- From: Joe Orton <joe@manyfish.co.uk>
- Date: Thu, 11 Jun 2020 20:13:14 +0100
- To: ietf-http-wg@w3.org
Apologies if this has already been discussed, I don't see anything in errata and the text is very confused. https://tools.ietf.org/html/rfc7616#section-3.4 says that the uri= parameter sent by the client should be the "effective request URI" as defined by 7230 and explicitly not the request-target: uri The Effective Request URI (Section 5.5 of [RFC7230]) of the HTTP request; duplicated here because proxies are allowed to change the request target ("request-target", Section 3.1.1 of [RFC7230]) in transit. 3.4.3 then says the A2 construction uses request-uri, a term which is never explicitly defined (except in RFC2616, which is not referenced): A2 = Method ":" request-uri ... unless you assume that request-uri means "effective request URI", which is non-obvious. 3.4.6 then defines "request-target" as if *that* was used in e.g. A2 construction, which it wasn't. All the examples then use absolute-path in uri=, which is not an "effective request URI" (right?), further muddying the waters. If request-target is supposed to be used in uri=, that makes sense to me; strike all mention of "effective URI" and "request-uri", use "request-target" everywhere and define that uri= takes the request-target. (This is then basically the same as 2617 as well) How have current 7616 implementations dealt with this? Stick with 2617-style request-target everywhere? Regards, Jos
Received on Thursday, 11 June 2020 19:13:30 UTC