deleted DELETE

[Bron Gondwana]

On Thu, Aug 30, 2007 at 12:27:39PM -0400, Ken Murchison wrote:
> David Carter wrote:
>> On Wed, 29 Aug 2007, David Carter wrote:
>>> mboxname_isusermailbox() works on internal mailbox names. I don't think 
>>> that it needs to worry about IMAPOPT_UNIXHIERARCHYSEP.
>> Here's a trivial patch to remove the IMAPOPT_UNIXHIERARCHYSEP stuff. 
>> Otherwise mboxname_isusermailbox() is broken when unixhierarchysep is set.
>
> I've modified your patch further, to make it look closer to the original 
> code.  But I'm now wondering what the 'isdel' clause is trying to do. Once 
> we've stripped the deletedprefix and found "user.", we shouldn't care about 
> any timestamp, because a deleted Inbox, isn't an Inbox anymore.  Correct?  
> Actually, can any deleted mailbox be considered a user mailbox since only 
> the admin can see them?
>
> if (!strncmp(start, "user.", 5) && (!isinbox || !strchr(start+5, '.'))
>     || (isdel && (p = strchr(start+5, '.')) && !strchr(p+1, '.')))
> return (char*) start+5;

Actually, that's buggy. Here's the corrected version (full patch
attached)

  /* starts with "user." AND
   * we don't care if it's an inbox OR
   * there's no dots after the username OR
   * it's deleted and there's only one more dot
   */
  if (strlen(start) > 5 && !strncmp(start, "user", 4) && start[4] == '.'
      && (!isinbox ||
          !strchr(start+5, '.') ||
          (isdel && (p = strchr(start+5, '.')) && !strchr(p+1, '.'))
         ))
     return (char*) start+5; /* could have trailing bits if isinbox+isdel */


Basically, a user's inbox looks like:

domain.com!user.username

a deleted inbox looks like:

domain.com!DELETED.user.username.46D16085

(or whatever timestamp)

So if isinbox, we want to check that either there's no more dots if it's
not deleted, or one more dot if it is deleted.

Bron.