PUT + If-Match over-constrained?

[mnot@…]

I found this old post by Roy at:

​http://tech.groups.yahoo.com/group/rest-discuss/message/10100

"A more interesting question is whether the spec over-constrains the case where a PUT is successful but tried again. In other words, should the server be allowed to accept the PUT if the etag differs but the current state matches what is being PUT? Subversion handles such cases nicely because it is common for two developers to patch the same bugs. I think the "MUST respond with 412" is yet another case of a bogus requirement being added in 2616.

Note: this is an HTTP spec issue, not a REST issue.

....Roy"

HTTPbis still requires 412 when the only thing that differs is etag. Any chance this will change?

[fielding@…]

From [2354]:

Allow an origin server to respond with a success status if the precondition is on a state change that already succeeded; also, make the description of If-Unmodified-Since have the same conditions as If-Match; addresses #455