March 24, 2026
6:22 a.m.
On Mon Mar 23 20:27:19 2026 +0000, Elizabeth Figura wrote:
Sure, I didn't look at the details of what the I/O is actually used for. If we're that corrupted it probably doesn't make sense to allocate the overlapped in heap either... Indeed.
Honestly I'd prefer just terminating immediately if possible. The heap allocation suggestion was more of a hack for the case when you cannot abort immediately and still need to run all other cleanup logic. (Meanwhile waiting for completion can report successful I/O even after CancelIoEx(), which I believe would complicate the logic for not much gain here.) Sorry for not being clear enough! -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10424#note_133501