[08:00] <seb128> good morning desktopers
[08:03] <Laney> lut
[08:03] <seb128> Laney, hey, wie gehts?
[08:04] <Laney> hey seb128, sehr gut, danke!
[08:05] <Laney> und dir?
[08:05] <seb128> Laney, larsu, desrt: good news for you, robert_ancell seems to have decided he was going to fix those nautilus bugs, he did SRU a segfault fix already and is looking at the bookmark one now (from launchpad comments)
[08:05] <Laney> nice
[08:05] <seb128> Laney, très bien merci ;-)
[08:05] <Laney> prima!
[08:05] <larsu> awesome, thanks!
[08:05] <larsu> hi Laney
[08:08] <Laney> hola larsu, ¿cómo estás?
[08:09] <larsu> Laney: muy bien, gracias! (thanks for getting me in the mood for my Spanish class tonight)
[08:09] <Laney> maracas!
[08:09]  * didrocks is upset about chromium loosing my tabs
[08:09] <didrocks> it doesn't seem to be something reported on launchpad though
[08:10] <didrocks> (it looses silently *some* tabs, not all)
[08:10]  * didrocks switched since yesterday to chrome unstable which is better in that regard
[08:10] <didrocks> (and yeah, all extensions disabled)
[08:10] <didrocks> let's see once qengho is around
[10:06] <mpt> seb128, mdeslaur: I was under the impression that Gnome made PolicyKit dialogs system-modal because that was the only way to ensure other programs weren’t listening to the keypresses. I might be wrong about that, though.
[10:08] <mpt> If so, though, my POV is that if you have a password-sniffing program running on your PC you’ve already lost. My online banking password is more important than my login password, for example, and browsers don’t go system-modal whenever they see <input type="password"> in an HTTPS page.
[10:27] <mlankhorst> well I found 2 or 3 bugs related to xserver switching now, fun! :P
[10:45] <mlankhorst> hm make that 4
[10:46] <mlankhorst> well, 5 but I haven't been able to reproduce the 5th one a second time
[10:49] <chrisccoulson> is anybody else able to adjust their laptop panel brightness in trusty?
[10:49] <chrisccoulson> the slider in the control centre does nothing here
[10:51] <chrisccoulson> and pressing the buttons on my laptop seems to work with around a 1-minute lag :/
[11:14] <Mirv> chrisccoulson: works fine here, on sandybridge
[11:32] <seb128> Laney, do you have an opinion on SRUing the gstreamer point release bugfix updates to trusty?
[11:34] <xnox> seb128: is there a way for me to launch qt5 ubuntu-one setup page?
[11:34] <seb128> xnox, what is "qt5 ubuntu-one setup page"?
[11:34] <Laney> seb128: Probably worth it
[11:34] <Laney> The fixes sound important
[11:34] <seb128> Laney, k, I'm going to have a look to gst/base to start
[11:35] <xnox> seb128: sorry, system-settings, online accounts, ubuntu one plugin page.
[11:35] <Laney> the synced ones should be easy
[11:35] <seb128> right
[11:35] <xnox> seb128: launch on the desktop (possibly stand alone), it works on the phone, but failing to launch it on the desktop
[11:35] <Mirv> I was hit by bug #1290368 to the extent I switched my shotwell hack from totem to ffmpegthumbnailer
[11:35] <ubot2> Launchpad bug 1290368 in gst-libav1.0 (Ubuntu) "totem-video-thumbnailer crashed with SIGSEGV in magazine_chain_pop_head()" [Medium,Confirmed] https://launchpad.net/bugs/1290368
[11:35] <Laney> just take them and change the version
[11:35] <Mirv> but I think it also occured when nautilus thumbnails
[11:36] <Laney> yes, it's the same program
[11:36] <seb128> Laney, right, I reading debdiff and I'm going to test build/run a bit first
[11:36] <Laney> can you reproduce it?
[11:36] <Laney> if so, debugging would be nice
[11:36] <seb128> talk to tjaalton maybe, I think he's looking after gst-libav
[11:36] <seb128> that's not something we install by default btw
[11:36] <Mirv> I'll try on my work laptop, I can probably at least provide a 0.1 second video from my camera if that's what triggers it (the format)
[11:37] <Laney> try running the program manually on it
[11:38] <seb128> xnox, dunno, check with mardy maybe ... doing "system-settings online-accounts" work, but I'm unsure about opening directly a specific account in there
[11:38] <tjaalton> haven't touched gst-libav..
[11:39] <Laney> deny all knowledge
[11:39]  * xnox ponders if i happen to be TIL
[11:39] <tjaalton> but it's probably same as debian 739579
[11:39] <ubot2> Debian bug 739579 in gst-libav1.0 "gstreamer1.0: totem crashes while seeking" [Important,Fixed] http://bugs.debian.org/739579
[11:39] <seb128> tjaalton, sorry, I'm probably confused, I though you had interest in it/looked at it before
[11:40] <Mirv> ok it was not as easy as "any file"
[11:40] <xnox> yeah, not me. However Laney does have a change in the TIL =)
[11:40] <tjaalton> seb128: it was -vaapi
[11:40] <seb128> tjaalton, oh,  right, thanks for correcting me ;-)
[11:40] <tjaalton> :)
[11:42] <Mirv> oh, yes it is, indeed seeking a h.264 file. there was just this another bug I noticed on my desktop computer that if I've gstreamer1.0-vaapi installed GStreamer thinks it doesn't have anything to play h.264 with
[11:44] <Mirv> yep, totem-video-thumbnailer -r 00017_2.MTS -t 1 out.png
[11:44] <Mirv> putting a clip of cat video to LP
[11:52] <Mirv> and not so easy, if I switch the container format so that I can get a small clip out of it, it doesn't crash anymore. so, shooting a new clip instead of using an old one
[11:52] <seb128> Mirv, opening the bug upstream would be nice
[11:57] <Mirv> well it sounds like it could be fixed in that new release in Debian, which I'm testing now
[11:58] <Mirv> so https://bugs.launchpad.net/ubuntu/+source/gst-libav1.0/+bug/1290368/+attachment/4095344/+files/00004.MTS is a shortish example that crashes when seeked
[11:58] <ubot2> Launchpad bug 1290368 in gst-libav1.0 (Ubuntu) "totem-video-thumbnailer crashed with SIGSEGV in magazine_chain_pop_head()" [Medium,Confirmed]
[12:03] <Mirv> yep, 1.2.4 from Debian fixed the issue
[12:03] <seb128> Mirv, 1.2.4 of what source?
[12:03] <Laney> libav
[12:03] <Laney> gst-
[12:03] <seb128> k
[12:04] <seb128> who wants to SRU that one? ;-)
[12:04] <Mirv> that
[12:04] <seb128> I just uploaded gst/base to the desktop ppa for testing btw
[12:04] <seb128> 1.2.4 of those I mean
[12:08] <Mirv> I SRUfied the bug description for now. I guess 1.2.4-1 will need to be autosynced to utopic first.
[12:08] <Laney> It can be uploaded first
[12:08] <Laney> I'll look at the rest of gst later if nobody else wants to
[12:15] <seb128> Laney, if you want to do those that would be welcome
[12:16] <Laney> nod
[12:19] <xnox> seb128: hm $ system-settings online-accounts -> click add account -> click ubuntu one -> empty screen
[12:20] <seb128> xnox, talk to kenvandine or mardy, I've nothing to do with online accounts
[12:21] <xnox> seb128: ok. thanks.
[12:21] <Laney> interesting, that doesn't list all of the same things that the u-c-c panel lists
[12:22] <Laney> like SIP says empathy there but not in u-s-s
[12:22] <Laney> if you click on an existing account that is
[12:46] <seb128> hum, launchpad having issues?
[12:47] <Laney> My grab-merge isn't grabbing the merge
[12:47] <Laney> so maybe ...
[12:47] <Laney> it's a sign
[12:47]  * Laney → lunch
[12:49] <xnox> Laney: yeah #is-outage is starting to heat up.
[13:12] <qengho> didrocks: "losing your tabs"?
[13:13] <didrocks> qengho: loosing*
[13:13] <didrocks> qengho: basically, if I quit chromium (and the notify)
[13:13] <didrocks> restarts it
[13:13] <didrocks> I can get from 17 tabs
[13:13] <didrocks> to 15, or even 2!
[13:13] <didrocks> it looses randomly some tabs on restart
[13:14] <didrocks> I disabled all extensions to ensure it's not the source of it
[13:14] <didrocks> and I can still reproduce it
[13:14] <didrocks> no issue with chrome unstable FYI
[13:14] <didrocks> from what I got popey also heard about that
[13:15] <popey> chromium 24 is a big step backwards
[13:15] <popey> you can't even horizontal scroll
[13:15] <popey> (worked in previous versions)
[13:15] <popey> https://code.google.com/p/chromium/issues/detail?id=338604
[13:16] <seb128> popey, talk to qengho
[13:16] <popey> hello qengho
[13:16] <seb128> tkamppeter, hey, do you remember https://bugs.launchpad.net/ubuntu/+source/libspectre/+bug/1242678?
[13:16] <ubot2> Launchpad bug 1242678 in libspectre (Ubuntu Trusty) "evince cannot render some EPS files" [High,Triaged]
[13:17] <seb128> tkamppeter, do you think you could/test the patch in https://bugs.freedesktop.org/show_bug.cgi?id=76450? that would be a nice one to backport to our package
[13:17] <ubot2> Freedesktop bug 76450 in general "Documents are not rotated correctly" [Normal,New]
[13:22] <qengho> didrocks: Hrm. Let's debug that.
[13:22] <qengho> didrocks: Let's go to #ubuntu-chromium so we don't bore these folks.
[13:23] <qengho> popey: horizontal scroll is on my to-do list.
[13:32] <bjsnider> pitti: what do you think about adding gdm to the systemd ppa? the service file lib/systemd/system/gdm.service is being built but not installed so that could be the only change
[13:32] <pitti> bjsnider: it probably also needs some dh_systemd_* invocations then
[13:33] <pitti> bjsnider: if you have a package, I'm happy to sponsor it to utopic and put it into the PPA (while utopic is frozen)
[13:34] <bjsnider> i'm the guinea pig i guess hahaha
[13:35] <bjsnider> dh_systemd isn't part of the loop at all at the moment
[13:59] <Laney> xnox: oh, what's that?
[13:59] <xnox> Laney: DoS or some such against something.
[13:59] <xnox> Laney: or do you mean the channel?
[13:59] <Laney> people are mean
[13:59] <Laney> yes but that's good enough info
[14:00] <xnox> Laney: #is-outage is #is with just #is folks talking when dealing with issues. It's like "real-time incident report reality TV show" =)
[14:00] <dobey> heh
[14:01] <Laney> sounds distracting :P
[15:29] <mlankhorst> just like real reality tv shows!
[16:05] <kenvandine> the PerformanceOverlay is simply amazing... that is all :)
[16:05] <Laney> what's that
[16:06] <kenvandine> developer.ubuntu.com/api/qml/sdk-14.04/Ubuntu.PerformanceMetrics.PerformanceOverlay/
[16:06] <kenvandine> Laney, ^^ although that link doesn't really give it justice :)
[16:06] <Laney> yeah ...
[16:06] <kenvandine> shows a slick running graph on top of your app
[16:06] <kenvandine> showing performance metrics live
[17:11] <Laney> night #ubuntu-desktop
[17:19] <tkamppeter> seb128, I will look into it. By the way, who is the maintainer/uploader of libspectre or does Ubuntu only auto-sync?
[17:20] <seb128> tkamppeter, thanks, we auto-sync from Debian
[17:20] <seb128> Laney, 'night
[17:21] <bjsnider> pitti: i built gdm with modifications that might work, but the ppa doesn't work here -- would you like me to email you the gdm packages so you can test them in your vm?
[17:55] <kenvandine> robru, i've been profiling the friends-app scrolling based on that bug from davmor2
[17:55] <robru> kenvandine, oh yeah?
[17:56] <kenvandine> robru, i've stepped back through the stack to the point where i have an empty flickable... no friends-app code at all
[17:56] <robru> kenvandine, what kind of profiling tools do we have for qml? I just know python stuff really
[17:56] <kenvandine> same performance :/
[17:56] <kenvandine> you can get debug output for object creation timing
[17:56] <kenvandine> and we have the slick PerformanceOverlay in the sdk
[17:56] <kenvandine> so you can see a pretty graph on top of you app
[17:56] <robru> kenvandine, it's not the first time I've heard that qml has poor performance :-/
[17:57] <kenvandine> i don't think that it's really "poor"
[17:57] <kenvandine> just while doing the constant movement it uses the cpu
[17:57] <kenvandine> it doesn't look bad :)
[17:57] <robru> heh
[17:57] <kenvandine> at one point Kaleo said we should keep all the timings down below 10ms while it scrolls
[17:58] <davmor2> robru, kenvandine: Yeah 102.2% of the cpu is the bit you missed there ;)
[17:58] <kenvandine> and i think i had it down to about 14
[17:58] <kenvandine> but now it's at 17ms
[17:58] <kenvandine> but i get the same with an empty flickable
[17:58] <robru> davmor2, but apparently it does the same with an empty flickable... ^^
[17:58] <kenvandine> set the contentHeight to 1000 with nothing rendering inside it
[17:58] <davmor2> robru: indeed which I think is even more scary :)
[17:58] <kenvandine> and drag it
[17:58] <robru> kenvandine, what do you mean by 'timings under 10ms'. timings of what? redraw speed?
[17:59] <kenvandine> it runs up over 100%
[17:59] <kenvandine> i think it's the total time to create each object
[17:59] <kenvandine> and it does that as it scrolls
[17:59] <kenvandine> you can increase the cache there
[17:59] <robru> ohhhh, object creation times
[17:59] <kenvandine> but it doesn't really help
[17:59] <kenvandine> it only creates the objects that are on screen
[17:59] <robru> right right
[17:59] <kenvandine> you can make it buffer more
[17:59] <kenvandine> but...
[18:00] <kenvandine> when you scroll, it still destroys and creates objects
[18:00] <robru> buffering more won't make them load any faster, just... sooner
[18:00] <kenvandine> it just does it further away from the view
[18:00] <robru> yeah
[18:00] <kenvandine> i'm amazed that it takes just as long with an empty flickable
[18:00] <robru> yeah, that's awful
[18:01] <robru> that can't be right, can it? how can it be just as slow if it's not creating anything?
[18:01] <kenvandine> dragging it around and keeping it pegged doesn't actually seem to make it noticably slow down though
[18:01] <kenvandine> good question...
[18:01] <kenvandine> but the graphs pegs constantly while it's flicking
[18:01] <kenvandine> but yeah, it's not creating objects...
[18:01] <kenvandine> humm
[18:02] <kenvandine> maybe that time isn't just creation
[18:02] <kenvandine> but also drawing
[18:03] <kenvandine> but regardless of the time that's reported
[18:03] <kenvandine> it does peg the cpu
[18:03] <kenvandine> with no contents
[18:03] <robru> yeah
[18:04] <kenvandine> now i wish i had started with an empty flickable :)
[18:04] <kenvandine> it took a long time to step backward from friends-app :)
[18:04] <robru> kenvandine, you should have bisected!
[18:04] <robru> ;-)
[18:05] <kenvandine> yeah... those tiles are so complex, i just assumed it had to do with anchoring and sizing stuff
[18:05] <robru> yeah, it seems reasonable to me. also because when I tried to reproduce this with calendar-app, it only went 50% CPU, not 102%. so I wonder if maybe they do something differently that's more efficient
[18:06] <kenvandine> interesting... i hadn't thought about that
[18:12] <kenvandine> http://paste.ubuntu.com/7324148/
[18:12] <kenvandine> robru, you can try that out
[18:12] <kenvandine> drag around in the black part of the left side of the screen
[18:12] <kenvandine> and watch top
[18:13] <robru> ok, one sec
[18:28] <robru> kenvandine, on my desktop, yeah, I see that consistently at 17ms
[18:28] <robru> trying on phone...
[18:29] <robru> kenvandine, hrm, how would I launch a simple qml file on the phone? qmlscene doesn't seem to do it
[18:36] <kenvandine> robru, build a click out of it
[18:36] <kenvandine> i didn't try on my phone
[18:37] <kenvandine> but comparing the test to friends-app on my desktop
[18:37] <kenvandine> they are about the same
[18:37] <kenvandine> so apples to apples :)
[18:37] <robru> ugh, I have to make a whole click? I can't just trick upstart-app-launch into running a qml file directly?
[18:37] <kenvandine> nope, needs to be registered with click for upstart to find it
[18:37] <kenvandine> want me to create one?
[18:37] <robru> kenvandine, yeah, it's just that the CPU usage on my quad-core hyperthreaded i7 is like 5% (even for friends-app) so it's tough to say if it's going 102% on the device or not without testing on the device ;-)
[18:38] <robru> kenvandine, if you can, please. I'm not sure how
[18:38] <kenvandine> on my dual core 2.93G desktop it goes over 100%
[18:39] <robru> kenvandine, oh, I get 30%, excuse me
[18:40] <robru> kenvandine, actually your pastebin gets me 8% CPU while friends-app does 30%, so I guess there is something in friends-app...
[18:43] <kenvandine> http://ubuntuone.com/4WQ1nkDWYxMyJ03NyMmc8N
[18:43] <kenvandine> robru, friends-app does do quite a bit more, so makes sense for it to go higher
[18:44] <kenvandine> i guess for me it's hard to go much higher than 100% :)
[18:45] <robru> right, but the point is that it goes to 100% on the device when other apps don't ;-)
[18:45] <robru> ok, I managed to install that click, thanks, how do I run it? it's not in the app list ;-)
[18:46] <kenvandine> search
[18:46] <kenvandine> it'll show up
[18:46] <kenvandine> interesting
[18:46] <robru> doesn't show up under "ken" or "listview"
[18:46] <kenvandine> that performs way better
[18:47] <kenvandine> i search for "lis"
[18:47] <kenvandine> and it showed up
[18:47] <kenvandine> you installed with pkcon right?
[18:47] <robru> uh "click install *.click" ;-)
[18:47] <kenvandine> and not as root or anything
[18:47] <robru> perhaps as root ;-)
[18:47] <kenvandine> pkcon install-local *.click
[18:47] <kenvandine> as phablet
[18:48] <robru> oh ok
[18:49] <robru> kenvandine, yeah I'm getting 30-50% CPU on device
[18:50] <robru> kenvandine, can you try friends-app just with avatar rendering turned off? I wonder if that's doing it
[18:56] <robru> kenvandine, as I recall, we had trouble with that in the past ;-)
[19:00] <kenvandine> yeah, it's going to make profiling this much more painful