=================================================================== RCS file: /cvs/openbsd/errata_scraper/errata_scraper.pl,v retrieving revision 1.2 retrieving revision 1.4 diff -u -r1.2 -r1.4 --- openbsd/errata_scraper/errata_scraper.pl 2011/03/21 17:28:15 1.2 +++ openbsd/errata_scraper/errata_scraper.pl 2011/05/29 02:04:30 1.4 @@ -1,5 +1,5 @@ #!/usr/bin/perl -# $AFresh1: errata_scraper.pl,v 1.1 2011/03/21 16:26:58 andrew Exp $ +# $AFresh1: errata_scraper.pl,v 1.3 2011/03/23 18:46:16 andrew Exp $ ######################################################################## # Copyright (c) 2011 Andrew Fresh # @@ -19,18 +19,26 @@ use warnings; use Mojo::UserAgent; -use Mojo::ByteStream 'b'; -Mojo::UserAgent->new->get('http://www.openbsd.org/errata48.html') - ->res->dom('li')->each( - sub { - my $e = shift; +my $base_uri = 'http://www.openbsd.org/'; - my $patch = $e->at('a[href$=".patch"]')->attrs->{href}; - my $title = b( $e->at('strong')->replace('')->all_text )->trim; - my $arch = b( $e->at('i')->replace('')->all_text )->trim; - my $descr = b( $e->all_text )->trim; +my $ua = Mojo::UserAgent->new; + +my $ls = $ua->get( $base_uri . 'errata.html' )->res->dom('a[href^="errata"]'); + +foreach my $l ( @{$ls}[ -2, -1 ] ) { + print 'Errata for OpenBSD ', $l->text, "\n"; + foreach my $e ( + reverse @{ $ua->get( $base_uri . $l->attrs->{'href'} )->res->dom('li') + } ) + { + my $patch = $e->at('a[href$=".patch"]')->replace('')->{href}; + my $title = $e->at('strong')->replace('')->all_text; + my $arch = $e->at('i')->replace('')->all_text; + my $descr = $e->all_text; $descr =~ s/\s+/ /gs; + $descr =~ s/\s(\.(?:\s|$))/$1/gs; + $descr =~ s/\.+$/./gs; print 'Title: ', $title, "\n"; print 'Arch: ', $arch, "\n"; @@ -38,4 +46,4 @@ print 'Descr: ', $descr, "\n"; print "\n"; } -); +}