<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 --- - IRForTarget"
href="https://llvm.org/bugs/show_bug.cgi?id=26790">26790</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>IRForTarget
</td>
</tr>
<tr>
<th>Product</th>
<td>lldb
</td>
</tr>
<tr>
<th>Version</th>
<td>unspecified
</td>
</tr>
<tr>
<th>Hardware</th>
<td>PC
</td>
</tr>
<tr>
<th>OS</th>
<td>All
</td>
</tr>
<tr>
<th>Status</th>
<td>NEW
</td>
</tr>
<tr>
<th>Severity</th>
<td>release blocker
</td>
</tr>
<tr>
<th>Priority</th>
<td>P
</td>
</tr>
<tr>
<th>Component</th>
<td>All Bugs
</td>
</tr>
<tr>
<th>Assignee</th>
<td>lldb-dev@lists.llvm.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>luke.drummond@codeplay.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>It looks like the recent changes committed in r260768 have caused regressions
when trying to call functions that take any arguments whose typename is the
result of a `typedef` involving untagged struct types.
e.g
```c++
typedef {
float f;
int i;
} mystruct;
mystruct s;
void somefunc(mystruct *s){}
int $__lldb_expr()
{
;
somefunc(&s);
}
```
lldb versions previous to r260768 resolved the anonymous struct without a
problem, though r260768 and later fail to match the function instead failing
with the following error:
<span class="quote">> "error: Couldn't lookup symbols:
> multiply($_0*)</span >
.
It seems that calling functions through pointers returned by
`IRForTarget::ResolveFunctionPointers` has been removed in favor of direct
calls through the function by name. That function previously called
IRForTarget::GetFunctionAddress` to resolve a function, and seems to have had
more robust lookups for matching functions did symbol/type resolution through.
I've attached a diff for a failing testcase, that I think would be helpful to
catch this regression in future. I haven't committed this test to the lldb
source tree yet, as I understand this may unduly upset the buildbots.
Apologies for the delay here, but it was a few days after this change went in
that we got this change locally.
Let me know if there's anything I can do to help test this, as I'm keen to get
this resolved as soon as possible - therwise it seems like the best course of
action is to revert r260768 as it seems there might still be some teething
problems with this commit (<a class="bz_bug_link
bz_status_NEW "
title="NEW --- - Expression evaluation broken for__attribute__((overloadable))"
href="show_bug.cgi?id=26694">https://llvm.org/bugs/show_bug.cgi?id=26694</a>
FindCodeSymbolInContext</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>