Next message: Geoffrey M. Clemm: "Re: postcondition for PUT"
Date: Fri, 25 Aug 2000 17:02:40 -0400 (EDT)
Message-Id: <200008252102.RAA05163@tantalum.atria.com>
From: "Geoffrey M. Clemm" <geoffrey.clemm@rational.com>
To: ietf-dav-versioning@w3.org
Subject: Re: checkout-fork
Date: Tue, 15 Aug 2000 02:28:00 -0700
From: Greg Stein <gstein@lyra.org>
I think we need to differentiate the CHECKOUT responses between the
"predecessor exists" and the "another working resource exists" cases for the
checkout-fork properties.
Specifically, I want to know whether the failure was caused by a predecessor
or by another working resource. Knowing the condition, I can issue the
appropriate warning to the user and/or retry with the "fork-ok" in the
request body.
OK, I'll split that into two separate preconditions.
Note: Before the draft goes to last call, we need some way to have a
way to let the client know which precondition failed ... using a separate
error code for each would use up too many error codes.
Currently, I'm inclined to add an "Error-Code" header, which takes a
comma-separated list of tokens and coded-url's. I'll follow up to
Jim's "status code" thread with this proposal. We could use the response
body, but somebody once said that this would be bad for some reason,
but I can't remember who, or what the reason was (:-).
Cheers,
Geoff