[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-panel-devel
Subject:    Minutes Monday Plasma Hangout
From:       Eike Hein <hein () kde ! org>
Date:       2017-08-21 10:30:44
Message-ID: 94106714-887c-e0a9-6f5b-a1bb6e44c81c () kde ! org
[Download RAW message or body]

What have the Plasma heroes been up to the past week? Read and find out!


Cheers,
Eike

["minutes-week-24.txt" (text/plain)]


Plasma Team meeting, 21-08-2017

Present: mgraesslin, fvogt, notmart, bshah, Sho_, d_ed

mgraesslin:
* I'm working on what I call the "XFree KWin" project
** trying to get KWin/Wayland to work without XWayland
** thus moving lots of code into the x11 standalone platform
** once we get KWin work without XWayland I hope to also be able to handle XWayland \
                crashing
** related to all of that, refactored events.cpp to get more dedicated event handlers
** and started moving the compositors into plugins
* oh and on the general Wayland stuff: ShellClient's caption is now feature complete \
with X11 Client's caption

fvogt:
* Tried to fix annoying warning on Qt startup about QQC1 style "org.kde.desktop" not \
being found. First attempt broke weirdly due to bugs in Qt's style selection method, \
                hacked around it with \
                QT_QUICK_CONTROLS_1_STYLE="non/existent/dir/Desktop"
* Noticed with notmart that while qqc2-style was supposed to be released as part of \
                Plasma 5.10, it was not. Nobody noticed apparently
* (Not really plasma) Did an experiment to get QColumnView into the KIO file dialog, \
QColumnView is extremely buggy and needs a *lot* of work

notmart:
* finalized kirigami layers
* national holiday on 15th
* ported koko to it as a proof of concept
* kirigami icon sizes in desktop style
* wayland: foreign protocol,
** refactoring in a single class server side
** more testing on fail cases, when surfaces or pieces of wrapper get deleted
** https://phabricator.kde.org/D7369
** got completely frustrated by random crashes, weird memory corruptions and useless \
                backtraces
** proposed an api change on wayland-devel list

bshah:
* [Neon] WIP still to add the armhf in mobile CI
** Mostly tooling fixes where things just assumed that there is only one architecture
* [Misc] Got a chance to test touch screen stuff on my new laptop, have some ideas \
                (some inspired from the windows though :p), for example:
** Converge kickoff/kicker to appdash when keyboard is folded away?
** Provide ability to hide mouse? Windows provides this "Tablet mode" toggle which \
                does this
** Some more not-so-concrete ideas, can discuss them post meeting in breakout
** This also means I can test screen rotation stuff from sebas :)
*‎ mostly time was spent doing non plasma/mobile or halium work (neon)

Sho_:
* [Desktop] Fixed 5.10 regressions causing desktop widget position loss at shell \
restart due to size hints vs. position code ping-pong, ate lots of time last week for \
                ultimately somewhat unsatisfying timer delay solution
** I have a suicidal impulse to rewrite the desktop containment widgets layer now, \
                trying to resist
* [Kicker] Fixed search field losing focus after first typed character when searching \
                while hovering top-level category item
* [Other] Submitted an entry into the KDE community Goals process that impacts Plasma \
                heavily: \
                https://mail.kde.org/pipermail/kde-community/2017q3/003913.html
* [Other] Some QtWS'17 coordination (help leinir with showing off Plasma!): \
                https://mail.kde.org/pipermail/kde-community/2017q3/003916.html
* [Other] Started Qt Quick/Kirigami version of Konversation on the side, \
http://i.imgur.com/RPFDPoN.png

d_ed:
* I modded the baloo runner to work out of process.
** which is a good test of my DBus runners
* made kwin tell wayland windows which output they're on
** which Qt now also has a patch for
* oh, and I redid the kscreen scaling dialogs to have the per-screen support.


Full log:

‎[19:01] ‎<‎Sho_‎>‎ meeting time! mgraesslin, bshah, d_ed, sebas, notmart, \
Riddell, kbroulik, fvogt, others: who be here ‎[19:01] ‎* ‎mgraesslin‎‎ is \
here ‎[19:01] ‎* ‎fvogt‎‎ is here as well
‎[19:01] ‎* ‎notmart‎‎ here
‎[19:01] ‎* ‎Sho_‎‎ has a feeling bshah is also here
‎[19:01] ‎* ‎bshah‎‎ is here
‎[19:02] ‎<‎Sho_‎>‎ ah there we go :P
‎[19:02] ‎<‎Sho_‎>‎ seems a bit light today, where are the others? humm
‎[19:02] ‎<‎Sho_‎>‎ anyhow, mgraesslin starts, fvogt next
‎[19:02] ‎<‎mgraesslin‎>‎ hello, first time around in a long time :-)
‎[19:02] ‎<‎mgraesslin‎>‎ I'm working on what I call the "XFree KWin" \
project ‎[19:03] ‎<‎mgraesslin‎>‎ trying to get KWin/Wayland to work \
without XWayland ‎[19:03] ‎<‎mgraesslin‎>‎ thus moving lots of code into \
the x11 standalone platform ‎[19:03] ‎<‎Sho_‎>‎ neat :)
‎[19:03] ‎<‎heliocastro‎>‎ mgraesslin:  Just congrats for your wedding ( \
sorry for interrupt ) ‎[19:03] ‎<‎bshah‎>‎ yay lovely
‎[19:03] ‎<‎mgraesslin‎>‎ general idea: code which isn't compiled in, \
cannot create issues ‎[19:03] ‎<‎mgraesslin‎>‎ heliocastro: thanks
‎[19:03] ‎<‎mgraesslin‎>‎ once we get KWin work without XWayland I hope to \
also be able to handle XWayland crashing ‎[19:04] ‎<‎mgraesslin‎>‎ related \
to all of that, refactored events.cpp to get more dedicated event handlers ‎[19:04] \
‎<‎mgraesslin‎>‎ and started moving the compositors into plugins ‎[19:04] \
‎<‎Sho_‎>‎ seems the event handlers/filter concept is really working out in \
general ‎[19:04] ‎<‎mgraesslin‎>‎ mostly yes
‎[19:05] ‎<‎mgraesslin‎>‎ a general problem is that we have an ordering in \
the event handlers currently ‎[19:05] ‎<‎mgraesslin‎>‎ and that ordering is \
partially important ‎[19:05] ‎<‎mgraesslin‎>‎ so these things need to stay \
unfortunately ‎[19:05] --> Delta-One \
(~zero@2001:2002:51e2:7ba1:6391:acc7:b55f:c617) has joined this channel. ‎[19:06] \
‎<‎mgraesslin‎>‎ oh and on the general Wayland stuff: ShellClient's caption \
is now feature complete with X11 Client's caption ‎[19:06] ‎<‎mgraesslin‎>‎ \
next() ‎[19:06] ‎<‎bshah‎>‎ mgraesslin: a question
‎[19:06] ‎<‎mgraesslin‎>‎ yes?
‎[19:06] ‎<‎bshah‎>‎ mgraesslin: are the tasks which needs to be done for \
XFree Kwin, listed somewhere? ‎[19:07] ‎<‎mgraesslin‎>‎ no
‎[19:07] ‎<‎mgraesslin‎>‎ I don't know the tasks myself
‎[19:07] ‎<‎mgraesslin‎>‎ I mostly read code, find an area and fix it
‎[19:07] ‎<‎bshah‎>‎ okay, fair enough
‎[19:07] ‎<‎mgraesslin‎>‎ code coverage helps
‎[19:07] ‎<‎mgraesslin‎>‎ files which have bad coverage are good candidates \
;-) ‎[19:07] ‎<‎bshah‎>‎ hehe
‎[19:07] ‎<‎Sho_‎>‎ i'd say we have enough wayland tasks that /are/ listed \
though, so while martin does that stuff if others want to do wayland bits it's not \
hard to find todos ;) ‎[19:08] ‎<‎Sho_‎>‎ fvogt is next btw, after that is \
notmart ‎[19:08] ‎<‎bshah‎>‎ Sho_: well I am actually quite interested in \
Xfree kwin thing.. (plasma mobile perspective) ‎[19:08] ‎<‎Sho_‎>‎ bshah: \
oh right, makes sense ‎[19:08] ‎<‎bshah‎>‎ but yeah, fvogt, go ahead
‎[19:08] <-- cotko (~ahmed@89-212-138-82.static.t-2.net) has left this server (Ping \
timeout: 246 seconds). ‎[19:08] ‎<‎fvogt‎>‎ - Tried to fix annoying warning \
on Qt startup about QQC1 style "org.kde.desktop" not being found. First attempt broke \
weirdly due to bugs in Qt's style selection method, hacked around it with \
QT_QUICK_CONTROLS_1_STYLE="non/existent/dir/Desktop" ‎[19:08] ‎<‎fvogt‎>‎ - \
Noticed with notmart that while qqc2-style was supposed to be released as part of \
Plasma 5.10, it was not. Nobody noticed apparently ‎[19:09] --> cotko \
(~ahmed@89-212-138-82.static.t-2.net) has joined this channel. ‎[19:09] \
‎<‎Sho_‎>‎ /o\ ‎[19:09] ‎<‎Sho_‎>‎ how can we fix the "nobody \
noticed" problem next time? ‎[19:09] ‎<‎fvogt‎>‎ - (Not really plasma) Did \
an experiment to get QColumnView into the KIO file dialog, QColumnView is extremely \
buggy and needs a *lot* of work ‎[19:09] ‎<‎Sho_‎>‎ can't we generate "what \
needs to be released" from the dep data? ‎[19:10] ‎<‎fvogt‎>‎ It's not a \
dep of anything ‎[19:10] ‎<‎notmart‎>‎ i tought that whatever was in \
w0rkspace/ must be in next release ‎[19:10] ‎<‎fvogt‎>‎ Only \
plasma-workspace references it by setting the QQC2 style to org.kde.desktop \
‎[19:10] ‎<‎Sho_‎>‎ i thought maybe the dep data had a "this is plasma" \
item ‎[19:10] <-- starbuck-office (~netrunner@p4FD07690.dip0.t-ipconnect.de) has \
left this server (Quit: Konversation terminated!). ‎[19:11] ‎<‎notmart‎>‎ \
which i tought was the location in the repo, apparently not ‎[19:11] \
‎<‎Sho_‎>‎ but maybe makes sense to discuss this next week with Riddell \
there? ‎[19:11] ‎<‎notmart‎>‎ sure
‎[19:11] ‎<‎Sho_‎>‎ alright, let's try not to forget :)
‎[19:11] ‎<‎Sho_‎>‎ notmart is next
‎[19:11] ‎<‎fvogt‎>‎ notmart: BTW, I installed it and it works fine here \
:-) ‎[19:11] ‎<‎notmart‎>‎ yay :)
‎[19:12] ‎<‎sKreamer‎>‎ ::plasma5:: Plasma 5 :: Re: Disable activation of \
Application Starter when hitting S @ \
https://forum.kde.org/viewtopic.php?f=289&t=141505&p=380186#p380186 (by Theoretiker) \
‎[19:12] ‎<‎notmart‎>‎ * finalized kirigami layers ‎[19:12] \
‎<‎notmart‎>‎ * national holiday on 15th ‎[19:12] ‎<‎notmart‎>‎ * \
ported koko to it as a proof of concept ‎[19:12] ‎<‎notmart‎>‎ * kirigami \
icon sizes in desktop style ‎[19:12] ‎<‎notmart‎>‎ * wayland: foreign \
protocol, ‎[19:12] ‎<‎notmart‎>‎ ** refactoring in a single class server \
side ‎[19:12] ‎<‎notmart‎>‎ ** more testing on fail cases, when surfaces or \
pieces of wrapper get deleted ‎[19:12] ‎<‎notmart‎>‎ ** \
https://phabricator.kde.org/D7369 ‎[19:12] ‎<‎sKreamer‎>‎ Diff 7369 "[WIP] \
Wayland foreign protocol" [Needs Review] https://phabricator.kde.org/D7369 ‎[19:12] \
‎<‎notmart‎>‎ ** got completely frustrated by random crashes, weird memory \
corruptions and useless backtraces ‎[19:12] ‎<‎notmart‎>‎ ** proposed an \
api change on wayland-devel list ‎[19:13] ‎<‎notmart‎>‎ that's pretty much \
my part ‎[19:13] ‎<‎Sho_‎>‎ what's kigirami layers?
‎[19:13] ‎<‎notmart‎>‎ but a lot of the time was spent into staring at \
valgrind or gdb to no avail :p ‎[19:13] ‎* ‎Sho_‎‎ appreciates enduring the \
frustration gauntlet :) ‎[19:13] ‎<‎notmart‎>‎ Sho_: is a stack view that \
now wrap arounds to the columns view ‎[19:13] ‎<‎Sho_‎>‎ what's it for?
‎[19:14] ‎<‎notmart‎>‎ so if the application needs to have a layer that \
covers all the columns, temporarly disabling the horizontal navigation it can \
‎[19:14] ‎<‎Sho_‎>‎ ah ‎[19:14] ‎<‎notmart‎>‎ mostly for things \
that are a viewer of something, like koko or peruse ‎[19:14] ‎<‎Sho_‎>‎ \
cool ‎[19:14] ‎<‎notmart‎>‎ where when you are in the mode of viewing the \
thing you are completely outside of the navigation mode ‎[19:15] ‎<‎Sho_‎>‎ \
right ‎[19:15] ‎<‎Sho_‎>‎ "breakout mode" :P
‎[19:15] ‎<‎Sho_‎>‎ alrighty, bshah is up
‎[19:15] ‎<‎bshah‎>‎ okay
‎[19:15] ‎<‎bshah‎>‎ - [Neon] WIP still to add the armhf in mobile CI
‎[19:15] ‎<‎bshah‎>‎     - Mostly tooling fixes where things just assumed \
that there is only one architecture ‎[19:15] ‎<‎bshah‎>‎ - [Misc] Got a \
chance to test touch screen stuff on my new laptop, have some ideas (some inspired \
from the windows though :p), for example: ‎[19:16] ‎<‎bshah‎>‎     - \
Converge kickoff/kicker to appdash when keyboard is folded away? ‎[19:16] \
‎<‎bshah‎>‎     - Provide ability to hide mouse? Windows provides this \
"Tablet mode" toggle which does this ‎[19:16] ‎<‎bshah‎>‎     - Some more \
not-so-concrete ideas, can discuss them post meeting in breakout ‎[19:16] \
‎<‎bshah‎>‎     - This also means I can test screen rotation stuff from sebas \
:) ‎[19:16] ‎<‎bshah‎>‎ mostly time was spent doing non plasma/mobile or \
halium work (neon) ‎[19:16] ‎<‎bshah‎>‎ questions? or next
‎[19:17] ‎<‎Sho_‎>‎ bshah: i also had a long-standing goal to add soem \
system toggles to dashboard (wifi, bluetooth, etc.) in hybrid/touch mode ‎[19:17] \
‎<‎Sho_‎>‎ somewhat thwarted by the difficulty of actually making my \
touchscreen work on my laptop ‎[19:18] ‎<‎bshah‎>‎ for me it just worked :p
‎[19:18] ‎<‎Sho_‎>‎ alright, i'll post my notes:
‎[19:18] ‎<‎Sho_‎>‎ * [Desktop] Fixed 5.10 regressions causing desktop \
widget position loss at shell restart due to size hints vs. position code ping-pong, \
ate lots of time last week for ultimately somewhat unsatisfying timer delay solution \
‎[19:18] ‎<‎Sho_‎>‎ ** I have a suicidal impulse to rewrite the desktop \
containment widgets layer now, trying to resist ‎[19:18] ‎<‎Sho_‎>‎ * \
[Kicker] Fixed search field losing focus after first typed character when searching \
while hovering top-level category item ‎[19:18] ‎<‎Sho_‎>‎ * [Other] \
Submitted an entry into the KDE community Goals process that impacts Plasma heavily: \
https://mail.kde.org/pipermail/kde-community/2017q3/003913.html ‎[19:18] \
‎<‎Sho_‎>‎ * [Other] Some QtWS'17 coordination (help leinir with showing off \
Plasma!): https://mail.kde.org/pipermail/kde-community/2017q3/003916.html ‎[19:18] \
‎<‎Sho_‎>‎ * [Other] Started Qt Quick/Kirigami version of Konversation on the \
side, http://i.imgur.com/RPFDPoN.png ‎[19:18] ‎<‎mgraesslin‎>‎ bshah: \
anything new on the failing kwin on the one device? ‎[19:20] ‎<‎d_ed‎>‎ \
notmart: I didn't have a chance to look at your xdg-foreign issue ‎[19:20] \
‎<‎d_ed‎>‎ I hope to do so this afternoon ‎[19:20] ‎<‎notmart‎>‎ no \
problem :) ‎[19:20] ‎<‎Sho_‎>‎ d_ed, wanna take a turn too?
‎[19:20] ‎<‎d_ed‎>‎ sure
‎[19:20] ‎<‎Sho_‎>‎ go go go
‎[19:20] ‎<‎d_ed‎>‎ I modded the baloo runner to work out of process.
‎[19:20] ‎<‎d_ed‎>‎ which is a good test of my DBus runners
‎[19:21] ‎<‎mgraesslin‎>‎ an awesome idea btw.
‎[19:21] ‎<‎d_ed‎>‎ Sho_: if you can test the kicker stuff is the same, \
that'd be really good ‎[19:21] ‎<‎notmart‎>‎ yay for baloo runner out of \
preocess \o/ ‎[19:21] ‎<‎Sho_‎>‎ d_ed: is it merged or testing the phabs?
‎[19:21] ‎<‎d_ed‎>‎ Sho_: will merge in a sec
‎[19:21] ‎<‎Sho_‎>‎ k
‎[19:21] ‎<‎d_ed‎>‎ made kwin tell wayland windows which output they're on
‎[19:22] ‎<‎d_ed‎>‎ which Qt now also has a patch for
‎[19:22] ‎<‎d_ed‎>‎ (what's nice is the Qt guy added me as a reviewer for \
his patch, and also subscribed to the kwin patch..so hopefully that's a sign of some \
better collab ‎[19:22] ‎<‎d_ed‎>‎ (
‎[19:22] ‎<‎mgraesslin‎>‎ is Qt also able to adjust the scale \
automagically? ‎[19:22] ‎<‎d_ed‎>‎ it doesn't yet
‎[19:22] ‎<‎d_ed‎>‎ nor does it adjust it when the output itself changes
‎[19:23] ‎<‎d_ed‎>‎ but that's obviously the next step
‎[19:23] ‎<‎mgraesslin‎>‎ what about GTK? Does it work for them?
‎[19:24] ‎<‎d_ed‎>‎ I tried doing that in Qt, it worked, but the frame \
looks weird for a while even though I retriggerered an expose event ‎[19:24] \
‎<‎d_ed‎>‎ so I'll wait for the Qt guy to do it :D ‎[19:24] \
‎<‎bshah‎>‎ mgraesslin: they managed to run kwin in framebuffer mode on qemu \
with their setup (consolekit + musl libc on alpine) \
https://www.reddit.com/r/postmarketOS/comments/6uoxlr/postmarketosdemos_running_on_kwin \
but AFAIK not tested on real device yet ‎[19:24] ‎<‎d_ed‎>‎ mgraesslin: I \
don't actually know. I'll test and let you know ‎[19:25] ‎<‎mgraesslin‎>‎ \
d_ed: from my understandig GTK should just work ‎[19:25] ‎<‎mgraesslin‎>‎ \
bshah: though honestly fbdev is pretty much pointless up to the point that I started \
considering to delete it ‎[19:26] ‎<‎bshah‎>‎ mgraesslin: kinda right... \
yeah ‎[19:26] ‎<‎fvogt‎>‎ Some HW doesn't have proper DRM though :-/
‎[19:26] ‎<‎mgraesslin‎>‎ that's the point why it's still there
‎[19:27] ‎<‎d_ed‎>‎ oh, and I redid the kscreen scaling dialogs to have the \
per-screen support. ‎[19:27] ‎<‎Sho_‎>‎ nifty
‎[19:27] ‎<‎Sho_‎>‎ alrighty
‎[19:27] ‎<‎Sho_‎>‎ i'll close the meeting now and get to notes, feel free \
to discuss more :) ‎[19:28] ‎<‎mgraesslin‎>‎ my normal "KDE times" got a \
problem currently: https://www.bo.de/sites/default/files/styles/artikeldetail/public/field/image/img_6254.jpg?itok=XnJAiX2y
 ‎[19:28] ‎<‎bshah‎>‎ Sho_: thanks for hosting
‎[19:28] ‎<‎mgraesslin‎>‎ due to that I have to commute by car
‎[19:28] ‎<‎notmart‎>‎ what happened?
‎[19:28] ‎<‎mgraesslin‎>‎ which of course makes it impossible to hack on \
KDE ‎[19:28] ‎* ‎Sho_‎‎ commutes by bike lately
‎[19:28] ‎<‎d_ed‎>‎ that tape will definitely stop a train
‎[19:29] ‎<‎mgraesslin‎>‎ notmart: German railway tried to build a tunnel \
on sand five meters below that railway track ‎[19:29] ‎<‎bshah‎>‎ d_ed: \
100% sure? :P ‎[19:29] ‎<‎mgraesslin‎>‎ and well it collapsed
‎[19:29] ‎<‎notmart‎>‎ lol



[prev in list] [next in list] [prev in thread] [next in thread] 

Configure | About | News | Add a list | Sponsored by KoreLogic