On Tue Jan 28 20:54:24 2025 +0000, Rémi Bernon wrote:
IMO it would be cleaner to avoid bothering drivers with surfaces for windows that don't exist anymore, that's also true for the `vulkan_surface_update` callback.
Looks like vulkan_surface_update callback is only called from win32u_vkCreateSwapchainKHR() where surface->hwnd is currently checked for being a valid window before that?