:34] <awe> TheMuso`, OK
[22:37] <chrisccoulson> has anyone here actually managed to use gwibber successfully?
[22:37] <Laney> held it at an older version
[22:37] <chrisccoulson> it doesn't work here at all
[22:37] <chrisccoulson> it hasn't ever done since i upgraded to karmic
[22:38] <Laney> Version: 1.2.0~bzr346-0ubuntu3
[22:38] <chrisccoulson> all it does is crash repeatedly and chew CPU
[22:38] <Laney> wfm
[22:38] <chrisccoulson> Laney - thanks. i might downgrade too then ;)
[22:43] <seb128> kenvandine, tedg1: do you have other updates coming today?
[22:45] <chrisccoulson> seb128 - i've just pushed a g-s-d change to fix the touchpad config migration
[22:45] <seb128> chrisccoulson, cool, I'm sponsoring now
[22:45] <seb128> I'm not working tomorrow so I will probably not be around much
[22:45] <chrisccoulson> thanks:)
[22:46] <seb128> doing some sponsoring now before going to bed ;-)
[22:46] <chrisccoulson> would you mind doing gnome-python-extras? (so i can get glom uploaded tomorrow):)
[22:46] <seb128> was the mir issue for libgda sorted?
[22:46] <chrisccoulson> seb128 - yeah, that's sorted
[22:47] <chrisccoulson> bug 432715
[22:47] <seb128> I didn't touch it since the most recent comment was from dholbach
[22:47] <seb128> you didn't update the bug after that
[22:47] <chrisccoulson> did i not?
[22:47] <chrisccoulson> ah
[22:47] <seb128> no
[22:47] <chrisccoulson> yes, i probably should have left a comment on there ;)
[22:47] <seb128> no worry doing that now too
[22:48] <chrisccoulson> thank you:)
[22:48]  * chrisccoulson looks at the remainder of milestoned bugs
[22:49] <seb128> we are running out of bug on that list I should add some new ones ;-)
[22:50] <chrisccoulson> yeah, we need a few more ;)
[22:51] <seb128> that's not the number of components missing, just pick one you know or want to work on and look at the bug list
[22:51] <seb128> ie crashers on gnome-session or gnome-settings-daemon are worth fixing
[22:51] <seb128> we are in good shape for karmic ;-)
[22:52] <chrisccoulson> yeah, definately. there's a crasher in gnome-settings-daemon i could probably fix, although it's not particularly urgent and users are unlikely to experience it
[22:52] <chrisccoulson> i could probably find more urgent crashers to work on
[22:53] <chrisccoulson> in fact, i could probably do another libxklavier update to fix the issues with freenx servers
[22:53] <seb128> good example of bug worth fixing for karmic
[22:53] <chrisccoulson> that will be the most commonly reported g-s-d crashes fixed then
[22:55] <seb128> chrisccoulson, gpe uploaded, it will need binary newing, you can ping pitti on irc about that tomorrow if needed
[22:55] <chrisccoulson> seb128 - thanks
[22:56] <seb128> I will probably be to bed before it's built and published
[22:56] <chrisccoulson> yeah, i think i'm going to have an early night as well
[22:56] <seb128> hey robert_ancell
[22:56] <robert_ancell> seb128, hey
[22:57] <seb128> robert_ancell, we got the gconf issue fixed
[22:57] <robert_ancell> seb128, what was it??? It was driving me insane :)
[22:57] <seb128> it was combined gconf and cdbs bugs, took a while to figure those
[22:57] <robert_ancell> cdbs?
[22:57] <seb128> robert_ancell, for one thing registering a schemas was dropping gettext_domain from all the keys of other schemas
[22:57] <seb128> that was a gconf bug
[22:58] <seb128> no gettext_domain -> not working gettext
[22:58] <seb128> once this one fixed and gconf-schemas --register-all run you get descriptions translated
[22:58] <seb128> gettext is not used for default value
[22:59] <seb128> rational is that you don't want to load extra mo files to get only a value
[22:59] <seb128> so those should be in the .schemas
[22:59] <robert_ancell> ok
[22:59] <seb128> but the magic pitti added to the schemas to clean those dropped those informations
[22:59] <seb128> to *cdbs* rather...
[23:00] <seb128> ie we clean translations from schemas at build to win space
[23:00] <seb128> but default value translations should stay there
[23:00] <seb128> that got fixed too
[23:00] <seb128> so from now rebuilds should get translations working again
[23:00] <robert_ancell> ah
[23:00] <robert_ancell> good effort guys!
[23:00] <seb128> thanks ;-)
[23:01] <seb128> robert_ancell, do you still have things to keep busy today?
[23:01] <robert_ancell> gdm and rhythmbox today but if there's anything important let me know and I'll look if I have time
[23:02] <seb128> robert_ancell, no, we were discussing it with chrisccoulson just before you joined
[23:02] <seb128> I'm running out of milestoned bugs to dispatch
[23:02] <seb128> I need to do some bug triage again to get some new ones on the list ;-)
[23:02] <robert_ancell> dbus is driving me crazy - I need gdm to make a few calls to ConsoleKit but the C API is so frustrating with any complex data types
[23:02] <seb128> what do you try to change on gdm?
[23:03] <robert_ancell> guest session - when you open one, switch back to your session, then try and open another it just locks your screen rather than changing to the open session
[23:03] <robert_ancell> bug 425952
[23:03] <chrisccoulson> robert_ancell - that's a pain, i'm surprised that consolekit doesn't have a proper C library for accessing the DBus interface
[23:04] <robert_ancell> chrisccoulson, maybe it does... I will look further
[23:04] <chrisccoulson> like devicekit-{disks,power} do
[23:04] <seb128> oh ok
[23:04] <chrisccoulson> robert_ancell - i don't think it does
[23:04] <robert_ancell> chrisccoulson, do you know how to get a dbus result of type 'ao'?
[23:04] <chrisccoulson> everything else that talks to consolekit (ie, gnome-session), has to just use dbus-glib
[23:04] <robert_ancell> (array of bus names)
[23:05] <chrisccoulson> robert_ancell - i'm not too sure, but i'll try and see if we already have an example somewhere
[23:05] <robert_ancell> I think after Karmic I'm going to propose some updates to the dbus tutorial...
[23:05] <seb128> robert_ancell, I would argue that this bug is probably a corner case and low priority but I've not lot of real karmic must be fixed bugs on my list so...
[23:05] <robert_ancell> and PK tutorial...
[23:05] <seb128> if one of you could look to the "gnome-panel freezes when opening > 8 tasks on a vertical task list"
[23:05] <robert_ancell> seb128, well, it means you can lock yourself out of guest and never return
[23:05] <seb128> it drives quite some user crazy for some cycles
[23:06] <robert_ancell> sure, #?
[23:06] <seb128> robert_ancell, if you don't know ctrl-atl-f<n>?
[23:06] <robert_ancell> seb128, yeah you can do that (once we fixed the screen locking bug) but users wont know
[23:06] <robert_ancell> that
[23:06] <seb128> one contributor sent a one char patch to fix spell checking broken on non ascii char since warty today
[23:06] <seb128> in gtkhtml
[23:07] <seb128> gchar -> gunichar
[23:07] <chrisccoulson> i just saw that - nice!
[23:07] <seb128> robert_ancell, bug #187540
[23:08] <seb128> robert_ancell, bug #187540
[23:08] <seb128> there is some work and patches on the bug too
[23:08] <robert_ancell> seb128, thanks for fixing bug 441111 - I know it's not high priority but I was making no progress on other bugs yesterday and it has been bugging me for months :)
[23:08] <seb128> so maybe it's just a matter of reviewing those
[23:08] <seb128> the user who sent that tried to fix some other things too
[23:08] <seb128> but maybe we just want to fix that freeze issue for karmic
[23:09] <seb128> robert_ancell, thank you for fixing it, I've a keyboard switcher and I noticed it was looking weird and that was annoying me
[23:09] <seb128> and after reading your comment I noticed the screensaver dialog has the same bug
[23:09] <seb128> so good to see that visual glitch fixed before karmic ;-)
[23:10] <robert_ancell> cool, yeah it did always look a little bright...
[23:10] <seb128> it did work on transparent panels there though
[23:10] <robert_ancell> Hey can you guys look at bug 435227 - I fixed the patch for me but pitti still gets it to crash for him.  I just can't reproduce here, very odd
[23:11] <chrisccoulson> robert_ancell - for getting an array of dbus objects with dbus_g_proxy_call, you can pass something like "dbus_g_type_get_collection("GPtrArray", DBUS_TYPE_G_OBJECT_PATH)" as the argument type for the return value
[23:11] <chrisccoulson> and you then just access the return array as a GPtrArray
[23:13] <robert_ancell> chrisccoulson, thanks! Worked in my test prog, I had got up to dbus_g_type_get_collection("GPtrArray", DBUS_TYPE_STRING)
[23:13] <Amaranth> robert_ancell: Can you do a new compiz-fusion-plugins-main snapshot? I guess mvo never got a chance to do so
[23:13] <robert_ancell> Amaranth, ok, I'll do that today
[23:13] <Amaranth> We're getting a lot of crash reports from staticswitcher that are already fixed
[23:13] <seb128> seems robert_ancell has a busy day already now
[23:13] <chrisccoulson> robert_ancell - i just copied that from somewhere else (the old FUSA applet)
[23:13] <chrisccoulson> ;)
[23:14] <chrisccoulson> generally for things like that, you can nearly always find good examples in the C libraries that are used to access certain daemons
[23:14] <seb128> robert_ancell, you are lucky the work hour overlap is only one hour ;-)
[23:14] <robert_ancell> seb128, lol
[23:15] <robert_ancell> MOTU application today...
[23:15] <Amaranth> But I'm here all day ;)
[23:15] <seb128> oh
[23:15] <robert_ancell> Amaranth, you too right?
[23:15] <Amaranth> robert_ancell: Hopefully, they still haven't updated the meeting time
[23:15] <Amaranth> Should probably poke someone about that...
[23:15] <seb128> robert_ancell, when is it? did I have to write something on a wikipage for you?
[23:16] <robert_ancell> seb128, my page is https://wiki.ubuntu.com/RobertAncell/MOTUApplication
[23:16] <seb128> thanks
[23:16] <Amaranth> Wow drizzle is hosted on launchpad?
[23:16] <chrisccoulson> Amaranth - yeah. i was confused by the logo
[23:16] <seb128> what is drizzle?
[23:17] <robert_ancell> 7 UTC I'm hoping
[23:17] <seb128> chrisccoulson, g-s-d uploaded
[23:17] <chrisccoulson> seb128 - i've no idea - i just saw the logo on launchpad though, and thought it was just a weather report
[23:17] <chrisccoulson> thanks:)
[23:17] <Amaranth> seb128: mysql fork
[23:17] <seb128> now let's write something on the wiki, try the change on bug #435227 and call it a week
[23:18] <Amaranth> Optimized for clustering and cloud computing and I've just had a buzzword overload reading about it
[23:20]  * chrisccoulson downloads some ubuntu-boot crack
[23:21] <Amaranth> robert_ancell: Your application says you're only going for MOTU but you put "MOTU and package uploader" like me :)
[23:22] <Amaranth> I'm going to get reapproved as MOTU and hoping to get approved for compiz upload rights at the same time
[23:22] <robert_ancell> Amaranth, oh I just copied it thinking that MOTU and package uploader are the same thing...
[23:22] <chrisccoulson> robert_ancell - sort of. we can only upload packages to universe though
[23:23] <robert_ancell> yes
[23:23]  * robert_ancell drowns under red tape
[23:23] <chrisccoulson> i don't really know where i fit in with the archive reorg :/
[23:23] <Amaranth> robert_ancell: mine says "I, Travis Watkins, apply for MOTU and upload rights for compiz packages."
[23:24] <jono> kenvandine, are those gwibber fixes uploaded now?
[23:24] <robert_ancell> Amaranth, hmm, I think I copied your template - better check I didn't copy that
[23:24] <Amaranth> I figured I'd try to take advantage of the fact that I was MOTU before and ask for both at the same time :)
[23:25] <Amaranth> robert_ancell: Hey if you're still around after the meeting you can give seb128 a load of things to work on when he wakes up
[23:25] <Amaranth> Should make for a nice change, no? :)
[23:25] <robert_ancell> especially since it's Friday for me so I wont be here to respond after doing it. hehehehe
[23:25] <chrisccoulson> when is your meeting?
[23:25] <robert_ancell> 7 UTC
[23:26] <chrisccoulson> cool! well, i wish you both all the best, but i'm sure you'll both be fine ;)
[23:26] <robert_ancell> 5pm for me
[23:26] <robert_ancell> chrisccoulson, thanks!
[23:26] <Amaranth> I've got to decide if I want to try to stay awake for it or sleep early and try to wake up for it
[23:26] <chrisccoulson> i did my meeting at work, when i was meant to be working ;)
[23:27] <Amaranth> When I did mine we were still using the TB to approve people, was a fun mess :)
[23:27] <chrisccoulson> how long ago was that?
[23:27] <seb128> Amaranth, robert_ancell: I'm on holiday tomorrow to luck for you guys
[23:28] <Amaranth> haha
[23:28] <Amaranth> chrisccoulson: *shrug*
[23:28] <robert_ancell> seb128, foiled again!
[23:28] <Amaranth> I'm bad with dates
[23:28] <seb128> robert_ancell, I dropped a note on your motu wikipage
[23:28] <robert_ancell> Amaranth, that's why I love the gnome-panel clock/calendar :)
[23:28] <robert_ancell> seb128, thanks
[23:28] <seb128> good luck for the meeting tomorrow ;-)
[23:29] <Amaranth> Yeah, I had it screaming at me about calling my grandfather all day yesterday. At 9pm I realized I forgot anyway :/
[23:29] <Amaranth> So much for that birthday
[23:30] <chrisccoulson> brb (if my machine boots again)
[23:31] <robert_ancell> "chrisccoulson - eternal optimist"
[23:31] <seb128> lol
[23:31] <Amaranth> Considering the fun the last time someone landed in the ubuntu-boot PPA I'd be thinking the same thing
[23:32] <Amaranth> s/someone/something/
[23:32] <seb128> I did comment that ppa yesterday
[23:32] <seb128> I noticed that I still had it thanks to update-manager refusing to install an update since I didn't have the key for the ppa
[23:41] <Amaranth> Travis Watkins	2007-09-21	 Expired on 2008-09-20
[23:41] <Amaranth> Why didn't I click the button back then? :)
[23:48] <seb128> robert_ancell, you 'select default user' change does nothing there
[23:48] <seb128> do I need to restart something?
[23:48] <seb128> polkit-gnome-authentication-agent-1 I guess?
[23:48] <robert_ancell> yes
[23:48] <robert_ancell> but then you need to run it manually (it really should be in the session)
[23:48] <seb128> crashed
[23:49] <robert_ancell> dammit!
[23:49] <Amaranth> robert_ancell: Any idea what we're going to do about polkit-gnome not getting the xid of the window calling it now that upstream has rejected the patch?
[23:50] <Amaranth> Just forget it and hope people don't trigger it? It only affects non-compiz users now
[23:50] <robert_ancell> Amaranth, yeah I don't know - everyone upstream is arguing philosophies rather than user experience
[23:50] <seb128> robert_ancell, http://paste.ubuntu.com/288886/
[23:51] <seb128> robert_ancell, I got the crash by "gdmsetup, unlock, cancel, unlock"
[23:51] <robert_ancell> it does effect compiz because if you change focus away and then back you can leave the authentication dialog underneath the window.  The WM needs to associate the two windows somehow
[23:51] <Amaranth> robert_ancell: I know from my experience helping a user with karmic over the phone they got focused between network manager asking for wifi password and software-store asking for user password since they look similar
[23:51] <Amaranth> which would argue for a system modal dialog
[23:52] <Amaranth> s/focused/confused
[23:52] <Amaranth> Starting to think going to bed early would be a good idea
[23:52] <robert_ancell> We really want a proper application modal dialog I think so it is attached to the window.  We really should take Apples lead and attach our modal dialogs to the parent windows
[23:52] <robert_ancell> Amaranth, good idea! See you later
[23:53] <Amaranth> Hey I can make compiz do that :)
[23:53] <robert_ancell> System modal is a pain in the ass - I don't want to be locked out my other apps when something needs authentication
[23:53] <robert_ancell> Amaranth, do it, do it!!
[23:53] <seb128> robert_ancell, the issue is not the deluser, you get the invalid read on every cancel
[23:54] <seb128> invalid free rather
[23:54] <robert_ancell> seb128, oh, I wasn't cancelling...
[23:54] <seb128> robert_ancell, did you see the url I just copied?
[23:54] <robert_ancell> looking now
[23:54] <seb128> robert_ancell, I get that in valgrind on every cancel
[23:57] <robert_ancell> seb128, thanks, can't see immediately why that is wrong but will look into it
[23:58] <robert_ancell> seb128, and you're definitely using the second patch in the bug report, right?
[23:59]  * robert_ancell thinks seb128 may have passed out at the keyboard
[23:59] <seb128> robert_ancell, I'm checking, I didn't notice there was 2 of those