As Encode 1.00 release getting close, I wondered something like Encode
needs and deservers a technology demonstrator. So I have written
piconv, iconv written in perl. See how simple it is written yet more
flexible than original iconv.
It will be under bin/ subdirectory but by default it won't be
installed (or jhi will whip me for tainting perl dist). To enable the
installation (in yet-to-come 1.00; Autrijus, where are you? I am
waiting for your opinion on ISO-2022-CN), say
breadperl Makefile.PL USE_SCRIPTS
from the prompt. This script should run even vanilla 5.7.3.
Dan the Encode Maintainer
#!/usr/bin/perl
# $Id$
#
use 5.7.3;
use strict;
use Encode ;
use Encode::Alias;
my %Scheme = map {$_ => 1} qw(from_to decode_encode perlio);
use Getopt::Std;
my %Opt; getopts("DS:lf:t:s:", \%Opt);
$Opt{l} and list_encodings();
my $locale = $ENV{LC_CTYPE} || $ENV{LC_ALL} || $ENV{LANG};
my $from = $Opt{f} || $locale or help("from_encoding unspecified");
my $to = $Opt{t} || $locale or help("to_encoding unspecified");
$Opt{s} and Encode::from_to($Opt{s}, $from, $to) and print $Opt{s} and
exit;
my $scheme = exists $Scheme{$Opt{S}} ? $Opt{S} : 'from_to';
if ($Opt{D}){
my $cfrom = Encode->getEncoding($from)->name;
my $cto = Encode->getEncoding($to)->name;
print STDERR <<"EOT";
Scheme: $scheme
From: $from => $cfrom
To: $to => $cto
EOT
}
# default
if ($scheme eq 'from_to'){
while(<>){
Encode::from_to($_, $from, $to); print;
};
# step-by-step
}elsif ($scheme eq 'decode_encode'){
while(<>){
my $decoded = decode($from, $_);
my $encoded = encode($to, $decoded);
print $encoded;
};
# NI-S favorite
}elsif ($scheme eq 'perlio'){
binmode(STDIN, ":encoding($from)");
binmode(STDOUT, ":encoding($to)");
while(<>){ print; }
}else{ # won't reach
die "unknown scheme: $scheme";
}
sub list_encodings{
print STDERR join("\n", Encode->encodings(":all")), "\n";
exit;
}
sub help{
my $message = shift;
use File::Basename;
my $name = basename($0);
$message and print STDERR "$name error: $message\n";
print STDERR <<"EOT";
$name [-f from_encoding] [-t to_encoding] [-s string] [files...]
$name -l
-l lists all available encodings.
-f from_encoding When omitted, the current locale will be used.
-t to_encoding When omitted, the current locale will be used.
-s string "string" will be converted instead of STDIN.
EOT
exit;
}
__END__
=head1 NAME
piconv -- iconv(1), reinvented in perl
=head1 SYNOPSIS
piconv [-f from_encoding] [-t to_encoding] [-s string] [files...]
piconv -l
=head1 DESCRIPTION
B<piconv> is perl version of F<iconv>, a character encoding converter
widely availabe for various unixen today. This script was primarily
a technology demostrator for Perl 5.8.0, you can use piconv in the
place of iconv for virtually any cases.
piconv converts character encoding of either STDIN or files specified
in the argument and prints out to STDOUT.
Here are list of options.
=over 4
=item -f from_encoding
Specifies the encoding you are converting from. Unlike F<iconv>,
this option can be ommited. In such cases the current locale is used.
=item -t to_encoding
Specifies the encoding you are converting to. Unlike F<iconv>,
this option can be ommited. In such cases the current locale is used.
Therefore when both -f and -t are omitted, F<piconv> just acts like
F<cat>.
=item -s I<string>
uses I<string> instead of file for the source of text. Same as F<iconv>.
=item -l
Lists all available encodings to STDERR. This feature is missing from
F<iconv>.
=item -D
Invokes debugging mode. primarily for Encode hackers.
=item -S scheme
Selects which scheme is to be used for conversion. Available schemes
are as follows;
=over 4
=item from_to
Uses Encode::from_to for conversion. This is the default.
=item decode_encode
Input strings are decode()ed then encode()ed. A straight step-by-step
implementation.
=item perlio
The new perlIO layer is used. NI-S favorite.
=back
Like I<-D> option, this is also for Encode hackers.
=back
=head1 SEE ALSO
L<iconv(1)>
L<locale(3)>
L<Encode>
L<PerlIO>
=cut