[07:38] <tsdgeos> Saviq: ok so tests are now run again, and obviously they don't pass again :D
[07:39] <Saviq> tsdgeos, that explains things ;)
[07:39] <tsdgeos> Saviq: fginther: and it seems we probably want to increase the 90 min limit?
[07:39] <tsdgeos> https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-utopic/59/consoleFull
[07:39] <Saviq> tsdgeos, probably, yeah
[07:43] <fginther> tsdgeos, Saviq, I've bumped it to 180 minutes and stated another test
[07:43] <Saviq> fginther, thanks!
[07:47] <tsdgeos> paulliu: you working on fixing the test errors regarding zoom stuff at https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-utopic/59/consoleFull ?
[07:59] <paulliu> tsdgeos: yes
[07:59] <tsdgeos> cool :)
[08:00] <paulliu> tsdgeos: Can you review it?
[08:01] <paulliu> tsdgeos: oh, sorry, I saw your comments.
[08:01] <paulliu> tsdgeos: I'll fix it.
[08:03] <Cimi> Saviq, pete-woods ping
[08:03] <pete-woods> Cimi: hi
[08:03] <Cimi> morning
[08:03] <Saviq> Cimi, wassup?
[08:04] <Cimi> Saviq, I was chatting yesterday with pete
[08:04] <Cimi> Saviq, basically with the current file infographics I only need to use a crossfadeImage
[08:05] <Cimi> Saviq, so we were wondering if we really need a model or not
[08:05] <pete-woods> Cimi, Saviq: http://paste.ubuntu.com/7500579/
[08:05] <Saviq> Cimi, it's safer future-wise
[08:05] <Cimi> mzanetti, if you want to join us again
[08:05] <pete-woods> could have that API?
[08:05] <pete-woods> it has the advantage that I know which infographic you looked at last, so I can save it
[08:05] <Saviq> pete-woods, I don't think the backend should have a notion of "current infographic"
[08:05] <pete-woods> (a requirement from design)
[08:06] <tsdgeos> Mirv: can you update/rebuild/something libunity-scopes-dev in the beta2 ppa?
[08:06] <Cimi> Saviq, issue with the model is that I need to use repeater or listview, which is not efficient on UI side
[08:06] <pete-woods> Saviq: it's not technically the backend, it's just a QML plugin that talks to the backend
[08:06] <pete-woods> "talks" = not really talks
[08:06] <Saviq> pete-woods, sure, but then if someone decides (like it was planned before) that we'll be swiping between infographics again
[08:07] <pete-woods> Saviq: okay, sure that's fair
[08:07] <Saviq> pete-woods, the API breaks
[08:07] <pete-woods> Saviq: obviously I would really prefer to do nothing :)
[08:07] <pete-woods> but I want to make Cimi's job easy, too
[08:07] <Saviq> Cimi, why do you say "repeater or listview are inefficient"? just for the sake of talking to the model I think that's fine
[08:08] <Saviq> Cimi, the delegate would just be a QtObject
[08:08] <Cimi> Saviq, so a listview with empty delegates?
[08:08] <Cimi> ah ok
[08:08] <Cimi> and the flickable?
[08:08] <mzanetti> Cimi: repeater, not listview
[08:08] <Saviq> Cimi, why a flickable at all?
[08:08] <Saviq> Cimi, and yeah, what mzanetti said
[08:08] <Cimi> Saviq, listview has one
[08:08] <Mirv> tsdgeos: isn't it already?
[08:08] <Mirv> unity-scopes-api
[08:09] <Saviq> mzanetti, one thing ListView has over Repeater, though, is currentIndex/Item
[08:09] <mzanetti> Saviq: but I guess in Cimi's case Repater.itemAt() is even more useful
[08:09] <tsdgeos> Mirv: it has 0.4.5 when utopic has 0.4.7
[08:09] <Mirv> tsdgeos: mm, weird
[08:09] <Cimi> mzanetti, yup
[08:09] <Saviq> mzanetti, yeah, I mean rather that if an index goes away from a Repeater, you don't know
[08:09] <tsdgeos> Mirv: what's even weirder is that i don't get the new version on dist-upgrade :S
[08:10] <Saviq> mzanetti, whereas in a ListView, it deals with it somehow
[08:10] <mzanetti> right...
[08:10] <Saviq> and you have currentItem, too
[08:10] <Mirv> tsdgeos: well of course because beta2 recipes have the epoch, to prevent that you'd suddenly get a non-rebuilt-against-qt5.3 version
[08:11]  * tsdgeos pretends he understands that ^_^
[08:11] <Mirv> tsdgeos: I'm trying to open https://code.launchpad.net/~canonical-qt5-edgers/+recipe/daily-qt53-unity-scopes-api ... LP just sucks with recipe pages
[08:11] <Mirv> tsdgeos: hehe, the "1:" in front of the version
[08:11] <Cimi> Saviq, currentItem relative to the user interaction on the flickable
[08:12] <Cimi> I would not add a listview with all it comes from it just to get a model working
[08:12] <Cimi> infographiocs with the current state should just be a crossfadeimage and one property with the url
[08:13] <Saviq> Cimi, again, model is just future proof, and you're just throwing words "listview is inefficient" without even knowing what that means
[08:13] <Saviq> pete-woods, do you ever update the model or replace it whole every time?
[08:14] <pete-woods> Saviq: only on first load
[08:14] <pete-woods> after that I only update the entries that actually change
[08:14] <pete-woods> I tried to make it a proper list model
[08:14] <Cimi> Saviq, it definitely adds a lot of stuff I don't need
[08:14] <pete-woods> so your QML can be nice and efficient
[08:14] <Cimi> Saviq, so we can call that a not efficient solution
[08:15] <Cimi> Saviq, also, you really expect to swipe between infographics?
[08:15] <Cimi> Saviq, we also swipe to unlock the lockscreen
[08:15] <pete-woods> I think that'd be much nicer (vertical swipe)
[08:15]  * mzanetti expect design to come up with something new in the infographic at least twice per year
[08:15] <Saviq> Cimi, that was the design until we got "swipe the whole greeter away"
[08:15] <Cimi> that could collide UX wise
[08:16] <Saviq> Cimi, that's the *only* reason why we're not swiping the infographics now
[08:16] <Saviq> I mean in what you're implementing now
[08:16] <Cimi> Saviq, I have in mind the old design
[08:16] <Saviq> pete-woods, Cimi, ok, let's just add the good 'ol Q_INVOKABLE get(int index)
[08:16] <Cimi> Saviq, when the infographics were like part of the dash?
[08:17] <Saviq> Cimi, huh?
[08:17] <pete-woods> Saviq: I've already got that method
[08:17] <Cimi> Saviq, ok I am talking of the very first design...
[08:17] <Saviq> Cimi, so you're talking tiles
[08:17] <Saviq> Cimi, not infographics
[08:17] <Saviq> pete-woods, Q_INVOKABLE is it?
[08:17] <pete-woods> yes
[08:17] <Cimi> Saviq, they were infographics 3 years ago :)
[08:18] <pete-woods> Saviq: this was at mzanetti's request :)
[08:18] <Cimi> Saviq, issue is that I call get
[08:18] <Cimi> Saviq, and the model is not ready
[08:18] <Cimi> Saviq, like
[08:18] <Cimi> Saviq, http://paste.ubuntu.com/7497997/
[08:19] <Cimi> Saviq, source is empty
[08:19] <Saviq> Cimi, yeah, which is why it's better to use something that understands updates on the model
[08:19] <Cimi> Saviq, and I agree with that
[08:20] <Saviq> Cimi, but I don't think moving it down to the plugin is the right thing to do
[08:20] <Cimi> Saviq, well, I will trust you who know those things better than me
[08:20] <Cimi> Saviq, let me try with repeater then
[08:20] <Mirv> tsdgeos: aaargh
[08:21] <tsdgeos> Mirv: wha?
[08:21] <Mirv> tsdgeos: unity-scopes-api does not auto-build becaause of " Missing build dependencies: apparmor:native "
[08:21] <Mirv> tsdgeos: and that's a feature of the recipes part of LP. I hope the ":native" build dependencies don't spread too much
[08:21] <Saviq> Cimi, yeah, remember you need to react to onCountChanged and check if your current index still "fits" in the model and such
[08:22] <Mirv> tsdgeos: so, I'll do a manual upload
[08:22] <Cimi> Saviq, we don't have count on the plugin, though
[08:23] <Cimi> Saviq, is count a qml property generated by repeater. right?
[08:23] <Saviq> Cimi, http://qt-project.org/doc/qt-5/qml-qtquick-repeater.html#count-prop
[08:23] <Cimi> I know that pro
[08:23] <mhr3> Mirv, Saviq wanted to look into it :)
[08:23] <Cimi> just wondering if it doesn't need that in the plugin too
[08:24] <Saviq> Mirv, mhr3, we can revert the :native from -scopes-api, apparmor will be fixed to cope...
[08:24] <mhr3> Saviq, what about python3?
[08:24] <Saviq> mhr3, :any
[08:24] <Mirv> Saviq: phew, nice. getting LP fixed would the best though.
[08:24] <Saviq> Mirv, indeed
[08:24] <Mirv> it only complains about the :native stuff
[08:24] <mhr3> Saviq, also we're moving to libapparmor
[08:25] <Saviq> Mirv, we did have that problem in jenkins, and fginther fixed it
[08:25] <mhr3> should workaround that
[08:25] <Saviq> mhr3, nice
[08:25] <Saviq> Mirv, so maybe it wouldn't be that difficult to update LP
[08:25] <Saviq> pete-woods, Cimi, now, about remembering the last selected one... obviously without domain knowledge (what _is_ the one that's displayed now)
[08:26] <Saviq> pete-woods, Cimi, we can't reliably remember which one to display :|
[08:26] <Cimi> nope
[08:26] <Mirv> Saviq: maybe, there are just so few people working on LP nowadays
[08:26] <Cimi> we need like an index
[08:26] <Cimi> too
[08:26] <Mirv> tsdgeos: ok, new unity-scopes-api building, should be available in 30+ mins or so
[08:26] <tsdgeos> tx
[08:27] <Cimi> so my starting index of the listview/repeater would be that index when we first launch unity
[08:27] <Saviq> Cimi, "like an index"?
[08:27] <Cimi> Saviq, the saved index
[08:27] <Saviq> Cimi, yeah, but index isn't good enough
[08:27] <mhr3> Saviq, btw i saw some oddity with u8 yesterday, it stopped rendering and was using 100% cpu
[08:27] <Saviq> Cimi, since the order might change
[08:28] <mhr3> Saviq, also, not reacting to any input
[08:28] <Cimi> Saviq, indeed and this index in the model should point at the correct one, no?
[08:28] <pete-woods> Saviq: indeed, but if you tell me to save an index, I can convert that to something persitent internally, though
[08:28] <Saviq> mhr3, not a great bug report
[08:28] <Saviq> ;)
[08:29] <Cimi> mhr3, you should remove your plastic cover from the display, might help on this 100% cpu usage
[08:29] <mhr3> Saviq, yea, sorry, things spinning in a loop are pain to debug
[08:29] <Saviq> pete-woods, yeah, but at that point the argument of moving more logic down to the plugin becomes more valid
[08:29] <mhr3> Cimi, :P
[08:29] <mhr3> Cimi, my work phone doesn't have it, you stripped it yourself
[08:30] <Cimi> Saviq, otherwise I save the content of the last url of the svg, and loop in the model to find it?
[08:30] <Cimi> Saviq, or maybe we can just add a search method to the model?
[08:31] <Saviq> Cimi, I have no strong opinion there, but yeah something of the sort
[08:31] <Saviq> pete-woods, Cimi, I'll let you sort this one out
[08:32] <fginther> Saviq, to be fair, the 'fix' was to strip ':native' since jenkins is already doing a native build
[08:32] <Saviq> fginther, oh ok...
[08:33] <pete-woods> Saviq: I have no problem switching to the API I proposed earlier
[08:33] <Saviq> pete-woods, I wonder if we could combine them then
[08:34] <pete-woods> Saviq: maybe, but that would probably start to make it convoluted
[08:34] <pete-woods> when we have a concrete use case right now
[08:35] <pete-woods> I'd rather add complexity when it's actually necessary
[08:35] <Saviq> pete-woods, I know... I know... we can always go bzr revert at some point can't we...
[08:35] <pete-woods> yep
[08:37] <Saviq> pete-woods, well, more work for you, then... and I'd actually still make the "new" api wrap the model anyway, but that's your call :)
[08:37] <pete-woods> Saviq: if we need the old model, I'd just copy and paste the original implementation and have it as a separate class, most of the work is already done by a child class anyway
[08:38] <Saviq> pete-woods, mhm
[08:38] <Saviq> Cimi, looks like the "remember last one shown" tipped the scale towards the simplified API
[08:39] <Cimi> yay
[08:39] <pete-woods> Cimi: I already have an implementation of the simple API for you, just fixing the unit tests
[08:43] <tsdgeos> guys, so i have this
[08:43] <tsdgeos> http://paste.ubuntu.com/7500659/
[08:43] <tsdgeos> where Department is a Qobject
[08:43] <tsdgeos> i was under assumption that JavaScript would delete the Department once i didn't reference to it in qml anymore
[08:44] <tsdgeos> but it's not happening
[08:44] <tsdgeos> what i am doing wrong?
[08:44] <Saviq> tsdgeos, nothing, garbage collection needs to kick in
[08:45] <tsdgeos> hmmm, so you say it's not getting deleted because the GC didn't run?
[08:45] <Saviq> tsdgeos, I'd expect so
[08:45] <tsdgeos> correct
[08:45] <Saviq> tsdgeos, QML deals a bit different, though
[08:45] <tsdgeos> after doing it a lot
[08:45] <Saviq> tsdgeos, if there's a delegate that gets destroyed, this should get deleted with it
[08:45] <tsdgeos> i got a bunch of "Department destroyed"
[08:46] <tsdgeos> so even the delegate was gone
[08:46] <tsdgeos> the var that the delegate referenced stayed in the limbo for a while
[08:46] <Saviq> tsdgeos, yeah, that's unfortunately possible with GC, and I don't think we can do anything about it, other than call into C++ to destroy it on delegate destruction
[08:47] <tsdgeos> ok, it's ok, it's not like we're going to have zillions of memory lost there
[08:48] <Saviq> tsdgeos, we could have a signal on Department that would be connected to in C++ and destroy it... stuff like that
[08:49] <Saviq> mhr3, but I believe Department objects will actually be rather light (proxies to a tree), mhr3?
[08:49] <Saviq> tsdgeos, ↑ I meant
[08:49] <tsdgeos> yeah
[08:51] <Saviq> tsdgeos, for reference http://qt-project.org/doc/qt-5/qqmlengine.html#ObjectOwnership-enum
[08:51] <tsdgeos> Saviq: yes i was reading that
[08:51] <tsdgeos> but then it wasn't doing what i expect it to do :D
[08:51] <Saviq> tsdgeos, well, according to that they will only get removed on engine deletion... :|
[08:52] <tsdgeos> i just expected the GC to kick in somewhat more often
[08:52] <tsdgeos> hmm, what?
[08:52] <mhr3> Saviq, it will have to dup the data, but still should be light-ish
[08:52] <tsdgeos> Saviq: that's the "An object tracked by one QQmlEngine" part
[08:52] <Saviq> "An object tracked by one QQmlEngine will be deleted during that QQmlEngine's destructor"
[08:53] <tsdgeos> yes, but if there's no references i assume it's not tracked by anyone
[08:53] <tsdgeos> no?
[08:53] <Saviq> hmm
[08:53] <Saviq> "QML will track it, and delete the object if there are no remaining JavaScript references to it and it has no QObject::parent()."
[08:53] <tsdgeos> i mean i'm getting the thing destroyed
[08:53] <Saviq> that doesn't, unfortunately say *when* it's going to be deleted... but maybe it has to be like that
[08:53] <Saviq> tsdgeos, ah ok, just after some time?
[08:54] <tsdgeos> yep
[08:54] <Saviq> tsdgeos, then yeah, GC explains
[08:54] <tsdgeos> http://paste.ubuntu.com/7500687/
[08:54] <tsdgeos> these qdebugs are all from the c++ side in constructor/destructor
[08:55] <Saviq> yeah sounds right
[08:55]  * greyback heard of a "gc()" function that's supposed to force trigger the garbage collector
[08:57] <tsdgeos> mhr3: any ETA for the review of https://code.launchpad.net/~aacid/unity-scopes-shell/use-unity-api/+merge/219221 ?
[08:57] <tsdgeos> Saviq: or you ↑
[08:57] <mhr3> tsdgeos, opened it a minute ago :)
[08:57] <tsdgeos> ah
[08:57] <tsdgeos> nice :)
[08:58] <Saviq> greyback, indeed this seems mentioned here and there, can't find actual docs though
[08:59] <greyback> Saviq: yeah, nor have I, but did see it in Qt Components at one stage - in their PagStack they called it to immediately release resources held by a popped page
[09:07] <mhr3> tsdgeos, added comments
[09:08] <Saviq> greyback, hah! http://qt-project.org/doc/qt-5/qtqml-javascript-qmlglobalobject.html
[09:08] <Saviq> greyback, 3rd bullet point
[09:08] <greyback> Saviq: good google-fu :)
[09:08] <Saviq> terse
[09:08] <tsdgeos> mhr3: tx
[09:09] <Saviq> greyback, grep over qtdeclarative rather ;)
[09:09] <greyback> :)
[09:11] <Saviq> tsdgeos, so yeah, if we decide we want to, we can call gc() after delegate destruction and the object should be dropped then and there
[09:11] <tsdgeos> ok
[09:13] <tsdgeos> mhr3: answred them
[09:16] <mhr3> tsdgeos, hm, i see, can you add a comment to cmakelists about the moc issue?
[09:16] <tsdgeos> mhr3: sure
[09:19] <tsdgeos> mhr3: i guess we could run moc on the other side of the lib, but that's how we do it for the other unity-api stuff so i mimic'ed it
[09:19] <tsdgeos> mhr3: comment added
[09:20] <mhr3> tsdgeos, would require installing autogenerated moc... /me not likey
[09:21] <tsdgeos> Saviq: mhr3: silo!
[09:24] <mhr3> ah, right, we're not landing anything
[09:24] <Cimi> pete-woods, tell me when you push
[09:25] <pete-woods> Cimi: will do
[09:25] <Cimi> pete-woods, even without unit tests
[09:25] <Cimi> pete-woods, I can work in the meanwhile :)
[09:28] <pete-woods> Cimi: sure, but I want to know it works
[09:28] <pete-woods> I found it crashed on me once
[09:28] <pete-woods> Cimi: don't worry, I'll be quick
[09:28]  * Cimi makes a coffee in the meanwhile :)
[09:31] <pete-woods> maybe not *that* quick!
[09:37] <sil2100> Saviq: hi! Do you know how's the split greeter going?
[09:38] <sil2100> Saviq: since there's a landing I wanted to assign that has indicator-network in it, can I ignore-conflicts or is this split ready for release soon?
[09:51] <Mirv> tsdgeos: unity-scopes-api up-to-date
[09:53] <Mirv> tsdgeos: unity-api is out of date though (fails tests on armhf, would need manual upload every time), and so is unity-scopes-shell (FTBFS since the last upload earlier this week, bug filed)
[09:55] <tsdgeos> oki
[09:58] <Saviq> sil2100, we're still a few issues away
[09:58] <Saviq> sil2100, so ignore it please and we'll manage
[10:01] <mhr3> Mirv, your fault :P https://bugs.launchpad.net/unity-scopes-shell/+bug/1321688/comments/1
[10:02] <Mirv> mhr3: oh so it's this same problem, that's actually great news! so the bug is invalid.
[10:02] <Mirv> so I'll just kick rebuild now 0.4.7 is there manually
[10:03] <Mirv> mhr3: no my fault, Launchpad's fault :)
[10:03] <mhr3> Mirv, sure, sure ;P
[10:03] <Mirv> so, unity-api only missing probably last update because it always fails on armhf + i386 (but not amd64...)
[10:04] <Saviq> thostr_, mhr3, I added a session on app installation progress in the app store/scope, who should be there? I added Rodney and you two from your side
[10:05] <mhr3> Saviq, +alecu & pstolowski
[10:05] <Saviq> done
[10:09] <pete-woods> Cimi: okay, have pushed the branch now
[10:09] <Cimi> thx¬
[10:09] <Cimi> !
[10:10] <pete-woods> Cimi: http://paste.ubuntu.com/7500892/ <— some example QML
[10:10] <Saviq> pete-woods, shall I rebuild the silo?
[10:10] <pete-woods> Saviq: sure, can do that too
[10:10] <pete-woods> I thought we'd got flushed
[10:10] <Cimi> pete-woods, awesome!
[10:11] <Saviq> pete-woods, indeed we did
[10:11] <Saviq> pete-woods, we're low on silos still, so let's do it when we're ready
[10:11] <pete-woods> that works for me
[10:12] <Cimi> I don't need it for testing it
[10:12] <Cimi> but phone would be good
[10:12] <Cimi> Saviq, I guess crossbuild doesn't work yet?
[10:12] <Cimi> tried yesterday and got issues
[10:12] <Saviq> Cimi, you'd have to manually install build deps, it works then
[10:13] <Cimi> Saviq, so I chroot and install deps?
[10:13] <Cimi> in sbuild
[10:13] <Saviq> Cimi, yeah, instead of relying on sbuild to resolve the deps
[10:13] <Saviq> Cimi, schroot -c foo-amd64-armhf -u root
[10:13] <Saviq> Cimi, apt-get build-dep -aarmhf unity8
[10:13] <pete-woods> Cimi: a quick warning, I think you'd going to find the CrossFadeImage looks a bit crappy
[10:13] <pete-woods> at least it did for me
[10:14] <Cimi> pete-woods, in which way?
[10:14] <Saviq> Cimi, and resolve anything it complains about
[10:14] <Cimi> pete-woods, I have a local version where I added transitions
[10:14] <pete-woods> Cimi: maybe you have improved it, but for me the transitions were pretty bad, like both images seemed to have 100% opacity at one point
[10:14] <Cimi> pete-woods, fixed thay
[10:14] <pete-woods> I think perhaps it's never been tested with transparent images?
[10:15] <pete-woods> Cimi: ah, cool, should be good then!
[10:15] <Cimi> Saviq, are you in Malta?
[10:15] <Saviq> Cimi, not until Saturday
[10:15] <Cimi> Saviq, good for you
[10:15] <Cimi> Saviq, I wanted to ask sdk guys how to name the property
[10:15] <Saviq> Cimi, so bad? ;)
[10:15] <Cimi> Saviq, for the crossfadeimage
[10:16] <Cimi> Saviq, in the meanwhile I have
[10:16] <Cimi> pastebinit qml/Components/CrossFadeInfographics.qml
[10:16] <Cimi> http://paste.ubuntu.com/7500920/
[10:16] <Saviq> /food
[10:17] <Saviq> Cimi, I'd go for "fadeStyle" or something - will allow supporting more modes later
[10:17] <Cimi> diff -urN ~/Development/sdk/ubuntu-ui-toolkit/modules/Ubuntu/Components/CrossFadeImage.qml qml/Components/CrossFadeInfographics.qml | pastebinit
[10:17] <Cimi> http://paste.ubuntu.com/7500923/
[10:18] <Cimi> Saviq, ok
[10:22] <boiko> hi, any uoa devs around?
[10:27] <boiko> mardy: hey, do you know of any changes released lately that could be breaking telepathy-ofono account creation on desktop and device?
[10:28] <Mirv> unity-scopes-shell built now too in qt5-beta2 PPA
[10:29] <boiko> mardy: if I create the account with no parameters, it works fine, but as soon as I try to set modem-objpath it fails saying ofono/ofono doesn't have such property
[10:32] <Mirv> tsdgeos: I also now pushed the newest version of unity-api to build with tests disabled, in case it's needed. clearly at least the unity-scopes-shell was required to be together with new unity-scopes-api.
[10:32] <tsdgeos> ty
[10:36] <mhr3> mzanetti, seen the unity-api test failure?
[10:36] <mzanetti> mhr3: no. looking...
[10:37] <mhr3> mzanetti, https://launchpadlibrarian.net/175993802/buildlog_ubuntu-utopic-armhf.unity-api_1%3A7.81%2B14.10.20140520-0~134%2B201405201819~ubuntu14.10.1_FAILEDTOBUILD.txt.gz
[10:37] <mzanetti> uh... what the...
[10:37]  * mzanetti fixing
[10:37] <mhr3> mzanetti, 5.3
[10:37] <mzanetti> ah
[10:38] <mzanetti> mhr3: can we install that already without breaking everything?
[10:38] <mhr3> Mirv, ^
[11:04] <Cimi> pete-woods, you mixed tabs and spaces in the modified files :)
[11:04] <Cimi> InfographicsList.h
[11:04] <pete-woods> Cimi: thanks for pointing that out! I'm switching between two different projects styles :)
[11:05] <Cimi> pete-woods, I personally prefer spaces in the end
[11:05] <Cimi> but here everything is tabs
[11:05] <pete-woods> Cimi: okay, I've fixed that, thanks for noticing
[11:06] <pete-woods> I only really care about consistency
[11:06] <Cimi> Saviq, so for our tests, I need to mock InfographicList with this new API, right?
[11:06] <Cimi> I don't think I can do this in js
[11:07] <mzanetti> mhr3: argh. can't reproduce there here with 5.3
[11:07] <pete-woods> can't you create a new QML type with a property and a slot?
[11:08] <mhr3> mzanetti, :(
[11:09] <Cimi> pete-woods, mmm
[11:10] <Cimi> pete-woods, I think is indeed possible
[11:10] <Cimi> basically wrapping
[11:10] <Cimi> a listmodel
[11:10] <Cimi> let's see
[11:10] <pete-woods> yeah, something like that
[11:10] <Cimi> haven't thought of that, thanks
[11:21] <Saviq> Cimi, you don't even need to wrap a ListModel, an array would be good enough
[11:21] <Cimi> Saviq, I did this
[11:22] <Cimi> Saviq, http://paste.ubuntu.com/7501064/
[11:22] <Cimi> Saviq, forget the next() function and the limited svgs I have now
[11:22] <Cimi> Saviq, this way I have model.path and model.next()
[11:24] <Saviq> Cimi, http://paste.ubuntu.com/7501070/
[11:24] <Saviq> Cimi, it should be property *url* path actually
[11:24] <Cimi> ok cppp
[11:24] <Cimi> cool
[11:28] <Cimi> Saviq, TypeError: Cannot read property 'path' of undefined
[11:28] <Mirv> mhr3: mzanetti yes, desktop (utopic) less tested though but QtC should continue to work
[11:29] <Cimi> forget me
[11:29] <mzanetti> Mirv: thanks
[11:29]  * Cimi hides
[11:29] <Saviq> Cimi, I will
[11:29] <Saviq> Cimi, really, IRC debugging is not useful :P
[11:29] <Mirv> instructions at https://launchpad.net/~canonical-qt5-edgers/+archive/qt5-beta2 - I should write to the mailing list as well
[11:29] <Cimi> Saviq, I'm trying to hide!!!!! :P
[11:29] <Mirv> note the packages are "ugly" but nothing that should directly concern you. I'm working in parallel (and much slower) on "proper" packaging
[11:30] <Saviq> tsdgeos, looks like we got uitests hanging on jenkins :|
[11:30] <Saviq> tsdgeos, https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-utopic/61/console
[11:30] <Saviq> tsdgeos, 3h is definitely too long
[11:30] <tsdgeos> wow
[11:31] <tsdgeos> aaaaaaaaaaactually it seems to hang here too :S
[11:31] <tsdgeos> well it crashed :S
[11:35] <tsdgeos> Saviq: can you see if it crashes for you? i'm not sure if this crash is becasue 5.3
[11:35] <tsdgeos> which is bad in itself
[11:35] <Saviq> tsdgeos, just all tests?
[11:35] <tsdgeos> ./tests/plugins/Dash/OrganicGridTestExec
[11:35] <tsdgeos> Saviq: ↑
[11:36] <Saviq> tsdgeos, yup, fatal 11
[11:36] <Saviq> tsdgeos, in delegate creation ranges...
[11:38] <tsdgeos> that's 5.2, no?
[11:40] <tsdgeos> Saviq: so what do we do? i go back to 5.2 and fix it? or given we don't use the organicgrid anywhere yet, just disable it? or?
[11:40] <Saviq> tsdgeos, yeah, I'm on 5.2... and we'll be enabling the other layouts asap, too
[11:41] <Saviq> tsdgeos, so...
[11:41] <tsdgeos> Saviq: ok, i guess the questions is we want to fix it in 5.2 or in 5.3?
[11:42] <Saviq> tsdgeos, 5.3
[11:42] <Saviq> tsdgeos, http://paste.ubuntu.com/7501125/
[11:43] <Saviq> died in GC ;|
[11:43] <tsdgeos> yep
[11:43] <Saviq> wonder if this is the same Olivier saw
[11:43] <Cimi> Saviq, tsdgeos I have different rendering with qml and imageviewer/inkscape
[11:44] <tsdgeos> Cimi: of course, i told you not to use SVG :D
[11:44] <Saviq> Cimi, yeah, qt doesn't render svgs correctly
[11:44] <tsdgeos> the svg render of Qt only renders in 1 pass
[11:44] <tsdgeos> so can't do lots of stuff
[11:44] <Saviq> Cimi, we'll have to deal with that for now, and render them with librsvg in the image cacher
[11:46] <Cimi> Saviq, but I receive svgs...
[11:47] <Saviq> Cimi, yes, image provider will then render them into bitmaps and cache them
[11:48] <Saviq> Cimi, until then, we'll have to live with not-so-great renders
[11:48] <Cimi> Saviq, call them broken
[11:48] <Saviq> Cimi, as we do in the dash for quite some time now
[11:48] <Cimi> Saviq, missing elements
[11:48] <Saviq> bug #1224998
[11:49] <Cimi> Saviq, if eog is using librsvg
[11:49] <Cimi> it is not good either
[11:49] <Cimi> I have different results inkscape/eog
[11:50] <tsdgeos> Saviq: it may actually be an infinite loop somewhre in the new delegate range code, checking
[11:50] <Saviq> Cimi, no idea what eog renders with
[11:51] <Saviq> Cimi, but we need to decide on one that we say is the right one, and afaict librsvg was the one recommended implementation
[11:52] <Cimi> Saviq, http://i.imgur.com/2ZMFNm8.png
[11:53] <Saviq> Cimi, heh, that's bad ;)
[11:53]  * tsdgeos lunches
[11:53] <Saviq> Cimi, in any case, we know what the *target* renderer is
[11:53] <Cimi> Saviq, inkscape uses cairo iirc
[11:54] <Saviq> Cimi, so people that write the visualizers are expected to verify their SVGs work well in Qt/QML
[11:54] <pete-woods> Saviq, Cimi: that's exactly what I did, yes
[11:54] <Cimi> pete-woods, ok
[11:54] <pete-woods> the gnome render seemed pretty weak at rendering SVGs
[11:54] <Cimi> pete-woods, this is an inkscape edited svg
[11:54] <pete-woods> I just made sure it worked well in Qt
[11:54] <pete-woods> oh, okay
[11:55] <pete-woods> you edited one of the ones that my program generated?
[11:56] <Cimi> pete-woods, yes
[11:56] <pete-woods> cool
[11:56] <pete-woods> glad the inkscape was happy with it
[11:56] <pete-woods> *that
[11:56] <Cimi> pete-woods, not qt though :)
[11:56] <pete-woods> ah
[11:57] <pete-woods> the SVGs I generate are pretty clean
[11:57] <Cimi> pete-woods, I don't know how people are supposed to create those infographics svgs
[11:57] <pete-woods> you could always tweak them by hand
[11:57] <pete-woods> Cimi: well realistically it's "write a script"
[11:57] <pete-woods> but yes, I understand completely what you're saying
[11:58] <Cimi> what's the point of using svgs if to edit them I need to write code?
[11:58] <pete-woods> Cimi: agreed
[11:58] <Cimi> we could have just used qml or css
[11:58] <pete-woods> yep
[11:58] <Saviq> Cimi, no we couldn't
[11:58] <Cimi> Saviq, ok for security
[11:58] <Saviq> exactly
[11:59] <pete-woods> it's still a completely rubbish development experience for someone who wants to write an infographic
[11:59] <Cimi> I agree
[11:59] <Saviq> pete-woods, Cimi, you're free to come up with something better
[11:59] <mzanetti> mhr3: https://code.launchpad.net/~mzanetti/unity-api/fix-launcher-tests-with-5.3/+merge/220623
[12:00] <pete-woods> Saviq: I think even a library to help write SVGs would help
[12:00] <pete-woods> I ended up just using QXmlStreamWriter
[12:00] <Saviq> pete-woods, you can use whatever library you want
[12:00] <Cimi> Saviq, pete-woods I don't know what this script it, but it feels to me it would be easier to do other things :)
[12:00] <Cimi> it's like a new language to learn
[12:00] <Saviq> pete-woods, provided you ship it with your visualizer and there's just an executable that spits the svg out
[12:00] <Cimi> I don't expect any designer learning to code to draw an infographics
[12:01] <Saviq> Cimi, designers are not meant to code them, they are meant to design them
[12:01] <pete-woods> Cimi: http://bazaar.launchpad.net/~unity-team/libusermetrics/file-based-infographics/view/head:/src/infographic/Infographic.cpp
[12:01] <pete-woods> that's the code that makes the default infographic in-case you're interested
[12:02] <pete-woods> it's not like it's intractable, you just need to understand SVG
[12:02] <Saviq> pete-woods, in time we'll bless a library that will become part of our SDK
[12:02] <Saviq> pete-woods, until then, use whatever floats your boat
[12:02] <pete-woods> that'd go a long way towards helping
[12:02] <pete-woods> yeah, I know *I* can do whatever I want
[12:02] <pete-woods> it's just making it nice for other people I worry about
[12:03] <pete-woods> anyway, I implemented the things as I was told, and here it is! :)
[12:03] <pete-woods> just wish there was some way to embed a sandboxed lump of QML into the greeter
[12:03] <Saviq> pete-woods, I think giving people the freedom is better than to say "here's the library to use", people are quite happy with choosing to do what they want
[12:04] <pete-woods> I don't mean to say they have to be restricted
[12:04] <Saviq> aaanyway
[12:04] <pete-woods> just some help along the way, like a library with classes representing the different SVG types
[12:04] <pete-woods> yep
[12:04] <pete-woods> it's all done now
[12:04] <pete-woods> so no point discussing
[12:05] <Cimi> Saviq, the original idea of using SVGs was "so designers can draw what they want" - which is not what will happen
[12:06] <Cimi> Saviq, I understand though the security benefits now
[12:06] <Saviq> Cimi, that idea was shot down years ago
[12:06] <Saviq> Cimi, because it's just impossible
[12:27] <Mirv> Saviq: that qtubuntu patch fixes half of the world
[12:27] <Mirv> like all ~rendering issues while the crash issues remain
[12:27] <Saviq> Mirv, yup, that's expected :)
[12:35] <Saviq> Mirv, is there a bug for settings app being empty?
[12:37] <Mirv> Saviq: no there isn't, it hasn't been debugged so far as to get eg. some error output to quote
[12:37] <Saviq> Mirv, I'll do
[12:37] <Mirv> Saviq: great, thanks!
[12:38] <Mirv> it's possibly related to some of the UITK issues seen in the unit tests
[12:38] <Mirv> also, performance of Qt 5.3 in Unity 8 side-by-side is visibly better
[12:39] <Mirv> I didn't realize it before davmor2 showed his 5.2 mako and we eg. scrolled app scope at the same time
[12:41] <Saviq> Mirv, orly? nice
[12:41] <Saviq> Mirv, bug #1322159
[12:41] <Mirv> thx
[12:41] <Saviq> Mirv, maybe his 5.2 was before the scope performance improvements? ;)
[12:42] <Mirv> davmor2: were you running ~latest on your mako?
[12:43] <davmor2> Mirv: yes
[12:43] <davmor2> Mirv: r43
[12:50] <Mirv> Saviq: ^ that's what I thought too, but it was worth checking
[12:50] <Mirv> thanks dave
[13:34] <tsdgeos> Saviq: soooo, it's actually two things
[13:34] <tsdgeos> there's a bug in porting the test
[13:34] <tsdgeos> and then there's a bug in the code that makes the "new" test crash
[13:35] <Cimi> tsdgeos, Saviq can I define enums from qml?
[13:35] <Saviq> Cimi, no
[13:35] <Saviq> Cimi, go for strings
[13:35] <Cimi> Saviq, ok
[13:38] <tsdgeos> Saviq: so i'll fix both :D
[13:38] <tsdgeos> Saviq: or c++ enums, no?
[13:38] <Saviq> tsdgeos, cool you found it :)
[13:39] <Saviq> tsdgeos, it's a QML-only API (CrossFadeImage)
[13:39] <tsdgeos> i see
[13:39] <Saviq> tsdgeos, we *could* expose the enums in some global object from C++ in Ubuntu.Components, but I don't think there is something like that yet
[13:40] <Saviq> so let's see what the SDK team says
[13:41] <Cimi> #namingishard
[13:41] <Cimi> FadeOutFadeIn?
[13:42] <Cimi> Saviq, we have two images
[13:43] <Saviq> Cimi, parallel, sequential
[13:43] <Cimi> right
[13:43] <Cimi> Saviq, for the current one?
[13:43] <Cimi> one fades in
[13:44] <Saviq> Cimi, overlay?
[13:44] <Cimi> good one
[13:51] <tsdgeos> Saviq: https://code.launchpad.net/~aacid/unity8/organiccgridcrash/+merge/220651
[14:13] <Saviq> tsdgeos, we have a leftover CardHeader entry in tests/qmltests/CMakeLists.txt :/
[14:13] <tsdgeos> wops
[14:14] <tsdgeos> Saviq: i'll do a quick MR to kill it
[14:14] <Saviq> tsdgeos, tx
[14:18] <tsdgeos> Saviq: https://code.launchpad.net/~aacid/unity8/nomorecardheader/+merge/220659
[14:20] <Saviq> yikes that's bad... over an hour and it's still building??
[14:20] <Saviq> http://s-jenkins.ubuntu-ci:8080/job/unity-phablet-qmluitests-utopic/62/consoleFull
[14:21] <Saviq> jeez we need timestamps in jenkins :(
[14:23] <Cimi> Saviq, ok infographics working fine
[14:23] <Cimi> Saviq, only issue I have is testing on the desktop
[14:23] <Cimi> Saviq, where UID is 1000
[14:24] <Cimi> and we're using mocks for lightdm
[14:27] <Saviq> Cimi, testing as in actual real testing?
[14:27] <Cimi> Saviq, using on the desktop
[14:27] <Saviq> Cimi, just tweak the mocks
[14:27] <Saviq> Cimi, so that it comes up with the right UID
[14:27] <Cimi> Saviq, mocks for what?
[14:28] <Saviq> Cimi, LightDM
[14:28] <Cimi> yes but the mock isn't for the phone as well?
[14:28] <Cimi> and on the phone has different uid
[14:28] <Cimi> https://code.launchpad.net/~cimi/unity8/infographics-new-lightdm/+merge/220662 in the meanwhile
[14:29] <Cimi> I have to test it on tablet too
[14:29] <Saviq> Cimi, I didn't mean to tweak it and commit it
[14:30] <Saviq> just tweak it for testing
[14:35] <mhr3> mzanetti, ah... parentless qobjects... nasty
[14:35] <mhr3> guess they changed how gc runs
[14:35] <mzanetti> mhr3: the nasty thing is that Qt < 5.2 didn't take ownership of those, while >= 5.2 does
[14:35] <mzanetti> mhr3: not sure why we only see the crash with 5.3 now
[14:35] <mzanetti> yeah... probably
[14:36] <mhr3> mzanetti, according to docs qml takes ownership of such objects since 5.0
[14:36] <mhr3> mzanetti, at least i remember reading about it in 5.0 docs
[14:36] <mzanetti> mhr3: it probably was just buggy in 5.0
[14:36] <mzanetti> mhr3: xbmcremote started to go all wonky wiht 5.2 only
[14:36] <mhr3> or as i said.. delayed :)
[14:37] <Cimi> Saviq, but when I run unity8, it uses lightdm mock afaics
[14:38] <dednick> charles: ping
[14:38] <mzanetti> mhr3: no... with 5.0 it wasn't delayed even... I would have noticed that
[14:38] <mzanetti> unless it was delayed for days :D
[14:39] <mhr3> mzanetti, perhaps delayed until QQmlEngine was about to be destroyed ;)
[14:39] <mzanetti> Saviq: https://bugs.launchpad.net/unity8/+bug/1309162
[14:39] <mzanetti> mhr3: yeah... that could have been, yes
[14:41] <Saviq> Cimi, yes, because it can't use the actual lightdm plugin, 'cause that's only allowed for the greeter
[14:42] <Saviq> Cimi, not sure what you want to achieve?
[14:42] <Cimi> Saviq, I run unity on the desktop, lightdm mock gives me uid 1000
[14:43] <Saviq> mzanetti, tx
[14:43] <Cimi> Saviq, I run unity on the phone, 32011
[14:44] <Saviq> Cimi, you might want to tweak the mock to get the actual UID, although I wonder if we'll need a mock of the infographics backend anyway, don't think it should be available in the user session anyway (just for greeter)
[14:45] <Saviq> Cimi, but I wouldn't care much about this now
[14:45] <Saviq> Cimi, as long as you have it tested in isolation
[14:45] <Cimi> Saviq, so as it is now, should work on the phone
[14:45] <Saviq> Cimi, and verified it works for real on the phone
[14:45] <Saviq> Cimi, that's good enough
[14:46] <Cimi> I have to test it again
[14:46] <Saviq> Cimi, the "windowed" unity8 isn't our top prio
[14:46] <Cimi> works on desktop if I manually set uid in GreeterContent.qml;
[14:46] <Saviq> Cimi, and when you run it with -f, it will load the mock infographics one anyway won't it?
[14:46] <Saviq> Cimi, good enough
[14:46] <Cimi> Saviq, we removed the mock infographics
[14:47] <Cimi> I merged your branch
[14:47] <Saviq> Cimi, didn't you just write a new mock infographic backend?
[14:47] <Cimi> Saviq, I don't
[14:47] <Saviq> Cimi, after mterry lands his greeter, you won't even see the greeter when running unity8 anyway, so
[14:47] <Cimi> Saviq, for tests I wroto those two qml lines
[14:47] <Saviq> Cimi, what's that then http://paste.ubuntu.com/7501070/ ?
[14:48] <Cimi> Saviq, it's the mock in the tests
[14:48] <Saviq> Cimi, yeah, how are you passing it to GreeterContent?
[14:48] <Cimi> actually I haven't tested greeterContent
[14:48] <Cimi> Saviq, I am passing to Infographics
[14:49] <Saviq> Cimi, we need this to be an object under the same import as the original one
[14:49] <Saviq> Cimi, otherwise wherever you go "import Infographics", it will load the actual backend - we don't want that
[14:49] <Cimi> ok
[14:59] <Saviq> mhr3, do you have the :native → :any change MP'd somewhere? or shall I?
[14:59] <mhr3> Saviq, pls do, you're the ma expert :)
[14:59] <Saviq> lol
[15:00] <mhr3> Saviq, but you pete's branch as base
[15:00] <mhr3> s/you/use/
[15:00] <mhr3> Saviq, lp:~unity-api-team/unity-scopes-api/use-libapparmor
[15:02] <Saviq> mhr3, kk
[15:10] <Saviq> Cimi, once you're done with infographics, could you please look at https://bugs.launchpad.net/unity8/+bug/1322233
[15:32] <dandrader> Cimi, can you approve it now? https://code.launchpad.net/~dandrader/unity8/unusedBackgroundImage/+merge/220288
[16:17] <popey> willcooke: hello!
[16:18] <willcooke> oh hai!
[16:19] <willcooke> I'm going to try and write a scope.  Be afraid.
[16:19] <willcooke> Do I have to learn C++ first?
[16:19]  * willcooke thinks this might take longer than a weekend
[16:21] <popey> yes
[16:21] <davidcalle> willcooke, Go works too
[16:21] <popey> and maybe go in the future
[16:21] <popey> \o/ We're in the future!
[16:21] <davidcalle> willcooke, or (don't tell anyone) Python.
[16:21] <willcooke> hrm.
[16:21] <willcooke> wait wat!
[16:21] <willcooke> Python eh
[16:21] <willcooke> Ok, now we're cooking
[16:22] <willcooke> well, we're warming things up in the microwave, but it's kinda the same thing
[16:22] <davidcalle> willcooke, look at lp:ubuntu-rest-scopes's README, all you need to know
[16:22] <willcooke> thx davidcalle!
[16:23] <davidcalle> willcooke, my pleasure. BUT, these won't ever be client side scopes, only smarscopes server side.
[16:23]  * willcooke reads about smart scopes
[16:23] <davidcalle> willcooke, device side is, only Go and C++.
[16:24] <popey> i think willcooke is looking at desktop (unity 7) not device (unity 8)
[16:24] <willcooke> I'm easy
[16:24] <willcooke> I wonder if I can learn Go enough to copy&paste some examples
[16:24] <willcooke> probably
[16:24] <dandrader> greyback, with qt compositor stuff on top of latest image I'm not getting any icons in the dash, only their titles and sections
[16:24] <dandrader> greyback, are you getting the same?
[16:25] <greyback> dandrader: same here, yes
[16:25] <davidcalle> popey, willcooke, aaaah... Then it's C++. Go scopes won't work for Unity7.
[16:26] <dandrader> wonder if rebasing on top of latest lp:unity8 will be of any help
[16:26] <davidcalle> willcooke, in any case, there is a soundcloud go scope (for unity8) in lp unity-scope-soundcloud, if you want to have a look
[16:26] <willcooke> I'll take a look.  What I'd like to implement is pretty simple.  I'll do some reading and see if I can make it work.  Thanks davidcalle
[16:27] <davidcalle> willcooke, np
[16:36] <dandrader> greyback, hah! rebasing on top of latest trunk solved it. pushed.
[16:36] <greyback> dandrader: yay
[16:37] <greyback> dandrader: had a breakthrough with the resizing problem too
[16:37] <dandrader> greyback, awesome. so, what's happening?
[16:41] <greyback> dandrader: hmm, maybe postpone that. Breakthrough not complete
[17:47] <mhr3> Saviq, is there in integration u8 branch for the unity-api+scopes-plugin changes?
[17:47] <mhr3> seems it's far from working
[17:47] <Saviq> mhr3, https://code.launchpad.net/~aacid/unity8/use-unity-api/+merge/219222
[17:48] <mhr3> someone forgot to add that to the landing
[17:48]  * mhr3 < someone
[17:48] <Saviq> ;-)