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 ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */
68 !(flags & UNICODE_ALLOW_FFFF))) &&
69 /* UNICODE_ALLOW_SUPER includes
70 * FFFEs and FFFFs beyond 0x10FFFF. */
71 ((uv <= PERL_UNICODE_MAX) ||
72 !(flags & UNICODE_ALLOW_SUPER))
74 Perl_warner(aTHX_ packWARN(WARN_UTF8),
75 "Unicode character 0x%04"UVxf" is illegal", uv);
77 if (UNI_IS_INVARIANT(uv)) {
78 *d++ = UTF_TO_NATIVE(uv);
83 STRLEN len = UNISKIP(uv);
86 *p-- = UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK);
87 uv >>= UTF_ACCUMULATION_SHIFT;
89 *p = UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len));
92 #else /* Non loop style */
94 *d++ = (( uv >> 6) | 0xc0);
95 *d++ = (( uv & 0x3f) | 0x80);
99 *d++ = (( uv >> 12) | 0xe0);
100 *d++ = (((uv >> 6) & 0x3f) | 0x80);
101 *d++ = (( uv & 0x3f) | 0x80);
105 *d++ = (( uv >> 18) | 0xf0);
106 *d++ = (((uv >> 12) & 0x3f) | 0x80);
107 *d++ = (((uv >> 6) & 0x3f) | 0x80);
108 *d++ = (( uv & 0x3f) | 0x80);
111 if (uv < 0x4000000) {
112 *d++ = (( uv >> 24) | 0xf8);
113 *d++ = (((uv >> 18) & 0x3f) | 0x80);
114 *d++ = (((uv >> 12) & 0x3f) | 0x80);
115 *d++ = (((uv >> 6) & 0x3f) | 0x80);
116 *d++ = (( uv & 0x3f) | 0x80);
119 if (uv < 0x80000000) {
120 *d++ = (( uv >> 30) | 0xfc);
121 *d++ = (((uv >> 24) & 0x3f) | 0x80);
122 *d++ = (((uv >> 18) & 0x3f) | 0x80);
123 *d++ = (((uv >> 12) & 0x3f) | 0x80);
124 *d++ = (((uv >> 6) & 0x3f) | 0x80);
125 *d++ = (( uv & 0x3f) | 0x80);
129 if (uv < UTF8_QUAD_MAX)
132 *d++ = 0xfe; /* Can't match U+FEFF! */
133 *d++ = (((uv >> 30) & 0x3f) | 0x80);
134 *d++ = (((uv >> 24) & 0x3f) | 0x80);
135 *d++ = (((uv >> 18) & 0x3f) | 0x80);
136 *d++ = (((uv >> 12) & 0x3f) | 0x80);
137 *d++ = (((uv >> 6) & 0x3f) | 0x80);
138 *d++ = (( uv & 0x3f) | 0x80);
143 *d++ = 0xff; /* Can't match U+FFFE! */
144 *d++ = 0x80; /* 6 Reserved bits */
145 *d++ = (((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */
146 *d++ = (((uv >> 54) & 0x3f) | 0x80);
147 *d++ = (((uv >> 48) & 0x3f) | 0x80);
148 *d++ = (((uv >> 42) & 0x3f) | 0x80);
149 *d++ = (((uv >> 36) & 0x3f) | 0x80);
150 *d++ = (((uv >> 30) & 0x3f) | 0x80);
151 *d++ = (((uv >> 24) & 0x3f) | 0x80);
152 *d++ = (((uv >> 18) & 0x3f) | 0x80);
153 *d++ = (((uv >> 12) & 0x3f) | 0x80);
154 *d++ = (((uv >> 6) & 0x3f) | 0x80);
155 *d++ = (( uv & 0x3f) | 0x80);
159 #endif /* Loop style */
163 Perl_uvuni_to_utf8(pTHX_ U8 *d, UV uv)
165 return Perl_uvuni_to_utf8_flags(aTHX_ d, uv, 0);
170 =for apidoc A|STRLEN|is_utf8_char|U8 *s
172 Tests if some arbitrary number of bytes begins in a valid UTF-8
173 character. Note that an INVARIANT (i.e. ASCII) character is a valid UTF-8 character.
174 The actual number of bytes in the UTF-8 character will be returned if
175 it is valid, otherwise 0.
180 Perl_is_utf8_char(pTHX_ U8 *s)
186 if (UTF8_IS_INVARIANT(u))
189 if (!UTF8_IS_START(u))
194 if (len < 2 || !UTF8_IS_CONTINUATION(s[1]))
199 u &= UTF_START_MASK(len);
203 if (!UTF8_IS_CONTINUATION(*s))
205 uv = UTF8_ACCUMULATE(uv, *s);
212 if (UNISKIP(uv) < len)
219 =for apidoc A|bool|is_utf8_string|U8 *s|STRLEN len
221 Returns true if first C<len> bytes of the given string form a valid UTF8
222 string, false otherwise. Note that 'a valid UTF8 string' does not mean
223 'a string that contains UTF8' because a valid ASCII string is a valid
230 Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len)
237 len = strlen((char *)s);
253 =for apidoc A|UV|utf8n_to_uvuni|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
255 Bottom level UTF-8 decode routine.
256 Returns the unicode code point value of the first character in the string C<s>
257 which is assumed to be in UTF8 encoding and no longer than C<curlen>;
258 C<retlen> will be set to the length, in bytes, of that character.
260 If C<s> does not point to a well-formed UTF8 character, the behaviour
261 is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY,
262 it is assumed that the caller will raise a warning, and this function
263 will silently just set C<retlen> to C<-1> and return zero. If the
264 C<flags> does not contain UTF8_CHECK_ONLY, warnings about
265 malformations will be given, C<retlen> will be set to the expected
266 length of the UTF-8 character in bytes, and zero will be returned.
268 The C<flags> can also contain various flags to allow deviations from
269 the strict UTF-8 encoding (see F<utf8.h>).
271 Most code should use utf8_to_uvchr() rather than call this directly.
277 Perl_utf8n_to_uvuni(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
282 bool dowarn = ckWARN_d(WARN_UTF8);
284 STRLEN expectlen = 0;
287 /* This list is a superset of the UTF8_ALLOW_XXX. */
289 #define UTF8_WARN_EMPTY 1
290 #define UTF8_WARN_CONTINUATION 2
291 #define UTF8_WARN_NON_CONTINUATION 3
292 #define UTF8_WARN_FE_FF 4
293 #define UTF8_WARN_SHORT 5
294 #define UTF8_WARN_OVERFLOW 6
295 #define UTF8_WARN_SURROGATE 7
296 #define UTF8_WARN_LONG 8
297 #define UTF8_WARN_FFFF 9 /* Also FFFE. */
300 !(flags & UTF8_ALLOW_EMPTY)) {
301 warning = UTF8_WARN_EMPTY;
305 if (UTF8_IS_INVARIANT(uv)) {
308 return (UV) (NATIVE_TO_UTF(*s));
311 if (UTF8_IS_CONTINUATION(uv) &&
312 !(flags & UTF8_ALLOW_CONTINUATION)) {
313 warning = UTF8_WARN_CONTINUATION;
317 if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) &&
318 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
319 warning = UTF8_WARN_NON_CONTINUATION;
324 uv = NATIVE_TO_UTF(uv);
326 if ((uv == 0xfe || uv == 0xff) &&
327 !(flags & UTF8_ALLOW_FE_FF)) {
328 warning = UTF8_WARN_FE_FF;
333 if (!(uv & 0x20)) { len = 2; uv &= 0x1f; }
334 else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; }
335 else if (!(uv & 0x08)) { len = 4; uv &= 0x07; }
336 else if (!(uv & 0x04)) { len = 5; uv &= 0x03; }
338 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
339 else { len = 7; uv &= 0x01; }
341 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
342 else if (!(uv & 0x01)) { len = 7; uv = 0; }
343 else { len = 13; uv = 0; } /* whoa! */
351 if ((curlen < expectlen) &&
352 !(flags & UTF8_ALLOW_SHORT)) {
353 warning = UTF8_WARN_SHORT;
362 if (!UTF8_IS_CONTINUATION(*s) &&
363 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
365 warning = UTF8_WARN_NON_CONTINUATION;
369 uv = UTF8_ACCUMULATE(uv, *s);
371 /* These cannot be allowed. */
373 if (!(flags & UTF8_ALLOW_LONG)) {
374 warning = UTF8_WARN_LONG;
378 else { /* uv < ouv */
379 /* This cannot be allowed. */
380 warning = UTF8_WARN_OVERFLOW;
388 if (UNICODE_IS_SURROGATE(uv) &&
389 !(flags & UTF8_ALLOW_SURROGATE)) {
390 warning = UTF8_WARN_SURROGATE;
392 } else if ((expectlen > UNISKIP(uv)) &&
393 !(flags & UTF8_ALLOW_LONG)) {
394 warning = UTF8_WARN_LONG;
396 } else if (UNICODE_IS_ILLEGAL(uv) &&
397 !(flags & UTF8_ALLOW_FFFF)) {
398 warning = UTF8_WARN_FFFF;
406 if (flags & UTF8_CHECK_ONLY) {
413 SV* sv = sv_2mortal(newSVpv("Malformed UTF-8 character ", 0));
416 case 0: /* Intentionally empty. */ break;
417 case UTF8_WARN_EMPTY:
418 Perl_sv_catpvf(aTHX_ sv, "(empty string)");
420 case UTF8_WARN_CONTINUATION:
421 Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv);
423 case UTF8_WARN_NON_CONTINUATION:
425 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")",
426 (UV)s[1], startbyte);
428 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)",
429 (UV)s[1], s - s0, s - s0 > 1 ? "s" : "", startbyte, expectlen);
432 case UTF8_WARN_FE_FF:
433 Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv);
435 case UTF8_WARN_SHORT:
436 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
437 curlen, curlen == 1 ? "" : "s", expectlen, startbyte);
438 expectlen = curlen; /* distance for caller to skip */
440 case UTF8_WARN_OVERFLOW:
441 Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")",
444 case UTF8_WARN_SURROGATE:
445 Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv);
448 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
449 expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte);
452 Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv);
455 Perl_sv_catpvf(aTHX_ sv, "(unknown reason)");
463 Perl_warner(aTHX_ packWARN(WARN_UTF8),
464 "%s in %s", s, OP_DESC(PL_op));
466 Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s);
471 *retlen = expectlen ? expectlen : len;
477 =for apidoc A|UV|utf8_to_uvchr|U8 *s|STRLEN *retlen
479 Returns the native character value of the first character in the string C<s>
480 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
481 length, in bytes, of that character.
483 If C<s> does not point to a well-formed UTF8 character, zero is
484 returned and retlen is set, if possible, to -1.
490 Perl_utf8_to_uvchr(pTHX_ U8 *s, STRLEN *retlen)
492 return Perl_utf8n_to_uvchr(aTHX_ s, UTF8_MAXLEN, retlen,
493 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
497 =for apidoc A|UV|utf8_to_uvuni|U8 *s|STRLEN *retlen
499 Returns the Unicode code point of the first character in the string C<s>
500 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
501 length, in bytes, of that character.
503 This function should only be used when returned UV is considered
504 an index into the Unicode semantic tables (e.g. swashes).
506 If C<s> does not point to a well-formed UTF8 character, zero is
507 returned and retlen is set, if possible, to -1.
513 Perl_utf8_to_uvuni(pTHX_ U8 *s, STRLEN *retlen)
515 /* Call the low level routine asking for checks */
516 return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXLEN, retlen,
517 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
521 =for apidoc A|STRLEN|utf8_length|U8 *s|U8 *e
523 Return the length of the UTF-8 char encoded string C<s> in characters.
524 Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end
525 up past C<e>, croaks.
531 Perl_utf8_length(pTHX_ U8 *s, U8 *e)
535 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
536 * the bitops (especially ~) can create illegal UTF-8.
537 * In other words: in Perl UTF-8 is not just for Unicode. */
540 if (ckWARN_d(WARN_UTF8)) {
542 Perl_warner(aTHX_ packWARN(WARN_UTF8),
543 "%s in %s", unees, OP_DESC(PL_op));
545 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
553 if (ckWARN_d(WARN_UTF8)) {
555 Perl_warner(aTHX_ packWARN(WARN_UTF8),
556 unees, OP_DESC(PL_op));
558 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
570 =for apidoc A|IV|utf8_distance|U8 *a|U8 *b
572 Returns the number of UTF8 characters between the UTF-8 pointers C<a>
575 WARNING: use only if you *know* that the pointers point inside the
582 Perl_utf8_distance(pTHX_ U8 *a, U8 *b)
586 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
587 * the bitops (especially ~) can create illegal UTF-8.
588 * In other words: in Perl UTF-8 is not just for Unicode. */
595 if (ckWARN_d(WARN_UTF8)) {
597 Perl_warner(aTHX_ packWARN(WARN_UTF8),
598 "%s in %s", unees, OP_DESC(PL_op));
600 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
613 if (ckWARN_d(WARN_UTF8)) {
615 Perl_warner(aTHX_ packWARN(WARN_UTF8),
616 "%s in %s", unees, OP_DESC(PL_op));
618 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
631 =for apidoc A|U8 *|utf8_hop|U8 *s|I32 off
633 Return the UTF-8 pointer C<s> displaced by C<off> characters, either
636 WARNING: do not use the following unless you *know* C<off> is within
637 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
638 on the first byte of character or just after the last byte of a character.
644 Perl_utf8_hop(pTHX_ U8 *s, I32 off)
646 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
647 * the bitops (especially ~) can create illegal UTF-8.
648 * In other words: in Perl UTF-8 is not just for Unicode. */
657 while (UTF8_IS_CONTINUATION(*s))
665 =for apidoc A|U8 *|utf8_to_bytes|U8 *s|STRLEN *len
667 Converts a string C<s> of length C<len> from UTF8 into byte encoding.
668 Unlike C<bytes_to_utf8>, this over-writes the original string, and
669 updates len to contain the new length.
670 Returns zero on failure, setting C<len> to -1.
676 Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len)
682 /* ensure valid UTF8 and chars < 256 before updating string */
683 for (send = s + *len; s < send; ) {
686 if (!UTF8_IS_INVARIANT(c) &&
687 (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send)
688 || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) {
697 *d++ = (U8)utf8_to_uvchr(s, &ulen);
706 =for apidoc A|U8 *|bytes_from_utf8|U8 *s|STRLEN *len|bool *is_utf8
708 Converts a string C<s> of length C<len> from UTF8 into byte encoding.
709 Unlike <utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to
710 the newly-created string, and updates C<len> to contain the new
711 length. Returns the original string if no conversion occurs, C<len>
712 is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to
713 0 if C<s> is converted or contains all 7bit characters.
719 Perl_bytes_from_utf8(pTHX_ U8 *s, STRLEN *len, bool *is_utf8)
729 /* ensure valid UTF8 and chars < 256 before converting string */
730 for (send = s + *len; s < send;) {
732 if (!UTF8_IS_INVARIANT(c)) {
733 if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send &&
734 (c = *s++) && UTF8_IS_CONTINUATION(c))
743 Newz(801, d, (*len) - count + 1, U8);
744 s = start; start = d;
747 if (!UTF8_IS_INVARIANT(c)) {
748 /* Then it is two-byte encoded */
749 c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++);
750 c = ASCII_TO_NATIVE(c);
760 =for apidoc A|U8 *|bytes_to_utf8|U8 *s|STRLEN *len
762 Converts a string C<s> of length C<len> from ASCII into UTF8 encoding.
763 Returns a pointer to the newly-created string, and sets C<len> to
764 reflect the new length.
770 Perl_bytes_to_utf8(pTHX_ U8 *s, STRLEN *len)
777 Newz(801, d, (*len) * 2 + 1, U8);
781 UV uv = NATIVE_TO_ASCII(*s++);
782 if (UNI_IS_INVARIANT(uv))
783 *d++ = UTF_TO_NATIVE(uv);
785 *d++ = UTF8_EIGHT_BIT_HI(uv);
786 *d++ = UTF8_EIGHT_BIT_LO(uv);
795 * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8.
797 * Destination must be pre-extended to 3/2 source. Do not use in-place.
798 * We optimize for native, for obvious reasons. */
801 Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
807 Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen");
812 UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */
819 *d++ = (( uv >> 6) | 0xc0);
820 *d++ = (( uv & 0x3f) | 0x80);
823 if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */
825 if (low < 0xdc00 || low >= 0xdfff)
826 Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
827 uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000;
830 *d++ = (( uv >> 12) | 0xe0);
831 *d++ = (((uv >> 6) & 0x3f) | 0x80);
832 *d++ = (( uv & 0x3f) | 0x80);
836 *d++ = (( uv >> 18) | 0xf0);
837 *d++ = (((uv >> 12) & 0x3f) | 0x80);
838 *d++ = (((uv >> 6) & 0x3f) | 0x80);
839 *d++ = (( uv & 0x3f) | 0x80);
843 *newlen = d - dstart;
847 /* Note: this one is slightly destructive of the source. */
850 Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
853 U8* send = s + bytelen;
860 return utf16_to_utf8(p, d, bytelen, newlen);
863 /* for now these are all defined (inefficiently) in terms of the utf8 versions */
866 Perl_is_uni_alnum(pTHX_ UV c)
868 U8 tmpbuf[UTF8_MAXLEN+1];
869 uvchr_to_utf8(tmpbuf, c);
870 return is_utf8_alnum(tmpbuf);
874 Perl_is_uni_alnumc(pTHX_ UV c)
876 U8 tmpbuf[UTF8_MAXLEN+1];
877 uvchr_to_utf8(tmpbuf, c);
878 return is_utf8_alnumc(tmpbuf);
882 Perl_is_uni_idfirst(pTHX_ UV c)
884 U8 tmpbuf[UTF8_MAXLEN+1];
885 uvchr_to_utf8(tmpbuf, c);
886 return is_utf8_idfirst(tmpbuf);
890 Perl_is_uni_alpha(pTHX_ UV c)
892 U8 tmpbuf[UTF8_MAXLEN+1];
893 uvchr_to_utf8(tmpbuf, c);
894 return is_utf8_alpha(tmpbuf);
898 Perl_is_uni_ascii(pTHX_ UV c)
900 U8 tmpbuf[UTF8_MAXLEN+1];
901 uvchr_to_utf8(tmpbuf, c);
902 return is_utf8_ascii(tmpbuf);
906 Perl_is_uni_space(pTHX_ UV c)
908 U8 tmpbuf[UTF8_MAXLEN+1];
909 uvchr_to_utf8(tmpbuf, c);
910 return is_utf8_space(tmpbuf);
914 Perl_is_uni_digit(pTHX_ UV c)
916 U8 tmpbuf[UTF8_MAXLEN+1];
917 uvchr_to_utf8(tmpbuf, c);
918 return is_utf8_digit(tmpbuf);
922 Perl_is_uni_upper(pTHX_ UV c)
924 U8 tmpbuf[UTF8_MAXLEN+1];
925 uvchr_to_utf8(tmpbuf, c);
926 return is_utf8_upper(tmpbuf);
930 Perl_is_uni_lower(pTHX_ UV c)
932 U8 tmpbuf[UTF8_MAXLEN+1];
933 uvchr_to_utf8(tmpbuf, c);
934 return is_utf8_lower(tmpbuf);
938 Perl_is_uni_cntrl(pTHX_ UV c)
940 U8 tmpbuf[UTF8_MAXLEN+1];
941 uvchr_to_utf8(tmpbuf, c);
942 return is_utf8_cntrl(tmpbuf);
946 Perl_is_uni_graph(pTHX_ UV c)
948 U8 tmpbuf[UTF8_MAXLEN+1];
949 uvchr_to_utf8(tmpbuf, c);
950 return is_utf8_graph(tmpbuf);
954 Perl_is_uni_print(pTHX_ UV c)
956 U8 tmpbuf[UTF8_MAXLEN+1];
957 uvchr_to_utf8(tmpbuf, c);
958 return is_utf8_print(tmpbuf);
962 Perl_is_uni_punct(pTHX_ UV c)
964 U8 tmpbuf[UTF8_MAXLEN+1];
965 uvchr_to_utf8(tmpbuf, c);
966 return is_utf8_punct(tmpbuf);
970 Perl_is_uni_xdigit(pTHX_ UV c)
972 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
973 uvchr_to_utf8(tmpbuf, c);
974 return is_utf8_xdigit(tmpbuf);
978 Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp)
981 return to_utf8_upper(p, p, lenp);
985 Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp)
988 return to_utf8_title(p, p, lenp);
992 Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp)
995 return to_utf8_lower(p, p, lenp);
999 Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp)
1001 uvchr_to_utf8(p, c);
1002 return to_utf8_fold(p, p, lenp);
1005 /* for now these all assume no locale info available for Unicode > 255 */
1008 Perl_is_uni_alnum_lc(pTHX_ UV c)
1010 return is_uni_alnum(c); /* XXX no locale support yet */
1014 Perl_is_uni_alnumc_lc(pTHX_ UV c)
1016 return is_uni_alnumc(c); /* XXX no locale support yet */
1020 Perl_is_uni_idfirst_lc(pTHX_ UV c)
1022 return is_uni_idfirst(c); /* XXX no locale support yet */
1026 Perl_is_uni_alpha_lc(pTHX_ UV c)
1028 return is_uni_alpha(c); /* XXX no locale support yet */
1032 Perl_is_uni_ascii_lc(pTHX_ UV c)
1034 return is_uni_ascii(c); /* XXX no locale support yet */
1038 Perl_is_uni_space_lc(pTHX_ UV c)
1040 return is_uni_space(c); /* XXX no locale support yet */
1044 Perl_is_uni_digit_lc(pTHX_ UV c)
1046 return is_uni_digit(c); /* XXX no locale support yet */
1050 Perl_is_uni_upper_lc(pTHX_ UV c)
1052 return is_uni_upper(c); /* XXX no locale support yet */
1056 Perl_is_uni_lower_lc(pTHX_ UV c)
1058 return is_uni_lower(c); /* XXX no locale support yet */
1062 Perl_is_uni_cntrl_lc(pTHX_ UV c)
1064 return is_uni_cntrl(c); /* XXX no locale support yet */
1068 Perl_is_uni_graph_lc(pTHX_ UV c)
1070 return is_uni_graph(c); /* XXX no locale support yet */
1074 Perl_is_uni_print_lc(pTHX_ UV c)
1076 return is_uni_print(c); /* XXX no locale support yet */
1080 Perl_is_uni_punct_lc(pTHX_ UV c)
1082 return is_uni_punct(c); /* XXX no locale support yet */
1086 Perl_is_uni_xdigit_lc(pTHX_ UV c)
1088 return is_uni_xdigit(c); /* XXX no locale support yet */
1092 Perl_to_uni_upper_lc(pTHX_ U32 c)
1094 /* XXX returns only the first character -- do not use XXX */
1095 /* XXX no locale support yet */
1097 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1098 return (U32)to_uni_upper(c, tmpbuf, &len);
1102 Perl_to_uni_title_lc(pTHX_ U32 c)
1104 /* XXX returns only the first character XXX -- do not use XXX */
1105 /* XXX no locale support yet */
1107 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1108 return (U32)to_uni_title(c, tmpbuf, &len);
1112 Perl_to_uni_lower_lc(pTHX_ U32 c)
1114 /* XXX returns only the first character -- do not use XXX */
1115 /* XXX no locale support yet */
1117 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1118 return (U32)to_uni_lower(c, tmpbuf, &len);
1122 Perl_is_utf8_alnum(pTHX_ U8 *p)
1124 if (!is_utf8_char(p))
1127 /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true
1128 * descendant of isalnum(3), in other words, it doesn't
1129 * contain the '_'. --jhi */
1130 PL_utf8_alnum = swash_init("utf8", "IsWord", &PL_sv_undef, 0, 0);
1131 return swash_fetch(PL_utf8_alnum, p, TRUE);
1132 /* return *p == '_' || is_utf8_alpha(p) || is_utf8_digit(p); */
1133 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1135 PL_utf8_alnum = swash_init("utf8", "",
1136 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1137 return swash_fetch(PL_utf8_alnum, p, TRUE);
1142 Perl_is_utf8_alnumc(pTHX_ U8 *p)
1144 if (!is_utf8_char(p))
1147 PL_utf8_alnum = swash_init("utf8", "IsAlnumC", &PL_sv_undef, 0, 0);
1148 return swash_fetch(PL_utf8_alnum, p, TRUE);
1149 /* return is_utf8_alpha(p) || is_utf8_digit(p); */
1150 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1152 PL_utf8_alnum = swash_init("utf8", "",
1153 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1154 return swash_fetch(PL_utf8_alnum, p, TRUE);
1159 Perl_is_utf8_idfirst(pTHX_ U8 *p)
1161 return *p == '_' || is_utf8_alpha(p);
1165 Perl_is_utf8_alpha(pTHX_ U8 *p)
1167 if (!is_utf8_char(p))
1170 PL_utf8_alpha = swash_init("utf8", "IsAlpha", &PL_sv_undef, 0, 0);
1171 return swash_fetch(PL_utf8_alpha, p, TRUE);
1175 Perl_is_utf8_ascii(pTHX_ U8 *p)
1177 if (!is_utf8_char(p))
1180 PL_utf8_ascii = swash_init("utf8", "IsAscii", &PL_sv_undef, 0, 0);
1181 return swash_fetch(PL_utf8_ascii, p, TRUE);
1185 Perl_is_utf8_space(pTHX_ U8 *p)
1187 if (!is_utf8_char(p))
1190 PL_utf8_space = swash_init("utf8", "IsSpacePerl", &PL_sv_undef, 0, 0);
1191 return swash_fetch(PL_utf8_space, p, TRUE);
1195 Perl_is_utf8_digit(pTHX_ U8 *p)
1197 if (!is_utf8_char(p))
1200 PL_utf8_digit = swash_init("utf8", "IsDigit", &PL_sv_undef, 0, 0);
1201 return swash_fetch(PL_utf8_digit, p, TRUE);
1205 Perl_is_utf8_upper(pTHX_ U8 *p)
1207 if (!is_utf8_char(p))
1210 PL_utf8_upper = swash_init("utf8", "IsUpper", &PL_sv_undef, 0, 0);
1211 return swash_fetch(PL_utf8_upper, p, TRUE);
1215 Perl_is_utf8_lower(pTHX_ U8 *p)
1217 if (!is_utf8_char(p))
1220 PL_utf8_lower = swash_init("utf8", "IsLower", &PL_sv_undef, 0, 0);
1221 return swash_fetch(PL_utf8_lower, p, TRUE);
1225 Perl_is_utf8_cntrl(pTHX_ U8 *p)
1227 if (!is_utf8_char(p))
1230 PL_utf8_cntrl = swash_init("utf8", "IsCntrl", &PL_sv_undef, 0, 0);
1231 return swash_fetch(PL_utf8_cntrl, p, TRUE);
1235 Perl_is_utf8_graph(pTHX_ U8 *p)
1237 if (!is_utf8_char(p))
1240 PL_utf8_graph = swash_init("utf8", "IsGraph", &PL_sv_undef, 0, 0);
1241 return swash_fetch(PL_utf8_graph, p, TRUE);
1245 Perl_is_utf8_print(pTHX_ U8 *p)
1247 if (!is_utf8_char(p))
1250 PL_utf8_print = swash_init("utf8", "IsPrint", &PL_sv_undef, 0, 0);
1251 return swash_fetch(PL_utf8_print, p, TRUE);
1255 Perl_is_utf8_punct(pTHX_ U8 *p)
1257 if (!is_utf8_char(p))
1260 PL_utf8_punct = swash_init("utf8", "IsPunct", &PL_sv_undef, 0, 0);
1261 return swash_fetch(PL_utf8_punct, p, TRUE);
1265 Perl_is_utf8_xdigit(pTHX_ U8 *p)
1267 if (!is_utf8_char(p))
1269 if (!PL_utf8_xdigit)
1270 PL_utf8_xdigit = swash_init("utf8", "IsXDigit", &PL_sv_undef, 0, 0);
1271 return swash_fetch(PL_utf8_xdigit, p, TRUE);
1275 Perl_is_utf8_mark(pTHX_ U8 *p)
1277 if (!is_utf8_char(p))
1280 PL_utf8_mark = swash_init("utf8", "IsM", &PL_sv_undef, 0, 0);
1281 return swash_fetch(PL_utf8_mark, p, TRUE);
1285 =for apidoc A|UV|to_utf8_case|U8 *p|U8* ustrp|STRLEN *lenp|SV **swash|char *normal|char *special
1287 The "p" contains the pointer to the UTF-8 string encoding
1288 the character that is being converted.
1290 The "ustrp" is a pointer to the character buffer to put the
1291 conversion result to. The "lenp" is a pointer to the length
1294 The "swashp" is a pointer to the swash to use.
1296 Both the special and normal mappings are stored lib/unicore/To/Foo.pl,
1297 and loaded by SWASHGET, using lib/utf8_heavy.pl. The special (usually,
1298 but not always, a multicharacter mapping), is tried first.
1300 The "special" is a string like "utf8::ToSpecLower", which means the
1301 hash %utf8::ToSpecLower. The access to the hash is through
1302 Perl_to_utf8_case().
1304 The "normal" is a string like "ToLower" which means the swash
1310 Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp, char *normal, char *special)
1313 U8 tmpbuf[UTF8_MAXLEN_FOLD+1];
1316 uv0 = utf8_to_uvchr(p, 0);
1317 /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings
1318 * are necessary in EBCDIC, they are redundant no-ops
1319 * in ASCII-ish platforms, and hopefully optimized away. */
1320 uv1 = NATIVE_TO_UNI(uv0);
1321 uvuni_to_utf8(tmpbuf, uv1);
1323 if (!*swashp) /* load on-demand */
1324 *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0);
1327 /* It might be "special" (sometimes, but not always,
1328 * a multicharacter mapping) */
1334 if ((hv = get_hv(special, FALSE)) &&
1335 (keysv = sv_2mortal(Perl_newSVpvf(aTHX_ "%04"UVXf, uv1))) &&
1336 (he = hv_fetch_ent(hv, keysv, FALSE, 0)) &&
1337 (val = HeVAL(he))) {
1342 len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp;
1345 /* If we have EBCDIC we need to remap the characters
1346 * since any characters in the low 256 are Unicode
1347 * code points, not EBCDIC. */
1348 U8 *t = (U8*)s, *tend = t + len, *d;
1355 UV c = utf8_to_uvchr(t, &tlen);
1357 d = uvchr_to_utf8(d, UNI_TO_NATIVE(c));
1366 d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t));
1371 Copy(tmpbuf, ustrp, len, U8);
1373 Copy(s, ustrp, len, U8);
1379 if (!len && *swashp) {
1380 UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE);
1383 /* It was "normal" (a single character mapping). */
1384 UV uv3 = UNI_TO_NATIVE(uv2);
1386 len = uvchr_to_utf8(ustrp, uv3) - ustrp;
1390 if (!len) /* Neither: just copy. */
1391 len = uvchr_to_utf8(ustrp, uv0) - ustrp;
1396 return len ? utf8_to_uvchr(ustrp, 0) : 0;
1400 =for apidoc A|UV|to_utf8_upper|U8 *p|U8 *ustrp|STRLEN *lenp
1402 Convert the UTF-8 encoded character at p to its uppercase version and
1403 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1404 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1405 uppercase version may be longer than the original character (up to two
1408 The first character of the uppercased version is returned
1409 (but note, as explained above, that there may be more.)
1414 Perl_to_utf8_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1416 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1417 &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper");
1421 =for apidoc A|UV|to_utf8_title|U8 *p|U8 *ustrp|STRLEN *lenp
1423 Convert the UTF-8 encoded character at p to its titlecase version and
1424 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1425 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1426 titlecase version may be longer than the original character (up to two
1429 The first character of the titlecased version is returned
1430 (but note, as explained above, that there may be more.)
1435 Perl_to_utf8_title(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1437 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1438 &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle");
1442 =for apidoc A|UV|to_utf8_lower|U8 *p|U8 *ustrp|STRLEN *lenp
1444 Convert the UTF-8 encoded character at p to its lowercase version and
1445 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1446 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1447 lowercase version may be longer than the original character (up to two
1450 The first character of the lowercased version is returned
1451 (but note, as explained above, that there may be more.)
1456 Perl_to_utf8_lower(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1458 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1459 &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower");
1463 =for apidoc A|UV|to_utf8_fold|U8 *p|U8 *ustrp|STRLEN *lenp
1465 Convert the UTF-8 encoded character at p to its foldcase version and
1466 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1467 that the ustrp needs to be at least UTF8_MAXLEN_FOLD+1 bytes since the
1468 foldcase version may be longer than the original character (up to
1471 The first character of the foldcased version is returned
1472 (but note, as explained above, that there may be more.)
1477 Perl_to_utf8_fold(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1479 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1480 &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold");
1483 /* a "swash" is a swatch hash */
1486 Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none)
1489 SV* tokenbufsv = sv_2mortal(NEWSV(0,0));
1491 HV *stash = gv_stashpvn(pkg, strlen(pkg), FALSE);
1494 if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */
1496 errsv_save = newSVsv(ERRSV);
1497 Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpv(pkg,0), Nullsv);
1499 sv_setsv(ERRSV, errsv_save);
1500 SvREFCNT_dec(errsv_save);
1504 PUSHSTACKi(PERLSI_MAGIC);
1507 PUSHs(sv_2mortal(newSVpvn(pkg, strlen(pkg))));
1508 PUSHs(sv_2mortal(newSVpvn(name, strlen(name))));
1510 PUSHs(sv_2mortal(newSViv(minbits)));
1511 PUSHs(sv_2mortal(newSViv(none)));
1517 if (PL_curcop == &PL_compiling)
1518 /* XXX ought to be handled by lex_start */
1519 sv_setpv(tokenbufsv, PL_tokenbuf);
1520 errsv_save = newSVsv(ERRSV);
1521 if (call_method("SWASHNEW", G_SCALAR))
1522 retval = newSVsv(*PL_stack_sp--);
1524 retval = &PL_sv_undef;
1526 sv_setsv(ERRSV, errsv_save);
1527 SvREFCNT_dec(errsv_save);
1530 if (PL_curcop == &PL_compiling) {
1532 char* pv = SvPV(tokenbufsv, len);
1534 Copy(pv, PL_tokenbuf, len+1, char);
1535 PL_curcop->op_private = PL_hints;
1537 if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) {
1539 Perl_croak(aTHX_ "Can't find Unicode property definition \"%s\"",
1540 SvPV_nolen(retval));
1541 Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref");
1547 /* This API is wrong for special case conversions since we may need to
1548 * return several Unicode characters for a single Unicode character
1549 * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is
1550 * the lower-level routine, and it is similarly broken for returning
1551 * multiple values. --jhi */
1553 Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr, bool do_utf8)
1555 HV* hv = (HV*)SvRV(sv);
1564 UV c = NATIVE_TO_ASCII(*ptr);
1566 if (!do_utf8 && !UNI_IS_INVARIANT(c)) {
1567 tmputf8[0] = UTF8_EIGHT_BIT_HI(c);
1568 tmputf8[1] = UTF8_EIGHT_BIT_LO(c);
1571 /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ
1572 * then the "swatch" is a vec() for al the chars which start
1574 * So the key in the hash (klen) is length of encoded char -1
1576 klen = UTF8SKIP(ptr) - 1;
1581 /* If char in invariant then swatch is for all the invariant chars
1582 * In both UTF-8 and UTF8-MOD that happens to be UTF_CONTINUATION_MARK
1584 needents = UTF_CONTINUATION_MARK;
1585 off = NATIVE_TO_UTF(ptr[klen]);
1589 /* If char is encoded then swatch is for the prefix */
1590 needents = (1 << UTF_ACCUMULATION_SHIFT);
1591 off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK;
1595 * This single-entry cache saves about 1/3 of the utf8 overhead in test
1596 * suite. (That is, only 7-8% overall over just a hash cache. Still,
1597 * it's nothing to sniff at.) Pity we usually come through at least
1598 * two function calls to get here...
1600 * NB: this code assumes that swatches are never modified, once generated!
1603 if (hv == PL_last_swash_hv &&
1604 klen == PL_last_swash_klen &&
1605 (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) )
1607 tmps = PL_last_swash_tmps;
1608 slen = PL_last_swash_slen;
1611 /* Try our second-level swatch cache, kept in a hash. */
1612 SV** svp = hv_fetch(hv, (char*)ptr, klen, FALSE);
1614 /* If not cached, generate it via utf8::SWASHGET */
1615 if (!svp || !SvPOK(*svp) || !(tmps = (U8*)SvPV(*svp, slen))) {
1617 /* We use utf8n_to_uvuni() as we want an index into
1618 Unicode tables, not a native character number.
1620 UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXLEN, 0,
1622 0 : UTF8_ALLOW_ANY);
1627 PUSHSTACKi(PERLSI_MAGIC);
1631 /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */
1632 PUSHs(sv_2mortal(newSViv((klen) ?
1633 (code_point & ~(needents - 1)) : 0)));
1634 PUSHs(sv_2mortal(newSViv(needents)));
1636 errsv_save = newSVsv(ERRSV);
1637 if (call_method("SWASHGET", G_SCALAR))
1638 retval = newSVsv(*PL_stack_sp--);
1640 retval = &PL_sv_undef;
1642 sv_setsv(ERRSV, errsv_save);
1643 SvREFCNT_dec(errsv_save);
1647 if (PL_curcop == &PL_compiling)
1648 PL_curcop->op_private = PL_hints;
1650 svp = hv_store(hv, (char*)ptr, klen, retval, 0);
1652 if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || (slen << 3) < needents)
1653 Perl_croak(aTHX_ "SWASHGET didn't return result of proper length");
1656 PL_last_swash_hv = hv;
1657 PL_last_swash_klen = klen;
1658 PL_last_swash_tmps = tmps;
1659 PL_last_swash_slen = slen;
1661 Copy(ptr, PL_last_swash_key, klen, U8);
1664 switch ((int)((slen << 3) / needents)) {
1666 bit = 1 << (off & 7);
1668 return (tmps[off] & bit) != 0;
1673 return (tmps[off] << 8) + tmps[off + 1] ;
1676 return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ;
1678 Perl_croak(aTHX_ "panic: swash_fetch");
1684 =for apidoc A|U8 *|uvchr_to_utf8|U8 *d|UV uv
1686 Adds the UTF8 representation of the Native codepoint C<uv> to the end
1687 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
1688 bytes available. The return value is the pointer to the byte after the
1689 end of the new character. In other words,
1691 d = uvchr_to_utf8(d, uv);
1693 is the recommended wide native character-aware way of saying
1700 /* On ASCII machines this is normally a macro but we want a
1701 real function in case XS code wants it
1703 #undef Perl_uvchr_to_utf8
1705 Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv)
1707 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0);
1711 Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
1713 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags);
1717 =for apidoc A|UV|utf8n_to_uvchr|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
1719 Returns the native character value of the first character in the string C<s>
1720 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
1721 length, in bytes, of that character.
1723 Allows length and flags to be passed to low level routine.
1727 /* On ASCII machines this is normally a macro but we want
1728 a real function in case XS code wants it
1730 #undef Perl_utf8n_to_uvchr
1732 Perl_utf8n_to_uvchr(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
1734 UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags);
1735 return UNI_TO_NATIVE(uv);
1739 =for apidoc A|char *|pv_uni_display|SV *dsv|U8 *spv|STRLEN len|STRLEN pvlim|UV flags
1741 Build to the scalar dsv a displayable version of the string spv,
1742 length len, the displayable version being at most pvlim bytes long
1743 (if longer, the rest is truncated and "..." will be appended).
1745 The flags argument can have UNI_DISPLAY_ISPRINT set to display
1746 isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH
1747 to display the \\[nrfta\\] as the backslashed versions (like '\n')
1748 (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\).
1749 UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both
1750 UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on.
1752 The pointer to the PV of the dsv is returned.
1756 Perl_pv_uni_display(pTHX_ SV *dsv, U8 *spv, STRLEN len, STRLEN pvlim, UV flags)
1761 sv_setpvn(dsv, "", 0);
1762 for (s = (char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) {
1766 if (pvlim && SvCUR(dsv) >= pvlim) {
1770 u = utf8_to_uvchr((U8*)s, 0);
1772 if (!ok && (flags & UNI_DISPLAY_BACKSLASH)) {
1775 Perl_sv_catpvf(aTHX_ dsv, "\\n"); ok = TRUE; break;
1777 Perl_sv_catpvf(aTHX_ dsv, "\\r"); ok = TRUE; break;
1779 Perl_sv_catpvf(aTHX_ dsv, "\\t"); ok = TRUE; break;
1781 Perl_sv_catpvf(aTHX_ dsv, "\\f"); ok = TRUE; break;
1783 Perl_sv_catpvf(aTHX_ dsv, "\\a"); ok = TRUE; break;
1785 Perl_sv_catpvf(aTHX_ dsv, "\\\\" ); ok = TRUE; break;
1789 /* isPRINT() is the locale-blind version. */
1790 if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(u & 0xFF)) {
1791 Perl_sv_catpvf(aTHX_ dsv, "%c", (char)(u & 0xFF));
1796 Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u);
1799 sv_catpvn(dsv, "...", 3);
1805 =for apidoc A|char *|sv_uni_display|SV *dsv|SV *ssv|STRLEN pvlim|UV flags
1807 Build to the scalar dsv a displayable version of the scalar sv,
1808 the displayable version being at most pvlim bytes long
1809 (if longer, the rest is truncated and "..." will be appended).
1811 The flags argument is as in pv_uni_display().
1813 The pointer to the PV of the dsv is returned.
1817 Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags)
1819 return Perl_pv_uni_display(aTHX_ dsv, (U8*)SvPVX(ssv), SvCUR(ssv),
1824 =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
1826 Return true if the strings s1 and s2 differ case-insensitively, false
1827 if not (if they are equal case-insensitively). If u1 is true, the
1828 string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true,
1829 the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2
1830 are false, the respective string is assumed to be in native 8-bit
1833 If the pe1 and pe2 are non-NULL, the scanning pointers will be copied
1834 in there (they will point at the beginning of the I<next> character).
1835 If the pointers behind pe1 or pe2 are non-NULL, they are the end
1836 pointers beyond which scanning will not continue under any
1837 circustances. If the byte lengths l1 and l2 are non-zero, s1+l1 and
1838 s2+l2 will be used as goal end pointers that will also stop the scan,
1839 and which qualify towards defining a successful match: all the scans
1840 that define an explicit length must reach their goal pointers for
1841 a match to succeed).
1843 For case-insensitiveness, the "casefolding" of Unicode is used
1844 instead of upper/lowercasing both the characters, see
1845 http://www.unicode.org/unicode/reports/tr21/ (Case Mappings).
1849 Perl_ibcmp_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2)
1851 register U8 *p1 = (U8*)s1;
1852 register U8 *p2 = (U8*)s2;
1853 register U8 *e1 = 0, *f1 = 0, *q1 = 0;
1854 register U8 *e2 = 0, *f2 = 0, *q2 = 0;
1855 STRLEN n1 = 0, n2 = 0;
1856 U8 foldbuf1[UTF8_MAXLEN_FOLD+1];
1857 U8 foldbuf2[UTF8_MAXLEN_FOLD+1];
1859 STRLEN foldlen1, foldlen2;
1864 if (e1 == 0 || (l1 && l1 < e1 - (U8*)s1))
1868 if (e2 == 0 || (l2 && l2 < e2 - (U8*)s2))
1871 if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0))
1872 return 1; /* mismatch; possible infinite loop or false positive */
1875 natbuf[1] = 0; /* Need to terminate the buffer. */
1877 while ((e1 == 0 || p1 < e1) &&
1878 (f1 == 0 || p1 < f1) &&
1879 (e2 == 0 || p2 < e2) &&
1880 (f2 == 0 || p2 < f2)) {
1883 to_utf8_fold(p1, foldbuf1, &foldlen1);
1886 to_utf8_fold(natbuf, foldbuf1, &foldlen1);
1893 to_utf8_fold(p2, foldbuf2, &foldlen2);
1896 to_utf8_fold(natbuf, foldbuf2, &foldlen2);
1902 if ( UTF8SKIP(q1) != UTF8SKIP(q2) ||
1903 (UTF8SKIP(q1) == 1 && *q1 != *q2) ||
1904 memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) )
1905 return 1; /* mismatch */
1912 p1 += u1 ? UTF8SKIP(p1) : 1;
1914 p2 += u2 ? UTF8SKIP(p2) : 1;
1918 /* A match is defined by all the scans that specified
1919 * an explicit length reaching their final goals. */
1920 match = (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2);
1929 return match ? 0 : 1; /* 0 match, 1 mismatch */