1 # !!!!!!! INTERNAL PERL USE ONLY !!!!!!!
2 # This helper module is for internal use by core Perl only. This module is
3 # subject to change or removal at any time without notice. Don't use it
4 # directly. Use the public <charnames> module instead.
10 our $VERSION = '1.32';
11 use unicore::Name; # mktables-generated algorithmically-defined names
13 use bytes (); # for $bytes::hint_bits
14 use re "/aa"; # Everything in here should be ASCII
16 $Carp::Internal{ (__PACKAGE__) } = 1;
18 # Translate between Unicode character names and their code points. This is a
19 # submodule of package <charnames>, used to allow \N{...} to be autoloaded,
20 # but it was decided not to autoload the various functions in charnames; the
21 # splitting allows this behavior.
23 # The official names with their code points are stored in a table in
24 # lib/unicore/Name.pl which is read in as a large string (almost 3/4 Mb in
25 # Unicode 6.0). Each code point/name combination is separated by a \n in the
26 # string. (Some of the CJK and the Hangul syllable names are determined
27 # instead algorithmically via subroutines stored instead in
28 # lib/unicore/Name.pm). Because of the large size of this table, it isn't
29 # converted into hashes for faster lookup.
31 # But, user defined aliases are stored in their own hashes, as are Perl
32 # extensions to the official names. These are checked first before looking at
35 # Basically, the table is grepped for the input code point (viacode()) or
36 # name (the other functions), and the corresponding value on the same line is
37 # returned. The grepping is done by turning the input into a regular
38 # expression. Thus, the same table does double duty, used by both name and
39 # code point lookup. (If we were to have hashes, we would need two, one for
40 # each lookup direction.)
42 # For loose name matching, the logical thing would be to have a table
43 # with all the ignorable characters squeezed out, and then grep it with the
44 # similiarly-squeezed input name. (And this is in fact how the lookups are
45 # done with the small Perl extension hashes.) But since we need to be able to
46 # go from code point to official name, the original table would still need to
47 # exist. Due to the large size of the table, it was decided to not read
48 # another very large string into memory for a second table. Instead, the
49 # regular expression of the input name is modified to have optional spaces and
50 # dashes between characters. For example, in strict matching, the regular
51 # expression would be:
53 # Under loose matching, the blank would be squeezed out, and the re would be:
54 # qr/\tD[- ]?I[- ]?G[- ]?I[- ]?T[- ]?O[- ]?N[- ]?E$/m
55 # which matches a blank or dash between any characters in the official table.
57 # This is also how script lookup is done. Basically the re looks like
58 # qr/ (?:LATIN|GREEK|CYRILLIC) (?:SMALL )?LETTER $name/
59 # where $name is the loose or strict regex for the remainder of the name.
61 # The hashes are stored as utf8 strings. This makes it easier to deal with
62 # sequences. I (khw) also tried making Name.pl utf8, but it slowed things
63 # down by a factor of 7. I then tried making Name.pl store the ut8
64 # equivalents but not calling them utf8. That led to similar speed as leaving
65 # it alone, but since that is harder for a human to parse, I left it as-is.
67 my %system_aliases = (
69 'SINGLE-SHIFT 2' => pack("U", 0x8E),
70 'SINGLE-SHIFT 3' => pack("U", 0x8F),
71 'PRIVATE USE 1' => pack("U", 0x91),
72 'PRIVATE USE 2' => pack("U", 0x92),
75 # These are the aliases above that differ under :loose and :full matching
76 # because the :full versions have blanks or hyphens in them.
77 #my %loose_system_aliases = (
80 #my %deprecated_aliases;
81 #$deprecated_aliases{'BELL'} = pack("U", 0x07) if $^V lt v5.17.0;
83 #my %loose_deprecated_aliases = (
86 # These are special cased in :loose matching, differing only in a medial
88 my $HANGUL_JUNGSEONG_O_E_utf8 = pack("U", 0x1180);
89 my $HANGUL_JUNGSEONG_OE_utf8 = pack("U", 0x116C);
92 my $txt; # The table of official character names
94 my %full_names_cache; # Holds already-looked-up names, so don't have to
95 # re-look them up again. The previous versions of charnames had scoping
96 # bugs. For example if we use script A in one scope and find and cache
97 # what Z resolves to, we can't use that cache in a different scope that
98 # uses script B instead of A, as Z might be an entirely different letter
99 # there; or there might be different aliases in effect in different
100 # scopes, or :short may be in effect or not effect in different scopes,
101 # or various combinations thereof. This was solved in this version
102 # mostly by moving things to %^H. But some things couldn't be moved
103 # there. One of them was the cache of runtime looked-up names, in part
104 # because %^H is read-only at runtime. I (khw) don't know why the cache
105 # was run-time only in the previous versions: perhaps oversight; perhaps
106 # that compile time looking doesn't happen in a loop so didn't think it
107 # was worthwhile; perhaps not wanting to make the cache too large. But
108 # I decided to make it compile time as well; this could easily be
110 # Anyway, this hash is not scoped, and is added to at runtime. It
111 # doesn't have scoping problems because the data in it is restricted to
112 # official names, which are always invariant, and we only set it and
113 # look at it at during :full lookups, so is unaffected by any other
114 # scoped options. I put this in to maintain parity with the older
115 # version. If desired, a %short_names cache could also be made, as well
116 # as one for each script, say in %script_names_cache, with each key
117 # being a hash for a script named in a 'use charnames' statement. I
118 # decided not to do that for now, just because it's added complication,
119 # and because I'm just trying to maintain parity, not extend it.
121 # Like %full_names_cache, but for use when :loose is in effect. There needs
122 # to be two caches because :loose may not be in effect for a scope, and a
123 # loose name could inappropriately be returned when only exact matching is
125 my %loose_names_cache;
127 # Designed so that test decimal first, and then hex. Leading zeros
128 # imply non-decimal, as do non-[0-9]
129 my $decimal_qr = qr/^[1-9]\d*$/;
131 # Returns the hex number in $1.
132 my $hex_qr = qr/^(?:[Uu]\+|0[xX])?([[:xdigit:]]+)$/;
136 require Carp; goto &Carp::croak;
141 require Carp; goto &Carp::carp;
144 sub alias (@) # Set up a single alias
146 my $alias = ref $_[0] ? $_[0] : { @_ };
147 foreach my $name (keys %$alias) {
148 my $value = $alias->{$name};
149 next unless defined $value; # Omit if screwed up.
151 # Is slightly slower to just after this statement see if it is
152 # decimal, since we already know it is after having converted from
153 # hex, but makes the code easier to maintain, and is called
154 # infrequently, only at compile-time
155 if ($value !~ $decimal_qr && $value =~ $hex_qr) {
156 $value = CORE::hex $1;
158 if ($value =~ $decimal_qr) {
159 no warnings qw(non_unicode surrogate nonchar); # Allow any non-malformed
160 $^H{charnames_ord_aliases}{$name} = pack("U", $value);
162 # Use a canonical form.
163 $^H{charnames_inverse_ords}{sprintf("%05X", $value)} = $name;
166 # XXX validate syntax when deprecation cycle complete. ie. start
167 # with an alpha only, etc.
168 $^H{charnames_name_aliases}{$name} = $value;
173 sub not_legal_use_bytes_msg {
174 my ($name, $utf8) = @_;
177 if (length($utf8) == 1) {
178 $return = sprintf("Character 0x%04x with name '%s' is", ord $utf8, $name);
180 $return = sprintf("String with name '%s' (and ordinals %s) contains character(s)", $name, join(" ", map { sprintf "0x%04X", ord $_ } split(//, $utf8)));
182 return $return . " above 0xFF with 'use bytes' in effect";
185 sub alias_file ($) # Reads a file containing alias definitions
187 my ($arg, $file) = @_;
188 if (-f $arg && File::Spec->file_name_is_absolute ($arg)) {
191 elsif ($arg =~ m/^\w+$/) {
192 $file = "unicore/${arg}_alias.pl";
195 croak "Charnames alias files can only have identifier characters";
197 if (my @alias = do $file) {
198 @alias == 1 && !defined $alias[0] and
199 croak "$file cannot be used as alias file for charnames";
201 croak "$file did not return a (valid) list of alias pairs";
208 # For use when don't import anything. This structure must be kept in
209 # sync with the one that import() fills up.
211 charnames_stringified_names => "",
212 charnames_stringified_ords => "",
213 charnames_scripts => "",
215 charnames_loose => 0,
216 charnames_short => 0,
220 sub lookup_name ($$$) {
221 my ($name, $wants_ord, $runtime) = @_;
223 # Lookup the name or sequence $name in the tables. If $wants_ord is false,
224 # returns the string equivalent of $name; if true, returns the ordinal value
225 # instead, but in this case $name must not be a sequence; otherwise undef is
226 # returned and a warning raised. $runtime is 0 if compiletime, otherwise
227 # gives the number of stack frames to go back to get the application caller
229 # If $name is not found, returns undef in runtime with no warning; and in
230 # compiletime, the Unicode replacement character, with a warning.
232 # It looks first in the aliases, then in the large table of official Unicode
235 my $utf8; # The string result
240 my $hints_ref = (caller($runtime))[10];
242 # If we didn't import anything (which happens with 'use charnames ()',
243 # substitute a dummy structure.
244 $hints_ref = \%dummy_H if ! defined $hints_ref
245 || (! defined $hints_ref->{charnames_full}
246 && ! defined $hints_ref->{charnames_loose});
248 # At runtime, but currently not at compile time, $^H gets
249 # stringified, so un-stringify back to the original data structures.
250 # These get thrown away by perl before the next invocation
251 # Also fill in the hash with the non-stringified data.
252 # N.B. New fields must be also added to %dummy_H
254 %{$^H{charnames_name_aliases}} = split ',',
255 $hints_ref->{charnames_stringified_names};
256 %{$^H{charnames_ord_aliases}} = split ',',
257 $hints_ref->{charnames_stringified_ords};
258 $^H{charnames_scripts} = $hints_ref->{charnames_scripts};
259 $^H{charnames_full} = $hints_ref->{charnames_full};
260 $^H{charnames_loose} = $hints_ref->{charnames_loose};
261 $^H{charnames_short} = $hints_ref->{charnames_short};
264 my $loose = $^H{charnames_loose};
265 my $lookup_name; # Input name suitably modified for grepping for in the
268 # User alias should be checked first or else can't override ours, and if we
269 # were to add any, could conflict with theirs.
270 if (exists $^H{charnames_ord_aliases}{$name}) {
271 $utf8 = $^H{charnames_ord_aliases}{$name};
273 elsif (exists $^H{charnames_name_aliases}{$name}) {
274 $name = $^H{charnames_name_aliases}{$name};
275 $save_input = $lookup_name = $name; # Cache the result for any error
277 # The aliases are documented to not match loosely, so change loose match
281 $^H{charnames_full} = 1;
286 # Here, not a user alias. That means that loose matching may be in
287 # effect; will have to modify the input name.
288 $lookup_name = $name;
290 $lookup_name = uc $lookup_name;
292 # Squeeze out all underscores
293 $lookup_name =~ s/_//g;
295 # Remove all medial hyphens
296 $lookup_name =~ s/ (?<= \S ) - (?= \S )//gx;
298 # Squeeze out all spaces
299 $lookup_name =~ s/\s//g;
302 # Here, $lookup_name has been modified as necessary for looking in the
303 # hashes. Check the system alias files next. Most of these aliases are
304 # the same for both strict and loose matching. To save space, the ones
305 # which differ are in their own separate hash, which is checked if loose
306 # matching is selected and the regular match fails. To save time, the
307 # loose hashes could be expanded to include all aliases, and there would
308 # only have to be one check. But if someone specifies :loose, they are
309 # interested in convenience over speed, and the time for this second check
310 # is miniscule compared to the rest of the routine.
311 if (exists $system_aliases{$lookup_name}) {
312 $utf8 = $system_aliases{$lookup_name};
314 # There are currently no entries in this hash, so don't waste time looking
315 # for them. But the code is retained for the unlikely possibility that
316 # some will be added in the future.
317 # elsif ($loose && exists $loose_system_aliases{$lookup_name}) {
318 # $utf8 = $loose_system_aliases{$lookup_name};
320 # if (exists $deprecated_aliases{$lookup_name}) {
322 # warnings::warnif('deprecated',
323 # "Unicode character name \"$name\" is deprecated, use \""
324 # . viacode(ord $deprecated_aliases{$lookup_name})
326 # $utf8 = $deprecated_aliases{$lookup_name};
328 # There are currently no entries in this hash, so don't waste time looking
329 # for them. But the code is retained for the unlikely possibility that
330 # some will be added in the future.
331 # elsif ($loose && exists $loose_deprecated_aliases{$lookup_name}) {
333 # warnings::warnif('deprecated',
334 # "Unicode character name \"$name\" is deprecated, use \""
335 # . viacode(ord $loose_deprecated_aliases{$lookup_name})
337 # $utf8 = $loose_deprecated_aliases{$lookup_name};
341 my @off; # Offsets into table of pattern match begin and end
343 # If haven't found it yet...
344 if (! defined $utf8) {
346 # See if has looked this input up earlier.
347 if (! $loose && $^H{charnames_full} && exists $full_names_cache{$name}) {
348 $utf8 = $full_names_cache{$name};
350 elsif ($loose && exists $loose_names_cache{$name}) {
351 $utf8 = $loose_names_cache{$name};
353 else { # Here, must do a look-up
355 # If full or loose matching succeeded, points to where to cache the
359 ## Suck in the code/name list as a big string.
361 ## "00052\tLATIN CAPITAL LETTER R\n"
363 # "0052 0303\tLATIN CAPITAL LETTER R WITH TILDE\n"
364 $txt = do "unicore/Name.pl" unless $txt;
366 ## @off will hold the index into the code/name string of the start and
367 ## end of the name as we find it.
369 ## If :loose, look for a loose match; if :full, look for the name
371 # First, see if the name is one which is algorithmically determinable.
372 # The subroutine is included in Name.pl. The table contained in
373 # $txt doesn't contain these. Experiments show that checking
374 # for these before checking for the regular names has no
375 # noticeable impact on performance for the regular names, but
376 # the other way around slows down finding these immensely.
377 # Algorithmically determinables are not placed in the cache because
378 # that uses up memory, and finding these again is fast.
379 if (($loose || $^H{charnames_full})
380 && (defined (my $ord = charnames::name_to_code_point_special($lookup_name, $loose))))
382 $utf8 = pack("U", $ord);
386 # Not algorithmically determinable; look up in the table. The name
387 # will be turned into a regex, so quote any meta characters.
388 $lookup_name = quotemeta $lookup_name;
392 # For loose matches, $lookup_name has already squeezed out the
393 # non-essential characters. We have to add in code to make the
394 # squeezed version match the non-squeezed equivalent in the table.
395 # The only remaining hyphens are ones that start or end a word in
396 # the original. They have been quoted in $lookup_name so they look
397 # like "\-". Change all other characters except the backslash
398 # quotes for any metacharacters, and the final character, so that
399 # e.g., COLON gets transformed into: /C[- ]?O[- ]?L[- ]?O[- ]?N/
400 $lookup_name =~ s/ (?! \\ -) # Don't do this to the \- sequence
401 ( [^-\\] ) # Nor the "-" within that sequence,
402 # nor the "\" that quotes metachars,
403 # but otherwise put the char into $1
404 (?=.) # And don't do it for the final char
405 /$1\[- \]?/gx; # And add an optional blank or
406 # '-' after each $1 char
408 # Those remaining hyphens were originally at the beginning or end of
409 # a word, so they can match either a blank before or after, but not
410 # both. (Keep in mind that they have been quoted, so are a '\-'
412 $lookup_name =~ s/\\ -/(?:- | -)/xg;
415 # Do the lookup in the full table if asked for, and if succeeds
416 # save the offsets and set where to cache the result.
417 if (($loose || $^H{charnames_full}) && $txt =~ /\t$lookup_name$/m) {
418 @off = ($-[0] + 1, $+[0]); # The 1 is for the tab
419 $cache_ref = ($loose) ? \%loose_names_cache : \%full_names_cache;
423 # Here, didn't look for, or didn't find the name.
424 # If :short is allowed, see if input is like "greek:Sigma".
425 # Keep in mind that $lookup_name has had the metas quoted.
426 my $scripts_trie = "";
427 my $name_has_uppercase;
428 if (($^H{charnames_short})
429 && $lookup_name =~ /^ (?: \\ \s)* # Quoted space
430 (.+?) # $1 = the script
434 (.+?) # $2 = the name
438 # Even in non-loose matching, the script traditionally has been
440 $scripts_trie = "\U$1";
443 # Use original name to find its input casing, but ignore the
444 # script part of that to make the determination.
445 $save_input = $name if ! defined $save_input;
447 $name_has_uppercase = $name =~ /[[:upper:]]/;
449 else { # Otherwise look in allowed scripts
450 $scripts_trie = $^H{charnames_scripts};
452 # Use original name to find its input casing
453 $name_has_uppercase = $name =~ /[[:upper:]]/;
456 my $case = $name_has_uppercase ? "CAPITAL" : "SMALL";
459 /\t (?: $scripts_trie ) \ (?:$case\ )? LETTER \ \U$lookup_name $/xm)
461 # Here we still don't have it, give up.
464 # May have zapped input name, get it again.
465 $name = (defined $save_input) ? $save_input : $_[0];
466 carp "Unknown charname '$name'";
467 return ($wants_ord) ? 0xFFFD : pack("U", 0xFFFD);
470 # Here have found the input name in the table.
471 @off = ($-[0] + 1, $+[0]); # The 1 is for the tab
474 # Here, the input name has been found; we haven't set up the output,
475 # but we know where in the string
476 # the name starts. The string is set up so that for single characters
477 # (and not named sequences), the name is preceded immediately by a
478 # tab and 5 hex digits for its code, with a \n before those. Named
479 # sequences won't have the 7th preceding character be a \n.
480 # (Actually, for the very first entry in the table this isn't strictly
481 # true: subtracting 7 will yield -1, and the substr below will
482 # therefore yield the very last character in the table, which should
483 # also be a \n, so the statement works anyway.)
484 if (substr($txt, $off[0] - 7, 1) eq "\n") {
485 $utf8 = pack("U", CORE::hex substr($txt, $off[0] - 6, 5));
487 # Handle the single loose matching special case, in which two names
488 # differ only by a single medial hyphen. If the original had a
489 # hyphen (or more) in the right place, then it is that one.
490 $utf8 = $HANGUL_JUNGSEONG_O_E_utf8
492 && $utf8 eq $HANGUL_JUNGSEONG_OE_utf8
493 && $name =~ m/O \s* - [-\s]* E/ix;
494 # Note that this wouldn't work if there were a 2nd
499 # Here, is a named sequence. Need to go looking for the beginning,
500 # which is just after the \n from the previous entry in the table.
501 # The +1 skips past that newline, or, if the rindex() fails, to put
502 # us to an offset of zero.
503 my $charstart = rindex($txt, "\n", $off[0] - 7) + 1;
504 $utf8 = pack("U*", map { CORE::hex }
505 split " ", substr($txt, $charstart, $off[0] - $charstart - 1));
509 # Cache the input so as to not have to search the large table
510 # again, but only if it came from the one search that we cache.
511 # (Haven't bothered with the pain of sorting out scoping issues for the
513 $cache_ref->{$name} = $utf8 if defined $cache_ref;
518 # Here, have the utf8. If the return is to be an ord, must be any single
521 return ord($utf8) if length $utf8 == 1;
525 # Here, wants string output. If utf8 is acceptable, just return what
526 # we've got; otherwise attempt to convert it to non-utf8 and return that.
527 my $in_bytes = ($runtime)
528 ? (caller $runtime)[8] & $bytes::hint_bits
529 : $^H & $bytes::hint_bits;
530 return $utf8 if (! $in_bytes || utf8::downgrade($utf8, 1)) # The 1 arg
531 # means don't die on failure
534 # Here, there is an error: either there are too many characters, or the
535 # result string needs to be non-utf8, and at least one character requires
536 # utf8. Prefer any official name over the input one for the error message.
538 $name = substr($txt, $off[0], $off[1] - $off[0]) if @off;
541 $name = (defined $save_input) ? $save_input : $_[0];
545 # Only way to get here in this case is if result too long. Message
546 # assumes that our only caller that requires single char result is
548 carp "charnames::vianame() doesn't handle named sequences ($name). Use charnames::string_vianame() instead";
552 # Only other possible failure here is from use bytes.
554 carp not_legal_use_bytes_msg($name, $utf8);
557 croak not_legal_use_bytes_msg($name, $utf8);
564 # For \N{...}. Looks up the character name and returns the string
565 # representation of it.
567 # The first 0 arg means wants a string returned; the second that we are in
569 return lookup_name($_[0], 0, 0);
574 shift; ## ignore class name
577 carp("'use charnames' needs explicit imports list");
579 $^H{charnames} = \&charnames ;
580 $^H{charnames_ord_aliases} = {};
581 $^H{charnames_name_aliases} = {};
582 $^H{charnames_inverse_ords} = {};
583 # New fields must be added to %dummy_H, and the code in lookup_name()
584 # that copies fields from the runtime structure
587 ## fill %h keys with our @_ args.
589 my ($promote, %h, @args) = (0);
590 while (my $arg = shift) {
591 if ($arg eq ":alias") {
593 croak ":alias needs an argument in charnames";
596 ref $alias eq "HASH" or
597 croak "Only HASH reference supported as argument to :alias";
601 if ($alias =~ m{:(\w+)$}) {
602 $1 eq "full" || $1 eq "loose" || $1 eq "short" and
603 croak ":alias cannot use existing pragma :$1 (reversed order?)";
604 alias_file ($1) and $promote = 1;
610 if (substr($arg, 0, 1) eq ':'
611 and ! ($arg eq ":full" || $arg eq ":short" || $arg eq ":loose"))
613 warn "unsupported special '$arg' in charnames";
619 @args == 0 && $promote and @args = (":full");
620 @h{@args} = (1) x @args;
622 # Don't leave these undefined as are tested for in lookup_names
623 $^H{charnames_full} = delete $h{':full'} || 0;
624 $^H{charnames_loose} = delete $h{':loose'} || 0;
625 $^H{charnames_short} = delete $h{':short'} || 0;
626 my @scripts = map { uc quotemeta } keys %h;
629 ## If utf8? warnings are enabled, and some scripts were given,
630 ## see if at least we can find one letter from each script.
632 if (warnings::enabled('utf8') && @scripts) {
633 $txt = do "unicore/Name.pl" unless $txt;
635 for my $script (@scripts) {
636 if (not $txt =~ m/\t$script (?:CAPITAL |SMALL )?LETTER /) {
637 warnings::warn('utf8', "No such script: '$script'");
638 $script = quotemeta $script; # Escape it, for use in the re.
643 # %^H gets stringified, so serialize it ourselves so can extract the
644 # real data back later.
645 $^H{charnames_stringified_ords} = join ",", %{$^H{charnames_ord_aliases}};
646 $^H{charnames_stringified_names} = join ",", %{$^H{charnames_name_aliases}};
647 $^H{charnames_stringified_inverse_ords} = join ",", %{$^H{charnames_inverse_ords}};
649 # Modify the input script names for loose name matching if that is also
650 # specified, similar to the way the base character name is prepared. They
651 # don't (currently, and hopefully never will) have dashes. These go into a
652 # regex, and have already been uppercased and quotemeta'd. Squeeze out all
653 # input underscores, blanks, and dashes. Then convert so will match a blank
654 # between any characters.
655 if ($^H{charnames_loose}) {
656 for (my $i = 0; $i < @scripts; $i++) {
657 $scripts[$i] =~ s/[_ -]//g;
658 $scripts[$i] =~ s/ ( [^\\] ) (?= . ) /$1\\ ?/gx;
662 $^H{charnames_scripts} = join "|", @scripts; # Stringifiy them as a trie
665 # Cache of already looked-up values. This is set to only contain
666 # official values, and user aliases can't override them, so scoping is
672 # Returns the name of the code point argument
675 carp "charnames::viacode() expects one argument";
681 # This is derived from Unicode::UCD, where it is nearly the same as the
682 # function _getcode(), but here it makes sure that even a hex argument
683 # has the proper number of leading zeros, which is critical in
684 # matching against $txt below
685 # Must check if decimal first; see comments at that definition
687 if ($arg =~ $decimal_qr) {
688 $hex = sprintf "%05X", $arg;
689 } elsif ($arg =~ $hex_qr) {
690 # Below is the line that differs from the _getcode() source
691 $hex = sprintf "%05X", hex $1;
693 carp("unexpected arg \"$arg\" to charnames::viacode()");
697 return $viacode{$hex} if exists $viacode{$hex};
701 # If the code point is above the max in the table, there's no point
702 # looking through it. Checking the length first is slightly faster
703 if (length($hex) <= 5 || CORE::hex($hex) <= 0x10FFFF) {
704 $txt = do "unicore/Name.pl" unless $txt;
706 # See if the name is algorithmically determinable.
707 my $algorithmic = charnames::code_point_to_name_special(CORE::hex $hex);
708 if (defined $algorithmic) {
709 $viacode{$hex} = $algorithmic;
713 # Return the official name, if exists. It's unclear to me (khw) at
714 # this juncture if it is better to return a user-defined override, so
715 # leaving it as is for now.
716 if ($txt =~ m/^$hex\t/m) {
718 # The name starts with the next character and goes up to the
719 # next new-line. Using capturing parentheses above instead of
720 # @+ more than doubles the execution time in Perl 5.13
721 $return = substr($txt, $+[0], index($txt, "\n", $+[0]) - $+[0]);
723 # If not one of these 4 code points, return what we've found.
724 if ($hex !~ / ^ 000 (?: 8[014] | 99 ) $ /x) {
725 $viacode{$hex} = $return;
729 # For backwards compatibility, we don't return the official name of
730 # the 4 code points if there are user-defined aliases for them -- so
735 # See if there is a user name for it, before giving up completely.
736 # First get the scoped aliases, give up if have none.
737 my $H_ref = (caller(1))[10];
738 return if ! defined $return
740 || ! exists $H_ref->{charnames_stringified_inverse_ords});
742 my %code_point_aliases;
743 if (defined $H_ref->{charnames_stringified_inverse_ords}) {
744 %code_point_aliases = split ',',
745 $H_ref->{charnames_stringified_inverse_ords};
746 return $code_point_aliases{$hex} if exists $code_point_aliases{$hex};
749 # Here there is no user-defined alias, return any official one.
750 return $return if defined $return;
752 if (CORE::hex($hex) > 0x10FFFF) {
753 carp "Unicode characters only allocated up to U+10FFFF (you asked for U+$hex)";
761 # ex: set ts=8 sts=2 sw=2 et: