RE: Level 2 iterators

John Cowan wrote,
> Actually, there is no guarantee that walking
> the tree yourself with parent, child, and sibling 
> attributes doesn't lead you into a cul-de-sac
> either, when arbitrary tree mutations are going
> on.

Err ... then 'don't do that' (TM).

Seriously tho', that sort of issue should be
resolved by defining the semantics of the
DOM under mutations from arbitrary threads.

For my money those semantics should be
defined as 'undefined', and synchronization
should be achieved via the use of synchronized
wrappers (as seen in the Java2 collections
API).

Cheers,


Miles

-- 
Miles Sabin                          Cromwell Media
Internet Systems Architect           5/6 Glenthorne Mews
+44 (0)181 410 2230                  London, W6 0LJ
msabin@cromwellmedia.co.uk           England

Received on Friday, 5 March 1999 12:16:22 UTC