Receipt of STOP_SENDING on already-closed stream

[dtikhonov]

As of today, the draft states the following:

Receiving a STOP_SENDING frame for a locally-initiated send stream that is “Ready” or non-existent MUST be treated as a connection error of type PROTOCOL_VIOLATION.

Due to packet loss, it is possible for a STOP_SENDING frame to arrive after the stream initiator closes the stream. In this case, it should not be an error.

[mikkelfj]

There are many other reasons why this could happen, but I don't think this is a problem. A stream does not cease to exist just because it is closed. From memory that state chart should be ok. But since you raise this point, perhaps clarification is needed.

[dtikhonov]

The term "non-existent" is only used once in the entire draft. I presumed it meant "not yet opened or already closed."

EDIT: I meant to type "not yet opened or already closed."