Re: Plasma Sprint discussion planning
Am 2017-01-16 10:48, schrieb Marco Martin: On Sunday 15 January 2017 21:36:32 Martin Gräßlin wrote: Just giving an idea: Some parts where KWin is responsible for on Wayland, it could also become responsible on X11 and take over the part of whatever component provided it before. E.g. KWin could manage the keyboard layout on X11 as it has all the code for Wayland. The glue code on X11 would be rather minimal. That's also what I want to focus on in KWin internally. My aim for 5.10 is to replace more and more of the X11 specific code through the new implementations added for Wayland. didn't occur to e that it would be possible.. if it is sounds interesting indeed.. would be the various things like the keyboard layout be independent enough to beadopted by different people? Yes, so in KWin this will be: * KeyboardLayout class (review open, platform independent) * which uses a KeyboardLayoutIndicator (can be taken from existing plasma-desktop/kcms/keyboard with all X11 specific code removed - platform independent) * Xkb class from KWin (platform independent) * An Xkb sub-class in platform/x11/standalone which has the X11 specific bits So I expect for the keyboard layout handling in the end something like 50 to 100 lines of X11-specific code. Thanks xkbcommon for making keyboard layouts sane! Cheers Martin
Re: Plasma Sprint discussion planning
On Sunday 15 January 2017 21:36:32 Martin Gräßlin wrote: > Just giving an idea: Some parts where KWin is responsible for on > Wayland, it could also become responsible on X11 and take over the part > of whatever component provided it before. E.g. KWin could manage the > keyboard layout on X11 as it has all the code for Wayland. The glue code > on X11 would be rather minimal. > > That's also what I want to focus on in KWin internally. My aim for 5.10 > is to replace more and more of the X11 specific code through the new > implementations added for Wayland. didn't occur to e that it would be possible.. if it is sounds interesting indeed.. would be the various things like the keyboard layout be independent enough to beadopted by different people? -- Marco Martin
Re: Plasma Sprint discussion planning
Am 2017-01-13 16:27, schrieb Marco Martin: 3. Maintaining X11 specific code: Over the last half year exclusively on Wayland I realized that I cannot maintain X11 specific code any more. KWin is mostly no problem as XWayland windows are used the same way as native X windows. But things which have a different implementation on Wayland I cannot maintain any more. This affects mostly kglobalaccel, klipper and partly the KWin-standalone platform. I want the team to think about plans how we can continue to maintain the X11 specific parts especially with more of the team (hopefully) moving to Wayland. This is one of the most important things that have to be discussed at the sprint, yes Just giving an idea: Some parts where KWin is responsible for on Wayland, it could also become responsible on X11 and take over the part of whatever component provided it before. E.g. KWin could manage the keyboard layout on X11 as it has all the code for Wayland. The glue code on X11 would be rather minimal. That's also what I want to focus on in KWin internally. My aim for 5.10 is to replace more and more of the X11 specific code through the new implementations added for Wayland. Cheers Martin
Re: Plasma Sprint discussion planning
On Thursday 12 January 2017 20:39:39 Martin Gräßlin wrote: > 1. Global Touchscreen gestures: > ... > 2. Global Touchpad gestures: > we started on that, unfortunately the design document went into the > world of gestures which cannot be supported, since then it didn't go > anywhere. I would love the team to have a look at it again and define > the gestures to implement. I'll give another go at obtaining an updated design document, for both if not, we'll talk about that at the sprint, as we have several developers with many years of experience in designing UI. I refrained from that so far, as I didn't want to prevaricate > 3. Maintaining X11 specific code: > Over the last half year exclusively on Wayland I realized that I cannot > maintain X11 specific code any more. KWin is mostly no problem as > XWayland windows are used the same way as native X windows. But things > which have a different implementation on Wayland I cannot maintain any > more. This affects mostly kglobalaccel, klipper and partly the > KWin-standalone platform. I want the team to think about plans how we > can continue to maintain the X11 specific parts especially with more of > the team (hopefully) moving to Wayland. This is one of the most important things that have to be discussed at the sprint, yes -- Marco Martin