On 2006-01-18, at 07:46:56 -0500, John E. Malmberg wrote:

> ext/Devel/PPPort/t/ppphtest - not investigated yet.

I'd bet that any failures in this test were introduced by the
changes made between version 3.06 and 3.06_01. The changes
between 3.06_01 and 3.07 that affect this test are rather
trivial, but the earlier changes include some VMS specific parts.

Nicholas, this may mean that the version integrated into maint
will show the same problems on VMS.

Here are two parts that changed between 3.06 and 3.06_01:

   if ($ENV{'PERL_CORE'}) {
     if (-d '../../lib') {
  -    $inc = $^O eq 'VMS' ? '-"I../../lib"' : '-I../../lib';
  +    if ($isVMS) {
  +      $inc = '"-I../../lib"';
  +    }
  +    elsif ($isMAC) {
  +      $inc = '-I:::lib';
  +    }
  +    else {
  +      $inc = '-I../../lib';
  +    }
       unshift @INC, '../../lib';
     }
   }

I think this change isn't the cause of the new failures.
The other part looks a bit more suspicious:

   sub ppport
   {
  -  my @args = @_;
  -  print "# *** running $perl $inc ppport.h @args ***\n";
  -  my $out = join '', `$perl $inc ppport.h @args`;
  -  my $copy = $out;
  -  $copy =~ s/^/# | /mg;
  -  print "$copy\n";
  -  return $out;
  +  my @args = ('ppport.h', @_);
  +  unshift @args, $inc if $inc;
  +  my $run = $perl =~ m/\s/ ? qq("$perl") : $perl;
  +  $run .= ' -MMac::err=unix' if $isMAC;
  +  for (@args) {
  +    $_ = qq("$_") if $isVMS && /^[^"]/;
  +    $run .= " $_";
  +  }
  +  print "# *** running $run ***\n";
  +  $run .= ' 2>&1' unless $isMAC;
  +  my @out = `$run`;
  +  my $out = join '', @out;
  +  comment($out);
  +  return wantarray ? @out : $out;
   }

This is the code that executes the ppport.h script. Most
of the tricks are from the _create_runperl routine in the
t/test.pl script, but maybe I messed something up.

I spent a significant part of this evening trying to
reproduce the failure on a testdrive system, but due to
my lack of the most basic VMS knowledge, I wasn't able
to a) make the system perl find its modules, b) make the
old perl I compiled earlier last year work and c) build
a new version of blead because of strange mms errors.

I'd really like to help in fixing this, but I think I will
need some input from a VMS expert.

Marcus

-- 
Newman's Discovery:
        Your best dreams may not come true; fortunately, neither will
        your worst dreams.

Reply via email to