System settings KCM redesign

[Martin Flöser]

Am 2017-07-31 16:15, schrieb kainz.a:
> Hi,
> 
> should we start a redesign of ALL KCMs. Is someone interested in
> participate the mockup discussion and afterwards the code work.
> 
> The idea is to redesign ALL kcms and not some specific ones so in the
> end the design language should be for all the same, but the layout
> could be different.

Just to give you some facts:
  * KWin has 20000 lines of KCM code
  * 12 % of KWin's git repository is KCM
  * Kcm code has no tests at all
  * KCM code is ancient (Copyright (c) 1998 Matthias Ettrich 
<ettrich at kde.org>)
  * The main window management configuration module had 3 (!) meaning 
full changes since the port to Qt

Personal interpretation:
This means just for KWin this is a huge project. Changing the KCM look 
and feel will require a complete rewrite of the configuration module. 
This will cause many regressions due to the fact that nobody knows the 
code and nobody knows how this currently actually works [1]. There is 
lots of magic going on between KWin core and the KCMs. The size of the 
project (20000 new lines of code) is comparable to the Wayland port in 
KWin (all the platform plugins are ~10 kSLOC, libinput support is 2 
kSLOC, new files added to KWin for input handling and Wayland server are 
about 7 kSLOC).

Please don't interpret that as I'm against it. I just want to give an 
overview of the work required for a redesign. Personally I think it 
means it's impossible to change all KCMs at once.

For KWin I would not accept a code dump changing everything as the risk 
of breakage is too high. We could talk about maybe one KCM per release 
cycle, otherwise chances are too high that regressions will kill us.

Cheers
Martin

[1] Been there, changed stuff, broke stuff - always.