<html>
<head>
<base href="https://bugs.webkit.org/" />
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_NEW "
title="NEW - DFG::safeToExecute() is wrong for MultiGetByOffset, doesn't consider the structures of the prototypes that get loaded from"
href="https://bugs.webkit.org/show_bug.cgi?id=147250#c1">Comment # 1</a>
on <a class="bz_bug_link
bz_status_NEW "
title="NEW - DFG::safeToExecute() is wrong for MultiGetByOffset, doesn't consider the structures of the prototypes that get loaded from"
href="https://bugs.webkit.org/show_bug.cgi?id=147250">bug 147250</a>
from <span class="vcard"><a class="email" href="mailto:fpizlo@apple.com" title="Filip Pizlo <fpizlo@apple.com>"> <span class="fn">Filip Pizlo</span></a>
</span></b>
<pre>Oh man, this is super hard to test. We end up not hoisting MultiGetByOffset in LICM in any of the situations where safeToExecute() would have been wrong - if the loop is responsible for adding or removing the property that is being loaded, then it will appear to clobber JSCell_structure or the property. This means that of MultiGetByOffset relies on some CheckStructure to prove the presence of a property in a prototype, then that CheckStructure will be hoisted anytime that MultiGetByOffset is hoisted.</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are the assignee for the bug.</li>
</ul>
</body>
</html>