File Coverage

blib/lib/JavaScript/Packer.pm

Criterion	Covered	Total	%
statement	338	359	94.1
branch	104	130	80.0
condition	29	33	87.8
subroutine	25	25	100.0
pod	1	3	33.3
total	497	550	90.3

package JavaScript::Packer;

use 5.008009;
use warnings;
use strict;
use Carp;
use Regexp::RegGrp;

# =========================================================================== #

our $VERSION = "2.10";

our @BOOLEAN_ACCESSORS = ( 'no_compress_comment', 'remove_copyright' );

our @COPYRIGHT_ACCESSORS = ( 'copyright', 'copyright_comment' );

our @COMPRESS_OPTS = ( 'clean', 'obfuscate', 'shrink', 'best' );
our $DEFAULT_COMPRESS = 'clean';

our $PACKER_COMMENT    = '\/\*\s*JavaScript::Packer\s*(\w+)\s*\*\/';
our $COPYRIGHT_COMMENT = '\/\*((?>[^\*]|\*[^\/])*copyright(?>[^\*]|\*[^\/])*)\*\/';

our $RESTORE_PATTERN     = qr~\x01(\d+)\x01~;
our $RESTORE_REPLACEMENT = "\x01%d\x01";

our $SHRINK_VARS = {
    BLOCK         => qr/(((catch|do|if|while|with|function)\b[^~{};]*(\(\s*[^{};]*\s*\))\s*)?(\{[^{}]*\}))/,    # function ( arg ) { ... }
    ENCODED_BLOCK => qr/~#?(\d+)~/,
    CALLER     => qr/((?>[a-zA-Z0-9_\x24\.]+)\s*\([^\(\)]*\))(?=[,\)])/,    # do_something( arg1, arg2 ) as argument of another function call
    BRACKETS   => qr/\{[^{}]*\}|\[[^\[\]]*\]|\([^\(\)]*\)|~[^~]+~/,
    IDENTIFIER => qr~[a-zA-Z_\x24][a-zA-Z_0-9\\x24]*~,
    SCOPED     => qr/~#(\d+)~/,
    VARS       => qr~\b(?:var|function)\s+((?>[a-zA-Z0-9_\x24]+))~,         # var x, funktion blah
    PREFIX     => qr~\x02~,
    SHRUNK     => qr~\x02\d+\b~
};

our $BASE62_VARS = {
    WORDS    => qr~(\b[0-9a-zA-Z]\b|(?>[a-zA-Z0-9_]{2,}))~,
    ENCODE10 => 'String',
    ENCODE36 => 'function(c){return c.toString(36)}',
    ENCODE62 => q~function(c){return(c<62?'':e(parseInt(c/62)))+((c=c%62)>35?String.fromCharCode(c+29):c.toString(36))}~,
    UNPACK =>
        q~eval(function(p,a,c,k,e,r){e=%s;if('0'.replace(0,e)==0){while(c--)r[e(c)]=k[c];k=[function(e){return r[e]||e}];e=function(){return'%s'};c=1};while(c--)if(k[c])p=p.replace(new RegExp('\\\\b'+e(c)+'\\\\b','g'),k[c]);return p}('%s',%s,%d,'%s'.split('|'),0,{}))~
};

our $DICTIONARY = {
    STRING1     => qr~"(?>(?:(?>[^"\\]+)|\\.|\\")*)"~,
    STRING2     => qr~'(?>(?:(?>[^'\\]+)|\\.|\\')*)'~,
    STRING3     => qr~`(?>(?:(?>[^`\\]+)|\\.|\\`)*)`~,
    REGEXP      => qr~\/(\\[\/\\]|[^*\/])(\\.|[^\/\n\\])*\/[gim]*~,
    OPERATOR    => qr'return|typeof|[\[(\^=,{}:;&|!*?]',
    CONDITIONAL => qr~\/\*\@\w*|\w*\@\*\/|\/\/\@\w*|\@(?>\w+)~,

    # single line comments
    COMMENT1    => qr~(:?)\/\/([\@#])?([^\n]*)?\n~,

    # multline comments
    COMMENT2    => qr~\/\*[^*]*\*+(?:[^\/][^*]*\*+)*\/~
};

our $DATA = [
    { regexp => $DICTIONARY->{STRING1} },
    { regexp => $DICTIONARY->{STRING2} },
    { regexp => $DICTIONARY->{STRING3} },
    { regexp => $DICTIONARY->{CONDITIONAL} },
    {
        regexp      => '(' . $DICTIONARY->{OPERATOR} . ')\s*(' . $DICTIONARY->{REGEXP} . ')',
        replacement => sub {
            return sprintf( "%s%s", $_[0]->{submatches}->[0], $_[0]->{submatches}->[1] );
        },
    },
];

our $COMMENTS = [
    {
        regexp      => ';;;[^\n]*\n',
        replacement => ''
    },
    { regexp => $DICTIONARY->{COMMENT1} . '\s*(' . $DICTIONARY->{REGEXP} . ')?', },
    { regexp => '(' . $DICTIONARY->{COMMENT2} . ')\s*(' . $DICTIONARY->{REGEXP} . ')?' }
];

our $CLEAN = [
    {
        regexp      => '\(\s*([^;)]*)\s*;\s*([^;)]*)\s*;\s*([^;)]*)\)',
        replacement => sub { return sprintf( "(%s;%s;%s)", @{ $_[0]->{submatches} } ); }
    },
    { regexp => 'throw[^};]+[};]' },
    {
        regexp      => ';+\s*([};])',
        replacement => sub { return $_[0]->{submatches}->[0]; }
    }
];

our $WHITESPACE = [
    { regexp => '\/\/@[^\n]*\n' },
    {
        regexp      => '@\s+\b',
        replacement => '@ '
    },
    {
        regexp      => '\b\s+@',
        replacement => ' @'
    },
    {
        regexp      => '(\d)\s+(\.\s*[a-z\x24_\[(])',
        replacement => sub { return sprintf( "%s %s", @{ $_[0]->{submatches} } ); }
    },
    {
        regexp      => '([+-])\s+([+-])',
        replacement => sub { return sprintf( "%s %s", @{ $_[0]->{submatches} } ); }
    },
    {
        regexp      => '(?>\s+)(\x24)(?>\s+)',
        replacement => sub { return sprintf( " %s ", $_[0]->{submatches}->[0] ); }
    },
    {
        regexp      => '(\x24)(?>\s+)(?!=)',
        replacement => sub { return sprintf( "%s ", $_[0]->{submatches}->[0] ); }
    },
    {
        regexp      => '(?\s+)(\x24)',
        replacement => sub { return sprintf( " %s", $_[0]->{submatches}->[0] ); }
    },
    {
        regexp      => '\b\s+\b',
        replacement => ' '
    },
    {
        regexp      => '\s+',
        replacement => ''
    }
];

our $TRIM = [
    {
        regexp      => '(\d)(?:\|\d)+\|(\d)',
        replacement => sub { return sprintf( "%d-%d", $_[0]->{submatches}->[0] || 0, $_[0]->{submatches}->[1] || 0 ); }
    },
    {
        regexp      => '([a-z])(?:\|[a-z])+\|([a-z])',
        replacement => sub { return sprintf( "%s-%s", $_[0]->{submatches}->[0], $_[0]->{submatches}->[1] ); }
    },
    {
        regexp      => '([A-Z])(?:\|[A-Z])+\|([A-Z])',
        replacement => sub { return sprintf( "%s-%s", $_[0]->{submatches}->[0], $_[0]->{submatches}->[1] ); }
    },
    {
        regexp      => '\|',
        replacement => ''
    }
];

our @REGGRPS = ( 'comments', 'clean', 'whitespace', 'concat', 'trim', 'data_store', 'concat_store' );

# --------------------------------------------------------------------------- #

{
    no strict 'refs';

    foreach my $field ( @BOOLEAN_ACCESSORS ) {
        next if defined *{ __PACKAGE__ . '::' . $field }{CODE};

        *{ __PACKAGE__ . '::' . $field } = sub {
            my ( $self, $value ) = @_;

            $self->{ '_' . $field } = $value ? 1 : undef if ( defined( $value ) );

            return $self->{ '_' . $field };
        };
    }

    foreach my $field ( @COPYRIGHT_ACCESSORS ) {
        $field = '_' . $field if ( $field eq 'copyright_comment' );
        next if defined *{ __PACKAGE__ . '::' . $field }{CODE};

        *{ __PACKAGE__ . '::' . $field } = sub {
            my ( $self, $value ) = @_;

            if ( defined( $value ) and not ref( $value ) ) {
                $value =~ s/^\s*|\s*$//gs;
                $self->{ '_' . $field } = $value;
            }

            my $ret = '';

            if ( $self->{ '_' . $field } ) {
                $ret = '/* ' . $self->{ '_' . $field } . ' */' . "\n";
            }

            return $ret;
        };
    }

    foreach my $reggrp ( @REGGRPS ) {
        next if defined *{ __PACKAGE__ . '::reggrp_' . $reggrp }{CODE};

        *{ __PACKAGE__ . '::reggrp_' . $reggrp } = sub {
            my ( $self ) = shift;

            return $self->{ '_reggrp_' . $reggrp };
        };
    }
}

sub compress {
    my ( $self, $value ) = @_;

    if ( defined( $value ) ) {
        if ( grep( $value eq $_, @COMPRESS_OPTS ) ) {
            $self->{_compress} = $value;
        }
        elsif ( !$value ) {
            $self->{_compress} = undef;
        }
    }

    $self->{_compress} ||= $DEFAULT_COMPRESS;

    return $self->{_compress};
}

# these variables are used in the closures defined in the init function
# below - we have to use globals as using $self within the closures leads
# to a reference cycle and thus memory leak, and we can't scope them to
# the init method as they may change. they are set by the minify sub
our $reggrp_comments;
our $reggrp_clean;
our $reggrp_whitespace;

sub init {
    my $class = shift;
    my $self  = {};

    bless( $self, $class );

    @{ $self->{clean}->{reggrp_data} } = ( @$DATA, @$CLEAN );
    @{ $self->{whitespace}->{reggrp_data} } = ( @$DATA[ 0, 1, 2, 4 ], @$WHITESPACE );
    $self->{trim}->{reggrp_data} = $TRIM;

    @{ $self->{data_store}->{reggrp_data} } = map {
        {
            regexp      => $_->{regexp},
            store       => sub { return sprintf( "%s", $_[0]->{match} ); },
            replacement => sub {
                return sprintf( $RESTORE_REPLACEMENT, $_[0]->{store_index} );
            },
        }
    } @$DATA;

    $self->{data_store}->{reggrp_data}->[-1]->{replacement} = sub {
        return sprintf( "%s$RESTORE_REPLACEMENT", $_[0]->{submatches}->[0], $_[0]->{store_index} );
    };

    $self->{data_store}->{reggrp_data}->[-1]->{store} = sub {
        return $_[0]->{submatches}->[1];
    };

    @{ $self->{concat_store}->{reggrp_data} } = map {
        my $data = $_;
        {
            regexp => $data->{regexp},
            store  => sub {
                my ( $quote, $string, $rest ) = $_[0]->{match} =~ /^(['"`])(.*)(['"`])$/;
				return $_[0]->{match} if ! $rest;

                return $string;
            },
            replacement => sub {
                my ( $quote, $string, $rest ) = $_[0]->{match} =~ /^(['"`])(.*)(['"`])$/;
				return $_[0]->{match} if ! $rest;

                return sprintf( "%s$RESTORE_REPLACEMENT%s", $quote, $_[0]->{store_index}, $quote );
            },

        };
    } @$DATA[ 0, 1, 2 ];

    @{ $self->{concat}->{reggrp_data} } = map {
        my $quote           = $_;
        my $restore_pattern = $RESTORE_PATTERN;
        $restore_pattern =~ s/\(\\d\+\)/\\d\+/g;

        my $regexp = '(' . $quote . $restore_pattern . $quote . ')((?:\+' . $quote . $restore_pattern . $quote . ')+)';
        $regexp = qr/$regexp/;

        {
            regexp      => $regexp,
            replacement => sub {
                my $submatches = $_[0]->{submatches};
                my $ret        = $submatches->[0];

                my $next_str = '^\+(' . $quote . $restore_pattern . $quote . ')';

                while ( my ( $next ) = $submatches->[1] =~ /$next_str/ ) {
                    chop( $ret );
                    $ret .= substr( $next, 1 );
                    $submatches->[1] =~ s/$next_str//;
                }

                return $ret;
            },
        };
    } ( '"', '\'', '\`' );

    @{ $self->{comments}->{reggrp_data} } = ( @$DATA[ 0, 1, 2, 4 ], @$COMMENTS );

    # single line comment
    $self->{comments}->{reggrp_data}->[-2]->{replacement} = sub {
        my $submatches = $_[0]->{submatches};

        if ( $submatches->[0] eq ':' ) {
            return sprintf( "://%s", $submatches->[2] );
        }
        elsif ( $submatches->[1] eq '#' ) {
            my $cmnt = sprintf( "%s//%s%s__NEW_LINE__", @{$submatches}[0 .. 2] );
            return $cmnt;
        }
        elsif ( $submatches->[1] eq '@' ) {
            $reggrp_comments->exec( \$submatches->[2] );
            $reggrp_clean->exec( \$submatches->[2] );
            $reggrp_whitespace->exec( \$submatches->[2] );

            return sprintf( "%s//%s%s\n%s", @{$submatches}[0 .. 3] );
        }
        return sprintf( "\n%s", $submatches->[3] );
    };

    # multi line comments
    $self->{comments}->{reggrp_data}->[-1]->{replacement} = sub {
        my $submatches = $_[0]->{submatches};
        if ( $submatches->[0] =~ /^\/\*\@(.*)\@\*\/$/sm ) {
            my $cmnt = $1;

            $reggrp_comments->exec( \$cmnt );
            $reggrp_clean->exec( \$cmnt );
            $reggrp_whitespace->exec( \$cmnt );

            return sprintf( '/*@%s@*/ %s', $cmnt, $submatches->[1] );
        }
        return sprintf( " %s", $submatches->[1] );
    };

    foreach my $reggrp ( @REGGRPS ) {
        my $reggrp_args = { reggrp => $self->{$reggrp}->{reggrp_data} };

        $reggrp_args->{restore_pattern} = $RESTORE_PATTERN if ( $reggrp eq 'data_store' or $reggrp eq 'concat_store' );

        $self->{ '_reggrp_' . $reggrp } = Regexp::RegGrp->new( $reggrp_args );
    }

    $self->{block_data} = [];

    return $self;
}

sub minify {
    my ( $self, $input, $opts );

    unless (ref( $_[0] )
        and ref( $_[0] ) eq __PACKAGE__ )
    {
        $self = __PACKAGE__->init();

        shift( @_ ) unless ( ref( $_[0] ) );

        ( $input, $opts ) = @_;
    }
    else {
        ( $self, $input, $opts ) = @_;
    }

    if ( ref( $input ) ne 'SCALAR' ) {
        carp( 'First argument must be a scalarref!' );
        return undef;
    }

    my $javascript = \'';
    my $cont       = 'void';

    if ( defined( wantarray ) ) {
        my $tmp_input = ref( $input ) ? ${$input} : $input;

        $javascript = \$tmp_input;
        $cont       = 'scalar';
    }
    else {
        $javascript = ref( $input ) ? $input : \$input;
    }

    if ( ref( $opts ) eq 'HASH' ) {
        foreach my $field ( @BOOLEAN_ACCESSORS ) {
            $self->$field( $opts->{$field} ) if ( defined( $opts->{$field} ) );
        }

        foreach my $field ( 'compress', 'copyright' ) {
            $self->$field( $opts->{$field} ) if ( defined( $opts->{$field} ) );
        }
    }

	# (re)initialize variables used in the closures
	$reggrp_comments = $self->reggrp_comments;
	$reggrp_clean = $self->reggrp_clean;
	$reggrp_whitespace = $self->reggrp_whitespace;

    my $copyright_comment = '';

    if ( ${$javascript} =~ /$COPYRIGHT_COMMENT/ism ) {
        $copyright_comment = $1;
    }

    # Resets copyright_comment() if there is no copyright comment
    $self->_copyright_comment( $copyright_comment );

    if ( not $self->no_compress_comment() and ${$javascript} =~ /$PACKER_COMMENT/ ) {
        my $compress = $1;
        if ( $compress eq '_no_compress_' ) {
            return ${$javascript} if ( $cont eq 'scalar' );
            return;
        }

        $self->compress( $compress );
    }

    ${$javascript} =~ s/\r//gsm;
    ${$javascript} .= "\n";
    $self->reggrp_comments()->exec( $javascript );
    $self->reggrp_clean()->exec( $javascript );
    $self->reggrp_whitespace()->exec( $javascript );
    $self->reggrp_concat_store()->exec( $javascript );
    $self->reggrp_concat()->exec( $javascript );
    $self->reggrp_concat_store()->restore_stored( $javascript );

    if ( $self->compress() ne 'clean' ) {
        $self->reggrp_data_store()->exec( $javascript );

        while ( ${$javascript} =~ /$SHRINK_VARS->{BLOCK}/ ) {
            ${$javascript} =~ s/$SHRINK_VARS->{BLOCK}/$self->_store_block_data( $1 )/egsm;
        }

        $self->_restore_data( $javascript, 'block_data', $SHRINK_VARS->{ENCODED_BLOCK} );

        my %shrunk_vars = map { $_ => 1 } ( ${$javascript} =~ /$SHRINK_VARS->{SHRUNK}/g );

        my $cnt = 0;
        foreach my $shrunk_var ( sort keys( %shrunk_vars ) ) {
            my $short_id;
            do {
                $short_id = $self->_encode52( $cnt++ );
            } while ( ${$javascript} =~ /[^a-zA-Z0-9_\\x24\.]\Q$short_id\E[^a-zA-Z0-9_\\x24:]/ );

            ${$javascript} =~ s/$shrunk_var/$short_id/g;
        }

        $self->reggrp_data_store()->restore_stored( $javascript );

        $self->{block_data} = [];
    }

    if ( $self->compress() eq 'obfuscate' or $self->compress() eq 'best' ) {
        my $words = {};

        my @words = ${$javascript} =~ /$BASE62_VARS->{WORDS}/g;

        my $idx = 0;

        foreach ( @words ) {
            $words->{$_}->{count}++;
        }

        WORD: foreach my $word ( sort { $words->{$b}->{count} <=> $words->{$a}->{count} } sort keys( %{$words} ) ) {

            if ( exists( $words->{$word}->{encoded} ) and $words->{$word}->{encoded} eq $word ) {
                next WORD;
            }

            my $encoded = $self->_encode62( $idx );

            if ( exists( $words->{$encoded} ) ) {
                my $next = 0;
                if ( exists( $words->{$encoded}->{encoded} ) ) {
                    $words->{$word}->{encoded} = $words->{$encoded}->{encoded};
                    $words->{$word}->{index}   = $words->{$encoded}->{index};
                    $words->{$word}->{minus}   = length( $word ) - length( $words->{$word}->{encoded} );
                    $next                      = 1;
                }
                $words->{$encoded}->{encoded} = $encoded;
                $words->{$encoded}->{index}   = $idx;
                $words->{$encoded}->{minus}   = 0;
                $idx++;
                next WORD if ( $next );
                redo WORD;
            }

            $words->{$word}->{encoded} = $encoded;
            $words->{$word}->{index}   = $idx;
            $words->{$word}->{minus}   = length( $word ) - length( $encoded );

            $idx++;
        }

        my $packed_length = length( ${$javascript} );

        my ( @pk, @pattern ) = ( (), () );

        foreach ( sort { $words->{$a}->{index} <=> $words->{$b}->{index} } sort keys( %{$words} ) ) {
            $packed_length -= ( $words->{$_}->{count} * $words->{$_}->{minus} );

            if ( $words->{$_}->{encoded} ne $_ ) {
                push( @pk,      $_ );
                push( @pattern, $words->{$_}->{encoded} );
            }
            else {
                push( @pk,      '' );
                push( @pattern, '' );
            }
        }

        my $size = scalar( @pattern );

        splice( @pattern, 62 ) if ( scalar( @pattern ) > 62 );

        my $pd = join( '|', @pattern );

        $self->reggrp_trim()->exec( \$pd );

        unless ( $pd ) {
            $pd = '^$';
        }
        else {
            $pd = '[' . $pd . ']';

            if ( $size > 62 ) {
                $pd = '(' . $pd . '|';

                my $enc = $self->_encode62( $size );

                my ( $c ) = $enc =~ /(^.)/;
                my $ord = ord( $c );

                my $mul = length( $enc ) - 1;

                my $is62 = 0;

                if ( $ord >= 65 ) {
                    if ( $c eq 'Z' ) {
                        $mul += 1;
                        $is62 = 1;
                    }
                    else {
                        $pd .= '[0-9a';
                        if ( $ord > 97 ) {
                            $pd .= '-' . $c;
                        }
                        elsif ( $ord > 65 ) {
                            $pd .= '-zA-' . $c;
                        }
                        elsif ( $ord == 65 ) {
                            $pd .= '-zA';
                        }
                        $pd .= ']';
                    }
                }
                elsif ( $ord == 57 ) {
                    $pd .= '[0-9]';
                }
                elsif ( $ord == 50 ) {
                    $pd .= '[12]';
                }
                elsif ( $ord == 49 ) {
                    $pd .= '1';
                }
                else {
                    $pd .= '[0-' . ( $ord - 48 ) . ']';
                }

                $pd .= '[0-9a-zA-Z]' . ( ( $mul > 1 ) ? '{' . $mul . '}' : '' );

                $mul-- if ( $is62 );

                if ( $mul > 1 ) {
                    for ( my $i = $mul; $i >= 2; $i-- ) {
                        $pd .= '|[0-9a-zA-Z]{' . $i . '}';
                    }
                }

                $pd .= ')';
            }
        }
        $packed_length += length( $pd );

        my $pk = join( '|', @pk );
        $pk =~ s/(?>\|+)$//;
        $packed_length += length( $pk );

        my $pc = length( $pk ) ? ( ( $pk =~ tr/|/|/ ) + 1 ) : 0;
        $packed_length += length( $pc );

        my $pa = '[]';
        $packed_length += length( $pa );

        my $pe = $BASE62_VARS->{ 'ENCODE' . ( $pc > 10 ? $pc > 36 ? 62 : 36 : 10 ) };
        $packed_length += length( $pe );

        $packed_length += length( $BASE62_VARS->{UNPACK} );
        $packed_length -= ( $BASE62_VARS->{UNPACK} =~ s/(%s|%d)/$1/g ) * 2;

        my ( @length_matches ) = ${$javascript} =~ s/((?>[\r\n]+))/$1/g;
        foreach ( @length_matches ) {
            $packed_length -= length( $_ ) - 3;
        }

        $packed_length += ${$javascript} =~ tr/\\\'/\\\'/;

        if ( $self->compress() eq 'obfuscate' or $packed_length <= length( ${$javascript} ) ) {

            ${$javascript} =~ s/$BASE62_VARS->{WORDS}/sprintf( "%s", $words->{$1}->{encoded} )/eg;

            ${$javascript} =~ s/([\\'])/\\$1/g;
            ${$javascript} =~ s/[\r\n]+/\\n/g;

            my $pp = ${$javascript};

            ${$javascript} = sprintf( $BASE62_VARS->{UNPACK}, $pe, $pd, $pp, $pa, $pc, $pk );
        }

    }

    if ( not $self->remove_copyright() ) {
        ${$javascript} = ( $self->copyright() || $self->_copyright_comment() ) . ${$javascript};
    }

    # GH #9 bodge for sourceMappingURL
    ${$javascript} =~ s/__NEW_LINE__/\n/xsmg;
    ${$javascript} =~ s!//#sourceMappingURL!//# sourceMappingURL!g;
    chomp( ${$javascript} );

    return ${$javascript} if ( $cont eq 'scalar' );
}

sub _restore_data {
    my ( $self, $string_ref, $data_name, $pattern ) = @_;

    while ( ${$string_ref} =~ /$pattern/ ) {
        ${$string_ref} =~ s/$pattern/$self->{$data_name}->[$1]/egsm;
    }
}

sub _store_block_data {
    my ( $self, $match ) = @_;

    my ( undef, $prefix, $blocktype, $args, $block ) = $match =~ /$SHRINK_VARS->{BLOCK}/;

    $prefix    ||= '';
    $blocktype ||= '';
    $args      ||= '';
    my $replacement = '';
    if ( $blocktype eq 'function' ) {

        $self->_restore_data( \$block, 'block_data', $SHRINK_VARS->{SCOPED} );

        $args =~ s/\s*//g;

        $block = $args . $block;
        $prefix =~ s/$SHRINK_VARS->{BRACKETS}//;

        $args =~ s/^\(|\)$//g;

        while ( $args =~ /$SHRINK_VARS->{CALLER}/ ) {
            $args =~ s/$SHRINK_VARS->{CALLER}//gsm;
        }

        my @vars = grep( $_, split( /\s*,\s*/, $args ) );
        my $do_shrink = grep( $_ eq '_no_shrink_', @vars ) ? 0 : 1;

        my %block_vars = ();
        if ( $do_shrink ) {
            %block_vars = map { $_ => 1 } ( $block =~ /$SHRINK_VARS->{VARS}/g ), grep( $_ ne '$super', @vars );
        }

        $self->_restore_data( \$block, 'block_data', $SHRINK_VARS->{ENCODED_BLOCK} );

        if ( $do_shrink ) {

            my $cnt = 0;
            foreach my $block_var ( sort keys( %block_vars ) ) {
                if ( length( $block_var ) ) {
                    while ( $block =~ /$SHRINK_VARS->{PREFIX}\Q$cnt\E\b/ ) {
                        $cnt++;
                    }

                    while ( $block =~ /[^a-zA-Z0-9_\\x24\.]\Q$block_var\E[^a-zA-Z0-9_\\x24:]/ ) {
                        $block =~ s/([^a-zA-Z0-9_\\x24\.])\Q$block_var\E([^a-zA-Z0-9_\\x24:])/sprintf( "%s\x02%d%s", $1, $cnt, $2 )/eg;
                    }

                    $block =~ s/([^{,a-zA-Z0-9_\\x24\.])\Q$block_var\E:/sprintf( "%s\x02%d:", $1, $cnt )/eg;

                    $cnt++;
                }
            }
        }
        $replacement = sprintf( "%s~%d~", $prefix, scalar( @{ $self->{block_data} } ) );

        push( @{ $self->{block_data} }, $block );
    }
    else {
        $replacement = sprintf( "~#%d~", scalar( @{ $self->{block_data} } ) );

        push( @{ $self->{block_data} }, $prefix . $block );
    }

    return $replacement;
}

sub _encode52 {
    my ( $self, $c ) = @_;

    my $m = $c % 52;

    my $ret = $m > 25 ? chr( $m + 39 ) : chr( $m + 97 );

    if ( $c >= 52 ) {
        $ret = $self->_encode52( int( $c / 52 ) ) . $ret;
    }

    $ret = substr( $ret, 1 ) . '0' if ( $ret =~ /^(do|if|in)$/ );

    return $ret;
}

sub _encode62 {
    my ( $self, $c ) = @_;

    my $m = $c % 62;

    my $ret = $m > 35 ? chr( $m + 29 ) : $m > 9 ? chr( $m + 87 ) : $m;

    if ( $c >= 62 ) {
        $ret = $self->_encode62( int( $c / 62 ) ) . $ret;
    }

    return $ret;
}

1;

__END__