Re: I-D Action: draft-ietf-httpbis-safe-method-w-body-08.txt

On 01.03.2025 12:57, Amos Jeffries wrote:
> ...
> I have implemented test changes for Squid to support this feature set.
> ...

Very good.

> Some editorial nits:
>
> Section 2.4 requires normalization of the content for cache key, with
> several suggestions/requirements:
>
>   * The first one mentions content encoding but does not name the
> relevant header(s) (Content-Encoding, Transfer-Encoding) as the other
> bullet points do.

Just Content-Encoding, right? T-C should be orthogonal to caching.

>   * second and third bullet point appear to be saying the same thing.
> Just differently worded.

The first one is about normalization based on common formats (as in JSON
or XML), as per the "+format" media subtype convention (do we have an
RFC for that?). That is, some normalization can happen even if the
*concrete* media type is unknown.

OTOH, the normalization based on the semantics requires understanding
the concrete type.

Best regards, Julian

Received on Sunday, 2 March 2025 09:25:05 UTC