Ian Douglas Scott

I think the only reason I marked this as a draft is because it's a breaking change. So if there's going to be a breaking release, it should be mergable.

CI passing, after also updating a line in `planes` example.

> Note that this only happens for the client side, since servers are expected to support protocols indefinitely (P.S. should this be added to the description, if so only on...

I've found the name a bit odd, and `RecieverSource` does seem a bit more descriptive.

Could your program be refactored to write into the `softbuffer::Buffer` instead of a separate buffer? At least on some backends, that should perform better than any possible implementation that presents...

If you want to write something specifically for Windows and have some very specific requirements, you can also always just do that directly with windows-rs instead of using Softbuffer. I...

Ah, fails to link on FreeBSD 12.1? Well, I guess that answers my question of when this was added. Not sure if there's any easy way to conditionally build based...

Would we need to `dlopen` `libc.so` to do that? A bit awkward. FreeBSD also has a `SHM_ANON` flag that can be passed as the path for `shm_open`. Which I think...

It might help to look at `WAYLAND_DEBUG=1` logs from both clients and see if there's anything prominently different about the calls being made. Smithay-client-toolkit is a relatively thin wrapper over...

> Also, is SlotPool::raw_data_mut really sound considering aliasing? You mean because the server may also be accessing the same memory? If the server isn't mutating the memory, this isn't unsound...