Re: Relative Module Path

[tadamec]

Why not combine P and M into a single unit? That way you're sure they won't be separated by a too-specific copy command.

Using your example, ~/P would become:

#!/usr/bin/perl -w
use strict;
use M;
my $m = M->new;
#
# End of main program
#
package M;
sub new{1};
1;

[pbeckingham]

Why not combine them? Because there are many modules, and many more programs that use them. The example shown was just that - an example. The real code is somewhat more complex.

[tadamec]

Ah, gotcha.

I use something like this for situations where I have a local "lib" directory. You could probably modify it somewhat:

#!/usr/bin/perl
use strict;                              # Always use strict.
my $BASEPATH;

#
# Include path magic.
#
BEGIN {
  use File::Basename;
  use File::Spec;
  my $PROGRAM = File::Spec->rel2abs( $0 );
  ( undef, $BASEPATH, undef ) = fileparse( $PROGRAM );
  $BASEPATH = File::Spec->catdir( $BASEPATH, "../" );
  my $LIBPATH = File::Spec->catdir( $BASEPATH, "lib" );
  eval " use lib '$LIBPATH'";
}

If you want to keep the module files in the same directory this BEGIN block should work (I haven't tested it):

BEGIN {
  use File::Basename;
  use File::Spec;
  my $PROGRAM = File::Spec->rel2abs( $0 );
  ( undef, $BASEPATH, undef ) = fileparse( $PROGRAM );
  eval " use lib '$BASEPATH'";
}