=== JanC_ is now known as JanC | ||
Mirv | tsdgeos: so I was considering publishing final Qt 5.7.1 to zesty, but it turns out Unity 8 has some new test issues - can you look at how serious bug #1654819 failures seem? | 08:31 |
---|---|---|
ubot5 | bug 1654819 in unity8 (Ubuntu) "new failing qmluitests with Unity 8 on Qt 5.7.1" [Critical,New] https://launchpad.net/bugs/1654819 | 08:31 |
tsdgeos | sure i saw your email/bug | 08:31 |
tsdgeos | still didn't get to it | 08:31 |
tsdgeos | let me skip the queue | 08:31 |
tsdgeos | look relatively bad | 08:32 |
tsdgeos | is that because you packager a newer Qt? | 08:32 |
tsdgeos | last time we did this the tests passed, no? | 08:32 |
tsdgeos | was it pre 5.7.1? | 08:32 |
Mirv | tsdgeos: yeah you closed the previous bug, that was near-final snapshot 20161022 | 08:32 |
Mirv | tsdgeos: but it seems they changed something (fixed bugs, hopefully) before releasing the final | 08:33 |
tsdgeos | let me see what's missing | 08:33 |
Mirv | considering it's an already published stable series, changes like that in final stretch of getting .1 out are a bit unfortunate | 08:33 |
Mirv | all seem surface sort of issues, in test failures group of 3, 4 and 4 failing tests | 08:35 |
Mirv | not much in qtdeclarative http://code.qt.io/cgit/qt/qtdeclarative.git/log/?h=v5.7.1 | 08:35 |
Mirv | not much in qtbase either http://code.qt.io/cgit/qt/qtbase.git/log/?h=v5.7.1 | 08:36 |
tsdgeos | there was someone reporting some shortcut issues in Qt 5.7.1. vs 5.7.0 in the kde mailing lists over the weekend | 08:36 |
tsdgeos | didn't really pay much attention though | 08:36 |
tsdgeos | was busy playing kingdom rush frontiers ^_^ | 08:36 |
Mirv | there are essentially 5 (five) real commits together in qtbase and qtdeclarative 5.7.1 since 20161021 | 08:37 |
Mirv | ^_^ | 08:37 |
Mirv | maybe that fbo commit in qtbase or V4 change in qtdeclarative could be related | 08:38 |
Mirv | I guess those surface tests wouldn't use jpegs | 08:39 |
Mirv | and of course there is always a possibility of something else in zesty changed, although Qt doesn't use much besides C library | 08:40 |
tsdgeos | the ubuntu browser killed my session ^_^ | 08:42 |
tsdgeos | oSoMoN: ↑ | 08:42 |
oSoMoN | tsdgeos, how did it do that? do you have logs / a crash file? | 09:05 |
tsdgeos | oSoMoN: so konsole has this "right click -> open url" when hovering a link | 09:05 |
tsdgeos | i clicked on it and somehow the ubuntu browser decided it was better than the firefox i had configured previously to open the link | 09:06 |
tsdgeos | and that's all i have | 09:06 |
tsdgeos | everything locked up and i needed to reboot the laptop | 09:06 |
oSoMoN | wow | 09:09 |
oSoMoN | tsdgeos, but your DE is configured to use firefox as default? or was the default somehow overridden? | 09:09 |
tsdgeos | if i click on a quassel link it'll open firefox | 09:10 |
tsdgeos | no idea why konsole decided to use the ubuntu browser | 09:10 |
oSoMoN | tsdgeos, is it reliably reproducible? | 09:12 |
tsdgeos | the crash? | 09:12 |
tsdgeos | i am not sure i want to try to lock my system again :D | 09:12 |
tsdgeos | but ok | 09:12 |
tsdgeos | i'll do it for you :) | 09:12 |
oSoMoN | tsdgeos, thanks :) | 09:12 |
oSoMoN | tsdgeos, I meant the fact that konsole will open the link in the browser | 09:13 |
tsdgeos | oSoMoN: both :D | 09:15 |
tsdgeos | oSoMoN: you've a telegram picture of the crash | 09:15 |
Mirv | tsdgeos: did you miss my ponderings about the only 5 commits in qtbase/qtdeclarative when it crashed? https://irclogs.ubuntu.com/2017/01/09/%23ubuntu-unity.html | 09:23 |
tsdgeos | i guess i did | 09:23 |
tsdgeos | was testing webbrowser-app crashyness | 09:24 |
tsdgeos | Mirv: there's also a weird error message about itemgrabber not being registered | 09:24 |
tsdgeos | let me see if i get that on current zesty | 09:24 |
tsdgeos | nope | 09:25 |
tsdgeos | file:///home/tsdgeos/unity8/qml/Stage/ApplicationWindow.qml:143:5: QML Image: Protocol "itemgrabber" is unknown | 09:25 |
tsdgeos | this is bad | 09:25 |
tsdgeos | also | 09:25 |
tsdgeos | Trying to pass extra request flags to provider but it is not a QQuickImageProviderWithOptions | 09:26 |
tsdgeos | is weird | 09:26 |
tsdgeos | Mirv: do we still carry any QQuickImageProviderWithOptions patch? | 09:26 |
Mirv | UITK has that providerwithoptions warning too in 2/3 of its tests | 09:27 |
Mirv | tsdgeos: here is the list https://anonscm.debian.org/cgit/pkg-kde/qt/qtdeclarative.git/tree/debian/patches?h=ubuntu%2b1 | 09:27 |
Mirv | I don't see any in qtdeclarative 5.6.1 either https://anonscm.debian.org/cgit/pkg-kde/qt/qtdeclarative.git/tree/debian/patches?h=ubuntu/5.6.1-7ubuntu2 | 09:28 |
tsdgeos | ok | 09:29 |
Mirv | the async stuff was merged in 5.6 | 09:29 |
tsdgeos | oSoMoN: https://bugs.launchpad.net/ubuntu/+source/webbrowser-app/+bug/1654974 | 09:30 |
ubot5 | Ubuntu bug 1654974 in webbrowser-app (Ubuntu) "webbrowser-app locks up my session (zesty)" [Undecided,New] | 09:30 |
oSoMoN | tsdgeos, thanks | 09:32 |
oSoMoN | tsdgeos, I will look into it today | 09:33 |
tsdgeos | oSoMoN: it *may* be related to using unity8-greeter as greeter | 09:33 |
tsdgeos | since i think one of the times it crashed | 09:33 |
tsdgeos | on reboot | 09:33 |
tsdgeos | it complained about unity-system-compositor having crashed | 09:34 |
tsdgeos | but you can't know if it did really cause the crash or it crashed because i hard rebooted the laptop and then things get antry | 09:34 |
tsdgeos | angry | 09:34 |
=== vigo_ is now known as vigo | ||
tsdgeos | Mirv: do you know the hash of the qtdeclarative of the last try? | 09:40 |
* tsdgeos hates vim decided to change behavoiur when you use the mouse on it | 09:43 | |
Saviq | tsdgeos, thought it's some plugin, but /me hates, too | 09:45 |
tsdgeos | i did changed it in my main account | 09:46 |
tsdgeos | but chroot still borked | 09:46 |
tsdgeos | set mouse-=a | 09:46 |
tsdgeos | fixed it afair | 09:46 |
Mirv | tsdgeos: as Debian named its orig tarball as 5.7.1~20161021, and looking at http://code.qt.io/cgit/qt/qtdeclarative.git/log/?h=v5.7.1 , I'd say it was very likely that last 2016-10-05 merge commit ie fff4477661ae240c43088fa6d9069ccf969dbee8 | 09:52 |
* tsdgeos kicks himself for not having made | 09:56 | |
tsdgeos | "Trying to pass extra request flags to provider but it is not a QQuickImageProviderWithOptions" be better | 09:56 |
tsdgeos | it's like "something is wrong" | 09:56 |
tsdgeos | but i'm not going to tell you where | 09:56 |
tsdgeos | thought that is a red herring! | 10:02 |
tsdgeos | Mirv: ok i can reproduce the problem in a smaller scale now, so we're making progress | 10:08 |
Mirv | tsdgeos: diff between 20161021 orig tarball and final 5.7.1: http://paste.ubuntu.com/23769521/ - additionally, added the SignalSpy fix and Scale-images-correctly patches | 10:08 |
Mirv | tsdgeos: ok | 10:09 |
tsdgeos | i am thinking Scale-images-correctly may be bad | 10:09 |
tsdgeos | yo i am a bad devel! | 10:10 |
tsdgeos | every patch i bring breaks stuff :D | 10:10 |
Mirv | someone requested it :) | 10:10 |
tsdgeos | wonder who that may be | 10:11 |
tsdgeos | ok at least it seems only dev upstream contains that | 10:11 |
tsdgeos | so i didn't break much outside dev and our wannabe packages | 10:12 |
tsdgeos | let me confirm it's it | 10:12 |
Mirv | zsombi: you may want to pay attention here if you started looking at that QQuickImageProviderWithOptions UITK warning | 10:12 |
tsdgeos | the warning itself is "ok" | 10:12 |
tsdgeos | it's not what causes the regression | 10:12 |
tsdgeos | but would be nice to get fixed too if possible of course | 10:13 |
tsdgeos | and yes i need to improve the warning | 10:13 |
* zsombi reading the logs | 10:14 | |
zsombi | tsdgeos: Mirv: ok, so what is the conclusion? is that warning a valid one? is something screwed that we need to fix still? | 10:22 |
tsdgeos | zsombi: it is a valid warning (even though the message can defenitely be improved) | 10:23 |
tsdgeos | you can also ignore it, since it's a "this could be better" | 10:23 |
tsdgeos | not a "this is wrong" kind of warning | 10:23 |
zsombi | tsdgeos: any hunch what actually causes it? | 10:23 |
zsombi | yeah, the warning text is a bit weird :D | 10:24 |
tsdgeos | the new QQuickImageProviderWithOptions code i added | 10:24 |
tsdgeos | basically image providers can (in dev + our patches) know they are being asked to do aspectImageFit/Crop | 10:24 |
tsdgeos | and provide better images | 10:24 |
tsdgeos | than the ones they did before | 10:24 |
zsombi | I see | 10:24 |
zsombi | and how are we supposed to let it know? | 10:26 |
zsombi | tsdgeos: the doc https://doc-snapshots.qt.io/qt5-dev/qquickimageprovideroptions.html is pretty weak... | 10:27 |
tsdgeos | ey | 10:28 |
tsdgeos | it has a see also to itself | 10:28 |
tsdgeos | pretty awesome | 10:28 |
tsdgeos | zsombi: you don't let it know, it let's you know | 10:29 |
tsdgeos | an no | 10:29 |
tsdgeos | ah no | 10:29 |
tsdgeos | it has a see also to QQuickImageProviderWithOptions | 10:29 |
tsdgeos | that i guess give it's hidden in a _p.h | 10:29 |
zsombi | haha | 10:29 |
tsdgeos | doesn't get the docu generated | 10:29 |
tsdgeos | Mirv: for now i'd say revert that patch, i need to figure out how to unbreak Item::grabToImage | 10:33 |
tsdgeos | Mirv: you added it to debian too? or only ubuntu? | 10:34 |
Mirv | tsdgeos: only Ubuntu, I tend to avoid automatically adding our patches to Debian unless they are from upstream stable branch or such | 10:39 |
Mirv | tsdgeos: ok then | 10:39 |
tsdgeos | ok, good | 10:39 |
tsdgeos | i mean i still need to come up with a fix since upstream is broken | 10:39 |
tsdgeos | but that should unblock you | 10:39 |
Mirv | fabulous. I wonder if I dare to then publish the silo without pre-running autopkgtests again, since I get flamed for occupying autopkgtest infra for 3+ days multiple times. | 10:42 |
Mirv | I think I will, but will ask to test once more from the PPA once it's built | 10:43 |
Mirv | also since UITK shows the warning in unit tests I can just see about its rebuild | 10:44 |
tsdgeos | that warning will also be gone if you revert that patch | 10:44 |
Mirv | doh, forgot to revert the symbol changes | 11:59 |
Mirv | rebuild | 11:59 |
Mirv | tsdgeos: ok 1985 now updated with qtdeclarative. | 13:08 |
tsdgeos | Mirv: tests seem to be back to good now | 13:44 |
Mirv | great | 14:07 |
rvr | Is there anyway to dismiss the gesture wizard in kvm? | 15:50 |
Saviq | rvr, for whatever reason KVM does not allow you to "trap" your pointer in the window, so you can't push against the edge... using USB forwarding was the only way I found | 15:54 |
Saviq | tried with the different graphics/input stacks and it was either that or no input at all... | 15:55 |
rvr | Saviq: USB forwarding? How does it work? | 15:55 |
rvr | Saviq: Oh, you mean, sharing the mouse device directly with the VM? | 15:55 |
Saviq | rvr, yes | 15:55 |
rvr | Interesting! | 15:55 |
Saviq | rvr, that's really the only way to fully use it anyway, as how else will you trigger launcher or right edge | 15:56 |
rvr | Saviq: Thanks for the tip | 15:56 |
roasted | Hi friends. Is there a "feature request form" of some sort in regard to Unity 8 features? | 17:24 |
dmj_s76 | hey Trevinho! | 17:33 |
greyback | roasted: really the best place is to log a bug https://bugs.launchpad.net/ubuntu/+source/unity8 | 17:40 |
roasted | greyback: I wondered that, but "bug" didn't feel right to me, so I felt compelled to dig around first before submitting. | 17:41 |
greyback | roasted: I agree it feels strange, but we do also keep track of features we add via bugs. We find it handiest to use a single database for bugs, features and wish-list things | 17:41 |
roasted | greyback: sounds good. thanks for the insight. :D | 17:42 |
greyback | np, thanks for dropping by :) | 17:42 |
roasted | greyback: I'm somewhat hesitant to submit this in the name of cluttering things up. Perhaps you know offhand. Do you know if Unity 8 has/is planned to get display scaling/desktop zoom support? | 17:44 |
roasted | I'm asking on behalf of students at my district. We have a few that are visually impaired and benefit massively from U7's zoom/scale abilities. | 17:44 |
greyback | roasted: accessibility features like scaling, zoom, high-contrast are indeed on our roadlist | 17:48 |
greyback | roasted: by adding a bug, it will help push those features up the priority queue :) | 17:48 |
roasted | oh? well hot dog I'll submit it then. | 17:48 |
roasted | Maybe a heart-warming success story of a few middle school kids who use them on a daily basis to better their education would help. :P | 17:49 |
dmj_s76 | roasted: Please do! This will have to be implemented in a new way. The U7 zoom functionality you're talking about is based on a Compiz plugin I believe. When I was teaching computers to visually impaired students, the compiz zoom was super useful to them. | 17:50 |
roasted | super useful doesn't even do it justice. It's *gold* to the few kids we have that depend on it. :D | 17:51 |
greyback | yep, and to many others. We for sure want to keep that feature working in unity8 | 17:54 |
roasted | submitted. thanks you awesome folks. <3 | 17:59 |
greyback | roasted: thank you for the feedback, it is always appreciated | 18:00 |
dmj_s76 | roasted: bug link? I may want to follow this | 18:00 |
roasted | https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1655099 | 18:03 |
ubot5 | Ubuntu bug 1655099 in unity8 (Ubuntu) "Feature Request - Scale/Zoom Features for Visually Impaired" [Undecided,New] | 18:03 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!