3 * Copyright (C) 2000, 2001, 2002, 2003, 2004, by Larry Wall and others
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 This file contains various utility functions for manipulating UTF8-encoded
33 strings. For the uninitiated, this is a method of representing arbitrary
34 Unicode characters as a variable number of bytes, in such a way that
35 characters in the ASCII range are unmodified, and a zero byte never appears
36 within non-zero characters.
38 =for apidoc A|U8 *|uvuni_to_utf8_flags|U8 *d|UV uv|UV flags
40 Adds the UTF-8 representation of the Unicode codepoint C<uv> to the end
41 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
42 bytes available. The return value is the pointer to the byte after the
43 end of the new character. In other words,
45 d = uvuni_to_utf8_flags(d, uv, flags);
49 d = uvuni_to_utf8(d, uv);
51 (which is equivalent to)
53 d = uvuni_to_utf8_flags(d, uv, 0);
55 is the recommended Unicode-aware way of saying
63 Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
65 if (ckWARN(WARN_UTF8)) {
66 if (UNICODE_IS_SURROGATE(uv) &&
67 !(flags & UNICODE_ALLOW_SURROGATE))
68 Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv);
70 ((uv >= 0xFDD0 && uv <= 0xFDEF &&
71 !(flags & UNICODE_ALLOW_FDD0))
73 ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */
74 !(flags & UNICODE_ALLOW_FFFF))) &&
75 /* UNICODE_ALLOW_SUPER includes
76 * FFFEs and FFFFs beyond 0x10FFFF. */
77 ((uv <= PERL_UNICODE_MAX) ||
78 !(flags & UNICODE_ALLOW_SUPER))
80 Perl_warner(aTHX_ packWARN(WARN_UTF8),
81 "Unicode character 0x%04"UVxf" is illegal", uv);
83 if (UNI_IS_INVARIANT(uv)) {
84 *d++ = (U8)UTF_TO_NATIVE(uv);
89 STRLEN len = UNISKIP(uv);
92 *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK);
93 uv >>= UTF_ACCUMULATION_SHIFT;
95 *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len));
98 #else /* Non loop style */
100 *d++ = (U8)(( uv >> 6) | 0xc0);
101 *d++ = (U8)(( uv & 0x3f) | 0x80);
105 *d++ = (U8)(( uv >> 12) | 0xe0);
106 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
107 *d++ = (U8)(( uv & 0x3f) | 0x80);
111 *d++ = (U8)(( uv >> 18) | 0xf0);
112 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
113 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
114 *d++ = (U8)(( uv & 0x3f) | 0x80);
117 if (uv < 0x4000000) {
118 *d++ = (U8)(( uv >> 24) | 0xf8);
119 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
120 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
121 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
122 *d++ = (U8)(( uv & 0x3f) | 0x80);
125 if (uv < 0x80000000) {
126 *d++ = (U8)(( uv >> 30) | 0xfc);
127 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
128 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
129 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
130 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
131 *d++ = (U8)(( uv & 0x3f) | 0x80);
135 if (uv < UTF8_QUAD_MAX)
138 *d++ = 0xfe; /* Can't match U+FEFF! */
139 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
140 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
141 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
142 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
143 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
144 *d++ = (U8)(( uv & 0x3f) | 0x80);
149 *d++ = 0xff; /* Can't match U+FFFE! */
150 *d++ = 0x80; /* 6 Reserved bits */
151 *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */
152 *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80);
153 *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80);
154 *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80);
155 *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80);
156 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
157 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
158 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
159 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
160 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
161 *d++ = (U8)(( uv & 0x3f) | 0x80);
165 #endif /* Loop style */
169 Perl_uvuni_to_utf8(pTHX_ U8 *d, UV uv)
171 return Perl_uvuni_to_utf8_flags(aTHX_ d, uv, 0);
176 =for apidoc A|STRLEN|is_utf8_char|U8 *s
178 Tests if some arbitrary number of bytes begins in a valid UTF-8
179 character. Note that an INVARIANT (i.e. ASCII) character is a valid
180 UTF-8 character. The actual number of bytes in the UTF-8 character
181 will be returned if it is valid, otherwise 0.
185 Perl_is_utf8_char(pTHX_ U8 *s)
191 if (UTF8_IS_INVARIANT(u))
194 if (!UTF8_IS_START(u))
199 if (len < 2 || !UTF8_IS_CONTINUATION(s[1]))
204 u &= UTF_START_MASK(len);
208 if (!UTF8_IS_CONTINUATION(*s))
210 uv = UTF8_ACCUMULATE(uv, *s);
217 if ((STRLEN)UNISKIP(uv) < len)
224 =for apidoc A|bool|is_utf8_string|U8 *s|STRLEN len
226 Returns true if first C<len> bytes of the given string form a valid
227 UTF-8 string, false otherwise. Note that 'a valid UTF-8 string' does
228 not mean 'a string that contains code points above 0x7F encoded in UTF-8'
229 because a valid ASCII string is a valid UTF-8 string.
235 Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len)
242 len = strlen((char *)s);
246 /* Inline the easy bits of is_utf8_char() here for speed... */
247 if (UTF8_IS_INVARIANT(*x))
249 else if (!UTF8_IS_START(*x))
252 /* ... and call is_utf8_char() only if really needed. */
266 =for apidoc A|bool|is_utf8_string_loc|U8 *s|STRLEN len|U8 **p
268 Like is_ut8_string but store the location of the failure in
275 Perl_is_utf8_string_loc(pTHX_ U8 *s, STRLEN len, U8 **p)
282 len = strlen((char *)s);
286 /* Inline the easy bits of is_utf8_char() here for speed... */
287 if (UTF8_IS_INVARIANT(*x))
289 else if (!UTF8_IS_START(*x)) {
295 /* ... and call is_utf8_char() only if really needed. */
315 =for apidoc A|UV|utf8n_to_uvuni|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
317 Bottom level UTF-8 decode routine.
318 Returns the unicode code point value of the first character in the string C<s>
319 which is assumed to be in UTF-8 encoding and no longer than C<curlen>;
320 C<retlen> will be set to the length, in bytes, of that character.
322 If C<s> does not point to a well-formed UTF-8 character, the behaviour
323 is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY,
324 it is assumed that the caller will raise a warning, and this function
325 will silently just set C<retlen> to C<-1> and return zero. If the
326 C<flags> does not contain UTF8_CHECK_ONLY, warnings about
327 malformations will be given, C<retlen> will be set to the expected
328 length of the UTF-8 character in bytes, and zero will be returned.
330 The C<flags> can also contain various flags to allow deviations from
331 the strict UTF-8 encoding (see F<utf8.h>).
333 Most code should use utf8_to_uvchr() rather than call this directly.
339 Perl_utf8n_to_uvuni(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
344 bool dowarn = ckWARN_d(WARN_UTF8);
346 STRLEN expectlen = 0;
349 /* This list is a superset of the UTF8_ALLOW_XXX. */
351 #define UTF8_WARN_EMPTY 1
352 #define UTF8_WARN_CONTINUATION 2
353 #define UTF8_WARN_NON_CONTINUATION 3
354 #define UTF8_WARN_FE_FF 4
355 #define UTF8_WARN_SHORT 5
356 #define UTF8_WARN_OVERFLOW 6
357 #define UTF8_WARN_SURROGATE 7
358 #define UTF8_WARN_LONG 8
359 #define UTF8_WARN_FFFF 9 /* Also FFFE. */
362 !(flags & UTF8_ALLOW_EMPTY)) {
363 warning = UTF8_WARN_EMPTY;
367 if (UTF8_IS_INVARIANT(uv)) {
370 return (UV) (NATIVE_TO_UTF(*s));
373 if (UTF8_IS_CONTINUATION(uv) &&
374 !(flags & UTF8_ALLOW_CONTINUATION)) {
375 warning = UTF8_WARN_CONTINUATION;
379 if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) &&
380 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
381 warning = UTF8_WARN_NON_CONTINUATION;
386 uv = NATIVE_TO_UTF(uv);
388 if ((uv == 0xfe || uv == 0xff) &&
389 !(flags & UTF8_ALLOW_FE_FF)) {
390 warning = UTF8_WARN_FE_FF;
395 if (!(uv & 0x20)) { len = 2; uv &= 0x1f; }
396 else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; }
397 else if (!(uv & 0x08)) { len = 4; uv &= 0x07; }
398 else if (!(uv & 0x04)) { len = 5; uv &= 0x03; }
400 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
401 else { len = 7; uv &= 0x01; }
403 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
404 else if (!(uv & 0x01)) { len = 7; uv = 0; }
405 else { len = 13; uv = 0; } /* whoa! */
413 if ((curlen < expectlen) &&
414 !(flags & UTF8_ALLOW_SHORT)) {
415 warning = UTF8_WARN_SHORT;
424 if (!UTF8_IS_CONTINUATION(*s) &&
425 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
427 warning = UTF8_WARN_NON_CONTINUATION;
431 uv = UTF8_ACCUMULATE(uv, *s);
433 /* These cannot be allowed. */
435 if (!(flags & UTF8_ALLOW_LONG)) {
436 warning = UTF8_WARN_LONG;
440 else { /* uv < ouv */
441 /* This cannot be allowed. */
442 warning = UTF8_WARN_OVERFLOW;
450 if (UNICODE_IS_SURROGATE(uv) &&
451 !(flags & UTF8_ALLOW_SURROGATE)) {
452 warning = UTF8_WARN_SURROGATE;
454 } else if ((expectlen > (STRLEN)UNISKIP(uv)) &&
455 !(flags & UTF8_ALLOW_LONG)) {
456 warning = UTF8_WARN_LONG;
458 } else if (UNICODE_IS_ILLEGAL(uv) &&
459 !(flags & UTF8_ALLOW_FFFF)) {
460 warning = UTF8_WARN_FFFF;
468 if (flags & UTF8_CHECK_ONLY) {
475 SV* sv = sv_2mortal(newSVpv("Malformed UTF-8 character ", 0));
478 case 0: /* Intentionally empty. */ break;
479 case UTF8_WARN_EMPTY:
480 Perl_sv_catpvf(aTHX_ sv, "(empty string)");
482 case UTF8_WARN_CONTINUATION:
483 Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv);
485 case UTF8_WARN_NON_CONTINUATION:
487 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")",
488 (UV)s[1], startbyte);
490 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)",
491 (UV)s[1], s - s0, s - s0 > 1 ? "s" : "", startbyte, expectlen);
494 case UTF8_WARN_FE_FF:
495 Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv);
497 case UTF8_WARN_SHORT:
498 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
499 curlen, curlen == 1 ? "" : "s", expectlen, startbyte);
500 expectlen = curlen; /* distance for caller to skip */
502 case UTF8_WARN_OVERFLOW:
503 Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")",
506 case UTF8_WARN_SURROGATE:
507 Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv);
510 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
511 expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte);
514 Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv);
517 Perl_sv_catpvf(aTHX_ sv, "(unknown reason)");
525 Perl_warner(aTHX_ packWARN(WARN_UTF8),
526 "%s in %s", s, OP_DESC(PL_op));
528 Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s);
533 *retlen = expectlen ? expectlen : len;
539 =for apidoc A|UV|utf8_to_uvchr|U8 *s|STRLEN *retlen
541 Returns the native character value of the first character in the string C<s>
542 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
543 length, in bytes, of that character.
545 If C<s> does not point to a well-formed UTF-8 character, zero is
546 returned and retlen is set, if possible, to -1.
552 Perl_utf8_to_uvchr(pTHX_ U8 *s, STRLEN *retlen)
554 return Perl_utf8n_to_uvchr(aTHX_ s, UTF8_MAXLEN, retlen,
555 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
559 =for apidoc A|UV|utf8_to_uvuni|U8 *s|STRLEN *retlen
561 Returns the Unicode code point of the first character in the string C<s>
562 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
563 length, in bytes, of that character.
565 This function should only be used when returned UV is considered
566 an index into the Unicode semantic tables (e.g. swashes).
568 If C<s> does not point to a well-formed UTF-8 character, zero is
569 returned and retlen is set, if possible, to -1.
575 Perl_utf8_to_uvuni(pTHX_ U8 *s, STRLEN *retlen)
577 /* Call the low level routine asking for checks */
578 return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXLEN, retlen,
579 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
583 =for apidoc A|STRLEN|utf8_length|U8 *s|U8 *e
585 Return the length of the UTF-8 char encoded string C<s> in characters.
586 Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end
587 up past C<e>, croaks.
593 Perl_utf8_length(pTHX_ U8 *s, U8 *e)
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. */
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);
615 if (ckWARN_d(WARN_UTF8)) {
617 Perl_warner(aTHX_ packWARN(WARN_UTF8),
618 unees, OP_DESC(PL_op));
620 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
632 =for apidoc A|IV|utf8_distance|U8 *a|U8 *b
634 Returns the number of UTF-8 characters between the UTF-8 pointers C<a>
637 WARNING: use only if you *know* that the pointers point inside the
644 Perl_utf8_distance(pTHX_ U8 *a, U8 *b)
648 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
649 * the bitops (especially ~) can create illegal UTF-8.
650 * In other words: in Perl UTF-8 is not just for Unicode. */
657 if (ckWARN_d(WARN_UTF8)) {
659 Perl_warner(aTHX_ packWARN(WARN_UTF8),
660 "%s in %s", unees, OP_DESC(PL_op));
662 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
675 if (ckWARN_d(WARN_UTF8)) {
677 Perl_warner(aTHX_ packWARN(WARN_UTF8),
678 "%s in %s", unees, OP_DESC(PL_op));
680 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
693 =for apidoc A|U8 *|utf8_hop|U8 *s|I32 off
695 Return the UTF-8 pointer C<s> displaced by C<off> characters, either
698 WARNING: do not use the following unless you *know* C<off> is within
699 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
700 on the first byte of character or just after the last byte of a character.
706 Perl_utf8_hop(pTHX_ U8 *s, I32 off)
708 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
709 * the bitops (especially ~) can create illegal UTF-8.
710 * In other words: in Perl UTF-8 is not just for Unicode. */
719 while (UTF8_IS_CONTINUATION(*s))
727 =for apidoc A|U8 *|utf8_to_bytes|U8 *s|STRLEN *len
729 Converts a string C<s> of length C<len> from UTF-8 into byte encoding.
730 Unlike C<bytes_to_utf8>, this over-writes the original string, and
731 updates len to contain the new length.
732 Returns zero on failure, setting C<len> to -1.
738 Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len)
744 /* ensure valid UTF-8 and chars < 256 before updating string */
745 for (send = s + *len; s < send; ) {
748 if (!UTF8_IS_INVARIANT(c) &&
749 (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send)
750 || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) {
759 *d++ = (U8)utf8_to_uvchr(s, &ulen);
768 =for apidoc A|U8 *|bytes_from_utf8|U8 *s|STRLEN *len|bool *is_utf8
770 Converts a string C<s> of length C<len> from UTF-8 into byte encoding.
771 Unlike <utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to
772 the newly-created string, and updates C<len> to contain the new
773 length. Returns the original string if no conversion occurs, C<len>
774 is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to
775 0 if C<s> is converted or contains all 7bit characters.
781 Perl_bytes_from_utf8(pTHX_ U8 *s, STRLEN *len, bool *is_utf8)
791 /* ensure valid UTF-8 and chars < 256 before converting string */
792 for (send = s + *len; s < send;) {
794 if (!UTF8_IS_INVARIANT(c)) {
795 if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send &&
796 (c = *s++) && UTF8_IS_CONTINUATION(c))
805 Newz(801, d, (*len) - count + 1, U8);
806 s = start; start = d;
809 if (!UTF8_IS_INVARIANT(c)) {
810 /* Then it is two-byte encoded */
811 c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++);
812 c = ASCII_TO_NATIVE(c);
822 =for apidoc A|U8 *|bytes_to_utf8|U8 *s|STRLEN *len
824 Converts a string C<s> of length C<len> from ASCII into UTF-8 encoding.
825 Returns a pointer to the newly-created string, and sets C<len> to
826 reflect the new length.
828 If you want to convert to UTF-8 from other encodings than ASCII,
829 see sv_recode_to_utf8().
835 Perl_bytes_to_utf8(pTHX_ U8 *s, STRLEN *len)
842 Newz(801, d, (*len) * 2 + 1, U8);
846 UV uv = NATIVE_TO_ASCII(*s++);
847 if (UNI_IS_INVARIANT(uv))
848 *d++ = (U8)UTF_TO_NATIVE(uv);
850 *d++ = (U8)UTF8_EIGHT_BIT_HI(uv);
851 *d++ = (U8)UTF8_EIGHT_BIT_LO(uv);
860 * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8.
862 * Destination must be pre-extended to 3/2 source. Do not use in-place.
863 * We optimize for native, for obvious reasons. */
866 Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
872 Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen");
877 UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */
884 *d++ = (U8)(( uv >> 6) | 0xc0);
885 *d++ = (U8)(( uv & 0x3f) | 0x80);
888 if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */
889 UV low = (p[0] << 8) + p[1];
891 if (low < 0xdc00 || low >= 0xdfff)
892 Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
893 uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000;
896 *d++ = (U8)(( uv >> 12) | 0xe0);
897 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
898 *d++ = (U8)(( uv & 0x3f) | 0x80);
902 *d++ = (U8)(( uv >> 18) | 0xf0);
903 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
904 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
905 *d++ = (U8)(( uv & 0x3f) | 0x80);
909 *newlen = d - dstart;
913 /* Note: this one is slightly destructive of the source. */
916 Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
919 U8* send = s + bytelen;
926 return utf16_to_utf8(p, d, bytelen, newlen);
929 /* for now these are all defined (inefficiently) in terms of the utf8 versions */
932 Perl_is_uni_alnum(pTHX_ UV c)
934 U8 tmpbuf[UTF8_MAXLEN+1];
935 uvchr_to_utf8(tmpbuf, c);
936 return is_utf8_alnum(tmpbuf);
940 Perl_is_uni_alnumc(pTHX_ UV c)
942 U8 tmpbuf[UTF8_MAXLEN+1];
943 uvchr_to_utf8(tmpbuf, c);
944 return is_utf8_alnumc(tmpbuf);
948 Perl_is_uni_idfirst(pTHX_ UV c)
950 U8 tmpbuf[UTF8_MAXLEN+1];
951 uvchr_to_utf8(tmpbuf, c);
952 return is_utf8_idfirst(tmpbuf);
956 Perl_is_uni_alpha(pTHX_ UV c)
958 U8 tmpbuf[UTF8_MAXLEN+1];
959 uvchr_to_utf8(tmpbuf, c);
960 return is_utf8_alpha(tmpbuf);
964 Perl_is_uni_ascii(pTHX_ UV c)
966 U8 tmpbuf[UTF8_MAXLEN+1];
967 uvchr_to_utf8(tmpbuf, c);
968 return is_utf8_ascii(tmpbuf);
972 Perl_is_uni_space(pTHX_ UV c)
974 U8 tmpbuf[UTF8_MAXLEN+1];
975 uvchr_to_utf8(tmpbuf, c);
976 return is_utf8_space(tmpbuf);
980 Perl_is_uni_digit(pTHX_ UV c)
982 U8 tmpbuf[UTF8_MAXLEN+1];
983 uvchr_to_utf8(tmpbuf, c);
984 return is_utf8_digit(tmpbuf);
988 Perl_is_uni_upper(pTHX_ UV c)
990 U8 tmpbuf[UTF8_MAXLEN+1];
991 uvchr_to_utf8(tmpbuf, c);
992 return is_utf8_upper(tmpbuf);
996 Perl_is_uni_lower(pTHX_ UV c)
998 U8 tmpbuf[UTF8_MAXLEN+1];
999 uvchr_to_utf8(tmpbuf, c);
1000 return is_utf8_lower(tmpbuf);
1004 Perl_is_uni_cntrl(pTHX_ UV c)
1006 U8 tmpbuf[UTF8_MAXLEN+1];
1007 uvchr_to_utf8(tmpbuf, c);
1008 return is_utf8_cntrl(tmpbuf);
1012 Perl_is_uni_graph(pTHX_ UV c)
1014 U8 tmpbuf[UTF8_MAXLEN+1];
1015 uvchr_to_utf8(tmpbuf, c);
1016 return is_utf8_graph(tmpbuf);
1020 Perl_is_uni_print(pTHX_ UV c)
1022 U8 tmpbuf[UTF8_MAXLEN+1];
1023 uvchr_to_utf8(tmpbuf, c);
1024 return is_utf8_print(tmpbuf);
1028 Perl_is_uni_punct(pTHX_ UV c)
1030 U8 tmpbuf[UTF8_MAXLEN+1];
1031 uvchr_to_utf8(tmpbuf, c);
1032 return is_utf8_punct(tmpbuf);
1036 Perl_is_uni_xdigit(pTHX_ UV c)
1038 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1039 uvchr_to_utf8(tmpbuf, c);
1040 return is_utf8_xdigit(tmpbuf);
1044 Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp)
1046 uvchr_to_utf8(p, c);
1047 return to_utf8_upper(p, p, lenp);
1051 Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp)
1053 uvchr_to_utf8(p, c);
1054 return to_utf8_title(p, p, lenp);
1058 Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp)
1060 uvchr_to_utf8(p, c);
1061 return to_utf8_lower(p, p, lenp);
1065 Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp)
1067 uvchr_to_utf8(p, c);
1068 return to_utf8_fold(p, p, lenp);
1071 /* for now these all assume no locale info available for Unicode > 255 */
1074 Perl_is_uni_alnum_lc(pTHX_ UV c)
1076 return is_uni_alnum(c); /* XXX no locale support yet */
1080 Perl_is_uni_alnumc_lc(pTHX_ UV c)
1082 return is_uni_alnumc(c); /* XXX no locale support yet */
1086 Perl_is_uni_idfirst_lc(pTHX_ UV c)
1088 return is_uni_idfirst(c); /* XXX no locale support yet */
1092 Perl_is_uni_alpha_lc(pTHX_ UV c)
1094 return is_uni_alpha(c); /* XXX no locale support yet */
1098 Perl_is_uni_ascii_lc(pTHX_ UV c)
1100 return is_uni_ascii(c); /* XXX no locale support yet */
1104 Perl_is_uni_space_lc(pTHX_ UV c)
1106 return is_uni_space(c); /* XXX no locale support yet */
1110 Perl_is_uni_digit_lc(pTHX_ UV c)
1112 return is_uni_digit(c); /* XXX no locale support yet */
1116 Perl_is_uni_upper_lc(pTHX_ UV c)
1118 return is_uni_upper(c); /* XXX no locale support yet */
1122 Perl_is_uni_lower_lc(pTHX_ UV c)
1124 return is_uni_lower(c); /* XXX no locale support yet */
1128 Perl_is_uni_cntrl_lc(pTHX_ UV c)
1130 return is_uni_cntrl(c); /* XXX no locale support yet */
1134 Perl_is_uni_graph_lc(pTHX_ UV c)
1136 return is_uni_graph(c); /* XXX no locale support yet */
1140 Perl_is_uni_print_lc(pTHX_ UV c)
1142 return is_uni_print(c); /* XXX no locale support yet */
1146 Perl_is_uni_punct_lc(pTHX_ UV c)
1148 return is_uni_punct(c); /* XXX no locale support yet */
1152 Perl_is_uni_xdigit_lc(pTHX_ UV c)
1154 return is_uni_xdigit(c); /* XXX no locale support yet */
1158 Perl_to_uni_upper_lc(pTHX_ U32 c)
1160 /* XXX returns only the first character -- do not use XXX */
1161 /* XXX no locale support yet */
1163 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1164 return (U32)to_uni_upper(c, tmpbuf, &len);
1168 Perl_to_uni_title_lc(pTHX_ U32 c)
1170 /* XXX returns only the first character XXX -- do not use XXX */
1171 /* XXX no locale support yet */
1173 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1174 return (U32)to_uni_title(c, tmpbuf, &len);
1178 Perl_to_uni_lower_lc(pTHX_ U32 c)
1180 /* XXX returns only the first character -- do not use XXX */
1181 /* XXX no locale support yet */
1183 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1184 return (U32)to_uni_lower(c, tmpbuf, &len);
1188 Perl_is_utf8_alnum(pTHX_ U8 *p)
1190 if (!is_utf8_char(p))
1193 /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true
1194 * descendant of isalnum(3), in other words, it doesn't
1195 * contain the '_'. --jhi */
1196 PL_utf8_alnum = swash_init("utf8", "IsWord", &PL_sv_undef, 0, 0);
1197 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1198 /* return *p == '_' || is_utf8_alpha(p) || is_utf8_digit(p); */
1199 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1201 PL_utf8_alnum = swash_init("utf8", "",
1202 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1203 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1208 Perl_is_utf8_alnumc(pTHX_ U8 *p)
1210 if (!is_utf8_char(p))
1213 PL_utf8_alnum = swash_init("utf8", "IsAlnumC", &PL_sv_undef, 0, 0);
1214 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1215 /* return is_utf8_alpha(p) || is_utf8_digit(p); */
1216 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1218 PL_utf8_alnum = swash_init("utf8", "",
1219 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1220 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1225 Perl_is_utf8_idfirst(pTHX_ U8 *p) /* The naming is historical. */
1229 if (!is_utf8_char(p))
1231 if (!PL_utf8_idstart) /* is_utf8_idstart would be more logical. */
1232 PL_utf8_idstart = swash_init("utf8", "IdStart", &PL_sv_undef, 0, 0);
1233 return swash_fetch(PL_utf8_idstart, p, TRUE) != 0;
1237 Perl_is_utf8_idcont(pTHX_ U8 *p)
1241 if (!is_utf8_char(p))
1243 if (!PL_utf8_idcont)
1244 PL_utf8_idcont = swash_init("utf8", "IdContinue", &PL_sv_undef, 0, 0);
1245 return swash_fetch(PL_utf8_idcont, p, TRUE) != 0;
1249 Perl_is_utf8_alpha(pTHX_ U8 *p)
1251 if (!is_utf8_char(p))
1254 PL_utf8_alpha = swash_init("utf8", "IsAlpha", &PL_sv_undef, 0, 0);
1255 return swash_fetch(PL_utf8_alpha, p, TRUE) != 0;
1259 Perl_is_utf8_ascii(pTHX_ U8 *p)
1261 if (!is_utf8_char(p))
1264 PL_utf8_ascii = swash_init("utf8", "IsAscii", &PL_sv_undef, 0, 0);
1265 return swash_fetch(PL_utf8_ascii, p, TRUE) != 0;
1269 Perl_is_utf8_space(pTHX_ U8 *p)
1271 if (!is_utf8_char(p))
1274 PL_utf8_space = swash_init("utf8", "IsSpacePerl", &PL_sv_undef, 0, 0);
1275 return swash_fetch(PL_utf8_space, p, TRUE) != 0;
1279 Perl_is_utf8_digit(pTHX_ U8 *p)
1281 if (!is_utf8_char(p))
1284 PL_utf8_digit = swash_init("utf8", "IsDigit", &PL_sv_undef, 0, 0);
1285 return swash_fetch(PL_utf8_digit, p, TRUE) != 0;
1289 Perl_is_utf8_upper(pTHX_ U8 *p)
1291 if (!is_utf8_char(p))
1294 PL_utf8_upper = swash_init("utf8", "IsUppercase", &PL_sv_undef, 0, 0);
1295 return swash_fetch(PL_utf8_upper, p, TRUE) != 0;
1299 Perl_is_utf8_lower(pTHX_ U8 *p)
1301 if (!is_utf8_char(p))
1304 PL_utf8_lower = swash_init("utf8", "IsLowercase", &PL_sv_undef, 0, 0);
1305 return swash_fetch(PL_utf8_lower, p, TRUE) != 0;
1309 Perl_is_utf8_cntrl(pTHX_ U8 *p)
1311 if (!is_utf8_char(p))
1314 PL_utf8_cntrl = swash_init("utf8", "IsCntrl", &PL_sv_undef, 0, 0);
1315 return swash_fetch(PL_utf8_cntrl, p, TRUE) != 0;
1319 Perl_is_utf8_graph(pTHX_ U8 *p)
1321 if (!is_utf8_char(p))
1324 PL_utf8_graph = swash_init("utf8", "IsGraph", &PL_sv_undef, 0, 0);
1325 return swash_fetch(PL_utf8_graph, p, TRUE) != 0;
1329 Perl_is_utf8_print(pTHX_ U8 *p)
1331 if (!is_utf8_char(p))
1334 PL_utf8_print = swash_init("utf8", "IsPrint", &PL_sv_undef, 0, 0);
1335 return swash_fetch(PL_utf8_print, p, TRUE) != 0;
1339 Perl_is_utf8_punct(pTHX_ U8 *p)
1341 if (!is_utf8_char(p))
1344 PL_utf8_punct = swash_init("utf8", "IsPunct", &PL_sv_undef, 0, 0);
1345 return swash_fetch(PL_utf8_punct, p, TRUE) != 0;
1349 Perl_is_utf8_xdigit(pTHX_ U8 *p)
1351 if (!is_utf8_char(p))
1353 if (!PL_utf8_xdigit)
1354 PL_utf8_xdigit = swash_init("utf8", "IsXDigit", &PL_sv_undef, 0, 0);
1355 return swash_fetch(PL_utf8_xdigit, p, TRUE) != 0;
1359 Perl_is_utf8_mark(pTHX_ U8 *p)
1361 if (!is_utf8_char(p))
1364 PL_utf8_mark = swash_init("utf8", "IsM", &PL_sv_undef, 0, 0);
1365 return swash_fetch(PL_utf8_mark, p, TRUE) != 0;
1369 =for apidoc A|UV|to_utf8_case|U8 *p|U8* ustrp|STRLEN *lenp|SV **swash|char *normal|char *special
1371 The "p" contains the pointer to the UTF-8 string encoding
1372 the character that is being converted.
1374 The "ustrp" is a pointer to the character buffer to put the
1375 conversion result to. The "lenp" is a pointer to the length
1378 The "swashp" is a pointer to the swash to use.
1380 Both the special and normal mappings are stored lib/unicore/To/Foo.pl,
1381 and loaded by SWASHGET, using lib/utf8_heavy.pl. The special (usually,
1382 but not always, a multicharacter mapping), is tried first.
1384 The "special" is a string like "utf8::ToSpecLower", which means the
1385 hash %utf8::ToSpecLower. The access to the hash is through
1386 Perl_to_utf8_case().
1388 The "normal" is a string like "ToLower" which means the swash
1394 Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp, char *normal, char *special)
1397 U8 tmpbuf[UTF8_MAXLEN_FOLD+1];
1400 uv0 = utf8_to_uvchr(p, 0);
1401 /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings
1402 * are necessary in EBCDIC, they are redundant no-ops
1403 * in ASCII-ish platforms, and hopefully optimized away. */
1404 uv1 = NATIVE_TO_UNI(uv0);
1405 uvuni_to_utf8(tmpbuf, uv1);
1407 if (!*swashp) /* load on-demand */
1408 *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0);
1410 /* The 0xDF is the only special casing Unicode code point below 0x100. */
1411 if (special && (uv1 == 0xDF || uv1 > 0xFF)) {
1412 /* It might be "special" (sometimes, but not always,
1413 * a multicharacter mapping) */
1417 if ((hv = get_hv(special, FALSE)) &&
1418 (svp = hv_fetch(hv, (const char*)tmpbuf, UNISKIP(uv1), FALSE)) &&
1422 s = SvPV(*svp, len);
1424 len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp;
1427 /* If we have EBCDIC we need to remap the characters
1428 * since any characters in the low 256 are Unicode
1429 * code points, not EBCDIC. */
1430 U8 *t = (U8*)s, *tend = t + len, *d;
1437 UV c = utf8_to_uvchr(t, &tlen);
1439 d = uvchr_to_utf8(d, UNI_TO_NATIVE(c));
1448 d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t));
1453 Copy(tmpbuf, ustrp, len, U8);
1455 Copy(s, ustrp, len, U8);
1461 if (!len && *swashp) {
1462 UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE);
1465 /* It was "normal" (a single character mapping). */
1466 UV uv3 = UNI_TO_NATIVE(uv2);
1468 len = uvchr_to_utf8(ustrp, uv3) - ustrp;
1472 if (!len) /* Neither: just copy. */
1473 len = uvchr_to_utf8(ustrp, uv0) - ustrp;
1478 return len ? utf8_to_uvchr(ustrp, 0) : 0;
1482 =for apidoc A|UV|to_utf8_upper|U8 *p|U8 *ustrp|STRLEN *lenp
1484 Convert the UTF-8 encoded character at p to its uppercase version and
1485 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1486 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1487 uppercase version may be longer than the original character (up to two
1490 The first character of the uppercased version is returned
1491 (but note, as explained above, that there may be more.)
1496 Perl_to_utf8_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1498 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1499 &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper");
1503 =for apidoc A|UV|to_utf8_title|U8 *p|U8 *ustrp|STRLEN *lenp
1505 Convert the UTF-8 encoded character at p to its titlecase version and
1506 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1507 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1508 titlecase version may be longer than the original character (up to two
1511 The first character of the titlecased version is returned
1512 (but note, as explained above, that there may be more.)
1517 Perl_to_utf8_title(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1519 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1520 &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle");
1524 =for apidoc A|UV|to_utf8_lower|U8 *p|U8 *ustrp|STRLEN *lenp
1526 Convert the UTF-8 encoded character at p to its lowercase version and
1527 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1528 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1529 lowercase version may be longer than the original character (up to two
1532 The first character of the lowercased version is returned
1533 (but note, as explained above, that there may be more.)
1538 Perl_to_utf8_lower(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1540 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1541 &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower");
1545 =for apidoc A|UV|to_utf8_fold|U8 *p|U8 *ustrp|STRLEN *lenp
1547 Convert the UTF-8 encoded character at p to its foldcase version and
1548 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1549 that the ustrp needs to be at least UTF8_MAXLEN_FOLD+1 bytes since the
1550 foldcase version may be longer than the original character (up to
1553 The first character of the foldcased version is returned
1554 (but note, as explained above, that there may be more.)
1559 Perl_to_utf8_fold(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1561 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1562 &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold");
1565 /* a "swash" is a swatch hash */
1568 Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none)
1571 SV* tokenbufsv = sv_2mortal(NEWSV(0,0));
1573 size_t pkg_len = strlen(pkg);
1574 size_t name_len = strlen(name);
1575 HV *stash = gv_stashpvn(pkg, pkg_len, FALSE);
1578 if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */
1580 errsv_save = newSVsv(ERRSV);
1581 Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpvn(pkg,pkg_len),
1584 sv_setsv(ERRSV, errsv_save);
1585 SvREFCNT_dec(errsv_save);
1589 PUSHSTACKi(PERLSI_MAGIC);
1592 PUSHs(sv_2mortal(newSVpvn(pkg, pkg_len)));
1593 PUSHs(sv_2mortal(newSVpvn(name, name_len)));
1595 PUSHs(sv_2mortal(newSViv(minbits)));
1596 PUSHs(sv_2mortal(newSViv(none)));
1602 if (IN_PERL_COMPILETIME) {
1603 /* XXX ought to be handled by lex_start */
1606 sv_setpv(tokenbufsv, PL_tokenbuf);
1608 errsv_save = newSVsv(ERRSV);
1609 if (call_method("SWASHNEW", G_SCALAR))
1610 retval = newSVsv(*PL_stack_sp--);
1612 retval = &PL_sv_undef;
1614 sv_setsv(ERRSV, errsv_save);
1615 SvREFCNT_dec(errsv_save);
1618 if (IN_PERL_COMPILETIME) {
1620 char* pv = SvPV(tokenbufsv, len);
1622 Copy(pv, PL_tokenbuf, len+1, char);
1623 PL_curcop->op_private = (U8)(PL_hints & HINT_PRIVATE_MASK);
1625 if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) {
1627 Perl_croak(aTHX_ "Can't find Unicode property definition \"%"SVf"\"",
1629 Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref");
1635 /* This API is wrong for special case conversions since we may need to
1636 * return several Unicode characters for a single Unicode character
1637 * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is
1638 * the lower-level routine, and it is similarly broken for returning
1639 * multiple values. --jhi */
1641 Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr, bool do_utf8)
1643 HV* hv = (HV*)SvRV(sv);
1652 UV c = NATIVE_TO_ASCII(*ptr);
1654 if (!do_utf8 && !UNI_IS_INVARIANT(c)) {
1655 tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c);
1656 tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c);
1659 /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ
1660 * then the "swatch" is a vec() for al the chars which start
1662 * So the key in the hash (klen) is length of encoded char -1
1664 klen = UTF8SKIP(ptr) - 1;
1669 /* If char in invariant then swatch is for all the invariant chars
1670 * In both UTF-8 and UTF-8-MOD that happens to be UTF_CONTINUATION_MARK
1672 needents = UTF_CONTINUATION_MARK;
1673 off = NATIVE_TO_UTF(ptr[klen]);
1677 /* If char is encoded then swatch is for the prefix */
1678 needents = (1 << UTF_ACCUMULATION_SHIFT);
1679 off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK;
1683 * This single-entry cache saves about 1/3 of the utf8 overhead in test
1684 * suite. (That is, only 7-8% overall over just a hash cache. Still,
1685 * it's nothing to sniff at.) Pity we usually come through at least
1686 * two function calls to get here...
1688 * NB: this code assumes that swatches are never modified, once generated!
1691 if (hv == PL_last_swash_hv &&
1692 klen == PL_last_swash_klen &&
1693 (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) )
1695 tmps = PL_last_swash_tmps;
1696 slen = PL_last_swash_slen;
1699 /* Try our second-level swatch cache, kept in a hash. */
1700 SV** svp = hv_fetch(hv, (char*)ptr, klen, FALSE);
1702 /* If not cached, generate it via utf8::SWASHGET */
1703 if (!svp || !SvPOK(*svp) || !(tmps = (U8*)SvPV(*svp, slen))) {
1705 /* We use utf8n_to_uvuni() as we want an index into
1706 Unicode tables, not a native character number.
1708 UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXLEN, 0,
1710 0 : UTF8_ALLOW_ANY);
1715 PUSHSTACKi(PERLSI_MAGIC);
1719 /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */
1720 PUSHs(sv_2mortal(newSViv((klen) ?
1721 (code_point & ~(needents - 1)) : 0)));
1722 PUSHs(sv_2mortal(newSViv(needents)));
1724 errsv_save = newSVsv(ERRSV);
1725 if (call_method("SWASHGET", G_SCALAR))
1726 retval = newSVsv(*PL_stack_sp--);
1728 retval = &PL_sv_undef;
1730 sv_setsv(ERRSV, errsv_save);
1731 SvREFCNT_dec(errsv_save);
1735 if (IN_PERL_COMPILETIME)
1736 PL_curcop->op_private = (U8)(PL_hints & HINT_PRIVATE_MASK);
1738 svp = hv_store(hv, (char*)ptr, klen, retval, 0);
1740 if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || (slen << 3) < needents)
1741 Perl_croak(aTHX_ "SWASHGET didn't return result of proper length");
1744 PL_last_swash_hv = hv;
1745 PL_last_swash_klen = klen;
1746 PL_last_swash_tmps = tmps;
1747 PL_last_swash_slen = slen;
1749 Copy(ptr, PL_last_swash_key, klen, U8);
1752 switch ((int)((slen << 3) / needents)) {
1754 bit = 1 << (off & 7);
1756 return (tmps[off] & bit) != 0;
1761 return (tmps[off] << 8) + tmps[off + 1] ;
1764 return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ;
1766 Perl_croak(aTHX_ "panic: swash_fetch");
1772 =for apidoc A|U8 *|uvchr_to_utf8|U8 *d|UV uv
1774 Adds the UTF-8 representation of the Native codepoint C<uv> to the end
1775 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
1776 bytes available. The return value is the pointer to the byte after the
1777 end of the new character. In other words,
1779 d = uvchr_to_utf8(d, uv);
1781 is the recommended wide native character-aware way of saying
1788 /* On ASCII machines this is normally a macro but we want a
1789 real function in case XS code wants it
1791 #undef Perl_uvchr_to_utf8
1793 Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv)
1795 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0);
1799 Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
1801 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags);
1805 =for apidoc A|UV|utf8n_to_uvchr|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
1807 Returns the native character value of the first character in the string C<s>
1808 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
1809 length, in bytes, of that character.
1811 Allows length and flags to be passed to low level routine.
1815 /* On ASCII machines this is normally a macro but we want
1816 a real function in case XS code wants it
1818 #undef Perl_utf8n_to_uvchr
1820 Perl_utf8n_to_uvchr(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
1822 UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags);
1823 return UNI_TO_NATIVE(uv);
1827 =for apidoc A|char *|pv_uni_display|SV *dsv|U8 *spv|STRLEN len|STRLEN pvlim|UV flags
1829 Build to the scalar dsv a displayable version of the string spv,
1830 length len, the displayable version being at most pvlim bytes long
1831 (if longer, the rest is truncated and "..." will be appended).
1833 The flags argument can have UNI_DISPLAY_ISPRINT set to display
1834 isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH
1835 to display the \\[nrfta\\] as the backslashed versions (like '\n')
1836 (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\).
1837 UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both
1838 UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on.
1840 The pointer to the PV of the dsv is returned.
1844 Perl_pv_uni_display(pTHX_ SV *dsv, U8 *spv, STRLEN len, STRLEN pvlim, UV flags)
1849 sv_setpvn(dsv, "", 0);
1850 for (s = (char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) {
1854 if (pvlim && SvCUR(dsv) >= pvlim) {
1858 u = utf8_to_uvchr((U8*)s, 0);
1860 if (!ok && (flags & UNI_DISPLAY_BACKSLASH)) {
1863 Perl_sv_catpvf(aTHX_ dsv, "\\n"); ok = TRUE; break;
1865 Perl_sv_catpvf(aTHX_ dsv, "\\r"); ok = TRUE; break;
1867 Perl_sv_catpvf(aTHX_ dsv, "\\t"); ok = TRUE; break;
1869 Perl_sv_catpvf(aTHX_ dsv, "\\f"); ok = TRUE; break;
1871 Perl_sv_catpvf(aTHX_ dsv, "\\a"); ok = TRUE; break;
1873 Perl_sv_catpvf(aTHX_ dsv, "\\\\" ); ok = TRUE; break;
1877 /* isPRINT() is the locale-blind version. */
1878 if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(u & 0xFF)) {
1879 Perl_sv_catpvf(aTHX_ dsv, "%c", (char)(u & 0xFF));
1884 Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u);
1887 sv_catpvn(dsv, "...", 3);
1893 =for apidoc A|char *|sv_uni_display|SV *dsv|SV *ssv|STRLEN pvlim|UV flags
1895 Build to the scalar dsv a displayable version of the scalar sv,
1896 the displayable version being at most pvlim bytes long
1897 (if longer, the rest is truncated and "..." will be appended).
1899 The flags argument is as in pv_uni_display().
1901 The pointer to the PV of the dsv is returned.
1905 Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags)
1907 return Perl_pv_uni_display(aTHX_ dsv, (U8*)SvPVX(ssv), SvCUR(ssv),
1912 =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
1914 Return true if the strings s1 and s2 differ case-insensitively, false
1915 if not (if they are equal case-insensitively). If u1 is true, the
1916 string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true,
1917 the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2
1918 are false, the respective string is assumed to be in native 8-bit
1921 If the pe1 and pe2 are non-NULL, the scanning pointers will be copied
1922 in there (they will point at the beginning of the I<next> character).
1923 If the pointers behind pe1 or pe2 are non-NULL, they are the end
1924 pointers beyond which scanning will not continue under any
1925 circustances. If the byte lengths l1 and l2 are non-zero, s1+l1 and
1926 s2+l2 will be used as goal end pointers that will also stop the scan,
1927 and which qualify towards defining a successful match: all the scans
1928 that define an explicit length must reach their goal pointers for
1929 a match to succeed).
1931 For case-insensitiveness, the "casefolding" of Unicode is used
1932 instead of upper/lowercasing both the characters, see
1933 http://www.unicode.org/unicode/reports/tr21/ (Case Mappings).
1937 Perl_ibcmp_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2)
1939 register U8 *p1 = (U8*)s1;
1940 register U8 *p2 = (U8*)s2;
1941 register U8 *e1 = 0, *f1 = 0, *q1 = 0;
1942 register U8 *e2 = 0, *f2 = 0, *q2 = 0;
1943 STRLEN n1 = 0, n2 = 0;
1944 U8 foldbuf1[UTF8_MAXLEN_FOLD+1];
1945 U8 foldbuf2[UTF8_MAXLEN_FOLD+1];
1947 STRLEN foldlen1, foldlen2;
1952 if (e1 == 0 || (l1 && l1 < (UV)(e1 - (U8*)s1)))
1956 if (e2 == 0 || (l2 && l2 < (UV)(e2 - (U8*)s2)))
1959 if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0))
1960 return 1; /* mismatch; possible infinite loop or false positive */
1963 natbuf[1] = 0; /* Need to terminate the buffer. */
1965 while ((e1 == 0 || p1 < e1) &&
1966 (f1 == 0 || p1 < f1) &&
1967 (e2 == 0 || p2 < e2) &&
1968 (f2 == 0 || p2 < f2)) {
1971 to_utf8_fold(p1, foldbuf1, &foldlen1);
1974 to_utf8_fold(natbuf, foldbuf1, &foldlen1);
1981 to_utf8_fold(p2, foldbuf2, &foldlen2);
1984 to_utf8_fold(natbuf, foldbuf2, &foldlen2);
1990 if ( UTF8SKIP(q1) != UTF8SKIP(q2) ||
1991 (UTF8SKIP(q1) == 1 && *q1 != *q2) ||
1992 memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) )
1993 return 1; /* mismatch */
2000 p1 += u1 ? UTF8SKIP(p1) : 1;
2002 p2 += u2 ? UTF8SKIP(p2) : 1;
2006 /* A match is defined by all the scans that specified
2007 * an explicit length reaching their final goals. */
2008 match = (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2);
2017 return match ? 0 : 1; /* 0 match, 1 mismatch */