John Baumgarten wrote:

> In the pending maintenance release running in our lab (not necessarily 
> what's up on .mac today), we return, respectively:
> HTTP/1.1 412 Precondition Failed: Lock-Token header missing
> HTTP/1.1 409 Conflict: Lock-Token is invalid
>> .Mac server implementors could tell us what they do...


thanks for the feedback. This contributes to the picture -- clients will 
have to expect almost everything between 400 and 499.

At the end of the day, a revised spec should

- define specific RFC3253-like precondition identifiers, so that new 
software can precisely find out what went wrong,

- give server implementors guidelines which 4xx code to choose (in 
particular, we should answer that question whether "lock-token" 
contributes to the set of request headers that can cause a 412 
Precondition Failed),

- give client implementors guidelines how to interpret server responses 
(such as first check for DAV:error response body, then possibly fall 
back to generic HTTP status usage).

I'll try to capture this in 

Best regards, Julian

<green/>bytes GmbH -- -- tel:+492512807760

Received on Tuesday, 8 June 2004 02:50:17 UTC