[webkit-reviews] review denied: [Bug 89000] [CSS Regions] Fix the lifecycle for the flow objects and their renderers : [Attachment 151436] Patch

[bugzilla-daemon at webkit.org]

Eric Seidel <eric at webkit.org> has denied Andrei Bucur <abucur at adobe.com>'s
request for review:
Bug 89000: [CSS Regions] Fix the lifecycle for the flow objects and their
renderers
https://bugs.webkit.org/show_bug.cgi?id=89000

Attachment 151436: Patch
https://bugs.webkit.org/attachment.cgi?id=151436&action=review

------- Additional Comments from Eric Seidel <eric at webkit.org>
I've read this another 3 times through. :)  I think I finally understand what
you're going for here.	Please answer my above questions and upload a new copy.
 I think this needs one more round.

Part of me would rather you did this in smaller pieces.  Like, for example, the
item() and length() stuff doesn't need to be added until it's
tested/used/exposed to the web.

None of these weak pointers are actually necessary, since the AtomicString
representing the name of the flow is sufficient to look up any flow or flow
renderer.  But the pointers may be easier to use in the end, unsure.

The test changes are not at all clear to me.  Are those just changes to the
tests themselves?  Or is there a behavior change hidden in here?  Could we do
the test chagnes separately?  Could we do whatever the behavior change is here
in the smallest patch possible?