Opened 12 years ago
Closed 12 years ago
#251 closed design (fixed)
message-body in CONNECT request
Reported by: | mnot@… | Owned by: | mnot@… |
---|---|---|---|
Priority: | normal | Milestone: | 14 |
Component: | p2-semantics | Severity: | Active WG Document |
Keywords: | Cc: |
Description
I've come across a problem with an HTTP client which sends message bodies and Content-Length when using the CONNECT method.
Is it even legal for a CONNECT command to have a body? If it does, what should the proxy do? The client (a panasonic blue-ray player) breaks if the body is forwarded, however RFC2817 Sec 5.2 last para implies it is legal to have a body here.
Does anyone know of any other client that sends a body on CONNECT and relies on it? It seems very odd to have a body there, since you don't even have a connection necessarily to the end server yet. I'm loath to break the proxy to support a broken client.
Change History (4)
comment:1 Changed 12 years ago by mnot@…
- Owner set to mnot@…
comment:2 Changed 12 years ago by mnot@…
- Milestone changed from unassigned to 14
- Resolution set to incorporated
- Status changed from new to closed
see [1261]
comment:3 Changed 12 years ago by mnot@…
- Resolution incorporated deleted
- Status changed from closed to reopened
comment:4 Changed 12 years ago by mnot@…
- Resolution set to fixed
- Status changed from reopened to closed
Proposal -- for each of our methods, define whether a request body has any semantics; where it doesn't, say that including a request body might break some implementations. Also mention in considerations for new request methods.