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

List:       kde-panel-devel
Subject:    Minutes Monday Plasma Hangout
From:       Sebastian =?UTF-8?B?S8O8Z2xlcg==?= <sebas () kde ! org>
Date:       2017-10-30 12:13:30
Message-ID: 26984801.MNLo20ncyC () miro
[Download RAW message or body]

Meeting notes, fun as always. Enjoy the read!
-- 
sebas

http://www.kde.org | http://vizZzion.org

Plasma Team meeting, 4-9-2017

Present: fvogt, d_ed, romangg, Sho_, einar77_work, bshah, notmart, kbroulik, sebas

fvogt:
- I'm not quite sure what to do with ivan's request on the simplefavoritesmodel
- Neither with https://phabricator.kde.org/D8455
Diff 8455 "Fix DBus startup notification" [Closed] https://phabricator.kde.org/D8455 \
                (Is the X-DBus-ServiceName even implemented by other DEs?)
- Any news on Qt 5.10 integration on the CI?
- https://bugs.kde.org/show_bug.cgi?id=382291#c22 seems to be broken still

d_ed:
Fixed wayland multiple screens where a screen has a y offset
Fixed glScissors calls on scaling (minimise animation now works nicely)
and added the wayland highDPI support for window decs (not merged that last one)

romangg:
reviewed patch from Xuetian Weng for refactoring of input (mouse) kcm. D8168 was the \
original one. Diff 8168 "kcm_input: Refactor the code and add support for libinput \
                under X11" [Needs Review] https://phabricator.kde.org/D8168
-> now split up into several ones:
The first one D8456 has already landed. the next one is D8460
Diff 8456 "Port kcm input to UI File." [Closed] https://phabricator.kde.org/D8456
Diff 8460 "Refactor kcm_input to enable having multiple backends." [Needs Review] \
https://phabricator.kde.org/D8460 The goal is to have the mouse kcm work like the \
touchpad kcm, such that we can have fully functional libinput backends for X and \
Wayland. Dived into reviewing the screen recording patches from Oleg.
some dev (Plagman) asked for a solution for his kvm issues this weekend. it's a very \
specific setup. but we need something like virtual screens in Wayland since xrandr \
not available anymore. created a task for now: T7331 Task 7331 "[kwin] Virtual \
screens" [Open,Low] {Plasma on Wayland} https://phabricator.kde.org/T7331 still need \
a review for my lib for night color: D5931 Martin reviewed the patch to KWin Diff \
5931 "ColorCorrect Library - for configuring KWin's native color correction (in \
particular Night Color)" [Needs Review] https://phabricator.kde.org/D5931


Sho:
* Fixed Simple Menu crashing when opening context menu for runner result items
** Planning 1.0.4 release this week
* Fixed Activity Pager not updating labels when activities are renamed, still in \
                review
* Working on libtm URL chaos fix
* Set up official KDE community on Matrix and the sysadmin stuff around that
* Contributed QtWS postmortem on Phab (https://phabricator.kde.org/T7315
Task 7315 "Qt World Summit: Postmortem" [Open,Normal] { KDE Promo} \
                https://phabricator.kde.org/T7315
* Off tomorrow, national holiday (but will actually spend it drafting the 2018 KDE \
e.V. budget)

einar77_work:
testing Qt 5.10 with Plasma
so far a couple of regressions, but need to be investigated further (file dialog is \
the major offender)

bhsah:
- Audio fixes
- Figured out how audio works without droid module if device is using ALSA
- One need to use haskell based tool xml2ucm to convert android's mixter_paths.xml to \
                ALSA ucm configuration or write it manually
- However sound quality by such ALSA driver is ... bad
- Investigated why pulseaudio crashes when loading droid-module
- Two different issues, droid-module is not compatible with new way of declaring \
                audio_policy
- And it needed a two liner patch to fix the incompatibility with newer pulseaudio \
                API
- Investigating the build system of pulseaudio-modules-droid to build it out-of-tree
- Overall have sound working locally, trying to fix the remaining packaging issues to \
                actually ship it on phone
- dealt with massive spam in #halium channel last weekend
- I fixed brightness-gets-low-at-start bug on phone

notmart:
 * Fix Dialog component in qqc2-desktop-style, now actually loads and works
 * plasma mobile: fixed a bug in opening plasmoids from the top panel
This week, mostly KCMS redesign things:
 * KCM redesigns: back and forward with Andy Betts using the cursor theme one as \
                prototype https://imgur.com/a/bL7w8
 * made a couple of basic components that can be reused across kcms
 * port of plasma mobile settings to kirigami
 * use a single shared qml engine for all kcms in the same process, solves a crash in \
plasma mobile settings: https://phabricator.kde.org/D8436 Diff 8436 "use a single QML \
                engine" [Needs Review] https://phabricator.kde.org/D8436
 * loader for qml kcms in systemsettings/kcmshell: a big hack for keyboard navigation \
(QQuickWidget has tab key navigation not working at all) \
https://phabricator.kde.org/D8522 Diff 8522 "keyboard navigation in and out QML kcms" \
[Needs Review] https://phabricator.kde.org/D8522

kbroulik:
* Not much from me, did a bit of code reviewing
* Will add a menu button to notification thumbnails to make it more obvious that \
there's a menu

sebas:
- plasma mobile roadmap blog to finish and publish
- trying to get some time in to look into why kscreen config isn't properly restored \
                on Wayland
- interview with purism's Todd and Swapnil scheduled somewhen next week




LOG:
[11:59:56] <sebas> dingdong meeting time!
[12:00:01] <sebas> Who who who?
[12:00:24] <d_ed> I am here
[12:00:26] <romangg> here
[12:00:27] <-- salamanderrake (~quassel@2605:a000:122a:32:48e9:f5b7:451b:b44e) has \
quit (Remote host closed the connection) [12:00:29] <Sho_> here
[12:00:35] <Sho_> from the dinner table
[12:00:42] <Sho_> (dst change ...)
[12:00:42] <sebas> make us jealous
[12:00:56] <einar77_work> kbroulik: ping (but after meeting)
[12:01:04] <kbroulik> here
[12:01:06] -*- fvogt is here, for approx. a few mins until lunch
[12:01:09] <bshah> hello
[12:01:17] <romangg> then fvogt first?
[12:01:32] <sebas> alright, fvogt to start to respect his food habits?
[12:01:58] <fvogt> Ok, I've only got a few items anyway
[12:02:35] <fvogt> - I'm not quite sure what to do with ivan's request on the \
simplefavoritesmodel [12:02:38] --> llucas \
(d4f4b91c@gateway/web/freenode/ip.212.244.185.28) has joined #plasma [12:02:40] \
<fvogt> - Neither with https://phabricator.kde.org/D8455 [12:02:41] <sKreamer> Diff \
8455 "Fix DBus startup notification" [Closed] https://phabricator.kde.org/D8455 \
[12:02:57] <fvogt> (Is the X-DBus-ServiceName even implemented by other DEs?) \
[12:02:58] --> xamindar (~quassel@172.56.15.88) has joined #plasma [12:03:10] <fvogt> \
- Any news on Qt 5.10 integration on the CI? [12:03:16] <d_ed> about the kconf \
script? Do .desktop files get copied about? [12:03:21] --> Son_Goku \
(~King_InuY@fedora/ngompa) has joined #plasma [12:03:40] <Sho_> Ben said building 5.9 \
and 5.10 was possible [12:03:48] <d_ed> Sho_: Qt?
[12:03:51] <fvogt> d_ed: Yes, into .local/share/applications
[12:03:55] <Sho_> yup
[12:04:07] <fvogt> d_ed: Also, the rename violates the startuptype=unique spec
[12:04:22] <fvogt> The dbus path has to follow the desktop filename
[12:04:52] <d_ed> so maybe I should look for another way to solve my bug?
[12:05:16] <fvogt> Yes, this is IMO more like a hack
[12:06:16] --> soee (~soee@bmx64.neoplus.adsl.tpnet.pl) has joined #plasma
[12:06:21] <fvogt> Ah, before I forget: \
https://bugs.kde.org/show_bug.cgi?id=382291#c22 seems to be broken still [12:06:21] \
<sKreamer> KDE bug 382291 in kde-gtk-config (general) "System settings GTK apps \
configurations resets." [normal,FIXED] https://bugs.kde.org/show_bug.cgi?id=382291 \
[12:07:14] <d_ed> the file name is \
/opt/kde5/share/applications/systemsettings.desktop  how is it a violation now and \
not before? [12:07:27] <fvogt> d_ed: It's always been
[12:07:31] <d_ed> ah, ok
[12:07:47] <d_ed> but it's slightly more correct now as it loses the 5.
[12:08:07] <d_ed> but I'll take a look
[12:08:10] <d_ed> worst case..kconf_update
[12:08:23] <-- NeKit (~nekit@85.143.20.2) has quit (Ping timeout: 255 seconds)
[12:08:35] <fvogt> If that even works, .desktop files might be anywhere
[12:08:53] <fvogt> (FTR: \
https://standards.freedesktop.org/desktop-entry-spec/latest/ar01s07.html) [12:09:12] \
<notmart> meeting? [12:09:22] <sebas> going on, notmart
[12:09:38] <notmart> crap, how i missed it
[12:09:38] -*- fvogt grabs lunch, will read backlog later
[12:09:45] --> fbeutel (~Thunderbi@2a02:8071:b95:c800:224:d7ff:fec6:470c) has joined \
#plasma [12:09:45] <notmart> wein queue :)
[12:09:49] <sebas> notmart: you haven't, you're right in it
[12:09:58] <sebas>  d_ed, romangg, Sho_, einar77_work, bshah, notmart, sebas are next \
up [12:10:42] <d_ed> in terms of Plasma stuff:
[12:11:05] <d_ed> Fixed wayland multiple screens where a screen has a y offset
[12:11:25] <sebas> (thanks muchos for that, got a pointer?)
[12:11:42] <sebas> Is that apol's bug?
[12:11:49] <d_ed> Fixed glScissors calls on scaling (minimise animation now works \
nicely) [12:12:01] <d_ed> it's the one he reported on kwin
[12:12:13] <Sho_> oh I think I have seen that anim bug
[12:12:16] <d_ed> it's not the config saving one he reported on kscreen
[12:12:28] <d_ed> and added the wayland highDPI support for window decs
[12:12:42] <d_ed> (not merged that last one)
[12:12:47] <Sho_> sweet
[12:13:18] <sebas> d_ed: ah ok
[12:13:19] <d_ed> sebas: 5d7e22e5dcf8981f9ba65088b81c1c6d456efef3 and \
019d4bf4257da2c5a601e6ec03c0134a606d5d74 and 9cafbb117984f746d4b71213effc3d27e1435f36 \
[12:13:27] <sebas> thanks! [12:14:20] <d_ed> That's mostly it for me on Plasma, but \
I'm doing some speedup work in Solid [12:14:29] <d_ed> next
[12:14:34] <sebas> romangg!
[12:14:46] <romangg> reviewed patch from Xuetian Weng for refactoring of input \
(mouse) kcm. D8168 was the original one. [12:14:47] <sKreamer> Diff 8168 "kcm_input: \
Refactor the code and add support for libinput under X11" [Needs Review] \
https://phabricator.kde.org/D8168 [12:14:56] <romangg> -> now split up into several \
ones: [12:15:04] --> james000- (~james000_@static.154.7.9.176.clients.your-server.de) \
has joined #plasma [12:15:05] <romangg> The first one D8456 has already landed. the \
next one is D8460 [12:15:05] <sKreamer> Diff 8456 "Port kcm input to UI File." \
[Closed] https://phabricator.kde.org/D8456 [12:15:05] <sKreamer> Diff 8460 "Refactor \
kcm_input to enable having multiple backends." [Needs Review] \
https://phabricator.kde.org/D8460 [12:15:14] <romangg> The goal is to have the mouse \
kcm work like the touchpad kcm, such that we can have fully functional libinput \
backends for X and Wayland. [12:15:40] <romangg> Dived into reviewing the screen \
recording patches from Oleg. [12:16:00] <romangg> I asked him to reevaluate if \
PipeWire support is a good idea. [12:16:15] <Sho_> what's his last name?
[12:16:19] <sebas> do you think pipewire is good?
[12:16:26] <Sho_> i havr patches from an Oleg for Konversation
[12:16:35] <romangg> @Sho_: Chernovskiy
[12:16:43] <Sho_> neat
[12:16:46] <Sho_> same person
[12:16:52] <sebas> assimilate!
[12:16:53] <d_ed> romangg: IMHO it is, but if we want glCopyPixels out of the \
compositor you need something the patch he did, and then a helper app to speak \
pipewire which still means merging this thing [12:17:14] <romangg> sebas: I would \
like to reduce the work load on screen recording app devs. So if we use the same \
system as GNOME, that's good. [12:17:35] <sebas> my impression as well, so I'd agree
[12:17:36] <romangg> d_ed: very right. in fact we came to the same conclusion in our \
discussion in T5653 [12:17:37] <sKreamer> Task 5653 "[kwin] Screen recording in \
Wayland session" [Open,High] {KWin,Plasma on Wayland} \
https://phabricator.kde.org/T5653 [12:17:52] <d_ed> romangg: <3
[12:18:10] <notmart> could that screen recording be used for remote access as well?
[12:18:19] <sebas> Coolio, on to The Sho?!
[12:18:21] <d_ed> notmart: yes
[12:18:24] <notmart> cool
[12:18:29] <sebas> Sho must go on
[12:18:35] <Sho_> * Fixed Simple Menu crashing when opening context menu for runner \
result items [12:18:36] <Sho_> ** Planning 1.0.4 release this week
[12:18:36] <Sho_> * Fixed Activity Pager not updating labels when activities are \
renamed, still in review [12:18:36] <Sho_> * Working on libtm URL chaos fix
[12:18:36] <Sho_> * Set up official KDE community on Matrix and the sysadmin stuff \
around that [12:18:37] <Sho_> * Contributed QtWS postmortem on Phab \
(https://phabricator.kde.org/T7315 [12:18:37] <sKreamer> Task 7315 "Qt World Summit: \
Postmortem" [Open,Normal] { KDE Promo} https://phabricator.kde.org/T7315 [12:18:38] \
<Sho_> * Off tomorrow, national holiday (but will actually spend it drafting the 2018 \
KDE e.V. budget) [12:18:39] <Sho_> * Jealousy for sebas: \
https://i.imgur.com/AZdb3tz.jpg [12:18:41] <romangg> I just would like to have the \
support for PipeWire (if reasonable) land at the same time as the GBM scanout. \
otherwise some people could misinterpret this in a way that we have a custom solution \
only. [12:18:48] <llucas> how is gnome solution performance wise?
[12:19:01] <romangg> hey, I had other stuff still!
[12:19:06] <kbroulik> Sho_: what's left with launcher url chaos? only thing i saw was \
the launch in place with activities or so [12:19:12] <sebas> Sho_: was just thinking \
that that is a proper monsterlunch ... but then it's dinner [12:19:14] <kbroulik> \
(plus some performance optimkzations on the fixes I did;) [12:19:15] <Sho_> next()
[12:19:22] <romangg> me again! :P
[12:19:23] <sebas> previous() then
[12:19:28] <romangg> ^^
[12:19:31] <Sho_> kbroulik: performance, and the spaces thing
[12:19:34] <kbroulik> oki
[12:19:42] <Sho_> :)
[12:20:31] <romangg> some dev (Plagman) asked for a solution for his kvm issues this \
weekend. it's a very specific setup. but we need something like virtual screens in \
Wayland since xrandr not available anymore. created a task for now: T7331 [12:20:32] \
<sKreamer> Task 7331 "[kwin] Virtual screens" [Open,Low] {Plasma on Wayland} \
https://phabricator.kde.org/T7331 [12:20:57] <romangg> still need a review for my lib \
for night color: D5931 Martin reviewed the patch to KWin [12:20:57] <sKreamer> Diff \
5931 "ColorCorrect Library - for configuring KWin's native color correction (in \
particular Night Color)" [Needs Review] https://phabricator.kde.org/D5931 [12:21:02] \
<romangg> -> d_ed :P [12:21:03] <romangg> done
[12:21:25] <romangg> -> next()
[12:22:11] <sebas> einar77_work is next
[12:22:29] <einar77_work> Nothing much to report
[12:22:35] <einar77_work> testing Qt 5.10 with Plasma
[12:22:54] <einar77_work> so far a couple of regressions, but need to be investigated \
further (file dialog is the major offender) [12:23:09] <einar77_work> but, I need to \
update to beta 2 and retest [12:23:11] <einar77_work> done();
[12:23:25] <sebas> thanks!
[12:23:28] <sebas> bshah!
[12:23:33] <bshah> okay
[12:23:55] <bshah> - Audio fixes
[12:23:56] <bshah> - Figured out how audio works without droid module if device is \
using ALSA [12:23:58] <bshah> - One need to use haskell based tool xml2ucm to convert \
android's mixter_paths.xml to ALSA ucm configuration or write it manually [12:24:00] \
<bshah> - However sound quality by such ALSA driver is ... bad [12:24:02] <bshah> - \
Investigated why pulseaudio crashes when loading droid-module [12:24:04] <bshah> - \
Two different issues, droid-module is not compatible with new way of declaring \
audio_policy [12:24:06] <bshah> - And it needed a two liner patch to fix the \
incompatibility with newer pulseaudio API [12:24:08] <bshah> - Investigating the \
build system of pulseaudio-modules-droid to build it out-of-tree [12:24:10] <bshah> - \
Overall have sound working locally, trying to fix the remaining packaging issues to \
actually ship it on phone [12:24:36] <bshah> And also... on other random note
[12:24:49] <bshah> dealt with massive spam in #halium channel last weekend
[12:25:04] <sebas> audio fixes are for halium / nexus5x?
[12:25:15] <bshah> sebas: yes
[12:25:17] <sebas> the one that accused us of being gay?
[12:25:24] <sebas> or prided us, depending
[12:25:30] <bshah> sebas: yes
[12:25:45] <sebas> aye
[12:25:45] <bshah> same user also changed topic of this channel to "eat less xxxxx"
[12:25:47] <sebas> annoying
[12:25:51] <bshah> so now topic lock here
[12:25:56] <sebas> sad
[12:26:05] <notmart> "that's why we can't have nice things"
[12:26:16] <sebas> yes, that, notmart
[12:26:54] <bshah> ... is a known spammer and person-with-too-much-time-in-hand \
according to freenode ops [12:27:09] <bshah> anyway.. next
[12:27:13] <sebas> notmart:
[12:27:16] <-- vpilo (~quassel@dhcp-089-098-130-086.chello.nl) has quit (Read error: \
Connection reset by peer) [12:27:24] <notmart> ok
[12:27:27] <notmart>  * Fix Dialog component in qqc2-desktop-style, now actually \
loads and works [12:27:28] <notmart>  * plasma mobile: fixed a bug in opening \
plasmoids from the top panel [12:27:29] <notmart> This week, mostly KCMS redesign \
things: [12:27:31] <notmart>  * KCM redesigns: back and forward with Andy Betts using \
the cursor theme one as prototype https://imgur.com/a/bL7w8 [12:27:32] <notmart>  * \
made a couple of basic components that can be reused across kcms [12:27:34] <notmart> \
* port of plasma mobile settings to kirigami [12:27:35] <notmart>  * use a single \
shared qml engine for all kcms in the same process, solves a crash in plasma mobile \
settings: https://phabricator.kde.org/D8436 [12:27:36] <sKreamer> Diff 8436 "use a \
single QML engine" [Needs Review] https://phabricator.kde.org/D8436 [12:27:37] \
<notmart>  * loader for qml kcms in systemsettings/kcmshell: a big hack for keyboard \
navigation (QQuickWidget has tab key navigation not working at all) \
https://phabricator.kde.org/D8522 [12:27:37] <sKreamer> Diff 8522 "keyboard \
navigation in and out QML kcms" [Needs Review] https://phabricator.kde.org/D8522 \
[12:27:55] <notmart> so next week probaly more on the kcms stuff [12:28:14] \
<kbroulik> notmart: pretty! (re cursor kcm) [12:28:18] <notmart> (also, away on \
wednesday as is national holiday here (as opposed to tomorrow) [12:28:29] <kbroulik> \
we have two holidays this week \o/ [12:28:31] <romangg> I really like the end result \
of the mouse cursor theme kcm [12:28:41] <bshah> notmart: question on kirigami \
plasma-settings, can it be merged to master? to test on phone? [12:28:45] <kbroulik> \
and I'm glad we traded the weird white border polariod thing for a proper rounded \
shadow [12:28:51] <sebas> that's sweet
[12:28:53] <notmart> bshah: yes, i think so
[12:29:15] <notmart> it will crash on the secon module never the less, as it needs \
https://phabricator.kde.org/D8436 [12:29:15] <sKreamer> Diff 8436 "use a single QML \
engine" [Needs Review] https://phabricator.kde.org/D8436 [12:29:19] <sebas> the \
rounded corners of that shadow aren't entirely breezy though? [12:29:22] <notmart> \
but not worse than the current one [12:29:47] <notmart> sebas: the delegate is \
completely custom drawn, yes [12:30:00] <notmart> as there aren't similar thengs in \
breeze so far [12:30:39] <sebas> ok
[12:30:43] <sebas> my turn?
[12:30:54] <romangg> what's the plan regarding the margins? it still feels a little \
bit overloaded because there is not enough space between the elements. [12:31:03] \
<notmart> biggest problem for now is the background of the scrollview, which i would \
like to get a bit less hacky on the things i'm doing on upstream qqc2... [12:31:18] \
<notmart> margins, i don't know.. [12:31:39] <romangg> you said it's so it aligns \
with the buttons, which makes sense. [12:31:40] <notmart> ..to me it feels overloaded \
because there is too much of it... go figure :) [12:31:53] <llucas> notmart how it \
looks with breeze dark? [12:32:09] <romangg> but maybe we can solve this differently
[12:32:15] <notmart> llucas: didn't try i yet, it should be well... dark :p
[12:32:16] <kbroulik> can we complain about the design later and get the meeting \
going? thx [12:32:21] <romangg> or shall we wait for a redesign of whole syse?
[12:32:45] <romangg> ok. lets go on
[12:33:02] <sebas> yeah, let's finish the meeting and talk design then, it's just me \
left anyway ... my stuff is smallish, since I'm caught up with \
not-entirely-plasma-things: [12:33:11] <sebas> - plasma mobile roadmap blog to finish \
and publish [12:33:11] <sebas> - trying to get some time in to look into why kscreen \
config isn't properly restored on Wayland [12:33:11] <sebas> - interview with \
purism's Todd and Swapnil scheduled somewhen next week [12:33:30] <kbroulik> you \
forgot me again :D [12:33:43] <llucas> lol
[12:33:47] <sebas> pff, don't wear IRC camo then
[12:33:55] <sebas> your holy turn, kbroulik :)
[12:34:01] <kbroulik> * Not much from me, did a bit of code reviewing
[12:34:01] <kbroulik> * Will add a menu button to notification thumbnails to make it \
more obvious that there's a menu [12:34:10] <kbroulik> EOM (endofmeeting)
[12:34:40] <sebas> aye! Will send log and minutes.
[12:34:54] <bshah> ah.. code reviewing reminds me.. I fixed \
brightness-gets-low-at-start bug on phone [12:34:56] <bshah> :P
[12:35:02] <kbroulik> oh you did?
[12:35:04] <kbroulik> without side-effects? :P
[12:35:17] <bshah> side-effect is.... slider doesn't update on phone
[12:35:20] <bshah> :P
[12:35:25] <bshah> investigating again



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

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