On Apr 20, 2006, at 02:31, Cameron McCormack wrote:
Two options:

  1. never send the individual mutation events (and hope that the
     implementation does support the MutationNameEvents feature), or

  2. send the individual mutation events only if renameNode falls back
     to node replacement.

I'm not sure that I have a preference, but (2) is easier to implement.

As an author I would much prefer (1), as it means consistency. Option 2 is also a bunch of events, remove element, add element, set attributes (the two latter stages being having invertible orders, which has different effects on bubbling, etc.).

--
Robin Berjon
   Senior Research Scientist
   Expway, http://expway.com/



Reply via email to