[01:04] <chrisccoulson> hmmm, so, it takes the HUD 8 minutes to fetch 364364 bookmarks from firefox
[01:07] <RAOF> That's… unfortunate.
[01:08] <RAOF> Do you actually have 364364 bookmarks in firefox? :)
[01:09] <chrisccoulson> RAOF, i just created a setup to investigate performance suckiness. i could create 10-15 second hangs with a normal of bookmarks :)
[01:09] <chrisccoulson> (which is what other people are seeing)
[01:09] <RAOF> Booo.
[01:10] <chrisccoulson> and is also the reason why google don't put bookmarks in the chrome menu (because it causes chrome to hang for minutes at startup)
[01:11] <chrisccoulson> dbusmenu has O(N) performance when looking up menu nodes to deliver events to, basically
[01:11] <TheMuso> I know may a person who goes the menus to look for bookmarks, not knowing about the bookmarks toolbar...
[01:11] <TheMuso> many
[01:11] <chrisccoulson> ie, "lets just iterate over the entire tree of menuitems until we find the right one" :(
[01:15] <stgraber> pitti: when you're around, cjwatson started looking at gnome-keyring failing to build with my last upload (though the actual change is not the cause of the FTBFS), would be great if someone from desktop who's familiar with gnome-keyring could have a look
[01:15] <stgraber> pitti: the FTBFS prevents us from having a fix for a race condition in the install that in my tests happens around 10% of the installs (might be specific to my environment though)
[01:16] <stgraber> (so not beta2 critical but definitely something nice to have if we can)
[01:23] <chrisccoulson> RAOF, as a more realistic use case, 8000 bookmarks causes firefox to hang for around 40 seconds
[01:23] <chrisccoulson> which is pretty terrible
[05:08] <pitti> Good morning
[05:08] <pitti> stgraber: yes, can do
[05:18] <ritz> heya, the oneiric branch of gtk+3 is broken
[05:18] <ritz> https://bugs.launchpad.net/udd/+bug/848064/comments/8
[05:18] <ubot2> Launchpad bug 848064 in udd "Revision not present branching from udd-imported branches on lp" [Critical,Confirmed]
[05:27] <pitti> robert_ancell: good morning, how are you?
[05:27] <robert_ancell> pitti, hey, good
[05:33] <robert_ancell> pitti, so I've been out of the packing loop for a while.  Can you remind me how we manage packages like d-conf?  i.e. they don't have a packaging branch, what is the best way to update them?
[05:34] <pitti> robert_ancell: UDD (bzr branch ubuntu:d-conf), or just plain apt-get source/uupdate
[05:34] <pitti> robert_ancell: btw, GNOME 3.4 is being landed in experimental, too
[05:34] <pitti> robert_ancell: so some updates might be just syncs
[05:35] <pitti> robert_ancell: i. e. librsvg migth be; if not, I'll commit the missing bits to Debian's svn
[05:35] <pitti> robert_ancell: (but don't worry about librsvg for now, we can sync later)
[05:35] <robert_ancell> the stuff I updated was already -0ubuntu stuff, so unless the merges are done they won't be trivial syncs
[05:35] <pitti> just saying to avoid duplicate and diverging work
[05:35] <pitti> robert_ancell: -0ubuntu just means that we went ahead in upstream versions, but Debian is catching up now
[05:39] <rickspencer3> good morning pitti, robert_ancell
[05:39] <robert_ancell> rickspencer3, hey
[05:39] <pitti> hey rickspencer3
[05:40] <rickspencer3> pitti, robert_ancell any idea there was progress on the unity upgrader crasher during hte night?
[05:40] <pitti> rickspencer3: we have the workaround in place now
[05:40] <pitti> so the current images have it
[05:40] <rickspencer3> pitti, just the workaround? did they diagnose the bug?
[05:40] <pitti> just the workaround for now
[05:42] <rickspencer3> ok
[06:07] <ritz> jasoncwarner_, hey :)
[06:07] <jasoncwarner_> hey ritz I'm OTP right now...be there in a bit.
[06:26] <didrocks> good morning
[06:30] <pitti> bonjour didrocks, ca va?
[06:30] <didrocks> pitti: fatigué, mais ça va ;) et toi?
[06:31] <pitti> didrocks: je suis bien, merci
[06:31] <pitti> didrocks: Porquois ne pas dormez plus?
[06:32] <pitti> didrocks: thanks for handling the unity regression yesterday! *hug*
[06:32] <didrocks> pitti: all the recent fuzz about unity/compiz isn't good for sleeping long and well enough ;)
[06:32]  * didrocks hugs back
[06:32] <didrocks> I still don't get what changed in compiz for that
[06:32] <didrocks> so, in oneiric, we had two plugins
[06:32] <didrocks> bailer and detection
[06:32] <didrocks> they were added by default
[06:33] <didrocks> at the start of the cycle, we decided to remove them
[06:33] <didrocks> i removed them, but compiz is bad in dealing with configuration
[06:33] <didrocks> so did some tests, all went fine with those plugins still listed but not installed
[06:33] <didrocks> … 4 months later …
[06:34] <didrocks> "now compiz gives corruptions if a plugin it lists don't exist"
[06:34] <didrocks> (with the new compiz having *no change* in the configuration/plugin detection part)
[06:34] <didrocks> I guess we triggered a race…
[06:35] <pitti> "fun"... not
[06:35] <didrocks> rickspencer3: jibel: some summary of the black magic that happened ^
[06:35] <RAOF> Oh, that was the problem?  Yay!
[06:35] <didrocks> pitti: yeah, I guess some optimization that duflu did helped on the loading speed
[06:35] <didrocks> pitti: and we triggered this cases, on some machines
[06:36] <didrocks> RAOF: well, the workaround is really a "workaround"
[06:36] <didrocks> RAOF: right now, on upgrade, we revert once the plugin list (not the config, just the plugin list)
[06:36] <didrocks> as a stop gap mesure
[06:36] <rickspencer3> didrocks, I'm a bit dense, here
[06:37] <rickspencer3> the problem is that the plugins existed in a previous version, but not in this version
[06:37] <didrocks> rickspencer3: right, we don't want them anymore
[06:37] <rickspencer3> now, unity comes along sees the list of plugins, and pukes when one of the plugins is in the list, but not actually installed?
[06:37] <didrocks> so, at the start of the cycle, in november, I remove them
[06:38] <rickspencer3> didrocks, right, and you removed them from the list, not just from the disk, right?
[06:38] <didrocks> well, I see unity has nothing to do with it
[06:38] <rickspencer3> sorry, compiz, I mean
[06:38] <didrocks> rickspencer3: no, it's removed from the disk, but not from the list
[06:38] <didrocks> as compiz has bad way to deal with such cases
[06:38] <rickspencer3> didrocks, oh, in a fresh install they are still on the disk?
[06:38] <rickspencer3> dang it
[06:38] <rickspencer3> I mean
[06:38] <didrocks> no :p
[06:39] <rickspencer3> on a fresh install they are still on the list
[06:39] <rickspencer3> ?
[06:39] <didrocks> I removed from the *default list*
[06:39] <rickspencer3> right, so:
[06:39] <didrocks> it's just that, when you upgrade, compiz has no good way to update an already existing list
[06:39] <didrocks> (even if the list still correspond to the default)
[06:39] <rickspencer3> so:
[06:39] <rickspencer3> 1. we should not have compliz plugins, it causes a huge amount of work and bugs and has limited user value
[06:40] <rickspencer3> 2. in any case, compiz should be robust in face of a mismatch between list and actually available plugins
[06:40] <didrocks> +1000
[06:40] <didrocks> right, and it was…
[06:40] <didrocks> until yesterday
[06:40] <didrocks> well, until the last update, I mean
[06:40] <rickspencer3> didrocks, well, I guess it wasnt robust, we were just lucky
[06:40] <didrocks> as we did a bunch of upgrade test from oneiric -> precise
[06:41] <didrocks> yeah, I mean, something was making it working…
[06:41] <didrocks> I'm puzzled, apart from the recent duflu optimization which can just make the plugin loading faster, I see no change that can impact that
[06:41] <rickspencer3> didrocks, so, now that you know where the problem lies, how confident are you that the dx team can do a proper fix?
[06:41] <rickspencer3> also, great work!
[06:44] <didrocks> thanks ;)
[06:44] <didrocks> rickspencer3: duflu is looking at it
[06:44] <didrocks> we are getting closer
[07:25] <pitti> Sweetshark: good morning
[07:25] <pitti> Sweetshark: btw, if you feel like doing an upload, please target it to precise-proposed
[07:25] <pitti> Sweetshark: we can then build it there as a staging area without affecting precise, and move it into precise when ready and built everywhere
[07:26] <pitti> s/if/if and when/
[07:26]  * pitti needs to run out for 30 mins
[07:32] <Sweetshark> pitti: right. well, I do the upload to a ppa anyway for dist-upgrade testing
[07:41] <BigWhale> hrmpf, latest updates messed up my grub
[07:47] <seb128> hey
[07:49] <didrocks> salut seb128
[07:49] <seb128> lut didrocks, ca va ?
[07:49] <seb128> what are the dx news this morning? ;-)
[07:50] <seb128> I see robert_ancell was on an upgrade day ;-)
[07:50] <didrocks> seb128: nothing really changed since yesterday, my workaround is still valid. jibel will confirm it this morning
[07:50] <seb128> didrocks, ok, nobody confirmed if the workaround is enough or not yet then, thanks
[07:50] <didrocks> seb128: right, but from all evidences, it should
[07:51] <seb128> great
[07:51] <didrocks> (basically, we narrowed down the issue to "if the unknown plugin is loaded between composite and opengl"
[07:52] <didrocks> some people confirms on the bug that it's fixed
[07:52] <didrocks> with the reset upgrade
[07:53] <didrocks> (only starting to look at bug mails now, I had a lot of emails to write this morning…)
[08:16] <didrocks> seb128: you're bad!
[08:16] <didrocks> seb128: you didn't remind pitti about the reminder email!
[08:16] <seb128> pitti, hey, it's reminder day !
[08:17] <seb128> didrocks, well at the same time he didn't talk on the channel since I joined :p
[08:17] <seb128> I guess he has already been working for 5 hours and it's lunch time for him ;-)
[08:17] <didrocks> seb128: still, I know we are tired, it's the end of the cycle, but we need to keep it up! :p
[08:22] <pitti> oh, that, thanks!
[08:22] <pitti> I said good morning a bit earlier, but then got bitten by DSL reconnect
[08:22] <pitti> seb128: nah,  more like 3 h today :)
[08:22] <pitti> as long as my wife is at home, I don't get up at 6
[08:22] <pitti> 7 is fine
[08:23] <seb128> pitti, ;-)
[08:23]  * pitti says "bonjour" to seb128 again
[08:23] <seb128> pitti, good morning, how are you?
[08:23] <pitti> splendid, thanks
[08:23] <pitti> went to TKD again yesterday; my jaw is still hurting a little, but workout was nice
[08:25] <seb128> great
[08:26] <pitti> hm, unity-lens-video taking all my CPU and causing a lot of IO apparently
[08:26] <pitti> *zap*
[08:41] <seb128> pitti, grep VIDEO .config/user-dirs.dirs
[08:41] <seb128> pitti, (just curious)
[08:41] <pitti> XDG_VIDEOS_DIR="$HOME/"
[08:41] <seb128> it's supposed to "index" that dir
[08:41] <pitti> ah, is it indexing my entire home dir then?
[08:41] <seb128> pitti, I though they had a workaround for that buggy case but maybe it's broken
[08:41] <seb128> pitti, it will teach you to delete the xdg dirs :p
[08:42] <seb128> it's supprising how many people do that and how they screw their system :-(
[08:42] <pitti> well, most are just clutter
[08:42] <pitti> I don't need a "templates", and I don't keep videos on my system really
[08:42] <seb128> like if you delete templates it fallback to home and nautilus try to index the whole user dir to build its template submenu
[08:42]  * pitti fixes it to "templates"
[08:42] <pitti> err, "$HOME/download"
[08:42] <seb128> pitti, right, well our desktop breaks in funny way when you do that
[08:43] <seb128> pitti, it means softwares that should "index"  $music or $video or $template start indexing your whole disk
[08:43] <pitti> oh, I thought templates was only for LibO and that stuff
[08:43] <seb128> and that's an hard performances hit
[08:43] <pitti> ok, thanks
[08:43] <seb128> pitti, no, it's the nautilus "create new document" submenu
[08:43] <pitti> hm, why does that need to be a top-level dir?
[08:44] <pitti> this could be in .local/share/templates/ somewhere
[08:44] <didrocks> I remembered as well DavidCalle told he would fix that case
[08:44] <didrocks> not sure, ken is uploading the video lens, maybe he would know more
[08:44] <pitti> ok, fixed my XDG dirs
[08:45] <pitti> and yay for the lenses being in pythohn
[08:45] <pitti> it seems like every time we eliminate a python daemon we get two back *sigh*
[08:46] <seb128> pitti, because you want easy access to those dirs
[08:47] <seb128> but maybe we should do it the microsoft way
[08:47] <seb128> i.e My Documents/My {Images,Video,...}
[08:47] <didrocks> hum, I hate that way TBH
[08:47] <seb128> so users wouldn't remove "My Documents"
[08:47] <didrocks> a nested dir by default
[08:48] <didrocks> not really found of it :)
[08:48] <seb128> didrocks, well, the "flat" way doesn't work
[08:48] <seb128> see even pitti "screwed" it without seing the implications
[08:48] <seb128> users just think "I don't use that, it's fine to delete it"
[08:49] <didrocks> yeah, I know… but that won't prevent people removing the "My Documents"
[08:49] <pitti> right, especially for "templates" -- it doesn't really tell you what it is for, and I never used it
[08:49] <seb128> didrocks, it's less likely that they would remove 1 dir with all the interesting subdirs
[08:49] <didrocks> not sure of this move though, I can understand the intent, but I don't think it's the way forward
[08:50] <seb128> can we "lock" those dirs on the fs?
[08:50] <seb128> i.e forbid users to delete them :p
[08:50] <didrocks> like you want to remove it, and then, getting a dialog "no no no :p"
[08:50] <seb128> well we should at least get nautilus to warn you when you delete them
[08:50] <didrocks> agreed
[08:51] <didrocks> it's a first step
[08:51] <seb128> it will not be enough to avoid geeks to rm screw it though
[08:51] <seb128> one other way would be to change the fallback to not be ~
[08:51] <didrocks> indeed, but then, we can patch the apps to avoidd indexing if it's home?
[08:51] <didrocks> hum
[08:51] <seb128> right now if you delete a dir it goes to the dir under it
[08:51] <seb128> so any ~/xdg fallbacks to ~
[08:52] <seb128> didrocks, yeah, we can I guess but again you lead then in stuff "not working" without knowing why
[08:52] <didrocks> patch xdg to recreate teh dir?
[08:52] <didrocks> indeed
[08:52] <didrocks> the*
[08:52] <seb128> it's a tricky problem
[08:52] <didrocks> yeah, I have no good solution either TBH ;)
[08:54] <pitti> perhaps nautilus, video lens, etc. should just not do anything if that dir is $HOME, instead of trying to index everything
[08:54] <pitti> or more fundamentally, if a user removes an XDG dir, it shouldn't be set to $HOME in the first place
[08:55] <pitti> then video lens etc. woudl just see thaht the dir doesn't exist
[08:55] <pitti> they need to handle that case anyway
[08:57] <seb128> pitti, well that still breaks applications without telling the user why
[08:58] <didrocks> it's up to the app to tells it, but yeah, for nautilus and templates, there is no good way apart from a banner…
[08:58] <seb128> pitti, or we need to teach every application to "warn" about the issue
[08:58] <seb128> but I think it's a gsettings like situation
[08:58] <seb128> if you garanty the api will always return a valid dir you avoid tons of code in clients to all do the same checks
[08:59] <seb128> same story than gsettings, the fact to enforce rules simplify client code, no need to do tons of error checking
[09:15] <Chipaca> morning all... something seems broken wrt ssl in precise this lovely morning
[09:16] <Chipaca> getting “SSL routines:SSL23_GET_SERVER_HELLO:unsupported protocol” on sites that work in lucid (and worked in precise until recently)
[09:17] <pitti> Chipaca: we recently got a new openssl, presumably related to thaht
[09:19] <jibel> didrocks, I hit the "background only" issue after an upgrade from oneiric :/
[09:20] <didrocks> jibel: what's your plugin list?
[09:20] <jibel> didrocks, same machine, same base image, same upgrade but different result than yesterday. The session was corrupted but started.
[09:20] <jibel> didrocks, hold on, I'm saving everything
[09:20] <didrocks> jibel: you mean, yesterday, the session was corrupted but started, but now you have nothing displayed?
[09:21] <jibel> didrocks, right
[09:21] <didrocks> let me know about the plugin list
[09:21] <jibel> didrocks, yesterday I had #963093 and today it's #963633
[09:21] <didrocks> some people confirmed the upgrade worked (when they got the broken state :/) maybe the upgrade from oneiric is different
[09:22] <duflu> didrocks: I'm very busy. Is there something new?
[09:23] <didrocks> duflu: right, even with the workaround, jibel ends up in a blank desktop, no launcher, no panel, just the bakcground
[09:23] <didrocks> background*
[09:23] <didrocks> I'm asking him to see the plugin list
[09:24] <duflu> didrocks: That's not surprising. The bug is not fixed yet. Working on it.
[09:24] <didrocks> duflu: hum, didn't you say the workaround should work? as we reset the plugin ist
[09:24] <didrocks> lsit*
[09:24] <didrocks> so no more "unknown plugin"
[09:25] <Chipaca> pitti: https://bugs.launchpad.net/ubuntu/+bug/966058
[09:25] <ubot2> Launchpad bug 966058 in ubuntu "SSL routines:SSL23_GET_SERVER_HELLO:unsupported protocol" [Undecided,New]
[09:25] <duflu> didrocks: I did observe it to work when I tested it. However the root cause is still there so things may still be unpredictable
[09:28] <chrisccoulson> good morning everyone
[09:30] <didrocks> hey chrisccoulson
[09:30] <chrisccoulson> hi didrocks, how are you?
[09:30] <didrocks> chrisccoulson: well, still on those compiz/unity issues
[09:31] <didrocks> and you?
[09:31] <seb128> hey chrisccoulson, how are you?
[09:31] <chrisccoulson> didrocks, seb128, i'm good thanks :)
[09:32] <seb128> chrisccoulson, did you get at the bottom of that hud slowness issue?
[09:32] <tkamppeter_> pitti, hi
[09:32] <chrisccoulson> seb128, yeah, it's the same issue that google reported for chrome last year. we've always escaped it in firefox because we build the menus on demand
[09:32] <chrisccoulson> seb128, https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/964584/comments/17
[09:32] <ubot2> Launchpad bug 964584 in firefox "Firefox hangs temporarily everytimes it receives the focus" [Undecided,Incomplete]
[09:33] <chrisccoulson> not many bookmarks at all ;)
[09:33] <seb128> chrisccoulson, but you said some people have 15s hang with 50 bookmarks
[09:33] <seb128> which doesn't seem a crazy number
[09:34] <chrisccoulson> seb128, if they are buried deep in nested subfolders, then it doesn't surprise me at all now
[09:34] <pitti> hello tkamppeter_
[09:34] <chrisccoulson> the whole "lets just iterate over the tree of menuitems until we happen to stumble on the one we want" when handling events doesn't scale all that well :)
[09:35] <seb128> chrisccoulson, "realistic use case of 8000 bookmarks"
[09:35] <seb128> srsly?
[09:35] <seb128> do people user over 100 bookmarks?
[09:35] <jibel> didrocks, [core,composite,opengl,compiztoolbox,decor,vpswitch,snap,mousepoll,resize,place,move,wall,grid,regex,imgpng,session,gnomecompat,animation,fade,unitymtgrabhandles,workarounds,scale,expo,ezoom,unityshell]
[09:35] <tkamppeter> pitti, can you upload cups and cups-filters into Debian and Ubuntu? I have done several small fixes.
[09:35] <seb128> I use none but I always find stuff impossible to find with hundred of bookmarks...
[09:35] <tjaalton> use folders :)
[09:36] <chrisccoulson> seb128, i have over 100 bookmarks. perhaps 8000 is too much. but, still, 1 minute of CPU for getting 8000 bookmarks is pretty bad.....
[09:36] <pitti> tkamppeter: ubuntu is frozen right now; so perhaps on Thursday evening? we might get another fix by then
[09:36] <didrocks> duflu: so jibel list is fine, seems there is still an issue, then… ^
[09:36] <jibel> didrocks, logout/login "fixes" the problem and next login works correctly
[09:36] <seb128> chrisccoulson, I don't understand why that didn't happen previous cycles with menus though
[09:36] <didrocks> jibel: ok, at least, it's a start, but the underlying issue is still racy then
[09:36] <chrisccoulson> seb128, we build menus on demand, so you take a hit only when opening a subfolder
[09:37] <chrisccoulson> this is why we build menus on demand :)
[09:37] <seb128> chrisccoulson, can't you cache menus?
[09:37] <duflu> didrocks: I know. I'm working on it.
[09:37] <didrocks> jibel: I guess that's acceptable for a beta2 and we should maybe document that on the beta2 wiki annoucement?
[09:37] <didrocks> duflu: just to ensre, thanks again ;)
[09:38] <chrisccoulson> seb128, dbusmenu probably could. the issue is that when the application gets an event targetted at a specific menuitem, dbusmenu iterates over the whole tree of menuitems to find the right node
[09:38] <seb128> chrisccoulson, is there a better way to find an item?
[09:38] <xclaesse> gnome-documents does not start on precise: ** (gnome-documents:27601): WARNING **: Failed to load shared library 'libgdprivate-1.0.so' referenced by the typelib: libevdocument3.so.3
[09:38] <xclaesse> missing dep I guess?
[09:39] <seb128> xclaesse, likely, talk to jbicha when he's online
[09:39] <zzecool> https://bugs.launchpad.net/ubuntu/+source/unity/+bug/966068
[09:39] <ubot2> Launchpad bug 966068 in unity "unity 3d crashes and fallback to 2d ( unhandled ConfigureNotify ) unity.glib-gobject <unknown>:0 invalid cast from `BamfWindow' to `BamfApplication'" [Undecided,New]
[09:39] <seb128> zzecool, that's not a dump your bug channel
[09:39] <zzecool> seb128:  ?
[09:39] <seb128> zzecool, try #ubuntu-unity for unity issues and maybe say hi when joining or something
[09:40] <seb128> zzecool, well, it's impolite the join and paste a bug url
[09:40] <didrocks> jibel: for the record, what kind of machine is it?
[09:40] <didrocks> jibel: rather slow or fast CPU?
[09:40] <chrisccoulson> seb128, i'm sure there are better ways to search for nodes other then iterating over lists. perhaps a binary tree?
[09:40] <seb128> zzecool, we are not there to be dumped bug numbers like that without any text, hi, or anything
[09:40] <tkamppeter> pitti, OK, it is all committed to the BZRs, it is not urgent, cups-filters is a fix for printing photos from iOS clients (AirPrint) and cups is bug 965124.
[09:40] <ubot2> Launchpad bug 965124 in cups "printers no longer found after re-connecting network" [Undecided,Fix committed] https://launchpad.net/bugs/965124
[09:40] <chrisccoulson> of course, changing that would probably require redesigning dbusmenu
[09:40] <zzecool> seb128: most of the times noone says hi back so i just paste the bug so maybe someone will notice that
[09:41] <zzecool> its seems that there is activity right now
[09:41] <xclaesse> gnome-boxes is not packaged (yet) ?
[09:41] <seb128> zzecool, right, too much activity maybe, try #ubuntu-unity for unity issues
[09:41] <zzecool> seb128: chill out i got the msg
[09:41] <zzecool> ok thanks
[09:41] <zzecool> :)
[09:41] <seb128> xclaesse, it depends on a libvirt version which is new and buggy
[09:41] <seb128> too new
[09:41] <xclaesse> ok
[09:41] <seb128> xclaesse, it's only in Debian experimental and has known regressions
[09:41] <tkamppeter> pitti, what would be the other fix which you mention?
[09:42] <chrisccoulson> seb128, the issue is this: http://bazaar.launchpad.net/~dbusmenu-team/dbusmenu/trunk.0.6/view/head:/libdbusmenu-glib/server.c#L1609
[09:42] <chrisccoulson> and http://bazaar.launchpad.net/~dbusmenu-team/dbusmenu/trunk.0.6/view/head:/libdbusmenu-glib/menuitem.c#L935
[09:42] <xclaesse> seb128, ok so unlikely to impossible for a LTS release I guess
[09:42] <seb128> chrisccoulson, I doubt we will get those fixed for precise though :-(
[09:42] <chrisccoulson> this is basically where all of the CPU time goes
[09:42] <pitti> tkamppeter: nothing in particular, just that a lot can happen in two days
[09:42] <seb128> xclaesse, right, I doubt that boxes is lts quality after its first cycle anyway
[09:42] <chrisccoulson> seb128, if we can't fix it, then i'll have to do what google did for chrome and remove the bookmarks menu :(
[09:42] <seb128> xclaesse, it will be in the GNOME3 ppa for those who like to play with buggy cracks ;-)
[09:43] <chrisccoulson> it's pretty much unusable for some people otherwise
[09:43] <seb128> chrisccoulson, can you disable bookmark export if the number of bookmark is > 100 or something?
[09:43] <tkamppeter> pitti, tu a lu bug 962226?
[09:43] <ubot2> Launchpad bug 962226 in cups "package cups 1.5.2-8bzr1 failed to install/upgrade: ErrorMessage: subprocess installed post-installation script returned error exit status 1" [Undecided,Incomplete] https://launchpad.net/bugs/962226
[09:43] <xclaesse> seb128, btw I was wondering if it is possible to add a ppa and only pick some of its packages (and its deps) but not everything. A bit like experimental works on debian?
[09:43] <Sweetshark> zzecool: seb128 is taking it cool, I remember the trainwreck when someone dared to dump a launchpad-bug uncommented upstream at #libreoffice and _rene_ was around. Poor chap didnt stand a chance.
[09:44] <chrisccoulson> seb128, i'm not sure. it would probably be easier to just disable it and have the bookmark button on the navigation bar instead
[09:44] <pitti> tkamppeter: need to put the latest responses through google translate :)
[09:44] <chrisccoulson> i think people will be confused if some people have bookmarks in the menu, and others have them on the navigation bar
[09:44] <jibel> didrocks, low-end laptop core i3 with an integrated graphics chipset
[09:44] <seb128> chrisccoulson, that would suck, it would break the hud for 90% of users who are not going crazy with bookmarks just because 10% users have years of cruft
[09:44] <zzecool> Sweetshark: i think we all got the msg allready
[09:45] <didrocks> jibel: ok thanks
[09:45] <Adri2000> if an application has two icons available: /usr/share/pixmaps/$APP.{png,svg}, which one will be used? is there a policy about this or does it depend on the implementation? (if so, what does unity?)
[09:45] <chrisccoulson> seb128, right, but it doesn't take many bookmarks to have a few seconds of CPU use already, which is still pretty bad
[09:47] <seb128> Adri2000, not sure, read the freedesktop icon spec?
[09:48] <seb128> Adri2000, http://standards.freedesktop.org/icon-theme-spec/icon-theme-spec-latest.html
[09:49] <seb128> Adri2000, the pseudo code in there seems to suggest it prefers png
[09:50] <Adri2000> right I should have looked at that page, reading now thanks
[09:55] <Adri2000> ok, it says svg support is optional, so anyway I have to keep shipping both png and svg
[09:55] <Adri2000> and choosing the best one is desktop environment's responsability :)
[10:18] <slomo__> seb128: you don't want to get cairo 1.12.0 from debian/experimental btw, it causes interesting problems due to a EXA bug
[10:19] <slomo__> seb128: do you know if something happened to the gtk scrolling stuff?
[10:19] <seb128> slomo__, thanks, I didn't plan to go for a new cairo 2 months after feature freeze anyway
[10:19] <seb128> slomo__, not a lot happened no, it's still slow, buggy and breaks applications and compat
[10:22] <slomo__> seb128: is there a plan how this will be handled upstream? i remember seeing a gtk 3.4.0 tarball... which imho shouldn't have happened before there was at least a plan how to handle this. breaking random applications is not something a library/framework should ever do...
[10:22] <seb128> slomo__, no, they seem to consider that gtk3 is not used out of GNOME yet so it's ok to break compat
[10:23] <seb128> I don't agree with that though
[10:23] <seb128> but that seems to be the status quo
[10:23] <slomo__> i don't agree either, that's simply wrong... and i remember seeing ebassi complaining about this attitude too
[10:41] <savvas> hi, can you please review libgxps MIR request? https://bugs.launchpad.net/ubuntu/+source/libgxps/+bug/965467
[10:41] <ubot2> Launchpad bug 965467 in libgxps "[MIR] Please transfer libgxps 0.2.2-1 (universe) to main" [Undecided,Confirmed]
[10:52] <pitti> seb128: FYI, you uploaded nautilus twice, rejecting one (so don't get confused)
[10:53] <seb128> pitti, oh, thanks, I got confused with all the ping crazyness earlier, I might have done that yes
[10:53] <seb128> pitti, thanks ;-)
[11:30] <jbicha> good morning!
[11:30] <jbicha> any ideas how to fix the gnome:NextVersion thing? it's trying to depend on libmutter0 (>= 3.3),  libmutter0 (<< 3.4)
[11:30] <jbicha> does gnome-shell just need a rebuild or what?
[11:32] <didrocks> testing g-s-d, brb
[11:38] <nessita> hello everyone!
[11:42] <BigWhale> Did anyone think about simple calculator in dash?
[11:43] <BigWhale> this would be awesome
[11:44] <slomo__> BigWhale: if you just need a commandline calculator... nickle is nice (and has also many advanced features)
[11:44] <BigWhale> slomo__, I use bc mostly, but calculator in dash would be awesome... just press super and then 1233+731
[11:45] <BigWhale> and the result is there
[11:45] <BigWhale> hit super again and your working what you were woking before
[11:49] <seb128> nessita, hey, how are you?
[11:50] <nessita> seb128: pretty good, you? :-)
[11:50] <nessita> congrats on the beta-2 btw
[11:50] <seb128> BigWhale, there is a unity-lens-utilities and a unity-scope-calculator
[11:50] <seb128> see http://askubuntu.com/questions/38772/what-lenses-for-unity-are-available
[11:51] <seb128> http://i.stack.imgur.com/zOrC3.png
[11:51] <seb128> nessita, I'm good thanks
[11:51] <seb128> nessita, thanks, the congrats will be for when it will be out though, i.e in 2 days ;-)
[11:51] <nessita> we're almost there ;-)
[11:52] <BigWhale> seb128, yeah I just found it....
[12:40] <trkemist> Anyone here familiar with Cisco VPN
[13:01] <BigWhale> I just found something disturbing. https://bugs.launchpad.net/ubuntu/+source/gnome-desktop3/+bug/951479
[13:01] <ubot2> Launchpad bug 951479 in gnome-desktop3 "Laptop with nvidia binary drivers does not suspend when lid is closed" [Undecided,Confirmed]
[13:02] <BigWhale> Will this be fixed somehow for 12.04?
[13:02] <BigWhale> I know it is essentially nvidia bug, but this will hit most of the users that use this driver.
[13:07] <seb128> BigWhale, if you provide a patch maybe it will be fixed?
[13:08] <seb128> but I personally don't own any nvidia hardware to work on it and nobody else seems interested so far
[13:09] <BigWhale> seb128, hah... I just asked this on ubuntu-desktop list
[13:09] <BigWhale> if someone provides a patch, will it get included....
[13:09] <jbicha> seb128: good morning, do you know anything about that gnome:NextVersion thing?
[13:09] <BigWhale> :))
[13:10] <seb128> BigWhale, no, we hate fixes and like buggy softwares, we will refuse any patch ;-)
[13:10] <BigWhale> seb128, now you're talking like someone who works for Gnome. ;>
[13:10] <seb128> jbicha, hey, it's basically computing the next serie
[13:10] <seb128> jbicha, i.e on 3.4 tarballs it would give 3.6
[13:11] <seb128> BigWhale, ;-)
[13:11] <BigWhale> seb128, the problem is that I am not sure how to address this correctly. There is a working patch for gnome-setting-daemon, but it will disable the whole 'stay active if there's an external monitor present' functionality.
[13:11] <seb128> jbicha, DEB_GNOME_NEXTVERSION := $(shell echo $(DEB_GNOME_VERSION) | awk -F. '{ printf("%s.%i\n",$$1,$$2+1) }')
[13:12] <seb128> jbicha, that's in gnome-pkg-tools: /usr/share/gnome-pkg-tools/1/rules/gnome-version.mk
[13:12] <BigWhale> seb128, the best solution would be a checkbox somewhere in the power settings
[13:12] <seb128> BigWhale, no checkbox no
[13:12] <BigWhale> but at this stage
[13:12] <seb128> no ui change
[13:12] <BigWhale> exactly
[13:12] <seb128> BigWhale, is the issue is that nvidia is crappy enough that they can't detect if a monitor is plugged?
[13:13] <BigWhale> seb128, it thinks that monitor is always plugged in
[13:14] <seb128> BigWhale, driver bug?
[13:14] <BigWhale> seb128, and it is quite annoying to pull your notebook from the bag with fans spinning at full speed
[13:14] <mdeslaur> BigWhale: the only sane work-around is to detect if it's using the proprietary nvidia driver and suspend anyway
[13:14] <mdeslaur> but even that will bug people who _are_ using external monitors with nvidia
[13:14] <seb128> or to display a big warning, if your laptop melt blame nvidia
[13:15] <mdeslaur> seb128: lol :)
[13:15] <seb128> ;-)
[13:15] <BigWhale> mdeslaur, I know.. the whole situation is really crappy and no easy solution
[13:15] <jbicha> seb128: will rebuilding gnome-shell fix it trying to depend on mutter 3.3 << mutter 3.4 ?
[13:15]  * mdeslaur shakes fist at binary driver
[13:16] <BigWhale> seb128, this is a related upstream bug https://bugzilla.gnome.org/show_bug.cgi?id=672030
[13:16] <ubot2> Gnome bug 672030 in libgnome-desktop "gnome-settings-daemon always prevents suspend on macbookpro5,3" [Normal,Resolved: notgnome]
[13:17] <seb128> BigWhale, well I've no good reply, I guess we could make nvidia laptops always suspend on lid close even if external monitors are plugged in
[13:18] <BigWhale> seb128, could this be passed as some config option somewhere, anywhere, that can be later removed when nvidia fixes the driver?
[13:18] <desrt> didrocks: hi
[13:18] <didrocks> hey desrt
[13:18] <seb128> BigWhale, not that I know about
[13:18] <desrt> how does the hud work in unity2d? :)
[13:18] <didrocks> desrt: hum, what do you mean by "how"? ;)
[13:19] <mdeslaur> BigWhale: if they fix the driver, and we manage to get it in Precise, we can just remove the patch as a SRU
[13:19] <didrocks> for the alt detection tap?
[13:19] <desrt> so i downloaded unity-2d-5.8.0
[13:19] <didrocks> or for something else? ;)
[13:19] <desrt> and i'm looking for 'StartQuery' anywhere in the source
[13:19] <desrt> i'm not finding it
[13:19] <BigWhale> seb128, I am sure that more people will notice this. telling them to use neuveau (or whatever the spelling) is not really an option :/
[13:19] <desrt> so i guess i mean: how does it communicate with the service?
[13:20] <desrt> in fact, i find the opposite here: what appears to _itself_ be a dbus service for making hud queries
[13:20] <seb128> BigWhale, yeah, I share the sentiment, still I've no good reply, nor nvidia hardware to work on it
[13:20] <didrocks> desrt: the StartQuery is in libunitycore
[13:20] <didrocks> desrt: which is a shared component between 2d and 3d
[13:21] <desrt> didrocks: which lives in unity's source module
[13:21] <didrocks> right
[13:21] <desrt> gotcha.  thanks
[13:21] <didrocks> yw ;)
[13:21] <desrt> engineering!
[13:21] <didrocks> desrt: see, sharing code, amazing! :)
[13:24] <BigWhale> seb128, so if I provide a patch by the end of the week?
[13:24] <seb128> BigWhale, we will review it
[13:25] <BigWhale> seb128, a patch that would force suspend for nvidia cards to enter suspend even if external monitors are present
[13:25] <BigWhale> force suspend to enter suspend... good one Whale..
[13:25] <BigWhale> seb128, ok... excellent
[13:26] <seb128> BigWhale, comment on bug #949296 maybe if you work on it
[13:26] <ubot2> Launchpad bug 949296 in gnome-settings-daemon "gnome-settings-daemon always prevents suspend when proprietary nvidia drivers are used." [Medium,In progress] https://launchpad.net/bugs/949296
[13:26] <seb128> pitti, there?
[13:26]  * pitti waves to seb128
[13:26] <seb128> line 349, in duplicate_db_publish
[13:26] <seb128>     os.mkdir(out)
[13:26] <seb128> OSError: [Errno 17] File exists: 'apport-duplicates.new'
[13:26] <seb128> pitti, did you see that before?
[13:26] <seb128> i386 retracers stopped on that
[13:26] <pitti> seb128: it's on my list
[13:26] <seb128> pitti, ok
[13:27] <seb128> pitti, thanks
[13:27] <pitti> seb128: apparetnly you restarted it after the previous crash
[13:27] <pitti> seb128: it crashed becauase a signature was too long for a file name
[13:27] <BigWhale> seb128, ok noted.
[13:27] <pitti> seb128: I need to fix that, otherwise it'll just crash again
[13:27] <seb128> pitti, yeah, restarting seems to work, he went through over 100 bugs backlog yesterday
[13:27] <seb128> pitti, ok
[13:27] <pitti> seb128: ah, so feel free to just nuke the dupe dir then, and restart
[13:28] <seb128> pitti, oh, ok, thanks
[13:38] <desrt> chrisccoulson: good morning
[13:39] <desrt> how does firefox find itself this morning? :)
[13:39] <chrisccoulson> hi desrt, how are you?
[13:40] <desrt> okay
[13:40] <desrt> closing lots of hud bugs saying "not my fault!"
[13:40] <desrt> ;)
[13:40] <chrisccoulson> heh :)
[13:40] <chrisccoulson> those are the types of bugs that i like!
[13:41] <desrt> so have you managed to get dbusmenu's appetite for CPU cycles under control?
[13:43] <chrisccoulson> desrt, i'm looking at that now :)
[13:44] <desrt> chrisccoulson: cool.  did you see what i mentioned about menuitems?
[13:44] <chrisccoulson> desrt, i don't think so
[13:45] <chrisccoulson> what was that?
[13:45] <desrt> i'm not sure if it was a case of you already fixed it and there hasn't been a release yet..
[13:45] <desrt> when creating a new submenu menuitem it first gets created as a normal item, and then changes properties to become a submenu after-the-fact
[13:46] <desrt> my code was kinda unhappy about that but i changed it to deal with it.... but it's some extra dbus traffic you could probably trim down on
[13:48] <chrisccoulson> ah, ok. yeah, i have a fix for one thing which might cause that
[13:52] <didrocks> "1000 characters if enough for everyone"
[15:24] <jbicha_> seb128: looks like a rebuild does fix the NextVersion problem
[15:25] <seb128> jbicha_, good
[16:14] <kenvandine> weee... compiz crash
[16:55] <Sweetshark> how long does it usually take for a upload to show up in a ppa? I first uploaded to an nonexistant ppa (missing one char from the ppa-name during copypaste) and then had to forceupload again to the real ppa. However, I dont see anything in the ppa, nor did I get any nagging mails from launchpad yet.
[16:55] <Sweetshark> heh
[16:55] <Sweetshark> ok, it just showed up.
[16:55] <kenvandine> Sweetshark, i've had it take 20m before
[16:56] <Sweetshark> kenvandine: thanks a lot!
[16:56] <rickspencer3> seb128, desrt, anybody, what's the package that the settings apps comes in?
[16:58] <kenvandine> rickspencer3, which settings?
[16:58] <kenvandine> the control center?
[16:58] <rickspencer3> kenvandine, never mind, it's gnome-control-setting
[16:58] <kenvandine> gnome-control-center, but some of the panels embedded in it come from other packages
[17:04] <Sweetshark> bryceh: https://launchpad.net/~bjoern-michaelsen/+archive/libreoffice-precisetest-20120327/+packages has a libreoffice package which should include the /tmp fixing. (I wouldnt usually ppa that, but need to for debugging bug 916291 anyway).
[17:04] <ubot2> Launchpad bug 916291 in libreoffice "failed to upgrade from Oneiric to Precise: ERROR: Cannot determine language! - exit status 134" [Critical,In progress] https://launchpad.net/bugs/916291
[17:05] <bryceh> Sweetshark, thanks, I'll try it out
[17:12] <bryceh> seb128, hey, one of the inkscape guys is asking about if there's any chance for cairo 1.12 in precise, as there's some nice performance improvements and mesh gradient support.  I told them with FF passed it couldn't be included now, but promised to doublecheck that with you.
[17:20] <didrocks> ok, time for some exercice and dinner, see you guys!
[18:59] <seb128> cyphermox, hey
[19:00] <pitti> good night everyone!
[19:00] <seb128> pitti, 'night
[19:00] <seb128> pitti, you only worked 15 hours today, what's going on with you?! ;-)
[19:09] <seb128> bryceh, yeah, no plan to ship cairo 1.12, it's like a 1.5 years of work and non trivial changes, not something to try to sneak in between beta2 and hard freeze
[19:09] <seb128> bryceh, not to mention that slomo (the Debian maintainer) warned me that he uploaded to experimental and there are some exa regressions
[19:10] <bryceh> seb128, thanks.  ok, thought so.
[19:25] <cyphermox> seb128: hey
[19:25] <seb128> cyphermox, hey, how are you?
[19:26] <cyphermox> doing alright, how about you?
[19:29] <seb128> cyphermox, I'm good thanks
[19:29] <seb128> cyphermox, how busy are you?
[19:29] <cyphermox> I have some free time I was fighting evo now
[19:30] <seb128> cyphermox, feel free to say you have enough to do ;-) I'm trying to find somebody to sign to update the vino appindicator patch which got commented in the GNOME3 update in oneiric because it needed to be updated to the upstream changes
[19:30] <seb128> it might turn out to be easy but it's on my lts bugslist
[19:30] <cyphermox> vino appindicator?
[19:30] <seb128> it's a somewhat security issue
[19:30] <cyphermox> yeah
[19:30] <seb128> it means the "show an icon when somebody connects to my box" doesn't show
[19:31] <cyphermox> ah, I see
[19:31] <seb128> cyphermox, vino has a gtkstatusicon, a dx contractor did the port to an indicator by then, rodrigo commented the patch when updating to GNOME3 because he wanted to update and it was not trivial
[19:31] <seb128> cyphermox, nobody picked it up sicne
[19:31] <cyphermox> I have some experience with appindicator anyway  :P
[19:31] <seb128> cyphermox, so it's on the "fix for the lts" list and I'm trying to find somebody to sign for it ;-)
[19:32] <cyphermox> sure,I'll take a look
[19:32] <seb128> cyphermox, it's still in the package,vcs but commented in the serie
[19:32] <cyphermox> cool, that should be easy
[19:32] <seb128> cyphermox, https://bugs.launchpad.net/ubuntu/precise/+source/vino/+bug/884003
[19:32] <ubot2> Launchpad bug 884003 in vino "The remote connection indicator is broken" [High,In progress]
[19:32] <seb128> cyphermox, "it's only code" right? ;-)
[19:32] <seb128> cyphermox, thanks!
[19:32] <seb128> cyphermox, feel free to assign the bug to yourself then ;-)
[19:33] <cyphermox> aye.
[19:36] <cyphermox> heh, it's one huge patch :)
[19:36] <seb128> cyphermox, yeah, if it was that trivial it would have been done at the time :p
[19:37] <cyphermox> well, it does kind of apply
[19:37] <seb128> cyphermox, the contractor who did it was not great so maybe it can be simplified a lot, but if you can refresh it easily maybe just do that
[19:38] <cyphermox> yeah that's where I'll start
[19:38] <cyphermox> it's hard to do both at the same time anyway :)
[19:51] <cyphermox> seb128: I see, there /is/ a lot of code duplication there
[19:51] <seb128> cyphermox, indeed
[19:51] <seb128> it's a bit monkey copy
[19:52] <cyphermox> I guess. I'm almost considering using the same kind of fallback trick I used for nm-applet to avoid replacing bunches of code with ifdefs, and just add appindicator support alongside statusicon
[23:30] <Laney> where should I push a branch to for merging into an ubuntu-desktop branch?
[23:30] <Laney> do LP MPs work for these?
[23:33] <micahg> Laney: yeah, you would push under you and propose a merge
[23:33] <Laney> i still don't really understand how launchpad's branch namespacing works
[23:33] <Laney> because these are ~ubuntu-desktop/foopkg/ubuntu branches
[23:34] <micahg> Laney: right, so you push under ~laney/foopkg/fix-foo
[23:34] <Laney> doesn't that stack on lp:ubuntu/foopkg?
[23:35] <micahg> Laney: no, that would be ~/laney/ubuntu/precise/foo/fix-foo
[23:35] <Laney> indeed
[23:35] <Laney> it stacked on the upstream branch instead :P
[23:35] <micahg> and stacking is based on branching in any event, not pushing
[23:35] <Laney> Stacked on: lp:gnome-settings-daemon
[23:37] <micahg> right, which wouldn't be right
[23:37] <Laney> indeed
[23:37] <micahg> well, I shouldn't say that :)
[23:37] <Laney> dunno why it did it
[23:37] <Laney> but it let me do the merge proposal anyway
[23:37] <micahg> yeah, LP's not finicky about that, you just get a very big diff
[23:38] <RAOF> Stacking is entirely a performance optimisation; it doesn't change anything you can touch about the branch.
[23:38] <Laney> o brave new world
[23:40] <RAOF> If you've branched of ~ubuntu-desktop/foopkg/ubuntu, you push to ~laney/foopkg/ubuntu, and then you request a merge into ~ubuntu-desktop/foopkg/ubuntu, that merge will contain only the diff between ~laney and ~ubuntu-desktop, regardless of what ~laney happens to be stacked on (or even if ~laney isn't stacked on anything).
[23:41] <Laney> Yep, I can see that the generated diff is correct
[23:41] <Laney> it just feels like an incorrect use of namespace
[23:42] <Laney> would it have worked if there wasn't a gsd upstream import?
[23:43] <Laney> never mind, it is late. goodnight :-)
[23:44] <RAOF> It depends on what the default branch is; it probably would have, yes.