Minutes Monday Plasma Meeting
Eike Hein
hein at kde.org
Mon Apr 3 12:25:25 UTC 2017
Plasma Meeting 2017-04-03
Present: romangg, notmart, d_ed, Riddell, Sho_
romangg:
last week I did D5270
needs review. but I would like to ask you if you have a solution to
raise/open the window directly instead of only invoking the context menu?
notmart:
* better management of minimum sizes for plasmawindowed (pushed)
* ctrl+q support for plasmawindowed
* better statusnotifieritem support for plasmawindowed:
https://phabricator.kde.org/D5242
Diff 5242 "fixes for plasmawindowed sni support" [Closed]
https://phabricator.kde.org/D5242
* retrocompatibility hackeroo for IconItem:
https://phabricator.kde.org/D5243
Diff 5243 "if the user did set an implicit size, keep it" [Needs
Review] https://phabricator.kde.org/D5243
* better layout and animate keyboard https://phabricator.kde.org/D5254
Diff 5254 "better layout and animate keyboard" [Needs Review]
https://phabricator.kde.org/D5254
* Kirigami2: introduce ApplicationItem when ApplicationWindow can't be
used (like gemini that neeeds to do its own window)
https://phabricator.kde.org/D5166
Diff 5166 "introduce ApplicationItem" [Needs Review]
https://phabricator.kde.org/D5166
* the 3 qt patches are still in limbo and i'm just getting frustrated by
that :/
d_ed:
I wrote support for bundling translations in plasmoids
the i18n part is up for review,
and I wrote support for QSvgIconEngine::isNull() which is needed for
properly fixing the QIcon::fromTheme fallback that Kai's been making
patches for
Riddell:
nothing from me I'm afraid, although amsterdam was nice
no releases this week or for the next three
Sho_:
* [TM] Fixed applet status not reverting to passive when the last
window in demands-attention state is closed (while keeping the status),
interfered with panel auto-hide
* [Other] Was mostly busy with Konversation and Yakuake releases this week
now i noticed the latte dock people did a major release ->
http://psifidotos.blogspot.kr/2017/04/latte-dock-v06-fresh-air.html
next there's two tricky things i wanted to inquire about because they
need some team think
the first is, ivan's been working on making kicker favorites
kastats-based, which makes them per-activity, but also global and shared
across all launcher instances
backporting TM group dialog from 5.9 to 5.8
Cheers,
Eike
-------------- next part --------------
[19:00] <Sho_> it's time for cool people to have a meeting
[19:00] <Sho_> who's cool?
[19:00] * notmart here
[19:00] <Sho_> bshah, notmart, d_ed, Riddell, romangg, ...
[19:00] <d_ed> I'm about, don't think I have much to say though
[19:01] <Riddell> hola
[19:01] <romangg> can I go second? need to leave early.
[19:01] <Riddell> likewise not much to add
[19:01] <Sho_> that's ok, i did app releases last week so my plasma list is a tad short too, but at least i have a question to bring
[19:01] <Sho_> sure romangg
[19:01] <Sho_> you can go first even
[19:01] <Sho_> let's start with romangg
[19:01] <romangg> Awesome! :D
[19:01] <Sho_> go go go
[19:01] <romangg> last week I did D5270
[19:01] <sKreamer> Diff 5270 "[System Tray] Fix regression on SNI fallback" [Needs Review] https://phabricator.kde.org/D5270
[19:01] --> [Enrico] (~chiccoroc at gentoo/contributor/Enrico) has joined this channel.
[19:02] <romangg> needs review. but I would like to ask you if you have a solution to raise/open the window directly instead of only invoking the context menu?
[19:02] <-- fbeutel (~Thunderbi at WL-POOL2-ONT-076.UNI-MUENSTER.DE) has left this server (Ping timeout: 256 seconds).
[19:02] --> llucas (~holoirc at host-89-230-174-106.dynamic.mm.pl) has joined this channel.
[19:02] <notmart> looking
[19:02] <romangg> the problem is that the needed DBus method Activate isn't implemented correctly by libappindicator
[19:03] <romangg> the workaround for now was/is now again to use the correctly implemented DBus method to raise the context menu instead.
[19:03] --> ltinkl (~ltinkl at 2a00:1028:83cc:7d32:79a4:34bf:148d:3a35) has joined this channel.
[19:03] <notmart> but the metoh is exported i guess...
[19:03] <romangg> but it would be better to raise the window directly on left click. XFCE can do it apparently.
[19:03] <notmart> is possible from dbus to see if the implementation is libappindicator?\
[19:04] <romangg> not sure. iirc yes. but in any case it's possible to see if the activate method worked or if it returned an error
[19:04] <romangg> it already uses that to raise the context menu instead.
[19:05] <romangg> ok, that's all for me. if somebody has a solution or more question regarding this issue, please reach out :)
[19:05] <Sho_> aye
[19:05] <Sho_> notmart is next
[19:05] <notmart> ok
[19:05] <notmart> * better management of minimum sizes for plasmawindowed (pushed)
[19:05] <notmart> * ctrl+q support for plasmawindowed
[19:06] <notmart> * better statusnotifieritem support for plasmawindowed: https://phabricator.kde.org/D5242
[19:06] <sKreamer> Diff 5242 "fixes for plasmawindowed sni support" [Closed] https://phabricator.kde.org/D5242
[19:06] <notmart> * retrocompatibility hackeroo for IconItem: https://phabricator.kde.org/D5243
[19:06] <sKreamer> Diff 5243 "if the user did set an implicit size, keep it" [Needs Review] https://phabricator.kde.org/D5243
[19:06] <notmart> * better layout and animate keyboard https://phabricator.kde.org/D5254
[19:06] <sKreamer> Diff 5254 "better layout and animate keyboard" [Needs Review] https://phabricator.kde.org/D5254
[19:06] <notmart> * Kirigami2: introduce ApplicationItem when ApplicationWindow can't be used (like gemini that neeeds to do its own window) https://phabricator.kde.org/D5166
[19:06] <sKreamer> Diff 5166 "introduce ApplicationItem" [Needs Review] https://phabricator.kde.org/D5166
[19:06] <notmart> * the 3 qt patches are still in limbo and i'm just getting frustrated by that :/
[19:06] <Sho_> sorry about the iconitem stuff :/
[19:07] <Sho_> reaching out to qt people didn't help?
[19:07] <Sho_> even on irc?
[19:07] <notmart> current status is frederik "will look at it"
[19:07] <notmart> will periodically pester people on #qt-labs about them this week
[19:08] <Sho_> alright
[19:08] <Sho_> d_ed would be next
[19:08] <notmart> i did have a couple of reviews on 2 of the patches, which i adressed in new revisions
[19:08] <notmart> so something moved
[19:08] <d_ed> I wrote support for bundling translations in plasmoids
[19:09] <d_ed> the i18n part is up for review,
[19:09] <Sho_> oh that's nifty
[19:09] <notmart> awesome
[19:10] <d_ed> and I wrote support for QSvgIconEngine::isNull() which is needed for properly fixing the QIcon::fromTheme fallback that Kai's been making patches for
[19:10] <Sho_> oh that's even niftier
[19:10] <d_ed> I put it in dev which is Qt5.11 now?
[19:10] <notmart> another patch in qt that will need some effort to get in?
[19:10] <d_ed> nah, got merged that day
[19:11] <notmart> oh, nice
[19:11] * notmart jealous
[19:11] * Sho_ too
[19:11] <Sho_> maybe d_ed needs to submit all our qt patches
[19:11] <Sho_> :)
[19:12] * bcooksley three.... getting patches into Qt is near impossible :(
[19:12] <d_ed> it depends on the patch, there's lots of places that people simply don't feel responsible for
[19:12] <d_ed> anyway, that's my Plasma week
[19:12] <Sho_> that's the advantage of kde's "common ownership" thing
[19:13] <notmart> eh, i have the misfortune of having a patch in qwidgets :)
[19:13] <Sho_> Riddell's next
[19:13] <Riddell> nothing from me I'm afraid, although amsterdam was nice
[19:13] <Riddell> no releases this week or for the next three
[19:13] <notmart> so everyone is like "oh, like we used to have qwidgets, lol is even a thing anymore?" :p
[19:13] <romangg> Riddell: any update on X Server update in Neon?
[19:13] <Sho_> boo hiss
[19:13] <Sho_> romangg you had to go
[19:13] <Sho_> !
[19:14] <Sho_> :)
[19:14] <romangg> yea, couldn't know you are all that quick :P
[19:14] <Riddell> ug, x server update is scary and not something I'm keen to do
[19:14] --> fekepp (~Thunderbi at 195.37.61.178) has joined this channel.
[19:14] <romangg> I mean what's the situation with Ubuntu 17.04 using X Server 1.19 or not. then it should be part of the next HWE stack.
[19:15] <Sho_> maybe this is a #kde-neon breakout?
[19:15] <Sho_> neon isn't technically plasma, sorry :)
[19:15] <romangg> ok, np.
[19:16] <Sho_> alright then it's my turn i guess
[19:16] <Sho_> my plasma list is short, but i have a discussion point and a few questions
[19:16] <Sho_> first the list
[19:16] <Sho_> * [TM] Fixed applet status not reverting to passive when the last window in demands-attention state is closed (while keeping the status), interfered with panel auto-hide
[19:16] <Sho_> * [Other] Was mostly busy with Konversation and Yakuake releases this week
[19:16] --> NicoRod (~nico at 200.127.54.64) has joined this channel.
[19:16] <Sho_> (i have more detailed notes on the app stuff if anyone cares)
[19:17] <Sho_> now i noticed the latte dock people did a major release -> http://psifidotos.blogspot.kr/2017/04/latte-dock-v06-fresh-air.html
[19:17] <Riddell> latte? now I want a coffee
[19:17] <Sho_> it looks pretty cool and they're getting pretty positive user feedback, and is a really nice showcase for plasma and frameworks (they use libtaskmanager extensively, etc.)
[19:17] <Sho_> might be worth trying out, and looking both at how we can help them and what we can steal :-)
[19:17] --> henriknj (~hnj at 5.103.133.127) has joined this channel.
[19:18] <Sho_> they even have a guy doing artwork, which begs the question why they're not part of the vdg
[19:18] <Sho_> :)
[19:18] --> JanKusanagi (~jank at 23.67.17.95.dynamic.jazztel.es) has joined this channel.
[19:18] <Sho_> let's see what we can maybe to do pull them in closer
[19:18] <romangg> the latte dock seems to be getting really good!
[19:18] <notmart> and does pretty crazy things with windows (afaik has even its own shell if they didn't change plans)
[19:18] <notmart> warned them in the past none of that would work on wayland, but whatever
[19:19] <Sho_> notmart: yeah, but there is value in doing crazy experimental work, especially if it's third-party and doesn't cause any technical debt to the core
[19:19] <-- fekepp (~Thunderbi at 195.37.61.178) has left this server (Ping timeout: 258 seconds).
[19:19] <Sho_> if they come up with a UX that's really winner, we can worry about how to do it on wayland later anyways
[19:19] <notmart> well
[19:19] <Riddell> romangg: it's being blocked in ubuntu 17.04 currently by qemu stuff https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/1671799
[19:19] <notmart> they were closer, and i really tried to help...
[19:19] <romangg> Riddell: nice find. Thank you!
[19:19] <-- pinheiro (~pinheiro at bl17-72-220.dsl.telepac.pt) has left this server (Remote host closed the connection).
[19:20] <Sho_> notmart: it's rather the cool dock everybody raves about is presented by a blog post saying "we extensively use plasma and frameworks for this" than $competition :)
[19:20] <notmart> then when one is told how to do a thing properly that would work and would look 100% the same, and goes "nah, i'll do my brittle hack instead and will go away"
[19:20] <notmart> then for me is a bit... over
[19:20] <-- jankusanagi_ (~jank at 151.25.23.95.dynamic.jazztel.es) has left this server (Ping timeout: 260 seconds).
[19:20] <llucas> Isn't latte a plasma fork, technically speaking?
[19:20] <Sho_> notmart: sometimes people do this because they want results now and still need to learn the lesson about why it's wrong
[19:21] <notmart> i guess so
[19:21] --> pinheiro (~pinheiro at bl17-72-220.dsl.telepac.pt) has joined this channel.
[19:21] <Sho_> but if they have the long-term motivation to keep working long enough to learn it and fix it later, ...
[19:21] <Sho_> like i said, as long as it's third-party it's not like there's any real harm
[19:21] <notmart> true
[19:21] <Sho_> and it creates buzz for us
[19:21] <-- friedrich (~friedrich at aextron.de) has left this server (Read error: Connection reset by peer).
[19:22] <Sho_> it's just natural, we're the old farts now who worry and know how to do things "properly" ;)
[19:22] <Sho_> anyways
[19:22] <notmart> llucas: not really, because isn't a standalone shell, it's just a dock
[19:22] <llucas> Ok
[19:22] <Sho_> llucas: it's no forking, they didn't copy any code, they use libs correctly at least
[19:23] <Sho_> i'm very very negative about calling anybody using our frameworks out for "forking" because plasma as a project has spent 10 years saying in public we've built these things to allow people to build crazy UIs
[19:23] <Sho_> so when that actually happens it should be welcomed
[19:23] <llucas> Yep understood, my mistake
[19:23] <Sho_> np
[19:23] <notmart> yep, indeed
[19:23] <Sho_> anyhooow
[19:23] <Sho_> next there's two tricky things i wanted to inquire about because they need some team think
[19:24] <Sho_> the first is, ivan's been working on making kicker favorites kastats-based, which makes them per-activity, but also global and shared across all launcher instances
[19:24] --> friedrich (~friedrich at aextron.de) has joined this channel.
[19:24] <Sho_> the work has been progressing nicely and is getting closer to being merged
[19:24] <Sho_> but i'm concerned about the migration process
[19:24] <-- Delta-One (~zero at 2001:2002:51e2:7ba1:6391:acc7:b55f:c617) has left this server (Ping timeout: 258 seconds).
[19:25] <Sho_> right now the way migration works is that the first kicker-based launcher that kicks it off gets its favorites migrated, and then all others use the migrated favorites
[19:25] <Sho_> which means there's a russian roulette race picking which faves to migrate and all the others get lost
[19:25] <Sho_> i don't think that's good enough, but it's hard to code up something else
[19:25] <Sho_> what do we do?
[19:25] <notmart> so cases with multiple kickers with different favorites will be extremely rare, but the handful of users that have it will have it broken
[19:26] <Sho_> not that rare actually
[19:26] <Sho_> quite a few users use a regular kicker + dashboard with diff faves
[19:26] <Sho_> (in fact, i expect angry bug reports about forcing them to be shared soon)
[19:26] <notmart> another thing, would be on plasma size completely forget about whatever is stored, but have another process (a kconfigupdate?)
[19:26] <notmart> that runs first and scans the whole configuration file
[19:27] <notmart> then "merges" the favorites in one big lists and dumps it into kastats?
[19:27] <Sho_> that sounds pretty good
[19:27] <Sho_> we'd still lose sorting, but at least get the superset
[19:27] <d_ed> why make it another process?
[19:27] <d_ed> plasma has it's own scripted updates
[19:27] <Sho_> it has? :D
[19:27] <notmart> ah, as a scripted updated
[19:27] <notmart> hm, yeah
[19:28] <Sho_> (fwiw, i'll dump this chat log into the phab ticket for the thing so ivan is in the loop later)
[19:28] <Sho_> i think we at the very least need the superset migration, data loss is imho not an option
[19:28] <d_ed> ShellCorona::processUpdateScripts()
[19:28] <notmart> a scripted updated may also be executed at the proper moment
[19:29] <notmart> tough, is there any desktop scripting way to populate that?
[19:29] <d_ed> to populate the activities DB?
[19:29] <notmart> (that -> favorites in kastats)
[19:29] <Sho_> i suddenly felt a dejavu
[19:29] <Sho_> i think i remember telling ivan we need a kastats scripting api to populate global favs in plasma scripting
[19:29] <Sho_> also for isv default favs ...
[19:30] <Sho_> because i don't know how he intends to do distro default favs
[19:30] <-- moofang (~lim at brk-24-218.tm.net.my) has left this server (Ping timeout: 246 seconds).
[19:30] * Sho_ is getting concerned we can really do this for 5.10 still, even as advanced as the work is, and sponsor pressure to get it in
[19:30] <-- friedrich (~friedrich at aextron.de) has left this server (Ping timeout: 264 seconds).
[19:31] <Sho_> and tbh, as i said as far as 1-2 years ago, i'm still not even sure forcing all launchers to have the same favs is good and serves all users correctly
[19:31] <notmart> are there places other people can work on?
[19:31] <Sho_> sure
[19:31] <Sho_> it's in a branch, so if you e.g. want to take a stab at the migration stuff, i'm sure ivan would appreciate it (and me)
[19:31] <Sho_> the phab ticket is also up to date with all our testing correspondence and debuging/testing instructions
[19:31] <Sho_> and even a testcase list
[19:32] <notmart> again, would need api from desktop scripting tough
[19:32] <notmart> but yeah, i can look into that
[19:32] --> friedrich (~friedrich at aextron.de) has joined this channel.
[19:32] --> stikonas (~gentoo at wesnoth/translator/stikonas) has joined this channel.
[19:33] <Sho_> thanks
[19:33] <-- palasso (~quassel at 2a02:214d:8240:d800:c829:e838:d135:50) has left this server (Ping timeout: 240 seconds).
[19:33] <Sho_> let's try to stay on that ball throughout this week
[19:33] <Sho_> second thing i need to inquire about
[19:34] <Sho_> due to me not paying attention, in the 5.8.x branch anthony got away with a dumb patch limiting the size of the TM group dialog which cuts off group popups and potentially doesn't show all tasks
[19:34] <Sho_> i rewrote the dialog for 5.9.x so it now has a scrollbar, and much better keyboard nav and other fixes
[19:34] <Sho_> i was going to backport this to 5.8.x after it saw some 5.9.x testing
[19:35] <Sho_> and this has now become more urgent because of https://bugs.kde.org/show_bug.cgi?id=378042
[19:35] <sKreamer> KDE bug 378042 in plasmashell (Task Manager) "Icons in Task Manager Dropdowns are too large, causing Dropdown to be clipped" [major,] https://bugs.kde.org/show_bug.cgi?id=378042
[19:35] <Sho_> however
[19:35] <Sho_> some users of the 5.9.x code report the rewritten dialog isn't scrollable for them -> https://bugs.kde.org/show_bug.cgi?id=375937
[19:35] <sKreamer> KDE bug 375937 in plasmashell (Task Manager) "regression: task manager group menu no longer shows all windows" [normal,WAITINGFORINFO] https://bugs.kde.org/show_bug.cgi?id=375937
[19:35] <Sho_> which makes the backport a shakier proposition
[19:35] <Sho_> however, i can't reproduce this
[19:35] <Sho_> i'd like others to test it and see if they can, so we can progress towards backporting it to 5.8.x
[19:36] <d_ed> Sho_: got a branch name?
[19:36] <d_ed> and/or phab link
[19:38] <notmart> how is the logic? after how many windows/what size should it get a scrollbar?
[19:38] <Sho_> d_ed: for the dialog or the faves?
[19:38] <llucas> What about backporting a fix for system settings crash with qt 5.7.1?
[19:38] <Sho_> notmart: the dialog is limited to like one third of screen height or something
[19:39] <notmart> in master, getting the scrollbar correctly
[19:39] <Sho_> d_ed: ivan/new-favourites-per-activity in plasma-desktop + https://phabricator.kde.org/D3805
[19:39] <sKreamer> Diff 3805 "Per-activity favorites (Final, again?)" [Needs Review] https://phabricator.kde.org/D3805
[19:40] <-- starbuck11 (~Blue_Syst at p57AF5881.dip0.t-ipconnect.de) has left this channel.
[19:42] <notmart> wonder if that gentoo user has some weird version of frameworks or qt..
[19:44] --> palasso (~quassel at 2a02:214d:8240:d800:c829:e838:d135:50) has joined this channel.
[19:44] <Sho_> notmart: i'm considering just taking the risk an backporting the dialog to 5.8.x
[19:44] <Sho_> worst case it won't break it /more/ since the 5.8.x code is already unscrollable
[19:44] <notmart> yep
[19:44] <Sho_> and if needs to be fixed later the same fix just goes into all branches
[19:44] <notmart> i'm fine withit
[19:45] <Sho_> ok, thanks
[19:45] <Sho_> that's it from my side then
[19:45] <Sho_> /meeting
[19:45] <Sho_> :)
More information about the Plasma-devel
mailing list