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.
37 =for apidoc A|U8 *|uvuni_to_utf8_flags|U8 *d|UV uv|UV flags
39 Adds the UTF-8 representation of the Unicode codepoint C<uv> to the end
40 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
41 bytes available. The return value is the pointer to the byte after the
42 end of the new character. In other words,
44 d = uvuni_to_utf8_flags(d, uv, flags);
48 d = uvuni_to_utf8(d, uv);
50 (which is equivalent to)
52 d = uvuni_to_utf8_flags(d, uv, 0);
54 is the recommended Unicode-aware way of saying
62 Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
64 if (ckWARN(WARN_UTF8)) {
65 if (UNICODE_IS_SURROGATE(uv) &&
66 !(flags & UNICODE_ALLOW_SURROGATE))
67 Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv);
69 ((uv >= 0xFDD0 && uv <= 0xFDEF &&
70 !(flags & UNICODE_ALLOW_FDD0))
72 ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */
73 !(flags & UNICODE_ALLOW_FFFF))) &&
74 /* UNICODE_ALLOW_SUPER includes
75 * FFFEs and FFFFs beyond 0x10FFFF. */
76 ((uv <= PERL_UNICODE_MAX) ||
77 !(flags & UNICODE_ALLOW_SUPER))
79 Perl_warner(aTHX_ packWARN(WARN_UTF8),
80 "Unicode character 0x%04"UVxf" is illegal", uv);
82 if (UNI_IS_INVARIANT(uv)) {
83 *d++ = (U8)UTF_TO_NATIVE(uv);
88 STRLEN len = UNISKIP(uv);
91 *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK);
92 uv >>= UTF_ACCUMULATION_SHIFT;
94 *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len));
97 #else /* Non loop style */
99 *d++ = (U8)(( uv >> 6) | 0xc0);
100 *d++ = (U8)(( uv & 0x3f) | 0x80);
104 *d++ = (U8)(( uv >> 12) | 0xe0);
105 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
106 *d++ = (U8)(( uv & 0x3f) | 0x80);
110 *d++ = (U8)(( uv >> 18) | 0xf0);
111 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
112 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
113 *d++ = (U8)(( uv & 0x3f) | 0x80);
116 if (uv < 0x4000000) {
117 *d++ = (U8)(( uv >> 24) | 0xf8);
118 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
119 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
120 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
121 *d++ = (U8)(( uv & 0x3f) | 0x80);
124 if (uv < 0x80000000) {
125 *d++ = (U8)(( uv >> 30) | 0xfc);
126 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
127 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
128 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
129 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
130 *d++ = (U8)(( uv & 0x3f) | 0x80);
134 if (uv < UTF8_QUAD_MAX)
137 *d++ = 0xfe; /* Can't match U+FEFF! */
138 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
139 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
140 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
141 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
142 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
143 *d++ = (U8)(( uv & 0x3f) | 0x80);
148 *d++ = 0xff; /* Can't match U+FFFE! */
149 *d++ = 0x80; /* 6 Reserved bits */
150 *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */
151 *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80);
152 *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80);
153 *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80);
154 *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80);
155 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
156 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
157 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
158 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
159 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
160 *d++ = (U8)(( uv & 0x3f) | 0x80);
164 #endif /* Loop style */
168 Perl_uvuni_to_utf8(pTHX_ U8 *d, UV uv)
170 return Perl_uvuni_to_utf8_flags(aTHX_ d, uv, 0);
175 =for apidoc A|STRLEN|is_utf8_char|U8 *s
177 Tests if some arbitrary number of bytes begins in a valid UTF-8
178 character. Note that an INVARIANT (i.e. ASCII) character is a valid
179 UTF-8 character. The actual number of bytes in the UTF-8 character
180 will be returned if it is valid, otherwise 0.
184 Perl_is_utf8_char(pTHX_ U8 *s)
190 if (UTF8_IS_INVARIANT(u))
193 if (!UTF8_IS_START(u))
198 if (len < 2 || !UTF8_IS_CONTINUATION(s[1]))
203 u &= UTF_START_MASK(len);
207 if (!UTF8_IS_CONTINUATION(*s))
209 uv = UTF8_ACCUMULATE(uv, *s);
216 if ((STRLEN)UNISKIP(uv) < len)
223 =for apidoc A|bool|is_utf8_string|U8 *s|STRLEN len
225 Returns true if first C<len> bytes of the given string form a valid
226 UTF-8 string, false otherwise. Note that 'a valid UTF-8 string' does
227 not mean 'a string that contains code points above 0x7F encoded in UTF-8'
228 because a valid ASCII string is a valid UTF-8 string.
234 Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len)
241 len = strlen((char *)s);
245 /* Inline the easy bits of is_utf8_char() here for speed... */
246 if (UTF8_IS_INVARIANT(*x))
248 else if (!UTF8_IS_START(*x))
251 /* ... and call is_utf8_char() only if really needed. */
265 =for apidoc A|bool|is_utf8_string_loc|U8 *s|STRLEN len|U8 **p
267 Like is_ut8_string but store the location of the failure in
274 Perl_is_utf8_string_loc(pTHX_ U8 *s, STRLEN len, U8 **p)
281 len = strlen((char *)s);
285 /* Inline the easy bits of is_utf8_char() here for speed... */
286 if (UTF8_IS_INVARIANT(*x))
288 else if (!UTF8_IS_START(*x)) {
294 /* ... and call is_utf8_char() only if really needed. */
314 =for apidoc A|UV|utf8n_to_uvuni|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
316 Bottom level UTF-8 decode routine.
317 Returns the unicode code point value of the first character in the string C<s>
318 which is assumed to be in UTF-8 encoding and no longer than C<curlen>;
319 C<retlen> will be set to the length, in bytes, of that character.
321 If C<s> does not point to a well-formed UTF-8 character, the behaviour
322 is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY,
323 it is assumed that the caller will raise a warning, and this function
324 will silently just set C<retlen> to C<-1> and return zero. If the
325 C<flags> does not contain UTF8_CHECK_ONLY, warnings about
326 malformations will be given, C<retlen> will be set to the expected
327 length of the UTF-8 character in bytes, and zero will be returned.
329 The C<flags> can also contain various flags to allow deviations from
330 the strict UTF-8 encoding (see F<utf8.h>).
332 Most code should use utf8_to_uvchr() rather than call this directly.
338 Perl_utf8n_to_uvuni(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
343 bool dowarn = ckWARN_d(WARN_UTF8);
345 STRLEN expectlen = 0;
348 /* This list is a superset of the UTF8_ALLOW_XXX. */
350 #define UTF8_WARN_EMPTY 1
351 #define UTF8_WARN_CONTINUATION 2
352 #define UTF8_WARN_NON_CONTINUATION 3
353 #define UTF8_WARN_FE_FF 4
354 #define UTF8_WARN_SHORT 5
355 #define UTF8_WARN_OVERFLOW 6
356 #define UTF8_WARN_SURROGATE 7
357 #define UTF8_WARN_LONG 8
358 #define UTF8_WARN_FFFF 9 /* Also FFFE. */
361 !(flags & UTF8_ALLOW_EMPTY)) {
362 warning = UTF8_WARN_EMPTY;
366 if (UTF8_IS_INVARIANT(uv)) {
369 return (UV) (NATIVE_TO_UTF(*s));
372 if (UTF8_IS_CONTINUATION(uv) &&
373 !(flags & UTF8_ALLOW_CONTINUATION)) {
374 warning = UTF8_WARN_CONTINUATION;
378 if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) &&
379 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
380 warning = UTF8_WARN_NON_CONTINUATION;
385 uv = NATIVE_TO_UTF(uv);
387 if ((uv == 0xfe || uv == 0xff) &&
388 !(flags & UTF8_ALLOW_FE_FF)) {
389 warning = UTF8_WARN_FE_FF;
394 if (!(uv & 0x20)) { len = 2; uv &= 0x1f; }
395 else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; }
396 else if (!(uv & 0x08)) { len = 4; uv &= 0x07; }
397 else if (!(uv & 0x04)) { len = 5; uv &= 0x03; }
399 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
400 else { len = 7; uv &= 0x01; }
402 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
403 else if (!(uv & 0x01)) { len = 7; uv = 0; }
404 else { len = 13; uv = 0; } /* whoa! */
412 if ((curlen < expectlen) &&
413 !(flags & UTF8_ALLOW_SHORT)) {
414 warning = UTF8_WARN_SHORT;
423 if (!UTF8_IS_CONTINUATION(*s) &&
424 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
426 warning = UTF8_WARN_NON_CONTINUATION;
430 uv = UTF8_ACCUMULATE(uv, *s);
432 /* These cannot be allowed. */
434 if (!(flags & UTF8_ALLOW_LONG)) {
435 warning = UTF8_WARN_LONG;
439 else { /* uv < ouv */
440 /* This cannot be allowed. */
441 warning = UTF8_WARN_OVERFLOW;
449 if (UNICODE_IS_SURROGATE(uv) &&
450 !(flags & UTF8_ALLOW_SURROGATE)) {
451 warning = UTF8_WARN_SURROGATE;
453 } else if ((expectlen > (STRLEN)UNISKIP(uv)) &&
454 !(flags & UTF8_ALLOW_LONG)) {
455 warning = UTF8_WARN_LONG;
457 } else if (UNICODE_IS_ILLEGAL(uv) &&
458 !(flags & UTF8_ALLOW_FFFF)) {
459 warning = UTF8_WARN_FFFF;
467 if (flags & UTF8_CHECK_ONLY) {
474 SV* sv = sv_2mortal(newSVpv("Malformed UTF-8 character ", 0));
477 case 0: /* Intentionally empty. */ break;
478 case UTF8_WARN_EMPTY:
479 Perl_sv_catpvf(aTHX_ sv, "(empty string)");
481 case UTF8_WARN_CONTINUATION:
482 Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv);
484 case UTF8_WARN_NON_CONTINUATION:
486 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")",
487 (UV)s[1], startbyte);
489 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)",
490 (UV)s[1], s - s0, s - s0 > 1 ? "s" : "", startbyte, expectlen);
493 case UTF8_WARN_FE_FF:
494 Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv);
496 case UTF8_WARN_SHORT:
497 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
498 curlen, curlen == 1 ? "" : "s", expectlen, startbyte);
499 expectlen = curlen; /* distance for caller to skip */
501 case UTF8_WARN_OVERFLOW:
502 Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")",
505 case UTF8_WARN_SURROGATE:
506 Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv);
509 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
510 expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte);
513 Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv);
516 Perl_sv_catpvf(aTHX_ sv, "(unknown reason)");
524 Perl_warner(aTHX_ packWARN(WARN_UTF8),
525 "%s in %s", s, OP_DESC(PL_op));
527 Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s);
532 *retlen = expectlen ? expectlen : len;
538 =for apidoc A|UV|utf8_to_uvchr|U8 *s|STRLEN *retlen
540 Returns the native character value of the first character in the string C<s>
541 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
542 length, in bytes, of that character.
544 If C<s> does not point to a well-formed UTF-8 character, zero is
545 returned and retlen is set, if possible, to -1.
551 Perl_utf8_to_uvchr(pTHX_ U8 *s, STRLEN *retlen)
553 return Perl_utf8n_to_uvchr(aTHX_ s, UTF8_MAXLEN, retlen,
554 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
558 =for apidoc A|UV|utf8_to_uvuni|U8 *s|STRLEN *retlen
560 Returns the Unicode code point of the first character in the string C<s>
561 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
562 length, in bytes, of that character.
564 This function should only be used when returned UV is considered
565 an index into the Unicode semantic tables (e.g. swashes).
567 If C<s> does not point to a well-formed UTF-8 character, zero is
568 returned and retlen is set, if possible, to -1.
574 Perl_utf8_to_uvuni(pTHX_ U8 *s, STRLEN *retlen)
576 /* Call the low level routine asking for checks */
577 return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXLEN, retlen,
578 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
582 =for apidoc A|STRLEN|utf8_length|U8 *s|U8 *e
584 Return the length of the UTF-8 char encoded string C<s> in characters.
585 Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end
586 up past C<e>, croaks.
592 Perl_utf8_length(pTHX_ U8 *s, U8 *e)
596 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
597 * the bitops (especially ~) can create illegal UTF-8.
598 * In other words: in Perl UTF-8 is not just for Unicode. */
601 if (ckWARN_d(WARN_UTF8)) {
603 Perl_warner(aTHX_ packWARN(WARN_UTF8),
604 "%s in %s", unees, OP_DESC(PL_op));
606 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
614 if (ckWARN_d(WARN_UTF8)) {
616 Perl_warner(aTHX_ packWARN(WARN_UTF8),
617 unees, OP_DESC(PL_op));
619 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
631 =for apidoc A|IV|utf8_distance|U8 *a|U8 *b
633 Returns the number of UTF-8 characters between the UTF-8 pointers C<a>
636 WARNING: use only if you *know* that the pointers point inside the
643 Perl_utf8_distance(pTHX_ U8 *a, U8 *b)
647 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
648 * the bitops (especially ~) can create illegal UTF-8.
649 * In other words: in Perl UTF-8 is not just for Unicode. */
656 if (ckWARN_d(WARN_UTF8)) {
658 Perl_warner(aTHX_ packWARN(WARN_UTF8),
659 "%s in %s", unees, OP_DESC(PL_op));
661 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
674 if (ckWARN_d(WARN_UTF8)) {
676 Perl_warner(aTHX_ packWARN(WARN_UTF8),
677 "%s in %s", unees, OP_DESC(PL_op));
679 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
692 =for apidoc A|U8 *|utf8_hop|U8 *s|I32 off
694 Return the UTF-8 pointer C<s> displaced by C<off> characters, either
697 WARNING: do not use the following unless you *know* C<off> is within
698 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
699 on the first byte of character or just after the last byte of a character.
705 Perl_utf8_hop(pTHX_ U8 *s, I32 off)
707 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
708 * the bitops (especially ~) can create illegal UTF-8.
709 * In other words: in Perl UTF-8 is not just for Unicode. */
718 while (UTF8_IS_CONTINUATION(*s))
726 =for apidoc A|U8 *|utf8_to_bytes|U8 *s|STRLEN *len
728 Converts a string C<s> of length C<len> from UTF-8 into byte encoding.
729 Unlike C<bytes_to_utf8>, this over-writes the original string, and
730 updates len to contain the new length.
731 Returns zero on failure, setting C<len> to -1.
737 Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len)
743 /* ensure valid UTF-8 and chars < 256 before updating string */
744 for (send = s + *len; s < send; ) {
747 if (!UTF8_IS_INVARIANT(c) &&
748 (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send)
749 || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) {
758 *d++ = (U8)utf8_to_uvchr(s, &ulen);
767 =for apidoc A|U8 *|bytes_from_utf8|U8 *s|STRLEN *len|bool *is_utf8
769 Converts a string C<s> of length C<len> from UTF-8 into byte encoding.
770 Unlike <utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to
771 the newly-created string, and updates C<len> to contain the new
772 length. Returns the original string if no conversion occurs, C<len>
773 is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to
774 0 if C<s> is converted or contains all 7bit characters.
780 Perl_bytes_from_utf8(pTHX_ U8 *s, STRLEN *len, bool *is_utf8)
790 /* ensure valid UTF-8 and chars < 256 before converting string */
791 for (send = s + *len; s < send;) {
793 if (!UTF8_IS_INVARIANT(c)) {
794 if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send &&
795 (c = *s++) && UTF8_IS_CONTINUATION(c))
804 Newz(801, d, (*len) - count + 1, U8);
805 s = start; start = d;
808 if (!UTF8_IS_INVARIANT(c)) {
809 /* Then it is two-byte encoded */
810 c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++);
811 c = ASCII_TO_NATIVE(c);
821 =for apidoc A|U8 *|bytes_to_utf8|U8 *s|STRLEN *len
823 Converts a string C<s> of length C<len> from ASCII into UTF-8 encoding.
824 Returns a pointer to the newly-created string, and sets C<len> to
825 reflect the new length.
827 If you want to convert to UTF-8 from other encodings than ASCII,
828 see sv_recode_to_utf8().
834 Perl_bytes_to_utf8(pTHX_ U8 *s, STRLEN *len)
841 Newz(801, d, (*len) * 2 + 1, U8);
845 UV uv = NATIVE_TO_ASCII(*s++);
846 if (UNI_IS_INVARIANT(uv))
847 *d++ = (U8)UTF_TO_NATIVE(uv);
849 *d++ = (U8)UTF8_EIGHT_BIT_HI(uv);
850 *d++ = (U8)UTF8_EIGHT_BIT_LO(uv);
859 * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8.
861 * Destination must be pre-extended to 3/2 source. Do not use in-place.
862 * We optimize for native, for obvious reasons. */
865 Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
871 Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen");
876 UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */
883 *d++ = (U8)(( uv >> 6) | 0xc0);
884 *d++ = (U8)(( uv & 0x3f) | 0x80);
887 if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */
888 UV low = (p[0] << 8) + p[1];
890 if (low < 0xdc00 || low >= 0xdfff)
891 Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
892 uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000;
895 *d++ = (U8)(( uv >> 12) | 0xe0);
896 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
897 *d++ = (U8)(( uv & 0x3f) | 0x80);
901 *d++ = (U8)(( uv >> 18) | 0xf0);
902 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
903 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
904 *d++ = (U8)(( uv & 0x3f) | 0x80);
908 *newlen = d - dstart;
912 /* Note: this one is slightly destructive of the source. */
915 Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
918 U8* send = s + bytelen;
925 return utf16_to_utf8(p, d, bytelen, newlen);
928 /* for now these are all defined (inefficiently) in terms of the utf8 versions */
931 Perl_is_uni_alnum(pTHX_ UV c)
933 U8 tmpbuf[UTF8_MAXLEN+1];
934 uvchr_to_utf8(tmpbuf, c);
935 return is_utf8_alnum(tmpbuf);
939 Perl_is_uni_alnumc(pTHX_ UV c)
941 U8 tmpbuf[UTF8_MAXLEN+1];
942 uvchr_to_utf8(tmpbuf, c);
943 return is_utf8_alnumc(tmpbuf);
947 Perl_is_uni_idfirst(pTHX_ UV c)
949 U8 tmpbuf[UTF8_MAXLEN+1];
950 uvchr_to_utf8(tmpbuf, c);
951 return is_utf8_idfirst(tmpbuf);
955 Perl_is_uni_alpha(pTHX_ UV c)
957 U8 tmpbuf[UTF8_MAXLEN+1];
958 uvchr_to_utf8(tmpbuf, c);
959 return is_utf8_alpha(tmpbuf);
963 Perl_is_uni_ascii(pTHX_ UV c)
965 U8 tmpbuf[UTF8_MAXLEN+1];
966 uvchr_to_utf8(tmpbuf, c);
967 return is_utf8_ascii(tmpbuf);
971 Perl_is_uni_space(pTHX_ UV c)
973 U8 tmpbuf[UTF8_MAXLEN+1];
974 uvchr_to_utf8(tmpbuf, c);
975 return is_utf8_space(tmpbuf);
979 Perl_is_uni_digit(pTHX_ UV c)
981 U8 tmpbuf[UTF8_MAXLEN+1];
982 uvchr_to_utf8(tmpbuf, c);
983 return is_utf8_digit(tmpbuf);
987 Perl_is_uni_upper(pTHX_ UV c)
989 U8 tmpbuf[UTF8_MAXLEN+1];
990 uvchr_to_utf8(tmpbuf, c);
991 return is_utf8_upper(tmpbuf);
995 Perl_is_uni_lower(pTHX_ UV c)
997 U8 tmpbuf[UTF8_MAXLEN+1];
998 uvchr_to_utf8(tmpbuf, c);
999 return is_utf8_lower(tmpbuf);
1003 Perl_is_uni_cntrl(pTHX_ UV c)
1005 U8 tmpbuf[UTF8_MAXLEN+1];
1006 uvchr_to_utf8(tmpbuf, c);
1007 return is_utf8_cntrl(tmpbuf);
1011 Perl_is_uni_graph(pTHX_ UV c)
1013 U8 tmpbuf[UTF8_MAXLEN+1];
1014 uvchr_to_utf8(tmpbuf, c);
1015 return is_utf8_graph(tmpbuf);
1019 Perl_is_uni_print(pTHX_ UV c)
1021 U8 tmpbuf[UTF8_MAXLEN+1];
1022 uvchr_to_utf8(tmpbuf, c);
1023 return is_utf8_print(tmpbuf);
1027 Perl_is_uni_punct(pTHX_ UV c)
1029 U8 tmpbuf[UTF8_MAXLEN+1];
1030 uvchr_to_utf8(tmpbuf, c);
1031 return is_utf8_punct(tmpbuf);
1035 Perl_is_uni_xdigit(pTHX_ UV c)
1037 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1038 uvchr_to_utf8(tmpbuf, c);
1039 return is_utf8_xdigit(tmpbuf);
1043 Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp)
1045 uvchr_to_utf8(p, c);
1046 return to_utf8_upper(p, p, lenp);
1050 Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp)
1052 uvchr_to_utf8(p, c);
1053 return to_utf8_title(p, p, lenp);
1057 Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp)
1059 uvchr_to_utf8(p, c);
1060 return to_utf8_lower(p, p, lenp);
1064 Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp)
1066 uvchr_to_utf8(p, c);
1067 return to_utf8_fold(p, p, lenp);
1070 /* for now these all assume no locale info available for Unicode > 255 */
1073 Perl_is_uni_alnum_lc(pTHX_ UV c)
1075 return is_uni_alnum(c); /* XXX no locale support yet */
1079 Perl_is_uni_alnumc_lc(pTHX_ UV c)
1081 return is_uni_alnumc(c); /* XXX no locale support yet */
1085 Perl_is_uni_idfirst_lc(pTHX_ UV c)
1087 return is_uni_idfirst(c); /* XXX no locale support yet */
1091 Perl_is_uni_alpha_lc(pTHX_ UV c)
1093 return is_uni_alpha(c); /* XXX no locale support yet */
1097 Perl_is_uni_ascii_lc(pTHX_ UV c)
1099 return is_uni_ascii(c); /* XXX no locale support yet */
1103 Perl_is_uni_space_lc(pTHX_ UV c)
1105 return is_uni_space(c); /* XXX no locale support yet */
1109 Perl_is_uni_digit_lc(pTHX_ UV c)
1111 return is_uni_digit(c); /* XXX no locale support yet */
1115 Perl_is_uni_upper_lc(pTHX_ UV c)
1117 return is_uni_upper(c); /* XXX no locale support yet */
1121 Perl_is_uni_lower_lc(pTHX_ UV c)
1123 return is_uni_lower(c); /* XXX no locale support yet */
1127 Perl_is_uni_cntrl_lc(pTHX_ UV c)
1129 return is_uni_cntrl(c); /* XXX no locale support yet */
1133 Perl_is_uni_graph_lc(pTHX_ UV c)
1135 return is_uni_graph(c); /* XXX no locale support yet */
1139 Perl_is_uni_print_lc(pTHX_ UV c)
1141 return is_uni_print(c); /* XXX no locale support yet */
1145 Perl_is_uni_punct_lc(pTHX_ UV c)
1147 return is_uni_punct(c); /* XXX no locale support yet */
1151 Perl_is_uni_xdigit_lc(pTHX_ UV c)
1153 return is_uni_xdigit(c); /* XXX no locale support yet */
1157 Perl_to_uni_upper_lc(pTHX_ U32 c)
1159 /* XXX returns only the first character -- do not use XXX */
1160 /* XXX no locale support yet */
1162 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1163 return (U32)to_uni_upper(c, tmpbuf, &len);
1167 Perl_to_uni_title_lc(pTHX_ U32 c)
1169 /* XXX returns only the first character XXX -- do not use XXX */
1170 /* XXX no locale support yet */
1172 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1173 return (U32)to_uni_title(c, tmpbuf, &len);
1177 Perl_to_uni_lower_lc(pTHX_ U32 c)
1179 /* XXX returns only the first character -- do not use XXX */
1180 /* XXX no locale support yet */
1182 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1183 return (U32)to_uni_lower(c, tmpbuf, &len);
1187 Perl_is_utf8_alnum(pTHX_ U8 *p)
1189 if (!is_utf8_char(p))
1192 /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true
1193 * descendant of isalnum(3), in other words, it doesn't
1194 * contain the '_'. --jhi */
1195 PL_utf8_alnum = swash_init("utf8", "IsWord", &PL_sv_undef, 0, 0);
1196 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1197 /* return *p == '_' || is_utf8_alpha(p) || is_utf8_digit(p); */
1198 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1200 PL_utf8_alnum = swash_init("utf8", "",
1201 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1202 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1207 Perl_is_utf8_alnumc(pTHX_ U8 *p)
1209 if (!is_utf8_char(p))
1212 PL_utf8_alnum = swash_init("utf8", "IsAlnumC", &PL_sv_undef, 0, 0);
1213 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1214 /* return is_utf8_alpha(p) || is_utf8_digit(p); */
1215 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1217 PL_utf8_alnum = swash_init("utf8", "",
1218 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1219 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1224 Perl_is_utf8_idfirst(pTHX_ U8 *p) /* The naming is historical. */
1228 if (!is_utf8_char(p))
1230 if (!PL_utf8_idstart) /* is_utf8_idstart would be more logical. */
1231 PL_utf8_idstart = swash_init("utf8", "IdStart", &PL_sv_undef, 0, 0);
1232 return swash_fetch(PL_utf8_idstart, p, TRUE) != 0;
1236 Perl_is_utf8_idcont(pTHX_ U8 *p)
1240 if (!is_utf8_char(p))
1242 if (!PL_utf8_idcont)
1243 PL_utf8_idcont = swash_init("utf8", "IdContinue", &PL_sv_undef, 0, 0);
1244 return swash_fetch(PL_utf8_idcont, p, TRUE) != 0;
1248 Perl_is_utf8_alpha(pTHX_ U8 *p)
1250 if (!is_utf8_char(p))
1253 PL_utf8_alpha = swash_init("utf8", "IsAlpha", &PL_sv_undef, 0, 0);
1254 return swash_fetch(PL_utf8_alpha, p, TRUE) != 0;
1258 Perl_is_utf8_ascii(pTHX_ U8 *p)
1260 if (!is_utf8_char(p))
1263 PL_utf8_ascii = swash_init("utf8", "IsAscii", &PL_sv_undef, 0, 0);
1264 return swash_fetch(PL_utf8_ascii, p, TRUE) != 0;
1268 Perl_is_utf8_space(pTHX_ U8 *p)
1270 if (!is_utf8_char(p))
1273 PL_utf8_space = swash_init("utf8", "IsSpacePerl", &PL_sv_undef, 0, 0);
1274 return swash_fetch(PL_utf8_space, p, TRUE) != 0;
1278 Perl_is_utf8_digit(pTHX_ U8 *p)
1280 if (!is_utf8_char(p))
1283 PL_utf8_digit = swash_init("utf8", "IsDigit", &PL_sv_undef, 0, 0);
1284 return swash_fetch(PL_utf8_digit, p, TRUE) != 0;
1288 Perl_is_utf8_upper(pTHX_ U8 *p)
1290 if (!is_utf8_char(p))
1293 PL_utf8_upper = swash_init("utf8", "IsUppercase", &PL_sv_undef, 0, 0);
1294 return swash_fetch(PL_utf8_upper, p, TRUE) != 0;
1298 Perl_is_utf8_lower(pTHX_ U8 *p)
1300 if (!is_utf8_char(p))
1303 PL_utf8_lower = swash_init("utf8", "IsLowercase", &PL_sv_undef, 0, 0);
1304 return swash_fetch(PL_utf8_lower, p, TRUE) != 0;
1308 Perl_is_utf8_cntrl(pTHX_ U8 *p)
1310 if (!is_utf8_char(p))
1313 PL_utf8_cntrl = swash_init("utf8", "IsCntrl", &PL_sv_undef, 0, 0);
1314 return swash_fetch(PL_utf8_cntrl, p, TRUE) != 0;
1318 Perl_is_utf8_graph(pTHX_ U8 *p)
1320 if (!is_utf8_char(p))
1323 PL_utf8_graph = swash_init("utf8", "IsGraph", &PL_sv_undef, 0, 0);
1324 return swash_fetch(PL_utf8_graph, p, TRUE) != 0;
1328 Perl_is_utf8_print(pTHX_ U8 *p)
1330 if (!is_utf8_char(p))
1333 PL_utf8_print = swash_init("utf8", "IsPrint", &PL_sv_undef, 0, 0);
1334 return swash_fetch(PL_utf8_print, p, TRUE) != 0;
1338 Perl_is_utf8_punct(pTHX_ U8 *p)
1340 if (!is_utf8_char(p))
1343 PL_utf8_punct = swash_init("utf8", "IsPunct", &PL_sv_undef, 0, 0);
1344 return swash_fetch(PL_utf8_punct, p, TRUE) != 0;
1348 Perl_is_utf8_xdigit(pTHX_ U8 *p)
1350 if (!is_utf8_char(p))
1352 if (!PL_utf8_xdigit)
1353 PL_utf8_xdigit = swash_init("utf8", "IsXDigit", &PL_sv_undef, 0, 0);
1354 return swash_fetch(PL_utf8_xdigit, p, TRUE) != 0;
1358 Perl_is_utf8_mark(pTHX_ U8 *p)
1360 if (!is_utf8_char(p))
1363 PL_utf8_mark = swash_init("utf8", "IsM", &PL_sv_undef, 0, 0);
1364 return swash_fetch(PL_utf8_mark, p, TRUE) != 0;
1368 =for apidoc A|UV|to_utf8_case|U8 *p|U8* ustrp|STRLEN *lenp|SV **swash|char *normal|char *special
1370 The "p" contains the pointer to the UTF-8 string encoding
1371 the character that is being converted.
1373 The "ustrp" is a pointer to the character buffer to put the
1374 conversion result to. The "lenp" is a pointer to the length
1377 The "swashp" is a pointer to the swash to use.
1379 Both the special and normal mappings are stored lib/unicore/To/Foo.pl,
1380 and loaded by SWASHGET, using lib/utf8_heavy.pl. The special (usually,
1381 but not always, a multicharacter mapping), is tried first.
1383 The "special" is a string like "utf8::ToSpecLower", which means the
1384 hash %utf8::ToSpecLower. The access to the hash is through
1385 Perl_to_utf8_case().
1387 The "normal" is a string like "ToLower" which means the swash
1393 Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp, char *normal, char *special)
1396 U8 tmpbuf[UTF8_MAXLEN_FOLD+1];
1399 uv0 = utf8_to_uvchr(p, 0);
1400 /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings
1401 * are necessary in EBCDIC, they are redundant no-ops
1402 * in ASCII-ish platforms, and hopefully optimized away. */
1403 uv1 = NATIVE_TO_UNI(uv0);
1404 uvuni_to_utf8(tmpbuf, uv1);
1406 if (!*swashp) /* load on-demand */
1407 *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0);
1409 /* The 0xDF is the only special casing Unicode code point below 0x100. */
1410 if (special && (uv1 == 0xDF || uv1 > 0xFF)) {
1411 /* It might be "special" (sometimes, but not always,
1412 * a multicharacter mapping) */
1416 if ((hv = get_hv(special, FALSE)) &&
1417 (svp = hv_fetch(hv, (const char*)tmpbuf, UNISKIP(uv1), FALSE)) &&
1421 s = SvPV(*svp, len);
1423 len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp;
1426 /* If we have EBCDIC we need to remap the characters
1427 * since any characters in the low 256 are Unicode
1428 * code points, not EBCDIC. */
1429 U8 *t = (U8*)s, *tend = t + len, *d;
1436 UV c = utf8_to_uvchr(t, &tlen);
1438 d = uvchr_to_utf8(d, UNI_TO_NATIVE(c));
1447 d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t));
1452 Copy(tmpbuf, ustrp, len, U8);
1454 Copy(s, ustrp, len, U8);
1460 if (!len && *swashp) {
1461 UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE);
1464 /* It was "normal" (a single character mapping). */
1465 UV uv3 = UNI_TO_NATIVE(uv2);
1467 len = uvchr_to_utf8(ustrp, uv3) - ustrp;
1471 if (!len) /* Neither: just copy. */
1472 len = uvchr_to_utf8(ustrp, uv0) - ustrp;
1477 return len ? utf8_to_uvchr(ustrp, 0) : 0;
1481 =for apidoc A|UV|to_utf8_upper|U8 *p|U8 *ustrp|STRLEN *lenp
1483 Convert the UTF-8 encoded character at p to its uppercase version and
1484 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1485 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1486 uppercase version may be longer than the original character (up to two
1489 The first character of the uppercased version is returned
1490 (but note, as explained above, that there may be more.)
1495 Perl_to_utf8_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1497 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1498 &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper");
1502 =for apidoc A|UV|to_utf8_title|U8 *p|U8 *ustrp|STRLEN *lenp
1504 Convert the UTF-8 encoded character at p to its titlecase version and
1505 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1506 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1507 titlecase version may be longer than the original character (up to two
1510 The first character of the titlecased version is returned
1511 (but note, as explained above, that there may be more.)
1516 Perl_to_utf8_title(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1518 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1519 &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle");
1523 =for apidoc A|UV|to_utf8_lower|U8 *p|U8 *ustrp|STRLEN *lenp
1525 Convert the UTF-8 encoded character at p to its lowercase version and
1526 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1527 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1528 lowercase version may be longer than the original character (up to two
1531 The first character of the lowercased version is returned
1532 (but note, as explained above, that there may be more.)
1537 Perl_to_utf8_lower(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1539 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1540 &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower");
1544 =for apidoc A|UV|to_utf8_fold|U8 *p|U8 *ustrp|STRLEN *lenp
1546 Convert the UTF-8 encoded character at p to its foldcase version and
1547 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1548 that the ustrp needs to be at least UTF8_MAXLEN_FOLD+1 bytes since the
1549 foldcase version may be longer than the original character (up to
1552 The first character of the foldcased version is returned
1553 (but note, as explained above, that there may be more.)
1558 Perl_to_utf8_fold(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1560 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1561 &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold");
1564 /* a "swash" is a swatch hash */
1567 Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none)
1570 SV* tokenbufsv = sv_2mortal(NEWSV(0,0));
1572 size_t pkg_len = strlen(pkg);
1573 size_t name_len = strlen(name);
1574 HV *stash = gv_stashpvn(pkg, pkg_len, FALSE);
1577 if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */
1579 errsv_save = newSVsv(ERRSV);
1580 Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpvn(pkg,pkg_len),
1583 sv_setsv(ERRSV, errsv_save);
1584 SvREFCNT_dec(errsv_save);
1588 PUSHSTACKi(PERLSI_MAGIC);
1591 PUSHs(sv_2mortal(newSVpvn(pkg, pkg_len)));
1592 PUSHs(sv_2mortal(newSVpvn(name, name_len)));
1594 PUSHs(sv_2mortal(newSViv(minbits)));
1595 PUSHs(sv_2mortal(newSViv(none)));
1601 if (IN_PERL_COMPILETIME) {
1602 /* XXX ought to be handled by lex_start */
1605 sv_setpv(tokenbufsv, PL_tokenbuf);
1607 errsv_save = newSVsv(ERRSV);
1608 if (call_method("SWASHNEW", G_SCALAR))
1609 retval = newSVsv(*PL_stack_sp--);
1611 retval = &PL_sv_undef;
1613 sv_setsv(ERRSV, errsv_save);
1614 SvREFCNT_dec(errsv_save);
1617 if (IN_PERL_COMPILETIME) {
1619 char* pv = SvPV(tokenbufsv, len);
1621 Copy(pv, PL_tokenbuf, len+1, char);
1622 PL_curcop->op_private = (U8)(PL_hints & HINT_PRIVATE_MASK);
1624 if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) {
1626 Perl_croak(aTHX_ "Can't find Unicode property definition \"%"SVf"\"",
1628 Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref");
1634 /* This API is wrong for special case conversions since we may need to
1635 * return several Unicode characters for a single Unicode character
1636 * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is
1637 * the lower-level routine, and it is similarly broken for returning
1638 * multiple values. --jhi */
1640 Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr, bool do_utf8)
1642 HV* hv = (HV*)SvRV(sv);
1651 UV c = NATIVE_TO_ASCII(*ptr);
1653 if (!do_utf8 && !UNI_IS_INVARIANT(c)) {
1654 tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c);
1655 tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c);
1658 /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ
1659 * then the "swatch" is a vec() for al the chars which start
1661 * So the key in the hash (klen) is length of encoded char -1
1663 klen = UTF8SKIP(ptr) - 1;
1668 /* If char in invariant then swatch is for all the invariant chars
1669 * In both UTF-8 and UTF-8-MOD that happens to be UTF_CONTINUATION_MARK
1671 needents = UTF_CONTINUATION_MARK;
1672 off = NATIVE_TO_UTF(ptr[klen]);
1676 /* If char is encoded then swatch is for the prefix */
1677 needents = (1 << UTF_ACCUMULATION_SHIFT);
1678 off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK;
1682 * This single-entry cache saves about 1/3 of the utf8 overhead in test
1683 * suite. (That is, only 7-8% overall over just a hash cache. Still,
1684 * it's nothing to sniff at.) Pity we usually come through at least
1685 * two function calls to get here...
1687 * NB: this code assumes that swatches are never modified, once generated!
1690 if (hv == PL_last_swash_hv &&
1691 klen == PL_last_swash_klen &&
1692 (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) )
1694 tmps = PL_last_swash_tmps;
1695 slen = PL_last_swash_slen;
1698 /* Try our second-level swatch cache, kept in a hash. */
1699 SV** svp = hv_fetch(hv, (char*)ptr, klen, FALSE);
1701 /* If not cached, generate it via utf8::SWASHGET */
1702 if (!svp || !SvPOK(*svp) || !(tmps = (U8*)SvPV(*svp, slen))) {
1704 /* We use utf8n_to_uvuni() as we want an index into
1705 Unicode tables, not a native character number.
1707 UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXLEN, 0,
1709 0 : UTF8_ALLOW_ANY);
1714 PUSHSTACKi(PERLSI_MAGIC);
1718 /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */
1719 PUSHs(sv_2mortal(newSViv((klen) ?
1720 (code_point & ~(needents - 1)) : 0)));
1721 PUSHs(sv_2mortal(newSViv(needents)));
1723 errsv_save = newSVsv(ERRSV);
1724 if (call_method("SWASHGET", G_SCALAR))
1725 retval = newSVsv(*PL_stack_sp--);
1727 retval = &PL_sv_undef;
1729 sv_setsv(ERRSV, errsv_save);
1730 SvREFCNT_dec(errsv_save);
1734 if (IN_PERL_COMPILETIME)
1735 PL_curcop->op_private = (U8)(PL_hints & HINT_PRIVATE_MASK);
1737 svp = hv_store(hv, (char*)ptr, klen, retval, 0);
1739 if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || (slen << 3) < needents)
1740 Perl_croak(aTHX_ "SWASHGET didn't return result of proper length");
1743 PL_last_swash_hv = hv;
1744 PL_last_swash_klen = klen;
1745 PL_last_swash_tmps = tmps;
1746 PL_last_swash_slen = slen;
1748 Copy(ptr, PL_last_swash_key, klen, U8);
1751 switch ((int)((slen << 3) / needents)) {
1753 bit = 1 << (off & 7);
1755 return (tmps[off] & bit) != 0;
1760 return (tmps[off] << 8) + tmps[off + 1] ;
1763 return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ;
1765 Perl_croak(aTHX_ "panic: swash_fetch");
1771 =for apidoc A|U8 *|uvchr_to_utf8|U8 *d|UV uv
1773 Adds the UTF-8 representation of the Native codepoint C<uv> to the end
1774 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
1775 bytes available. The return value is the pointer to the byte after the
1776 end of the new character. In other words,
1778 d = uvchr_to_utf8(d, uv);
1780 is the recommended wide native character-aware way of saying
1787 /* On ASCII machines this is normally a macro but we want a
1788 real function in case XS code wants it
1790 #undef Perl_uvchr_to_utf8
1792 Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv)
1794 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0);
1798 Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
1800 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags);
1804 =for apidoc A|UV|utf8n_to_uvchr|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
1806 Returns the native character value of the first character in the string C<s>
1807 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
1808 length, in bytes, of that character.
1810 Allows length and flags to be passed to low level routine.
1814 /* On ASCII machines this is normally a macro but we want
1815 a real function in case XS code wants it
1817 #undef Perl_utf8n_to_uvchr
1819 Perl_utf8n_to_uvchr(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
1821 UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags);
1822 return UNI_TO_NATIVE(uv);
1826 =for apidoc A|char *|pv_uni_display|SV *dsv|U8 *spv|STRLEN len|STRLEN pvlim|UV flags
1828 Build to the scalar dsv a displayable version of the string spv,
1829 length len, the displayable version being at most pvlim bytes long
1830 (if longer, the rest is truncated and "..." will be appended).
1832 The flags argument can have UNI_DISPLAY_ISPRINT set to display
1833 isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH
1834 to display the \\[nrfta\\] as the backslashed versions (like '\n')
1835 (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\).
1836 UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both
1837 UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on.
1839 The pointer to the PV of the dsv is returned.
1843 Perl_pv_uni_display(pTHX_ SV *dsv, U8 *spv, STRLEN len, STRLEN pvlim, UV flags)
1848 sv_setpvn(dsv, "", 0);
1849 for (s = (char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) {
1853 if (pvlim && SvCUR(dsv) >= pvlim) {
1857 u = utf8_to_uvchr((U8*)s, 0);
1859 if (!ok && (flags & UNI_DISPLAY_BACKSLASH)) {
1862 Perl_sv_catpvf(aTHX_ dsv, "\\n"); ok = TRUE; break;
1864 Perl_sv_catpvf(aTHX_ dsv, "\\r"); ok = TRUE; break;
1866 Perl_sv_catpvf(aTHX_ dsv, "\\t"); ok = TRUE; break;
1868 Perl_sv_catpvf(aTHX_ dsv, "\\f"); ok = TRUE; break;
1870 Perl_sv_catpvf(aTHX_ dsv, "\\a"); ok = TRUE; break;
1872 Perl_sv_catpvf(aTHX_ dsv, "\\\\" ); ok = TRUE; break;
1876 /* isPRINT() is the locale-blind version. */
1877 if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(u & 0xFF)) {
1878 Perl_sv_catpvf(aTHX_ dsv, "%c", (char)(u & 0xFF));
1883 Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u);
1886 sv_catpvn(dsv, "...", 3);
1892 =for apidoc A|char *|sv_uni_display|SV *dsv|SV *ssv|STRLEN pvlim|UV flags
1894 Build to the scalar dsv a displayable version of the scalar sv,
1895 the displayable version being at most pvlim bytes long
1896 (if longer, the rest is truncated and "..." will be appended).
1898 The flags argument is as in pv_uni_display().
1900 The pointer to the PV of the dsv is returned.
1904 Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags)
1906 return Perl_pv_uni_display(aTHX_ dsv, (U8*)SvPVX(ssv), SvCUR(ssv),
1911 =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
1913 Return true if the strings s1 and s2 differ case-insensitively, false
1914 if not (if they are equal case-insensitively). If u1 is true, the
1915 string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true,
1916 the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2
1917 are false, the respective string is assumed to be in native 8-bit
1920 If the pe1 and pe2 are non-NULL, the scanning pointers will be copied
1921 in there (they will point at the beginning of the I<next> character).
1922 If the pointers behind pe1 or pe2 are non-NULL, they are the end
1923 pointers beyond which scanning will not continue under any
1924 circustances. If the byte lengths l1 and l2 are non-zero, s1+l1 and
1925 s2+l2 will be used as goal end pointers that will also stop the scan,
1926 and which qualify towards defining a successful match: all the scans
1927 that define an explicit length must reach their goal pointers for
1928 a match to succeed).
1930 For case-insensitiveness, the "casefolding" of Unicode is used
1931 instead of upper/lowercasing both the characters, see
1932 http://www.unicode.org/unicode/reports/tr21/ (Case Mappings).
1936 Perl_ibcmp_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2)
1938 register U8 *p1 = (U8*)s1;
1939 register U8 *p2 = (U8*)s2;
1940 register U8 *e1 = 0, *f1 = 0, *q1 = 0;
1941 register U8 *e2 = 0, *f2 = 0, *q2 = 0;
1942 STRLEN n1 = 0, n2 = 0;
1943 U8 foldbuf1[UTF8_MAXLEN_FOLD+1];
1944 U8 foldbuf2[UTF8_MAXLEN_FOLD+1];
1946 STRLEN foldlen1, foldlen2;
1951 if (e1 == 0 || (l1 && l1 < (UV)(e1 - (U8*)s1)))
1955 if (e2 == 0 || (l2 && l2 < (UV)(e2 - (U8*)s2)))
1958 if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0))
1959 return 1; /* mismatch; possible infinite loop or false positive */
1962 natbuf[1] = 0; /* Need to terminate the buffer. */
1964 while ((e1 == 0 || p1 < e1) &&
1965 (f1 == 0 || p1 < f1) &&
1966 (e2 == 0 || p2 < e2) &&
1967 (f2 == 0 || p2 < f2)) {
1970 to_utf8_fold(p1, foldbuf1, &foldlen1);
1973 to_utf8_fold(natbuf, foldbuf1, &foldlen1);
1980 to_utf8_fold(p2, foldbuf2, &foldlen2);
1983 to_utf8_fold(natbuf, foldbuf2, &foldlen2);
1989 if ( UTF8SKIP(q1) != UTF8SKIP(q2) ||
1990 (UTF8SKIP(q1) == 1 && *q1 != *q2) ||
1991 memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) )
1992 return 1; /* mismatch */
1999 p1 += u1 ? UTF8SKIP(p1) : 1;
2001 p2 += u2 ? UTF8SKIP(p2) : 1;
2005 /* A match is defined by all the scans that specified
2006 * an explicit length reaching their final goals. */
2007 match = (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2);
2016 return match ? 0 : 1; /* 0 match, 1 mismatch */