/srv/irclogs.ubuntu.com/2013/11/13/#ubuntu-unity.txt

=== Guest12534 is now known as jacky
=== jacky is now known as jalcine
=== _salem is now known as salem_
=== salem_ is now known as _salem
=== duflu_ is now known as duflu
=== iahmad_ is now known as iahmad|afk
=== iahmad|afk is now known as iahmad
tvossmzanetti, you fine if I answer to your latest reply publicly. Seems like you only sent it to me instead of the list09:42
=== tvoss is now known as tvoss|test
=== tvoss|test is now known as tvoss
tsdgeosSaviq: damn, my 5 lines patch fixes https://bugreports.qt-project.org/browse/QTBUG-30730 and https://bugreports.qt-project.org/browse/QTBUG-34617 but not https://bugreports.qt-project.org/browse/QTBUG-3465110:10
tsdgeosi'll have a look at that one too10:10
mzanettitvoss: oh... yeah, that wasn't intentionally10:17
tvossmzanetti, ack, can you resend publicly, will answer then10:18
mzanettitvoss: done10:19
Saviqtsdgeos, oh interesting...10:25
tsdgeosyeah10:25
tsdgeossimilar but not same thing10:25
Saviqtsdgeos, 1 != 1?...10:25
tsdgeosyep10:25
tsdgeoskind of10:25
Saviqit probably tries to compare the wrapper against the object or something?10:25
tsdgeosyeah10:25
tsdgeosneed to find out a comparison that works and then fix this one to go there too :D10:26
Saviq;)10:26
dednickCimi: ping10:51
tsdgeoswoa, adding a debug in qv4runtime_p.h wasn't a good idea :D its recompiling almost everything :S10:57
=== _salem is now known as salem_
dednickCimi: https://bugs.launchpad.net/unity8/+bug/125081511:14
ubot5Ubuntu bug 1250815 in Unity 8 "Carousel click/pressAndHold not working" [Critical,New]11:14
Saviqmhr3, any reason why lp:unity-scopes-shell source package is unity8-plugin-scopes? why can't it be the same name?11:36
Saviqmhr3, also, qtdeclarative5-foo-plugin is the preferred name for QML plugins11:37
Saviqmhr3, OTOH we might want a different naming, 'cause it's meant to be installed in a different import path...11:37
Saviqdidrocks, thoughts ↑? where should we install plugins directed at the shell?11:38
Saviqand whether to change the package naming scheme?11:38
SaviqMirv, I'd like your opinion, too ↑11:38
didrocksthe source package needs to match the project name11:39
didrocksfor non confusing reason :)11:39
didrocksare those only unity8 plugins?11:39
didrocksit's not in the standard QML import path?11:39
MirvSaviq: the preferred naming is actually qtdeclarative5-foo0.1, as decided by kenvandine & co, so that also the install path contains the API version specific path - see for example qtdeclarative5-friends0.211:40
MirvSaviq: the idea is that one can co-install older API package as well, for example than that was released with 13.1011:40
didrocks(I decided that one actually :p)11:40
Saviqdidrocks, yeah, we want to put them on a unity8- (well, shell-) specific path11:41
Saviqdidrocks, so that apps don't have access there11:41
didrocksmaybe it shouldn't follow that naming scheme then11:41
Mirvdidrocks: oh, I never figured out how the decision process went, I only heard the end result :)11:41
didrocksSaviq: will it follow your virtual unity-apis versions?11:41
Saviqdidrocks, yes11:41
didrocksso I would say unity8.<APIversion>-<plugin-name>-plugin11:42
didrockshoping that one day we can make 8 and API version matching :)11:42
Saviqdidrocks, there's no one global API version, though...11:43
Saviqdidrocks, or what is that APIversion meant to be?11:43
Saviqdidrocks, I mean scopes have different API version than notifications, for examlpe11:44
Saviqas we didn't want people to have to bump if changes were unrelated11:44
didrocksSaviq: well, as long as you are backward compatible11:44
didrockswe should really use semantic versionning11:44
didrocksand "integrating with the shell" is just one version11:45
didrocksbeing scope or integration with the launcher or indicators11:45
Saviqdidrocks, yeah, we're not backwards compatible, remember? ;P11:45
didrocksSaviq: I know… ;)11:46
Saviqdidrocks, what would APIversion be, then? unity-api version number?11:46
didrockswhen do you think you will ensure backward compatibility?11:47
didrocksthat can maybe help shaping the versioning11:47
Saviqdidrocks, backwards compatibility of Unity8 with older API versions?11:57
Saviqdidrocks, not until 16.04 :P11:57
Saviqdidrocks, those are not APIs public enough that it warrants backwards compat for quite some time11:58
mhr3Saviq, i knew the naming will be the primary thing of discussions :)12:01
mhr3anyway, i don't care, just tell me what to change it to12:01
mhr3didrocks, but we'll need to push the scopes plugin as a separate pkg to the distro12:02
mhr3ideally asap :)12:02
mhr3Saviq, though i guess we should bump unity8 minor/micro version12:03
mhr3and then make the scopes plugin break unity8 << that_version12:04
Saviqmhr3, yeah probably12:04
didrocksSaviq: so I would say, let's version it to 0 and handle the breakage manually12:10
didrockseverytime you break the API, you Breaks: all components << version12:11
Saviqdidrocks, so, unity.0-scopes-plugin in that instance?12:11
Saviq(I wouldn't want to put the 8 in there)12:11
didrocksSaviq: unity0 then12:11
didrocksno .12:12
Saviqdidrocks, Mirv how about the path where we install shell-facing plugins?12:12
didrocksSaviq: same concept to me?12:12
didrocksall unity0- something12:12
didrocksthen, the patch itself12:12
didrockspath*12:12
mhr3"unity0-scopes-plugin"12:13
didrocksshould be libexec I guess12:13
mhr3looks wrong :P12:13
Saviqyeah, /me no likes either :/12:13
didrocksmhr3: yeah, we need to prepend12:13
didrockssomething12:13
didrocks-scopes- is the name of the extension, right?12:13
mhr3also the plugin seems pretty redundant12:13
mhr3qtdeclarative-unity0-scopes12:14
didrocksmhr3: no12:14
Saviqdidrocks, we need a common /usr/lib/*/unity/qml, or similar, path where those will get installed12:14
didrocksmhr3: that sounds like it's an official qt things12:14
mhr3didrocks, we already have dozens of qml plugins like that that aren't official12:14
didrocksSaviq: yeah, sounds good to me (unversionned or unity0 until we have something)12:15
didrocksmhr3: yeah, but it's a plugin of a shell12:15
didrocksnot a qt plugin for using without the shell12:15
mhr3well you wanted a prefix :)12:15
didrocksnot that one :p12:15
=== alan_g is now known as alan_g|afk
didrocksor unityplugin0-scope ?12:16
mhr3fine with me12:16
didrocksunity-plugin0-scope? (putting the API version next to plugin)12:16
Saviq+112:16
mhr3what if the plugin supports multiple versions of the api?12:17
didrocksmhr3: it's unrelated from a packaging perspective12:17
mhr3ok12:17
didrocksunity-plugin21-scope can provides unity-plugin20-scope12:17
didrocksmhr3: ^12:17
mhr3ok, renaming to unity-plugin0-scopes12:18
didrockshum, in fact, we're wrong12:18
didrocksmhr3: you are not providing the plugin12:18
didrocksunity8 is12:18
didrocksso unity8 needs to provides unity-plugin012:18
didrocksand you plugin is just unity-plugin-scopes12:18
didrocksyour*12:18
didrocksyou won't have plugin of your plugin, right? :)12:18
mhr3hm?12:18
mhr3ah, i see what you mean12:19
didrockswill the scopes plugin won't have backward compatibility for the scopes?12:19
didrocksso, basically, do you provide an API in it?12:19
mhr3no12:20
didrocksor is it just an extension like "show battery level"12:20
didrocksok, so you don't need to version in12:20
didrocksyou just need to dep on unity-plugin012:20
didrockswhich will be provided by unity812:20
Saviqhuhuh?12:20
mhr3this last part is huh indeed12:21
Saviqdidrocks, that package will provide the binary plugin12:21
didrocksok, I think there is an inception effect :)12:21
didrocksah ok12:21
didrocksso yeah, versioning is needed12:21
didrocksit will provide an API to scopes?12:21
Saviqdidrocks, no12:22
Saviqdidrocks, it's the shell side of it12:22
mhr3it will implement interface defined in lp:unity-api, used by unity812:22
* didrocks is lost by what you with "will provide the binary plugin"12:22
Saviqdidrocks, will provide the plugin unity8 will use to talk to scopes12:22
Saviqdidrocks, not the APIs scopes will use12:23
* didrocks is probably dumb, but not seeing the difference in those 2 statements12:23
didrocksif the plugin implements the code for unity8 to talke to scopes12:24
didrocksit's exposing/creating the API for scopes, right?12:24
Saviqno12:24
Saviqit's the shell side of that transaction12:24
Saviqthe scope side of that transaction will come from lp:unity-scopes-api12:24
mhr3and it will actually dep on libunity-scopes-api... soon12:25
didrocksah, so basically, this plugin will only decide "this is how I render/handle the signals I'm getting through unity-scopes-api"12:25
Saviqdidrocks, scope → unity-scopes-api → unity-scopes-shell → unity812:26
didrocksyeah, so that's what I told12:26
Saviqdidrocks, yeah, you can think of this as a shell-facing QML wrapper of unity-scopes-api12:26
didrocksunity8 as a private API for its plugins12:26
=== alan_g|afk is now known as alan_g
didrockswhich isn't unity-scopes-api12:26
didrockshas*12:26
didrocksso unity8 should provides a unity-plugins0 package12:27
didrocksthat the unity-scopes-shell will depend on12:27
Saviqdidrocks, there is nothing like that ;(12:27
mhr3not really private, it's defined in unity-api12:27
Saviqor well, in that sense12:27
mhr3or... will be :)12:28
Saviqunity-api defines the APIs unity8 consumes12:28
didrocksok, so all will be exposed in unity-api?12:28
mhr3the problem is that there's no compile/link dependency there12:28
Saviqdidrocks, unity-api just provides headers12:29
Saviqdidrocks, interfaces12:29
Saviqwe will probably make a .so of them soon-ish12:29
didrocksright, and that's the interface that the unity-scopes-shell will use?12:29
Saviqdidrocks, will expose12:29
Saviqdidrocks, unity8 will use12:29
didrocksso there is no need for yet-another-versioning? We have through unity-api what unity8 version the unity-scopes will be compatible with12:30
Saviqdidrocks, yes12:30
didrocksand so unity-plugin-scopes is enough12:30
Saviqyay12:30
didrocks(and we can name all those plugins unity-plugin-*)12:31
Saviqdidrocks, +112:31
mhr3damn it, if i didn't put the 8 there i'd have it already right :P12:31
didrocksok, nice :)12:31
Saviqlol12:31
didrocksheh mhr3 :)12:31
Saviqmhr3, rename the source, please, too12:33
mhr3Saviq, pushed12:34
Saviqmhr3, and make it Provides: unity-scopes-impl, unity-scopes-impl-012:34
mhr3k12:34
Saviqmhr3, and then in the lp:unity8 part, Depends: unity-plugin-scopes | unity-scopes-impl, unity-scopes-impl-012:35
Saviqmhr3, and the version bump and Breaks: you mentioned12:36
mhr3eh12:37
mhr3dpkg-source: error: source package has two conflicting values - unity-scopes-shell and unity-plugin-scopes12:37
mhr3are you sure src pkg can be different to the binary pkg name?12:37
didrocksmhr3: it means you changed in either debian/control or debian/changelog but not in both place12:38
tsdgeosSaviq: this one is a bit more complicated https://codereview.qt-project.org/#patch,all,71155,3 let's see how it goes12:39
mhr3didrocks, i have to use the src pkg name in changelog?12:39
Saviqtsdgeos, lol, so the original code was isEqual() { return false; }? :D12:40
mhr3yea.. ok12:40
Saviqmhr3, just the last entry12:40
mhr3there's just one :)12:40
tsdgeosSaviq: well, there's lots of classes that reimplement the isEqual12:40
tsdgeosit just doesn't hit them if you're comparing singletons12:40
Saviqtsdgeos, mhm12:41
tsdgeosthat's why i'm not sure it's the best place to do it12:41
tsdgeosbut here it works12:41
tsdgeoslet them comment :D12:41
didrocksmhr3: in the top one, yeah12:42
mhr3didrocks, since both unity-plugin-scopes and unity8-private provide the same file, do i add Breaks to unity-scopes-shell, or Conflicts/Replaces needed too?12:46
didrocksmhr3: the files are getting removed from unity8-private, right?12:47
didrocksso we need to uninstall it?12:47
mhr3didrocks, yea, latest version of unity8-private won't have them12:47
didrocksmhr3: Breaks/Replaces in that case12:48
didrockson the version << unity8-private without that file12:48
mhr3thx12:48
didrocksyw!12:48
* didrocks goes for a run12:48
Cimiseb128, if I run system settings on the desktop, I don't have wifi... do you?12:48
mhr3didrocks, if you could check once you're back http://bazaar.launchpad.net/~unity-team/unity-scopes-shell/trunk/view/head:/debian/control12:50
mhr3Saviq, ^ you too12:51
SaviqCimi, network uses different indicator backends12:52
SaviqCimi, on desktop it's nm-applet, not indicator-network12:52
seb128Cimi, I have it just fine12:52
seb128Saviq, the package depends on indicator-network so it should be there12:52
Cimimmm12:52
Saviqseb128, right, if you have two network indicators, yeah12:53
Cimino wifi here on system settings12:53
Cimitrusty12:53
seb128is indicator-network running?12:53
seb128did you restart your session since you installed it?12:53
Saviqmhr3, I don't think you need -dev-tools build-dep, though12:53
Cimino12:53
seb128so restart the session12:54
=== alan_g is now known as alan_g|lunch
Saviqseb128, so 1980's solution! ;D12:54
SaviqCimi, seb128 restarting unity-panel-service should be enough, no?12:54
seb128restart unity-panel-service12:54
seb128Saviq, snap :p12:55
Saviq:)12:55
mhr3Saviq, qmlplugindump12:55
Saviqmhr3, ah12:55
Cimiseb128, ok works now, but I have two12:55
SaviqCimi, yes, expected12:55
Cimion unity12:55
Cimihow can I remove one?12:55
Cimiblacklist in unity?12:55
mhr3Saviq, but i wouldn't be surprised if something was missing/extra, didn't try to build it on clean chroot12:55
SaviqCimi, you'll lose functionality12:56
seb128can you stop complaining? :p12:56
Saviqmhr3, let me12:56
mhr3Saviq, thx12:56
Saviqseb128, me? :D12:56
seb128you can edit /usr/share/unity/indicators/com.canonical.indicator.network and delete the desktop profile12:56
seb128Saviq, no, Cimi12:56
mhr3Saviq, as there's no ci for it yet, you can push fixes directly to trunk ;)12:56
Saviqseb128, ah yeah, no, he can't ;)12:56
seb128Saviq, seems so ;-)12:57
CimiSaviq, I'm half polish :)12:57
SaviqCimi, indeed, that would explain things12:57
seb128Cimi, btw can you spend some time to fix the GTK themes before the LTS? they have some issues/stuff not looking right on new widgets12:57
Cimiseb128, saw that12:58
Cimiseb128, was actually thinking of doing an ubuntu touch theme12:58
seb128Cimi, LTS is not about doing new stuff, it's about polishing what we have and shipping a solid product...12:59
seb128Cimi, thanks for looking at those bugs/fixing the issues12:59
Cimiseb128, depends when polishing is more work12:59
seb128I doubt fixing a few bugs is that much work12:59
mhr3seb128, then you didn't do gtk theming :)13:00
seb128mhr3, don't get me started ... can we have a working dash for the LTS? ;-)13:00
mhr3seb128, we don't?13:01
seb128mhr3, for some value of "working" I guess13:01
seb128mhr3, I still get the "no result found" empty dash issue several time a week13:02
mhr3hmm, i still don't know what's causing that13:02
seb128still no idea of what info I could provide to help debugging it?13:02
mhr3seb128, steps to reproduce13:03
seb128"use the dash"13:03
mhr3works fine... not a bug :P13:03
seb128mhr3, http://ubuntuone.com/0RPwctnL0nRftR9hWFTMuS13:05
seb128mhr3, I clear the text by pressing "esc"13:06
Saviqdednick, any steps to repro bug #1243146 ?13:07
ubot5bug 1243146 in Unity 8 "SIGSEGV when clearing message indicators" [Medium,In progress] https://launchpad.net/bugs/124314613:07
Saviqmhr3, you need qt5-default13:07
seb128mhr3, also my homescreen only lists apps, not files atm ... is that normal?13:07
Saviqmhr3, let me add13:07
mhr3seb128, do you have online sources on?13:08
seb128mhr3, yes13:08
mhr3wonder if that would make difference13:08
mhr3hm, so do i13:08
dednickSaviq: it's pretty random13:08
mhr3seb128, as for files, check if the category is enabled13:09
seb128mhr3, oh, fun13:09
seb128mhr3, only "help" is enabled in the filters, that's why my dash is empty13:09
dednickSaviq: have a message in the indicator expanded and clear them all.13:09
seb128mhr3, if I enable applications it works13:09
mhr3solved, yey! :)13:10
seb128mhr3, is the smart server trying to be too smart and telling my dash to disable apps?13:10
seb128mhr3, well, it's not happening on session start/by my doing, I didn't even know we had filter in the homescreen13:10
mhr3no, empty search filters are remembered13:11
mhr3and i think there's a bug somewhere where it saves incorrect state of those13:11
mhr3pstolowski, ^13:11
Saviqdednick, thanks13:12
mhr3pstolowski, maybe, search for "foo" (which takes a while), change filters while it's in progress and quickly press esc can make it confused?13:21
pstolowskiseb128, as mhr3 said, the only way it can disable apps is when user manually de-selects apps in the filter while the search string is empty. if it happens on any other interactions, then this is a bug13:29
pstolowskiseb128, if you find any way to reproduce it, then please include the output of 'gsettings get com.canonical.Unity.Lenses home-lens-default-view' in the bug report13:29
seb128pstolowski, hey, there is a bug then...13:29
seb128pstolowski, ok, the issue is that I've not been able to find obvious steps to reproduce, but I run into that several time a week13:30
seb128the dash is not empty all the time, and not on session start, so I doubt it's a buggy config13:30
seb128it happens a some point during the day13:30
seb128pstolowski, in that case the filter had "help" only selected13:31
pstolowskiseb128, but when it's gone, it's gone for good, until you enable it back via filters?13:31
seb128pstolowski,  http://ubuntuone.com/0RPwctnL0nRftR9hWFTMuS is what happens from an user view13:31
seb128pstolowski, I just noticed, thanks to mhr3, that the homescreen has filters and that only "help" was selected13:32
pstolowskiseb128, and you're hitting esc on that screencast?13:33
seb128pstolowski, yes13:33
seb128easiest way to clean the text entry13:34
pstolowskiseb128, can you report it?13:36
seb128pstolowski, sure, is there enough info in that description to make an useful bug?13:36
seb128pstolowski, I never reported it because I don't know how I land in the buggy state and mhr3 always said he didn't have enough info/a clue how to debug13:37
mhr3i think there are two bugs actually13:37
mhr3one that makes the home scope save an odd state of the filters for empty searches (not requested by the user)13:38
mhr3and another where you're getting results you shouldn't be13:38
mhr3so for example you should have never seen app results if they're disabled13:38
mhr3yet you somehow do from time to time13:39
mhr3(with empty search string)13:39
seb128right13:39
seb128seems like doing "one" -> esc was giving me a view ignoring the filters13:40
mhr3yet i can't reproduce that here13:40
Saviqmzanetti, reject https://code.launchpad.net/~unity-team/unity8/fix-1238232/+merge/191424 ?13:40
mzanettiSaviq: imo yes13:41
=== alan_g|lunch is now known as alan_g
Saviqmzanetti, please do13:41
mzanettiajmitch: done13:41
mzanettiSaviq: ^13:42
mzanettisorry ajmitch13:42
Saviqmhr3, pushed tweaks to debian/control - builds in chroot now13:54
mhr3Saviq, great, thx a lot13:59
mhr3didrocks, now we just need to push it in t :)13:59
mhr3well... once the unity branch is merged13:59
Saviqtsdgeos, yay, we got +1 on the bugs :D14:03
tsdgeosyeah14:03
tsdgeosi've been wondering were tronical was14:03
tsdgeosbeen trying to reach him on irc to no avail14:03
tsdgeosi randomly remember someone telling me he is/was on paternal leave14:04
tsdgeosbut may be a different guy it was about14:04
tsdgeosmzanetti: what's the units.gu(5) in  lp:~mzanetti/unity8/fix-switching-previews-positioning  ?14:05
mzanettitsdgeos: yeah. that sucks. it's the height of the section header. Is there a way to get to it?14:05
tsdgeosmzanetti: category header?14:05
tsdgeosshould be possible14:05
tsdgeoslet me see14:06
tsdgeosmzanetti: can't you just use sectionDelegate.height?14:07
tsdgeosah no it's a qqmlcomonent14:08
tsdgeosdamnit14:08
Saviqtsdgeos, yeah, you can't get to the delegate14:08
Saviqtsdgeos, mzanetti, only thing I can think of: use a prop on the *view14:08
Saviqdednick, we need https://code.launchpad.net/~nick-dedekind/unity8/StrFTimeFormatter/+merge/192343 so that we support non-qt time format strings?14:09
tsdgeosmzanetti: are you trying to do with that, is it something regarding the clipping of the top section header?14:10
pstolowskiseb128, sorry, I was itm. yes, I think whatever you described so far is enough for a bug report, it's important we don't forget about it and try to repro14:10
mzanettitsdgeos: with what exactly? the header height?14:11
seb128pstolowski, ok, I'm going to file it in a bit14:11
tsdgeosmzanetti: yep, i see you have that max in there14:11
mzanettitsdgeos: I make sure the split for the openEffect doesn't happen at the top where the content could be hidden by the header14:11
tsdgeosright14:11
tsdgeosok14:11
mzanettitsdgeos: let me add a comment in all places where I use the 5 grid units14:12
tsdgeoswe either need a context property that chaneges all the time (to not actually change)14:12
tsdgeosor introduce the idea of "i promise i'll be well behaved" and just have one value14:13
tsdgeosor not sure :D14:13
tsdgeosbut that units.gu(5) looks to fragile tbh14:13
tsdgeostoo14:13
mzanettiyes. I agree14:13
dednickSaviq: yeah, i think it's the alarm indicator that gives time format in gdatetime (strftime)14:16
mzanettitsdgeos: such a context property would be per delegate, right?14:17
Saviqdednick, any idea why it wouldn't give up a preformatted string?14:17
Saviqdednick, same as we expect translated strings and such...14:17
tsdgeosmzanetti: we can see how to expose it14:17
tsdgeosone per delegate14:17
tsdgeosor one just for the top section delegate14:17
tsdgeosbut yeah one per delegate makes sense14:18
dednickSaviq: tz changes14:18
dednickpossibly14:18
mzanettitsdgeos: just like the other section.* properties probably14:18
mzanettisection.headerHeight14:18
Saviqdednick, well, wouldn't the indicator service know?14:18
tsdgeosmzanetti: we don't really have secion. properties but yes14:19
tsdgeoswe have "section" and "delegateIndex"14:19
tsdgeosmzanetti: do oyu want to have a look at it?14:20
mzanettitsdgeos: on it14:21
didrocksmhr3: what's going to dep on unity-scopes-impl-0? (priority should be optional btw and I prefer libs just after just after priority)14:21
dednickSaviq: it would then have to repoll all the calendar events from the source when a change occurs. Plus I think it's better that the view decides how to display the data. I don't even think the dackend supplying the format is right...14:21
Saviqdednick, top-approve https://code.launchpad.net/~aacid/unity8/noModelInSignal/+merge/194672 ?14:21
Saviqdednick, right, it feels weird to get the format string from the backend14:22
dednickSaviq: yeah.. i know, we will probably override it anyway at some point.14:23
dednickSaviq: Do you know when CI is going to be back?14:24
Saviqdednick, like... last Tuesday...14:26
Saviqdednick, there were issues, guys are working day/night to get everything back14:26
dednickSaviq: but there's no jenkins approve right? because services are down...14:27
dednick?14:27
Saviqdednick, yes14:27
dednickSo how does CI work with no jenkins?14:28
Saviqdednick, does it?14:28
dednick<dednick> Saviq: Do you know when CI is going to be back?14:28
Saviqdednick, I meant it was supposed to be back Tuesday14:29
Saviqdednick, but it's not :)14:29
dednickoh.. right :)14:29
tsdgeosguys i need someone to review/topApprove this too https://code.launchpad.net/~aacid/unity8/lvwph_bad_header_position_1240118/+merge/19320014:30
tsdgeosSaviq: standip?14:32
tsdgeoskgunn: ↑↑14:33
Saviqcrap14:33
Saviqgreyback, got that cron script yet? ;P14:33
greyback:)14:33
dednicktsdgeos: Just testing your branch now14:35
greybackmterry: how usable is nested mir right now? Can it be used inside xmir?14:36
mterrygreyback, yeah I think so.  There are some u-s-c branches that need to land for Touch, but I don't think they'd affect desktop+xmir14:37
mhr3didrocks, unity8 deps on the impl-014:38
didrocksmhr3: but the scope plugin deps on unity-api-impl-014:38
didrockswhich is implemented by unity814:38
didrocksseems like a circular dep to me14:38
greybackmterry: hmm, there instructions anywhere on how to do it?14:38
mhr3didrocks, no it doesn't, it provides it14:39
mterrygreyback, with xmir, I don't know.  I believe in current trusty, it'd be as simple as installing unity-system-compositor?14:39
=== dandrader is now known as dandrader|afkl
=== dandrader|afkl is now known as dandrader|afk
greybackmterry: ok. Might play around with it this evening14:40
mhr3didrocks, well, it provides unity-scopes-impl, it will build-dep on unity-api-dev at some point in the future14:40
didrocksmhr3: and what will dep on the unity-scopes-impl?14:41
mhr3didrocks, unity8 does :)14:41
mhr3it's not the deps that are circular, this discussion is :P14:42
didrocksmhr3: but both are going to depend on libunity-api0, right?14:43
mhr3didrocks, probably, yes14:43
mhr3if we find some actual code to put in there14:44
didrocksI think there is something fuzzy in the way you handle the deps TBH14:44
mhr3it might end up header-only pkg14:44
mhr3didrocks, fuzzy in what way?14:44
=== greyback is now known as greyback|away
mhr3didrocks, it's simple, unity8 deps on scopes-impl, scopes-plugin provides it14:45
mhr3that's all that matters right now14:45
didrocksmhr3: you expect to have other packages providing scopes-impl?14:45
mhr3hm14:45
mhr3i don't think so, but one binary pkg could provide different versions14:46
mhr3s/different/multiple/14:46
didrocksand you expect different unity8 which can be installed in parallel?14:47
didrocksand supporting different versions?14:47
Saviqdidrocks, no14:47
didrocksso I don't think this -impl is needed14:47
mhr3well ultimately Saviq wanted it this way, he'll know why :)14:47
Saviqdidrocks, if you applied that logic we wouldn't need it anywhere14:47
Saviqdidrocks, until someone else did another implementation14:48
didrocksSaviq: well, maybe that was a mistake, but on the other pieces, you told me that you would see other implementations14:48
didrocksbeing compatible with unity814:48
didrockshere, it doesn't seem to be the case from my question ^14:48
Saviqdidrocks, sure, that's the idea - doesn't mean *we* would provide them14:48
Saviqdidrocks, there could be a mock -impl14:49
Saviqdidrocks, used for ap tests14:49
dednicktsdgeos: you getting that crash in test?14:49
tsdgeosyep14:49
didrocksok, so we will have multiple implementations14:49
didrocksand so, it makes sense in that way :)14:49
dednicktsdgeos: ok. cool. thought i might be going crazy14:49
didrocks(even a mock is just an implementation)14:49
Saviqdidrocks, good :)14:49
tsdgeosdednick: i'm pretty sure i wasn't14:49
tsdgeosbut now i am14:49
tsdgeosso that's eough :D14:50
dednickhehe14:50
dednickCimi:14:54
dednickCimi: ping14:54
Cimidednick, pong14:55
=== dandrader|afk is now known as dandrader
dednickCimi: you got my bug I see. have you checked it out?14:55
didrocksmhr3: did you change on my other (small) requests?14:55
Ciminot yet back14:55
dednickback wherE?14:55
dandradertsdgeos, wjat14:58
dandradertsdgeos, in qt, what's dev branch as opposed to stable branch?14:58
tsdgeos5.314:58
dandraderhmm14:58
mhr3didrocks, oh the priority? fixing14:59
mhr3didrocks, not sure what you meant with the other thing... Build-Deps should be directly after Prio?15:01
didrocksmhr3: Prio, Section, Build-Deps15:01
mhr3didrocks, pushed15:02
didrocksmhr3: <3 thanks!15:02
mhr3sil2100, how are looking on the capnp and zmqpp front?15:04
dednickyikes. app scope is scary15:04
dednickthe view i mean15:04
sil2100mhr3: capnp is in universe already, zmqpp is ready in my PPA - Ken was to make a review of it and *maybe* sponsor ;)15:04
sil2100mhr3: while zmq3 is in the archive already15:05
=== dandrader is now known as dandrader|lunch
=== alan_g is now known as alan_g|tea
mhr3sil2100, awesome, do we bother you to setup ci and autolanding and stuff for the new scopes lib then?15:27
sil2100mhr3: which one it will be?15:27
mhr3sil2100, lp:unity-scopes-api15:28
sil2100mhr3: once all the infra is back, I'll take care of doing that stuff ;)15:28
mhr3sil2100, great15:29
mhr3so now i'll just bother kenvandine to sponsor zmqpp :)15:29
=== alan_g|tea is now known as alan_g
kenvandinehey mhr315:31
kenvandinesil2100, oh... sorry forgot to sponsor that :)15:31
mhr3that was quick :)15:32
mhr3thx kenvandine :)15:32
sil2100kenvandine: I updated it a bit, fixing some stuff :)15:33
sil2100kenvandine: now the debian/watch file is more correct and the build-deps are better15:33
sil2100kenvandine: there is a package built out of it in ppa:sil2100/ppa if you want to check15:33
sil2100kenvandine: thanks! :)15:33
sil2100kenvandine: (I'll create lp:zmqpp/ubuntu once you give a green light)15:34
mhr3didrocks, unity8 is now seeded in default trusty images? it seems to bring lots of the phone-only stuff15:35
mhr3or well.. stuff that was meant phone-only15:36
didrocksmhr3: hum, it shouldn't, where do you see it seeded?15:36
mhr3didrocks, i'm assuming cause thomas mentioned that there's mediascanner and related stuff in T15:37
didrocksmhr3: they are in T, but universe, I think they are not seeded15:39
didrocksunity8 as well is in universe15:39
didrocksso very very very very… (very?) few chance it's in the iso15:39
didrocksor there is a big bug15:39
mhr3didrocks, hmm, i'll follow up with thomas then15:39
=== greyback|away is now known as greyback
dednickgraaa. my phone wont start after a flash :(15:47
kenvandinesil2100, mhr3: sponsored, it's in NEW now15:54
sil2100kenvandine: waaaa!15:54
sil2100kenvandine: thank you!15:54
kenvandinen[15:55
kenvandinenp15:55
mhr3yey, thx kenvandine15:55
mhr3Saviq, didrocks, could we do the transition to the separate scopes plugin tomorrow or... since it won't be possible to do that automatically?15:56
didrocksmhr3: better to transition once CI is back15:58
didrocksmhr3: can you file that in the landing ask? so that we don't loose it from sight15:58
mhr3didrocks, but CI won't be able to test nor merge that15:58
didrockswhy?15:58
didrockswe'll need to land your packages15:58
didrockswhich is cu2d/daily release15:58
mhr3didrocks, cause it's a new pkg that replaces part of old pkg?15:58
didrockswhich is down as well due to ci15:58
didrocksyeah, we need to land unity8 + scope15:59
mhr3exactly15:59
mhr3so it needs to be done manually, no?15:59
didrockswell, it will be done with cu2d15:59
Cimidednick, what shall it do?15:59
Cimidednick, the carousel?15:59
mhr3didrocks, hm, ok, i guess i underestimated it :)15:59
didrocks:)16:00
Cimion click and hold16:00
didrocksmhr3: just write in the landing ask please ;)16:00
didrocksotherwise, I'll forget for sure16:00
mhr3didrocks, i don't think i have perms for that16:00
didrocksmhr3: ask thostr_ then, he's a specialist16:01
mhr3heh, k16:01
mhr3thostr_, could you? landing ask to land updated unity8 + unity-scopes-shell16:01
mhr3needs to be done in sync, cause the latter replaces part of the former16:01
didrocksunity-scopes-hell :)16:02
mhr3didrocks, you should be glad we still manage to provide you with some challenges! :)16:02
didrockshehe!16:03
Saviqmzanetti, for you https://code.launchpad.net/~mzanetti/unity8/music-preview/+merge/193803/comments/45034716:06
Saviqmhr3, so it doesn't have to be done in sync16:07
Saviqmhr3, because until you upgrade unity8, nothing will pull the new package in16:07
mzanettiSaviq: thanks16:08
mhr3Saviq, right, that actually makes sense :)16:08
dednickCimi: when i click on a result on the carousel, nothing happens....16:17
dednickso should probably either open a preview, or activate the result presumably16:18
tsdgeosdednick: that crash is nasty :-S16:46
tsdgeosdednick: the listview seems it's not getting deleted  when the parent goes away16:46
tsdgeosand then all kind of funky stuff happens16:47
dednick_tsdgeos: did you get the reply?16:56
mhr3according to my irc he didn't :)16:57
dednicktsdgeos: hm. sounds like  problem i've seen before. Is the listview deleted during a glib callback? eg. dee ?16:57
tsdgeosdednick: hmm16:58
tsdgeosdoes look like16:58
tsdgeoswhy?16:58
dednickanything that is deleted by calling deleteLater from an interal glib callback will not be deleted.16:59
tsdgeoshmmmm17:00
tsdgeoslil check17:00
mhr3ehm, that sound like it would introduce lots of problems17:01
dednicktsdgeos: i found it with delegates being deleted from a listview17:01
dednickmhr3: yeah...17:01
mhr3dednick, but i mean random crashing all the time, and we do not see that... fortunately17:01
mhr3cause afterall any external event is reported via a glib callback17:02
mhr3well... lots of them anyway17:02
dednickwell it was in the indicators, but i fixed it. Dont know how dee-qt works17:02
mhr3dee-qt will emit q-signals from within callbacks17:02
dednickmhr3: not problem with external events. It's only if we pick up through a callback directly17:03
mhr3dednick, so how does the call stack have to look like for it to be an issue?17:04
dednickmhr3: https://bugreports.qt-project.org/browse/QTBUG-3285917:04
dednickmhr3: tbh, i'm supprised we don't see problems with dee17:08
mhr3well, dee is not using timeout nor idle sources17:09
mhr3but i still don't fully get why that is a problem17:09
dednickmhr3: using g_signal17:09
dednickmhr3: it's to do with the qt event loop17:10
mhr3yea, but i don't understand the interaction that causes the issue17:10
mhr3dednick, hm, so if glib dispatches a source directly without qt knowing about it, deleteLater() doesn't work?17:15
mhr3...if you call it from inside that dispatch17:16
mhr3is that right?17:16
seb128mhr3, pstolowski: oh, I had open a bug about that back then: https://bugs.launchpad.net/ubuntu/+source/unity-scope-home/+bug/122393317:17
ubot5Ubuntu bug 1223933 in libunity "sometimes the dash home list "no result matching your query" string" [Undecided,Triaged]17:17
dednickmhr3: not only a source. any async callback that goes through the gmainloop.17:17
mhr3async callback is an idle source ;)17:17
mhr3but that is... really bad17:18
dednickmhr3: a gaction callback is a idle source ?17:19
mhr3i'm not into gactions really, but they sound like dbus signals and dbus signals are delivered via idle sources17:20
dednickmhr3: ok. then yes :)17:20
dednickmhr3: and yeah, it's bad. I've looked at the dee code, and it looks like it just goes directly from glib to qt.17:21
mhr3indeed, it responds to dbus signals :)17:21
mhr3so again idle sources17:21
dednickyup17:21
mhr3maybe someone who likes compiling qt should add some debug code in deleteLater and figure out how bad are we talking about17:22
mhr3but tsdgeos ran off :P17:23
dednickmhr3: what's really bad is when it happens during an infinite animation. It never ends. So cpu stays at 100%.17:25
dednickwhich is what was happening in indicators17:25
pstolowskiseb128, ack, thanks, will try to repro17:25
mhr3dednick, well sounds like we should come up with a real patch17:26
seb128pstolowski, thanks, same here17:26
mhr3dednick, not just workarounds like you did with indicators17:26
mhr3and upstream qt apparently doesn't care17:27
dednickmhr3: yeah, i was meaning to look into it, but never got there...17:27
dednickthen forgot...17:28
mhr3dednick, wondering now what was it that you were complaining about yesterday :P17:28
dednickheh17:28
mhr3dednick, but yea, talk to tsdgeos about it tomorrow, he likes patching qt :)17:29
dednickmhr3: aiight17:30
mhr3seb128, thanks for updating it17:33
seb128mhr3, yw ;-)17:33
=== dandrader|lunch is now known as dandrader
mhr3dednick, https://bugreports.qt-project.org/browse/QTBUG-18434?focusedCommentId=173857&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-17385717:54
mhr3scary17:54
dednickmhr3: heh, nice find17:56
mhr3dednick, it's actually linked from your bug :)17:56
dednickdednick: i'm not even sure it's possible to patch. It's a pretty fundimental issue17:56
dednicktalking to myself again17:56
mhr3well, their stance is - anything like this should be posted to qt event loop17:57
dednickyeah17:57
mhr3but that'd be massively expensive for things like dee17:57
mhr3so i think we should just write a downstream patch17:58
dednickwhich is probably their way of saying its really diffucult to fix...17:58
mhr3while trying to not break too much stuff :)17:58
dednickmhr3: no need to post, can just send.17:58
mhr3dednick, is there a difference?17:58
dednickpost is async17:58
mhr3ah17:59
mhr3yea, send would work17:59
mhr3still, quite a lot of patching though17:59
dednickyeah.. it was a pain just for the unitymenumodel.17:59
mhr3indeed, that's why i'd rather go the nasty downstream patch route18:00
mhr3anyway.. eod, cyas18:02
dednicki dont think it can be fixed. the only way i could think is pushing the g_main_loop and using another context, but it doesnt work for g_idle.18:02
dednickmhr3: ^ thats for qt patch i mean18:02
=== boiko_ is now known as boiko
dednickmhr3: using pushing the g_main_context_push_thread_default i mean.18:03
mhr3everything's hackable18:04
dednickmhr3: heh.18:04
dednickmhr3: ok, i'm off too18:04
dednickmhr3: you in office tomorrow?18:05
mhr3dednick, yea, for the afternoon ;)18:05
dednickmhr3: :) ok18:05
mhr3got meeting at 8, no way i'd get to the office so early :)18:05
=== alan_g is now known as alan_g|EOD
=== dandrader is now known as dandrader|afk
=== dandrader|afk is now known as dandrader
=== dandrader_ is now known as dandrader
=== salem_ is now known as _salem
veebersrobru: ping22:39
Saviqveebers, hey, just a ping about https://code.launchpad.net/~veebers/unity8/ap_make_use_of_helpers_in_tests/+merge/191575 so you don't forget ;)22:48
veebersHi Saviq22:48
* veebers looks22:48
veebersSaviq: ugh, that one slipped through my fingers somehow :-\ I'll get that updated and sorted today. Thanks for that22:49
Saviqveebers, cheers22:49
veebersSaviq: also fyi, it looks like I was too slow getting another of my MRs reviewed: https://code.launchpad.net/~veebers/unity8/update_ap_tests_ready_for_py3/+merge/194655 vs https://code.launchpad.net/~robru/unity8/autopilot-py3k/+merge/19512822:49
Saviqveebers, yeah, just saw that22:50
Saviqveebers, that's 'cause it's marked WiP ;)22:50
veebersSaviq: ugh, d'oh. My bad again :-|22:50
Saviqveebers, robru, lp:~veebers/unity8/update_ap_tests_ready_for_py3 seems to work fine, lp:~robru/unity8/autopilot-py3k complains:23:02
SaviqTraceback (most recent call last):23:03
Saviq  File "/home/michal/dev/canonical/unity8/repo/tests/autopilot/unity8/shell/tests/__init__.py", line 111, in setUpClass23:03
Saviq    if "start/" in output:23:03
SaviqTypeError: Type str doesn't support the buffer API23:03
Saviqfor each test23:03
Saviqso please guys fight amongst each other which one should be merged ;)23:03
veebersSaviq: :-)23:03
veebersrobru: that would be due to the subprocess call needing: universal_newlines=True23:03

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