7 use bytes (); # for $bytes::hint_bits
10 # Icky 3.2 names with parentheses.
11 'LINE FEED' => 'LINE FEED (LF)',
12 'FORM FEED' => 'FORM FEED (FF)',
13 'CARRIAGE RETURN' => 'CARRIAGE RETURN (CR)',
14 'NEXT LINE' => 'NEXT LINE (NEL)',
16 'LF' => 'LINE FEED (LF)',
17 'FF' => 'FORM FEED (FF)',
18 'CR' => 'CARRIAGE RETURN (CR)',
19 'NEL' => 'NEXT LINE (NEL)',
20 # More convenience. For futher convencience,
21 # it is suggested some way using using the NamesList
22 # aliases is implemented.
23 'ZWNJ' => 'ZERO WIDTH NON-JOINER',
24 'ZWJ' => 'ZERO WIDTH JOINER',
25 'BOM' => 'BYTE ORDER MARK',
29 # Pre-3.2 compatibility (only for the first 256 characters).
30 'HORIZONTAL TABULATION' => 'CHARACTER TABULATION',
31 'VERTICAL TABULATION' => 'LINE TABULATION',
32 'FILE SEPARATOR' => 'INFORMATION SEPARATOR FOUR',
33 'GROUP SEPARATOR' => 'INFORMATION SEPARATOR THREE',
34 'RECORD SEPARATOR' => 'INFORMATION SEPARATOR TWO',
35 'UNIT SEPARATOR' => 'INFORMATION SEPARATOR ONE',
36 'PARTIAL LINE DOWN' => 'PARTIAL LINE FORWARD',
37 'PARTIAL LINE UP' => 'PARTIAL LINE BACKWARD',
41 # User defined aliasses. Even more convenient :)
47 require Carp; goto &Carp::croak;
52 require Carp; goto &Carp::carp;
58 my $alias = ref $_[0] ? $_[0] : { @_ };
59 @alias3{keys %$alias} = values %$alias;
64 my ($arg, $file) = @_;
65 if (-f $arg && File::Spec->file_name_is_absolute ($arg)) {
68 elsif ($arg =~ m/^\w+$/) {
69 $file = "unicore/${arg}_alias.pl";
72 croak "Charnames alias files can only have identifier characters";
74 if (my @alias = do $file) {
75 @alias == 1 && !defined $alias[0] and
76 croak "$file cannot be used as alias file for charnames";
78 croak "$file did not return a (valid) list of alias pairs";
85 # This is not optimized in any way yet
90 if (exists $alias1{$name}) {
91 $name = $alias1{$name};
93 elsif (exists $alias2{$name}) {
95 warnings::warnif('deprecated', qq{Unicode character name "$name" is deprecated, use "$alias2{$name}" instead});
96 $name = $alias2{$name};
98 elsif (exists $alias3{$name}) {
99 $name = $alias3{$name};
106 if ($name eq "BYTE ORDER MARK") {
110 ## Suck in the code/name list as a big string.
112 ## "0052\t\tLATIN CAPITAL LETTER R\n"
113 $txt = do "unicore/Name.pl" unless $txt;
115 ## @off will hold the index into the code/name string of the start and
116 ## end of the name as we find it.
118 ## If :full, look for the name exactly
119 if ($^H{charnames_full} and $txt =~ /\t\t\Q$name\E$/m) {
120 @off = ($-[0], $+[0]);
123 ## If we didn't get above, and :short allowed, look for the short name.
124 ## The short name is like "greek:Sigma"
126 if ($^H{charnames_short} and $name =~ /^(.+?):(.+)/s) {
127 my ($script, $cname) = ($1, $2);
128 my $case = $cname =~ /[[:upper:]]/ ? "CAPITAL" : "SMALL";
129 if ($txt =~ m/\t\t\U$script\E (?:$case )?LETTER \U\Q$cname\E$/m) {
130 @off = ($-[0], $+[0]);
135 ## If we still don't have it, check for the name among the loaded
138 my $case = $name =~ /[[:upper:]]/ ? "CAPITAL" : "SMALL";
139 for my $script (@{$^H{charnames_scripts}}) {
140 if ($txt =~ m/\t\t$script (?:$case )?LETTER \U\Q$name\E$/m) {
141 @off = ($-[0], $+[0]);
147 ## If we don't have it by now, give up.
149 carp "Unknown charname '$name'";
154 ## Now know where in the string the name starts.
155 ## The code, in hex, is before that.
157 ## The code can be 4-6 characters long, so we've got to sort of
158 ## go look for it, just after the newline that comes before $off[0].
160 ## This would be much easier if unicore/Name.pl had info in
161 ## a name/code order, instead of code/name order.
163 ## The +1 after the rindex() is to skip past the newline we're finding,
164 ## or, if the rindex() fails, to put us to an offset of zero.
166 my $hexstart = rindex($txt, "\n", $off[0]) + 1;
168 ## we know where it starts, so turn into number -
169 ## the ordinal for the char.
170 $ord = CORE::hex substr($txt, $hexstart, $off[0] - $hexstart);
173 if ($^H & $bytes::hint_bits) { # "use bytes" in effect?
175 return chr $ord if $ord <= 255;
176 my $hex = sprintf "%04x", $ord;
177 if (not defined $fname) {
178 $fname = substr $txt, $off[0] + 2, $off[1] - $off[0] - 2;
180 croak "Character 0x$hex with name '$fname' is above 0xFF";
183 no warnings 'utf8'; # allow even illegal characters
184 return pack "U", $ord;
189 shift; ## ignore class name
192 carp("`use charnames' needs explicit imports list");
194 $^H{charnames} = \&charnames ;
197 ## fill %h keys with our @_ args.
199 my ($promote, %h, @args) = (0);
200 while (my $arg = shift) {
201 if ($arg eq ":alias") {
203 croak ":alias needs an argument in charnames";
206 ref $alias eq "HASH" or
207 croak "Only HASH reference supported as argument to :alias";
211 if ($alias =~ m{:(\w+)$}) {
212 $1 eq "full" || $1 eq "short" and
213 croak ":alias cannot use existing pragma :$1 (reversed order?)";
214 alias_file ($1) and $promote = 1;
220 if (substr($arg, 0, 1) eq ':' and ! ($arg eq ":full" || $arg eq ":short")) {
221 warn "unsupported special '$arg' in charnames";
226 @args == 0 && $promote and @args = (":full");
227 @h{@args} = (1) x @args;
229 $^H{charnames_full} = delete $h{':full'};
230 $^H{charnames_short} = delete $h{':short'};
231 $^H{charnames_scripts} = [map uc, keys %h];
234 ## If utf8? warnings are enabled, and some scripts were given,
235 ## see if at least we can find one letter of each script.
237 if (warnings::enabled('utf8') && @{$^H{charnames_scripts}}) {
238 $txt = do "unicore/Name.pl" unless $txt;
240 for my $script (@{$^H{charnames_scripts}}) {
241 if (not $txt =~ m/\t\t$script (?:CAPITAL |SMALL )?LETTER /) {
242 warnings::warn('utf8', "No such script: '$script'");
253 carp "charnames::viacode() expects one argument";
259 # this is derived from Unicode::UCD, where it is nearly the same as the
260 # function _getcode(), but it makes sure that even a hex argument has the
261 # proper number of leading zeros, which is critical in matching against $txt
264 if ($arg =~ /^[1-9]\d*$/) {
265 $hex = sprintf "%04X", $arg;
266 } elsif ($arg =~ /^(?:[Uu]\+|0[xX])?([[:xdigit:]]+)$/) {
267 # Below is the line that differs from the _getcode() source
268 $hex = sprintf "%04X", hex $arg;
270 carp("unexpected arg \"$arg\" to charnames::viacode()");
274 # checking the length first is slightly faster
275 if (length($hex) > 5 && hex($hex) > 0x10FFFF) {
276 carp "Unicode characters only allocated up to U+10FFFF (you asked for U+$hex)";
280 return $viacode{$hex} if exists $viacode{$hex};
282 $txt = do "unicore/Name.pl" unless $txt;
284 return unless $txt =~ m/^$hex\t\t(.+)/m;
294 carp "charnames::vianame() expects one name argument";
300 return chr CORE::hex $1 if $arg =~ /^U\+([0-9a-fA-F]+)$/;
302 return $vianame{$arg} if exists $vianame{$arg};
304 $txt = do "unicore/Name.pl" unless $txt;
306 my $pos = index $txt, "\t\t$arg\n";
308 my $posLF = rindex $txt, "\n", $pos;
309 (my $code = substr $txt, $posLF + 1, 6) =~ tr/\t//d;
310 return $vianame{$arg} = CORE::hex $code;
312 # If $pos is at the 1st line, $posLF must be -1 (not found);
313 # then $posLF + 1 equals to 0 (at the beginning of $txt).
314 # Otherwise $posLF is the position of "\n";
315 # then $posLF + 1 must be the position of the next to "\n"
316 # (the beginning of the line).
317 # substr($txt, $posLF + 1, 6) may be "0000\t\t", "00A1\t\t",
318 # "10300\t", "100000", etc. So we can get the code via removing TAB.
330 charnames - define character names for C<\N{named}> string literal escapes
334 use charnames ':full';
335 print "\N{GREEK SMALL LETTER SIGMA} is called sigma.\n";
337 use charnames ':short';
338 print "\N{greek:Sigma} is an upper-case sigma.\n";
340 use charnames qw(cyrillic greek);
341 print "\N{sigma} is Greek sigma, and \N{be} is Cyrillic b.\n";
343 use charnames ":full", ":alias" => {
344 e_ACUTE => "LATIN SMALL LETTER E WITH ACUTE",
346 print "\N{e_ACUTE} is a small letter e with an acute.\n";
349 print charnames::viacode(0x1234); # prints "ETHIOPIC SYLLABLE SEE"
350 printf "%04X", charnames::vianame("GOTHIC LETTER AHSA"); # prints "10330"
354 Pragma C<use charnames> supports arguments C<:full>, C<:short>, script
355 names and customized aliases. If C<:full> is present, for expansion of
356 C<\N{CHARNAME}>, the string C<CHARNAME> is first looked up in the list of
357 standard Unicode character names. If C<:short> is present, and
358 C<CHARNAME> has the form C<SCRIPT:CNAME>, then C<CNAME> is looked up
359 as a letter in script C<SCRIPT>. If pragma C<use charnames> is used
360 with script name arguments, then for C<\N{CHARNAME}> the name
361 C<CHARNAME> is looked up as a letter in the given scripts (in the
362 specified order). Customized aliases are explained in L</CUSTOM ALIASES>.
364 For lookup of C<CHARNAME> inside a given script C<SCRIPTNAME>
365 this pragma looks for the names
367 SCRIPTNAME CAPITAL LETTER CHARNAME
368 SCRIPTNAME SMALL LETTER CHARNAME
369 SCRIPTNAME LETTER CHARNAME
371 in the table of standard Unicode names. If C<CHARNAME> is lowercase,
372 then the C<CAPITAL> variant is ignored, otherwise the C<SMALL> variant
375 Note that C<\N{...}> is compile-time, it's a special form of string
376 constant used inside double-quoted strings: in other words, you cannot
377 use variables inside the C<\N{...}>. If you want similar run-time
378 functionality, use charnames::vianame().
380 For the C0 and C1 control characters (U+0000..U+001F, U+0080..U+009F)
381 as of Unicode 3.1, there are no official Unicode names but you can use
382 instead the ISO 6429 names (LINE FEED, ESCAPE, and so forth). In
383 Unicode 3.2 (as of Perl 5.8) some naming changes take place ISO 6429
384 has been updated, see L</ALIASES>. Also note that the U+UU80, U+0081,
385 U+0084, and U+0099 do not have names even in ISO 6429.
387 Since the Unicode standard uses "U+HHHH", so can you: "\N{U+263a}"
388 is the Unicode smiley face, or "\N{WHITE SMILING FACE}".
392 A few aliases have been defined for convenience: instead of having
393 to use the official names
400 (yes, with parentheses) one can use
421 for ZERO WIDTH NON-JOINER and ZERO WIDTH JOINER.
423 For backward compatibility one can use the old names for
424 certain C0 and C1 controls
428 HORIZONTAL TABULATION CHARACTER TABULATION
429 VERTICAL TABULATION LINE TABULATION
430 FILE SEPARATOR INFORMATION SEPARATOR FOUR
431 GROUP SEPARATOR INFORMATION SEPARATOR THREE
432 RECORD SEPARATOR INFORMATION SEPARATOR TWO
433 UNIT SEPARATOR INFORMATION SEPARATOR ONE
434 PARTIAL LINE DOWN PARTIAL LINE FORWARD
435 PARTIAL LINE UP PARTIAL LINE BACKWARD
437 but the old names in addition to giving the character
438 will also give a warning about being deprecated.
440 =head1 CUSTOM ALIASES
442 This version of charnames supports three mechanisms of adding local
443 or customized aliases to standard Unicode naming conventions (:full).
445 Note that an alias should not be something that is a legal curly
446 brace-enclosed quantifier (see L<perlreref/QUANTIFIERS>). For example
447 C<\N{123}> means to match 123 non-newline characters, and is not treated as an
448 alias. Aliases are discouraged from beginning with anything other than an
449 alphabetic character and from containing anything other than alphanumerics,
450 spaces, dashes, colons, parentheses, and underscores. Currently they must be
453 =head2 Anonymous hashes
455 use charnames ":full", ":alias" => {
456 e_ACUTE => "LATIN SMALL LETTER E WITH ACUTE",
458 my $str = "\N{e_ACUTE}";
462 use charnames ":full", ":alias" => "pro";
464 will try to read "unicore/pro_alias.pl" from the @INC path. This
465 file should return a list in plain perl:
468 A_GRAVE => "LATIN CAPITAL LETTER A WITH GRAVE",
469 A_CIRCUM => "LATIN CAPITAL LETTER A WITH CIRCUMFLEX",
470 A_DIAERES => "LATIN CAPITAL LETTER A WITH DIAERESIS",
471 A_TILDE => "LATIN CAPITAL LETTER A WITH TILDE",
472 A_BREVE => "LATIN CAPITAL LETTER A WITH BREVE",
473 A_RING => "LATIN CAPITAL LETTER A WITH RING ABOVE",
474 A_MACRON => "LATIN CAPITAL LETTER A WITH MACRON",
477 =head2 Alias shortcut
479 use charnames ":alias" => ":pro";
481 works exactly the same as the alias pairs, only this time,
482 ":full" is inserted automatically as first argument (if no
483 other argument is given).
485 =head1 charnames::viacode(code)
487 Returns the full name of the character indicated by the numeric code.
490 print charnames::viacode(0x2722);
492 prints "FOUR TEARDROP-SPOKED ASTERISK".
494 Returns undef if no name is known for the code.
496 This works only for the standard names, and does not yet apply
497 to custom translators.
499 Notice that the name returned for of U+FEFF is "ZERO WIDTH NO-BREAK
500 SPACE", not "BYTE ORDER MARK".
502 =head1 charnames::vianame(name)
504 Returns the code point indicated by the name.
507 printf "%04X", charnames::vianame("FOUR TEARDROP-SPOKED ASTERISK");
511 Returns undef if the name is unknown.
513 This works only for the standard names, and does not yet apply
514 to custom translators.
516 =head1 CUSTOM TRANSLATORS
518 The mechanism of translation of C<\N{...}> escapes is general and not
519 hardwired into F<charnames.pm>. A module can install custom
520 translations (inside the scope which C<use>s the module) with the
521 following magic incantation:
525 $^H{charnames} = \&translator;
528 Here translator() is a subroutine which takes C<CHARNAME> as an
529 argument, and returns text to insert into the string instead of the
530 C<\N{CHARNAME}> escape. Since the text to insert should be different
531 in C<bytes> mode and out of it, the function should check the current
532 state of C<bytes>-flag as in:
534 use bytes (); # for $bytes::hint_bits
536 if ($^H & $bytes::hint_bits) {
537 return bytes_translator(@_);
540 return utf8_translator(@_);
544 See L</CUSTOM ALIASES> above for restrictions on C<CHARNAME>.
546 =head1 ILLEGAL CHARACTERS
548 If you ask by name for a character that does not exist, a warning is given and
549 the Unicode I<replacement character> "\x{FFFD}" is returned.
551 If you ask by code for a character that is unassigned, no warning is
552 given and C<undef> is returned. (Though if you ask for a code point
553 past U+10FFFF you do get a warning.) See L</BUGS> below.
557 viacode should return an empty string for unassigned in-range Unicode code
558 points, as that is their correct current name.
560 viacode(0) doesn't return C<NULL>, but C<undef>
562 vianame returns a chr if the input name is of the form C<U+...>, and an ord
563 otherwise. It is planned to change this to always return an ord.
565 None of the functions work on almost all the Hangul syllable and CJK Unicode
566 characters that have their code points as part of their names.
568 Names must be ASCII characters only.
570 Unicode standard named sequences are not recognized, such as
571 C<LATIN CAPITAL LETTER A WITH MACRON AND GRAVE>
572 (which should mean C<LATIN CAPITAL LETTER A WITH MACRON> with an additional
573 C<COMBINING GRAVE ACCENT>).
575 Since evaluation of the translation function happens in the middle of
576 compilation (of a string literal), the translation function should not
577 do any C<eval>s or C<require>s. This restriction should be lifted in
578 a future version of Perl.