Monday meeting notes for 2018-09-24 (week 39)

[Kai Uwe Broulik]

==== Eike ====

---- DONE ----

* [Wayland] Virtual desktops progress:
** "Switch Desktop" containment action plugin ported to libtaskmanager, 
bugs fixed
** "Switch Window" containment action updated to new libtaskmanager API, 
bugs fixed

---- PLAN ----

* working on making the prototype virtual desktops KCM better (e.g. not 
instant apply)
* trying to get the wayland virtual desktop stuff complete enough for an 
early-cycle 5.15 merge
** what this will eventually need:
** completing the above KCM work, and subsequently VDG input
** more code reviews, e.g. my libtm patch hasn't been reviewed at all so far
** i think there's outstanding review and maybe coding on marco's code 
too (I recall experiencing a recurring crash)
* figuring out what to do with the windows krunner plugin, which needs 
porting as well,
** but has the problem of being not in the plasmahsell process and 
needing the exclusive wayland protocols
** we should really try to merge this early in 5.15, because it is very 
likely going to need a lot of testing and bugfixing,
** and since we originally wanted this to be in 5.14 it'd suck a lot if 
it misses out another cycle

* we need to poke Riddell  to set up a 5.15 kickoff meeting, but this 
was important to discuss now already

---- Q & A ----

<kbroulik> Sho_: maybe dbusrunner can help, plasmashell exposing that 
stuff to krunner, but we must think about (not neccessarily address 
right now but have that in mind) security implications
<Sho_> kbroulik: I think a more pragmatic solution would be adopting the 
KCM's "only in=x11" plugin metadata thing to disable the windows runner 
on Wayland for now
<Sho_> kbroulik: And then asking KWin to figure out the auth thing 
before long

--

<romangg> In regards to not instant apply in the VD KCM: Is this 
alligned with what the VDG wants? There were several discussions in the 
past about (not) instant apply
<kbroulik> even if we were to change that (which I'm not saying we 
should) we must be consistent and not have a single new KCM randomly 
beahve differently
<Sho_> romangg: I think the VDG wants instant-apply, but it's not been 
decided yet and we can't just make individual KCMs step out of line randomly
<notmart_> it should use correctly the api needs apply, load, save
<Sho_> I agree with kbroulik
<notmart_> once it uses it really correctly, making the shell doing 
apply to save or instant apply is a switch
<Sho_> but the KCM was a prototype so far meant to help Marco mainly, 
and the code gets a lot more complex with delayed-apply, so i delayed 
the delayed-apply :D
<notmart_> unless the kcm does things by itself, then is a problem


==== Jonathan ====

!! Plasma 5.12.7 LTS is due tomorrow !!

Plasma 5.15 Kickoff meeting shortly after 5.14 release


==== Roman ====

---- DONE ----

* Rework of input redirection: D15595
* Rebased my Wayland touch drag code onto this: D15466
* My Xwayland DND patch series is now complete with addition of D15627 
D15628 D15629.

---- PLAN ----

* XDC 2018
* Respond to reviews
* Look into some issue in Xwayland's Present code discussed on xorg-devel

---- Q & A ----

<Sho_> romangg: So with the KRunner plugin thing we ran into the issue 
that it needs to use PlasmaWindowManagement, but that can only be used 
once right now by plasmashell
<Sho_> martin floeser had a plan there for some sort of auth system etc.
<Sho_> that's probably something to take up again in the mid-term future
<romangg> Do we really have this restriction in place already that it 
can only be used once?
<notmart_> the globals? not sure
<romangg> The question is if we really want to have auth via Wayland 
protocol at all. Nothing materialized in this direction anyways with our 
competitors/partners.
<romangg> So DBus only it might be in the end.
<notmart_> but exposing window tasks via dbus, kinda misses the point


==== Marco ====

---- DONE ----

* fix ripple effect for SwipeListItem on Material style
* enhancements in Card:
** possible to control the sourceSize and all Image proeprties for the 
header image
** possible to control wrapMode and size for the title
* fix icons:
** fix a regressions of icons that became smaller on mobile
** improve the heuristics on what icons should be monochrome and what 
shouldn't


==== Kai Uwe ====

---- DONE ----

* Made "Open..." action for finished copy/downloads also work in history
* Added Open Document thumbnailer (odt, ods, odp, etc)
* Fixed device actions not udpating live in Device Notifier
* Fixed Discover not showing update notification (also in 5.12)
* Made a patch for Krita to set decoration colorscheme: D15640
* Raise configuration window when reusing it: D15705
* Fixed (worked around) Ark opening Open Document files (that are 
*technically* archives) as archives


==== Fabian ====

---- Q & A ----

<fvogt> Someone asked this weekend why plasma-pa had only the 
contents.rcc file anymore, apparently it's indeed annoying for some 
"code editing users"
<fvogt> I had a look at that file and noticed that it's not compressed 
at all - so why are bundled packages still a thing?
<kbroulik> we had experimented with using qrc for packaging plasmoids 
but ended up not using it, perhaps was forgotten to remove again?=
<notmart_> yeah, can be removed
<notmart_> i still think it should be used, but ooooh, well
<kbroulik> "had only the contents"
<kbroulik> so something's wrong in cmake?
<fvogt> Would it be possible to just alias the _bundled variant in 
kpackage to the non-bundled one?
<notmart_> simply in plasma-pa is using the cmake macro to do bundled 
packages, should be changed that
-------------- next part --------------
A non-text attachment was scrubbed...
Name: meeting20180924.log
Type: text/x-log
Size: 9448 bytes
Desc: not available
URL: <http://mail.kde.org/pipermail/plasma-devel/attachments/20180924/2f0d024b/attachment-0001.bin>