--- trunk/Fast.pm 2007/11/06 20:06:07 24 +++ trunk/Fast.pm 2010/02/04 17:08:03 30 @@ -2,7 +2,7 @@ use strict; use Carp; -use Data::Dumper; +use Data::Dump qw/dump/; BEGIN { use Exporter (); @@ -124,6 +124,7 @@ # skip to next record my $o = substr($leader,0,5); + warn "# in record ", $self->{count}," record length isn't number but: ",dump($o),"\n" unless $o =~ m/^\d+$/; if ($o > 24) { seek($self->{fh},$o-24,1) if ($o); } else { @@ -383,6 +384,27 @@ 1; __END__ +=head1 UTF-8 ENCODING + +This module does nothing with encoding. But, since MARC format is byte +oriented even when using UTF-8 which has variable number of bytes for each +character, file is opened in binary mode. + +As a result, all scalars recturned to perl don't have utf-8 flag. Solution is +to use C and L to decode utf-8 encoding like this: + + use Encode; + + my $marc = new MARC::Fast( + marcdb => 'utf8.marc', + hash_filter => sub { + Encode::decode( 'utf-8', $_[0] ); + }, + ); + +This will affect C, but C will still return binary representation +since it doesn't support C. + =head1 AUTHOR Dobrica Pavlinusic