<html>
<head>
<base href="https://llvm.org/bugs/" />
</head>
<body><table border="1" cellspacing="0" cellpadding="8">
<tr>
<th>Bug ID</th>
<td><a class="bz_bug_link
bz_status_NEW "
title="NEW --- - __builtin_object_size is not consistently conservative with C++ inheritance"
href="https://llvm.org/bugs/show_bug.cgi?id=26741">26741</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>__builtin_object_size is not consistently conservative with C++ inheritance
</td>
</tr>
<tr>
<th>Product</th>
<td>clang
</td>
</tr>
<tr>
<th>Version</th>
<td>trunk
</td>
</tr>
<tr>
<th>Hardware</th>
<td>Macintosh
</td>
</tr>
<tr>
<th>OS</th>
<td>All
</td>
</tr>
<tr>
<th>Status</th>
<td>NEW
</td>
</tr>
<tr>
<th>Severity</th>
<td>normal
</td>
</tr>
<tr>
<th>Priority</th>
<td>P
</td>
</tr>
<tr>
<th>Component</th>
<td>Frontend
</td>
</tr>
<tr>
<th>Assignee</th>
<td>unassignedclangbugs@nondot.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>bob.wilson@apple.com
</td>
</tr>
<tr>
<th>CC</th>
<td>llvm-bugs@lists.llvm.org
</td>
</tr>
<tr>
<th>Classification</th>
<td>Unclassified
</td>
</tr></table>
<p>
<div>
<pre>Created <span class=""><a href="attachment.cgi?id=15946" name="attach_15946" title="patch for the object-size.cpp test">attachment 15946</a> <a href="attachment.cgi?id=15946&action=edit" title="patch for the object-size.cpp test">[details]</a></span>
patch for the object-size.cpp test
Clang r250488 added code to make __builtin_object_size more conservative when
dealing with an array of size 0 or 1 as the last member in a struct to support
the "writing off the end" idiom. The CodeGen/object-size.cpp test was added
along with that patch. I noticed that if I change the test2() function in that
test to use a buffer size of 1, then it does not report the object size in the
same conservative way. It didn't look like that was intentional. The
isDesignatorAtObjectEnd() function in lib/AST/ExprConstant.cpp is not handling
the case with C++ inheritance and so it gives up. The catch is that "giving up"
in this case does not lead to the conservative behavior. The attached patch to
the test shows what I would expect to happen.</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are on the CC list for the bug.</li>
</ul>
</body>
</html>