Ian Douglas Scott

WIP Use euclid crate for geometry types

2

This does not yet compile. It seems one difference between the Euclid types and the ones Smithay currently has is that `euclid`'s points cannot be added, and there's a seperate...

Compositor panic with `zwlr_layer_surface_v1::set_size(0, 0)`

1

This can be reproduced by modifying https://github.com/Smithay/client-toolkit/blob/master/examples/simple_layer.rs to use `set_size(0, 0)` and running it under a compositor like Anvil: ``` thread 'main' panicked at 'called `Option::unwrap()` on a `None` value',...

I'm not entirely sure how tablet events should be handled, and how this interacts with applications using `tablet-unstable-v2`. But it seems unexpected that moving the cursor using the tablet moves...

`glBlitFramebuffer` fails if default framebuffer doesn't use `STENCIL_INDEX8`

2

This seems to be an issue running with ANGLE: https://github.com/Smithay/smithay/pull/1082. Changing `STENCIL_INDEX8` to `DEPTH24_STENCIL8` makes it run without errors and render correctly. Should Smithay somehow detect the format to use...

xwm: Fix setting `_NET_ACTIVE_WINDOW`

7

This is probably not the best way to handle this, but I'm not sure what is. It requires a new root argument to `X11Surface::new`, which is a change to the...

Provide way to configure default decoration mode

2

`KdeDecorationStateKdeDecorationState::new` takes an argument for the default mode. `XdgDecorationState::new` doesn't take a default, and `send_configure` defaults to `ClientSide` if `decoration_mode` hasn't been set. I think there should be a way...

This occurs with anvil and cosmic-comp. Presumably this should generate a protocol error and not crash the compositor. This can be reproduced easily by modifying https://github.com/Smithay/wayland-rs/blob/master/wayland-client/examples/simple_window.rs to remove the line...

A start on trying to implement https://github.com/Smithay/smithay/issues/751. Some improvement here will be needed to make both the X and Wayland backends behave properly. As seen in https://github.com/Smithay/smithay/issues/646, the current way...

Use a wrapper type implementing `Sync`, instead of `static mut`

2

`static mut` is generally best avoided. It seems to only be necessary currently to allow the use of non-`Sync` types, since pointers are not sync. This also removes the `NULLPTR`...

Not sure how best to test this but it would be good to handle: * Partial writes * Reads and writes that aren't multiples of 4 bytes * EINTR *...