While I like the goal of this, why not errata the existing spec to allow GET requests to have bodies? Would that be too much of a breaking change?
Adding stuff is easier than patching existing implementations is probably the reason why. Because up until this point, the webserver can just ignore the GET body request but if the client starts sending with body then who knows what bug might surface that turns into security vulnerability.
I like it. Let’s do it.
This is pretty good, though I expect even if it’s accepted it’s going to be a long journey before you can reliably use it
If we’re adding stuff to http, it would be nice for some additional status codes, things have moved on a bit since the early days
There should be a 400 that just says “fuck you and your AI”
I believe you can use 422: Unprocessable Content for this.
The request was well-formed (i.e., syntactically correct) but could not be processed.
“I know what you said, but I’m not answering you.”
That’s what 418 is for imo
