But there is the limitation that each handler you install this way
removes any other handlers that might want to listen for the event.
That's the part that worries me. If there is a way to do the same
thing with attachEvent, we ought to fix these, IMO.
On 2009-06-18, at 14:58EDT, Max Carlson wrote:
When I wrote that comment, I was under the impression that
event.preventDefault() didn't work reliably in all browsers.
Apparently, M$ has and event.returnValue property which does the
same thing when set to false.
So, I used the idiom of returning false to prevent the default
behavior, which can only be done by directly assigning a function to
an event handler. It's old skool, but it works everywhere!
See http://www.quirksmode.org/js/events_early.html for more details.
P T Withington wrote:
On 2009-06-17, at 21:10EDT, [email protected] wrote:
+// can't use lz.embed.attachEventHandler because we need to
cancel events
Can someone explain this comment to me? I see it in a number of
places, and it worries me. I'm concerned that we are using the
"assign event handler as property" interface, because of the
possibility that more than one place in the code will try to assign
to the same event and end up losing. I thought attaching event
handlers supported both cancelling bubbling and default
suppression, so I don't understand the comment.
Please enlighten me.
--
Regards,
Max Carlson
OpenLaszlo.org
