Fill the windows in the examples with a solid color

[notgull]

• Tested on all platforms changed (only tested on X11)
• Added an entry to CHANGELOG.md if knowledge of this change could be valuable to users
• Updated documentation to reflect any user-facing changes, including notes of platform-specific behavior
• Created or updated an example program if it would help users understand this functionality
• Updated feature matrix, if new features were added or implemented

When you run the examples for winit, the contents of the window will be undefined depending on the platform. To the uninformed user, this appears to be a bug in some cases. In the past we've gotten quite a few issues because of this (#2807 comes to mind) that aren't issues at all.

This PR fills the windows with a solid gray background in order to make sure that there is something inside of the window.

Before (X11):

After (X11):

[fredizzimo]

I think this should be merged sooner rather than later. Currently the examples are unusable on Wayland, you can't even activate the window.

If users encounters bugs in Neovide, I would sometimes like to have the repeat the bug using one of the Winit examples, but if the examples don't work, that's not possible.

@notgull, Could you update to the latest master and also fix the new key_binding.rs example? That's a new example and does not have the fix.

[kchibisov]

That's true, though there's always glutin with window example, which I use as a testing or alacritty branches which I always open to target latest winit dev, since that's what I use myself as a maintainer to test winit API.

[notgull]

I've rebased on the latest master. So far softbuffer doesn't support Android (rust-windowing/softbuffer#44) or iOS (rust-windowing/softbuffer#43) yet, so I've just made fill_window a no-op on those platforms. The Windows CI errors are due to winit using windows-sys version 0.45 and should be fixed by #2842

[kchibisov]

All examples must handle Resize events, otherwise Wayland won't fly.

Probably macOS won't fly as well.

[kchibisov]

All examples must handle Resize events, otherwise Wayland won't fly.

Probably macOS won't fly as well.

Actually, you do Resize(inner_size), but the thing is that it's not really resized on Wayland, even when winit is telling the right sizes. I'd assume we'd need a new release of softbuffer?

it also busy drawing on Wayland even when you do wait and remove any requests for redraw, so I'd assume it's not working.

cc @ids1024

[ids1024]

At this point we might as well do another release of softbuffer, from the current master, instead of having things relying on 0.2. It's a breaking release, and it will probably have another breaking release for other things, but that's not really an issue at this stage.

[kchibisov]

At this point we might as well do another release of softbuffer, from the current master, instead of having things relying on 0.2. It's a breaking release, and it will probably have another breaking release for other things, but that's not really an issue at this stage.

That would be nice, given that it's un-usable in the current state on Wayland from what I've seen in this PR. The example on softbuffer master does work fine.

[notgull]

Now cargo-deny is flagging the memmap2 dependency. Shouldn't be a real issue.

[kchibisov]

Could you add memmap2 into deny and open an issue on client-toolkit(I can do this myself if you don't want to)?