proto: provide reason when refusing connections #2235
Conversation
| pub fn refuse_reason( | ||
| &mut self, | ||
| incoming: Incoming, | ||
| reason: Option<String>, |
There was a problem hiding this comment.
Connection::close takes Bytes. Should we do the same here?
There was a problem hiding this comment.
Why? ApplicationClose::reason is (publicly) Bytes but TransportError::reason is (publicly) String -- I feel like String is probably more reasonable in terms of the public API, and it doesn't feel like there's a bunch of performance to be had from refcounting error reasons instead of allocating?
There was a problem hiding this comment.
It gets baked down into ConnectionClose::reason, which exactly mirrors ApplicationClose::reason as Bytes, and is not strictly required to be UTF-8. The involvement of TransportError is an artifact of the private initial_close API, which we can easily revisit. Most important, though, is just that it would be more consistent with close, which is morally the same operation, just at a different time.
There was a problem hiding this comment.
- Reason Phrase: Additional diagnostic information for the closure. This can be zero length if the sender chooses not to give details beyond the Error Code value. This SHOULD be a UTF-8 encoded string [RFC3629], though the frame does not carry information, such as language tags, that would aid comprehension by any entity other than the one that created the text.
Which makes me feel that String is a better API than Bytes.
|
@gretchenfrage you mentioned you want to review this -- friendly ping? |
|
Sorry, I hadn't meant to block this as long as I have! I'll try not to much longer. |
No description provided.