<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
<title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Thompson, David C wrote:
<blockquote
cite="midC01B7DAB36AB864E9C99F5B79655413A0ADF9E88@ES20SNLNT.srn.sandia.gov"
type="cite">
<meta http-equiv="Content-Type" content="text/html; ">
<meta name="Generator"
content="MS Exchange Server version 6.5.7650.28">
<title>RE: [CMake] Fortran-C mixed code (possible solution)</title>
<!-- Converted from text/plain format -->
<p><font size="2"><br>
<br>
Correct me if I'm wrong, but I believe there are some<br>
FORTRAN compilers which change the number of underscores<br>
appended depending on whether the name already contains<br>
an underscore (the g77 documentation on -funderscoring,
-fno-underscoring, and -fno-second-underscore mentions<br>
this). So perhaps more effort is required. :-(<br>
</font></p>
</blockquote>
<p><font size="2">They indeed exist. A typical example is g77:<br>
</font></p>
<p><font size="2">my_sub becomes my_sub__<br>
mysub becomes mysub_<br>
</font></p>
<p><font size="2">but as you need to provide a list of translations
manually, this is not a programmatic <br>
problem. Besides the scheme would provide the following names:<br>
my_sub, my_sub_ and my_sub__ (plus uppercase variants)<br>
mysub, mysub_ and mysub__ (plus uppercase variants)<br>
</font></p>
<p><font size="2">The linker would pick out whatever is suitable :)<br>
</font></p>
<p><font size="2">Regards,<br>
</font></p>
<p><font size="2">Arjen<br>
</font></p>
<br>
</body>
</html>