Update: <br>I've ported the bootstrap script to DCL (yuck) and it has most of the functionality of the bash version of the script.<br>I decided to take a dependency on the A9 porting library a.k.a. "The Jackets" (
<a href="http://h71000.www7.hp.com/openvms/products/ips/porting.html">http://h71000.www7.hp.com/openvms/products/ips/porting.html</a>) to help with some of the small issues I was having with compatibility of unix header files. It also includes a port of gmake (
3.76.1).<br><br>Compiling the bootstrap I run into only one issue:<br>In ProcessUNIX.c, there is a sigaction struct, and the version on VMS does not have the sa_sigaction member of the sigaction struct, and correspondingly the SA_SIGINFO constant is not defined. The sigaction struct looks like this:
<br> void (*sa_handler)(int);<br> sigset_t sa_mask;<br> int sa_flags;<br><br>As a fix would I get correct results if I were to instead of sa_sigaction, assign to sa_handler (the Unix manpage seems to claim that on some implementations sa_sigaction and sa_handler are a union - so they end up being called for the same reasons. My logic maybe flawed here.
<br><br>-Alan<br><br><div><span class="gmail_quote">On 9/20/07, <b class="gmail_sendername">Bill Hoffman</b> <<a href="mailto:bill.hoffman@kitware.com">bill.hoffman@kitware.com</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Alan Antonuk wrote:<br>> Hi;<br>><br>> I'm trying to get cmake to run on the OpenVMS operating system with a<br>> goal of eventually getting ITK to run on VMS. Searching through this<br>> list and on the website it doesn't appear as if this effort has been
<br>> made. My first question: has anyone successfully got cmake to run on<br>> VMS?<br>><br>> I am prepared to put in the effort to get cmake to run on<br>> OpenVMS. Looking at a mail-thread<br>> (<a href="http://public.kitware.com/pipermail/cmake/2006-October/011622.html">
http://public.kitware.com/pipermail/cmake/2006-October/011622.html</a><br>> <<a href="http://public.kitware.com/pipermail/cmake/2006-October/011622.html">http://public.kitware.com/pipermail/cmake/2006-October/011622.html
</a>>)<br>> posted earlier I can answer those questions:<br>> 1.) Is it command-line based? Does it support scripting in any form?<br>> ANSWER: Yes it is command line based, and it does support scripting<br>> called DCL - which is not compatible in anyway unix versions of *sh
<br>> 2.) What would the build sequence look like for a simple hello-world.c<br>> program?<br>> ANSWER:<br>> $ CC /OBJECT=hello-world.OBJ hello-world.c<br>> $ LINK hello-world.obj /EXECUTABLE=hello-world.EXE
<br>> C++ is similar<br>> $ CXX /OBJECT= hello-world.OBJ hello-world.cpp<br>> $ CXXLINK hello-world.OBJ /EXECUTABLE=hello-world.EXE<br>><br>> 3.) Is there a make tool or native build tool of some kind?<br>> ANSWER: Yes there's a utility called MMS which uses a syntax very
<br>> similar to make with some key differences (<br>> <a href="http://h71000.www7.hp.com/doc/73FINAL/5825/5825_014.htm">http://h71000.www7.hp.com/doc/73FINAL/5825/5825_014.htm</a> - look at<br>> Appendix B)<br>>
<br>> My plan of attack for porting cmake - would be to port the bootstrap<br>> file to use the DCL scripting language, and then go from there trying<br>> to build it. Is this a sensible way to go?<br>I looked into this in the past. My approach was going to be to first
<br>use gmake on OpenVMS which is ported. So, if you used the cmake "Unix<br>Makefiles" generator and gmake, you should be able to just create some<br>platform files to get this going.<br><br>-Bill<br><br></blockquote>
</div><br><br clear="all"><br>-- <br>+––– –– – – – – –<br>| Alan Antonuk<br>| <a href="mailto:antonuka@msu.edu">antonuka@msu.edu</a>, <a href="mailto:aega@umich.edu">aega@umich.edu</a><br> tel cell: (734) 646-4476
<br>| tel work: (734) 647-1386<br> addr: 1705 Morehead Dr., Ann Arbor, MI 48103<br>| aim: alanxz1