search

httpwg / http-core

Core HTTP Specifications
https://httpwg.org/http-core/
470 stars 43 forks source link

make preconditions consistent on when they are required to be evaluated … #960

Closed royfielding closed 3 years ago

royfielding commented 3 years ago

… and when they can be ignored

fixes #954

royfielding commented 3 years ago

I do not disagree with the changes, but I am slightly concerned that we are changing normative requirements here:

* some requirements now only apply when the origin server actually process the conditional header field; that makes the mechanism essentially optional

The text above those requirements already requires that the origin server evaluate the condition, so there shouldn't be a change.

* at least one MUST became a SHOULD

Oops, that was my mistake when I pasted some text. Fixed.

If we agree on these changes we may also need to document as change from RFC723x.

I don't think there are any, unless the part about caches and intermediaries ignoring IUS and If-Match was not documented previously.

reschke commented 3 years ago

The text above those requirements already requires that the origin server evaluate the condition, so there shouldn't be a change.

I see. And after going back to the actual issue I also realize why we have this change.

reschke commented 3 years ago

I don't think there are any, unless the part about caches and intermediaries ignoring IUS and If-Match was not documented previously.

Ack. I see https://greenbytes.de/tech/webdav/rfc7232.html#rfc.section.3.1.p.7 and https://greenbytes.de/tech/webdav/rfc7232.html#rfc.section.3.4.p.8.