[webkit-reviews] review denied: [Bug 77135] [chromium] canvas demo is slow due to unnecessary resource cleanups : [Attachment 124764] proposed patch

[bugzilla-daemon at webkit.org]

James Robinson <jamesr at chromium.org> has denied Alok Priyadarshi
<alokp at chromium.org>'s request for review:
Bug 77135: [chromium] canvas demo is slow due to unnecessary resource cleanups
https://bugs.webkit.org/show_bug.cgi?id=77135

Attachment 124764: proposed patch
https://bugs.webkit.org/attachment.cgi?id=124764&action=review

------- Additional Comments from James Robinson <jamesr at chromium.org>
I'm afraid this doesn't work either, pointer identity != object identity in
C++.  Consider:

Layer removed from LTH A
LTH A destroyed
new LTH B allocated at the same addresses as LTH A
layer added to LTH B
pointer comparison passes, but all other state on the layer is still bogus!

This also doesn't address the problems I pointed out earlier with leaving the
LTH pointer set while the layer is out of the tree.