You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've seen drivers return VK_ERROR_UNKNOWN from random places, so all functions also return this error. I would suggest to just assert that there is no write instead of handling it. Maybe it would sense to do that at a library level? It said clean up some of these things, what do you think?
A future extension could add a possible error code (e.g. some of the "maintenance" extensions modify existing functions like this), and given that it returns a VkResult, Khronos might think it's a possibility. Although encountering the "unknown" case is a bit contrived as the user would need to have enabled the extension by name but be using an old version of the bindings that doesn't support that extension (i.e. they would not be able to use its functions).
EDIT: Also in this case where an error result is added later, if freeDescriptorSets() currently returned void then that would be a breaking change, even if you're not using that extension.
really? can something go wrong when freeing these?
edit: question remains, but I remembered that these are allocated out of a pool so it's uncommon to need to free them individually.
The text was updated successfully, but these errors were encountered: