<html>
<head>
<meta http-equiv="Content-Type" content="text/html;
charset=windows-1252">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<p><br>
</p>
<br>
<div class="moz-cite-prefix">On 2017-06-09 15:45, Ulrich Weigand
wrote:<br>
</div>
<blockquote type="cite"
cite="mid:OF3D3393D8.65829B81-ONC125813A.004B6B07-C125813A.004B93CE@notes.na.collabserv.com">
<p><tt><font size="2">Jonas Paulsson
<a class="moz-txt-link-rfc2396E" href="mailto:paulsson@linux.vnet.ibm.com"><paulsson@linux.vnet.ibm.com></a> wrote on 09.06.2017
15:40:03:<br>
<br>
> At -O1 and higher these are both optimized to have the
ret removed:<br>
> <br>
> *** IR Dump After Simplify the CFG ***<br>
> ; Function Attrs: nounwind<br>
> define void @test0() #0 {<br>
> entry:<br>
> call void @llvm.trap()<br>
> unreachable<br>
> }<br>
> <br>
> , so only at -O0 is there a ret remaining.<br>
</font></tt><br>
<tt><font size="2">So if we run the whole thing through clang at
-O0 (with expensive</font></tt><br>
<tt><font size="2">checks), do we still get errors?</font></tt><br>
<br>
<tt><font size="2">Bye,</font></tt><br>
<tt><font size="2">Ulrich</font></tt><br>
<br>
</p>
</blockquote>
<br>
Yes, with the terminator / barrier flags, it fails at -O0 ,but not
at -O3 (function returning an int).<br>
<br>
/Jonas<br>
<br>
<br>
</body>
</html>