When a shell client got mapped, unmapped and mapped again we emitted
the shellClientAdded signal in WaylandServer again. This resulted in
e.g. Workspace, EffectsHandler, etc. to start managing the window again.
This can be a reason for problems we see with such windows like the
Plasma panel dialog when opened the second time.
Details
Details
- Reviewers
sebas - Group Reviewers
Plasma on Wayland KWin - Commits
- R108:c7828eab8f61: Ensure that WaylandServer::shellClientAdded only gets emitted once
Needs extensive testing on real world system as that changes
behavior now.
Diff Detail
Diff Detail
- Repository
- R108 KWin
- Lint
Automatic diff as part of commit; lint not applicable. - Unit
Automatic diff as part of commit; unit tests not applicable.
Comment Actions
a few questions and suggestions, feedback welcome...
| autotests/wayland/shell_client_test.cpp | ||
|---|---|---|
| 130 | aren't you going overboard here? I don't think this makes the backtrace a lot easier to read... | |
| 181 | I wonder if this should be the default arg in the API. What do you think? | |
| 191 | Also check windowShownSpy.count()? | |
| 203 | Check count here? | |
| wayland_server.cpp | ||
| 251 | worth warning here, or is this a valid case? (Looks like an application bug if this return is hit) | |
| autotests/wayland/shell_client_test.cpp | ||
|---|---|---|
| 130 | backtraces here? This is an autotest. If that code crashes at that place we have a bigger problem than the readability. It's also a pattern I started to use in many tests as some tests showed to do the cleanup in a wrong way. | |
| 181 |
Not ABI compatible to change. And no, it should not. The idea is to limit the rendering of a client with a frame callback. That can only work if we set the frame callback. So the default should be with a frame callback. | |