It’s just data

Transfer Patches

Roy Fielding: My answer was to use PATCH, of course, but the standards issue is still that there are no registered diff formats.  So, I said to myself, I know how to fix that ... [via Rob Sayre]

A much more interesting question to me is whether PATCH will operate at the content level or the transfer level.  Or, to put it anther way, will patch operate at the infoset level, or will it be able to be directly applied to HTML as she is written?


Wouldn’t the difference between an infoset PATCH and a “bytes on wire” PATCH be found in the mime-type of the PATCH representation?

Posted by Joe Gregorio at

Lisa’s spec for PATCH uses Content-Type to reflect the type of the content.

Posted by Sam Ruby at

[link]

Posted by Joe Gregorio at

Sam, you’re looking at a completely outdated version of the draft.

As a matter of fact, the wrong usage of Content-Type has been one of Roy’s biggest complaints about the early PATCH drafts.

The current version is the one cited by Joe (or [link] if you prefer HTML).

Posted by Julian Reschke at

The Content-Type issue was one of the first things I changed when I updated Lisa’s original draft.  As currently written, PATCH carries a description of changes to be made to the content.

Posted by James Snell at

PATCH the message

Sam Ruby writes; A much more interesting question to me is whether PATCH will operate at the content level or the transfer level. Or, to put it anther way, will patch operate at the infoset level, or will it be able to be directly applied to HTML as...

Excerpt from Web Things, by Mark Baker at

PATCHes and Cream

It took several months, but we’re finally getting some really good discussions going about PATCH…. Rob Sayre 1 and 2 Roy Fielding Tim Bray Mark Nottingham Dare Obasanjo Sam Ruby 1 and 2 Joe Gregorio Aristotle Pagaltzis Update: Subbu Allamaraju For...

Excerpt from snellspace.com at

I need to reread this version of PATCH a few times, I don’t really see why Accept-Patch should be needed - surely Accept with the PATCH method should cover that.

But I do get the impression the patch operation is keying off the media type, which makes sense - especially given the need to do graph patches, e.g. see timbl’s old Delta doc [link] also (POST-based) ChangeSets [link]

Posted by Danny at

Danny:

Accept: PATCH” merely advertises the fact that the server will process PATCH requests at all. Accept-Patch has a completely different purpose. As you noted, knowing which media type to use in the PATCH request is of the essence; if the server can process application/xupdate+xml, that doesn’t mean it can equally process text/x-diff, much less your own home-grown concoction. So Accept-Patch is provided as a means for the server to advertise which patch formats it can process.

Posted by Aristotle Pagaltzis at

Aristotle, thanks, and I’ve now reread a couple of times. Ok, Accept-Patch is info on capabilities from the server (response) rather than from the client (request).

Dunno, having to make such a statement still seems a bit redundant somehow, almost like the smell of dependency. I mean, there doesn’t seem to have been a need for OPTIONS to have say Accept-Post: text/html. No doubt nothing worth worrying about :-)

Posted by Danny at

Add your comment