[webkit-reviews] review granted: [Bug 11729] REGRESSION: Crash closing page with frames after selection : [Attachment 12055] Minimal fix

[bugzilla-request-daemon at macosforge.org]

Darin Adler <darin at apple.com> has granted Darin Adler <darin at apple.com>'s
request for review:
Bug 11729: REGRESSION: Crash closing page with frames after selection
http://bugs.webkit.org/show_bug.cgi?id=11729

Attachment 12055: Minimal fix
http://bugs.webkit.org/attachment.cgi?id=12055&action=edit

------- Additional Comments from Darin Adler <darin at apple.com>
The function name pageDestroyed is no longer a good one if it's used in these
cases where the page is not being destroyed.

And if it's an error to call setView(0) without also calling pageDestroyed(),
perhaps a little refactoring of that is in order to make that mistake
impossible.

Since FrameLoader::detachFromParent does some of the same kind of work that
removeChild does, I'd like to see some refactoring so that code does not exist
in two different place.

It's a bad thing that the page's frame count is now managed both by Frame and
FrameTree -- that should not be a shared responsibility.

Still, r=me, because we want this fixed in the tree; lets see if we can clean
things up a bit after the fact.

If maybe Geoff is doing some bigger fix maybe that will do the trick.