[02:47] <chris__> i installed ubuntu 15.04 yesterday i had no issues shutting down and restarting so forth and today when i boot it up i get usb enumurate error -71 and i cant get to the desktop i know this is the wrong channel but i was in the ubuntu channel and i was waiting for almost 40 min
[04:48] <pitti> Good morning
[05:28] <didrocks> good morning
[05:29] <didrocks> back in a sec
[05:45]  * didrocks retries a reboot after upgrade
[06:50] <didrocks> pitti: that's really weird!
[06:50] <didrocks> I tried to downgrade as much things as possible - no change
[06:50] <pitti> didrocks: wb!
[06:50] <didrocks> (but gnome shell is working fine with multi-screen)
[06:50] <didrocks> and lightdm
[06:50] <didrocks> just not unity on "live change"
[06:50] <didrocks> so I need to set up the monitor xml file
[06:50] <didrocks> and restart the session
[06:51] <didrocks> to not have artefacts
[06:52] <RAOF> didrocks: Do your displays flash?
[06:52] <didrocks> yep
[06:52] <RAOF> Change workspace :)
[06:52] <didrocks> and not a specially pleasant FX :p
[06:52] <didrocks> seriously?
[06:52] <didrocks> don't tell me I tried to downgrade the world for this :p
[06:52] <RAOF> Yeah, it gets compiz to refresh whatever is stupid.
[06:52]  * didrocks tries to change settings
[06:52]  * RAOF thought this was a DRI3 problem for a while, but it isn't.
[06:53] <didrocks> RAOF: zomg!
[06:53] <didrocks> you're right
[06:53] <didrocks> I really wonder why I never had that before today thogh
[06:53] <didrocks> though*
[06:53] <didrocks> as I plug/unplug my laptop quite regularly
[06:53] <RAOF> I've had it for a while. I don't *think* I was running anything particularly out-of-archive?
[06:54] <didrocks> can be yesterday's xorg fix for my crashers which then trigger it for me
[06:54] <didrocks> RAOF: thanks for the tip! I saw that the bug was reported and on the hot list!
[06:57] <larsu> good morning!
[06:57] <larsu> fuck yeah sun!
[06:58] <didrocks> hey larsu
[06:58] <didrocks> sun never left here :p
[06:59] <larsu> yeah, south of france has longer summers than we do ;)
[07:00]  * larsu stops trolling and wishes didrocks a good sunny morning
[07:01] <pitti> hey larsu
[07:01] <larsu> morning pitti!
[07:01] <pitti> larsu: we already had sun yesterday, which was quite a relief indeed :)
[07:02] <larsu> pitti: we did a bit as well, but not as nice as right now
[07:19] <willcooke> Morning
[07:19] <willcooke> Trevinho, is this something you can look at?  https://bugs.launchpad.net/ubuntu/+source/libunity/+bug/1491542
[07:20] <larsu> morning willcooke
[07:20] <larsu> ya, airbnb, I'd like to go to the same cities I've recently been to...
[07:21]  * larsu wonders if he can turn off that newsletter
[07:21] <didrocks> hey willcooke
[07:38] <larsu> pitti: how do I kick off a new CI run for dbus-test-runner? Laney's not around yet and I'd like to know if my midnight hack was worth it
[07:39] <pitti> larsu: I can do it -- is it in -proposed already?
[07:40] <larsu> pitti: no, it's on a branch. I don't have archive-chops
[07:40] <larsu> lp:~larsu/dbus-test-runner/fix-test-race
[07:40] <pitti> larsu: we can't run the production autopkgtests against branches; the near-time plan is to add PPA support, though
[07:41] <larsu> pitti: ah that would be nice. I'll just wait for Laney to upload to proposed,then
[07:41] <pitti> larsu: but you can of course run the test locally, or you toss me a .dsc and I build it on ppc64el and run the tests on the production machines with that
[07:41] <larsu> pitti: I already tested on a porter
[07:42] <larsu> I'm pretty sure it's fixed, just want to see the green thing ;)
[07:43] <pitti> larsu: ah, nice!
[08:02] <Laney> hi
[08:03] <larsu> morning Laney!
[08:03] <larsu> how are you?
[08:03] <didrocks> hey larsu!
[08:03] <didrocks> argh
[08:04] <didrocks> larsu: stop talking!
[08:04] <didrocks> hey Laney ;)
[08:04] <larsu> didrocks: does your irc client do tab completion based on recency?
[08:04] <larsu> that would be pretty cool...
[08:04] <pitti> hey Laney, how are you?
[08:04] <pitti> larsu: yes, it does
[08:05] <didrocks> larsu: yep
[08:05]  * larsu should switch
[08:05] <larsu> or get work done..
[08:11] <Laney> didrocks: larsu: good, thanks!
[08:11] <Laney> played MGS and went to the pub quiz last night
[08:11] <Laney> I almost got wiped out on the way there though...
[08:11] <Laney> but STILL ALIVE
[08:11] <larsu> oh what happened?
[08:12] <Laney> someone drove a car at speed without lights out of some side street
[08:12] <Laney> while i was crossing in front of it...
[08:12] <Laney> burning rubber and all that
[08:13] <pitti> urgh!
[08:13] <pitti> weren't you all shaking after that?
[08:13] <Laney> I suspect they were doing some crime
[08:13] <Laney> yes
[08:13] <willcooke> Sounds like you do need a helmet cam
[08:13] <Laney> especially when they turned around in the street and some bloke got out ...
[08:13] <larsu> woah. Glad you're ok!
[08:13] <Laney> but he just ran off
[08:14] <pitti> Laney: glad that nothing serious happened to you!
[08:14] <pitti> gosh -- cars and cities just don't fit together
[08:18] <Laney> :)
[08:19] <Laney> larsu: didn't dezrt talk about a way to fix this race in dbus-test-runner itself too?
[08:41] <larsu> Laney: yes and I started doing that, but it really doesn't work with the way tasks work in there
[08:41] <larsu> they're always started in parallel
[08:41] <larsu> and it relies on that
[08:41] <larsu> we could add something like --wait-for-other-task-to-be-initialized
[08:41] <larsu> (and in fact, I started working on a patch that did that)
[08:42] <larsu> (but it became unwieldy quickly - so I decided on the saner solution last night)
[08:42] <pitti> didrocks: NB that pandora-build has a binary dependency on quickly-ubuntu-template; otherwise there are no rdepends left, we could remove quickly and quickly-lens-templates
[08:43] <didrocks> pitti: interesting about pandora-build, I need to look at why (there is no reason to dep on a template)
[08:43] <Laney> larsu: isn't it making it do all of the watching before starting the tasks?
[08:43] <didrocks> pitti: I'll have a look later this week, finishing up some snappy-related things
[08:44] <Laney> that already breaks things?
[08:44] <larsu> Laney: no, it waits only for that *one* tasks
[08:44] <pitti> didrocks: it's also dep'ing on quickly itself
[08:44] <larsu> *task
[08:44] <larsu> all other are started immediately
[08:44] <didrocks> pitti: I'll have a look
[08:44] <larsu> and obviously we can't add a --wait for the task that owns the name
[08:44] <pitti> didrocks: merci
[08:45] <Laney> but you can say don't start any tasks until I've processed all of the wait-fors
[08:45] <larsu> yes, I thought about this as well
[08:45] <larsu> good luck doing that if you're up for it
[08:45] <Laney> HAHA
[08:45] <didrocks> de rien ;)
[08:45]  * Laney goes to sit in the corner
[08:45] <larsu> Laney: it's possible, but requires major refactor
[08:46] <larsu> Laney: if you're entirely unhappy with the current solution, we could go for that
[08:46] <Laney> I guess this doesn't happen too much in reality
[08:47] <larsu> I agree that this patch is a bit fix-the-race-with-a-timeout-y
[08:52] <Laney> let's get tedg to have a look later & we can upload it today
[08:52] <Laney> thanks for working on it!
[08:52] <larsu> sure
[08:53] <larsu> thanks for helping me get on porter ;)
[08:53] <Laney> np
[08:54] <Laney> it comes in helpful sometimes
[08:54] <larsu> indeed
[09:36] <hikiko> hello :)
[09:37] <didrocks> hey hikiko!
[09:37] <darkxst> hey didrocks
[09:38] <hikiko> hi didrocks
[09:38] <hikiko> je suis a Paris!
[09:38] <hikiko> a CDG :p
[09:39] <didrocks> evening darkxst :)
[09:39] <didrocks> hikiko: oh, hoping that your connection will not make you stay there too long :)
[09:39] <hikiko> no :) the flight will be on time
[09:39] <hikiko> I just arrived early
[09:40] <darkxst> didrocks, had an interview today, hopefully I did ok
[09:42] <didrocks> darkxst: oh, good luck! :)
[09:43] <darkxst> and of course steve handballed me a essentially 'NEW' review, first ever and under interview
[09:45] <didrocks> waow, directly a NEW review, I hope at least it wasn't anything like Qt :p
[09:47] <darkxst> well no, and apparently the package is actaully in debian, its bad, and I generally haven't dealt with random packages!
[09:48] <Laney> sounds like a fun experience
[09:50] <darkxst> if i get the job sure, though having an interview and packaging review at the same time was a bit hard
[09:50] <darkxst> I think I picked out the worst of it though
[09:52] <darkxst> and if i get the job, I'm coming to europe
[09:53] <darkxst> not that its a requirement, but because I can
[09:53] <Laney> won't you need to get a permit?
[09:54] <darkxst> Laney, why? I have a UK passport
[09:55] <Laney> oh right
[09:55] <Laney> the best one
[09:56] <darkxst> yes, certainly much better than my AUS passport
[09:56] <Laney> http://www.movehub.com/sites/default/files/filemanager/passport-power-fixed.png
[09:58] <darkxst> but a lot of those would just be tourism visas?
[09:58] <Laney> right
[09:58] <Laney> still the best
[09:59] <Laney> and: http://www.bbc.co.uk/news/uk-34177107
[09:59]  * Laney is dressed entirely in union flags today
[10:00] <darkxst> Laney, your looking a little old there!
[10:01] <willcooke> Laney, phone:   01286 86 8081
[10:01] <Laney> 0181 811 8181
[10:01] <willcooke> \o/
[10:02] <willcooke> Going, going, going LIVE
[10:02] <willcooke> I still use that number when filling in online forms that I dont want to have my real number
[10:05] <pitti> Laney: heh, your Queen was all over the news yesterday in Germany even
[10:05] <Laney> \o/
[10:05] <Laney> I actually didn't know that until I heard it on the radio this morning...
[10:05] <Laney> probably would have guessed she already had the record
[10:10] <Sweet5hark> pitti: No wonder: German want to be kept informed about the news from house Coburg-Sachsen and Gotha! Esp. in relation to the Battenbergs!
[10:11]  * larsu wonders wtf Sweet5hark is talking about
[10:11] <pitti> Sweet5hark: I'm afraid I don't read enough "Klatschzeitungen" (in fact, none at all) to comprehend that :-(
[10:11] <larsu> not German enough I guess
[10:12] <pitti> Sweet5hark: is some British prince related to these families?
[10:12] <ogra_> are we discussion SuperIllu ?
[10:12] <ogra_> *discussing
[10:12] <pitti> /join #ubuntu-tabloids
[10:12] <ogra_> this channel seriously degraded recently :P
[10:12] <Sweet5hark> lol, this is hilarious
[10:12] <willcooke> \o/
[10:13] <pitti> non, cette chaîne est encore la mieux ! :-)
[10:15] <larsu> ogra_: recently?
[10:15] <ogra_> heh
[10:16] <pitti> I switched teams twice since my desktop days, and the guys here haven't kicked me out yet → happy ☺
[10:16] <larsu> that's because you haven't *really* switched
[10:17] <larsu> desktop at heart
[10:17] <pitti> well, I haven't done a gnome commit or desktop package upload for like two years :/
[10:17] <Sweet5hark> larsu: up until 1914, the house windsor was actually called house coburg-sachsen and gotha. some minor conflict in europe caused the namechange for reasons of popularity with the common people. same with the queens husbands name battenberg->mountbatten. <end of tabloid section/>
[10:17] <larsu> pitti: psst
[10:17] <larsu> Sweet5hark: "interesting"
[10:18] <pitti> Sweet5hark: thanks for the education :)  (... "some minor conflict"...)
[10:18] <larsu> it actually is interesting
[10:18] <larsu> for some weird definition of that word
[10:18] <pitti> yeah, it's history much more than tabloid indeed
[10:18]  * larsu congratulates the queen and all her ... err ... subjects?!
[10:18] <pitti> I guess gossip just turns into history for people who became sufficiently important
[10:19] <pitti> larsu: subordinates? not sure
[10:19] <larsu> Laney, willcooke: what are you to the queen?
[10:19] <Laney> subject
[10:19] <larsu> subjects? subordinates?
[10:19] <larsu> slaves?
[10:19] <willcooke> next door neighbour
[10:19] <pitti> admirer?
[10:19] <larsu> Laney: thanks :)
[10:19] <Laney> biggest fan and borderline stalker
[10:19] <larsu> willcooke: ooh. Did she ever come over because she was out of eggs?
[10:19] <Sweet5hark> larsu: subjects of her majesty. I challenge you to address people like that in scotland though.
[10:20]  * Laney has a tattoo of the queen's face
[10:20]  * pitti imagines Laney picking up the phone "Hey Liz, wanna come over for a beer?"
[10:20] <Laney> don't ask where
[10:20] <larsu> where?
[10:20] <Laney> on the underside of my feet
[10:20] <Laney> I did it to maximise pain
[10:20] <willcooke> larsu, *all* the time.  Seriously.  And when she runs out of king size cigarette papers at 3 in the morning
[10:20]  * larsu ← doesn't follow orders like a subject
[10:20] <Laney> that is how much I love her
[10:20] <larsu> willcooke: haha "king size"
[10:20] <Sweet5hark> krkr
[10:21] <willcooke> haha
[10:21] <larsu> Laney: dunno if tatoos on feet symbolize love...
[10:21] <pitti> larsu: not from sabdfl either? :-)
[10:22] <larsu> pitti: of course not^W^W^Wyes, I do!
[10:24] <Sweet5hark> one could put a tattoo on the underside of the foot, depicting a GNOME foot, which has a GNOME foot tattoo, which has a GNOME foot tattoo, which has a GNOME foot tattoo ...
[10:31] <darkxst> Sweet5hark, there is a new logo in the wild,
[12:06] <Trevinho> larsu: you want re-ack the change on ubuntu-themes? I've fixed the terminal style... The change also had the side effect of fixing the issue that caused it to resize when using the scrollbars (that was caused by widget size changes we had before)
[12:07] <larsu> Trevinho: sure
[12:08] <Trevinho> ta
[12:09] <didrocks> willcooke: oh, btw, I experienced the display monitor bug when changing config, do you have the bug # handy?
[12:09] <Trevinho> didrocks: which one? Blank areas?
[12:09] <didrocks> willcooke: as RAOF mentionned, it's a compiz/unity issue (you change workspace and all artefacts are gone)
[12:09] <didrocks> yep
[12:09] <Trevinho> didrocks: andyrock is on it
[12:09] <didrocks> oh excellent!
[12:09] <didrocks> blank and flashing, right?
[12:09] <willcooke> didrocks, changing work spaces didnt make any difference to me
[12:10] <didrocks> it did for me, tried multiple time
[12:10] <didrocks> however, dunno why it started affecting me today
[12:10] <Trevinho> didrocks: ah, I didn't try by changing WS, the only fix was to kill unity :D
[12:10] <didrocks> (tried to rollback this morning every updates for an hour, didn't find anything)
[12:10] <didrocks> Trevinho: seems it did work for RAOF and I
[12:10] <didrocks> and yeah, I would have never tried :p
[12:10] <willcooke> I /think/ that Seb and I revert to an old kernel to fix it
[12:10] <Trevinho> didrocks: might be, I just didn't try :)
[12:10] <willcooke> lemme see if I can find the bug
[12:11] <didrocks> can be 2 issues…
[12:11] <larsu> Trevinho: I get a black background behind the scrollbar in terminal
[12:11] <didrocks> willcooke: was it flashing for oyu?
[12:11] <didrocks> you*
[12:11] <Trevinho> larsu: isn't your terminal black anyway? :o
[12:11] <willcooke> didrocks, it was
[12:11] <larsu> Trevinho: nope
[12:11] <Trevinho> larsu: well, that's for radiance :) it should be black..
[12:12] <Trevinho> larsu: unfortunately it seems that it's impossible to make the actual BG transparent
[12:12] <Trevinho> so.... Mh let me see... if the bg can be different...
[12:12] <Trevinho> I didn't think it could be still changed without theming
[12:12] <willcooke> didrocks, I think we haven't opened a bug yet because we couldn't work out what had changed
[12:12] <willcooke> looking at seb's bugs I don't see anything
[12:13] <larsu> Trevinho: I think last time I tried it I found out the same thing
[12:13] <Trevinho> oh... So yes. the workspace changes fixes it
[12:14] <larsu> Trevinho: but we need to support differently-colored terminal backgrounds
[12:14] <Trevinho> larsu: let me see if I can find something by patching it...
[12:14] <larsu> thanks
[12:15] <Trevinho> didrocks, willcooke: the issue has been caused by a fix in unity, that unfortunately didn't take care of a bug in compiz :D. But andyrock should have a solution already.
[12:15] <didrocks> Trevinho: classical "fixing there, breaking here"? :)
[12:15] <willcooke> Trevinho, sweet.  I'll happily give it a go and report back
[12:15] <didrocks> Trevinho: do you have a bug for it? (I guess would be nice to scrubscribe)
[12:16] <willcooke> oh, I remember... popey - do you have a bug handy for the flickering screen issue?
[12:16] <Trevinho> didrocks: I had it (since it was reported sooo long ago), but I can't find it -_-
[12:16] <Trevinho> didrocks: the easier way to reproduce is just resizing a vmware window
[12:17] <popey> willcooke: the one we looked at at the sprint?
[12:18] <willcooke> popey, yeah
[12:18] <popey> https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1421575
[12:19] <willcooke> thanks popey
[12:19] <willcooke> didrocks, Trevinho ^^
[12:20] <larsu> so ... ubuntu-system-settings, unity-control-center, ubiquity, indicator-datetime (and probably other things) depend on libtimezonemap
[12:20] <larsu> which in turn depends on gtk
[12:20] <larsu> which we don't want on the phone by default
[12:20] <Trevinho> mh, ok... I'm sure there was another bug (from june)... It was something like "new mesa breaks unity...." but I can't find it anymore. I also replied, but no luck :(
[12:21] <larsu> u-s-s only uses it to parse the timezone file and maybe in the future get it from geonames.ubuntu.com
[12:21] <didrocks> willcooke: unsure if we should use that one for the one andyrock is looking at
[12:21] <didrocks> willcooke: maybe, let's wait for andy and the fix, we try it and see if you still have your issue
[12:21] <didrocks> (and so different cause)
[12:21] <Trevinho> Mh, the picture looks different from what I have BTW:..
[12:21] <willcooke> didrocks, sounds good
[12:21] <larsu> we have two options: (1) split the library or (2) reimplement (or copy) the parsing
[12:21] <Trevinho> Let's see
[12:21] <larsu> any suggestiongs?
[12:21] <larsu> *suggestions
[12:22] <andyrock> yeah I'm blocked on another compiz bug
[12:22] <larsu> Laney and I are of different opinion
[12:22] <larsu> maybe qt already has something for that?
[12:22] <didrocks> larsu: that's what I was going to say…
[12:22] <Laney> No
[12:23] <didrocks> they don't?
[12:23] <Laney> no
[12:23] <didrocks> :(
[12:23] <Laney> otherwise I wouldn't have made it use this
[12:23] <larsu> I guess Laney looked into that when he developed the time-date plugin
[12:23] <didrocks> I know it's not in QML, I would have hoped for the C++ version at least
[12:23] <didrocks> hum
[12:23] <Laney> The logic gets more complicated when someone implements the changes to weed out the shitty things from the results
[12:24] <larsu> right, this is a good point
[12:25] <larsu> also when we start doing a hybrid local/geonames approach
[12:25] <Laney> and all of these benefits are desirable for the installer, system-settings and everywhere else
[12:25] <Laney> IMHO
[12:25] <larsu> so we need a library in C, wrapped by C++, wrapped by qml
[12:25] <didrocks> agreed, you have multiple place where you want that behavior
[12:26] <larsu> or a daemon...
[12:26]  * larsu hides
[12:26] <Laney> you can just call the C from u-s-s
[12:26] <Laney> like we do currently
[12:26] <Laney> I don't imagine it would require much, if any, more glue there
[12:27] <larsu> at some point people will want c++
[12:27] <larsu> like they did for gsettings-qt
[12:27] <Laney> one step at a time
[12:28] <larsu> meh
[12:29] <larsu> I bet the next person using this will just implement it themselves
[12:29] <larsu> we should get the unity api team to do this
[12:29] <larsu> it's their job!
[12:29] <larsu> at least we need their input
[12:41]  * larsu feels ignored
[12:43]  * Laney hugs larsu
[12:44] <larsu> :)
[12:44] <Laney> I think the C++/QML thing could be done by them or whoever wants it
[12:44]  * larsu hugs back
[12:44] <Laney> in the meantime we can write the glib thing
[12:48] <larsu> libtimezone-locator?
[12:50] <jhodapp> Laney, hey, would you be able to prepare a silo against vivid+overlay for gstreamer 1.5.2...we're wanting to try to make sure that both vivid+overlay and wily are completely synced for the media stuff and this is the last change necessary
[12:51] <Laney> jhodapp: does it build against that?
[12:51] <Laney> jhodapp: also, are you sure that you want a pre-release there?
[12:51] <jhodapp> Laney, not currently, it would need a platform-api sync as well
[12:51] <Laney> and finally, it's .90 not and not .2
[12:51] <jhodapp> Laney, yes
[12:51] <Laney> ok...
[12:53] <jhodapp> Laney, didn't it used to be .2? I didn't pull that from nowhere :)
[12:53] <Laney> yes
[12:53] <Laney> wily moves on
[12:53] <jhodapp> ok good, I'm not going insane :)
[12:54] <jhodapp> Laney, thank you sir
[12:54] <Laney> I still think you're taking a risk if this is going to go to customers
[12:54] <Laney> jhodapp: but can you get a silo with the pre-requisites in it?
[12:55] <jhodapp> Laney, we'll give it a good test, it won't land for customers unless we're completely satisfied
[12:56] <Laney> the 1.6.0 final should come soon
[12:56] <jhodapp> Laney, any idea when?
[12:57] <desrt> hi desktop
[12:57]  * desrt rejoins EST
[12:58] <larsu> hi desrt!
[12:59] <Laney> jhodapp: they're following the gnome schedule more or less, so a couple of weeks
[12:59] <Laney> howdy desrt
[13:00] <jhodapp> Laney, ok, let's go with 1.5.90 right now and we can start testing with it...abeato is going to get platform-api into a silo against vivid and I'll let you know when that's ready. But we could hold off landing until 1.6 then
[13:01] <larsu> Laney: oh wow system-settings only fetches a list and does all the matching itself?!
[13:02] <larsu> maybe we should simply ship a gvariant blob (desrt will be happy) of the list and load that?
[13:02] <larsu> or even mmap it
[13:02] <desrt> gvariant on mars!!
[13:02] <larsu> desrt: this is specifically about earth
[13:02] <larsu> as in "cities on earth"
[13:02] <Laney> the matching should probably go back to the library I guess
[13:03] <desrt> you want something with compression
[13:03] <desrt> this is a giant list that will take up a lot of space and compress well
[13:03] <larsu> how does that work with qabstracttablemodel?
[13:03] <desrt> and if you use a block compressor and put it in sorted order, you can binary search it
[13:03] <desrt> probably don't.
[13:03] <desrt> *doesn't
[13:03] <larsu> doesn't what?
[13:04] <desrt> the compressed approach
[13:04] <larsu> we seem to be set on writing a new library anyway
[13:04] <Laney> I'll help work out the system settings side once a good API exists
[13:04] <larsu> so we might as well do it right
[13:04] <larsu> desrt: why not?
[13:04] <desrt> i have to assume the existing interface of the tabelmodel will not be easy to implement against a binary-searchable seekable compressed stream in an efficient way
[13:05] <desrt> and in particular, the way in which frontends use that interface could be shite
[13:05] <larsu> all they want to do is get a list of cities for a substr
[13:06] <desrt> how about prefix? :)
[13:06] <larsu> might be overkill for 5 megs of data
[13:06] <desrt> well
[13:06] <desrt> honestly, at 5MB of data, you can just keep it compressed on disk and then extract it into RAM when you're using it
[13:06] <desrt> which is rarely
[13:07] <larsu> exactly
[13:07] <desrt> so forget mmap :)
[13:07] <larsu> this is what we're doing now (except parse from a .txt file)
[13:07] <larsu> but if we only doing this
[13:07] <larsu> we might as well sktip the lib and load a gvariant
[13:07] <larsu> *skip
[13:07] <desrt> .gv.xz :)
[13:07]  * larsu would be happier about that
[13:07] <larsu> but Laney might not be
[13:08] <desrt> the streaming decompression in glib is very excellent
[13:08] <Laney> need to implement mpt's skip-the-shit algorithm
[13:08] <larsu> skip the shit?
[13:09] <larsu> ya, I'd do that at creation time
[13:09] <larsu> that will reduce file size as well
[13:09] <desrt> what if we want the shit?
[13:09] <larsu> nobody wants the shit
[13:09] <Laney> In addition, to avoid duplication, a suburb (PPLX) should not be included in search results whenever a larger settlement (PPL, PPLA, PPLC) that has the same time zone is present in the same search results
[13:09] <larsu> hence the nmae
[13:09] <desrt> i know a guy who lives outside of LA...
[13:09] <larsu> in a city that is not a PPLX
[13:10] <larsu> I would hope
[13:10]  * desrt reads pplx as peopleplex
[13:10] <larsu> gzip gets it down to 1.9mb
[13:10] <larsu> nice
[13:12] <desrt> larsu: would accept a patch for g_variant_new_from_input_stream() if you can figure out the fucked-up-ness of the library layering
[13:12] <desrt> ie: because of gio/glib would probably need to look more like g_input_stream_read_gvariant() or something dumb like that
[13:13] <larsu> g_variant_input_stream_new()
[13:13] <desrt> no.
[13:13] <desrt> that's something else :)
[13:13] <desrt> although i guess you could use it as this, with some very minimal extra framing
[13:13] <larsu> what would that be?
[13:13] <desrt> ie: a 8-byte size record at the start of the file
[13:14] <larsu> this is how gfileinputstream works, right?
[13:14] <desrt> no.  that works by posix...
[13:14] <larsu> y posix?
[13:14] <larsu> man I hate glib's streams
[13:14] <desrt> GInputStream API is basically a posix copy
[13:14] <desrt> heh
[13:14] <desrt> you were talking about how good they are only a week ago :)
[13:14]  * larsu wishes we had something like node's streams
[13:14] <desrt> in context of "GIO is such shit... but at least the streams are OK" :)
[13:15] <larsu> desrt: and you started to point out how shitty they are
[13:15] <larsu> with read_all
[13:15] <desrt> it's a bit of a disaster
[13:15] <desrt> but it's actually not beyond repair
[13:15] <larsu> and ginputstream is quite ok, it's the subclasses that are bad
[13:15] <desrt> dunno
[13:15] <desrt> subclasses are useful
[13:15] <desrt> the buffered and data streams are pretty much what you want them to be
[13:15] <desrt> the converter stream is pretty much what you want it to be
[13:16] <desrt> and the converters themselves are good
[13:16] <desrt> and there is a lot of conveinent utility API around all of this stuff
[13:16] <desrt> (although there are annoying gaps in that API that could use to be filled in)
[13:16] <desrt> it's _pouring_
[13:16] <desrt> nice view of that from here :)
[13:17] <larsu> but back to the topic
[13:17] <larsu> why is gvariantinputstream different from gfileinputstream?
[13:17] <desrt> GVariantInputStream would be a filter stream
[13:17] <desrt> well, in fact, GVariantInputStream would not be an input stream at all
[13:17] <desrt> but rather would contain an input stream
[13:18] <larsu> ah, so the naming's weird
[13:18] <desrt> ie: there would be no read() that gets you bytes
[13:18] <larsu> right
[13:18] <desrt> the filterinputstream model is a bit sideways, i will give you
[13:18] <desrt> since filters must also be streams in the "i can read some bytes" sense
[13:18] <desrt> whereas i would tend to think of a GVariantInputStream as a filter....
[13:19] <desrt> but this is pretty much a limitation of using subclasses instead of interfaces
[13:20] <desrt> anyway... this is all pretty theoretical... all you need to do is to make a fileinputstream, wrap it in a converter (decompressor) stream, and splice() that into a memoryinputstream
[13:20] <desrt> .steal_bytes() from that stream and shove them into a GVariant
[13:20] <desrt> done
[13:21] <desrt> *into a memoryoutputstream
[13:21] <larsu> this sounds more complicated than I'd expect...
[13:21] <desrt> the "nice new API" would take care of the creating of the memory output stream, splicing, stealing, and construction of GVariant
[13:22] <desrt> you could also read_all_as_bytes() but that doesn't exist yet :)
[13:25] <larsu> Laney: would you be ok with this solution as well?
[13:25] <larsu> desrt: why do we need a memory output stream in between?
[13:25] <desrt> you could also just use libxz or whatever it is externally
[13:25] <desrt> larsu: you don't.  it's just a convenient way of getting some bytes
[13:25] <desrt> splice() is nice
[13:26] <larsu> ah, got it
[13:28] <desrt> anyway
[13:28] <desrt> gotta catch a flight
[13:28] <desrt> so airplane mode
[13:29] <Laney> larsu: what does it look like from the client side?
[13:32] <larsu> Laney: iterating over a variant
[13:33] <larsu> desrt: have a safe trip!
[13:33] <larsu> Laney: maybe that's a shitty interface though...
[13:39] <Laney> meh, it's probably alright
[13:39] <davmor2> Laney: hey, I upgraded to wily on the first beta release and I noticed that the power indicator isn't showing an icon on fully charged with power connected.  If I detatch the power I get an icon and if I plug it back in after 30 minutes then I get the charging icon till it shows as full again
[13:40] <Laney> what's the preference in Power?
[13:40] <davmor2> Laney: when Battery is present
[13:41] <Laney> it sounds like the behaviour of the second optino
[13:41] <davmor2> http://people.canonical.com/~davmor2/power-indicator.png this is what I see note the icon is missing not that the indicator is, it is the square next to the sound indicator
[13:42] <Laney> you have the *not found* icon
[13:42] <davmor2> Laney: hence saying it isn't showing the icon
[13:43] <davmor2> Laney: it is only that full one with power though
[13:43] <Laney> isn't showing the icon sounds like it is showing nothing
[13:44] <davmor2> Laney: ah sorry :)
[13:44] <Laney> can't make this happen
[13:45] <davmor2> Laney: is there a log I can look at so I know the path it is looking at?
[13:45] <davmor2> Laney: and see if the icon is actually there
[13:45] <Laney> stop indicator-power; G_MESSAGES_DEBUG=all /usr/lib/x86_64-linux-gnu/indicator-power/indicator-power-service
[13:46] <davmor2> Laney: thanks I'll give it a go
[13:46] <Laney> or maybe .cache/upstart/indicator-power.log
[13:47] <Laney> wait
[13:47] <Laney> there is a power icon in that screenshot?
[13:47] <Laney> 3 from left
[13:49] <davmor2> Laney: that is the Solaar one for the logitech keyboard and mouse
[13:49] <davmor2> Laney: the missing icon one is the power indicator
[13:52] <Trevinho> davmor2: also checking UPS can be a good way to check it
[13:52] <Trevinho> gdbus call --session --dest com.canonical.Unity.Panel.Service.Desktop --object-path /com/canonical/Unity/Panel/Service --method com.canonical.Unity.Panel.Service.SyncOne com.canonical.indicator.power
[13:52] <Trevinho> For instance, here it returns
[13:52] <Trevinho> ([('com.canonical.indicator.power', '0x1966400', 'indicator-power', '', true, false, uint32 5, '/usr/share/icons/ubuntu-mono-dark/status/22/battery-full-charged-symbolic.svg', true, true, 1460)],)
[13:53] <Trevinho> larsu: so... I've complied the terminal 3.18, just to check if anything changed... And actually has not
[13:53] <larsu> Trevinho: the change would be in vte
[13:54] <Trevinho> larsu: mh, I thought about that... let me see
[13:55] <davmor2> Laney: Trevinho: I still see no mention of the icon it is looking for but I do see deprecated commands http://paste.ubuntu.com/12320941/
[13:56] <Trevinho> larsu: let me check if I can get that scrollbar transparent in some way.... Otherwise we have to do a choice. Make it light or dark. ;/
[13:56] <Trevinho> I'd vote for dark since it's default, and even with light BGs it doens't look too bad (http://i.imgur.com/479fsmQ.png)
[13:56] <Trevinho> davmor2: check that gdbus command I pasted before
[13:56] <Trevinho> err sorry it's there
[13:56] <Laney> image-missing
[13:56] <davmor2> :)
[13:57] <Laney> so do you have /usr/share/icons/ubuntu-mono-dark/status/22/battery-full-charged-symbolic.svg ?
[13:57] <larsu> Trevinho: we have the background transparent by default, no?
[13:57] <Trevinho> larsu: nope
[13:57] <larsu> Trevinho: we should turn o-s off for gnome-terminal if we need to give it a bg
[13:58] <larsu> how does Adwaita do it?
[13:58] <Trevinho> larsu: adwaita is all opaque... and it doesn't change size
[13:59] <Trevinho> larsu: we can just always use the fat-mode for terminal, but I don't like that much
[13:59] <larsu> Trevinho: better than always a background color
[13:59] <Trevinho> larsu: well, we need a bg anyway...
[13:59] <davmor2> Laney: oh should there be 2 of them?
[13:59] <Laney> 2 what?
[14:00] <davmor2> Laney: /usr/share/icons/ubuntu-mono-dark/status/22/battery-full-charged-symbolic.svg
[14:00] <Trevinho> larsu: it's just that it would be always dark or light
[14:00] <larsu> ya...
[14:00]  * Trevinho checks if coding can help
[14:00] <larsu> meh, I don't actually care
[14:00] <Laney> davmor2: don't get it, that's just one icon no?
[14:00] <davmor2> Laney: nevermind charged and charging
[14:01] <davmor2> Laney: so I have both let me open it up in nautilus and make sure there is an actual image to it
[14:03] <davmor2> Laney: yeap I have the icon and it has an image :)
[14:04] <Trevinho> with default bg I think they look quite nice (http://imgur.com/PI10vDG,qDk2OGS#1, there are both images for mouse hovering and not)
[14:05] <larsu> I suspect many people will not use the default bg
[14:05] <larsu> and also that especially those people will care a lot about this issue :D
[14:05] <larsu> but feel free to try ;)
[14:23] <Laney> Trevinho: can you help davmor2 maybe?
[14:24] <Trevinho> mh, yeah... let me see
[14:24]  * willcooke is charging his test laptop to see if it effects him too
[14:24]  * Laney has to go eat lunch
[14:24] <Laney> before my face falls off
[14:25] <Trevinho> Laney: eating lunch is too mainstream!
[14:25] <davmor2> Laney: Lunch it's getting on for tea time
[14:25] <Laney> just realised the cleaner is going to come in the middle of me trying to eat
[14:25] <Laney> #middleclassproblems
[14:26] <Trevinho> As for the icon... Mh, it's weird since I guess that that's the icon that the indicator provides
[14:26] <Laney> it's from the theme
[14:26] <Trevinho> davmor2: if you want to debug, the right spot is gtk_image_to_data inside panel-service.c
[14:26] <Trevinho> (what a bad function name...)
[14:28] <Laney> indicator-power calls g_themed_icon_new_from_names itself
[14:29] <willcooke> embed irc hyperlinks
[14:29] <willcooke> err, wrong window
[14:29]  * didrocks was trying to think hard about a link :)
[14:30] <willcooke> I assume its irc:// but how do you specify the server
[14:30] <willcooke> ah, found a page that answers that
[14:30] <Trevinho> irc://irc.freenode.org/channel
[14:30] <willcooke> :) thanks Trevinho
[14:30] <willcooke> I think you are my new Google
[14:30] <Trevinho> :)
[14:31] <willcooke> Trevinho, what is the gross domestic product of The Vatican
[14:31] <larsu> $666
[14:31] <willcooke> \o/
[14:31] <Trevinho> lol..
[14:32] <Trevinho> Ah, I think it's ... /#channel btw... or ..... /nickname (and can also be ircs://)
[14:33] <willcooke> meh GDocs doesn't allow it
[14:33] <Trevinho> willcooke: use http://is.gd to workjaround it
[14:34] <willcooke> hahaha
[14:34] <willcooke> nice
[14:37] <Trevinho> willcooke: i.e http://is.gd/ubuntu_desktop_irc points here
[14:37] <Trevinho> with stats :D http://is.gd/stats.php?url=ubuntu_desktop_irc
[14:50] <Sweet5hark> urgh, great. there goes the "getting to the libreoffice conference comfortably by train".
[14:50] <willcooke> Sweet5hark, oh noez!  What happened?
[14:50] <Sweet5hark> seems denmark just canceled all trains from germany.
[14:50] <ogra_> oh my
[14:51] <Sweet5hark> "Welcome to the schengen area. There are refugees about. Please wait until we have reerected national borders Berlin-wall style."
[14:52] <ogra_> and they just want to go through, they dont even aim for denmark
[15:08] <Sweet5hark> https://translate.google.com/translate?sl=da&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fwww.dsb.dk%2Ftrafikinformation%2Fandringer-i-trafik-og-drift%2Fandringer-i-trafik-og-drift%2Ffr-internationale-tog%2F&edit-text=&act=url
[15:09] <Trevinho> larsu: tried even git vte... Still buggy scrollbar (i.e. can't be transparent and doesn't support .hovering class)
[15:10] <larsu> :/
[15:18] <willcooke> davmor2, larsu Trevinho - my laptop is now fully charged, and I *do* have the correct icons
[15:18] <willcooke> fwiw
[15:19] <davmor2> willcooke: thanks I think it just got screwed up in the upgrade to be honest I'll try a live cd later
[15:22] <willcooke> Parcel to Brazil is now in the right country.  Parcel to Germany, still at Heathrow.  Sent via the same service.
[15:47] <Laney> larsu: conflict when merging the two i-datetime branches; think that one needs to be merged with the other (prereq)
[15:47] <larsu> Laney: oh, isn't it?
[15:47]  * larsu should not use git-bzr for such things...
[15:47] <Laney> seems not
[15:48] <Laney> it's usually sane but this time they touch the same stuff
[15:49]  * larsu fixes
[15:49] <Laney> fantastisch
[15:50] <larsu> :)
[15:51] <larsu> ah conflicts indeed
[15:58] <Sweet5hark> cyphermox: does https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1491964/comments/7 clear remaining issues?
[16:01] <cyphermox> Sweet5hark: not really
[16:02] <Trevinho> larsu: I think i've the right fix... But i need to polish it
[16:02] <cyphermox> the gcc5 transition doesn't mean that we should duplicate code if it's already in the archive
[16:02] <larsu> Trevinho: cool!
[16:03] <cyphermox> the whole changelog is very confusing, there's changes in there that mean nothing if it's already been applied to earlier releases and there's essentially no change
[16:03] <Trevinho> larsu: basically the reason is that the scrollbar is in an hbox, together with the terminal screen... So the background is transparent. But the window has white bg.... Basically making the hbox bg the same of the terminal one works
[16:05] <Trevinho> Ah, since the override bg function is deprecated, really isn't an utility function to do this with a GdkARGB without translating it into a string? :o
[16:06] <larsu> Laney: my own branches confuse me :/
[16:06] <larsu> it's like ... what's going on, I didn't even touch this!
[16:07] <Sweet5hark> cyphermox: as written in the FFE comment, a changelog cannot really represent the three-way branch merge that is a new major LibreOffice version. Of course, we shouldnt duplicate code without a reason. Are you suggesting we should sponsor a package with system libs even though its not widely tested by endusers? That is the alternative.
[16:08] <Sweet5hark> cyphermox: For me, thats a trivial change
[16:08] <cyphermox> I'm never saying untested things should land, that's wrong too
[16:08] <Sweet5hark> cyphermox: But I really would prefer not to do so with the first upload of a new major series package.
[16:09] <cyphermox> so are those fully external libraries or are they things normally built by libreoffice source?
[16:11] <Sweet5hark> cyphermox: Well, it either rock or a hard place -- as usual with LibreOffice. And as said: I want to go back to system libs. Right after this is in wily, I'll switch back here, testbuild with that and then only wait for 5.0.2.2 to be tagged upstream to land both together.
[16:12] <cyphermox> that isn't really answering the question though, are these libs from libreoffice source normally? or are they completely separate projects?
[16:12] <Sweet5hark> cyphermox: I'd rather not put even more change in _this_ upload (as that is the alternate).
[16:12] <cyphermox> I'm trying to get enough state about the whole thing to properly review libreoffice, it's far from the types of project I usually look at
[16:16] <cyphermox> it's also unclear to me whether using the internal copy of these libs is deviating from Debian
[16:16] <Sweet5hark> cyphermox: LibreOffice bundles ~all the tarballs of the libs we usually ship in the distro in the libreoffice.orig-src*.tar.xz tarball (a tarball of ~50 tarballs). Ideally we dont use any of those, but we always bundle them in the source, so that, if for some reason an internal copy is needed we can use it without doing a fake upsteam version (like 5.0.1+somenastyfix1) for adding the bundled external.
[16:16] <cyphermox> so collada is something that comes from libreoffice / TDF?
[16:17] <Sweet5hark> cyphermox: that is a practice that is shared with debian: they also have ~all the tarballs in their libreoffice.orig-src*.tar.xz.
[16:17] <larsu> Laney: hm, not sure which branch to overwrite now ... prereq doesn't work because I had to fix up some commits
[16:19] <Sweet5hark> cyphermox: in a way, it is coming from TDF: e.g. on window, TDF has to build all those too. when you do a TDF-style upstream or developer build, you are downloading all those externals from a TDF server.
[16:19] <cyphermox> but who is upstream?
[16:21] <Sweet5hark> cyphermox: various. Consider TDF/LibreOffice being a small source distro on its own for these. See e.g. the upstream download.lst: http://cgit.freedesktop.org/libreoffice/core/tree/download.lst
[16:23] <Laney> larsu: you need more commits in the remove warnings branch or what?
[16:24] <larsu> Laney: yes
[16:24] <larsu> well, not more
[16:24] <larsu> I amended the first one....
[16:24] <Laney> then that one
[16:24] <Laney> oh...
[16:24] <larsu> shall I just push a new one>
[16:24] <larsu> >
[16:24] <larsu> ?
[16:25] <Laney> just overwrite it
[16:25] <Laney> I would have probably done another one on top
[16:25] <larsu> yeah that was unwieldy
[16:25] <Sweet5hark> cyphermox: (for some of those TDF is upstream itself, e.g. for the document liberation libraries like librevenge, libvisio etc.)
[16:25] <larsu> (also, I was using git rebase \o)
[16:25] <Laney> /o\
[16:25]  * desrt gets annoyed at inotify some more
[16:27] <Sweet5hark> cyphermox: note none of this is new: it has been like that since before there even was libreoffice in openoffice/go-oo and is so in Debian too.
[16:27] <cyphermox> well, you know, but I don't
[16:28] <cyphermox> so I need to find that and that helps me review the whole thing
[16:28] <Sweet5hark> cyphermox: sorry, Im quite blind to the ugliness by now ;)
[16:29] <cyphermox> there isn't enough beer in the world...
[16:29] <larsu> Laney: pushed
[16:29] <cyphermox> I don't know how you manage to stay sane and/or sober.
[16:30] <Sweet5hark> cyphermox: why do you assume me to be either?
[16:30] <larsu> Laney: weird, it shows my last commit from the other branch, but not yours
[16:30]  * larsu is confused
[16:30] <larsu> but this branch hs everything in it
[16:30] <larsu> it's just launchpad
[16:31] <cyphermox> Sweet5hark: you're neither in the hospital (you're online) or tied up in a long-sleeved jacket (able to type) :D
[16:32] <ogra_> he's got a fast nose for typing ;)
[16:36] <Laney> this sucks
[16:36] <Laney> can I just take the one branch?
[16:37] <Laney> if it has everything, then yes
[16:37] <Laney> doing that
[16:37] <larsu> yep
[16:37] <Sweet5hark> cyphermox: In the old days, being able to switch between using the internal or external version of a libs was essential, as with some 50 libs one of them would always break openoffice (esp. if openoffice vendor-patched their libs).
[16:40] <Sweet5hark> cyphermox: while upstream libreoffice still does some small vendor patches, the problem isnt that huge anymore luckily. Still we have all that foo in place for it: You add or remove something from the SYSTEM_STUFF variable in the ./debian/rules file, regenerate the control file with "./debian/rules control" to update build/runtime deps and then rebuild the source package ...
[16:41] <Sweet5hark> (Yes, ./debian/control for LibreOffice is generated by a rule in ./debian/rules in case you didnt see this before.)
[16:44] <cyphermox> yeah, I saw
[16:56] <desrt> new nightmares in inotify land: watch file /home/desrt/foo, rm file /home/desrt/foo
[16:56] <desrt> what event gets reported?
[16:56] <desrt> "attributes changed", of course
[16:57] <desrt> because /home/desrt/foo and /home/desrt/bar were both hardlinks to the same inode, and because of that the file wasn't really deleted -- only its link count changed
[17:00] <willcooke> right, time to go.  l8r
[17:55] <Trevinho> larsu: I've added the patch https://bugzilla.gnome.org/show_bug.cgi?id=754796
[17:55] <Trevinho> Laney: would you be ok on patching our version with that, waiting upstream?