Re: Semantics: clarify handling of invalid or unsatisfiable range requests

Am 03.03.2022 um 18:38 schrieb Roy T. Fielding:
>> On Mar 2, 2022, at 10:54 PM, Eric J Bowman <
>> <>> wrote:
>> Wait, what? I feel like I'm speaking out-of-turn here nowadays. But
>> if, to satisfy some edge cases, PUT would now have the semantics of
>> PATCH then maybe we need a new method, let's call it PLACE, that
>> absolutely reflects the save-file-to-disk user intent, specifically
>> disallowing any sort of range request. Thought this was settled ages ago.
> This is just explaining what to do with an invalid or unsatisfiable GET
> range request
> in terms that are less vague. It does not change the protocol at all. It
> explains
> which error codes are preferred, which was already implied by the code
> semantics.
> It does not have anything to do with PUT ranges, which is not defined by
> this spec.
> If you think that needs to be clarified within this change, then why and
> where would
> be a more useful comment.

FWIW, a slightly modified iteration of this PR was just merged, and we
expect to finish publication preparations this week.

So please have a look at <>.

Best regards, Julian

Received on Monday, 14 March 2022 09:06:20 UTC