3 * Copyright (c) 1998-2002, Larry Wall
5 * You may distribute under the terms of either the GNU General Public
6 * License or the Artistic License, as specified in the README file.
11 * 'What a fix!' said Sam. 'That's the one place in all the lands we've ever
12 * heard of that we don't want to see any closer; and that's the one place
13 * we're trying to get to! And that's just where we can't get, nohow.'
15 * 'Well do I understand your speech,' he answered in the same language;
16 * 'yet few strangers do so. Why then do you not speak in the Common Tongue,
17 * as is the custom in the West, if you wish to be answered?'
19 * ...the travellers perceived that the floor was paved with stones of many
20 * hues; branching runes and strange devices intertwined beneath their feet.
24 #define PERL_IN_UTF8_C
27 static char unees[] = "Malformed UTF-8 character (unexpected end of string)";
30 =head1 Unicode Support
32 =for apidoc A|U8 *|uvuni_to_utf8_flags|U8 *d|UV uv|UV flags
34 Adds the UTF8 representation of the Unicode codepoint C<uv> to the end
35 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
36 bytes available. The return value is the pointer to the byte after the
37 end of the new character. In other words,
39 d = uvuni_to_utf8_flags(d, uv, flags);
43 d = uvuni_to_utf8(d, uv);
45 (which is equivalent to)
47 d = uvuni_to_utf8_flags(d, uv, 0);
49 is the recommended Unicode-aware way of saying
57 Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
59 if (ckWARN(WARN_UTF8)) {
60 if (UNICODE_IS_SURROGATE(uv) &&
61 !(flags & UNICODE_ALLOW_SURROGATE))
62 Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv);
64 ((uv >= 0xFDD0 && uv <= 0xFDEF &&
65 !(flags & UNICODE_ALLOW_FDD0))
67 (UNICODE_IS_BYTE_ORDER_MARK(uv) &&
68 !(flags & UNICODE_ALLOW_BOM))
70 ((uv & 0xFFFF) == 0xFFFF &&
71 !(flags & UNICODE_ALLOW_FFFF))) &&
72 /* UNICODE_ALLOW_SUPER includes
73 * FFFFs beyond 0x10FFFF. */
74 ((uv <= PERL_UNICODE_MAX) ||
75 !(flags & UNICODE_ALLOW_SUPER))
77 Perl_warner(aTHX_ packWARN(WARN_UTF8),
78 "Unicode character 0x%04"UVxf" is illegal", uv);
80 if (UNI_IS_INVARIANT(uv)) {
81 *d++ = UTF_TO_NATIVE(uv);
86 STRLEN len = UNISKIP(uv);
89 *p-- = UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK);
90 uv >>= UTF_ACCUMULATION_SHIFT;
92 *p = UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len));
95 #else /* Non loop style */
97 *d++ = (( uv >> 6) | 0xc0);
98 *d++ = (( uv & 0x3f) | 0x80);
102 *d++ = (( uv >> 12) | 0xe0);
103 *d++ = (((uv >> 6) & 0x3f) | 0x80);
104 *d++ = (( uv & 0x3f) | 0x80);
108 *d++ = (( uv >> 18) | 0xf0);
109 *d++ = (((uv >> 12) & 0x3f) | 0x80);
110 *d++ = (((uv >> 6) & 0x3f) | 0x80);
111 *d++ = (( uv & 0x3f) | 0x80);
114 if (uv < 0x4000000) {
115 *d++ = (( uv >> 24) | 0xf8);
116 *d++ = (((uv >> 18) & 0x3f) | 0x80);
117 *d++ = (((uv >> 12) & 0x3f) | 0x80);
118 *d++ = (((uv >> 6) & 0x3f) | 0x80);
119 *d++ = (( uv & 0x3f) | 0x80);
122 if (uv < 0x80000000) {
123 *d++ = (( uv >> 30) | 0xfc);
124 *d++ = (((uv >> 24) & 0x3f) | 0x80);
125 *d++ = (((uv >> 18) & 0x3f) | 0x80);
126 *d++ = (((uv >> 12) & 0x3f) | 0x80);
127 *d++ = (((uv >> 6) & 0x3f) | 0x80);
128 *d++ = (( uv & 0x3f) | 0x80);
132 if (uv < UTF8_QUAD_MAX)
135 *d++ = 0xfe; /* Can't match U+FEFF! */
136 *d++ = (((uv >> 30) & 0x3f) | 0x80);
137 *d++ = (((uv >> 24) & 0x3f) | 0x80);
138 *d++ = (((uv >> 18) & 0x3f) | 0x80);
139 *d++ = (((uv >> 12) & 0x3f) | 0x80);
140 *d++ = (((uv >> 6) & 0x3f) | 0x80);
141 *d++ = (( uv & 0x3f) | 0x80);
146 *d++ = 0xff; /* Can't match U+FFFE! */
147 *d++ = 0x80; /* 6 Reserved bits */
148 *d++ = (((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */
149 *d++ = (((uv >> 54) & 0x3f) | 0x80);
150 *d++ = (((uv >> 48) & 0x3f) | 0x80);
151 *d++ = (((uv >> 42) & 0x3f) | 0x80);
152 *d++ = (((uv >> 36) & 0x3f) | 0x80);
153 *d++ = (((uv >> 30) & 0x3f) | 0x80);
154 *d++ = (((uv >> 24) & 0x3f) | 0x80);
155 *d++ = (((uv >> 18) & 0x3f) | 0x80);
156 *d++ = (((uv >> 12) & 0x3f) | 0x80);
157 *d++ = (((uv >> 6) & 0x3f) | 0x80);
158 *d++ = (( uv & 0x3f) | 0x80);
162 #endif /* Loop style */
166 Perl_uvuni_to_utf8(pTHX_ U8 *d, UV uv)
168 return Perl_uvuni_to_utf8_flags(aTHX_ d, uv, 0);
173 =for apidoc A|STRLEN|is_utf8_char|U8 *s
175 Tests if some arbitrary number of bytes begins in a valid UTF-8
176 character. Note that an INVARIANT (i.e. ASCII) character is a valid UTF-8 character.
177 The actual number of bytes in the UTF-8 character will be returned if
178 it is valid, otherwise 0.
183 Perl_is_utf8_char(pTHX_ U8 *s)
189 if (UTF8_IS_INVARIANT(u))
192 if (!UTF8_IS_START(u))
197 if (len < 2 || !UTF8_IS_CONTINUATION(s[1]))
202 u &= UTF_START_MASK(len);
206 if (!UTF8_IS_CONTINUATION(*s))
208 uv = UTF8_ACCUMULATE(uv, *s);
215 if (UNISKIP(uv) < len)
222 =for apidoc A|bool|is_utf8_string|U8 *s|STRLEN len
224 Returns true if first C<len> bytes of the given string form a valid UTF8
225 string, false otherwise. Note that 'a valid UTF8 string' does not mean
226 'a string that contains UTF8' because a valid ASCII string is a valid
233 Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len)
240 len = strlen((char *)s);
256 =for apidoc A|UV|utf8n_to_uvuni|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
258 Bottom level UTF-8 decode routine.
259 Returns the unicode code point value of the first character in the string C<s>
260 which is assumed to be in UTF8 encoding and no longer than C<curlen>;
261 C<retlen> will be set to the length, in bytes, of that character.
263 If C<s> does not point to a well-formed UTF8 character, the behaviour
264 is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY,
265 it is assumed that the caller will raise a warning, and this function
266 will silently just set C<retlen> to C<-1> and return zero. If the
267 C<flags> does not contain UTF8_CHECK_ONLY, warnings about
268 malformations will be given, C<retlen> will be set to the expected
269 length of the UTF-8 character in bytes, and zero will be returned.
271 The C<flags> can also contain various flags to allow deviations from
272 the strict UTF-8 encoding (see F<utf8.h>).
274 Most code should use utf8_to_uvchr() rather than call this directly.
280 Perl_utf8n_to_uvuni(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
285 bool dowarn = ckWARN_d(WARN_UTF8);
287 STRLEN expectlen = 0;
290 /* This list is a superset of the UTF8_ALLOW_XXX. */
292 #define UTF8_WARN_EMPTY 1
293 #define UTF8_WARN_CONTINUATION 2
294 #define UTF8_WARN_NON_CONTINUATION 3
295 #define UTF8_WARN_FE_FF 4
296 #define UTF8_WARN_SHORT 5
297 #define UTF8_WARN_OVERFLOW 6
298 #define UTF8_WARN_SURROGATE 7
299 #define UTF8_WARN_BOM 8
300 #define UTF8_WARN_LONG 9
301 #define UTF8_WARN_FFFF 10
304 !(flags & UTF8_ALLOW_EMPTY)) {
305 warning = UTF8_WARN_EMPTY;
309 if (UTF8_IS_INVARIANT(uv)) {
312 return (UV) (NATIVE_TO_UTF(*s));
315 if (UTF8_IS_CONTINUATION(uv) &&
316 !(flags & UTF8_ALLOW_CONTINUATION)) {
317 warning = UTF8_WARN_CONTINUATION;
321 if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) &&
322 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
323 warning = UTF8_WARN_NON_CONTINUATION;
328 uv = NATIVE_TO_UTF(uv);
330 if ((uv == 0xfe || uv == 0xff) &&
331 !(flags & UTF8_ALLOW_FE_FF)) {
332 warning = UTF8_WARN_FE_FF;
337 if (!(uv & 0x20)) { len = 2; uv &= 0x1f; }
338 else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; }
339 else if (!(uv & 0x08)) { len = 4; uv &= 0x07; }
340 else if (!(uv & 0x04)) { len = 5; uv &= 0x03; }
342 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
343 else { len = 7; uv &= 0x01; }
345 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
346 else if (!(uv & 0x01)) { len = 7; uv = 0; }
347 else { len = 13; uv = 0; } /* whoa! */
355 if ((curlen < expectlen) &&
356 !(flags & UTF8_ALLOW_SHORT)) {
357 warning = UTF8_WARN_SHORT;
366 if (!UTF8_IS_CONTINUATION(*s) &&
367 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
369 warning = UTF8_WARN_NON_CONTINUATION;
373 uv = UTF8_ACCUMULATE(uv, *s);
375 /* These cannot be allowed. */
377 if (!(flags & UTF8_ALLOW_LONG)) {
378 warning = UTF8_WARN_LONG;
382 else { /* uv < ouv */
383 /* This cannot be allowed. */
384 warning = UTF8_WARN_OVERFLOW;
392 if (UNICODE_IS_SURROGATE(uv) &&
393 !(flags & UTF8_ALLOW_SURROGATE)) {
394 warning = UTF8_WARN_SURROGATE;
396 } else if (UNICODE_IS_BYTE_ORDER_MARK(uv) &&
397 !(flags & UTF8_ALLOW_BOM)) {
398 warning = UTF8_WARN_BOM;
400 } else if ((expectlen > UNISKIP(uv)) &&
401 !(flags & UTF8_ALLOW_LONG)) {
402 warning = UTF8_WARN_LONG;
404 } else if (UNICODE_IS_ILLEGAL(uv) &&
405 !(flags & UTF8_ALLOW_FFFF)) {
406 warning = UTF8_WARN_FFFF;
414 if (flags & UTF8_CHECK_ONLY) {
421 SV* sv = sv_2mortal(newSVpv("Malformed UTF-8 character ", 0));
424 case 0: /* Intentionally empty. */ break;
425 case UTF8_WARN_EMPTY:
426 Perl_sv_catpvf(aTHX_ sv, "(empty string)");
428 case UTF8_WARN_CONTINUATION:
429 Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv);
431 case UTF8_WARN_NON_CONTINUATION:
433 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")",
434 (UV)s[1], startbyte);
436 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)",
437 (UV)s[1], s - s0, s - s0 > 1 ? "s" : "", startbyte, expectlen);
440 case UTF8_WARN_FE_FF:
441 Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv);
443 case UTF8_WARN_SHORT:
444 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
445 curlen, curlen == 1 ? "" : "s", expectlen, startbyte);
446 expectlen = curlen; /* distance for caller to skip */
448 case UTF8_WARN_OVERFLOW:
449 Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")",
452 case UTF8_WARN_SURROGATE:
453 Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv);
456 Perl_sv_catpvf(aTHX_ sv, "(byte order mark 0x%04"UVxf")", uv);
459 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
460 expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte);
463 Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv);
466 Perl_sv_catpvf(aTHX_ sv, "(unknown reason)");
474 Perl_warner(aTHX_ packWARN(WARN_UTF8),
475 "%s in %s", s, OP_DESC(PL_op));
477 Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s);
482 *retlen = expectlen ? expectlen : len;
488 =for apidoc A|UV|utf8_to_uvchr|U8 *s|STRLEN *retlen
490 Returns the native character value of the first character in the string C<s>
491 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
492 length, in bytes, of that character.
494 If C<s> does not point to a well-formed UTF8 character, zero is
495 returned and retlen is set, if possible, to -1.
501 Perl_utf8_to_uvchr(pTHX_ U8 *s, STRLEN *retlen)
503 return Perl_utf8n_to_uvchr(aTHX_ s, UTF8_MAXLEN, retlen,
504 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
508 =for apidoc A|UV|utf8_to_uvuni|U8 *s|STRLEN *retlen
510 Returns the Unicode code point of the first character in the string C<s>
511 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
512 length, in bytes, of that character.
514 This function should only be used when returned UV is considered
515 an index into the Unicode semantic tables (e.g. swashes).
517 If C<s> does not point to a well-formed UTF8 character, zero is
518 returned and retlen is set, if possible, to -1.
524 Perl_utf8_to_uvuni(pTHX_ U8 *s, STRLEN *retlen)
526 /* Call the low level routine asking for checks */
527 return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXLEN, retlen,
528 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
532 =for apidoc A|STRLEN|utf8_length|U8 *s|U8 *e
534 Return the length of the UTF-8 char encoded string C<s> in characters.
535 Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end
536 up past C<e>, croaks.
542 Perl_utf8_length(pTHX_ U8 *s, U8 *e)
546 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
547 * the bitops (especially ~) can create illegal UTF-8.
548 * In other words: in Perl UTF-8 is not just for Unicode. */
551 if (ckWARN_d(WARN_UTF8)) {
553 Perl_warner(aTHX_ packWARN(WARN_UTF8),
554 "%s in %s", unees, OP_DESC(PL_op));
556 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
564 if (ckWARN_d(WARN_UTF8)) {
566 Perl_warner(aTHX_ packWARN(WARN_UTF8),
567 unees, OP_DESC(PL_op));
569 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
581 =for apidoc A|IV|utf8_distance|U8 *a|U8 *b
583 Returns the number of UTF8 characters between the UTF-8 pointers C<a>
586 WARNING: use only if you *know* that the pointers point inside the
593 Perl_utf8_distance(pTHX_ U8 *a, U8 *b)
597 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
598 * the bitops (especially ~) can create illegal UTF-8.
599 * In other words: in Perl UTF-8 is not just for Unicode. */
606 if (ckWARN_d(WARN_UTF8)) {
608 Perl_warner(aTHX_ packWARN(WARN_UTF8),
609 "%s in %s", unees, OP_DESC(PL_op));
611 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
624 if (ckWARN_d(WARN_UTF8)) {
626 Perl_warner(aTHX_ packWARN(WARN_UTF8),
627 "%s in %s", unees, OP_DESC(PL_op));
629 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
642 =for apidoc A|U8 *|utf8_hop|U8 *s|I32 off
644 Return the UTF-8 pointer C<s> displaced by C<off> characters, either
647 WARNING: do not use the following unless you *know* C<off> is within
648 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
649 on the first byte of character or just after the last byte of a character.
655 Perl_utf8_hop(pTHX_ U8 *s, I32 off)
657 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
658 * the bitops (especially ~) can create illegal UTF-8.
659 * In other words: in Perl UTF-8 is not just for Unicode. */
668 while (UTF8_IS_CONTINUATION(*s))
676 =for apidoc A|U8 *|utf8_to_bytes|U8 *s|STRLEN *len
678 Converts a string C<s> of length C<len> from UTF8 into byte encoding.
679 Unlike C<bytes_to_utf8>, this over-writes the original string, and
680 updates len to contain the new length.
681 Returns zero on failure, setting C<len> to -1.
687 Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len)
693 /* ensure valid UTF8 and chars < 256 before updating string */
694 for (send = s + *len; s < send; ) {
697 if (!UTF8_IS_INVARIANT(c) &&
698 (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send)
699 || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) {
708 *d++ = (U8)utf8_to_uvchr(s, &ulen);
717 =for apidoc A|U8 *|bytes_from_utf8|U8 *s|STRLEN *len|bool *is_utf8
719 Converts a string C<s> of length C<len> from UTF8 into byte encoding.
720 Unlike <utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to
721 the newly-created string, and updates C<len> to contain the new
722 length. Returns the original string if no conversion occurs, C<len>
723 is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to
724 0 if C<s> is converted or contains all 7bit characters.
730 Perl_bytes_from_utf8(pTHX_ U8 *s, STRLEN *len, bool *is_utf8)
740 /* ensure valid UTF8 and chars < 256 before converting string */
741 for (send = s + *len; s < send;) {
743 if (!UTF8_IS_INVARIANT(c)) {
744 if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send &&
745 (c = *s++) && UTF8_IS_CONTINUATION(c))
754 Newz(801, d, (*len) - count + 1, U8);
755 s = start; start = d;
758 if (!UTF8_IS_INVARIANT(c)) {
759 /* Then it is two-byte encoded */
760 c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++);
761 c = ASCII_TO_NATIVE(c);
771 =for apidoc A|U8 *|bytes_to_utf8|U8 *s|STRLEN *len
773 Converts a string C<s> of length C<len> from ASCII into UTF8 encoding.
774 Returns a pointer to the newly-created string, and sets C<len> to
775 reflect the new length.
781 Perl_bytes_to_utf8(pTHX_ U8 *s, STRLEN *len)
788 Newz(801, d, (*len) * 2 + 1, U8);
792 UV uv = NATIVE_TO_ASCII(*s++);
793 if (UNI_IS_INVARIANT(uv))
794 *d++ = UTF_TO_NATIVE(uv);
796 *d++ = UTF8_EIGHT_BIT_HI(uv);
797 *d++ = UTF8_EIGHT_BIT_LO(uv);
806 * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8.
808 * Destination must be pre-extended to 3/2 source. Do not use in-place.
809 * We optimize for native, for obvious reasons. */
812 Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
818 Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen");
823 UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */
830 *d++ = (( uv >> 6) | 0xc0);
831 *d++ = (( uv & 0x3f) | 0x80);
834 if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */
836 if (low < 0xdc00 || low >= 0xdfff)
837 Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
838 uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000;
841 *d++ = (( uv >> 12) | 0xe0);
842 *d++ = (((uv >> 6) & 0x3f) | 0x80);
843 *d++ = (( uv & 0x3f) | 0x80);
847 *d++ = (( uv >> 18) | 0xf0);
848 *d++ = (((uv >> 12) & 0x3f) | 0x80);
849 *d++ = (((uv >> 6) & 0x3f) | 0x80);
850 *d++ = (( uv & 0x3f) | 0x80);
854 *newlen = d - dstart;
858 /* Note: this one is slightly destructive of the source. */
861 Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
864 U8* send = s + bytelen;
871 return utf16_to_utf8(p, d, bytelen, newlen);
874 /* for now these are all defined (inefficiently) in terms of the utf8 versions */
877 Perl_is_uni_alnum(pTHX_ UV c)
879 U8 tmpbuf[UTF8_MAXLEN+1];
880 uvchr_to_utf8(tmpbuf, c);
881 return is_utf8_alnum(tmpbuf);
885 Perl_is_uni_alnumc(pTHX_ UV c)
887 U8 tmpbuf[UTF8_MAXLEN+1];
888 uvchr_to_utf8(tmpbuf, c);
889 return is_utf8_alnumc(tmpbuf);
893 Perl_is_uni_idfirst(pTHX_ UV c)
895 U8 tmpbuf[UTF8_MAXLEN+1];
896 uvchr_to_utf8(tmpbuf, c);
897 return is_utf8_idfirst(tmpbuf);
901 Perl_is_uni_alpha(pTHX_ UV c)
903 U8 tmpbuf[UTF8_MAXLEN+1];
904 uvchr_to_utf8(tmpbuf, c);
905 return is_utf8_alpha(tmpbuf);
909 Perl_is_uni_ascii(pTHX_ UV c)
911 U8 tmpbuf[UTF8_MAXLEN+1];
912 uvchr_to_utf8(tmpbuf, c);
913 return is_utf8_ascii(tmpbuf);
917 Perl_is_uni_space(pTHX_ UV c)
919 U8 tmpbuf[UTF8_MAXLEN+1];
920 uvchr_to_utf8(tmpbuf, c);
921 return is_utf8_space(tmpbuf);
925 Perl_is_uni_digit(pTHX_ UV c)
927 U8 tmpbuf[UTF8_MAXLEN+1];
928 uvchr_to_utf8(tmpbuf, c);
929 return is_utf8_digit(tmpbuf);
933 Perl_is_uni_upper(pTHX_ UV c)
935 U8 tmpbuf[UTF8_MAXLEN+1];
936 uvchr_to_utf8(tmpbuf, c);
937 return is_utf8_upper(tmpbuf);
941 Perl_is_uni_lower(pTHX_ UV c)
943 U8 tmpbuf[UTF8_MAXLEN+1];
944 uvchr_to_utf8(tmpbuf, c);
945 return is_utf8_lower(tmpbuf);
949 Perl_is_uni_cntrl(pTHX_ UV c)
951 U8 tmpbuf[UTF8_MAXLEN+1];
952 uvchr_to_utf8(tmpbuf, c);
953 return is_utf8_cntrl(tmpbuf);
957 Perl_is_uni_graph(pTHX_ UV c)
959 U8 tmpbuf[UTF8_MAXLEN+1];
960 uvchr_to_utf8(tmpbuf, c);
961 return is_utf8_graph(tmpbuf);
965 Perl_is_uni_print(pTHX_ UV c)
967 U8 tmpbuf[UTF8_MAXLEN+1];
968 uvchr_to_utf8(tmpbuf, c);
969 return is_utf8_print(tmpbuf);
973 Perl_is_uni_punct(pTHX_ UV c)
975 U8 tmpbuf[UTF8_MAXLEN+1];
976 uvchr_to_utf8(tmpbuf, c);
977 return is_utf8_punct(tmpbuf);
981 Perl_is_uni_xdigit(pTHX_ UV c)
983 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
984 uvchr_to_utf8(tmpbuf, c);
985 return is_utf8_xdigit(tmpbuf);
989 Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp)
992 return to_utf8_upper(p, p, lenp);
996 Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp)
999 return to_utf8_title(p, p, lenp);
1003 Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp)
1005 uvchr_to_utf8(p, c);
1006 return to_utf8_lower(p, p, lenp);
1010 Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp)
1012 uvchr_to_utf8(p, c);
1013 return to_utf8_fold(p, p, lenp);
1016 /* for now these all assume no locale info available for Unicode > 255 */
1019 Perl_is_uni_alnum_lc(pTHX_ UV c)
1021 return is_uni_alnum(c); /* XXX no locale support yet */
1025 Perl_is_uni_alnumc_lc(pTHX_ UV c)
1027 return is_uni_alnumc(c); /* XXX no locale support yet */
1031 Perl_is_uni_idfirst_lc(pTHX_ UV c)
1033 return is_uni_idfirst(c); /* XXX no locale support yet */
1037 Perl_is_uni_alpha_lc(pTHX_ UV c)
1039 return is_uni_alpha(c); /* XXX no locale support yet */
1043 Perl_is_uni_ascii_lc(pTHX_ UV c)
1045 return is_uni_ascii(c); /* XXX no locale support yet */
1049 Perl_is_uni_space_lc(pTHX_ UV c)
1051 return is_uni_space(c); /* XXX no locale support yet */
1055 Perl_is_uni_digit_lc(pTHX_ UV c)
1057 return is_uni_digit(c); /* XXX no locale support yet */
1061 Perl_is_uni_upper_lc(pTHX_ UV c)
1063 return is_uni_upper(c); /* XXX no locale support yet */
1067 Perl_is_uni_lower_lc(pTHX_ UV c)
1069 return is_uni_lower(c); /* XXX no locale support yet */
1073 Perl_is_uni_cntrl_lc(pTHX_ UV c)
1075 return is_uni_cntrl(c); /* XXX no locale support yet */
1079 Perl_is_uni_graph_lc(pTHX_ UV c)
1081 return is_uni_graph(c); /* XXX no locale support yet */
1085 Perl_is_uni_print_lc(pTHX_ UV c)
1087 return is_uni_print(c); /* XXX no locale support yet */
1091 Perl_is_uni_punct_lc(pTHX_ UV c)
1093 return is_uni_punct(c); /* XXX no locale support yet */
1097 Perl_is_uni_xdigit_lc(pTHX_ UV c)
1099 return is_uni_xdigit(c); /* XXX no locale support yet */
1103 Perl_to_uni_upper_lc(pTHX_ U32 c)
1105 /* XXX returns only the first character -- do not use XXX */
1106 /* XXX no locale support yet */
1108 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1109 return (U32)to_uni_upper(c, tmpbuf, &len);
1113 Perl_to_uni_title_lc(pTHX_ U32 c)
1115 /* XXX returns only the first character XXX -- do not use XXX */
1116 /* XXX no locale support yet */
1118 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1119 return (U32)to_uni_title(c, tmpbuf, &len);
1123 Perl_to_uni_lower_lc(pTHX_ U32 c)
1125 /* XXX returns only the first character -- do not use XXX */
1126 /* XXX no locale support yet */
1128 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1129 return (U32)to_uni_lower(c, tmpbuf, &len);
1133 Perl_is_utf8_alnum(pTHX_ U8 *p)
1135 if (!is_utf8_char(p))
1138 /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true
1139 * descendant of isalnum(3), in other words, it doesn't
1140 * contain the '_'. --jhi */
1141 PL_utf8_alnum = swash_init("utf8", "IsWord", &PL_sv_undef, 0, 0);
1142 return swash_fetch(PL_utf8_alnum, p, TRUE);
1143 /* return *p == '_' || is_utf8_alpha(p) || is_utf8_digit(p); */
1144 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1146 PL_utf8_alnum = swash_init("utf8", "",
1147 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1148 return swash_fetch(PL_utf8_alnum, p, TRUE);
1153 Perl_is_utf8_alnumc(pTHX_ U8 *p)
1155 if (!is_utf8_char(p))
1158 PL_utf8_alnum = swash_init("utf8", "IsAlnumC", &PL_sv_undef, 0, 0);
1159 return swash_fetch(PL_utf8_alnum, p, TRUE);
1160 /* return is_utf8_alpha(p) || is_utf8_digit(p); */
1161 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1163 PL_utf8_alnum = swash_init("utf8", "",
1164 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1165 return swash_fetch(PL_utf8_alnum, p, TRUE);
1170 Perl_is_utf8_idfirst(pTHX_ U8 *p)
1172 return *p == '_' || is_utf8_alpha(p);
1176 Perl_is_utf8_alpha(pTHX_ U8 *p)
1178 if (!is_utf8_char(p))
1181 PL_utf8_alpha = swash_init("utf8", "IsAlpha", &PL_sv_undef, 0, 0);
1182 return swash_fetch(PL_utf8_alpha, p, TRUE);
1186 Perl_is_utf8_ascii(pTHX_ U8 *p)
1188 if (!is_utf8_char(p))
1191 PL_utf8_ascii = swash_init("utf8", "IsAscii", &PL_sv_undef, 0, 0);
1192 return swash_fetch(PL_utf8_ascii, p, TRUE);
1196 Perl_is_utf8_space(pTHX_ U8 *p)
1198 if (!is_utf8_char(p))
1201 PL_utf8_space = swash_init("utf8", "IsSpacePerl", &PL_sv_undef, 0, 0);
1202 return swash_fetch(PL_utf8_space, p, TRUE);
1206 Perl_is_utf8_digit(pTHX_ U8 *p)
1208 if (!is_utf8_char(p))
1211 PL_utf8_digit = swash_init("utf8", "IsDigit", &PL_sv_undef, 0, 0);
1212 return swash_fetch(PL_utf8_digit, p, TRUE);
1216 Perl_is_utf8_upper(pTHX_ U8 *p)
1218 if (!is_utf8_char(p))
1221 PL_utf8_upper = swash_init("utf8", "IsUpper", &PL_sv_undef, 0, 0);
1222 return swash_fetch(PL_utf8_upper, p, TRUE);
1226 Perl_is_utf8_lower(pTHX_ U8 *p)
1228 if (!is_utf8_char(p))
1231 PL_utf8_lower = swash_init("utf8", "IsLower", &PL_sv_undef, 0, 0);
1232 return swash_fetch(PL_utf8_lower, p, TRUE);
1236 Perl_is_utf8_cntrl(pTHX_ U8 *p)
1238 if (!is_utf8_char(p))
1241 PL_utf8_cntrl = swash_init("utf8", "IsCntrl", &PL_sv_undef, 0, 0);
1242 return swash_fetch(PL_utf8_cntrl, p, TRUE);
1246 Perl_is_utf8_graph(pTHX_ U8 *p)
1248 if (!is_utf8_char(p))
1251 PL_utf8_graph = swash_init("utf8", "IsGraph", &PL_sv_undef, 0, 0);
1252 return swash_fetch(PL_utf8_graph, p, TRUE);
1256 Perl_is_utf8_print(pTHX_ U8 *p)
1258 if (!is_utf8_char(p))
1261 PL_utf8_print = swash_init("utf8", "IsPrint", &PL_sv_undef, 0, 0);
1262 return swash_fetch(PL_utf8_print, p, TRUE);
1266 Perl_is_utf8_punct(pTHX_ U8 *p)
1268 if (!is_utf8_char(p))
1271 PL_utf8_punct = swash_init("utf8", "IsPunct", &PL_sv_undef, 0, 0);
1272 return swash_fetch(PL_utf8_punct, p, TRUE);
1276 Perl_is_utf8_xdigit(pTHX_ U8 *p)
1278 if (!is_utf8_char(p))
1280 if (!PL_utf8_xdigit)
1281 PL_utf8_xdigit = swash_init("utf8", "IsXDigit", &PL_sv_undef, 0, 0);
1282 return swash_fetch(PL_utf8_xdigit, p, TRUE);
1286 Perl_is_utf8_mark(pTHX_ U8 *p)
1288 if (!is_utf8_char(p))
1291 PL_utf8_mark = swash_init("utf8", "IsM", &PL_sv_undef, 0, 0);
1292 return swash_fetch(PL_utf8_mark, p, TRUE);
1296 =for apidoc A|UV|to_utf8_case|U8 *p|U8* ustrp|STRLEN *lenp|SV **swash|char *normal|char *special
1298 The "p" contains the pointer to the UTF-8 string encoding
1299 the character that is being converted.
1301 The "ustrp" is a pointer to the character buffer to put the
1302 conversion result to. The "lenp" is a pointer to the length
1305 The "swashp" is a pointer to the swash to use.
1307 Both the special and normal mappings are stored lib/unicore/To/Foo.pl,
1308 and loaded by SWASHGET, using lib/utf8_heavy.pl. The special (usually,
1309 but not always, a multicharacter mapping), is tried first.
1311 The "special" is a string like "utf8::ToSpecLower", which means the
1312 hash %utf8::ToSpecLower. The access to the hash is through
1313 Perl_to_utf8_case().
1315 The "normal" is a string like "ToLower" which means the swash
1321 Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp, char *normal, char *special)
1324 U8 tmpbuf[UTF8_MAXLEN_FOLD+1];
1327 uv0 = utf8_to_uvchr(p, 0);
1328 /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings
1329 * are necessary in EBCDIC, they are redundant no-ops
1330 * in ASCII-ish platforms, and hopefully optimized away. */
1331 uv1 = NATIVE_TO_UNI(uv0);
1332 uvuni_to_utf8(tmpbuf, uv1);
1334 if (!*swashp) /* load on-demand */
1335 *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0);
1338 /* It might be "special" (sometimes, but not always,
1339 * a multicharacter mapping) */
1345 if ((hv = get_hv(special, FALSE)) &&
1346 (keysv = sv_2mortal(Perl_newSVpvf(aTHX_ "%04"UVXf, uv1))) &&
1347 (he = hv_fetch_ent(hv, keysv, FALSE, 0)) &&
1348 (val = HeVAL(he))) {
1353 len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp;
1356 /* If we have EBCDIC we need to remap the characters
1357 * since any characters in the low 256 are Unicode
1358 * code points, not EBCDIC. */
1359 U8 *t = (U8*)s, *tend = t + len, *d;
1366 UV c = utf8_to_uvchr(t, &tlen);
1368 d = uvchr_to_utf8(d, UNI_TO_NATIVE(c));
1377 d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t));
1382 Copy(tmpbuf, ustrp, len, U8);
1384 Copy(s, ustrp, len, U8);
1390 if (!len && *swashp) {
1391 UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE);
1394 /* It was "normal" (a single character mapping). */
1395 UV uv3 = UNI_TO_NATIVE(uv2);
1397 len = uvchr_to_utf8(ustrp, uv3) - ustrp;
1401 if (!len) /* Neither: just copy. */
1402 len = uvchr_to_utf8(ustrp, uv0) - ustrp;
1407 return len ? utf8_to_uvchr(ustrp, 0) : 0;
1411 =for apidoc A|UV|to_utf8_upper|U8 *p|U8 *ustrp|STRLEN *lenp
1413 Convert the UTF-8 encoded character at p to its uppercase version and
1414 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1415 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1416 uppercase version may be longer than the original character (up to two
1419 The first character of the uppercased version is returned
1420 (but note, as explained above, that there may be more.)
1425 Perl_to_utf8_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1427 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1428 &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper");
1432 =for apidoc A|UV|to_utf8_title|U8 *p|U8 *ustrp|STRLEN *lenp
1434 Convert the UTF-8 encoded character at p to its titlecase version and
1435 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1436 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1437 titlecase version may be longer than the original character (up to two
1440 The first character of the titlecased version is returned
1441 (but note, as explained above, that there may be more.)
1446 Perl_to_utf8_title(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1448 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1449 &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle");
1453 =for apidoc A|UV|to_utf8_lower|U8 *p|U8 *ustrp|STRLEN *lenp
1455 Convert the UTF-8 encoded character at p to its lowercase version and
1456 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1457 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1458 lowercase version may be longer than the original character (up to two
1461 The first character of the lowercased version is returned
1462 (but note, as explained above, that there may be more.)
1467 Perl_to_utf8_lower(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1469 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1470 &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower");
1474 =for apidoc A|UV|to_utf8_fold|U8 *p|U8 *ustrp|STRLEN *lenp
1476 Convert the UTF-8 encoded character at p to its foldcase version and
1477 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1478 that the ustrp needs to be at least UTF8_MAXLEN_FOLD+1 bytes since the
1479 foldcase version may be longer than the original character (up to
1482 The first character of the foldcased version is returned
1483 (but note, as explained above, that there may be more.)
1488 Perl_to_utf8_fold(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1490 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1491 &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold");
1494 /* a "swash" is a swatch hash */
1497 Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none)
1500 SV* tokenbufsv = sv_2mortal(NEWSV(0,0));
1502 HV *stash = gv_stashpvn(pkg, strlen(pkg), FALSE);
1505 if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */
1507 errsv_save = newSVsv(ERRSV);
1508 Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpv(pkg,0), Nullsv);
1510 sv_setsv(ERRSV, errsv_save);
1511 SvREFCNT_dec(errsv_save);
1515 PUSHSTACKi(PERLSI_MAGIC);
1518 PUSHs(sv_2mortal(newSVpvn(pkg, strlen(pkg))));
1519 PUSHs(sv_2mortal(newSVpvn(name, strlen(name))));
1521 PUSHs(sv_2mortal(newSViv(minbits)));
1522 PUSHs(sv_2mortal(newSViv(none)));
1528 if (PL_curcop == &PL_compiling)
1529 /* XXX ought to be handled by lex_start */
1530 sv_setpv(tokenbufsv, PL_tokenbuf);
1531 errsv_save = newSVsv(ERRSV);
1532 if (call_method("SWASHNEW", G_SCALAR))
1533 retval = newSVsv(*PL_stack_sp--);
1535 retval = &PL_sv_undef;
1537 sv_setsv(ERRSV, errsv_save);
1538 SvREFCNT_dec(errsv_save);
1541 if (PL_curcop == &PL_compiling) {
1543 char* pv = SvPV(tokenbufsv, len);
1545 Copy(pv, PL_tokenbuf, len+1, char);
1546 PL_curcop->op_private = PL_hints;
1548 if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) {
1550 Perl_croak(aTHX_ "Can't find Unicode property definition \"%s\"",
1551 SvPV_nolen(retval));
1552 Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref");
1558 /* This API is wrong for special case conversions since we may need to
1559 * return several Unicode characters for a single Unicode character
1560 * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is
1561 * the lower-level routine, and it is similarly broken for returning
1562 * multiple values. --jhi */
1564 Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr, bool do_utf8)
1566 HV* hv = (HV*)SvRV(sv);
1575 UV c = NATIVE_TO_ASCII(*ptr);
1577 if (!do_utf8 && !UNI_IS_INVARIANT(c)) {
1578 tmputf8[0] = UTF8_EIGHT_BIT_HI(c);
1579 tmputf8[1] = UTF8_EIGHT_BIT_LO(c);
1582 /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ
1583 * then the "swatch" is a vec() for al the chars which start
1585 * So the key in the hash (klen) is length of encoded char -1
1587 klen = UTF8SKIP(ptr) - 1;
1592 /* If char in invariant then swatch is for all the invariant chars
1593 * In both UTF-8 and UTF8-MOD that happens to be UTF_CONTINUATION_MARK
1595 needents = UTF_CONTINUATION_MARK;
1596 off = NATIVE_TO_UTF(ptr[klen]);
1600 /* If char is encoded then swatch is for the prefix */
1601 needents = (1 << UTF_ACCUMULATION_SHIFT);
1602 off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK;
1606 * This single-entry cache saves about 1/3 of the utf8 overhead in test
1607 * suite. (That is, only 7-8% overall over just a hash cache. Still,
1608 * it's nothing to sniff at.) Pity we usually come through at least
1609 * two function calls to get here...
1611 * NB: this code assumes that swatches are never modified, once generated!
1614 if (hv == PL_last_swash_hv &&
1615 klen == PL_last_swash_klen &&
1616 (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) )
1618 tmps = PL_last_swash_tmps;
1619 slen = PL_last_swash_slen;
1622 /* Try our second-level swatch cache, kept in a hash. */
1623 SV** svp = hv_fetch(hv, (char*)ptr, klen, FALSE);
1625 /* If not cached, generate it via utf8::SWASHGET */
1626 if (!svp || !SvPOK(*svp) || !(tmps = (U8*)SvPV(*svp, slen))) {
1628 /* We use utf8n_to_uvuni() as we want an index into
1629 Unicode tables, not a native character number.
1631 UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXLEN, 0,
1633 0 : UTF8_ALLOW_ANY);
1638 PUSHSTACKi(PERLSI_MAGIC);
1642 /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */
1643 PUSHs(sv_2mortal(newSViv((klen) ?
1644 (code_point & ~(needents - 1)) : 0)));
1645 PUSHs(sv_2mortal(newSViv(needents)));
1647 errsv_save = newSVsv(ERRSV);
1648 if (call_method("SWASHGET", G_SCALAR))
1649 retval = newSVsv(*PL_stack_sp--);
1651 retval = &PL_sv_undef;
1653 sv_setsv(ERRSV, errsv_save);
1654 SvREFCNT_dec(errsv_save);
1658 if (PL_curcop == &PL_compiling)
1659 PL_curcop->op_private = PL_hints;
1661 svp = hv_store(hv, (char*)ptr, klen, retval, 0);
1663 if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || (slen << 3) < needents)
1664 Perl_croak(aTHX_ "SWASHGET didn't return result of proper length");
1667 PL_last_swash_hv = hv;
1668 PL_last_swash_klen = klen;
1669 PL_last_swash_tmps = tmps;
1670 PL_last_swash_slen = slen;
1672 Copy(ptr, PL_last_swash_key, klen, U8);
1675 switch ((int)((slen << 3) / needents)) {
1677 bit = 1 << (off & 7);
1679 return (tmps[off] & bit) != 0;
1684 return (tmps[off] << 8) + tmps[off + 1] ;
1687 return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ;
1689 Perl_croak(aTHX_ "panic: swash_fetch");
1695 =for apidoc A|U8 *|uvchr_to_utf8|U8 *d|UV uv
1697 Adds the UTF8 representation of the Native codepoint C<uv> to the end
1698 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
1699 bytes available. The return value is the pointer to the byte after the
1700 end of the new character. In other words,
1702 d = uvchr_to_utf8(d, uv);
1704 is the recommended wide native character-aware way of saying
1711 /* On ASCII machines this is normally a macro but we want a
1712 real function in case XS code wants it
1714 #undef Perl_uvchr_to_utf8
1716 Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv)
1718 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0);
1722 Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
1724 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags);
1728 =for apidoc A|UV|utf8n_to_uvchr|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
1730 Returns the native character value of the first character in the string C<s>
1731 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
1732 length, in bytes, of that character.
1734 Allows length and flags to be passed to low level routine.
1738 /* On ASCII machines this is normally a macro but we want
1739 a real function in case XS code wants it
1741 #undef Perl_utf8n_to_uvchr
1743 Perl_utf8n_to_uvchr(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
1745 UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags);
1746 return UNI_TO_NATIVE(uv);
1750 =for apidoc A|char *|pv_uni_display|SV *dsv|U8 *spv|STRLEN len|STRLEN pvlim|UV flags
1752 Build to the scalar dsv a displayable version of the string spv,
1753 length len, the displayable version being at most pvlim bytes long
1754 (if longer, the rest is truncated and "..." will be appended).
1756 The flags argument can have UNI_DISPLAY_ISPRINT set to display
1757 isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH
1758 to display the \\[nrfta\\] as the backslashed versions (like '\n')
1759 (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\).
1760 UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both
1761 UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on.
1763 The pointer to the PV of the dsv is returned.
1767 Perl_pv_uni_display(pTHX_ SV *dsv, U8 *spv, STRLEN len, STRLEN pvlim, UV flags)
1772 sv_setpvn(dsv, "", 0);
1773 for (s = (char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) {
1777 if (pvlim && SvCUR(dsv) >= pvlim) {
1781 u = utf8_to_uvchr((U8*)s, 0);
1783 if (!ok && (flags & UNI_DISPLAY_BACKSLASH)) {
1786 Perl_sv_catpvf(aTHX_ dsv, "\\n"); ok = TRUE; break;
1788 Perl_sv_catpvf(aTHX_ dsv, "\\r"); ok = TRUE; break;
1790 Perl_sv_catpvf(aTHX_ dsv, "\\t"); ok = TRUE; break;
1792 Perl_sv_catpvf(aTHX_ dsv, "\\f"); ok = TRUE; break;
1794 Perl_sv_catpvf(aTHX_ dsv, "\\a"); ok = TRUE; break;
1796 Perl_sv_catpvf(aTHX_ dsv, "\\\\" ); ok = TRUE; break;
1800 /* isPRINT() is the locale-blind version. */
1801 if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(u & 0xFF)) {
1802 Perl_sv_catpvf(aTHX_ dsv, "%c", (char)(u & 0xFF));
1807 Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u);
1810 sv_catpvn(dsv, "...", 3);
1816 =for apidoc A|char *|sv_uni_display|SV *dsv|SV *ssv|STRLEN pvlim|UV flags
1818 Build to the scalar dsv a displayable version of the scalar sv,
1819 the displayable version being at most pvlim bytes long
1820 (if longer, the rest is truncated and "..." will be appended).
1822 The flags argument is as in pv_uni_display().
1824 The pointer to the PV of the dsv is returned.
1828 Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags)
1830 return Perl_pv_uni_display(aTHX_ dsv, (U8*)SvPVX(ssv), SvCUR(ssv),
1835 =for apidoc A|I32|ibcmp_utf8|const char *s1|char **pe1|register UV l1|bool u1|const char *s2|char **pe2|register UV l2|bool u2
1837 Return true if the strings s1 and s2 differ case-insensitively, false
1838 if not (if they are equal case-insensitively). If u1 is true, the
1839 string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true,
1840 the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2
1841 are false, the respective string is assumed to be in native 8-bit
1844 If the pe1 and pe2 are non-NULL, the scanning pointers will be copied
1845 in there (they will point at the beginning of the I<next> character).
1846 If the pointers behind pe1 or pe2 are non-NULL, they are the end
1847 pointers beyond which scanning will not continue under any
1848 circustances. If the byte lengths l1 and l2 are non-zero, s1+l1 and
1849 s2+l2 will be used as goal end pointers that will also stop the scan,
1850 and which qualify towards defining a successful match: all the scans
1851 that define an explicit length must reach their goal pointers for
1852 a match to succeed).
1854 For case-insensitiveness, the "casefolding" of Unicode is used
1855 instead of upper/lowercasing both the characters, see
1856 http://www.unicode.org/unicode/reports/tr21/ (Case Mappings).
1860 Perl_ibcmp_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2)
1862 register U8 *p1 = (U8*)s1;
1863 register U8 *p2 = (U8*)s2;
1864 register U8 *e1 = 0, *f1 = 0, *q1 = 0;
1865 register U8 *e2 = 0, *f2 = 0, *q2 = 0;
1866 STRLEN n1 = 0, n2 = 0;
1867 U8 foldbuf1[UTF8_MAXLEN_FOLD+1];
1868 U8 foldbuf2[UTF8_MAXLEN_FOLD+1];
1870 STRLEN foldlen1, foldlen2;
1875 if (e1 == 0 || (l1 && l1 < e1 - (U8*)s1))
1879 if (e2 == 0 || (l2 && l2 < e2 - (U8*)s2))
1882 if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0))
1883 return 1; /* mismatch; possible infinite loop or false positive */
1886 natbuf[1] = 0; /* Need to terminate the buffer. */
1888 while ((e1 == 0 || p1 < e1) &&
1889 (f1 == 0 || p1 < f1) &&
1890 (e2 == 0 || p2 < e2) &&
1891 (f2 == 0 || p2 < f2)) {
1894 to_utf8_fold(p1, foldbuf1, &foldlen1);
1897 to_utf8_fold(natbuf, foldbuf1, &foldlen1);
1904 to_utf8_fold(p2, foldbuf2, &foldlen2);
1907 to_utf8_fold(natbuf, foldbuf2, &foldlen2);
1913 if ( UTF8SKIP(q1) != UTF8SKIP(q2) ||
1914 (UTF8SKIP(q1) == 1 && *q1 != *q2) ||
1915 memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) )
1916 return 1; /* mismatch */
1923 p1 += u1 ? UTF8SKIP(p1) : 1;
1925 p2 += u2 ? UTF8SKIP(p2) : 1;
1929 /* A match is defined by all the scans that specified
1930 * an explicit length reaching their final goals. */
1931 match = (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2);
1940 return match ? 0 : 1; /* 0 match, 1 mismatch */