Definition for ethernet flow control packets.

[AJMansfield]

This is a pretty obsolete protocol and ended up not being useful for what I'd hoped it would be, but I wrote an implementation for it anyway so the world might as well have it.

[stevedoyle]

Missing unit test?

[AJMansfield]

Missing unit test?

AFAIK the standard doesn't give any test vectors, so I'm not sure how to meaningfully test this.

Inventing a test vector myself wouldn't add anything: if there were an error, given the declarative nature of the code here the error would inevitably be a misinterpretation of the standard rather than a mis-implementation of that interpretation -- so adding a test vector I've concocted myself would likely amount to just writing down the same mistake twice.

Would using Wireshark to create the test vector be appropriate here? Given that it's a much more mature project it might not be unreasonable to consider its behavior as "authoritative enough".

[stevedoyle]

Would using Wireshark to create the test vector be appropriate here? Given that it's a much more mature project it might not be unreasonable to consider its behavior as "authoritative enough".

I think using Wireshark or tcpdump to capture a reference/test packet is a good approach.