mzanetti | tsdgeos, good morning. how's the landing going? need any help? | 08:12 |
---|---|---|
tsdgeos | mzanetti: so been runnning the test plan | 08:15 |
tsdgeos | all good for now | 08:15 |
tsdgeos | mzanetti: the test plan doesn't mention autopilot tests | 08:17 |
tsdgeos | but i guess we still run them? | 08:17 |
mzanetti | tsdgeos, yes, please run them. ideally according to the doc Saviq linked yesterday... | 08:19 |
tsdgeos | we should link that from the test plan | 08:19 |
mzanetti | http://ubuntu-test-cases-touch.readthedocs.org/en/latest/ | 08:19 |
mzanetti | yes, lemme do that | 08:19 |
Saviq | tsdgeos, mzanetti, FYI: they added a 1001 priority for the overlay ppa, an add-apt-repository + upgrade might not work proper | 08:21 |
Saviq | the citrain tool should still work I think | 08:21 |
mzanetti | yep, using citrain tool | 08:21 |
Saviq | but pay close attention to what's upgraded | 08:21 |
mzanetti | ack | 08:21 |
Saviq | if doesn't work, drop the extra-ppas.pref file from /etc/preferences.d/ | 08:23 |
Saviq | mzanetti, tsdgeos, also, top tip: `sudo stop whoopsie`, this way unity8 crash on exit won't slow down testing | 08:24 |
mzanetti | hah! good one | 08:24 |
Saviq | (well, just the first time it will | 08:25 |
Saviq | ) | 08:25 |
tsdgeos | this ./scripts/run-autopilot-tests.sh doesn't work at all | 08:43 |
tsdgeos | asks me for the password of the phone when doing things | 08:43 |
tsdgeos | and gets stuck there | 08:43 |
Saviq | tsdgeos, did you use the provision script too? | 08:57 |
tsdgeos | i think i did | 08:57 |
tsdgeos | but maybe not | 08:57 |
tsdgeos | anyhow added myself as ultra sudoer and contienued | 08:58 |
tsdgeos | now it's stuck somewhere else | 08:58 |
tsdgeos | ah no it's actualyl doing stuff | 08:58 |
=== greyback__ is now known as greyback | ||
MacSlow | Saviq, can't get the test to run now (image 191 on mako) was fine yesterday with 190. | 10:16 |
MacSlow | Saviq, which silo do you use for testing? | 10:18 |
Saviq | MacSlow, demo-stuff, you probably did not get the upgraded bits because of a package pin | 10:18 |
Saviq | MacSlow, drop /etc/apt/preferences.d/extra-ppas.pref, apt update and dist-upgrade again | 10:18 |
MacSlow | trying that now... | 10:20 |
tsdgeos | pstolowski: ping | 10:58 |
pstolowski | tsdgeos, pong | 10:59 |
tsdgeos | pstolowski: you know anything about https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1324142 ? | 10:59 |
ubot5 | Ubuntu bug 1324142 in unity8 (Ubuntu) "Support providing fallback images" [High,In progress] | 10:59 |
tsdgeos | i'm doing it in the unity8 side | 11:00 |
tsdgeos | and was wondering if we need to cater for this usecase | 11:00 |
tsdgeos | widgetData["source"] is invalid, so we load widgetData["fallback"] | 11:01 |
tsdgeos | can widgetData["source"] change after that? | 11:01 |
tsdgeos | because at the moment if it changes, i ignore it and stay in widgetData["fallback"] | 11:01 |
dandrader | greyback, as for lifecycle in unity8, the non-trivial part is when you switch between usage modes. on staged -> windowed, I don't think we want to resume all the suspended applications at once. I thinkit would be better to resume them as they get focused | 11:04 |
pstolowski | tsdgeos, change how? like a thumbnail was generated in the meantime? | 11:04 |
tsdgeos | like the model gets changed | 11:04 |
tsdgeos | i guess no right | 11:04 |
tsdgeos | we don't have updating models | 11:05 |
greyback | dandrader: fair enough | 11:05 |
pstolowski | tsdgeos, that would be a new search | 11:05 |
tsdgeos | which would mean starting over and stuff | 11:05 |
pstolowski | tsdgeos, yeah | 11:06 |
tsdgeos | so i guess i'll go with the current code, now i have to convince mzanetti :D | 11:07 |
pstolowski | tsdgeos, btw, please check with michi as the new thumbnailer is in the works. may be relevant for error cases? | 11:07 |
tsdgeos | that's fine | 11:07 |
tsdgeos | the can just return QImage() instaead of the fallback | 11:07 |
tsdgeos | but there's a branch about that | 11:07 |
tsdgeos | we talked about it already in london | 11:08 |
pstolowski | tsdgeos, ah, cool | 11:08 |
MacSlow | Saviq, even after updating (and rebooting) I'm getting the "no process found with pid..." errors | 11:10 |
Saviq | MacSlow, did you remove ~/.cache/QML? | 11:10 |
MacSlow | Saviq, yup always do that before each test-run | 11:10 |
MacSlow | Saviq, no crash... no syslog-entry... | 11:11 |
MacSlow | Saviq, ok... I didn't dump all DBus-traffic yet... | 11:11 |
Saviq | never do that, it's useless | 11:11 |
Saviq | MacSlow, just check that you have the correct version of unity8 installed | 11:11 |
Saviq | and all the other packages from the silo | 11:12 |
MacSlow | Saviq, which one... in sources.list.d I've only unity-team/demo-stuff and ci-train-ppa-service/stable-phone-overlay | 11:15 |
Saviq | MacSlow, apt-cache policy unity8? | 11:15 |
MacSlow | Saviq, lists unity8 8.10+rot1676+ap1615~ubuntu15.04.1 | 11:15 |
Saviq | MacSlow, full output please | 11:16 |
MacSlow | Saviq, http://pastebin.ubuntu.com/10996651 | 11:16 |
Saviq | MacSlow, ok, you need to find out why unity8 exits/crashes, sorry but I won't be able to help | 11:17 |
Saviq | in any case, it's something in your environment, I've never had that issue on either krillin or mako | 11:18 |
MacSlow | Saviq, I'll dig into it | 11:18 |
Saviq | MacSlow, obvious place to look is `autopilot run -v...` and ~/.cache/upstart/unity8.log | 11:21 |
=== dandrader is now known as dandrader|afk | ||
=== JMulholland_ is now known as JMulholland | ||
=== MacSlow is now known as MacSlow|lunch | ||
=== dandrader|afk is now known as dandrader | ||
mzanetti | dandrader, hey, do you know why uqmlscene is stuck on 20% cpu all the time? | 12:04 |
dandrader | mzanetti, never noticed it | 12:04 |
* dandrader checks | 12:04 | |
mzanetti | dandrader, it depends on the size of the window how much it uses, but it constantly spins the cpu to some degree | 12:05 |
mzanetti | as if it would keep on repainting all the time | 12:05 |
dandrader | mzanetti, I get 0% CPU here on an idle "make tryShell" with lp:unity8 | 12:06 |
mzanetti | odd | 12:06 |
* mzanetti tries with a clean trunk | 12:07 | |
mzanetti | dandrader, with clean trunk and make tryShell I get like 7% ide cpu usage. if I maximize the window it goes up to 10% | 12:23 |
dandrader | mzanetti, yeah, I see it now. top shows around 5% on a maximized window | 12:25 |
dandrader | mzanetti, I think there are some QT env vars that will print out render loop info, right greyback? | 12:25 |
mzanetti | there is... can't remember it atm | 12:26 |
greyback | dandrader: yeah, QSG_RENDERER_DEBUG=render,build,change,upload,roots,dump,noalpha,noopaque,noclip | 12:26 |
greyback | pick & choose the log you want | 12:26 |
greyback | QSG_RENDER_TIMING=1 just to prove it is renderering continually | 12:27 |
mzanetti | yep. it is constantly rendering | 12:27 |
greyback | QSG_VISUALIZE=changes will show why, if it is visible | 12:29 |
mzanetti | seems invisible :/ | 12:30 |
=== MacSlow|lunch is now known as MacSlow | ||
greyback | QSG_RENDERER_DEBUG=change then will indicate what of the scenegraph changed requiring a render | 12:31 |
greyback | harder to figure out what qml element that corresponds to though | 12:31 |
greyback | Gammaray might be able to help | 12:31 |
mzanetti | hmm... that's not printing anything | 12:32 |
greyback | no? but it is rendering? would there be an Animator working? | 12:35 |
mzanetti | not that I know of... it happens with all tryXXX things | 12:36 |
mzanetti | so maybe the c++ code calls update() in a loop or so | 12:36 |
greyback | lemme see | 12:36 |
mzanetti | anpok, hey, could you please merge ~mir-team/unity-system-compositor/toggle-cursor and trigger a rebuild in silo0 | 12:41 |
dandrader | mzanetti, we want a property in Application so that unity8 can use it to suspend/resume apps. Properties are good instead of funcitons (suspend(), resume()) because you can write declarative (bindings!) code | 13:02 |
mzanetti | dandrader, is that a question? | 13:02 |
dandrader | mzanetti, I want to avoid this: http://paste.ubuntu.com/10996980/ | 13:02 |
dandrader | mzanetti, so I was discussing with greyback a name for that property | 13:02 |
mzanetti | yeah, definitely +1 on properties | 13:02 |
dandrader | mzanetti, my suggestion is Application.requestedState | 13:03 |
greyback | dandrader: let's give him some context | 13:03 |
dandrader | mzanetti, but it has the drawback that setting it to Starting doesn't make sense (ie, Application will ignore it) | 13:03 |
dandrader | mzanetti, and that greyback said there's no use for requesting the Stopped state | 13:03 |
greyback | mzanetti: we're discussing an API for application running state. We have a readonly property for each application which indicates: starting, running, suspended, stopped | 13:04 |
dandrader | mzanetti, so unity8 really only needs to switch between Running and Suspended | 13:04 |
dandrader | mzanetti, so a boolean would suffice. but we didn't find a good name for a boolean | 13:04 |
greyback | mzanetti: what we want is to have a way for shell to suggest to qtmir that a certain app can be suspended or not | 13:04 |
mzanetti | property bool canSuspend | 13:05 |
mzanetti | perhaps | 13:05 |
mzanetti | or rather, suspendAllowed | 13:05 |
dandrader | canSuspend, suspendAllowed implies that this is just a suggestion, not a request | 13:05 |
mzanetti | yeah well, isn't that what it is? | 13:05 |
mzanetti | qtmir still does what it feels like, i.e. ignoring it because of the whitelist | 13:06 |
dandrader | mzanetti, no, unity8 is controlling the lifecycle. it wants/needs and app to suspend or resume | 13:06 |
greyback | "can be suspend" is a suggestion. "should be running right now" is a demand | 13:06 |
dandrader | not just suggest/hint it | 13:06 |
greyback | but yeah, I agree that a boolean is enough | 13:06 |
mzanetti | then lets invert it... active or something | 13:06 |
mzanetti | whereas active == true indicates it must not be suspended | 13:07 |
mzanetti | active == false means qtmir might suspend it | 13:07 |
dandrader | the term "active" is taken | 13:07 |
dandrader | I think | 13:07 |
mzanetti | :D | 13:07 |
dandrader | the good thing about requestedState is that it obvious it affects the read-only state property | 13:07 |
mzanetti | well... I'm sure we can find a name for that property, bottomline is, I'm voting for a boolean too | 13:07 |
dandrader | naming is the hardest part | 13:08 |
mzanetti | I tend to disagree :D | 13:08 |
greyback | "active" is simple, I don't see it being used | 13:09 |
dandrader | https://en.wikipedia.org/wiki/Active_window | 13:10 |
greyback | it's not a property on a window though, it is an application | 13:10 |
dandrader | still... | 13:11 |
dandrader | hmmm, maybe active would do... | 13:13 |
mzanetti | +1 | 13:13 |
mzanetti | after all this isn't public api... we can always rename it if we feel like | 13:13 |
dandrader | mzanetti, greyback, ok, so I will go for Application.active then | 13:13 |
mzanetti | wfm | 13:13 |
greyback | cool | 13:17 |
tsdgeos | cimi: can you confirm that make tryPreviewZoomableImage doesn't work? | 13:18 |
tsdgeos | when clicking on the image the overlay oepns but nothing happens | 13:18 |
tsdgeos | i.e. the image is lost | 13:18 |
tsdgeos | mzanetti: or you ↑ ? | 13:21 |
* mzanetti tries | 13:22 | |
mzanetti | yeah, the overlay seems black here | 13:23 |
mzanetti | tsdgeos, ^ | 13:23 |
tsdgeos | :/ | 13:23 |
tsdgeos | lol | 13:27 |
tsdgeos | it's because it's black on black | 13:27 |
* tsdgeos shakes fist | 13:30 | |
tsdgeos | and changes the asset | 13:30 |
MacSlow | tsdgeos, the qml-cache related fix you packported from Qt 5.5 ... do we carry that already as a distro-patch? | 13:35 |
tsdgeos | MacSlow: not sure which fix you mean | 13:36 |
MacSlow | tsdgeos, the one that forces me to wipe ~/.cache/QML before each ap-test-run for shell-rotation | 13:36 |
tsdgeos | MacSlow: that code is totally ours | 13:37 |
tsdgeos | it doesn't exist in Qt | 13:37 |
MacSlow | tsdgeos, ah ok | 13:37 |
MacSlow | tsdgeos, can I globally/system-wide disable it via an env-var? | 13:37 |
tsdgeos | no idea | 13:37 |
tsdgeos | ask ricm | 13:37 |
tsdgeos | he coded it | 13:38 |
MacSlow | tsdgeos, just pinged him... thx | 13:38 |
=== dandrader is now known as dandrader|afk | ||
seb128 | hey there | 14:00 |
seb128 | who is going to join the UOS session now for unity8 by default | 14:00 |
seb128 | https://blueprints.launchpad.net/ubuntu/+spec/convergence-1505-unity-8-as-default-desktop | 14:00 |
seb128 | Saviq, ^ in memory of the good old time? ;-) | 14:02 |
Saviq | mzanetti, greyback ↑ | 14:03 |
greyback | yep | 14:03 |
seb128 | thanks | 14:05 |
cimi | tsdgeos, was having lunch :) | 14:08 |
=== dandrader|afk is now known as dandrader | ||
tsdgeos | pstolowski: ping | 14:53 |
pstolowski | tsdgeos, pong | 14:54 |
tsdgeos | pstolowski: i'm confused about the fallback machinery | 14:54 |
tsdgeos | for images | 14:54 |
tsdgeos | for card mascot | 14:54 |
tsdgeos | reading https://code.launchpad.net/~michihenning/unity-scopes-api/fallback-image/+merge/245622 | 14:54 |
tsdgeos | it says | 14:54 |
tsdgeos | - The JSON for a CategoryRenderer now supports a "fallback" field in the "art" and "mascot" entries of the "components" dictionary. | 14:55 |
tsdgeos | but was mascot ever a dictionary? | 14:55 |
tsdgeos | i guess ti doesn't matter really now that i think :D | 14:57 |
tsdgeos | ignore me plz | 14:57 |
tsdgeos | pstolowski: about the "edit review" thing we were speaking with Patricia, is there a bug or something about it? | 15:08 |
pstolowski | tsdgeos, i think there is. i'm in uds session right now, will get back to you later | 15:09 |
tsdgeos | ok | 15:09 |
pstolowski | tsdgeos, https://bugs.launchpad.net/ubuntu/+source/unity-scope-click/+bug/1318144 | 15:49 |
ubot5 | Ubuntu bug 1318144 in unity-scope-click (Ubuntu) "Can't edit reviews" [Medium,Confirmed] | 15:49 |
tsdgeos | pstolowski: so i guess we need a unity8 entry in there, since there's changes needed from our side too, right? | 15:51 |
pstolowski | tsdgeos, sure | 15:57 |
tsdgeos | k added | 15:57 |
tsdgeos | pstolowski: so will you do the scopes side and then tell me the new widget formats i have to implement? | 15:58 |
pstolowski | tsdgeos, do you need the attribute names etc? | 15:59 |
tsdgeos | preview widget name and attrbute names yeah | 16:00 |
pstolowski | tsdgeos, ok, yeah, i'm currently working on the scope, will let you know all that tomorrow morning | 16:00 |
tsdgeos | also this is high/medium | 16:00 |
tsdgeos | but then Patricia made it seem as it was quite more important | 16:00 |
pstolowski | tsdgeos, yes. this bug was opened one year. i got it recently and was told it needs fixing now. so, kinda important | 16:01 |
pstolowski | * one year ago | 16:02 |
tsdgeos | k | 16:03 |
pstolowski | tsdgeos, unfortunately, i cannot set priorities on most of the bugs for some reason (probably these opened in Ubuntu project) | 16:03 |
tsdgeos | ah | 16:03 |
tsdgeos | we tricked the ubuntu bug control to give us power over changing stuff | 16:04 |
tsdgeos | want me to change it? | 16:04 |
pstolowski | tsdgeos, yeah, feel free to | 16:04 |
tsdgeos | High? | 16:04 |
tsdgeos | Critical? | 16:04 |
pstolowski | high is ok | 16:04 |
tsdgeos | done | 16:04 |
pstolowski | let's not put even more pressure on ourselves ;') | 16:04 |
tsdgeos | he he | 16:04 |
* tsdgeos eods | 16:04 | |
tsdgeos | tty tomorrow | 16:05 |
=== dandrader is now known as dandrader|lunch | ||
=== alan_g is now known as alan_g|EOD | ||
=== dandrader|lunch is now known as dandrader | ||
=== ljp is now known as lpotter | ||
=== om26er_ is now known as om26er | ||
Saviq | mzanetti, thanks, hope I didn't talk too much ;) | 18:55 |
mzanetti | Saviq, no worries :D | 18:56 |
mzanetti | all fine | 18:56 |
mzanetti | anpok, hey, did you have a chance to update the u-s-c branch by now? | 18:57 |
mzanetti | olli wants to install the silo | 18:57 |
davmor2 | Saviq: were there no magic pendant talks? | 18:57 |
Saviq | davmor2, I got invited, couldn't refuse ;) | 18:58 |
davmor2 | Saviq: hahaha | 19:00 |
=== boiko_ is now known as boiko | ||
balloons | for those who asked today; I pushed a new build of terminal and filemanager for x86 to the store. Ready for your unity8 desktops | 19:42 |
Saviq | \o/ | 19:44 |
gQuigs | for unity8 devel, so ./run --what would give the right input devices to give me a desktop/windowed view | 20:45 |
gQuigs | ah, apparently just fullscreening, nvm :P | 20:46 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!