azul
azul copied to clipboard
fschutt
Fix wayland build
This is basically two fixes, the first one is because using winits hidpi is the proper method for wayland and using get_current_monitor() on a wayland window that hasnt yet been displayed leads to a panic.
I have absolutely not idea why the second fix works but it consists of not ignoring Refresh events and swapping the buffers before resizing the gl_window. Perhaps @vberger has an idea about this?
Pull Request Test Coverage Report for Build 887
- 0 of 0 changed or added relevant lines in 0 files are covered.
- No unchanged relevant lines lost coverage.
- Overall coverage remained the same at 0.0%
| Totals | |
|---|---|
| Change from base Build 831: | 0.0% |
| Covered Lines: | 0 |
| Relevant Lines: | 1409 |
💛 - Coveralls
I won't merge this right now, because this problem might be already fixed on the positioned_words branch - there are a lot of changes in that branch that may already fix this problem, right now is a bad time to merge. Also there is a reason why I'm filtering out Refresh events, that's not a bug, so please don't change that.
@fschutt both fixes are still needed despite the changes from the positioned_words branch.
Filtering out refresh events may be intentional however it causes the problem of the window not appearing on wayland, perhaps there's another way to fix this however.
Using actual master does not work when using Wayland. With this fix it seems to be ok. Need help updating the branch onto actual master?
@neyb I’d happily update this PR to master.
Unfortunately it seems that this PR is considered inadequate for the moment, which is fair enough considering it changes the Refresh event handling.
Saying that, I hope this issue will be resolved shortly as lacking support for wayland systems (gnome, sway, etc) is a huge issue for a GUI toolkit imo