I run http://archive.dojotoolkit.org/nightly/dojotoolkit/dijit/tests/form/test_Button.html and I get a lot of show events for not in document DOM nodes, therefore I get lot of assertions. So we could check this and ignore these events but I'm not sure in nature of this, I'm afraid we can miss major issues. Originally Ginn moved event processing into FlushPendingEvents in bug 418845, so we try to coalesce events when they are not in document and it doesn't make sense. I'm not sure really it helps in optimization because main optimization stuff should be ApplyToSiblings hack. That's more clear things should be if we will get back to event processing when delayed event is about to be fired (FireDelayedEvent) but we should have more smart event processing so that we should take into account the event type.

taking this bug

Hi Surkov, do you have a WIP for this one?

(In reply to comment #2) > Hi Surkov, do you have a WIP for this one? Actually I didn't touch this few months already. I tried to understand what's going wrong here and how to fix this but iirc I didn't write any robust WIP patch.

Making this bug kind of meta since "something strange" is complex issue, at least it looks so.

I don't see assertions any more and all dependent bugs were fixed, mark this bug as fixed.