svn-log-github.pl

#!/usr/bin/perl
# http://www.simplisticcomplexity.com/2008/03/05/cleanly-migrate-your-subversion-repository-to-a-git-repository/
use strict;
use warnings;
use Web::Scraper;
use URI::Escape;
use URI;

$| = 1;
my %authors;

while (<>) {
    /^r\d+\s*\|\s*(\S+)/ or next;
    unless (exists $authors{$1}) {
        $authors{$1} = github_info($1);
    }
}

sub github_info {
    my $username = shift;

    my $scraper = scraper {
        process "#profile_name", name => 'TEXT';
        process ".email", email => [ 'HTML', \&unmangle_js ];
    };
    my $res = eval { $scraper->scrape(URI->new("http://github.com/$username")) };

    if ($res->{name} && $res->{email}) {
        print "$username = $res->{name} <$res->{email}>\n";
    }

    return "blah";
}

sub unmangle_js {
    /decodeURIComponent\(&#39;([%0-9a-f]+)/;
    my $code = uri_unescape($1);
    $code =~ /mailto:(.*?)"/ and return $1;
}