- From: <marjorie@us.ibm.com>
- Date: Mon, 15 Nov 1999 19:13:05 -0500
- To: w3c-dist-auth@w3.org
This isn't perhaps that unexpected. If one models the lifecycle of a resource as a finite state machine, then it goes from non-existant (i.e., a reference to resource whose current state is undefined) to lock-null on LOCK. More locks leave it in this state. It only leaves this state and goes back to non-existant if there are no more locks. A PUT moves it from the lock-null state to the existing-resource state, etc. Greg Stein <gstein@lyra.org> on 11/11/99 07:58:06 PM To: w3c-dist-auth@w3.org cc: Subject: locknull resources Given the *current* spec :-), has anybody noticed that you can actually have a locknull resource that does not have any *direct* locks on it? For example: Establish a locknull as /a/b with a shared lock. Now, lock /a with a Depth: infinity lock, shared. Finally, unlock /a/b with the first locktoken. Just wanted to mention this to others, as it was pretty unexpected for me... Cheers, -g -- Greg Stein, http://www.lyra.org/
Received on Monday, 15 November 1999 19:15:46 UTC