This file is indexed.

/usr/share/perl5/String/ToIdentifier/EN.pm is in libstring-toidentifier-en-perl 0.11-1.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
package String::ToIdentifier::EN;
BEGIN {
  $String::ToIdentifier::EN::AUTHORITY = 'cpan:RKITOVER';
}
{
  $String::ToIdentifier::EN::VERSION = '0.11';
}

use 5.008001;
use strict;
use warnings;
use Text::Unidecode 'unidecode';
use Lingua::EN::Inflect::Phrase 'to_PL';
use Unicode::UCD 'charinfo';
use namespace::clean;
use Exporter 'import';

=head1 NAME

String::ToIdentifier::EN - Convert Strings to English Program Identifiers

=head1 SYNOPSIS

    use utf8;
    use String::ToIdentifier::EN 'to_identifier';

    to_identifier 'foo-bar';             # fooDashBar
    to_identifier 'foo-bar', '_';        # foo_dash_bar
    to_identifier 'foo.bar', '_';        # foo_dot_bar
    to_identifier "foo\x{4EB0}bar";      # fooJingBar
    to_identifier "foo\x00bar";          # fooNullCharBar
    to_identifier "foo\x00\x00bar";      # foo2NullCharsBar
    to_identifier "foo\x00\x00bar", '_'; # foo_2_null_chars_bar

    {
        no utf8;
        to_identifier "foo\xFF\xFFbar.baz";      # foo_2_0xFF_BarDotBaz
        to_identifier "foo\xFF\xFFbar.baz", '_'; # foo_2_0xFF_bar_dot_baz
    }

=head1 DESCRIPTION

This module provides a utility method, L</to_identifier> for converting an
arbitrary string into a readable representation using the ASCII subset of C<\w>
for use as an identifier in a computer program. The intent is to make unique
identifier names from which the content of the original string can be easily
inferred by a human just by reading the identifier.

If you need the full set of C<\w> including Unicode, see
the subclass L<String::ToIdentifier::EN::Unicode>.

Currently, this process is one way only, and will likely remain this way.

The default is to create camelCase identifiers, or you may pass in a separator
char of your choice such as C<_>.

Binary char groups will be separated by C<_> even in camelCase identifiers to
make them easier to read, e.g.: C<foo_2_0xFF_Bar>.

=head1 EXPORT

Optionally exports the L</to_identifier> function.

=cut

our @EXPORT_OK = qw/to_identifier/;

=head1 SUBROUTINES

=cut

our %ASCII_MAP = (
    0x00 => ['null'],
    0x01 => ['start', 'of', 'heading'],
    0x02 => ['start', 'of', 'text'],
    0x03 => ['end', 'of', 'text'],
    0x04 => ['end', 'of', 'transmission'],
    0x05 => ['enquiry', 'char'],
    0x06 => ['ack'],
    0x07 => ['bell', 'char'],
    0x08 => ['backspace'],
    0x09 => ['tab', 'char'],
    0x0A => ['newline'],
    0x0B => ['vertical', 'tab'],
    0x0C => ['form', 'feed'],
    0x0D => ['carriage', 'return'],
    0x0E => ['shift', 'out'],
    0x0F => ['shift', 'in'],
    0x10 => ['data', 'link', 'escape'],
    0x11 => ['device', 'control1'],
    0x12 => ['device', 'control2'],
    0x13 => ['device', 'control3'],
    0x14 => ['device', 'control4'],
    0x15 => ['negative', 'ack'],
    0x16 => ['synchronous', 'idle'],
    0x17 => ['end', 'of', 'transmission', 'block'],
    0x18 => ['cancel', 'char'],
    0x19 => ['end', 'of', 'medium'],
    0x1A => ['substitute', 'char'],
    0x1B => ['escape', 'char'],
    0x1C => ['file', 'separator'],
    0x1D => ['group', 'separator'],
    0x1E => ['record', 'separator'],
    0x1F => ['unit', 'separator'],
    0x20 => ['space', 'char'],
    0x21 => ['exclamation', 'mark'],
    0x22 => ['double', 'quote'],
    0x23 => ['hash', 'mark'],
    0x24 => ['dollar', 'sign'],
    0x25 => ['percent', 'sign'],
    0x26 => ['ampersand'],
    0x27 => ['single', 'quote'],
    0x28 => ['left', 'paren'],
    0x29 => ['right', 'paren'],
    0x2A => ['asterisk'],
    0x2B => ['plus', 'sign'],
    0x2C => ['comma'],
    0x2D => ['dash'],
    0x2E => ['dot'],
    0x2F => ['slash'],
    0x3A => ['colon'],
    0x3B => ['semicolon'],
    0x3C => ['left', 'angle', 'bracket'],
    0x3D => ['equals', 'sign'],
    0x3E => ['right', 'angle', 'bracket'],
    0x3F => ['question', 'mark'],
    0x40 => ['at', 'sign'],
    0x5B => ['left', 'bracket'],
    0x5C => ['backslash'],
    0x5D => ['right', 'bracket'],
    0x5E => ['caret'],
    0x60 => ['backtick'],
    0x7B => ['left', 'brace'],
    0x7C => ['pipe', 'char'],
    0x7D => ['right', 'brace'],
    0x7E => ['tilde'],
    0x7F => ['delete', 'char'],
);

# fixup for perl <= 5.8.3
$ASCII_MAP{0} = ['null'];

=head2 to_identifier

Takes the string to be converted to an identifier, and optionally a separator
char such as C<_>. If a separator char is not provided, a camelCase identifier
will be returned.

=cut

sub to_identifier {
    return __PACKAGE__->string_to_identifier(@_);
}

# Override some pluralizations Lingua::EN::Inflect::Phrase gets wrong here, if
# needed.
sub _pluralize_phrase {
    my ($self, $phrase) = @_;

    return to_PL($phrase);
}

# for overriding in ::Unicode
sub _non_identifier_char {
    return qr/[^0-9a-zA-Z_]/;
}

=head1 METHODS

=head2 string_to_identifier

The class method version of L</to_identifier>, if you want to use the object
oriented interface.

=cut

sub string_to_identifier {
    my ($self, $str, $sep_char) = @_;

    my $is_utf8 = utf8::is_utf8($str);

    my $char_to_match = $self->_non_identifier_char;

    my $phrase_at_start = 0;

    while ($str =~ /((${char_to_match})\2*)/sg) {
        my $to_replace = $1;
        my $pos        = $-[1];

        my $count = length $to_replace;
        my $char  = substr $to_replace, 0, 1;

        my $replacement_phrase;
        my $use_underscore = 0;

        if (ord $char < 128) {
            $replacement_phrase = join ' ', @{ $ASCII_MAP{ord $char} };
        }
        elsif ($is_utf8) {
            my $decoded = lcfirst unidecode $char;

            $decoded =~ s/^\s+//;
            $decoded =~ s/\s+\z//;

            (my $decoded_without_spaces = $decoded) =~ s/\s+//g;

            my $bad_chars =()= $decoded_without_spaces =~ /$char_to_match/sg;

            # If Text::Unidecode gives us non-identifier chars, we use
            # either it or the UCD charname, whichever has fewer
            # non-identifier chars, after recursively passing the strings
            # through ->string_to_identifier.
            if ($bad_chars) {
                my $charname = lc charinfo(ord $char)->{name};

                $charname =~ s/^\s+//;
                $charname =~ s/\s+\z//;

                (my $charname_without_spaces = $charname) =~ s/\s+//g;

                my $charname_bad_chars =()=
                    $charname_without_spaces =~ /$char_to_match/sg;

                $decoded = $charname if $charname_bad_chars < $bad_chars;

                $decoded =
                    join ' ',
                    map $self->string_to_identifier($_),
                        split /\s+/, $decoded;
            }

            $replacement_phrase = $decoded;
        }
        else { # binary
            $replacement_phrase = sprintf '0x%X', ord $char;
            $use_underscore     = 1;
        }

        # For single char replacements, no separation or camelcasing is
        # necessary.
        if (length $replacement_phrase > 1) {
            $phrase_at_start = 1 if $pos == 0;

            $replacement_phrase = $self->_pluralize_phrase("$count $replacement_phrase")
                if $count > 1;

            {
                my $sep_char = $use_underscore ? '_' : $sep_char;

                if ($sep_char) {
                    $replacement_phrase = 
                        join($sep_char, split /\s+/, $replacement_phrase);

                    $replacement_phrase = $sep_char . $replacement_phrase
                        unless $pos == 0;

                    # Insert sep_char at the end of replacement text unless
                    # position is at the end of the string.
                    $replacement_phrase .= $sep_char
                        unless $pos + length($to_replace) == length($str);
                }
                else {
                    $replacement_phrase =
                        join '', map "\u$_", split /\s+/, $replacement_phrase;
                }
            }

            # titlecase the following text for camelCase identifiers
            substr($str, $pos + length($to_replace), 1) =
                ucfirst substr($str, $pos + length($to_replace), 1)
                if not $sep_char;
        }
        else {
            # For single char replacements we want to match the case.
            if (substr($str, $pos, 1) =~ /^\p{Lu}\z/) {
                $replacement_phrase = ucfirst $replacement_phrase;
            }
            else {
                $replacement_phrase = lcfirst $replacement_phrase;
            }
        }

        substr($str, $pos, length($to_replace)) = $replacement_phrase;
    }

    $str = lcfirst $str if $phrase_at_start;

    return $str;
}

=head1 SEE ALSO

L<String::ToIdentifier::EN::Unicode>,
L<Text::Unidecode>,
L<Lingua::EN::Inflect::Phrase>

=head1 AUTHOR

Rafael Kitover, C<< <rkitover@cpan.org> >>

=head1 BUGS

Please report any bugs or feature requests to C<bug-string-toidentifier-en at
rt.cpan.org>, or through the web interface at
L<http://rt.cpan.org/NoAuth/ReportBug.html?Queue=String-ToIdentifier-EN>.  I
will be notified, and then you'll automatically be notified of progress on your
bug as I make changes.

=head1 REPOSITORY

L<http://github.com/rkitover/string-toidentifier-en>

=head1 SUPPORT

More information on this module is available at:

=over 4

=item * RT: CPAN's request tracker (report bugs here)

L<http://rt.cpan.org/NoAuth/Bugs.html?Dist=String-ToIdentifier-EN>

=item * AnnoCPAN: Annotated CPAN documentation

L<http://annocpan.org/dist/String-ToIdentifier-EN>

=item * CPAN Ratings

L<http://cpanratings.perl.org/d/String-ToIdentifier-EN>

=item * MetaCPAN

L<https://metacpan.org/module/String::ToIdentifier::EN>

=item * Search CPAN

L<http://search.cpan.org/dist/String-ToIdentifier-EN/>

=back

=head1 LICENSE AND COPYRIGHT

Copyright (c) 2011 Rafael Kitover <rkitover@cpan.org>.

This program is free software; you can redistribute it and/or modify it
under the terms of either: the GNU General Public License as published
by the Free Software Foundation; or the Artistic License.

See http://dev.perl.org/licenses/ for more information.

=cut

1; # End of String::ToIdentifier::EN