Cocoa_GetDisplayUsableBounds: Remove SDL_assert?

[past-due]

I'm not entirely sure why this SDL_assert call is here (possibly left-over from some debugging?), but this is getting triggered on my test system - despite the fact that the parent SDL calls handle a failure return code from Cocoa_GetDisplayUsableBounds properly. (And the public SDL_GetDisplayUsableBounds function also returns success / failure.)

Regardless of why this assert is getting hit, it doesn't seem like there should be an assert here, given everything that calls this is supposed to handle success / failure or return a success / failure value?

[slouken]

Can you debug into it and find out why the display can't be found?

[icculus]

I don't know why this is failing, but that assert is there because the heuristic is flaky and we wanted to know if/when it failed, so it did its job.

Is there anything unusual about your configuration? I'd like to know if screens was nil, or what thisDisplay was for each iteration vs cgdisplay.

Also, as the FIXME notes: we should just figure out the NSScreen when starting out, so we don't have this loop at all, which would also solve the problem.

[past-due]

I don't know why this is failing, but that assert is there because the heuristic is flaky and we wanted to know if/when it failed, so it did its job.

Is there anything unusual about your configuration? I'd like to know if screens was nil, or what thisDisplay was for each iteration vs cgdisplay.

I believe this was around the time the window was being transitioned (or had just transitioned?) from fullscreen to windowed.

Will be a bit before I have a chance to debug further, but it was pretty consistent on an (old) test system with macOS 10.15.