=== daek_ is now known as Daekdroom | ||
=== duflu_ is now known as duflu | ||
=== marcusto_ is now known as marcustomlinson | ||
=== marcusto_ is now known as marcustomlinson_ | ||
=== marcustomlinson_ is now known as marcustomlinson | ||
=== marcusto_ is now known as marcustomlinson | ||
tsdgeos | pstolowski: Saviq: should the unity-scope-tool package depend on pkg-config and libunity-scopes-dev ? | 08:42 |
---|---|---|
tsdgeos | tools/registry-tracker.cpp is invoking pkg-config | 08:42 |
Saviq | tsdgeos, then sounds like it should | 08:42 |
tsdgeos | k, will prepare a MR | 08:43 |
Saviq | tsdgeos, but why on libunity-scopes-dev? | 08:43 |
tsdgeos | Saviq: because it's the thing it's pkg-config'ing | 08:43 |
tsdgeos | arguments << "--variable=scopesdir"; | 08:43 |
tsdgeos | arguments << "libunity-scopes"; | 08:43 |
tsdgeos | pkg_config.start("pkg-config", arguments); | 08:43 |
pstolowski | otp | 08:45 |
Saviq | tsdgeos, ugh, shouldn't it be doing that build time? | 08:46 |
tsdgeos | don't know :D | 08:46 |
Mirv | Qt 5.4.1 breaking stuff!!11 ...not really, seems pretty smooth, but qtmir test is broken probably because a dummy qtsensors plugin went away bug #1427529 | 08:49 |
ubot5 | bug 1427529 in qtmir (Ubuntu) "qtmir fails a test against Qt 5.4.1" [Undecided,New] https://launchpad.net/bugs/1427529 | 08:49 |
Saviq | Mirv, thanks | 08:49 |
Mirv | tsdgeos: if you want to give a whirl, the PPA is now usable on phone | 08:49 |
Mirv | Saviq: I've assumed you're in some sort of deadline limbo lately, but I guess now actually you can be reached possibly again :) | 08:50 |
Saviq | Mirv, oh? nah, I was just in CPT the week before last | 08:51 |
Mirv | Saviq: ok, good that you're available for my eternal Qt pings too! | 08:57 |
Saviq | :) | 08:57 |
Saviq | brb | 08:57 |
Mirv | 5.4.1 should be smooth, of course, but you never know if we adjusted to some .0 bug already too hard. | 08:57 |
Saviq | seb128, hey, have you tried running the unity8 session recently? it started locking up before I can log in here :/ | 09:08 |
Saviq | @unity ↑ | 09:08 |
* Saviq grabs dbg symbols | 09:08 | |
seb128 | Saviq, hey | 09:08 |
seb128 | Saviq, likely https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1422835 | 09:08 |
ubot5 | Launchpad bug 1422835 in ubuntu-ui-toolkit (Ubuntu) "Unity8 stucks in icon loading code" [Undecided,Fix committed] | 09:08 |
Saviq | hmm right | 09:08 |
tsdgeos | didn't that land? | 09:09 |
seb128 | Saviq, there is an uitk fix in staging for 10 days, but uitk landings... no comment... | 09:09 |
Saviq | tsdgeos, it got committed to staging, not released | 09:09 |
seb128 | tsdgeos, no, there has been no uitk landing since mid-feb | 09:09 |
tsdgeos | Saviq: show the backtrace, that one is very disctintive | 09:09 |
Saviq | tsdgeos, comin' right up | 09:12 |
Saviq | tsdgeos, http://pastebin.ubuntu.com/10513059/ | 09:15 |
tsdgeos | hmmm | 09:16 |
tsdgeos | doens't look like it i'd say | 09:17 |
Saviq | yeah | 09:17 |
* Saviq gets more symbols | 09:17 | |
tsdgeos | Saviq: are we blocked on pam? | 09:18 |
Saviq | tsdgeos, ah, there's a branch from mterry that might be helping there | 09:18 |
Saviq | although that was a crash | 09:18 |
Saviq | https://code.launchpad.net/~mterry/unity8/cancel-pam-harder/+merge/251174 | 09:18 |
* Saviq tries anyway | 09:18 | |
tsdgeos | don't know saw two threads in pam-like waits | 09:19 |
Saviq | yeah | 09:19 |
pstolowski | tsdgeos, +1 for adding these dependencies to unity-scope-tool; i can't find good explanation for the pkg-config invocation there, could very well be determined at build-time; perhaps mhr3 had a reason to do it that way back then | 09:43 |
=== Malsasa_ is now known as Malsasa | ||
tsdgeos | Saviq: pstolowski: https://code.launchpad.net/~aacid/unity8/scope-tool-dependencies/+merge/251573 | 10:16 |
pstolowski | tsdgeos, looks good; is # comment allowed in debian/control? | 10:25 |
tsdgeos | pstolowski: yep, there's others and it did build :D | 10:26 |
pstolowski | cool :) | 10:27 |
tsdgeos | greyback: did we release the fix for the startTimer in another thread thing? | 10:47 |
tsdgeos | i think we did | 10:48 |
tsdgeos | i'm getting some others when launching apps | 10:48 |
greyback | tsdgeos: yep we did | 10:48 |
tsdgeos | but on the app process not in unity8 | 10:48 |
tsdgeos | so may be different issue | 10:48 |
greyback | yeah | 10:48 |
greyback | I guess it wasn't just qtmir that was making that mistake | 10:49 |
Saviq | dandrader, hey, here's the mirevent 2.0 silo http://people.canonical.com/~platform/citrain_dashboard/#?distro=ubuntu&q=landing-006 | 11:05 |
Saviq | dandrader, I had some doubts about package dependencies there, added comments to the relevant MPs | 11:06 |
Saviq | /food | 11:07 |
dandrader | Saviq, ok | 11:10 |
dandrader | Saviq, I saw your comments yesterday. racarr solved them. Unless you made new ones today | 11:10 |
=== MacSlow is now known as MacSlow|lunch | ||
tsdgeos | pete-woods: can you review/top-approve https://code.launchpad.net/~aacid/thumbnailer/fix_dbus_blocking/+merge/251065 ? | 11:53 |
pete-woods | tsdgeos: it seems like a bunch of different people are trying to fix the same bug here | 11:54 |
tsdgeos | well the bug is assigned to me :D | 11:54 |
tsdgeos | pete-woods: who else do you know is trying to fix it? | 11:54 |
pete-woods | tsdgeos: could you liaise with michi about this? he's not the owner of thumbnailer | 11:54 |
pete-woods | tsdgeos: afaik kaleo is doing something funky with some new image provider codepath | 11:55 |
tsdgeos | pete-woods: s/not/now ? | 11:55 |
tsdgeos | pete-woods: oh yeah Kaleo patch is crazy :) | 11:55 |
pete-woods | tsdgeos: correct :) | 11:55 |
pete-woods | tsdgeos: also michi wants to fix this at a lower level inside the thumbnailer service | 11:55 |
tsdgeos | and at the "wrong" level too imho, https://codereview.qt-project.org/#/c/107427/ is a better fix | 11:56 |
tsdgeos | pete-woods: he can't fix that at the lower level | 11:56 |
tsdgeos | at least not this bug | 11:56 |
pete-woods | tsdgeos: I guess I'm just saying, I think we should talk to everyone involved | 11:56 |
tsdgeos | i mean this bug is the provider is broken | 11:56 |
pete-woods | so we get the right solution, which could well be yours | 11:56 |
tsdgeos | so it blocks the main ui | 11:56 |
tsdgeos | you can make it faster and what not | 11:56 |
tsdgeos | but still should not block the main ui | 11:56 |
pete-woods | tsdgeos: I'm not saying you're wrong. I just want only one person to fix this thing | 11:56 |
pete-woods | instead of 3 competing and probably interfering patches | 11:57 |
tsdgeos | sure | 11:57 |
pete-woods | your fix is appealing to me, because I actually understand the code :) | 11:57 |
tsdgeos | i added michi to https://code.launchpad.net/~aacid/thumbnailer/fix_dbus_blocking/+merge/251065 | 11:57 |
pete-woods | thanks! | 11:58 |
pete-woods | if this is all that is required to fix the overall issue, that will save michi from diving down a large rabbit hole | 11:58 |
tsdgeos | this fixes "a lot" | 11:58 |
tsdgeos | of course doesn't fix the a slow thiumbnail "blocking" a quick one from appearing | 11:59 |
tsdgeos | since there's still only one thread for thumbnails | 11:59 |
tsdgeos | that'd be either Kaleo's or my patch | 11:59 |
tsdgeos | but it's not "the bug" itself | 11:59 |
pete-woods | right, so you don't paralellise, but at least you don't block | 11:59 |
tsdgeos | it's just a way to make it better | 11:59 |
pete-woods | I think it's really worth you having a conversation with michi to explain what exactly the main issue is, the details of how image providers work, etc. otherwise I worry he may do a load of complicated work with possibly little benefit | 12:01 |
tsdgeos | pete-woods: do you think it's best if i mail him? not sure if he'll read the MR mail | 12:01 |
pete-woods | tsdgeos: definitely | 12:01 |
tsdgeos | and we hardly co-indicide in time, no? | 12:01 |
pete-woods | probably the same as me | 12:01 |
pete-woods | he's there in the mornings | 12:01 |
pete-woods | but sure, it's hard | 12:01 |
pete-woods | just you know like a million times more than me about the image provider stuff | 12:02 |
pete-woods | I can only speculate, rather then know for certain | 12:02 |
pete-woods | I think he's planning a super parallelised vision of the thumbnailer | 12:02 |
tsdgeos | sure, i'll mail him | 12:02 |
pete-woods | and I want to make sure it will expose an API that is actually usable by the QML image provider engine | 12:02 |
pete-woods | and also that he doesn't waste 2 weeks of effort | 12:03 |
Mirv | tsdgeos: Saviq: any possibility of looking at Qt 5.4.1? keyboard seems broken, maliit-framework crashing. any problem like this without a good explanation makes it a bit more unlikely to get 5.4.1 in with the schedule tight.. | 12:10 |
Mirv | crashes at QV4::ExecutionContext::setProperty() | 12:10 |
tsdgeos | pete-woods: sent, hope i was clear enough :D | 12:15 |
tsdgeos | Mirv: which silo is it on? | 12:16 |
Mirv | tsdgeos: 012 | 12:20 |
Mirv | (https://wiki.ubuntu.com/Touch/QtTesting) | 12:21 |
tsdgeos | i'll try to have a quick look after lunc | 12:21 |
tsdgeos | h | 12:21 |
dandrader | Saviq, installed silo 006 on my N4 and all seems fine to me | 12:28 |
Saviq | dandrader, kk | 12:40 |
pete-woods | tsdgeos: thanks for that! :) | 12:50 |
=== MacSlow|lunch is now known as MacSlow | ||
=== dandrader is now known as dandrader|afk | ||
=== dandrader|afk is now known as dandrader | ||
tsdgeos | Mirv: the internal structures of qtdeclarative have changed in 5.4.1 vs 5.4.0, have we rebuilt everything that needs qtdeclarative5-private-dev to build? | 14:17 |
Mirv | tsdgeos: we should have, since everything that uses qtdeclarative5-private-dev, meaning a symbol from there marked so in the .symbols file, should depend on qtdeclarative-abi-5-4-0. so, maliit-framework is one of them and was rebuilt. | 14:21 |
Mirv | but with current knowledge that includes, in addition to Qt packages itself, only ciborium gsettings-qt maliit-framework qtmir qtubuntu ubuntu-ui-toolkit unity8 | 14:22 |
Mirv | I've tested browser, video and music playback, everything so far otherwise seems correct but the keyboard | 14:23 |
tsdgeos | the crash is weird | 14:23 |
tsdgeos | Mirv: how pressing is this? | 14:33 |
Mirv | tsdgeos: not pressing if you don't want 5.4.1 to vivid-rtm. in that case I'd land it only after the branching has happened. I just thought https://qt.gitorious.org/qt/qtbase/source/69196b38c481610ef30bfe8ce8e7ba6826729ab8:dist/changes-5.4.1 sounds pretty good. | 14:34 |
Mirv | tsdgeos: but also not pressing in the sense I'm gone Thu - Tue, ie back next Wed to see if 5.4.1 finalization can happen | 14:35 |
Mirv | but if wanted to vivid-rtm it'd need to be non-regressions of course and also the fixed regressions should be understood that they are not a cause of any risky changes in 5.4.1 | 14:35 |
tsdgeos | Mirv: ok | 14:35 |
Mirv | tsdgeos: filed bug #1427710 - and noted 5.4.0 -> 5.4.1 recompiles slight less than 5.3.2 -> 5.4.0. however, there are recompiles available in a separate PPA if paranoid, and I just upgraded eg ubuntu-keyboard without help | 14:51 |
ubot5 | bug 1427710 in maliit-framework (Ubuntu) "A maliit-framework segfault with Qt 5.4.1" [Undecided,New] https://launchpad.net/bugs/1427710 | 14:51 |
=== dandrader is now known as dandrader|lunch | ||
om26er | tsdgeos, Hi! | 16:20 |
om26er | tsdgeos, can you tell how can i verify the fix for bug 1410131 ? | 16:20 |
ubot5 | bug 1410131 in unity8 (Ubuntu) "Visibility ranges/buffer are not set correctly for the visible scope" [Medium,In progress] https://launchpad.net/bugs/1410131 | 16:20 |
tsdgeos | om26er: i'm going to go with you can't :/ | 16:21 |
tsdgeos | not without editing the code | 16:21 |
tsdgeos | to ouput some debug | 16:21 |
om26er | tsdgeos, hm, need moar unit tests :) | 16:22 |
om26er | tsdgeos, does it affect the phone or is it desktop specific ? | 16:23 |
tsdgeos | om26er: it does totally affect the phone | 16:25 |
tsdgeos | om26er: and yes, more unit tests there won't hurt | 16:26 |
om26er | tsdgeos, I have to verify the fix before landing, I am willing to edit some files if needed | 16:26 |
om26er | as long as its qml only (no compilation ;) | 16:27 |
tsdgeos | om26er: is this a task you need to do now? and how much qml do you know? i can give you the point were adding some code needs to be added, if you can carry on from there nice, if i need to give you the whole of qml code to add, i'm a bit blocked somehwere else atm | 16:30 |
Saviq | tsdgeos, this is QA validation of landings (as we have for RTM), enabled for vivid starting yesterday | 16:31 |
tsdgeos | right, just don't land that one :D | 16:32 |
tsdgeos | i honestly think it's still a bit risky and probably don't want to delay other landings because of this | 16:32 |
Saviq | tsdgeos, well, I already tested it, would have to rebuild and retest ;P | 16:32 |
Saviq | om26er, ultimately the validation is that you can't see the difference... | 16:34 |
Saviq | you could compare that mem usage is lower, right tsdgeos? | 16:35 |
Saviq | but it's not by a whole lot | 16:35 |
tsdgeos | Saviq: yes mem should be a bit smaller | 16:35 |
tsdgeos | for very long scopes | 16:35 |
om26er | Saviq, tsdgeos ok hope it doesn't introduce regressions. | 16:37 |
om26er | (and sorry for late response, the top indicator never blinked) | 16:37 |
tsdgeos | for a definition of regression it will introduce regressions, it changes the caching behaviour, so we have much less stuff on memory now | 16:37 |
tsdgeos | so it is possible that on fast flicks of long scopes the icons will have to be reloaded | 16:38 |
tsdgeos | if that's a regression of not | 16:38 |
tsdgeos | it's up to the definiion of a regression | 16:38 |
om26er | we can call that a "trade-off" | 16:39 |
tsdgeos | yep, and as Saviq now (and me tested back then when i made this) | 16:42 |
tsdgeos | you should not even see it | 16:42 |
om26er | tsdgeos, so need to confirm, are the qtmir changes in the silo only for desktop ? | 16:46 |
tsdgeos | om26er: do not know | 16:46 |
tsdgeos | Saviq: ↑↑ | 16:46 |
Saviq | om26er, qtmir isn't in the silo any more | 16:46 |
Saviq | or well... *should* not be | 16:47 |
Saviq | om26er, sorry, /me removes, but there is nothing in there affecting your testing | 16:47 |
=== dandrader|lunch is now known as dandrader | ||
Cimi | do we need to protect from division by zero in qml properties? | 17:26 |
dandrader | greyback, https://code.launchpad.net/~dandrader/qtubuntu/shellRotation/+merge/242215 is still on "needs fixing" from your side | 18:03 |
=== alan_g is now known as alan_g|EOD | ||
=== anpok_ is now known as anpok |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!