- From: Willy Tarreau <w@1wt.eu>
- Date: Tue, 1 Mar 2022 05:38:35 +0100
- To: "Backman, Annabelle" <richanna@amazon.com>
- Cc: HTTP Working Group <ietf-http-wg@w3.org>
Hello Annabelle,
On Mon, Feb 28, 2022 at 11:30:28PM +0000, Backman, Annabelle wrote:
> In light of the discussion regarding support for absent header fields in HTTP
> Message Signatures<https://github.com/httpwg/http-extensions/pull/1976>, I'd
> like to ask the perhaps obvious question: does HTTP permit header fields to
> make a semantic distinction between the header field being absent from a
> message, and the header field being present with field value equal to the
> empty string? If so, are there real-world use cases that people are aware of
> for this that might help inform that discussion?
We have such an example with the Host header field. Its presence is
mandatory in HTTP/1.1, though if you don't know what value to put in
it, an empty field will often work with plenty of servers and will
give you a response instead of a "400 bad request". There's also
Accept-Encoding (and probably more generally other Accept* though I'm
not sure), where absence means anything is accepted while emptiness
indicates nothing is accepted:
https://datatracker.ietf.org/doc/html/rfc7231#section-5.3.4
Hoping this helps,
Willy
Received on Tuesday, 1 March 2022 04:38:52 UTC