No such problem with the same module on my MinGW/dmake-built perl 5.8.8 (built from CPAN source). It compiles and tests fine. Apart from the fact that I'm using 'dmake' and you're (apparently) using 'nmake', I'm also using MinGW's port of gcc-3.4.2 (as opposed to your gcc-3.2.3). I also steadfastly refuse to use CPAN.pm (as I've got enough to do without having to come to terms with CPAN.pm bugs/features). However, I'm not inclined to think that any of those differences accounts for the problem ... could this be a PXPerl issue ? (PXPerl is one of the many things that I know absolutely nothing about.)

With ActiveState perl (build 817) I do, however, get exactly the same mind-numbing output as you reported ... but ... aaaarrgh ... this is one of those modules that wants to 'use inc::Module::Install' and there are problems with that wrt ActiveState perl. (ActiveState know about this - or at least Jan Dubois does. I spent some time digging around with this wrt PAR - not sure that I've fully recovered to the extent that I'm ready to re-visit the issue - see the thread "PAR problems" at http://www.mail-archive.com/par@perl.org/. The post at http://www.mail-archive.com/par@perl.org/msg02184.html is probably the most edifying of the posts within that thread - and should provide the fix. Let us know if it doesn't.)

Cheers,
Rob