[04:45] <xnox> popey: there is no migration assistant in precise and up....
[04:45] <xnox> popey: well we removed migration assistant in 12.04.1 and up.
[04:45] <xnox> desrt: specifically which ones are affecting you?
[04:45] <xnox> (ubiquity fixes)
[05:58] <BigWhale> Morning all ...
[06:55] <didrocks> good morning
[07:03] <pitti> Good morning
[07:05] <didrocks> hey pitti, how are you?
[07:06] <pitti> didrocks: very well, thanks! I had a fine weekend
[07:06] <pitti> didrocks: how about you?
[07:06] <didrocks> pitti: very nice as well, a lot of snow, so most of the time was video gaming (mario galaxy II, continuing to get my backlog in shape!), but relaxing :)
[07:35]  * Sweetshark reports back from skiing. No major injuries, now completed 10 of 16 red tracks done for that area with no intentions for the rest as they are too far off.
[07:35] <Sweetshark> Also: Inbox 4300 -- if you have anything urgent I need to look at, give me a hint here on IRC.
[07:39] <didrocks> Sweetshark: welcome back! :)
[07:44] <Sweetshark> didrocks: thanks, southern france was an excellent host! (Which also means that despite only one day without skiing, I didnt loose any weight: magret de canard sauce framboise et the gourmand were just too tempting.)
[07:45] <Sweetshark> Its all muscles and heavy bones by now of course!
[07:46] <didrocks> Sweetshark: ahah, yeah, that's the trap. For me, it was more fondue savoyarde, raclette… :-)
[07:47] <didrocks> Sweetshark: how was the weather? It wasn't splendid for me though
[07:55] <jibel> good morning
[08:01] <Sweetshark> didrocks: First week was very nice, only one day with rain and fog and wind (all the higher lifts were closed, I did a training session in lower altitudes (1700-1400m) that day. Second week had more new snow and less sun, so I didnt get the traditional drive drown the Massif Lauziere (was closed: they are more careful with that track because of multiples avalanches in the recent years) ... "I love the sound of anti-avalanche detonations in 
[08:01] <Sweetshark> didrocks: I can evade raclette et fondue, they are part of the organized progam: If you want to join the ride down the tracks at night with a torch, you have to take part ;)
[08:03] <didrocks> salut jibel :)
[08:03] <didrocks> Sweetshark: ahah ;)
[08:05] <jibel> Salut didrocks
[08:05] <Sweetshark> jibel: morning!
[08:07] <jibel> good morning Sweetshark !
[08:11] <BigWhale> pitti, I owe you a beer or two ... :)
[08:11] <pitti> hey BigWhale
[08:14] <pitti> BigWhale: for what? :)
[08:16] <BigWhale> pitti, for the pygobjecy/pycairo stuff/
[08:16] <pitti> BigWhale: ah, I got it to build now, will update the upstream patch in a bit
[08:16] <BigWhale> pitti, I'll go and bug them to make a new pycairo release.
[08:17] <BigWhale> It was last released in 2011 ...
[08:42] <chrisccoulson> good morning everyone
[08:44] <didrocks> good morning chrisccoulson :)
[08:44] <chrisccoulson> hi didrocks, how are you?
[08:46] <didrocks> chrisccoulson: I'm good, thanks! yourself?
[08:49] <chrisccoulson> didrocks, yeah, i'm not too bad thanks
[09:03] <Laney> helloooo
[09:03] <pitti> hey Laney
[09:03] <pitti> hey chrisccoulson
[09:03] <chrisccoulson> hi pitti, how are you?
[09:03] <chrisccoulson> hi Laney
[09:03] <pitti> chrisccoulson: I'm great, thanks!
[09:03] <pitti> how about you guys?
[09:05] <chrisccoulson> pitti, yeah, not too bad thanks. ruby had her first sleep over at her grandparents house at the weekend
[09:05] <chrisccoulson> we spent the night in a hotel a few miles away
[09:05] <chrisccoulson> just in case we were needed ;)
[09:05] <pitti> chrisccoulson: hah, as a trial run? how did it go?
[09:06] <chrisccoulson> yeah, she was fine with it
[09:06] <Laney> the main highlight of my weekend was finding a beer shop had opened in the city and going to visit it :-)
[09:06] <pitti> chrisccoulson: sweet! so, you've got the prospect of a quiet night every now and then now :)
[09:07] <chrisccoulson> pitti, sort of. we've still got maisie as well ;)
[09:07] <pitti> Laney: ah, I saw your G+ post; already tried some?
[09:07] <chrisccoulson> Laney, oh, what sort of beer? :)
[09:07] <Laney> nope, didn't have a chance yet
[09:07] <Laney> chrisccoulson: http://twitpic.com/c27ejy !
[09:08] <chrisccoulson> oh, nice :)
[09:09] <Laney> http://www.flippinggoodbeershop.co.uk/
[09:10] <seb128> hey pitti, chrisccoulson, Laney
[09:11] <chrisccoulson> hi seb128
[09:11] <pitti> bonjour seb128, ça va?
[09:11] <seb128> pitti, oui, très bien, et toi ?
[09:11] <pitti> seb128: je vais bien aussi, merci! j'ai eu une bonne we, avec Taekwondo, théâtre, et soleil :)
[09:12] <seb128> pitti, j'ai vu ton post sur le taekwondo sur g+
[09:12] <seb128> brrr, it's freezing (and snowing) here
[09:12] <pitti> ici aussi
[09:12] <pitti> -10
[09:12] <seb128> I've enough of winter, please bring warmer weather
[09:13] <Laney> indeed
[09:13] <chrisccoulson> i like winter
[09:13] <chrisccoulson> ;)
[09:14] <ogra_> seb128, ++
[09:14] <seb128> I like winter, but only so long ...
[09:18] <czajkowski> morning folks, did an upgrade to raring last night and this was the first message I saw, http://ubuntuone.com/5VKWYSD1q9TfkPfkqTcv56
[09:19] <Laney> true
[09:19] <BigWhale> czajkowski, you got auto trolled by your desktop background ... :)
[09:20] <czajkowski> well it;s not the most helpful of messages :/
[09:20] <seb128> pitti, looking to that udisks/eject/power off bug, do you plan to get udisks 2.1 in raring?
[09:21] <pitti> seb128: now that we have gvfs 1.15.x, I actually do
[09:21] <seb128> pitti, excellent ;-)
[09:21] <ogra_> czajkowski, yeah, there are so much more baeutiful 404s like http://kvartirakrasivo.ru/404/index.php
[09:22] <Laney> http://archive.ubuntu.com/ubuntu/dists/raring/main/dist-upgrader-all/current/
[09:22] <Laney> how are those made?
[09:22] <pitti> seb128: I didn't upload it yet as David still did quite some architectural changes and I wanted to let it settle for a bit
[09:22] <pitti> seb128: but I'll get it in before FF
[09:23] <seb128> pitti, ok, makes sense, I was wondering if we should look at backporting the patches needed for the gvfs poweroff stuf or not ... I will just wait for the update, danke ;-)
[09:23] <czajkowski> ogra_: yer a bit evil :)
[09:23] <ogra_> haha
[09:24] <czajkowski> hmm I also have two U1 clouds now also :/ feck
[09:24]  * ogra_ wishes he had one that worked 
[09:24] <ogra_> how do you guys use U1 with 2fa ?
[09:53] <Laney> .
[09:53] <Laney> ogra_: what problems do you have? I don't remember anything untoward in that regard
[09:53] <Laney> also there is #ubuntuone ;-)
[09:54] <ogra_> Laney, U1 uses my SSO account .... which is tied to 2Fa
[09:54] <ogra_> if i want to be able to use U1 on a non registered device i always have to have my ubikey or a phone with the google tool with me
[09:54] <ogra_> that kind of defeats the purpose
[09:56] <Laney> oh to register new devices
[09:56] <ogra_> to quickly push or pull something from my GFs or MOMs computers
[09:57] <Laney> I never considered that as a problem really
[09:58] <ogra_> i have a ton of machines i definitely dont want to have registered with LP but from which i would like to use shared files
[09:58] <ogra_> so i never use U1 after all
[10:04] <xnox> ogra_: use online/web interface instead on non-registered devices.
[10:05] <xnox> ogra_: you can login donwload/upload stuff.
[10:05] <ogra_> log in without 2Fa ?
[10:05]  * ogra_ never used the web interface, i thought it uses the same account
[10:10] <xnox> no, it still has 2Fa
[10:11] <ogra_> right
[10:11] <ogra_> i guess to effectively get around that i can only create an extra account (which i dont really want)
[10:12] <Laney> can't you just log out when you're finished?
[10:12] <ogra_> and that helps how ?
[10:13] <ogra_> sure i can log out, but that only means i need to log in again
[10:13] <Laney> Then I don't really understand what your problem is.
[10:13] <ogra_> with another annoying 2Fa run
[10:13] <Laney> just that it's annoying?
[10:13] <ogra_> Laney, i would like to use U1 without having to have my 2Fa key with me all the time
[10:14] <ogra_> or registering devices etc ...
[10:14] <ogra_> U1 is way more complex to use than it would need to be which makes me simply not use it
[10:14] <pitti> hm, I have my mobile pretty much everywhere; that works fine for 2fa?
[10:15] <ogra_> i just changed to a new mobile that i hadnt registered yet ... if you then are somewhere else than home and want to access your files you are screwed
[10:15] <Laney> it's even worse for me - I need a password manager to know my SSO password ;-)
[10:16] <ogra_> thats beyond any kind of usability imho
[10:16] <ogra_> (well, the whole 2Fa concept is imho)
[10:16] <pitti> I thought you only needed 2fa for the U1 web uI?
[10:17] <ogra_> you need your SSO account for U1
[10:17] <pitti> I'm using U1 to sync my music, and the desktop never asked me for 2fa AFAIR
[10:17] <ogra_> even in the native UI
[10:17] <pitti> ah, perhaps when I signed up first
[10:17] <ogra_> right
[10:17] <pitti> but not in day-to-day operation
[10:17] <ogra_> which means you effectively register the device for your account
[10:18] <ogra_> i cant just sit on my moms PC, quickly push and pull some files without jumping through a lot of hoops
[10:18] <pitti> right
[10:18] <ogra_> so in the end i use sshfs with my webserver because its so much less effort
[10:18] <ogra_> which kind of defeats the purpose of U1
[10:18] <pitti> yeah, for that use case a network storage is much better than syncing files locally
[10:20] <ogra_> well, but thats not how i understand personal cloud :)
[10:21] <ogra_> it should make me want to use it over old methods ;)
[10:21] <pitti> a gvfs backend for U1 would be quite nice indeed :)
[10:21] <pitti> to directly access files in the cloud instead of having to download everything
[10:21] <pitti> and/or a fuse module
[10:22] <ogra_> yeah
[10:46] <pitti> robru: did you ever make "valadoc" work for libfriends?
[10:47]  * pitti found bug 1097111 in his quest to use valadoc for umockdev
[10:47] <ubot2> Launchpad bug 1097111 in libfriends "Fix valadoc ;-)" [Medium,Triaged] https://launchpad.net/bugs/1097111
[10:52] <pitti> hm, it fails the same way for libgee, so I guess our version just doesn't get along with vala 0.18
[10:55] <pitti> indeed, it looks in /usr/share/vala-0.14/, so it needs an explicit --vapidir
[11:11] <jibel> pitti, would 4h be a reasonable default for timeout of automated runs of jhbuild? the longest build is webkit it takes 2h or so
[11:12] <jibel> pitti, I nearly finished implementing the feature and will submit the patch upstream
[11:12] <pitti> jibel: did we ever actually see a build hang? I think we mostly need a timeout for "make check"?
[11:12] <pitti> jibel: \o/
[11:12] <pitti> jibel: 4 h sounds good to me, but we might want to start with 2.5 then?
[11:13] <jibel> pitti, I meant 'make check' but the timeout will apply to every phase and is configurable
[11:14] <jibel> the way jhbuild is structured make it difficult to implement a 'per phase' timeout
[11:14] <pitti> *nod*
[11:48]  * Laney eyes webkit with suspicion
[11:48] <Laney> y u drop symbols
[11:49] <czajkowski> Laney: swap you sympbols for odd lookin icons
[11:49] <czajkowski> not sure if this is what it's meant to be
[11:49] <czajkowski> or just odd
[11:49] <Laney> I mean library symbols :P
[11:49] <czajkowski> I'm still willing to swap :)
[11:50] <Laney> el bug report
[12:29] <Laney> bah, too much churn http://paste.ubuntu.com/1636182/
[14:09] <glatzor> hello mpt
[14:17] <cyphermox> good morning!
[14:23] <marga> desrt, hey there!
[14:23] <marga> desrt, it's me with dconf bugs again :)
[14:24] <marga> desrt, https://bugs.launchpad.net/ubuntu/+source/glib2.0/+bug/1122028
[14:24] <ubot2> Ubuntu bug 1122028 in glib2.0 (Ubuntu) "gsettings does not respect dconf locks" [Undecided,Confirmed]
[14:24] <marga> I'm trying to figure out what's going on.  Sometimes, gsettings values don't match dconf values.
[14:44]  * xnox ponders if connection information should show hostname or not
[14:47] <desrt> marga: hi
[14:47] <psivaa> hello, bug 1122072 is affecting precise installations on VirtualBox.
[14:47] <ubot2> Launchpad bug 1122072 in xorg (Ubuntu) "[Precise amd64 on VirtualBox] "Fatal server error: no screen found" in Xorg.0.log" [Undecided,Confirmed] https://launchpad.net/bugs/1122072
[14:47] <desrt> marga: see https://bugzilla.gnome.org/show_bug.cgi?id=693149
[14:47] <ubot2> Gnome bug 693149 in daemon "idle-delay value ignored when locked" [Normal,Resolved: notabug]
[14:48] <marga> uhm.
[14:48] <marga> really?
[14:48] <marga> uint32 10 ?
[14:48] <desrt> yup
[14:50] <jibel> psivaa, what's the kernel on the installation media?
[14:51] <psivaa> jibel: 3.5.0-23-generic
[14:52] <psivaa> jibel: 'init: lightdm main process (2405) terminated with status 1' in dmesg
[14:53] <marga> desrt, :-/  I don't think this is right.  Why does dconf think it's a signed number?  Shouldn't it default to unsigned unless there's a sign there?
[14:58] <desrt> marga: no.  just as in C, the default is to assume signed
[14:59] <marga> Except this is a config file
[14:59] <marga> Not a C program.
[14:59] <desrt> marga: it's using the GVariant parser
[14:59] <desrt> which is the same everywhere...
[15:00] <desrt> when you run 'dconf update' it doesn't have access to the schemas
[15:00] <desrt> so it doesn't know about what type is expected
[16:13] <marga> desrt, is there no way that dconf could find out about the non-matching types?
[16:13] <desrt> no.  dconf is schemaless
[16:14] <marga> how does dconf communicate with gsettings?
[16:14] <desrt> gsettings has a pluggable backend mechanism based on the GSettingsBackend interface
[16:14] <desrt> dconf implements that
[16:15] <desrt> (it's not a capital-I Interface-- just abstract base class.... same idea, tho)
[16:18] <marga> Ok.  I understand, but I still think this is something that needs to be fixed
[16:19] <marga> I'm not sure if it needs to be fixed on the dconf side, or the gsettings side.
[16:19] <desrt> it's an easy fix: you just need to write the correct type in the file
[16:19] <marga> But it really makes little sense to cast a value in a config file.
[16:19] <marga> Yes, yes, but that's to fix our current setup
[16:19] <desrt> it's not a cast -- it's a type annotation
[16:19] <marga> I want to fix it so that other people are not biten by the same thing.
[16:20] <desrt> one thing that could be done is to emit a g_warning() when an incorrectly typed value is found
[16:20] <desrt> but i'm not sure how much good that would do
[16:20] <desrt> since it would just go to the .xsession-errors where you may or may not see it (and probably not)
[16:21] <marga> Right, there are waay to many warnings being emitted
[16:22] <marga> So, it's difficult to find a really important one
[16:22] <desrt> meanwhile i can't change the gvariant parser -- it's far too widely used in places where it is assumed that it will remain compatible
[16:22] <desrt> and i can't change the dconf db file format for the same reason
[16:23] <desrt> i also can't really introduce a hard dependency on schemas being present during 'dconf update' both because it would again break compatibility and would be a weird layering violation
[16:23] <marga> this incorrectly typed value... Who would find it? dconf or gsettings?
[16:23] <desrt> and i don't want an optional dependency (ie: providing 'hints' if the schema is found) because the output of 'dconf update' would change depending on if or if not these files in another directory happened to be installed
[16:23] <desrt> gsettings is finding it, in context of the program doing the read
[16:24] <desrt> http://git.gnome.org/browse/glib/tree/gio/gsettingsbackend.c#n725
[16:24] <desrt>   if G_UNLIKELY (value && !g_variant_is_of_type (value, expected_type))
[16:24] <desrt>     ....
[16:24] <marga> what's G_UNLIKELY? :)
[16:25] <desrt> it'a a static branch prediction hint
[16:25] <marga> ah, heh
[16:25] <desrt> we have a bit of a convention of decorating our unexpected error cases with it
[16:25] <desrt> such that the primary purpose is probably not even related to the performance improvements but improving readability
[16:26] <marga> would it be too messy to check for casting possibilities in that if?
[16:26] <desrt> ya... i definitely don't want to do that
[16:26] <desrt> because then you have a problem but you never find out about it
[16:26] <marga> So... We are left with adding a warning?
[16:26] <marga> :-/
[16:27] <desrt> i'm not even really crazy about a warning
[16:27] <marga> Well, I've been having this problem for 2 months or so
[16:27] <marga> And I couldn't fix it until I contacted you
[16:27] <desrt> you usually don't expect stderr output from properly-behaving GUI applications because (for example) a config file has a bad format
[16:27] <desrt> hmm
[16:27] <desrt> you should really ping sooner :)
[16:27] <desrt> but ya... i understand your frustration
[16:28] <marga> heh, we didn't know whose fault it was.  We thought it was gnome-screensaver or something, until Jake found the gsettings discrepancy
[16:28] <marga> I don't even know how he thought of checking that.
[16:29] <desrt> so one thing that could possibly be kinda OK is for 'dconf update' to open the gsettings schemas only for the purpose of issuing warnings in the case of a type mismatch
[16:29] <desrt> ie: no impact on the output... just an additional diagnostic printed
[16:29] <desrt> you're much more likely to see stderr output from 'dconf update' than some GUI program
[16:30] <desrt> (particularly something like the screensaver, running as part of the session)
[16:30] <desrt> anyway.... adding this check would be non-trivial and annoying
[16:31] <desrt> but if someone wanted to do the work, i think it would be helpful to prevent these types of situations
[16:31] <happyaron> seb128: I guess it's not quite possible to have ibus 1.5 for Raring.
[16:32] <marga> desrt, alright.
[16:32] <desrt> in short: if you look at how 'gsettings list-recursively' works, we could take each schema object visited by that command and check its path against the value that you're trying to set in 'dconf update'
[16:32] <happyaron> seb128: we've encoutered some difficult problems because of the gconf -> dconf move, which is... actually in a mess and need to be sorted out.
[16:32] <desrt> and do a typecheck
[16:32] <seb128> happyaron, you know better about IMs than me, what do you think? It's going to be difficult to update the gnome-settings-daemon/gnome-control-center keyboard layout handling code because it would mean having to rewrite our indicator
[16:33] <marga> although "non-trivial and annoying"\ doesn't sound too tempting :)
[16:33] <seb128> happyaron, oh, that was a question or a statement? ;-)
[16:34] <happyaron> seb128: I guess statement... though we're trying to fix every bit, but new problems come out from time to time. Use a newer but not tested version isn't a good idea.
[16:34] <desrt> marga: the big take away, though, is don't spend 2 months stuck on _anything_ without starting to ping people
[16:34] <seb128> happyaron, right
[16:34] <desrt> marga: even if just randomly for them to say "this is not my problem"
[16:34] <seb128> happyaron, who is "we" there (just curious) ... do you have details on the issues (bug reports, list discussions?)
[16:34] <marga> desrt, sure, it's just that it's not always easy to know who to ping.
[16:35] <marga> like... I have this very weird bug that I've been chasing for months... Let me look it up...
[16:35] <happyaron> seb128: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=700076
[16:35] <ubot2> Debian bug 700076 in ibus "ibus: non-functional, setup breaks" [Grave,Open]
[16:36] <marga> https://bugs.launchpad.net/ubuntu/+source/gnome-power-manager/+bug/998735
[16:36] <ubot2> Ubuntu bug 998735 in gnome-power-manager (Ubuntu) "Light blue screen on primary monitor when waking up from inactivity" [Undecided,Confirmed]
[16:36] <happyaron> seb128: I'm working with another people in ~pkg-ime, and I guess you know him, Yunqiang Su.
[16:36] <seb128> happyaron, right
[16:37] <seb128> happyaron, thanks for the update, we are still on ibus 1.4 but I would like to go for the new version and gtk3/gsettings at some point
[16:37] <seb128> we just need to figure how to get there
[16:37] <seb128> but probably not this cycle right
[16:37] <marga> This bug affects cinnamon and gnome, but not unity (in precise).  The bug is caused when using 2 monitors, regardless of the graphics card, regardless of the screensaver.  When the screen is blocked and the user receives a notification, the left screen becomes blue.
[16:38] <desrt> marga: for issues like that i'd bug RAOF or bryce
[16:38] <marga> RAOF is a nickname?
[16:38] <desrt> yes
[16:38] <desrt> he's on aussietime, though
[16:38] <marga> oh... :-/
[16:38] <marga> Ok.
[16:39] <desrt> but next week he'll be on brittime, i think
[16:39] <happyaron> seb128: I know the point, but I think let's wait for a bit more time. During my research of recent days I find that upstream is confused by the difference of gobject versions, which leads to quite a lot of problems.
[16:40] <marga> ok, then.
[16:40] <marga> tnx
[16:40] <seb128> happyaron, ok, do you plan to open bugs against ibus upstream about the issues you find? please give me the numbers if you do, I'm interested to follow the progresses on those
[16:40] <desrt> marga: about the other issue, i'd hate to say 'patches accepted' but i probably won't have time to implement it myself soon
[16:41] <desrt> (and i still don't _totally_ love the idea)
[16:41] <happyaron> seb128: I'm still in the progress of tracking issues that are already reported/fixed in RH bugzilla, that takes time...
[16:41] <seb128> ok
[16:42] <marga> desrt, well, who knows.  Maybe we'll come up with something better
[17:16] <xnox> Laney: so I set all the keys for g-s-d to draw background and it doesn't.... I ssh in, fetch the DBUS_SESSION_BUS_ADDRESS, rerun gsettings set org.gnome.desktop.background draw-background true and boom it draws it.
[17:16]  * xnox gtk why u no draw early
[17:20] <jbicha_> seb128: what do you think of https://bazaar.launchpad.net/~ubuntu-desktop/gnome-terminal/ubuntu/view/head:/debian/patches/02_add_transparency_properties.patch
[17:20] <larsu> seb128: I'm running attente's ppa, and it works well.  What's your plan on when to get it into the desktop ppa (or even raring)?
[17:21] <larsu> (I mean the unity-gtk-module thing)
[17:21] <jbicha_> gnome-terminal 3.7 dropped support for translucent backgrounds and since that patch was disabled for so long, I'm wondering whether it would be better to turn it off again
[17:21] <seb128> jbicha_, no opinion, feel free to drop it if the feature is gone
[17:21] <Laney> xnox: is this for the installer? ship a gsettings override?
[17:21] <larsu> people still want to use transparent terminals? What is this, the 90s?
[17:21] <seb128> larsu, hey, I was planning to see with him in London next week, maybe having it in the desktop ppa this week and land to archive next week would be good
[17:21] <ogra_> xnox, did you compile the scheme etc ?
[17:22] <ogra_> xnox, also in the scheme file the right quoting counts :)
[17:22] <ogra_> iirc
[17:22] <larsu> seb128: sounds awesome!
[17:24]  * didrocks waves good evening
[17:24] <xnox> Laney: in the installer, and it kind of executes `gsettings set schema key val` with python subprocess
[17:25]  * xnox ponders if i can ship overrides and load them upon launching gnome-settings-daemon
[17:25] <smspillaz> bug 1111111
[17:25] <ogra_> ubiquity should just ship them
[17:25] <smspillaz> (cmon)
[17:25] <ogra_> so they get removed with the package again
[17:26] <Laney> indeed
[17:27] <ogra_> there is a lot of stuff in casper that should just be in an override file from ubiquity actually
[17:27] <ogra_> we should consolidate that at some point and clean it up
[17:27] <ogra_> (casper really only needs to set settings that arent ubiquity related imho)
[17:29] <xnox> Laney: which override "wins" for the same key? first one or last one?
[17:31] <Laney> xnox: higher number I believe
[17:31] <Laney> call it 99_ubiquity-hacks.gstreamer.override or so
[17:31] <Laney> err not gstreamer, YKWIM
[17:31] <Laney> too much sessioninstaller on the brain
[17:34] <xnox> hehe ok.
[17:36] <robru> pitti, kenvandine was looking into that and I think he got it working for him, but I was never able to build it for unknown reasons.
[18:08] <xnox> Laney: right, so i think I'm getting to the bottom of this. g-s-d background plugin expects SessionManager on the dbus and listens to SessionRunning signal, but ubiquity runs without a session manager hence that is never fired and the background is not drawn.
[18:08] <xnox> option one is to use dbusmock to fake session manager =) (pitti will love this)
[18:08] <xnox> or i'm just gonna take g-s-d background plugin and refactor it into the ubiquity/wallpaper app.
[18:09] <xnox> (g-s-d implementation looks cleaner and supports screen size changes \o/ )
[18:30] <jbicha_> how much benefit is disabling the background plugin giving us?
[18:32] <ogra_> disabling ?
[18:32] <ogra_> xnox tries to enable it
[18:33] <jbicha_> ogra_: we disable the plugin by default which by default breaks setups where nautilus doesn't draw the desktop
[18:33] <notgary_> Question for the room at large: Regarding the closure of bug reports as invalid, what's the etiquette for using the 'No longer affects project' button?
[18:34] <ogra_> oh. that, that needs reviewing anyway
[18:34] <jbicha_> I'm curious if it saves 1 second on startup or saves 20k of RAM…
[18:34] <notgary_> I'm thinking about this for the paper cuts project
[18:34] <ogra_> jbicha_, there is a WI to make nautilus only start if there are actual icons on the desktop
[18:34] <notgary_> Since taking our task off the report won't affect the Ubuntu or upstream task
[18:34] <ogra_> jbicha_, that wouldnt work without the plugin enabled
[18:35] <ogra_> (unless we can just switch it on on the fly, but i guess that adds startup time)
[18:36] <seb128> jbicha_, it doesn't save a lot but small wins add, and users who go to gsettings to turn nautilus desktop drawing off can as well turn something else on
[18:37] <jbicha_> seb128: ok but Ubuntu GNOME ships with desktop-icons-off by default
[18:37] <seb128> well maybe it should override the g-s-d plugin back to on then?
[18:37] <jbicha_> so should ubuntu-gnome-default-settings explicitly override ubuntu-settings then?
[18:37] <seb128> you already do if you turn nautilus off...
[18:38] <jbicha_> seb128: not currently, we just hope people haven't installed ubuntu-settings :|
[18:38] <seb128> well, broken one way or the another ;-)
[18:38] <seb128> I'm fine saying that people who install the gnome settings want the gnome settings
[18:38] <seb128> so overriding the ubuntu ones
[18:39] <jbicha_> ok I'll try that then
[18:42] <jbicha_> I wish Tweak Tool didn't use an autogenerated UI as I don't know how to patch it to explicitly turn on the background plugin if desktop-icons are turned off
[18:44] <desrt> erm
[18:44] <desrt> we should never use the background plugin
[18:45] <desrt> either we should have nautilus drawing the background or compiz
[18:45] <desrt> never gnome-settings-daemon
[18:45] <desrt> background plugin just doesn't make sense when you have a compositor
[18:47] <xnox> desrt: can compiz draw background for me?
[18:47] <desrt> i don't know
[18:47] <desrt> but it ought to be able to and it certainly _should_
[18:47] <desrt> i mean.... one way or another compiz _is_ drawing your background for you as the first thing that it does when compositing your screen
[18:48] <xnox> desrt: the manpage '--bg-image IMAGE' is a lie. All other references to bg image in compiz refer to integration with nautilus/kwin (pre gtk3 world) to change bg_images based on which virtual desktop is active
[18:48] <desrt> things are more efficient if it just draws your wallpaper directly rather than some other process drawing it and compiz picking it up and copying it
[18:48] <xnox> desrt: well, it's just black if I don't have wallpaper app running in ubiquity - gsd manages to draw a background. I didn't manage compiz to trick into drawing a background.
[18:49] <desrt> isn't there some compiz plugin for this?
[18:49] <xnox> my next step is to fix up wallpaper app - currently it draws noise.
[18:49]  * desrt has a hard time believing that for all the plugins compiz has it would lack such a simple thing
[18:49] <xnox> desrt: well, I've search and the only bit I found was that integrates into nautilus. Didn't find a basic - slap this image on the screen background plugin.
[18:50] <desrt> :(
[18:50] <desrt> this is definitely the proper solution -- even if it doesn't exist, i'd suggest writing it :)
[18:50] <xnox> maybe I'm not up-to-date with how to enable / configure such a plugin from command-line (as ubiquity spawns it's own compiz with only basic plugins enabled to create a bare minimum / lightweight environment for ubiquity window to run in)
[18:50] <desrt> smspillaz or duflu or even racarr ought to know more about this...
[18:51] <xnox> well currently we have a botched up single C file wallpaper app which stopped playing nice with rarings gtk/gdk stack =)
[18:51] <seb128> I would just use the g-s-d plugin in the installer
[18:51] <seb128> not worth the work to write a compiz plugin
[18:51] <seb128> (if there is none for that)
[18:52] <desrt> seb128: do you really think that compiz would be missing a plugin for _anything_? :)
[18:52] <xnox> seb128: turns out background plugin in g-s-d wants session manager which we are not running. So my current plan is to refactor g-s-d background plugin back into wallpaper app.
[18:52] <seb128> desrt, well from a quick glance in ccsm I don't find one for that
[18:52] <xnox> seb128: plus it doesn't help with Xubuntu/Mythbuntu/Studio as they don't have g-s-d nor compiz.
[18:52] <seb128> hum, k
[18:53] <desrt> http://wiki.compiz.org/Plugins/Wallpaper <- what is this?
[18:53] <desrt> according to the docs on that page, the proper way to get it to work is to kill off the wallpaper drawer of your desktop environment
[18:53] <desrt> https://launchpad.net/compiz-wallpaper-plugin
[18:53] <desrt> you should definitely talk to smspillaz about this
[18:53] <xnox> hmm... do we ship that?
[18:54] <seb128> seems not
[18:54] <seb128> but that wouldn't help much, as you said most images don't use compiz
[18:54] <seb128> so for the installer it makes sense to simply draw the background
[18:54] <jbicha_> https://bugzilla.gnome.org/show_bug.cgi?id=686549
[18:54] <seb128> without having to bother about configuration, scalling, etc
[18:54] <ubot2> Gnome bug 686549 in background "background: remove code that deals with nautilus drawing the desktop" [Normal,Unconfirmed]
[18:55] <xnox> seb128: yeah.... =(
[18:56] <desrt> what is the goal here?  what installer?
[18:57] <seb128> desrt, ubiquity in install only mode
[18:57] <seb128> desrt, that's used by xubuntu, lubuntu, etc as wel
[18:57] <seb128> l
[18:57]  * desrt would guess that the non-unity/gnome desktops have their own ...
[18:57] <desrt> oh
[18:57] <desrt> xubuntu is not installed by a livecd running a xubuntu desktop
[18:57] <seb128> same as ubuntu
[18:57] <xnox> kubuntu using python-qt/X api to draw a picture in the background fullscreen =)
[18:57] <seb128> you have both items
[18:57] <desrt> interesting
[18:57] <seb128> "try desktop" and install from there
[18:57] <seb128> or install only
[18:57] <desrt> if this is just for the installer then ya... stop wasting time
[18:57] <seb128> which just runs ubiquity
[18:58] <desrt> but for the desktop we're gonna need to kick our g-s-d background dep. for the case that nautilus desktop drawing is disabled
[18:58] <xnox> desrt: install-only is like an upstart jobs which spawns an X server, paints background creates a fake panel with faked network-manager and opens ubiquity window. The "best" entertainment one gets is the slideshow =) it is meant to be low-RAM friendly yet prettier than alternate cd
[18:59] <desrt> neat.  got a11y?
[18:59] <xnox> desrt: yeah we spawn that as well based on kernel / boot params or if Ctrl+L gets clicked.
[18:59] <xnox> we have magnifier/alternative input/screenreader/low-contrast.
[19:00] <xnox> it's like a mininiature desktop environment with smoke and mirrors =)
[19:00] <xnox> supporting kwin, lxde, xcfe, compiz, metacity, etc.....
[19:01] <desrt> seb128: btw: i was wondering what you will do about the keyboard stuff this cycle.  did you plan to try to stay on old ibus and the 3.2/3.4 era panel?
[19:01]  * desrt tried out F18 over the weekend... wow...
[19:01] <seb128> desrt, yes
[19:02] <seb128> desrt, it might not be the nicest ui but it's allowing people in china to use their computer... ;-)
[19:02]  * desrt wanted to stab someone when trying to ues it
[19:02] <desrt> turns out that all the features that i thought they removed were just recategorised.... arguably in better places
[19:02]  * desrt still has an uneasy feeling
[19:03] <desrt> when/if we migrate we're gonna have to deal better with it than did upstream
[19:03] <desrt> like 3rd level switcher and compose keys are no longer accessible via xkb options panel... but they added items to the 'shortcut keys'
[19:03] <desrt> which kinda makes sense, but only if you've never used it before
[19:04] <desrt> for existing users it's like *argghgh*
[19:05] <desrt> they also removed esperanto support from the primary UI and stuck it in tweak tool.  a huge setback to the international agenda, imho :(
[19:07] <bryce> desrt, "they" being xkb, or gnome?
[19:07] <desrt> gnome
[19:07] <bryce> desrt, ah.
[19:07] <desrt> you know that big ugly xkb options panel that you can get to now by clicking on 'options'?
[19:07] <bryce> desrt, yeah
[19:07] <desrt> that's completely gone as of gnome 3.6
[19:07] <bryce> :-(
[19:08] <desrt> a select few things are lifted out of it into a nicer UI (like compose, 3rd level chooser, etc)
[19:08] <desrt> the rest are dumped into the tweaktool
[19:09] <bryce> desrt, I know there's some options from that which many people switch on (e.g. compose key, 3rd level, etc.)  I've been wondering if we ought to consider making some of those on by default.  Like R-Alt is compose by default, or whatever.
[19:09] <desrt> i use r-alt for 3rd level fwiw
[19:09] <desrt> which i guess is a pretty common thing to do
[19:09] <desrt> some keycaps even have 'AltGr' written on that key
[19:10] <desrt> but i guess at the same time some users would be quite surprised to find that alt+f (chorded using the right alt key) is no longer doing what it used to
[19:10] <bryce> desrt, well, this is why I think this.  _Because_ they're so widely used, and because they're not on by default, and because we expose so many options, there's huge diversity in what people have gotten used to using.
[19:11] <desrt> you're starting to sound like a gnome designer :)
[19:11] <bryce> desrt, maybe the inverse :-)
[19:11] <jasoncwarner_> chrisccoulson hey dude...rickspencer3 just mentioned that he sometimes has the goog hangout plugin freeze and completely take down firefox
[19:12] <chrisccoulson> jasoncwarner_, srsly? :(
[19:12] <bryce> desrt, I'd rather see better defaults with a detailed config tool, than no defaults and a braindead config tool
[19:12] <chrisccoulson> jasoncwarner_, any steps to reproduce?
[19:12] <desrt> bryce: ya.. so you're probably kinda on-page with the new gnome 3.6 approach already, then
[19:12] <desrt> meanwhile, tweak-tool allows access to the 'other' options
[19:13] <desrt> (although it does it with an interface that's even worse than the old xkb options interface from the control centre)
[19:13] <jasoncwarner_> chrisccoulson not sure on steps, he just said it crashes, then firefox becomes unresponsive and he has to killall it
[19:13] <bryce> desrt, worse?
[19:13] <desrt> ya.. seriously
[19:13] <bryce> (how??)
[19:13] <desrt> ll
[19:13] <bryce> anyway, ok...
[19:13] <chrisccoulson> jasoncwarner_, it might be helpful if rickspencer3 could attach gdb to it once it's hung :)
[19:14] <desrt> can't find a screenshot and don't want to boot fedora to get you one
[19:14] <desrt> but rest assured, it's substantially worse
[19:14] <bryce> desrt, don't worry about it, I prefer to live life in denial
[19:14] <desrt> basically it's a list of drop-downs with labels to the left
[19:14] <desrt> the label on the left is the xkb option name (as you would write it in the XKBOPTIONS string)
[19:14] <desrt> the pulldown has the strings that would appear on the right side in the XKBOPTIONS string
[19:15] <rickspencer3> chrisccoulson, ack, I'll see if I can get a stacktrace next time
[19:15] <desrt> so like lv3     [ ralt_switch ]
[19:15] <bryce> desrt, bet it was easy to implement!
[19:15] <chrisccoulson> rickspencer3, cool, thanks :)
[19:15] <desrt> bryce: i imagine so :)
[19:16] <bryce> desrt, thankfully our users *love* it when we change the config tools
[19:17] <desrt> ya.... i really can't recall the last time i heard a complaint in response to something like that
[19:51] <thumper> morning
[19:59] <jbicha_> desrt: have you tried gnome-control-center 3.7.5 yet? I think they moved several of the layout options back
[19:59] <desrt> no
[19:59] <desrt> just F18
[20:02] <seb128> it's funny to see how GNOME is adding back stuff we distro patched by then when we updated to GNOME3, after they wontfixed those for over a year
[20:02] <seb128> like they are adding a "never" option back to to "screen off after" list
[20:03] <seb128> it's not the first one to happen this cycle ;-)
[20:03] <seb128> they also decided to keep nautilus rendering of the desktop for their new "fallback"
[20:07]  * desrt disappears to the dentist for an hour or so
[21:07] <dobey> ugh nautilus; is the requirement to have nautilus running for drawing the background in raring now, a bug in gnome or compiz?
[21:07] <dobey> (or unity)
[21:30] <robert_ancell> jdstrand, should 09_lp577919-fix-chromium-launch.patch from lightdm be upstream?
[21:32] <jdstrand> robert_ancell: hey, which upstream?
[21:32] <robert_ancell> jdstrand, i.e. lightdm trunk
[21:33] <jdstrand> robert_ancell: lightdm upstream already has the guest session apparmor in trunk?
[21:33] <jdstrand> apparmor profile*
[21:33] <robert_ancell> jdstrand, yes (we're upstream)
[21:33] <robert_ancell> bug 577919
[21:33] <ubot2> Launchpad bug 577919 in Light Display Manager "chromium-browser fails to start (guest account, OpenVZ): "Failed to move to new PID namespace: Operation not permitted"" [Medium,Triaged] https://launchpad.net/bugs/577919
[21:33] <jdstrand> robert_ancell: yes (I know we/you are upstream :)
[21:34] <robert_ancell> jdstrand, the plan is to split the guest session out so then the apparmour will be separate but for now it's in trunk
[21:34] <jdstrand> robert_ancell: what I was getting at is that you could collapse all the apparmor patches into one if you don't have them upstream. otherwise, I think including upstream is best
[21:35] <robert_ancell> jdstrand, I hadn't noticed they were there. Is there any reason not to just merge them into trunk?
[21:35] <jdstrand> not that I can think of
[21:36] <jdstrand> anyone else who wants lightdm and uses apparmor would benefit
[21:36] <robert_ancell> ok, done
[21:36] <jdstrand> robert_ancell: thanks
[21:43] <jbicha> dobey: I believe it's a compiz bug since GNOME Shell doesn't need the g-s-d bg plugin & GNOME plans to get rid of the plugin
[21:46] <dobey> jbicha: is the plug-in already gone?
[21:46] <dobey> i guess gnome-shell does it in the WM, which is just silly
[21:47] <jbicha> dobey: no and we won't be taking g-s-d 3.8 for raring anyway
[21:47] <dobey> jbicha: then shouldn't it still work with show-desktop-icons = false?
[21:49] <Laney> if you re-enable the plugin
[21:52] <jbicha> hmm, actually I guess it doesn't work correctly with raring's gnome-shell
[21:55] <dobey> Laney: so the correct answer to "was it removed" is "basically, yes; it's still there but not enabled by default" ?
[21:57] <Laney> it's disabled because, by default, it's not useful as nautilus handles it
[21:57] <Laney> the show-desktop-icons thing is an annoying gotcha, i'll grant you
[22:01] <stgraber> Laney: don't we have some clever gsettings migration scripts for that? surely we can explicitly turn on the plugin if we detect show-desktop-icons is false on upgrade
[22:01] <stgraber> (won't work for anyone changing this after the upgrade script has run, but that should fix the obvious upgrade case at least)
[22:04] <Laney> we do have the session-migration stuff, yeah
[23:34] <BigWhale> The fact, that windows can't be placed over Panel and Launcher with Gtk.window.move() is seriously disturbing.