Background: The parsing of the begin and end attributes of animations is
tolerant to forward references and elements added later by allowing for a retry
of the parsing upon certain events. This is complicated by the fact that begin
targets and end targets can be different and resolved at different times.
The current implementation is simplistic in that it just keeps parsing
everything until it resolves everything. Unfortunately, this can result in
event listeners being added more than once for the target nodes as they are
resolved. The resulting events listeners are being called more than once,
resulting in excessive work being done.
The solution is to keep track of whether targets have been resolved already,
individually, and only add the event listeners once.
Original issue reported on code.google.com by grick23@gmail.com on 20 Nov 2010 at 10:06
Original issue reported on code.google.com by
grick23@gmail.com
on 20 Nov 2010 at 10:06