I have completely misunderstood the intent of USBIP_CMD_UNLINK to mean something else than it's supposed to be. In the current implementations, the correct behavior would be close to a no-op, because we handle all requests in a serial fashion (although #34 might provide the possibility of cancelling URBs).
I have removed the test that reflects the misunderstanding. I do not think it's worth writing a new test for the no-op behavior.
@jiegec: I'm not sure how many client libraries actively send these unlink requests, but I think this would warrant a patch release nonetheless. It might be a quite severe regression, rendering the library unusable for many people.
I have completely misunderstood the intent of
USBIP_CMD_UNLINKto mean something else than it's supposed to be. In the current implementations, the correct behavior would be close to a no-op, because we handle all requests in a serial fashion (although #34 might provide the possibility of cancelling URBs).I have removed the test that reflects the misunderstanding. I do not think it's worth writing a new test for the no-op behavior.
@jiegec: I'm not sure how many client libraries actively send these unlink requests, but I think this would warrant a patch release nonetheless. It might be a quite severe regression, rendering the library unusable for many people.
Fixes #39