7 use bytes (); # for $bytes::hint_bits
8 $charnames::hint_bits = 0x20000;
11 # Icky 3.2 names with parentheses.
12 'LINE FEED' => 'LINE FEED (LF)',
13 'FORM FEED' => 'FORM FEED (FF)',
14 'CARRIAGE RETURN' => 'CARRIAGE RETURN (CR)',
15 'NEXT LINE' => 'NEXT LINE (NEL)',
17 'LF' => 'LINE FEED (LF)',
18 'FF' => 'FORM FEED (FF)',
19 'CR' => 'CARRIAGE RETURN (CR)',
20 'NEL' => 'NEXT LINE (NEL)',
21 # More convenience. For futher convencience,
22 # it is suggested some way using using the NamesList
23 # aliases is implemented.
24 'ZWNJ' => 'ZERO WIDTH NON-JOINER',
25 'ZWJ' => 'ZERO WIDTH JOINER',
26 'BOM' => 'BYTE ORDER MARK',
30 # Pre-3.2 compatibility (only for the first 256 characters).
31 'HORIZONTAL TABULATION' => 'CHARACTER TABULATION',
32 'VERTICAL TABULATION' => 'LINE TABULATION',
33 'FILE SEPARATOR' => 'INFORMATION SEPARATOR FOUR',
34 'GROUP SEPARATOR' => 'INFORMATION SEPARATOR THREE',
35 'RECORD SEPARATOR' => 'INFORMATION SEPARATOR TWO',
36 'UNIT SEPARATOR' => 'INFORMATION SEPARATOR ONE',
37 'PARTIAL LINE DOWN' => 'PARTIAL LINE FORWARD',
38 'PARTIAL LINE UP' => 'PARTIAL LINE BACKWARD',
42 # User defined aliasses. Even more convenient :)
49 my $alias = ref $_[0] ? $_[0] : { @_ };
50 @alias3{keys %$alias} = values %$alias;
56 my $file = -f $arg ? $arg : "unicore/${arg}_alias.pl";
57 if (my @alias = do $file) {
64 # This is not optimized in any way yet
69 if (exists $alias1{$name}) {
70 $name = $alias1{$name};
72 elsif (exists $alias2{$name}) {
74 warnings::warnif('deprecated', qq{Unicode character name "$name" is deprecated, use "$alias2{$name}" instead});
75 $name = $alias2{$name};
77 elsif (exists $alias3{$name}) {
78 $name = $alias3{$name};
85 if ($name eq "BYTE ORDER MARK") {
89 ## Suck in the code/name list as a big string.
91 ## "0052\t\tLATIN CAPITAL LETTER R\n"
92 $txt = do "unicore/Name.pl" unless $txt;
94 ## @off will hold the index into the code/name string of the start and
95 ## end of the name as we find it.
97 ## If :full, look for the the name exactly
98 if ($^H{charnames_full} and $txt =~ /\t\t\Q$name\E$/m) {
99 @off = ($-[0], $+[0]);
102 ## If we didn't get above, and :short allowed, look for the short name.
103 ## The short name is like "greek:Sigma"
105 if ($^H{charnames_short} and $name =~ /^(.+?):(.+)/s) {
106 my ($script, $cname) = ($1, $2);
107 my $case = $cname =~ /[[:upper:]]/ ? "CAPITAL" : "SMALL";
108 if ($txt =~ m/\t\t\U$script\E (?:$case )?LETTER \U\Q$cname\E$/m) {
109 @off = ($-[0], $+[0]);
114 ## If we still don't have it, check for the name among the loaded
117 my $case = $name =~ /[[:upper:]]/ ? "CAPITAL" : "SMALL";
118 for my $script (@{$^H{charnames_scripts}}) {
119 if ($txt =~ m/\t\t$script (?:$case )?LETTER \U\Q$name\E$/m) {
120 @off = ($-[0], $+[0]);
126 ## If we don't have it by now, give up.
128 carp "Unknown charname '$name'";
133 ## Now know where in the string the name starts.
134 ## The code, in hex, is before that.
136 ## The code can be 4-6 characters long, so we've got to sort of
137 ## go look for it, just after the newline that comes before $off[0].
139 ## This would be much easier if unicore/Name.pl had info in
140 ## a name/code order, instead of code/name order.
142 ## The +1 after the rindex() is to skip past the newline we're finding,
143 ## or, if the rindex() fails, to put us to an offset of zero.
145 my $hexstart = rindex($txt, "\n", $off[0]) + 1;
147 ## we know where it starts, so turn into number -
148 ## the ordinal for the char.
149 $ord = hex substr($txt, $hexstart, $off[0] - $hexstart);
152 if ($^H & $bytes::hint_bits) { # "use bytes" in effect?
154 return chr $ord if $ord <= 255;
155 my $hex = sprintf "%04x", $ord;
156 if (not defined $fname) {
157 $fname = substr $txt, $off[0] + 2, $off[1] - $off[0] - 2;
159 croak "Character 0x$hex with name '$fname' is above 0xFF";
162 no warnings 'utf8'; # allow even illegal characters
163 return pack "U", $ord;
168 shift; ## ignore class name
171 carp("`use charnames' needs explicit imports list");
173 $^H |= $charnames::hint_bits;
174 $^H{charnames} = \&charnames ;
177 ## fill %h keys with our @_ args.
179 my ($promote, %h, @args) = (0);
180 while (@_ and $_ = shift) {
181 if ($_ eq ":alias" && @_) {
184 ref $alias eq "HASH" or
185 die "Only HASH reference supported as argument to :alias";
189 if ($alias =~ m{:(\w+)$} and $1 ne "full" && $1 ne "short") {
190 alias_file ($1) and $promote = 1, next;
192 alias_file ($alias) and next;
196 @args == 0 && $promote and @args = (":full");
197 @h{@args} = (1) x @args;
199 $^H{charnames_full} = delete $h{':full'};
200 $^H{charnames_short} = delete $h{':short'};
201 $^H{charnames_scripts} = [map uc, keys %h];
204 ## If utf8? warnings are enabled, and some scripts were given,
205 ## see if at least we can find one letter of each script.
207 if (warnings::enabled('utf8') && @{$^H{charnames_scripts}}) {
208 $txt = do "unicore/Name.pl" unless $txt;
210 for my $script (@{$^H{charnames_scripts}}) {
211 if (not $txt =~ m/\t\t$script (?:CAPITAL |SMALL )?LETTER /) {
212 warnings::warn('utf8', "No such script: '$script'");
218 require Unicode::UCD; # for Unicode::UCD::_getcode()
225 carp "charnames::viacode() expects one argument";
230 my $code = Unicode::UCD::_getcode($arg);
235 $hex = sprintf "%04X", $arg;
237 carp("unexpected arg \"$arg\" to charnames::viacode()");
241 if ($code > 0x10FFFF) {
242 carp sprintf "Unicode characters only allocated up to U+10FFFF (you asked for U+%X)", $hex;
246 return $viacode{$hex} if exists $viacode{$hex};
248 $txt = do "unicore/Name.pl" unless $txt;
250 if ($txt =~ m/^$hex\t\t(.+)/m) {
251 return $viacode{$hex} = $1;
262 carp "charnames::vianame() expects one name argument";
268 return chr hex $1 if $arg =~ /^U\+([0-9a-fA-F]+)$/;
270 return $vianame{$arg} if exists $vianame{$arg};
272 $txt = do "unicore/Name.pl" unless $txt;
274 my $pos = index $txt, "\t\t$arg\n";
276 my $posLF = rindex $txt, "\n", $pos;
277 (my $code = substr $txt, $posLF + 1, 6) =~ tr/\t//d;
278 return $vianame{$arg} = hex $code;
280 # If $pos is at the 1st line, $posLF must be $[ - 1 (not found);
281 # then $posLF + 1 equals to $[ (at the beginning of $txt).
282 # Otherwise $posLF is the position of "\n";
283 # then $posLF + 1 must be the position of the next to "\n"
284 # (the beginning of the line).
285 # substr($txt, $posLF + 1, 6) may be "0000\t\t", "00A1\t\t",
286 # "10300\t", "100000", etc. So we can get the code via removing TAB.
298 charnames - define character names for C<\N{named}> string literal escapes
302 use charnames ':full';
303 print "\N{GREEK SMALL LETTER SIGMA} is called sigma.\n";
305 use charnames ':short';
306 print "\N{greek:Sigma} is an upper-case sigma.\n";
308 use charnames qw(cyrillic greek);
309 print "\N{sigma} is Greek sigma, and \N{be} is Cyrillic b.\n";
311 use charnames ":full", ":alias" => {
312 e_ACUTE => "LATIN SMALL LETTER E WITH ACUTE",
314 print "\N{e_ACUTE} is a small letter e with an acute.\n";
316 print charnames::viacode(0x1234); # prints "ETHIOPIC SYLLABLE SEE"
317 printf "%04X", charnames::vianame("GOTHIC LETTER AHSA"); # prints "10330"
321 Pragma C<use charnames> supports arguments C<:full>, C<:short>, script
322 names and customized aliases. If C<:full> is present, for expansion of
323 C<\N{CHARNAME}> string C<CHARNAME> is first looked in the list of
324 standard Unicode names of chars. If C<:short> is present, and
325 C<CHARNAME> has the form C<SCRIPT:CNAME>, then C<CNAME> is looked up
326 as a letter in script C<SCRIPT>. If pragma C<use charnames> is used
327 with script name arguments, then for C<\N{CHARNAME}> the name
328 C<CHARNAME> is looked up as a letter in the given scripts (in the
329 specified order). Customized aliases are explained in L</CUSTOM ALIASES>.
331 For lookup of C<CHARNAME> inside a given script C<SCRIPTNAME>
332 this pragma looks for the names
334 SCRIPTNAME CAPITAL LETTER CHARNAME
335 SCRIPTNAME SMALL LETTER CHARNAME
336 SCRIPTNAME LETTER CHARNAME
338 in the table of standard Unicode names. If C<CHARNAME> is lowercase,
339 then the C<CAPITAL> variant is ignored, otherwise the C<SMALL> variant
342 Note that C<\N{...}> is compile-time, it's a special form of string
343 constant used inside double-quoted strings: in other words, you cannot
344 use variables inside the C<\N{...}>. If you want similar run-time
345 functionality, use charnames::vianame().
347 For the C0 and C1 control characters (U+0000..U+001F, U+0080..U+009F)
348 as of Unicode 3.1, there are no official Unicode names but you can use
349 instead the ISO 6429 names (LINE FEED, ESCAPE, and so forth). In
350 Unicode 3.2 (as of Perl 5.8) some naming changes take place ISO 6429
351 has been updated, see L</ALIASES>. Also note that the U+UU80, U+0081,
352 U+0084, and U+0099 do not have names even in ISO 6429.
354 Since the Unicode standard uses "U+HHHH", so can you: "\N{U+263a}"
355 is the Unicode smiley face, or "\N{WHITE SMILING FACE}".
357 =head1 CUSTOM TRANSLATORS
359 The mechanism of translation of C<\N{...}> escapes is general and not
360 hardwired into F<charnames.pm>. A module can install custom
361 translations (inside the scope which C<use>s the module) with the
362 following magic incantation:
364 use charnames (); # for $charnames::hint_bits
367 $^H |= $charnames::hint_bits;
368 $^H{charnames} = \&translator;
371 Here translator() is a subroutine which takes C<CHARNAME> as an
372 argument, and returns text to insert into the string instead of the
373 C<\N{CHARNAME}> escape. Since the text to insert should be different
374 in C<bytes> mode and out of it, the function should check the current
375 state of C<bytes>-flag as in:
377 use bytes (); # for $bytes::hint_bits
379 if ($^H & $bytes::hint_bits) {
380 return bytes_translator(@_);
383 return utf8_translator(@_);
387 =head1 CUSTOM ALIASES
389 This version of charnames supports three mechanisms of adding local
390 or customized aliases to standard Unicode naming conventions (:full)
392 =head2 Anonymous hashes
394 use charnames ":full", ":alias" => {
395 e_ACUTE => "LATIN SMALL LETTER E WITH ACUTE",
397 my $str = "\N{e_ACUTE}";
401 use charnames ":full", ":alias" => "pro";
403 will try to read "unicore/pro_alias.pl" from the @INC path. This
404 file should return a list in plain perl:
407 A_GRAVE => "LATIN CAPITAL LETTER A WITH GRAVE",
408 A_CIRCUM => "LATIN CAPITAL LETTER A WITH CIRCUMFLEX",
409 A_DIAERES => "LATIN CAPITAL LETTER A WITH DIAERESIS",
410 A_TILDE => "LATIN CAPITAL LETTER A WITH TILDE",
411 A_BREVE => "LATIN CAPITAL LETTER A WITH BREVE",
412 A_RING => "LATIN CAPITAL LETTER A WITH RING ABOVE",
413 A_MACRON => "LATIN CAPITAL LETTER A WITH MACRON",
416 =head2 Alias shortcut
418 use charnames ":alias" => ":pro";
420 works exactly the same as the alias pairs, only this time,
421 ":full" is inserted automatically as first argument (if no
422 other argument is given).
424 =head1 charnames::viacode(code)
426 Returns the full name of the character indicated by the numeric code.
429 print charnames::viacode(0x2722);
431 prints "FOUR TEARDROP-SPOKED ASTERISK".
433 Returns undef if no name is known for the code.
435 This works only for the standard names, and does not yet apply
436 to custom translators.
438 Notice that the name returned for of U+FEFF is "ZERO WIDTH NO-BREAK
439 SPACE", not "BYTE ORDER MARK".
441 =head1 charnames::vianame(name)
443 Returns the code point indicated by the name.
446 printf "%04X", charnames::vianame("FOUR TEARDROP-SPOKED ASTERISK");
450 Returns undef if the name is unknown.
452 This works only for the standard names, and does not yet apply
453 to custom translators.
457 A few aliases have been defined for convenience: instead of having
458 to use the official names
465 (yes, with parentheses) one can use
486 for ZERO WIDTH NON-JOINER and ZERO WIDTH JOINER.
488 For backward compatibility one can use the old names for
489 certain C0 and C1 controls
493 HORIZONTAL TABULATION CHARACTER TABULATION
494 VERTICAL TABULATION LINE TABULATION
495 FILE SEPARATOR INFORMATION SEPARATOR FOUR
496 GROUP SEPARATOR INFORMATION SEPARATOR THREE
497 RECORD SEPARATOR INFORMATION SEPARATOR TWO
498 UNIT SEPARATOR INFORMATION SEPARATOR ONE
499 PARTIAL LINE DOWN PARTIAL LINE FORWARD
500 PARTIAL LINE UP PARTIAL LINE BACKWARD
502 but the old names in addition to giving the character
503 will also give a warning about being deprecated.
505 =head1 ILLEGAL CHARACTERS
507 If you ask by name for a character that does not exist, a warning is
508 given and the Unicode I<replacement character> "\x{FFFD}" is returned.
510 If you ask by code for a character that does not exist, no warning is
511 given and C<undef> is returned. (Though if you ask for a code point
512 past U+10FFFF you do get a warning.)
516 Since evaluation of the translation function happens in a middle of
517 compilation (of a string literal), the translation function should not
518 do any C<eval>s or C<require>s. This restriction should be lifted in
519 a future version of Perl.