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++ = (U8)UTF_TO_NATIVE(uv);
83 STRLEN len = UNISKIP(uv);
86 *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK);
87 uv >>= UTF_ACCUMULATION_SHIFT;
89 *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len));
92 #else /* Non loop style */
94 *d++ = (U8)(( uv >> 6) | 0xc0);
95 *d++ = (U8)(( uv & 0x3f) | 0x80);
99 *d++ = (U8)(( uv >> 12) | 0xe0);
100 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
101 *d++ = (U8)(( uv & 0x3f) | 0x80);
105 *d++ = (U8)(( uv >> 18) | 0xf0);
106 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
107 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
108 *d++ = (U8)(( uv & 0x3f) | 0x80);
111 if (uv < 0x4000000) {
112 *d++ = (U8)(( uv >> 24) | 0xf8);
113 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
114 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
115 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
116 *d++ = (U8)(( uv & 0x3f) | 0x80);
119 if (uv < 0x80000000) {
120 *d++ = (U8)(( uv >> 30) | 0xfc);
121 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
122 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
123 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
124 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
125 *d++ = (U8)(( uv & 0x3f) | 0x80);
129 if (uv < UTF8_QUAD_MAX)
132 *d++ = 0xfe; /* Can't match U+FEFF! */
133 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
134 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
135 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
136 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
137 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
138 *d++ = (U8)(( uv & 0x3f) | 0x80);
143 *d++ = 0xff; /* Can't match U+FFFE! */
144 *d++ = 0x80; /* 6 Reserved bits */
145 *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */
146 *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80);
147 *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80);
148 *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80);
149 *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80);
150 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
151 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
152 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
153 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
154 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
155 *d++ = (U8)(( 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
174 UTF-8 character. The actual number of bytes in the UTF-8 character
175 will be returned if it is valid, otherwise 0.
179 Perl_is_utf8_char(pTHX_ U8 *s)
185 if (UTF8_IS_INVARIANT(u))
188 if (!UTF8_IS_START(u))
193 if (len < 2 || !UTF8_IS_CONTINUATION(s[1]))
198 u &= UTF_START_MASK(len);
202 if (!UTF8_IS_CONTINUATION(*s))
204 uv = UTF8_ACCUMULATE(uv, *s);
211 if ((STRLEN)UNISKIP(uv) < len)
218 =for apidoc A|bool|is_utf8_string|U8 *s|STRLEN len
220 Returns true if first C<len> bytes of the given string form a valid UTF8
221 string, false otherwise. Note that 'a valid UTF8 string' does not mean
222 'a string that contains UTF8' because a valid ASCII string is a valid
229 Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len)
236 len = strlen((char *)s);
240 /* Inline the easy bits of is_utf8_char() here for speed... */
241 if (UTF8_IS_INVARIANT(*x))
243 else if (!UTF8_IS_START(*x))
246 /* ... and call is_utf8_char() only if really needed. */
260 =for apidoc A|UV|utf8n_to_uvuni|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
262 Bottom level UTF-8 decode routine.
263 Returns the unicode code point value of the first character in the string C<s>
264 which is assumed to be in UTF8 encoding and no longer than C<curlen>;
265 C<retlen> will be set to the length, in bytes, of that character.
267 If C<s> does not point to a well-formed UTF8 character, the behaviour
268 is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY,
269 it is assumed that the caller will raise a warning, and this function
270 will silently just set C<retlen> to C<-1> and return zero. If the
271 C<flags> does not contain UTF8_CHECK_ONLY, warnings about
272 malformations will be given, C<retlen> will be set to the expected
273 length of the UTF-8 character in bytes, and zero will be returned.
275 The C<flags> can also contain various flags to allow deviations from
276 the strict UTF-8 encoding (see F<utf8.h>).
278 Most code should use utf8_to_uvchr() rather than call this directly.
284 Perl_utf8n_to_uvuni(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
289 bool dowarn = ckWARN_d(WARN_UTF8);
291 STRLEN expectlen = 0;
294 /* This list is a superset of the UTF8_ALLOW_XXX. */
296 #define UTF8_WARN_EMPTY 1
297 #define UTF8_WARN_CONTINUATION 2
298 #define UTF8_WARN_NON_CONTINUATION 3
299 #define UTF8_WARN_FE_FF 4
300 #define UTF8_WARN_SHORT 5
301 #define UTF8_WARN_OVERFLOW 6
302 #define UTF8_WARN_SURROGATE 7
303 #define UTF8_WARN_LONG 8
304 #define UTF8_WARN_FFFF 9 /* Also FFFE. */
307 !(flags & UTF8_ALLOW_EMPTY)) {
308 warning = UTF8_WARN_EMPTY;
312 if (UTF8_IS_INVARIANT(uv)) {
315 return (UV) (NATIVE_TO_UTF(*s));
318 if (UTF8_IS_CONTINUATION(uv) &&
319 !(flags & UTF8_ALLOW_CONTINUATION)) {
320 warning = UTF8_WARN_CONTINUATION;
324 if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) &&
325 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
326 warning = UTF8_WARN_NON_CONTINUATION;
331 uv = NATIVE_TO_UTF(uv);
333 if ((uv == 0xfe || uv == 0xff) &&
334 !(flags & UTF8_ALLOW_FE_FF)) {
335 warning = UTF8_WARN_FE_FF;
340 if (!(uv & 0x20)) { len = 2; uv &= 0x1f; }
341 else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; }
342 else if (!(uv & 0x08)) { len = 4; uv &= 0x07; }
343 else if (!(uv & 0x04)) { len = 5; uv &= 0x03; }
345 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
346 else { len = 7; uv &= 0x01; }
348 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
349 else if (!(uv & 0x01)) { len = 7; uv = 0; }
350 else { len = 13; uv = 0; } /* whoa! */
358 if ((curlen < expectlen) &&
359 !(flags & UTF8_ALLOW_SHORT)) {
360 warning = UTF8_WARN_SHORT;
369 if (!UTF8_IS_CONTINUATION(*s) &&
370 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
372 warning = UTF8_WARN_NON_CONTINUATION;
376 uv = UTF8_ACCUMULATE(uv, *s);
378 /* These cannot be allowed. */
380 if (!(flags & UTF8_ALLOW_LONG)) {
381 warning = UTF8_WARN_LONG;
385 else { /* uv < ouv */
386 /* This cannot be allowed. */
387 warning = UTF8_WARN_OVERFLOW;
395 if (UNICODE_IS_SURROGATE(uv) &&
396 !(flags & UTF8_ALLOW_SURROGATE)) {
397 warning = UTF8_WARN_SURROGATE;
399 } else if ((expectlen > (STRLEN)UNISKIP(uv)) &&
400 !(flags & UTF8_ALLOW_LONG)) {
401 warning = UTF8_WARN_LONG;
403 } else if (UNICODE_IS_ILLEGAL(uv) &&
404 !(flags & UTF8_ALLOW_FFFF)) {
405 warning = UTF8_WARN_FFFF;
413 if (flags & UTF8_CHECK_ONLY) {
420 SV* sv = sv_2mortal(newSVpv("Malformed UTF-8 character ", 0));
423 case 0: /* Intentionally empty. */ break;
424 case UTF8_WARN_EMPTY:
425 Perl_sv_catpvf(aTHX_ sv, "(empty string)");
427 case UTF8_WARN_CONTINUATION:
428 Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv);
430 case UTF8_WARN_NON_CONTINUATION:
432 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")",
433 (UV)s[1], startbyte);
435 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)",
436 (UV)s[1], s - s0, s - s0 > 1 ? "s" : "", startbyte, expectlen);
439 case UTF8_WARN_FE_FF:
440 Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv);
442 case UTF8_WARN_SHORT:
443 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
444 curlen, curlen == 1 ? "" : "s", expectlen, startbyte);
445 expectlen = curlen; /* distance for caller to skip */
447 case UTF8_WARN_OVERFLOW:
448 Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")",
451 case UTF8_WARN_SURROGATE:
452 Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv);
455 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
456 expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte);
459 Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv);
462 Perl_sv_catpvf(aTHX_ sv, "(unknown reason)");
470 Perl_warner(aTHX_ packWARN(WARN_UTF8),
471 "%s in %s", s, OP_DESC(PL_op));
473 Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s);
478 *retlen = expectlen ? expectlen : len;
484 =for apidoc A|UV|utf8_to_uvchr|U8 *s|STRLEN *retlen
486 Returns the native character value of the first character in the string C<s>
487 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
488 length, in bytes, of that character.
490 If C<s> does not point to a well-formed UTF8 character, zero is
491 returned and retlen is set, if possible, to -1.
497 Perl_utf8_to_uvchr(pTHX_ U8 *s, STRLEN *retlen)
499 return Perl_utf8n_to_uvchr(aTHX_ s, UTF8_MAXLEN, retlen,
500 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
504 =for apidoc A|UV|utf8_to_uvuni|U8 *s|STRLEN *retlen
506 Returns the Unicode code point of the first character in the string C<s>
507 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
508 length, in bytes, of that character.
510 This function should only be used when returned UV is considered
511 an index into the Unicode semantic tables (e.g. swashes).
513 If C<s> does not point to a well-formed UTF8 character, zero is
514 returned and retlen is set, if possible, to -1.
520 Perl_utf8_to_uvuni(pTHX_ U8 *s, STRLEN *retlen)
522 /* Call the low level routine asking for checks */
523 return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXLEN, retlen,
524 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
528 =for apidoc A|STRLEN|utf8_length|U8 *s|U8 *e
530 Return the length of the UTF-8 char encoded string C<s> in characters.
531 Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end
532 up past C<e>, croaks.
538 Perl_utf8_length(pTHX_ U8 *s, U8 *e)
542 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
543 * the bitops (especially ~) can create illegal UTF-8.
544 * In other words: in Perl UTF-8 is not just for Unicode. */
547 if (ckWARN_d(WARN_UTF8)) {
549 Perl_warner(aTHX_ packWARN(WARN_UTF8),
550 "%s in %s", unees, OP_DESC(PL_op));
552 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
560 if (ckWARN_d(WARN_UTF8)) {
562 Perl_warner(aTHX_ packWARN(WARN_UTF8),
563 unees, OP_DESC(PL_op));
565 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
577 =for apidoc A|IV|utf8_distance|U8 *a|U8 *b
579 Returns the number of UTF8 characters between the UTF-8 pointers C<a>
582 WARNING: use only if you *know* that the pointers point inside the
589 Perl_utf8_distance(pTHX_ U8 *a, U8 *b)
593 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
594 * the bitops (especially ~) can create illegal UTF-8.
595 * In other words: in Perl UTF-8 is not just for Unicode. */
602 if (ckWARN_d(WARN_UTF8)) {
604 Perl_warner(aTHX_ packWARN(WARN_UTF8),
605 "%s in %s", unees, OP_DESC(PL_op));
607 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
620 if (ckWARN_d(WARN_UTF8)) {
622 Perl_warner(aTHX_ packWARN(WARN_UTF8),
623 "%s in %s", unees, OP_DESC(PL_op));
625 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
638 =for apidoc A|U8 *|utf8_hop|U8 *s|I32 off
640 Return the UTF-8 pointer C<s> displaced by C<off> characters, either
643 WARNING: do not use the following unless you *know* C<off> is within
644 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
645 on the first byte of character or just after the last byte of a character.
651 Perl_utf8_hop(pTHX_ U8 *s, I32 off)
653 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
654 * the bitops (especially ~) can create illegal UTF-8.
655 * In other words: in Perl UTF-8 is not just for Unicode. */
664 while (UTF8_IS_CONTINUATION(*s))
672 =for apidoc A|U8 *|utf8_to_bytes|U8 *s|STRLEN *len
674 Converts a string C<s> of length C<len> from UTF8 into byte encoding.
675 Unlike C<bytes_to_utf8>, this over-writes the original string, and
676 updates len to contain the new length.
677 Returns zero on failure, setting C<len> to -1.
683 Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len)
689 /* ensure valid UTF8 and chars < 256 before updating string */
690 for (send = s + *len; s < send; ) {
693 if (!UTF8_IS_INVARIANT(c) &&
694 (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send)
695 || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) {
704 *d++ = (U8)utf8_to_uvchr(s, &ulen);
713 =for apidoc A|U8 *|bytes_from_utf8|U8 *s|STRLEN *len|bool *is_utf8
715 Converts a string C<s> of length C<len> from UTF8 into byte encoding.
716 Unlike <utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to
717 the newly-created string, and updates C<len> to contain the new
718 length. Returns the original string if no conversion occurs, C<len>
719 is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to
720 0 if C<s> is converted or contains all 7bit characters.
726 Perl_bytes_from_utf8(pTHX_ U8 *s, STRLEN *len, bool *is_utf8)
736 /* ensure valid UTF8 and chars < 256 before converting string */
737 for (send = s + *len; s < send;) {
739 if (!UTF8_IS_INVARIANT(c)) {
740 if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send &&
741 (c = *s++) && UTF8_IS_CONTINUATION(c))
750 Newz(801, d, (*len) - count + 1, U8);
751 s = start; start = d;
754 if (!UTF8_IS_INVARIANT(c)) {
755 /* Then it is two-byte encoded */
756 c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++);
757 c = ASCII_TO_NATIVE(c);
767 =for apidoc A|U8 *|bytes_to_utf8|U8 *s|STRLEN *len
769 Converts a string C<s> of length C<len> from ASCII into UTF8 encoding.
770 Returns a pointer to the newly-created string, and sets C<len> to
771 reflect the new length.
777 Perl_bytes_to_utf8(pTHX_ U8 *s, STRLEN *len)
784 Newz(801, d, (*len) * 2 + 1, U8);
788 UV uv = NATIVE_TO_ASCII(*s++);
789 if (UNI_IS_INVARIANT(uv))
790 *d++ = (U8)UTF_TO_NATIVE(uv);
792 *d++ = (U8)UTF8_EIGHT_BIT_HI(uv);
793 *d++ = (U8)UTF8_EIGHT_BIT_LO(uv);
802 * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8.
804 * Destination must be pre-extended to 3/2 source. Do not use in-place.
805 * We optimize for native, for obvious reasons. */
808 Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
814 Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen");
819 UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */
826 *d++ = (U8)(( uv >> 6) | 0xc0);
827 *d++ = (U8)(( uv & 0x3f) | 0x80);
830 if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */
832 if (low < 0xdc00 || low >= 0xdfff)
833 Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
834 uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000;
837 *d++ = (U8)(( uv >> 12) | 0xe0);
838 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
839 *d++ = (U8)(( uv & 0x3f) | 0x80);
843 *d++ = (U8)(( uv >> 18) | 0xf0);
844 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
845 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
846 *d++ = (U8)(( uv & 0x3f) | 0x80);
850 *newlen = d - dstart;
854 /* Note: this one is slightly destructive of the source. */
857 Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
860 U8* send = s + bytelen;
867 return utf16_to_utf8(p, d, bytelen, newlen);
870 /* for now these are all defined (inefficiently) in terms of the utf8 versions */
873 Perl_is_uni_alnum(pTHX_ UV c)
875 U8 tmpbuf[UTF8_MAXLEN+1];
876 uvchr_to_utf8(tmpbuf, c);
877 return is_utf8_alnum(tmpbuf);
881 Perl_is_uni_alnumc(pTHX_ UV c)
883 U8 tmpbuf[UTF8_MAXLEN+1];
884 uvchr_to_utf8(tmpbuf, c);
885 return is_utf8_alnumc(tmpbuf);
889 Perl_is_uni_idfirst(pTHX_ UV c)
891 U8 tmpbuf[UTF8_MAXLEN+1];
892 uvchr_to_utf8(tmpbuf, c);
893 return is_utf8_idfirst(tmpbuf);
897 Perl_is_uni_alpha(pTHX_ UV c)
899 U8 tmpbuf[UTF8_MAXLEN+1];
900 uvchr_to_utf8(tmpbuf, c);
901 return is_utf8_alpha(tmpbuf);
905 Perl_is_uni_ascii(pTHX_ UV c)
907 U8 tmpbuf[UTF8_MAXLEN+1];
908 uvchr_to_utf8(tmpbuf, c);
909 return is_utf8_ascii(tmpbuf);
913 Perl_is_uni_space(pTHX_ UV c)
915 U8 tmpbuf[UTF8_MAXLEN+1];
916 uvchr_to_utf8(tmpbuf, c);
917 return is_utf8_space(tmpbuf);
921 Perl_is_uni_digit(pTHX_ UV c)
923 U8 tmpbuf[UTF8_MAXLEN+1];
924 uvchr_to_utf8(tmpbuf, c);
925 return is_utf8_digit(tmpbuf);
929 Perl_is_uni_upper(pTHX_ UV c)
931 U8 tmpbuf[UTF8_MAXLEN+1];
932 uvchr_to_utf8(tmpbuf, c);
933 return is_utf8_upper(tmpbuf);
937 Perl_is_uni_lower(pTHX_ UV c)
939 U8 tmpbuf[UTF8_MAXLEN+1];
940 uvchr_to_utf8(tmpbuf, c);
941 return is_utf8_lower(tmpbuf);
945 Perl_is_uni_cntrl(pTHX_ UV c)
947 U8 tmpbuf[UTF8_MAXLEN+1];
948 uvchr_to_utf8(tmpbuf, c);
949 return is_utf8_cntrl(tmpbuf);
953 Perl_is_uni_graph(pTHX_ UV c)
955 U8 tmpbuf[UTF8_MAXLEN+1];
956 uvchr_to_utf8(tmpbuf, c);
957 return is_utf8_graph(tmpbuf);
961 Perl_is_uni_print(pTHX_ UV c)
963 U8 tmpbuf[UTF8_MAXLEN+1];
964 uvchr_to_utf8(tmpbuf, c);
965 return is_utf8_print(tmpbuf);
969 Perl_is_uni_punct(pTHX_ UV c)
971 U8 tmpbuf[UTF8_MAXLEN+1];
972 uvchr_to_utf8(tmpbuf, c);
973 return is_utf8_punct(tmpbuf);
977 Perl_is_uni_xdigit(pTHX_ UV c)
979 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
980 uvchr_to_utf8(tmpbuf, c);
981 return is_utf8_xdigit(tmpbuf);
985 Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp)
988 return to_utf8_upper(p, p, lenp);
992 Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp)
995 return to_utf8_title(p, p, lenp);
999 Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp)
1001 uvchr_to_utf8(p, c);
1002 return to_utf8_lower(p, p, lenp);
1006 Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp)
1008 uvchr_to_utf8(p, c);
1009 return to_utf8_fold(p, p, lenp);
1012 /* for now these all assume no locale info available for Unicode > 255 */
1015 Perl_is_uni_alnum_lc(pTHX_ UV c)
1017 return is_uni_alnum(c); /* XXX no locale support yet */
1021 Perl_is_uni_alnumc_lc(pTHX_ UV c)
1023 return is_uni_alnumc(c); /* XXX no locale support yet */
1027 Perl_is_uni_idfirst_lc(pTHX_ UV c)
1029 return is_uni_idfirst(c); /* XXX no locale support yet */
1033 Perl_is_uni_alpha_lc(pTHX_ UV c)
1035 return is_uni_alpha(c); /* XXX no locale support yet */
1039 Perl_is_uni_ascii_lc(pTHX_ UV c)
1041 return is_uni_ascii(c); /* XXX no locale support yet */
1045 Perl_is_uni_space_lc(pTHX_ UV c)
1047 return is_uni_space(c); /* XXX no locale support yet */
1051 Perl_is_uni_digit_lc(pTHX_ UV c)
1053 return is_uni_digit(c); /* XXX no locale support yet */
1057 Perl_is_uni_upper_lc(pTHX_ UV c)
1059 return is_uni_upper(c); /* XXX no locale support yet */
1063 Perl_is_uni_lower_lc(pTHX_ UV c)
1065 return is_uni_lower(c); /* XXX no locale support yet */
1069 Perl_is_uni_cntrl_lc(pTHX_ UV c)
1071 return is_uni_cntrl(c); /* XXX no locale support yet */
1075 Perl_is_uni_graph_lc(pTHX_ UV c)
1077 return is_uni_graph(c); /* XXX no locale support yet */
1081 Perl_is_uni_print_lc(pTHX_ UV c)
1083 return is_uni_print(c); /* XXX no locale support yet */
1087 Perl_is_uni_punct_lc(pTHX_ UV c)
1089 return is_uni_punct(c); /* XXX no locale support yet */
1093 Perl_is_uni_xdigit_lc(pTHX_ UV c)
1095 return is_uni_xdigit(c); /* XXX no locale support yet */
1099 Perl_to_uni_upper_lc(pTHX_ U32 c)
1101 /* XXX returns only the first character -- do not use XXX */
1102 /* XXX no locale support yet */
1104 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1105 return (U32)to_uni_upper(c, tmpbuf, &len);
1109 Perl_to_uni_title_lc(pTHX_ U32 c)
1111 /* XXX returns only the first character XXX -- do not use XXX */
1112 /* XXX no locale support yet */
1114 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1115 return (U32)to_uni_title(c, tmpbuf, &len);
1119 Perl_to_uni_lower_lc(pTHX_ U32 c)
1121 /* XXX returns only the first character -- do not use XXX */
1122 /* XXX no locale support yet */
1124 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1125 return (U32)to_uni_lower(c, tmpbuf, &len);
1129 Perl_is_utf8_alnum(pTHX_ U8 *p)
1131 if (!is_utf8_char(p))
1134 /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true
1135 * descendant of isalnum(3), in other words, it doesn't
1136 * contain the '_'. --jhi */
1137 PL_utf8_alnum = swash_init("utf8", "IsWord", &PL_sv_undef, 0, 0);
1138 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1139 /* return *p == '_' || is_utf8_alpha(p) || is_utf8_digit(p); */
1140 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1142 PL_utf8_alnum = swash_init("utf8", "",
1143 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1144 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1149 Perl_is_utf8_alnumc(pTHX_ U8 *p)
1151 if (!is_utf8_char(p))
1154 PL_utf8_alnum = swash_init("utf8", "IsAlnumC", &PL_sv_undef, 0, 0);
1155 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1156 /* return is_utf8_alpha(p) || is_utf8_digit(p); */
1157 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1159 PL_utf8_alnum = swash_init("utf8", "",
1160 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1161 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1166 Perl_is_utf8_idfirst(pTHX_ U8 *p) /* The naming is historical. */
1170 if (!is_utf8_char(p))
1172 if (!PL_utf8_idstart) /* is_utf8_idstart would be more logical. */
1173 PL_utf8_idstart = swash_init("utf8", "IdStart", &PL_sv_undef, 0, 0);
1174 return swash_fetch(PL_utf8_idstart, p, TRUE) != 0;
1178 Perl_is_utf8_idcont(pTHX_ U8 *p)
1182 if (!is_utf8_char(p))
1184 if (!PL_utf8_idcont)
1185 PL_utf8_idcont = swash_init("utf8", "IdContinue", &PL_sv_undef, 0, 0);
1186 return swash_fetch(PL_utf8_idcont, p, TRUE) != 0;
1190 Perl_is_utf8_alpha(pTHX_ U8 *p)
1192 if (!is_utf8_char(p))
1195 PL_utf8_alpha = swash_init("utf8", "IsAlpha", &PL_sv_undef, 0, 0);
1196 return swash_fetch(PL_utf8_alpha, p, TRUE) != 0;
1200 Perl_is_utf8_ascii(pTHX_ U8 *p)
1202 if (!is_utf8_char(p))
1205 PL_utf8_ascii = swash_init("utf8", "IsAscii", &PL_sv_undef, 0, 0);
1206 return swash_fetch(PL_utf8_ascii, p, TRUE) != 0;
1210 Perl_is_utf8_space(pTHX_ U8 *p)
1212 if (!is_utf8_char(p))
1215 PL_utf8_space = swash_init("utf8", "IsSpacePerl", &PL_sv_undef, 0, 0);
1216 return swash_fetch(PL_utf8_space, p, TRUE) != 0;
1220 Perl_is_utf8_digit(pTHX_ U8 *p)
1222 if (!is_utf8_char(p))
1225 PL_utf8_digit = swash_init("utf8", "IsDigit", &PL_sv_undef, 0, 0);
1226 return swash_fetch(PL_utf8_digit, p, TRUE) != 0;
1230 Perl_is_utf8_upper(pTHX_ U8 *p)
1232 if (!is_utf8_char(p))
1235 PL_utf8_upper = swash_init("utf8", "IsUpper", &PL_sv_undef, 0, 0);
1236 return swash_fetch(PL_utf8_upper, p, TRUE) != 0;
1240 Perl_is_utf8_lower(pTHX_ U8 *p)
1242 if (!is_utf8_char(p))
1245 PL_utf8_lower = swash_init("utf8", "IsLower", &PL_sv_undef, 0, 0);
1246 return swash_fetch(PL_utf8_lower, p, TRUE) != 0;
1250 Perl_is_utf8_cntrl(pTHX_ U8 *p)
1252 if (!is_utf8_char(p))
1255 PL_utf8_cntrl = swash_init("utf8", "IsCntrl", &PL_sv_undef, 0, 0);
1256 return swash_fetch(PL_utf8_cntrl, p, TRUE) != 0;
1260 Perl_is_utf8_graph(pTHX_ U8 *p)
1262 if (!is_utf8_char(p))
1265 PL_utf8_graph = swash_init("utf8", "IsGraph", &PL_sv_undef, 0, 0);
1266 return swash_fetch(PL_utf8_graph, p, TRUE) != 0;
1270 Perl_is_utf8_print(pTHX_ U8 *p)
1272 if (!is_utf8_char(p))
1275 PL_utf8_print = swash_init("utf8", "IsPrint", &PL_sv_undef, 0, 0);
1276 return swash_fetch(PL_utf8_print, p, TRUE) != 0;
1280 Perl_is_utf8_punct(pTHX_ U8 *p)
1282 if (!is_utf8_char(p))
1285 PL_utf8_punct = swash_init("utf8", "IsPunct", &PL_sv_undef, 0, 0);
1286 return swash_fetch(PL_utf8_punct, p, TRUE) != 0;
1290 Perl_is_utf8_xdigit(pTHX_ U8 *p)
1292 if (!is_utf8_char(p))
1294 if (!PL_utf8_xdigit)
1295 PL_utf8_xdigit = swash_init("utf8", "IsXDigit", &PL_sv_undef, 0, 0);
1296 return swash_fetch(PL_utf8_xdigit, p, TRUE) != 0;
1300 Perl_is_utf8_mark(pTHX_ U8 *p)
1302 if (!is_utf8_char(p))
1305 PL_utf8_mark = swash_init("utf8", "IsM", &PL_sv_undef, 0, 0);
1306 return swash_fetch(PL_utf8_mark, p, TRUE) != 0;
1310 =for apidoc A|UV|to_utf8_case|U8 *p|U8* ustrp|STRLEN *lenp|SV **swash|char *normal|char *special
1312 The "p" contains the pointer to the UTF-8 string encoding
1313 the character that is being converted.
1315 The "ustrp" is a pointer to the character buffer to put the
1316 conversion result to. The "lenp" is a pointer to the length
1319 The "swashp" is a pointer to the swash to use.
1321 Both the special and normal mappings are stored lib/unicore/To/Foo.pl,
1322 and loaded by SWASHGET, using lib/utf8_heavy.pl. The special (usually,
1323 but not always, a multicharacter mapping), is tried first.
1325 The "special" is a string like "utf8::ToSpecLower", which means the
1326 hash %utf8::ToSpecLower. The access to the hash is through
1327 Perl_to_utf8_case().
1329 The "normal" is a string like "ToLower" which means the swash
1335 Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp, char *normal, char *special)
1338 U8 tmpbuf[UTF8_MAXLEN_FOLD+1];
1341 uv0 = utf8_to_uvchr(p, 0);
1342 /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings
1343 * are necessary in EBCDIC, they are redundant no-ops
1344 * in ASCII-ish platforms, and hopefully optimized away. */
1345 uv1 = NATIVE_TO_UNI(uv0);
1346 uvuni_to_utf8(tmpbuf, uv1);
1348 if (!*swashp) /* load on-demand */
1349 *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0);
1352 /* It might be "special" (sometimes, but not always,
1353 * a multicharacter mapping) */
1359 if ((hv = get_hv(special, FALSE)) &&
1360 (keysv = sv_2mortal(Perl_newSVpvf(aTHX_ "%04"UVXf, uv1))) &&
1361 (he = hv_fetch_ent(hv, keysv, FALSE, 0)) &&
1362 (val = HeVAL(he))) {
1367 len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp;
1370 /* If we have EBCDIC we need to remap the characters
1371 * since any characters in the low 256 are Unicode
1372 * code points, not EBCDIC. */
1373 U8 *t = (U8*)s, *tend = t + len, *d;
1380 UV c = utf8_to_uvchr(t, &tlen);
1382 d = uvchr_to_utf8(d, UNI_TO_NATIVE(c));
1391 d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t));
1396 Copy(tmpbuf, ustrp, len, U8);
1398 Copy(s, ustrp, len, U8);
1404 if (!len && *swashp) {
1405 UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE);
1408 /* It was "normal" (a single character mapping). */
1409 UV uv3 = UNI_TO_NATIVE(uv2);
1411 len = uvchr_to_utf8(ustrp, uv3) - ustrp;
1415 if (!len) /* Neither: just copy. */
1416 len = uvchr_to_utf8(ustrp, uv0) - ustrp;
1421 return len ? utf8_to_uvchr(ustrp, 0) : 0;
1425 =for apidoc A|UV|to_utf8_upper|U8 *p|U8 *ustrp|STRLEN *lenp
1427 Convert the UTF-8 encoded character at p to its uppercase version and
1428 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1429 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1430 uppercase version may be longer than the original character (up to two
1433 The first character of the uppercased version is returned
1434 (but note, as explained above, that there may be more.)
1439 Perl_to_utf8_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1441 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1442 &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper");
1446 =for apidoc A|UV|to_utf8_title|U8 *p|U8 *ustrp|STRLEN *lenp
1448 Convert the UTF-8 encoded character at p to its titlecase version and
1449 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1450 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1451 titlecase version may be longer than the original character (up to two
1454 The first character of the titlecased version is returned
1455 (but note, as explained above, that there may be more.)
1460 Perl_to_utf8_title(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1462 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1463 &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle");
1467 =for apidoc A|UV|to_utf8_lower|U8 *p|U8 *ustrp|STRLEN *lenp
1469 Convert the UTF-8 encoded character at p to its lowercase version and
1470 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1471 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1472 lowercase version may be longer than the original character (up to two
1475 The first character of the lowercased version is returned
1476 (but note, as explained above, that there may be more.)
1481 Perl_to_utf8_lower(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1483 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1484 &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower");
1488 =for apidoc A|UV|to_utf8_fold|U8 *p|U8 *ustrp|STRLEN *lenp
1490 Convert the UTF-8 encoded character at p to its foldcase version and
1491 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1492 that the ustrp needs to be at least UTF8_MAXLEN_FOLD+1 bytes since the
1493 foldcase version may be longer than the original character (up to
1496 The first character of the foldcased version is returned
1497 (but note, as explained above, that there may be more.)
1502 Perl_to_utf8_fold(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1504 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1505 &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold");
1508 /* a "swash" is a swatch hash */
1511 Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none)
1514 SV* tokenbufsv = sv_2mortal(NEWSV(0,0));
1516 HV *stash = gv_stashpvn(pkg, strlen(pkg), FALSE);
1519 if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */
1521 errsv_save = newSVsv(ERRSV);
1522 Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpv(pkg,0), Nullsv);
1524 sv_setsv(ERRSV, errsv_save);
1525 SvREFCNT_dec(errsv_save);
1529 PUSHSTACKi(PERLSI_MAGIC);
1532 PUSHs(sv_2mortal(newSVpvn(pkg, strlen(pkg))));
1533 PUSHs(sv_2mortal(newSVpvn(name, strlen(name))));
1535 PUSHs(sv_2mortal(newSViv(minbits)));
1536 PUSHs(sv_2mortal(newSViv(none)));
1542 if (PL_curcop == &PL_compiling) {
1543 /* XXX ought to be handled by lex_start */
1545 sv_setpv(tokenbufsv, PL_tokenbuf);
1547 errsv_save = newSVsv(ERRSV);
1548 if (call_method("SWASHNEW", G_SCALAR))
1549 retval = newSVsv(*PL_stack_sp--);
1551 retval = &PL_sv_undef;
1553 sv_setsv(ERRSV, errsv_save);
1554 SvREFCNT_dec(errsv_save);
1557 if (PL_curcop == &PL_compiling) {
1559 char* pv = SvPV(tokenbufsv, len);
1561 Copy(pv, PL_tokenbuf, len+1, char);
1562 PL_curcop->op_private = (U8)(PL_hints & HINT_PRIVATE_MASK);
1564 if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) {
1566 Perl_croak(aTHX_ "Can't find Unicode property definition \"%s\"",
1567 SvPV_nolen(retval));
1568 Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref");
1574 /* This API is wrong for special case conversions since we may need to
1575 * return several Unicode characters for a single Unicode character
1576 * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is
1577 * the lower-level routine, and it is similarly broken for returning
1578 * multiple values. --jhi */
1580 Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr, bool do_utf8)
1582 HV* hv = (HV*)SvRV(sv);
1591 UV c = NATIVE_TO_ASCII(*ptr);
1593 if (!do_utf8 && !UNI_IS_INVARIANT(c)) {
1594 tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c);
1595 tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c);
1598 /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ
1599 * then the "swatch" is a vec() for al the chars which start
1601 * So the key in the hash (klen) is length of encoded char -1
1603 klen = UTF8SKIP(ptr) - 1;
1608 /* If char in invariant then swatch is for all the invariant chars
1609 * In both UTF-8 and UTF8-MOD that happens to be UTF_CONTINUATION_MARK
1611 needents = UTF_CONTINUATION_MARK;
1612 off = NATIVE_TO_UTF(ptr[klen]);
1616 /* If char is encoded then swatch is for the prefix */
1617 needents = (1 << UTF_ACCUMULATION_SHIFT);
1618 off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK;
1622 * This single-entry cache saves about 1/3 of the utf8 overhead in test
1623 * suite. (That is, only 7-8% overall over just a hash cache. Still,
1624 * it's nothing to sniff at.) Pity we usually come through at least
1625 * two function calls to get here...
1627 * NB: this code assumes that swatches are never modified, once generated!
1630 if (hv == PL_last_swash_hv &&
1631 klen == PL_last_swash_klen &&
1632 (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) )
1634 tmps = PL_last_swash_tmps;
1635 slen = PL_last_swash_slen;
1638 /* Try our second-level swatch cache, kept in a hash. */
1639 SV** svp = hv_fetch(hv, (char*)ptr, klen, FALSE);
1641 /* If not cached, generate it via utf8::SWASHGET */
1642 if (!svp || !SvPOK(*svp) || !(tmps = (U8*)SvPV(*svp, slen))) {
1644 /* We use utf8n_to_uvuni() as we want an index into
1645 Unicode tables, not a native character number.
1647 UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXLEN, 0,
1649 0 : UTF8_ALLOW_ANY);
1654 PUSHSTACKi(PERLSI_MAGIC);
1658 /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */
1659 PUSHs(sv_2mortal(newSViv((klen) ?
1660 (code_point & ~(needents - 1)) : 0)));
1661 PUSHs(sv_2mortal(newSViv(needents)));
1663 errsv_save = newSVsv(ERRSV);
1664 if (call_method("SWASHGET", G_SCALAR))
1665 retval = newSVsv(*PL_stack_sp--);
1667 retval = &PL_sv_undef;
1669 sv_setsv(ERRSV, errsv_save);
1670 SvREFCNT_dec(errsv_save);
1674 if (PL_curcop == &PL_compiling)
1675 PL_curcop->op_private = (U8)(PL_hints & HINT_PRIVATE_MASK);
1677 svp = hv_store(hv, (char*)ptr, klen, retval, 0);
1679 if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || (slen << 3) < needents)
1680 Perl_croak(aTHX_ "SWASHGET didn't return result of proper length");
1683 PL_last_swash_hv = hv;
1684 PL_last_swash_klen = klen;
1685 PL_last_swash_tmps = tmps;
1686 PL_last_swash_slen = slen;
1688 Copy(ptr, PL_last_swash_key, klen, U8);
1691 switch ((int)((slen << 3) / needents)) {
1693 bit = 1 << (off & 7);
1695 return (tmps[off] & bit) != 0;
1700 return (tmps[off] << 8) + tmps[off + 1] ;
1703 return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ;
1705 Perl_croak(aTHX_ "panic: swash_fetch");
1711 =for apidoc A|U8 *|uvchr_to_utf8|U8 *d|UV uv
1713 Adds the UTF8 representation of the Native codepoint C<uv> to the end
1714 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
1715 bytes available. The return value is the pointer to the byte after the
1716 end of the new character. In other words,
1718 d = uvchr_to_utf8(d, uv);
1720 is the recommended wide native character-aware way of saying
1727 /* On ASCII machines this is normally a macro but we want a
1728 real function in case XS code wants it
1730 #undef Perl_uvchr_to_utf8
1732 Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv)
1734 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0);
1738 Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
1740 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags);
1744 =for apidoc A|UV|utf8n_to_uvchr|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
1746 Returns the native character value of the first character in the string C<s>
1747 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
1748 length, in bytes, of that character.
1750 Allows length and flags to be passed to low level routine.
1754 /* On ASCII machines this is normally a macro but we want
1755 a real function in case XS code wants it
1757 #undef Perl_utf8n_to_uvchr
1759 Perl_utf8n_to_uvchr(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
1761 UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags);
1762 return UNI_TO_NATIVE(uv);
1766 =for apidoc A|char *|pv_uni_display|SV *dsv|U8 *spv|STRLEN len|STRLEN pvlim|UV flags
1768 Build to the scalar dsv a displayable version of the string spv,
1769 length len, the displayable version being at most pvlim bytes long
1770 (if longer, the rest is truncated and "..." will be appended).
1772 The flags argument can have UNI_DISPLAY_ISPRINT set to display
1773 isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH
1774 to display the \\[nrfta\\] as the backslashed versions (like '\n')
1775 (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\).
1776 UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both
1777 UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on.
1779 The pointer to the PV of the dsv is returned.
1783 Perl_pv_uni_display(pTHX_ SV *dsv, U8 *spv, STRLEN len, STRLEN pvlim, UV flags)
1788 sv_setpvn(dsv, "", 0);
1789 for (s = (char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) {
1793 if (pvlim && SvCUR(dsv) >= pvlim) {
1797 u = utf8_to_uvchr((U8*)s, 0);
1799 if (!ok && (flags & UNI_DISPLAY_BACKSLASH)) {
1802 Perl_sv_catpvf(aTHX_ dsv, "\\n"); ok = TRUE; break;
1804 Perl_sv_catpvf(aTHX_ dsv, "\\r"); ok = TRUE; break;
1806 Perl_sv_catpvf(aTHX_ dsv, "\\t"); ok = TRUE; break;
1808 Perl_sv_catpvf(aTHX_ dsv, "\\f"); ok = TRUE; break;
1810 Perl_sv_catpvf(aTHX_ dsv, "\\a"); ok = TRUE; break;
1812 Perl_sv_catpvf(aTHX_ dsv, "\\\\" ); ok = TRUE; break;
1816 /* isPRINT() is the locale-blind version. */
1817 if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(u & 0xFF)) {
1818 Perl_sv_catpvf(aTHX_ dsv, "%c", (char)(u & 0xFF));
1823 Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u);
1826 sv_catpvn(dsv, "...", 3);
1832 =for apidoc A|char *|sv_uni_display|SV *dsv|SV *ssv|STRLEN pvlim|UV flags
1834 Build to the scalar dsv a displayable version of the scalar sv,
1835 the displayable version being at most pvlim bytes long
1836 (if longer, the rest is truncated and "..." will be appended).
1838 The flags argument is as in pv_uni_display().
1840 The pointer to the PV of the dsv is returned.
1844 Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags)
1846 return Perl_pv_uni_display(aTHX_ dsv, (U8*)SvPVX(ssv), SvCUR(ssv),
1851 =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
1853 Return true if the strings s1 and s2 differ case-insensitively, false
1854 if not (if they are equal case-insensitively). If u1 is true, the
1855 string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true,
1856 the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2
1857 are false, the respective string is assumed to be in native 8-bit
1860 If the pe1 and pe2 are non-NULL, the scanning pointers will be copied
1861 in there (they will point at the beginning of the I<next> character).
1862 If the pointers behind pe1 or pe2 are non-NULL, they are the end
1863 pointers beyond which scanning will not continue under any
1864 circustances. If the byte lengths l1 and l2 are non-zero, s1+l1 and
1865 s2+l2 will be used as goal end pointers that will also stop the scan,
1866 and which qualify towards defining a successful match: all the scans
1867 that define an explicit length must reach their goal pointers for
1868 a match to succeed).
1870 For case-insensitiveness, the "casefolding" of Unicode is used
1871 instead of upper/lowercasing both the characters, see
1872 http://www.unicode.org/unicode/reports/tr21/ (Case Mappings).
1876 Perl_ibcmp_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2)
1878 register U8 *p1 = (U8*)s1;
1879 register U8 *p2 = (U8*)s2;
1880 register U8 *e1 = 0, *f1 = 0, *q1 = 0;
1881 register U8 *e2 = 0, *f2 = 0, *q2 = 0;
1882 STRLEN n1 = 0, n2 = 0;
1883 U8 foldbuf1[UTF8_MAXLEN_FOLD+1];
1884 U8 foldbuf2[UTF8_MAXLEN_FOLD+1];
1886 STRLEN foldlen1, foldlen2;
1891 if (e1 == 0 || (l1 && l1 < (UV)(e1 - (U8*)s1)))
1895 if (e2 == 0 || (l2 && l2 < (UV)(e2 - (U8*)s2)))
1898 if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0))
1899 return 1; /* mismatch; possible infinite loop or false positive */
1902 natbuf[1] = 0; /* Need to terminate the buffer. */
1904 while ((e1 == 0 || p1 < e1) &&
1905 (f1 == 0 || p1 < f1) &&
1906 (e2 == 0 || p2 < e2) &&
1907 (f2 == 0 || p2 < f2)) {
1910 to_utf8_fold(p1, foldbuf1, &foldlen1);
1913 to_utf8_fold(natbuf, foldbuf1, &foldlen1);
1920 to_utf8_fold(p2, foldbuf2, &foldlen2);
1923 to_utf8_fold(natbuf, foldbuf2, &foldlen2);
1929 if ( UTF8SKIP(q1) != UTF8SKIP(q2) ||
1930 (UTF8SKIP(q1) == 1 && *q1 != *q2) ||
1931 memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) )
1932 return 1; /* mismatch */
1939 p1 += u1 ? UTF8SKIP(p1) : 1;
1941 p2 += u2 ? UTF8SKIP(p2) : 1;
1945 /* A match is defined by all the scans that specified
1946 * an explicit length reaching their final goals. */
1947 match = (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2);
1956 return match ? 0 : 1; /* 0 match, 1 mismatch */