Some time (2 weeks?) ago, Michael G Schwern wrote:
> Movement on the MakeMaker front. I've integrated all the changes p5p
> made into the snapshot along with a bunch of other fixes. Give it a
> shot.
> The more reports I get, the faster I can get 6.22 out.
>
> Some of the patched folks have sent in over the past few months
> haven't been applied, there's been too much change in this release
> already, but they've
> been stored in RT.
> Let me know if I missed anything.
Between perl-5.6.2 and 5.8.1 an exception for dmake was made
in the writing of an implicit xs_o rule in MM_Win95.pm.
Between ExtUtils-MakeMaker-6.21 and 6.21_03 (?), this exception
to the implicit xs_o rule was withdrawn from MM_Win95.pm.
This is preventing bleadperl building on Win95.
--- perl-5.8.x/lib/ExtUtils/MM_Win95.pm 2004-02-27 19:47:46.000000000 +0800
+++ perl-current/lib/ExtUtils/MM_Win95.pm 2004-04-04 21:32:36.000000000 +0800
@@ -1,7 +1,7 @@
package ExtUtils::MM_Win95;
use vars qw($VERSION @ISA);
-$VERSION = 0.03;
+$VERSION = 0.03_01;
require ExtUtils::MM_Win32;
@ISA = qw(ExtUtils::MM_Win32);
@@ -125,8 +125,6 @@
sub xs_o {
my($self) = shift;
return '' unless $self->needs_linking();
- # Having to choose between .xs -> .c -> .o and .xs -> .o confuses dmake.
- return '' if $DMAKE;
'
Meanwhile, the xs_o method in MM_Win32.pm is just a stub. No
implicit xs_o rule is written into non-Win98 Win32 makefiles.
See some discussion at
http://www.mail-archive.com/[EMAIL PROTECTED]/msg01519.html
--
Greg Matheson, Taiwan
