STEPS TO REPRODUCE 1. Load http://www.hixie.ch/tests/evil/state/001.html 2. Look at the number in the IFRAME at the bottom. 3. Change stylesheet (View | Use Stylesheet | Alternate) 4. Look at the number again. ACTUAL RESULTS Number changes. EXPECTED RESULTS We shouldn't lose the current scroll position, we shouldn't lose the current selection, we shouldn't reset any running scripts -- basically, we shouldn't touch anything about subdocuments during a reframe. This is probably another symptom of bug 52334. Marking dependency.

not a table specific bug, reassigning to core owner.

Yea, when we break the commection between layout and iframe document loading this will be fixed. It might be good to keep this bug open though, to makes ure we update the layout correctly WRT scroll-position, selection, etc. of the IFRAME.

So... the subdocument is no longer reloaded. Its layout is redone from scratch, however, since the layout object containing it is destroyed and recreated. I suppose we should make this a tracking bug for the ensuing symptoms.... for example, fixing scroll position is easy; doing selection is harder.

This bug breaks the Style Editor and Orion. (see bug 583041) The scroll position of a DIV inside Orion (which inside an iframe) is lost once the editor is hidden. When the iframe is made visible the Orion code becomes confused after the user tries to interact with the editor. This bug only affects Gecko - things work fine in Opera and Chrome.