Re: [whatwg/fetch] non-null body with GET/HEAD requests (Issue #1705)

> Isn't this kind of spec-noncompliance what wintercg/fetch is for?

Yes and no. The goal is not really to maintain a forked spec that collects a bunch of non-compliant changes. Ideally if changes are needed an effort would be made to influence the direction of the core spec *first*, thus this issue. If this effort to introduce this modification to the core fetch spec fails, then yes, my next step would be to seek consensus among wintercg participating runtimes to collectively agree that such a change is worthwhile.

> I don't think there's a reason to get rid of this error on the web, since such bodies are explicitly disallowed there.

By "on the web" here I think you are specifically referring to Web browsers in particular, correct? The core specification of HTTP (which I would argue is generally more fundamentally "the web" but that's a debate for another time) specifically *allows* bodies for `GET` and `HEAD` even if those same specifications say they don't have any semantic meaning and should be avoided. The reality is that there are real world, in-production applications in use today on the web that do, in fact, attach a payload body to a `GET` request and that the fetch API as it is defined today is incapable of supporting those.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/fetch/issues/1705#issuecomment-1736578318
You are receiving this because you are subscribed to this thread.

Message ID: <whatwg/fetch/issues/1705/1736578318@github.com>