Grettings, All--
Has anyone successfully installed the PGPLOT.pm module on a win32 box? I have tried numerous methods, including trying to compile it myself. It seems to be beyond my current abilities-- if anyone could take pity on me and point me to a working install method, I'd be much obliged.
This is something I've thought of tackling before .... but you need to build the PGPLOT library first (don't you ?) - and that has always been the part that has deterred me.
Have you successfully built the PGPLOT library - if so, from which sources ?
I just found what looked like a promising source of PGPLOT at
http://jrfonseca.dyndns.org/projects/gnu-win32/software/ported/ in the
form of a file named pgplot-5.2.2-mingw32.tar.bz2. It contains the pgplot binaries - and since I have MinGW (and a MinGW-built perl) I was hopeful that I'd be able to get it to work with the PGPLOT module.
The PGPLOT module has, as a prerequisite, ExtUtils::F77. I found that for it to test succesfully I had to make the following modification to F77.pm. The subroutine 'find_in_path' starts:
sub find_in_path {
my @names = @_;
my @path = split(":",$ENV{PATH});I changed that to:
sub find_in_path {
my @names = @_;
my @path;
if($^O =~ /mswin32/i) {
for(@names) {$_ .= '.exe'}
@path = split(";", $ENV{PATH});
}
[EMAIL PROTECTED] = split(":",$ENV{PATH})}The module was then able to find my g77 compiler (which is part of MinGW).
With ExtUtils::F77 installed, it was simply a matter of editting the PGPLOT Makefile.PL to let the build process know where cpgplot.h, libcpgplot.a and libpgplot.a were located. (Those 3 files came with the PGPLOT binaries).
The PGPLOT module then compiled fine, but when I came to run 'dmake test' the first test complains that the "Procedure entry point pgarro_ could not be located in the dynamic link library pgplot.dll" - from which I deduce that there's some sort of incompatibility between the PGPLOT module and the pgplot.dll that I have. Or maybe there's just simply a problem with that dll. I think it's quite possible that PGPLOT-2.18 needs to be built against PGPLOT-5.2.0 (as per the README instructions) - not against version 5.2.2 as I have tried.
The subsequent tests complain that they can't find any code for the pgqinf() function - which is strange because it's in @EXPORT, and the function is coded in the xs file. The pgqinf() function doesn't do much, so I commented it out of the test files .... which results in the "Procedure entry point ..." error presenting itself again.
That's about as far as I got with that approach.
For anyone wanting to build the PGPLOT module with MSVC++ 6.0, then Karl Glazebrook's page at http://www.pha.jhu.edu/~kgb/PGPLOT-Win32/ is probably a good place to start. There's a link to precompiled static pgplot libs among other stuff. Given that Karl is also the author of the perl module, it's surprising that he doesn't provide Win32 binaries for that perl module ... or maybe he does, and I just failed to find them.
Maybe there's something there that helps ... maybe not :-)
Cheers, Rob
_______________________________________________ Perl-Win32-Users mailing list [EMAIL PROTECTED] To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
