----- Original Message ----- From: "Sisyphus" <[email protected]>
You could then maybe see if this approach works and it would help in tracking down your test failures.Definitely - hopefully I can find time to get onto this today.
Yes - everything seems fine with EU::FC, though I've not yet tracked down the source of the DynaLoader failures.
As soon as I had installed EU::FC I modified my ActivePerl alterations to use the libperl512.a that EU::FC had built and installed. I was hoping that would be sufficient to eradicate the failures I had been getting, but no such luck - still those same damned DynaLoader errors. However, as soon as I set PERL5OPT to -MConfig_w64 everything goes fine. I'm hard pressed to understand how there's any siginificant difference between the 2 approaches (now that they're both linking to the same libperl512.a).
Iff you're interested && iff you have the time, I'm curious to know what happens if you:
1) unset PERL5OPT (so that EU::FC does not get involved) 2) rename lib/ActivePerl/Config.pm to Config.pm_orig (or somesuch)3) install the attached Config.pm_new in lib/ActivePerl and rename it to Config.pm 4) then try to build Inline. (Maybe first check that 'perl-V:make', 'perl -V:libperl' and perl -V:cc' are reporting correct values.)
You should then see the same failures as I've been getting (and that would ease my mind a little as to my own sanity :-)
The first such failure should occur when the C/t/01syntax.t script is run. (No need for you to make any alterations to your MM_Win32.pm.)After you've finished, just reset PERL5OPT and restore the original lib/ActivePerl/Config.pm - and you're back to where you started from.
Also, if you (or anyone else) have any ideas as to why one approach (EU::FC) is working for me, but the other isn't, I'd be interested to hear them.
Cheers,Rob
Config.pm_new
Description: Binary data
_______________________________________________ Perl-Win32-Users mailing list [email protected] To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
