Re: [gentoo-perl] g-cpan -u

[Yuval Yaari]

Michael Cummings wrote:

> I could try the old developer trick of "in the next version..." - but
> the short of it is, you're right, we cheat here. Instead of expending
> time on checking module versions (which seem to change numbering
> schemes between releases sometimes), we just check to see if we need
> to make an ebuild for it or if one already exists (after making sure
> we are in $OVERLAY/perl-gcpan of course, and not upgrading your 'real'
> tree) - then we just rely on portage to handle the -u for anything
> that has an ebuild greater than what's installed. Cheap, tawdry - sure
> beat writing version checkers (or adding deps for one)(or stealing
> more code from how cpan handles it, which I will show off in the next
> version with some new fangled dep checking).
>
Stealing & hacking code from CPAN:

#!/usr/bin/perl -w
use strict;
use CPAN;

my ($module) = CPAN::Shell->expand("Module", 'DBD::mysql');
my $latest = $module->cpan_version;
my $have = $module->inst_version;

# Skipping version_zeros part (which can also be easily stolen from CPAN.pm)

if (CPAN::Version->vgt($latest, $have)) {
    print "There's a newer version on CPAN\n";
    print "CPAN version: $latest\n";
    print "You have: $have\n";
} else {
    print "No newer version!\n";
}
__END__

Can't we rely on that if the CPAN shell relies on that?

  --Yuval
-- 
gentoo-perl@g.o mailing list