[v8] Do not install accessors if element could not be retrieved with it
Created attachment 76694 [details] Patch
See http://code.google.com/p/chromium/issues/detail?id=63241
What's the reason for this patch? Should it have a layout test?
I didn't manage (yet) to create a proper layout test. The reason is to fix Chromium bug http://code.google.com/p/chromium/issues/detail?id=63241 (see comment #2) And it does fix it :) (In reply to comment #3) > What's the reason for this patch? Should it have a layout test?
Given that this is a bugfix, we really should have either a layout test or a good case for why one isn't feasible. I'm happy to help in any way I can to reduce the case provided in the chromium bug report.
Comment on attachment 76694 [details] Patch r- for lack of test or explanation why testing is impossible in the ChangeLog.
I'll try another time to either test or explain better why it's involved. But I don't have spare cycles right now, sorry. (In reply to comment #6) > (From update of attachment 76694 [details]) > r- for lack of test or explanation why testing is impossible in the ChangeLog.
Comment on attachment 76694 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=76694&action=review > WebCore/bindings/v8/V8DOMWindowShell.cpp:532 > + // If there is no matching elements, do not install accessor. > + if (!doc->documentNamedItems(name)->length()) > + return; Why does this happen? I think we need to figure out how this can happen to proceed.
V8 is gone from WebKit.