Bjoern Hoehrmann wrote:
So if we add a requirement that the event must occur at least once every hour it would be sufficiently well-defined to be implemented (in an implementation of the minimum requirements to successfully execute a minimal test program that tests this)?
That would be testable, yes. But as you pointed out, this is not really a useful requirement. The "fire before painting the changes" requirement is a much more useful one, so if we can actually make that well-defined we should, imo.
My basic problem with implementing this is that if I don't fire the DOMSubtreeModified event sync, then the question becomes "when is it OK to fire it?" What if I do a DOM mutation, then put up an alert and wait for two hours? Should the event fire while the alert is up? Or not?
-Boris
