On Wed Jun 28 03:03:57 2023 +0000, Davide Beatrici wrote:
I wonder whether this is the actual behavior on Windows. `winepulse` is the only driver that sets the pointer to `NULL` (and only after validating the device).
It's rather difficult to test on Windows since the failures generally shouldn't happen. I'd probably move the `NULL` init to the beginning.
Also, I'd probably use `E_UNEXPECTED` for the `dataflow` error - again it's hard to test since this really shouldn't fail.