/srv/irclogs.ubuntu.com/2013/11/18/#ubuntu-unity.txt

tsdgeoslol, it shows the CI is back08:01
tsdgeos170 emails :D08:01
tsdgeosanyone knows what's the new s-jenkins ip?08:15
tsdgeosok 10.98.3.13 says Mirv08:16
om26erSaviq, Hi!08:31
mzanettio/08:33
Saviqom26er, o/08:39
Saviqtsdgeos, don't use IPs08:39
om26erSaviq, I created a branch for phablet-tools to unlock screen before running tests, I was asked to get it reviewed by you.08:39
Saviqtsdgeos, just set up the dns https://wiki.canonical.com/UbuntuEngineering/QA/VPN08:39
om26erSaviq, its using the unlock helpers from unity808:40
om26erhttps://code.launchpad.net/~om26er/phablet-tools/unlock_screen/+merge/19552908:40
Saviqom26er, will do08:40
tsdgeosSaviq: but that dns tech is black magic! a good hand written hosts file is the future!08:40
SaviqCimi, re: units.gu and no import - yeah, that's what we get for using context props - as long as you import it once somewhere, it's available for everyone everywhere08:44
SaviqCimi, not ideal, and yeah, we should have the import everywhere we are using units08:45
tsdgeosSaviq: do you know if our autolander is really back? we had some autolanding failures due to some jenkins stuff, shall i re-approve them or wait?08:50
Saviqtsdgeos, yeah it should be working now08:50
tsdgeosok, so i'll top-approve them08:50
Saviqrotfl08:51
Saviqhttps://jenkins.qa.ubuntu.com/job/autopilot-testrunner-otto-trusty/682/artifact/results/autopilot/artifacts/unity8.shell.tests.test_hud.TestHud.test_show_hud_button_appears%20%28Desktop%20Nexus%2010%29.ogv08:51
tsdgeos:D08:51
tsdgeoscome on you can do it, type ubuntu in there!08:51
tsdgeosbut no, it never happens08:52
Saviqthat seems to be the most prevailing failure reason08:52
tsdgeosi've seen a few with timed out scp08:53
tsdgeoslike https://jenkins.qa.ubuntu.com/job/unity8-trusty-armhf-autolanding/74/console08:53
Saviqtsdgeos, https://code.launchpad.net/~aacid/unity8/killApplicationsFilterGrid.qml/+merge/194511 conflicted08:54
tsdgeosbooo08:54
tsdgeosok, tx08:54
Saviqtsdgeos, scp should be fixed hopefully since then08:55
tsdgeos:)08:56
tsdgeosSaviq: reapprove https://code.launchpad.net/~aacid/unity8/killApplicationsFilterGrid.qml/+merge/194511 ? otherwise it'll complain it was approved in 515 but there's a 516 now, no?09:01
Saviqtsdgeos, right, shouldn't have approved09:01
Saviqtsdgeos, done09:01
tsdgeostx09:01
Saviqom26er, here you go https://code.launchpad.net/~om26er/phablet-tools/unlock_screen/+merge/195529/comments/45162709:08
Saviqmzanetti, Cimi, what's the status of https://code.launchpad.net/~cimi/unity8/carousel-music-video/+merge/192118 ?09:16
SaviqCimi, if you could reply to my comment as to what was fixed, would help progress on this09:18
Saviqmzanetti, this should be ready, right? the last two things got fixed https://code.launchpad.net/~mzanetti/unity8/music-preview/+merge/193803 ?09:22
Saviqmzanetti, filed a bug for SDK yet? could we link it in a FIXME there?09:23
mzanettiSaviq: re Cimi's branch: still no tests. do you think it would make sense to have some?09:23
mzanettiSaviq: re my branch: yes. last 2 things are fixed09:24
mzanettiSaviq: yes, bug is filed. can add a FIXME09:24
Saviqmzanetti, re: Cimi's branch, what would you test there? it's mostly really layout changes, apart from maybe one or two "visible" checks09:25
mzanettiSaviq: yeah... ok. fine with me. do we have at least some test that loads up the loader once? to see if components actually compile09:29
Saviqmzanetti, don't think we do, that'd be a simple test, though09:30
Saviqmzanetti, thinking `source = data.source; tryCompare(loader, "status", Loader.Ready)` or so?09:31
mzanettilet me read it once more09:31
Saviqtvoss_, https://code.launchpad.net/~thomas-voss/unity-mir/refactor-oom-score-adj-to-rely-on-process-cpp/+merge/194797/comments/45144009:33
Saviqtvoss_, https://code.launchpad.net/~thomas-voss/unity-mir/refactor-process-group-operations-to-rely-on-process-cpp/+merge/194804/comments/45144109:33
Saviqtvoss_, re: https://code.launchpad.net/~thomas-voss/unity-mir/replace_get_env_with_thread_safe_variant/+merge/195037 - do we have a corresponding set-env? as the actual thread-non-safeness wasn't from two threads getenv'ing, but from one of them getenv'ing while the other was setenv'ing09:34
Saviqtvoss_, resolved by just moving the setenvs to execute earlier09:35
mzanettiSaviq: nah... I think its ok...09:35
Saviqmzanetti, k09:36
mzanettiSaviq: do you know which designer I should contact regarding the expanding list items?09:57
Saviqmzanetti, 'fraid not09:59
mzanettiSaviq: so maybe you can help me then. have a minute?09:59
Saviqmzanetti, poke the usual suspects09:59
Saviqmzanetti, sure09:59
mzanettiSaviq: I'm thinking how to make the API of it. I guess I start off with a ListItem.Base09:59
mzanettiSaviq: now, it sucks if I force the user to position stuff attached to the top. as he'd need to hardcode the collapsed size and he cannot vCenter the text in there for example10:00
mzanettiSaviq: also I'm a bit unsure what should happen on expansion. should the item just grow  and with that reveal any previously clipped content? (/me not likes)10:01
mzanettiSaviq: or should it have some sort of "collapsedContent" and "expandedContent" properties which get exchanged on collapsed/epxanded (/me likes a bit more)10:02
Saviqmzanetti, as said before, that last part should, IMO, be component-specific10:02
mzanettiSaviq: or should the expandedContent just be loaded additionally below the content (I think the nicest API and easy to use, but restricts it a bit on the looks&feel)10:02
Saviqmzanetti, think OptionSelector - it shouldn't have its content replaced, as the content is just a ListView that's being expanded10:03
mzanettiSaviq: which is not really the case in a expanding list item10:03
mzanettiSaviq: as the epanded stuff is most likely not a list10:03
mzanettibut more like previewData10:04
Saviqmzanetti, the OptionSelector *is* an example of expanding list item10:04
Saviqwhere you have an OptionSelector in a ListItem, and it just grows when you "focus" it10:04
mzanettiSaviq: no. it's the other way round10:04
Saviqmzanetti, no, it is not ;)10:04
Saviqmzanetti, ok, let's put it another way10:05
Saviqmzanetti, an OptionSelector is an example of the *content* of an expandable10:05
Saviqwhere that content just grows to fill the expandable10:05
Saviqmzanetti, we're not after a behavior like we have with the preview (OpenEffect)10:06
Saviqmzanetti, but instead just an item growing in height when focused, revealing more content10:06
mzanettiso option one10:07
mzanettiok10:07
Saviqmzanetti, as for "sucks if I force the user to position stuff..." not sure I get what you mean10:09
mzanettiSaviq: I think most of the times the stuff in there will not be a list10:09
mzanettiSaviq: so you have the item that looks like a regular ListItem when collapsed10:10
mzanettiSaviq: but you can't vCenter text in there, as the actually size will change when expanded for example10:11
mzanettiso you need to anchor it to top/bottom with hardcoding margins etc10:11
mzanettistill, the most flexible10:11
mzanettiand that's what we need. so...10:12
Saviqmzanetti, yeah, nothing better comes to mind10:13
om26ercan I check in python or through bash if the screen is turned on or not? i.e. is there  a powerd api for python or a commandline option without root access ?10:21
Saviqom26er, I don't think there is, no, sforshee should know more10:35
=== om26er_ is now known as om26er
nic-doffaySaviq, getting device not found errors in these jenkins reports, haven't seen those yet https://code.launchpad.net/~nicolas-doffay/unity8/search-history-persist10:46
Saviqnic-doffay, it's in queue to run one more time - there were some configuration issues a few hours ago still10:48
nic-doffaySaviq, kk10:48
Saviqnic-doffay, but it's better now, we got two green ones already this morning http://s-jenkins:8080/job/unity8-autolanding/10:49
Saviqnic-doffay, https://code.launchpad.net/~nicolas-doffay/unity8/filter-selector/+merge/191145 fails due to "ListItem.ItemSelectorDelegate is not a type"10:53
Saviqnic-doffay, did that get renamed or something?10:53
nic-doffaySaviq, that's waiting on an sdk branch.10:53
Saviqnic-doffay, ok10:53
nic-doffaySaviq, I still need that icon from design, everything else has been seen to though.10:54
CimiSaviq, don't really know how to fix the carousel...10:54
Cimiasin landscape10:54
Cimi*in10:55
Cimibecause it scales....10:56
Cimiit's dynamic10:56
=== _salem is now known as salem_
Cimiit's bigger than when it's in portrait10:57
Cimiso the assets and spacing/padding/margins are different10:57
Cimifonts too10:57
SaviqCimi, well it should just scale with them, should it not (i.e. margins and such should be a factor of the size)11:01
CimiSaviq, but they won't be pixel grid aligned?11:01
CimiSaviq, when it involves operations11:01
SaviqCimi, dednick had a "align-to-grid" function somewhere11:01
SaviqCimi, in indicators11:02
SaviqCimi, that would find the closest grid-aligned value for an arbitrary value11:02
CimiSaviq, but the asset is not aligned maybe11:02
Cimidunno11:02
Cimihave to try11:02
SaviqCimi, font size should be good as-is, we don't want to scale fonts11:02
SaviqCimi, please do11:02
SaviqCimi, as this will break badly on the tablet, too11:02
CimiSaviq, I think we need better ubuntu shape11:04
Cimilike when we can finally fill inside11:04
dednickPanel/Indicators/DefaultIndicatorWidget.qml::guRoundUp11:04
dednickCimi: ^11:04
SaviqCimi, please ping loicm on the status of it11:04
CimiSaviq, will land before january iirc11:04
CimiSaviq, in december11:04
SaviqCimi, if you can't find a ~simple "will-do" solution, we'll have to wait for that, then11:05
CimiSaviq, I'd wait11:05
CimiSaviq, can't think of something that will work now, unless asking for more assets11:06
Cimiwith a .sci file11:06
SaviqCimi, just ponder it, if you can't - mark the branch as WiP and comment on the bug accordingly11:06
dednickCimi: can you do a review when you have some time?11:10
dednickCimi: https://code.launchpad.net/~nick-dedekind/ubuntu-settings-components/fix-broken-calendar-tests11:10
mhr3sil2100, seems like ci is working now, could you setup ci and autolanding for lp:unity-scopes-api?11:12
mhr3sil2100, althought zmqpp seems to be stuck in proposed?11:12
didrocksmhr3: releasing to distro is not ready yet though11:13
mhr3didrocks, not needed to enable ci though, is it?11:14
didrocksmhr3: well, for upstream merger, it still worth it I guess11:16
tsdgeosmzanetti: you forgot to fix a typo in the same sentence :D re https://code.launchpad.net/~mzanetti/ubuntu-ui-toolkit/fix-listitems-empty-typo/+merge/19540511:23
mzanettitsdgeos: :D11:24
mzanettitypo overflow11:25
tvoss_Saviq, we have set_env, too11:28
tvoss_Saviq, working on a queue of mps right now, will get to that specific one shortly11:29
Saviqtvoss_, cheers11:29
tsdgeosdednick: aren't we better served by fixing that in Qt or glib or wherever the root cause is?11:46
Saviqgreyback, ping11:46
greybackSaviq: pong11:46
Saviqgreyback, icanhazchangelogbump on https://code.launchpad.net/~gerboland/unity-mir/listen-for-server-start-stop-ready/+merge/191224 ?11:47
Saviqgreyback, otherwise the ap tests I've just added to https://code.launchpad.net/~saviq/unity8/raise-sigstop/+merge/191212 will fail if we don't upgrade libunity-mir11:47
greybackokies11:47
Saviqgood we don't have .symbols files, 'cause that wouldn't help at that point, we'd need an explicit >= on libunity-mir111:48
dednicktsdgeos: yeah, well that would be great. But as far as I can tell, it's not actually bug. It's done that way on purpose.11:48
Saviqdidrocks, here's a question, unity8 build-deps on libunity-mir-dev, there's a behavioral change in unity-mir that unity8 wants to depend on, so we're bumping unity-mir's version and build-depending on >= $that_version11:49
Saviqdidrocks, shlibs will then make sure that there's a Depends on >= $that_version11:50
Saviqdidrocks, but if it was C, and there was a .symbols file, and we wouldn't have changed ABI, would that still work?11:50
Saviqdidrocks, or would shlibs find the lowest possible symbol set and use that for the Depends: unity-mir >= $version ?11:51
Saviq/food11:51
greybackSaviq: done11:53
Saviqgreyback, cheers11:53
Saviqgreyback, cu2d would've taken care of the commit message (and the +foo in the version number)11:54
Saviqgreyback, just FYI11:54
greybackSaviq: didn't know that. Ta11:54
tsdgeosdednick: really?11:55
Saviqgreyback, hmm why do we have an explicit Depends on libunity-mir1 for unity8?11:55
Saviqgreyback, wouldn't shlibs take care of that?11:56
tsdgeosdednick: why would it be on purpose?11:56
greybackSaviq: it would. Not sure why.11:56
Saviqgreyback, k, will prep a branch11:56
om26erSaviq, fixed per suggestions: https://code.launchpad.net/~om26er/phablet-tools/unlock_screen/+merge/19552912:01
Saviqom26er, yup, saw that, will look/test soon12:02
om26ercool12:02
dednicktsdgeos: explanation given to me:12:03
dednick<shanek> dednick: the loop level condition for deleteLater processing is to protect you from calling deleteLater and having it destroyed by a nested event loop while still in scope12:03
tsdgeosdednick: and why should i care about internal stuff?12:04
dednicktsdgeos: what we're doing in glib callbacks bypasses the event system.12:04
tsdgeosso bascially i can't use deleteLater and assume it'll delete stuff later?12:05
tsdgeosthat's a big bug, no?12:05
dednicktsdgeos: it will delete it. when the current event loop finsishes. Unfortunately that's when we shut the shell down.12:07
tsdgeosdednick: :D not very helpful12:07
dednicktsdgeos: yes, I'm painfully aware.12:08
dednicktsdgeos: unfortunately that's what happens when you twist a framework.12:09
tsdgeosdednick: so the problem is when we do stuff directly in a glib callback?12:12
dednicktsdgeos: yeah12:12
mhr3can't we just detect the case where the callback is called outside of the qt system and deal with that inside deleteLater?12:12
mhr3tsdgeos, dednick ^?12:12
tsdgeosdednick: is it just deletelater or everything else that fails?12:12
tsdgeosmhr3: it's what i was thinking, but i've had no look at the code at all12:13
dednickmhr3: detect? not that I know of. It's using the same mainloop. There is no difference between an event coming from keyboard and one coming from something we have initiated.12:13
mhr3dednick, well one goes through some more qt :)12:14
dednickmhr3: because the keyboard event pushes it through the qt. Which is what we're supposed to do :)12:14
mhr3dednick, in that case we could deal with it - set a flag when it gets into the qt event system, and unset when it leaves it12:15
mhr3then inside deleteLater we'd look at that flag12:15
mhr3problem solved12:16
dednickmhr3: ? when what goes into qt event system?12:16
mhr3a qevent12:17
dednickmhr3: the qevent for the delete ?12:18
dednickor what caused the delete ?12:19
dednicktsdgeos: it's only deleteLater12:19
tsdgeosdednick: i see12:20
tsdgeosdednick: any idea why deelistmodeltest doesn't get built here?12:20
dednicktsdgeos: need to compile with -DWITHQT5=112:20
tsdgeosah right12:20
dandradermzanetti, got a quick one for you: https://code.launchpad.net/~dandrader/unity8/greeter_edge_hint/+merge/19559112:21
tsdgeosdednick: seen that CI failed?12:21
tsdgeosactually fails here too12:21
tsdgeoshmm12:21
dednickhuh. odd12:22
mhr3dednick, ok, from start 1) app receives key event (or any other) 2) there's a GSource that gets dispatched 3) that GSource usually sends a QEvent 4) stuff happens (q signals etc etc)12:22
mzanettidandrader: what does this exactly do?12:23
mhr3dednick, now somewhere inside the qt event system we'd set a flag (in sendEvent?), if deleteLater is called with that flag set, use the current behaviour, if not - which should mean that this was an direct glib dispatch - don't wait for the next event loop12:23
dandradermzanetti, adds a hinting animation when you press on the right edge of the greeter. try it12:23
mhr3dednick, makes sense?12:24
dandradermzanetti, I think it fits nicely with the teasing animation that happens when you hit the body of the Greeter12:24
mzanettidandrader: ah got it. when you press inside the drag area12:24
dandradermzanetti, exactly12:25
tsdgeosdednick: somehow moc is not getting the -DWITHQT5 flag12:25
tsdgeosmaybe a regression somewhere12:25
tsdgeos?12:25
dednicktsdgeos: eh. it may be because i had compiled before i made changes12:25
mzanettidandrader: shouldn't the DDA in that case reject the gesture and pass it on to any mousearea behind it?12:25
dednickmhr3: yeah, makes sense. could possibly work.12:27
tsdgeosdednick: it's not your fault, the lp:dee-qt one doesn't build here either12:27
tsdgeosdednick: are you on trusty?12:27
dednicktsdgeos: no12:27
tsdgeosi am12:27
tsdgeosso there you go12:27
dandradermzanetti, no. it's the same as what happens when you press on the indicator panel12:27
dednicki thought i had upgraded...12:27
tsdgeossomething changed in cmake probably12:28
tsdgeosor somewhere else12:28
tsdgeosdednick: can you do this for me plz12:28
tsdgeos rm modules/Dee/moc_plugin.cpp12:28
tsdgeosand then12:28
mzanettidandrader: well, ok with me. I was just wondering. as my understanding was it would pass on the input to the next handler if it isn't a gesture. in which case the greeter teasing would already work12:28
tsdgeosmake VERBOSE=112:29
tsdgeosdednick: and paste the moc line somewhere12:29
tsdgeos/usr/lib/x86_64-linux-gnu/qt5/bin/moc bla bla12:29
dandradermzanetti, 1- the framework for having that forwarding scheme working is not in trunk yet12:29
dandradermzanetti, 2- hitting the border of the greeter is different than hitting its main body12:30
dandraderbecause you cannot drag it from the main body12:30
dandradermzanetti,  whereas you can when you hit its right edge12:30
dandradermzanetti, so it makes sense for those two animations to be like they are with this patch12:31
mzanettidandrader: yes. I agree. was just curious. but if there is an app in foreground the right edge would pass non-recognized gestures on to the app once the required stuff is in trunk?12:32
dandradermzanetti, the fist animation says "hey, hit the border to drag me!" whereas the second tells "you can drag me if you start dragging your finger leftwards"12:32
dednicktsdgeos: http://pastebin.ubuntu.com/6437235/12:32
dandradermzanetti, yes of course. in that case a DragHandle is not used. but a EdgeDragArea12:32
tsdgeosdednick: so the -DWITHQT5 is there12:33
tsdgeosnot for me12:33
tsdgeos:-/12:33
mzanettidandrader: ah, got it. yeah. thanks12:33
dednicktsdgeos: from builddir "cmake .. -DWITHQT5=1"12:34
tsdgeosdednick: sure, but not propagating to moc12:34
tsdgeosfor me12:34
dednickweird12:34
tsdgeoswell cmake may have regressed12:35
tsdgeos    Merge branch 'fix-automoc-compile-definitions' into release12:44
tsdgeosthat looks like it :D12:44
dednickcould be, could be :)12:52
=== dandrader is now known as dandrader|afk
=== alan_g is now known as alan_g|lunch
=== dandrader|afk is now known as dandrader
dednickmhr3: trying a patch now. unfortunately it requires me to build qtbase. may take awhile :(13:17
mhr3dednick, yea, i'll ask about it tomorrow ;)13:18
tsdgeosdednick: ok, got the new cmake packaged and on prosoped, going to try if that fixes the build for me after rebasing for nth time my qt patch14:03
* mzanetti celebrates jenkins-is-back-day :)14:05
=== alan_g|lunch is now known as alan_g
tsdgeosdednick: yep, problem gone with cmake from trusty-proposed14:13
Saviqmzanetti, if only it was jenkins-is-really-back-day and not jenkins-is-sometimes-back-day ;P14:14
dednicktsdgeos: cool14:15
mzanettilol14:15
Saviqmzanetti, too many things still not in their best shape14:15
mzanettiSaviq: https://code.launchpad.net/~mzanetti/ubuntu-ui-toolkit/expanding-listitem/+merge/19560214:42
Saviqdednick, standup?14:43
mzanettiMacSlow: we should get this reviewed in order to land your fullscreen branch: https://code.launchpad.net/~mzanetti/unity8/generic-lockscreen/+merge/19195114:45
greybackkgunn: when is my scenegraph session? I can't find it...14:48
MacSlowmzanetti, yup14:48
mzanettiMacSlow: can you review this? I'll take your fullscreen branch then14:49
kgunngreyback: mmm...let me check....i followed instructions....which may not have resulted in a "planned session" :)14:49
MacSlowmzanetti, ok14:49
* greyback totally thought USD was next week14:49
Saviqkgunn, you should also mark folks (me/ricmm) as essential where we're supposed to run them14:49
didrockskgunn: hey, just coming to the info from u-s-c not building :)14:49
didrockskgunn: also, it seems that platform-api doesn't have the bump build-dep merged?14:49
kgunnSaviq: so picky...:).... i thot i had...sorry14:50
Saviqkgunn, I just don't want to miss it, ya know ;)14:50
kgunnSaviq: i know you were super pumped about it! ;) (...Borat _not_)14:50
tsdgeoshmmm14:51
didrocksargh, I saw a kgunn and was thinking it was the Mir channel, let's move there :)14:51
tsdgeosshould we delay our standup tomorrow given it is at the same time of vUDS opening so we can listen to Jono/Mark?14:51
Saviqmzanetti, there should be only one expanded item per container (i.e. others collapse when you expand a new one)14:54
Saviqmzanetti, that's not the case yet - that expected?14:54
mzanettiSaviq: ah right. sure. no forgot that. but its easy14:54
mzanettiSaviq: but the general approach seems ok, right?14:54
Saviqmzanetti, just looking at the code - it behaved correctly, yeah14:54
Saviqmzanetti, looking good, yeah14:56
mzanettiok, cool14:56
Saviqmzanetti, we'll need an ExpandingColumn, too, in case you don't want a ListView14:57
mzanettiyep14:57
mzanettiSaviq: altough that would be just setting height: contentHeight; interactive: false14:57
Saviqmzanetti, nope14:58
Saviqmzanetti, you'd have to use a Loader to get different delegates14:58
MacSlowmzanetti, there are three merge-conflicts with  your generic-lockscreen branch atm14:58
Saviqmzanetti, in a ListView14:58
Saviqmzanetti, also, for naming I'd go Expandable, ExpandablesListView / ExpandablesColumn or so14:58
mzanettiMacSlow: oh... fixing asap14:59
Saviqas ExpandingListView feels like it itself would be expandable14:59
Saviqmzanetti, for ExpandablesColumn, think of:14:59
mzanettiyeah, I got it14:59
mzanettino repeater15:00
SaviqExpandablesColumn { ExpandableFoo { } ExpandableBar { } ExpandableBaz {} }15:00
Saviqyeah, no model15:00
mzanettiack15:00
Saviqso slightly trickier, but not a whole lot15:00
mzanettiMacSlow: I'm wondering why the SIM pin lock in trunk still is 4 digits only given that we merged this already: https://code.launchpad.net/~mzanetti/unity8/sim-pin-variable-length/+merge/19162515:01
Saviqmzanetti, one more thing, though, you need to limit expandedHeight15:01
mzanettiyep.15:01
mzanettithere's no fllickable in there yet15:01
Saviqmzanetti, ok coll15:01
mzanettithat's what I'm waiting on design for15:01
Saviqcool15:01
mzanettiSaviq: there are a few issues with closing it in that case15:02
mzanettidesign-wise15:02
Saviqmzanetti, yeah I know15:02
Saviqmzanetti, progressing good :)15:02
mzanettithanks15:03
kgunngreyback: weird...the scenegraph session didn't get approved...don't know why15:07
MacSlowmzanetti, I can't trigger the pin-entry on my device (unlocked sim).15:07
greybackkgunn: ok. Bit late to add now?15:08
MacSlowmzanetti, I used the simentry-example from lp:unity-notifications15:08
MacSlowmzanetti, and there it works correctly.15:08
mzanettiMacSlow: not sure what you mean15:08
dandradertsdgeos, what motivated you to report https://bugs.launchpad.net/bugs/125041215:09
ubot5Ubuntu bug 1250412 in Ubuntu UI Toolkit "InverseMouseArea does not work with TouchEvents" [Undecided,New]15:09
dandradertsdgeos, ie., what unity8 use case is affected by it15:10
tsdgeosdednick: cimi's15:10
tsdgeosCimi: ↑↑15:10
mhall119hey guys, who's still developing on Unity 7 in here?  I have some non-work-related questions15:10
dandradertsdgeos, you mean this? https://bugs.launchpad.net/unity8/+bug/119341415:10
ubot5Ubuntu bug 1193414 in unity8 (Ubuntu) "[DASH] "quit" mode for 'recent apps' should be less persistent" [High,Triaged]15:10
Cimimhall119, Trevinho15:10
dednicktsdgeos: eh?15:10
tsdgeosdednick: sorry, wrong d-autocompletion15:11
dednickah15:11
kgunngreyback: i think so...15:11
mhall119Trevinho: ping when you're around15:11
tsdgeosdandrader: well, the branch he's working on to fix that15:11
MacSlowmzanetti, what did you run to check what pinlength is used?15:11
kgunngreyback: for sure...even it does get add late...no prep...just a review of where we're at15:11
Trevinhomhall119: pong15:11
mzanettiMacSlow: there is no such thing as a pinlength any more15:11
greybackkgunn: okies, keep me posted anyway15:12
tsdgeosdandrader: i don't remember the branch exactly15:12
tsdgeosdandrader: but basically anything that needs the topmostitem feature in IMA15:12
tsdgeoss/enables/needs15:13
dednickdamned qt private data exports...15:13
mhall119Trevinho: hey, how hard would it be to build a new desktop session (an alternative to ubuntu-desktop) that uses Unity, without it's settings conflicting with ubuntu-desktop's Unity settings?15:13
mhall119I tried to do this with Xfce years ago with moderate success15:14
MacSlowmzanetti, but in Greeter/PinLockscreen.qml there's still this property "pinLength"... I'm confused now.15:14
=== dandrader is now known as dandrader|lunch
mzanettiMacSlow: yeah. ignore that. just don't use it in the notification15:15
Trevinhomhall119: mhhm... well unity settings are stored all in gsettings, but some via comipz other directly...15:15
Saviqdidrocks, (how) can you mark people as essential for a session on UDS?15:15
Trevinhomhall119: so... for compiz side all you need is just to add a new compiz profile15:15
Saviqdidrocks, pinging you as you're marked as the person to ask for that on the session participation page15:15
Trevinhomhall119: for gsetttings side I think you need to force a different backend when starting the session15:15
mhall119Saviq: only the meeting's creator (and possibly track lead) can mark people as essential15:16
seb128Saviq, just subscribe them to the bluepprint and tick the "presence required" box?15:16
Trevinhoalthough, this last one would probably affect even compiz settings, so should apply to all15:16
mhall119Trevinho: blegh, I was going to have to do that with Xfce too, which was hit-or-miss15:16
Saviqseb128, hmm there seem to be sessions without blueprints (http://summit.ubuntu.com/uds-1311/meeting/22084/unity8-shell-discussion/)15:17
mhall119which means in order to have separate Unity/Compiz settings, I also necessarily have to have different GEdit settings (and different settings for any app using gsettings)15:17
seb128Saviq, ok, so what mhall119 said then15:17
mhall119seb128: the Launchpad "require" option is no longer strictly enforced in Summit15:17
mhall119LP "required" = Summit "would really like to attend but not absolutely critical"15:18
Trevinhomhall119: I see... mh, well, I didn't check gsettings much, probably other guys are much more experienced with it than I am, but I don't know if it's possible to define a different profile for  just a class of setings... it seems tha compiz is doing that in some way, but I didn't study that much15:18
seb128mhall119, oh? since when? what's the right way to make sure people don't have conflicts?15:18
MacSlowmzanetti, so there also is no "x-canonical-pin-length" anymore?15:18
mzanettiMacSlow: nope15:19
mzanettiMacSlow: that merge actually already removed that... I guess it came back somehow which is what it breaks again15:19
seb128Saviq, I can add participants I think ... who do you need?15:19
didrocksSaviq: can do as well if needed15:20
MacSlowmzanetti, no idea what what merge might have reverted that15:20
mzanettiMacSlow: fixed the merge conflicts15:21
* mzanetti -> food15:21
mhall119seb128: for a while now, since before Copenhagen at least15:21
mhall119seb128: Summit will warn about conflicts with "really would like to attend", but it won't block them15:22
seb128mhall119, ok, that's good enough I guess (if people schedule do take the warnings into accounts and tweak to resolve those)15:22
MacSlowmzanetti, ok15:23
mhall119seb128: with auto-scheduling turned off for vUDS, it really doesn't make much of a difference either, it was mainly removed to make auto-scheduling/rescheduling less disruptive to the schedule15:25
seb128right15:26
seb128mhall119, thanks15:26
=== alan_g is now known as alan_g|tea
CimiSaviq, tsdgeos so I have API shared between carousel and filter grid, but when I use a loader to load those I guess I need to alias all those properties15:27
Cimiis there a way to avoid that and duplicating code?15:27
tsdgeosi don't think so15:27
tsdgeosCimi: do you have the branch where you need the topmostItem enabled IMA ?15:30
tsdgeosi forgot the url15:30
Cimitsdgeos, it was a patch15:31
* Cimi looks history15:31
Cimihttp://paste.ubuntu.com/6382346/15:31
tsdgeosdandrader|lunch: ↑↑↑15:31
tsdgeosdednick: ping15:32
dednicktsdgeos: plop15:32
tsdgeosdednick: i was wondering why you went that weird QGlibCallbackEvent instead of just doing a QMetaObject::invokeMethod that my quick test shows it seems to work too15:33
tsdgeosi think it's simpler (doesn't need a new class)15:33
=== om26er is now known as om26er1
=== om26er1 is now known as om26er
dednicktsdgeos: you use queud connection?15:33
tsdgeosyep15:33
dednicktsdgeos: because it goes onto the post queue. we don't want that15:34
tsdgeosah right, you're sending it instead of posting it15:35
=== alan_g|tea is now known as alan_g
tsdgeoswhich tbh in this specific case i don't see why it should make any difference15:35
dednicktsdgeos: yeah, otherwise it's async, and the values we use in the callbacks may not be valid fot the dee model. Would get much more compicated to cache.15:36
tsdgeoshmmm15:37
tsdgeosok15:37
dednickqt stacktrace over 125 levels deep. sigh...16:03
tsdgeosmzanetti: what do we do with https://bugreports.qt-project.org/browse/QTBUG-32251 ?16:06
tsdgeoskill it?16:06
tsdgeosor?16:06
tsdgeosdednick: so there's no way we can fix it inside deleteLater? mhr3's idea no good?16:08
dednicktsdgeos: mhr3's seems to be working, but i'm just testing and getting a few "out of context" deletes (this special case) which i don't know where are coming from in unity8.16:09
dednicktsdgeos: and the stack is so deep, that i can't get to the bottom :(16:09
tsdgeoslol16:10
dednickbut it seems to be to do with changing the theme name16:10
mhr3dednick, you sure you're not getting stack overflow?16:10
tsdgeostbh i'd prefer a fix in deleteLater than this fix16:10
mhr3such deep stack sounds like something broken16:11
dednicktsdgeos: yes, well we would have to fix it elsewhere if not in qt. gsettings-qt is another place i've found which will suffer.16:11
tsdgeosdednick: if you want i can try to help you with this effort, i even fixed the qt event loop ordering events incorrectly once (took me two days to fix it and two weeks to convince the Qt guys they had a bug in there) :D16:11
dednicktsdgeos, mhr3: http://pastebin.ubuntu.com/6438140/16:12
mhr3some debug symbols would be nice :)16:12
dednicktsdgeos: yeah, help would be appreciated.16:12
mhr3dednick, can you pastebin the patch?16:12
mzanettitsdgeos: good question...16:13
tsdgeosmzanetti: i'm voting to kill it16:13
tsdgeosand open a new clean one if we need later16:13
mzanettitsdgeos: ok16:13
tsdgeosthis one's already linked to a dead review request16:13
dednickmhr3: um... i apt-source'd it.16:13
tsdgeoswhich makes it "unclean it"16:13
dednickmhr3: how do i gen changes?16:14
mhr3dednick, apt-source again and diff the dirs?16:14
tsdgeosdednick: my "i know nothing about deb" way is doing what mhr3 says :D16:14
tsdgeosnot sure if there's a more correct way16:14
dednick:)16:14
mhr3tsdgeos, yey, i'm not the only one! :)16:14
mhr3dednick, also, consider installing pkg-create-dbgsym for the next build :)16:18
mhr3(that will give you .ddebs in your build-area)16:18
dednickmhr3: http://pastebin.ubuntu.com/6438169/16:20
dednickmhr3: meh. most of that is from libqt5declarative, which i'm not building myself16:22
=== dandrader|lunch is now known as dandrader
tsdgeosdednick: can you explain a bit the logic of the patch?16:31
mhr3i was about to ask about the isWidgetType16:32
mhr3why events are not delivered to non-widgets?16:32
mhr3i guess those don't go through coreapplication?16:32
dednickhm, yeah, i'm not sure about that16:33
dandraderSaviq, ping16:37
dandraderCimi, I'll investigate https://bugs.launchpad.net/unity8/+bug/1193414 taking into consideration the patch I've written to Qt (https://codereview.qt-project.org/71241) and if/how InverseMouseArea would benefit from it16:43
ubot5Ubuntu bug 1193414 in unity8 (Ubuntu) "[DASH] "quit" mode for 'recent apps' should be less persistent" [High,Triaged]16:43
om26ermterry, hey! can you point me to the code in unity8 which turns on the screen when a call arrives ?17:07
mterryom26er, I'm guessing that's deeper than unity8.  the way screen display is controlled is that powerd talks over dbus to unity-mir.  I'm not sure which piece of code pokes powerd upon a call arriving though17:08
mterryom26er, something in the ofono stack?17:09
om26ermterry, I am thinking telephony-service. will as salem_17:09
om26er*ask17:10
cwayneSaviq: hey, re: our discussion awhile ago about choosing a background based on whether the device is a phone/tablet/desktop.. is there a UDS session to address this?17:10
=== boiko_ is now known as boiko
Saviqdandrader, pong17:11
Saviqcwayne, I don't think there is, but we might hijack one of the convergence sessions17:11
dandraderSaviq, so what's the procedure to have a qt patch backported?17:11
salem_mterry, the code currently lives in powerd itself. but we should move it to another place.17:12
salem_om26er, ^17:12
Saviqdandrader, I'd say easiest is to:17:12
Saviqa) branch lp:ubuntu/qtdeclarative-opensource-src17:12
om26ersalem_, so telephony-service asks powerd to turn on the display when a call arrives ?17:12
Saviqb) quilt -import <patch file>17:12
Saviqc) resolve any patch issues17:13
Saviqd) quilt refresh17:13
salem_om26er, no, powerd talks directly with ofono.17:13
Saviqe) bzr commit / propose merge17:13
Saviqdandrader, or, if you don't want to learn what quilt is and stuff ;)17:13
Saviqdandrader, file a bug against lp:ubuntu/qtdeclarative-opensource-src17:14
Saviqdandrader, and attach the patch file - ideally backported to 5.2 if required17:14
Saviqdandrader, but if you don't backport it - whoever will integrate it, will17:14
dandraderSaviq, ok, thanks17:15
Saviqdandrader, don't hesitate to ping if you need pointers17:15
=== dandrader is now known as dandrader|afk
sil2100mhr3: hi! lp:unity-scopes-api is to be daily-released sooner or later?17:46
sil2100(although the term daily-release is not as accurate now as it was before)17:46
mhr3sil2100, the sooner the better17:46
sil2100mhr3: ok, I'll review the packaging then and prepare17:47
mhr3sil2100, but at least having autolanding would be nice17:47
mhr3sil2100, i mean automerging of branches to trunk17:48
sil2100mhr3: right, let me work on that as well17:49
=== dandrader|afk is now known as dandrader
=== alan_g is now known as alan_g|EOD
=== salem_ is now known as _salem
=== _salem is now known as salem_
dednickmhr3: so you coming to the office tomorrow?18:43
mhr3dednick, yep, afternoon20:01
Cimimhr3, you're worse than me!20:48
Cimi:P20:48
veebersSaviq: Hey I saw the unity-ci runs succeeded finally. Was it an infrastructure thing, unity thing or generally unstable?20:50
Saviqveebers, depends - last issue I saw was https://bugs.launchpad.net/ubuntu-ci-services-itself/+bug/125238620:51
ubot5Ubuntu bug 1252386 in Ubuntu CI Services "otto runner has locked unity7 session from time to time" [Undecided,New]20:51
veebersSaviq: ah ok, cheers20:53
Saviqveebers, but yeah, it's getting better20:53
Saviqveebers, just slow - stuff's queued up20:53
veebersSaviq: Ack, it's good to have the labs back though :-)20:54
Saviqindeed20:55
=== salem_ is now known as _salem
mhr3Cimi, nope, you still hold the record :P23:13

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!