3 * Copyright (C) 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
4 * by Larry Wall and others
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
12 * 'What a fix!' said Sam. 'That's the one place in all the lands we've ever
13 * heard of that we don't want to see any closer; and that's the one place
14 * we're trying to get to! And that's just where we can't get, nohow.'
16 * [p.603 of _The Lord of the Rings_, IV/I: "The Taming of Sméagol"]
18 * 'Well do I understand your speech,' he answered in the same language;
19 * 'yet few strangers do so. Why then do you not speak in the Common Tongue,
20 * as is the custom in the West, if you wish to be answered?'
21 * --Gandalf, addressing Théoden's door wardens
23 * [p.508 of _The Lord of the Rings_, III/vi: "The King of the Golden Hall"]
25 * ...the travellers perceived that the floor was paved with stones of many
26 * hues; branching runes and strange devices intertwined beneath their feet.
28 * [p.512 of _The Lord of the Rings_, III/vi: "The King of the Golden Hall"]
32 #define PERL_IN_UTF8_C
36 /* Separate prototypes needed because in ASCII systems these
37 * usually macros but they still are compiled as code, too. */
38 PERL_CALLCONV UV Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags);
39 PERL_CALLCONV U8* Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv);
42 static const char unees[] =
43 "Malformed UTF-8 character (unexpected end of string)";
46 =head1 Unicode Support
48 This file contains various utility functions for manipulating UTF8-encoded
49 strings. For the uninitiated, this is a method of representing arbitrary
50 Unicode characters as a variable number of bytes, in such a way that
51 characters in the ASCII range are unmodified, and a zero byte never appears
52 within non-zero characters.
58 =for apidoc is_ascii_string
60 Returns true if first C<len> bytes of the given string are ASCII (i.e. none
61 of them even raise the question of UTF-8-ness).
63 See also is_utf8_string(), is_utf8_string_loclen(), and is_utf8_string_loc().
69 Perl_is_ascii_string(const U8 *s, STRLEN len)
71 const U8* const send = s + (len ? len : strlen((const char *)s));
74 PERL_ARGS_ASSERT_IS_ASCII_STRING;
76 for (; x < send; ++x) {
77 if (!UTF8_IS_INVARIANT(*x))
85 =for apidoc uvuni_to_utf8_flags
87 Adds the UTF-8 representation of the Unicode codepoint C<uv> to the end
88 of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free
89 bytes available. The return value is the pointer to the byte after the
90 end of the new character. In other words,
92 d = uvuni_to_utf8_flags(d, uv, flags);
96 d = uvuni_to_utf8(d, uv);
98 (which is equivalent to)
100 d = uvuni_to_utf8_flags(d, uv, 0);
102 is the recommended Unicode-aware way of saying
110 Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
112 PERL_ARGS_ASSERT_UVUNI_TO_UTF8_FLAGS;
114 if (ckWARN(WARN_UTF8)) {
115 if (UNICODE_IS_SURROGATE(uv) &&
116 !(flags & UNICODE_ALLOW_SURROGATE))
117 Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv);
119 ((uv >= 0xFDD0 && uv <= 0xFDEF &&
120 !(flags & UNICODE_ALLOW_FDD0))
122 ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */
123 !(flags & UNICODE_ALLOW_FFFF))) &&
124 /* UNICODE_ALLOW_SUPER includes
125 * FFFEs and FFFFs beyond 0x10FFFF. */
126 ((uv <= PERL_UNICODE_MAX) ||
127 !(flags & UNICODE_ALLOW_SUPER))
129 Perl_warner(aTHX_ packWARN(WARN_UTF8),
130 "Unicode non-character 0x%04"UVxf" is illegal for interchange", uv);
132 if (UNI_IS_INVARIANT(uv)) {
133 *d++ = (U8)UTF_TO_NATIVE(uv);
138 STRLEN len = UNISKIP(uv);
141 *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK);
142 uv >>= UTF_ACCUMULATION_SHIFT;
144 *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len));
147 #else /* Non loop style */
149 *d++ = (U8)(( uv >> 6) | 0xc0);
150 *d++ = (U8)(( uv & 0x3f) | 0x80);
154 *d++ = (U8)(( uv >> 12) | 0xe0);
155 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
156 *d++ = (U8)(( uv & 0x3f) | 0x80);
160 *d++ = (U8)(( uv >> 18) | 0xf0);
161 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
162 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
163 *d++ = (U8)(( uv & 0x3f) | 0x80);
166 if (uv < 0x4000000) {
167 *d++ = (U8)(( uv >> 24) | 0xf8);
168 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
169 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
170 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
171 *d++ = (U8)(( uv & 0x3f) | 0x80);
174 if (uv < 0x80000000) {
175 *d++ = (U8)(( uv >> 30) | 0xfc);
176 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
177 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
178 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
179 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
180 *d++ = (U8)(( uv & 0x3f) | 0x80);
184 if (uv < UTF8_QUAD_MAX)
187 *d++ = 0xfe; /* Can't match U+FEFF! */
188 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
189 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
190 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
191 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
192 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
193 *d++ = (U8)(( uv & 0x3f) | 0x80);
198 *d++ = 0xff; /* Can't match U+FFFE! */
199 *d++ = 0x80; /* 6 Reserved bits */
200 *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */
201 *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80);
202 *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80);
203 *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80);
204 *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80);
205 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
206 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
207 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
208 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
209 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
210 *d++ = (U8)(( uv & 0x3f) | 0x80);
214 #endif /* Loop style */
219 Tests if some arbitrary number of bytes begins in a valid UTF-8
220 character. Note that an INVARIANT (i.e. ASCII) character is a valid
221 UTF-8 character. The actual number of bytes in the UTF-8 character
222 will be returned if it is valid, otherwise 0.
224 This is the "slow" version as opposed to the "fast" version which is
225 the "unrolled" IS_UTF8_CHAR(). E.g. for t/uni/class.t the speed
226 difference is a factor of 2 to 3. For lengths (UTF8SKIP(s)) of four
227 or less you should use the IS_UTF8_CHAR(), for lengths of five or more
228 you should use the _slow(). In practice this means that the _slow()
229 will be used very rarely, since the maximum Unicode code point (as of
230 Unicode 4.1) is U+10FFFF, which encodes in UTF-8 to four bytes. Only
231 the "Perl extended UTF-8" (the infamous 'v-strings') will encode into
236 S_is_utf8_char_slow(const U8 *s, const STRLEN len)
242 PERL_ARGS_ASSERT_IS_UTF8_CHAR_SLOW;
244 if (UTF8_IS_INVARIANT(u))
247 if (!UTF8_IS_START(u))
250 if (len < 2 || !UTF8_IS_CONTINUATION(s[1]))
256 u = NATIVE_TO_UTF(u);
258 u &= UTF_START_MASK(len);
262 if (!UTF8_IS_CONTINUATION(*s))
264 uv = UTF8_ACCUMULATE(uv, *s);
271 if ((STRLEN)UNISKIP(uv) < len)
278 =for apidoc is_utf8_char
280 Tests if some arbitrary number of bytes begins in a valid UTF-8
281 character. Note that an INVARIANT (i.e. ASCII on non-EBCDIC machines)
282 character is a valid UTF-8 character. The actual number of bytes in the UTF-8
283 character will be returned if it is valid, otherwise 0.
287 Perl_is_utf8_char(const U8 *s)
289 const STRLEN len = UTF8SKIP(s);
291 PERL_ARGS_ASSERT_IS_UTF8_CHAR;
293 if (IS_UTF8_CHAR_FAST(len))
294 return IS_UTF8_CHAR(s, len) ? len : 0;
295 #endif /* #ifdef IS_UTF8_CHAR */
296 return is_utf8_char_slow(s, len);
301 =for apidoc is_utf8_string
303 Returns true if first C<len> bytes of the given string form a valid
304 UTF-8 string, false otherwise. Note that 'a valid UTF-8 string' does
305 not mean 'a string that contains code points above 0x7F encoded in UTF-8'
306 because a valid ASCII string is a valid UTF-8 string.
308 See also is_ascii_string(), is_utf8_string_loclen(), and is_utf8_string_loc().
314 Perl_is_utf8_string(const U8 *s, STRLEN len)
316 const U8* const send = s + (len ? len : strlen((const char *)s));
319 PERL_ARGS_ASSERT_IS_UTF8_STRING;
323 /* Inline the easy bits of is_utf8_char() here for speed... */
324 if (UTF8_IS_INVARIANT(*x))
326 else if (!UTF8_IS_START(*x))
329 /* ... and call is_utf8_char() only if really needed. */
332 if (IS_UTF8_CHAR_FAST(c)) {
333 if (!IS_UTF8_CHAR(x, c))
337 c = is_utf8_char_slow(x, c);
340 #endif /* #ifdef IS_UTF8_CHAR */
355 Implemented as a macro in utf8.h
357 =for apidoc is_utf8_string_loc
359 Like is_utf8_string() but stores the location of the failure (in the
360 case of "utf8ness failure") or the location s+len (in the case of
361 "utf8ness success") in the C<ep>.
363 See also is_utf8_string_loclen() and is_utf8_string().
365 =for apidoc is_utf8_string_loclen
367 Like is_utf8_string() but stores the location of the failure (in the
368 case of "utf8ness failure") or the location s+len (in the case of
369 "utf8ness success") in the C<ep>, and the number of UTF-8
370 encoded characters in the C<el>.
372 See also is_utf8_string_loc() and is_utf8_string().
378 Perl_is_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el)
380 const U8* const send = s + (len ? len : strlen((const char *)s));
385 PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN;
388 /* Inline the easy bits of is_utf8_char() here for speed... */
389 if (UTF8_IS_INVARIANT(*x))
391 else if (!UTF8_IS_START(*x))
394 /* ... and call is_utf8_char() only if really needed. */
397 if (IS_UTF8_CHAR_FAST(c)) {
398 if (!IS_UTF8_CHAR(x, c))
401 c = is_utf8_char_slow(x, c);
404 #endif /* #ifdef IS_UTF8_CHAR */
423 =for apidoc utf8n_to_uvuni
425 Bottom level UTF-8 decode routine.
426 Returns the Unicode code point value of the first character in the string C<s>
427 which is assumed to be in UTF-8 encoding and no longer than C<curlen>;
428 C<retlen> will be set to the length, in bytes, of that character.
430 If C<s> does not point to a well-formed UTF-8 character, the behaviour
431 is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY,
432 it is assumed that the caller will raise a warning, and this function
433 will silently just set C<retlen> to C<-1> and return zero. If the
434 C<flags> does not contain UTF8_CHECK_ONLY, warnings about
435 malformations will be given, C<retlen> will be set to the expected
436 length of the UTF-8 character in bytes, and zero will be returned.
438 The C<flags> can also contain various flags to allow deviations from
439 the strict UTF-8 encoding (see F<utf8.h>).
441 Most code should use utf8_to_uvchr() rather than call this directly.
447 Perl_utf8n_to_uvuni(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
450 const U8 * const s0 = s;
453 const bool dowarn = ckWARN_d(WARN_UTF8);
454 const UV startbyte = *s;
455 STRLEN expectlen = 0;
459 PERL_ARGS_ASSERT_UTF8N_TO_UVUNI;
461 /* This list is a superset of the UTF8_ALLOW_XXX. BUT it isn't, eg SUPER missing XXX */
463 #define UTF8_WARN_EMPTY 1
464 #define UTF8_WARN_CONTINUATION 2
465 #define UTF8_WARN_NON_CONTINUATION 3
466 #define UTF8_WARN_FE_FF 4
467 #define UTF8_WARN_SHORT 5
468 #define UTF8_WARN_OVERFLOW 6
469 #define UTF8_WARN_SURROGATE 7
470 #define UTF8_WARN_LONG 8
471 #define UTF8_WARN_FFFF 9 /* Also FFFE. */
474 !(flags & UTF8_ALLOW_EMPTY)) {
475 warning = UTF8_WARN_EMPTY;
479 if (UTF8_IS_INVARIANT(uv)) {
482 return (UV) (NATIVE_TO_UTF(*s));
485 if (UTF8_IS_CONTINUATION(uv) &&
486 !(flags & UTF8_ALLOW_CONTINUATION)) {
487 warning = UTF8_WARN_CONTINUATION;
491 if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) &&
492 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
493 warning = UTF8_WARN_NON_CONTINUATION;
498 uv = NATIVE_TO_UTF(uv);
500 if ((uv == 0xfe || uv == 0xff) &&
501 !(flags & UTF8_ALLOW_FE_FF)) {
502 warning = UTF8_WARN_FE_FF;
507 if (!(uv & 0x20)) { len = 2; uv &= 0x1f; }
508 else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; }
509 else if (!(uv & 0x08)) { len = 4; uv &= 0x07; }
510 else if (!(uv & 0x04)) { len = 5; uv &= 0x03; }
512 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
513 else { len = 7; uv &= 0x01; }
515 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
516 else if (!(uv & 0x01)) { len = 7; uv = 0; }
517 else { len = 13; uv = 0; } /* whoa! */
525 if ((curlen < expectlen) &&
526 !(flags & UTF8_ALLOW_SHORT)) {
527 warning = UTF8_WARN_SHORT;
536 if (!UTF8_IS_CONTINUATION(*s) &&
537 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
539 warning = UTF8_WARN_NON_CONTINUATION;
543 uv = UTF8_ACCUMULATE(uv, *s);
545 /* These cannot be allowed. */
547 if (expectlen != 13 && !(flags & UTF8_ALLOW_LONG)) {
548 warning = UTF8_WARN_LONG;
552 else { /* uv < ouv */
553 /* This cannot be allowed. */
554 warning = UTF8_WARN_OVERFLOW;
562 if (UNICODE_IS_SURROGATE(uv) &&
563 !(flags & UTF8_ALLOW_SURROGATE)) {
564 warning = UTF8_WARN_SURROGATE;
566 } else if ((expectlen > (STRLEN)UNISKIP(uv)) &&
567 !(flags & UTF8_ALLOW_LONG)) {
568 warning = UTF8_WARN_LONG;
570 } else if (UNICODE_IS_ILLEGAL(uv) &&
571 !(flags & UTF8_ALLOW_FFFF)) {
572 warning = UTF8_WARN_FFFF;
580 if (flags & UTF8_CHECK_ONLY) {
582 *retlen = ((STRLEN) -1);
587 if (warning == UTF8_WARN_FFFF) {
588 sv = newSVpvs_flags("Unicode non-character ", SVs_TEMP);
589 Perl_sv_catpvf(aTHX_ sv, "0x%04"UVxf" is illegal for interchange", uv);
592 sv = newSVpvs_flags("Malformed UTF-8 character ", SVs_TEMP);
595 case 0: /* Intentionally empty. */ break;
596 case UTF8_WARN_EMPTY:
597 sv_catpvs(sv, "(empty string)");
599 case UTF8_WARN_CONTINUATION:
600 Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv);
602 case UTF8_WARN_NON_CONTINUATION:
604 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")",
605 (UV)s[1], startbyte);
607 const int len = (int)(s-s0);
608 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)",
609 (UV)s[1], len, len > 1 ? "s" : "", startbyte, (int)expectlen);
613 case UTF8_WARN_FE_FF:
614 Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv);
616 case UTF8_WARN_SHORT:
617 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
618 (int)curlen, curlen == 1 ? "" : "s", (int)expectlen, startbyte);
619 expectlen = curlen; /* distance for caller to skip */
621 case UTF8_WARN_OVERFLOW:
622 Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")",
625 case UTF8_WARN_SURROGATE:
626 Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv);
629 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
630 (int)expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte);
633 sv_catpvs(sv, "(unknown reason)");
639 const char * const s = SvPVX_const(sv);
642 Perl_warner(aTHX_ packWARN(WARN_UTF8),
643 "%s in %s", s, OP_DESC(PL_op));
645 Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s);
650 *retlen = expectlen ? expectlen : len;
656 =for apidoc utf8_to_uvchr
658 Returns the native character value of the first character in the string C<s>
659 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
660 length, in bytes, of that character.
662 If C<s> does not point to a well-formed UTF-8 character, zero is
663 returned and retlen is set, if possible, to -1.
669 Perl_utf8_to_uvchr(pTHX_ const U8 *s, STRLEN *retlen)
671 PERL_ARGS_ASSERT_UTF8_TO_UVCHR;
673 return utf8n_to_uvchr(s, UTF8_MAXBYTES, retlen,
674 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
678 =for apidoc utf8_to_uvuni
680 Returns the Unicode code point of the first character in the string C<s>
681 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
682 length, in bytes, of that character.
684 This function should only be used when the returned UV is considered
685 an index into the Unicode semantic tables (e.g. swashes).
687 If C<s> does not point to a well-formed UTF-8 character, zero is
688 returned and retlen is set, if possible, to -1.
694 Perl_utf8_to_uvuni(pTHX_ const U8 *s, STRLEN *retlen)
696 PERL_ARGS_ASSERT_UTF8_TO_UVUNI;
698 /* Call the low level routine asking for checks */
699 return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXBYTES, retlen,
700 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
704 =for apidoc utf8_length
706 Return the length of the UTF-8 char encoded string C<s> in characters.
707 Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end
708 up past C<e>, croaks.
714 Perl_utf8_length(pTHX_ const U8 *s, const U8 *e)
719 PERL_ARGS_ASSERT_UTF8_LENGTH;
721 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
722 * the bitops (especially ~) can create illegal UTF-8.
723 * In other words: in Perl UTF-8 is not just for Unicode. */
726 goto warn_and_return;
728 if (!UTF8_IS_INVARIANT(*s))
739 Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8),
740 "%s in %s", unees, OP_DESC(PL_op));
742 Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), unees);
749 =for apidoc utf8_distance
751 Returns the number of UTF-8 characters between the UTF-8 pointers C<a>
754 WARNING: use only if you *know* that the pointers point inside the
761 Perl_utf8_distance(pTHX_ const U8 *a, const U8 *b)
763 PERL_ARGS_ASSERT_UTF8_DISTANCE;
765 return (a < b) ? -1 * (IV) utf8_length(a, b) : (IV) utf8_length(b, a);
771 Return the UTF-8 pointer C<s> displaced by C<off> characters, either
774 WARNING: do not use the following unless you *know* C<off> is within
775 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
776 on the first byte of character or just after the last byte of a character.
782 Perl_utf8_hop(pTHX_ const U8 *s, I32 off)
784 PERL_ARGS_ASSERT_UTF8_HOP;
787 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
788 * the bitops (especially ~) can create illegal UTF-8.
789 * In other words: in Perl UTF-8 is not just for Unicode. */
798 while (UTF8_IS_CONTINUATION(*s))
806 =for apidoc utf8_to_bytes
808 Converts a string C<s> of length C<len> from UTF-8 into native byte encoding.
809 Unlike C<bytes_to_utf8>, this over-writes the original string, and
810 updates len to contain the new length.
811 Returns zero on failure, setting C<len> to -1.
813 If you need a copy of the string, see C<bytes_from_utf8>.
819 Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len)
822 U8 * const send = s + *len;
825 PERL_ARGS_ASSERT_UTF8_TO_BYTES;
827 /* ensure valid UTF-8 and chars < 256 before updating string */
831 if (!UTF8_IS_INVARIANT(c) &&
832 (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send)
833 || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) {
834 *len = ((STRLEN) -1);
842 *d++ = (U8)utf8_to_uvchr(s, &ulen);
851 =for apidoc bytes_from_utf8
853 Converts a string C<s> of length C<len> from UTF-8 into native byte encoding.
854 Unlike C<utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to
855 the newly-created string, and updates C<len> to contain the new
856 length. Returns the original string if no conversion occurs, C<len>
857 is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to
858 0 if C<s> is converted or consisted entirely of characters that are invariant
859 in utf8 (i.e., US-ASCII on non-EBCDIC machines).
865 Perl_bytes_from_utf8(pTHX_ const U8 *s, STRLEN *len, bool *is_utf8)
872 PERL_ARGS_ASSERT_BYTES_FROM_UTF8;
878 /* ensure valid UTF-8 and chars < 256 before converting string */
879 for (send = s + *len; s < send;) {
881 if (!UTF8_IS_INVARIANT(c)) {
882 if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send &&
883 (c = *s++) && UTF8_IS_CONTINUATION(c))
892 Newx(d, (*len) - count + 1, U8);
893 s = start; start = d;
896 if (!UTF8_IS_INVARIANT(c)) {
897 /* Then it is two-byte encoded */
898 c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++);
899 c = ASCII_TO_NATIVE(c);
909 =for apidoc bytes_to_utf8
911 Converts a string C<s> of length C<len> from the native encoding into UTF-8.
912 Returns a pointer to the newly-created string, and sets C<len> to
913 reflect the new length.
915 A NUL character will be written after the end of the string.
917 If you want to convert to UTF-8 from encodings other than
918 the native (Latin1 or EBCDIC),
919 see sv_recode_to_utf8().
925 Perl_bytes_to_utf8(pTHX_ const U8 *s, STRLEN *len)
927 const U8 * const send = s + (*len);
931 PERL_ARGS_ASSERT_BYTES_TO_UTF8;
934 Newx(d, (*len) * 2 + 1, U8);
938 const UV uv = NATIVE_TO_ASCII(*s++);
939 if (UNI_IS_INVARIANT(uv))
940 *d++ = (U8)UTF_TO_NATIVE(uv);
942 *d++ = (U8)UTF8_EIGHT_BIT_HI(uv);
943 *d++ = (U8)UTF8_EIGHT_BIT_LO(uv);
952 * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8.
954 * Destination must be pre-extended to 3/2 source. Do not use in-place.
955 * We optimize for native, for obvious reasons. */
958 Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
963 PERL_ARGS_ASSERT_UTF16_TO_UTF8;
966 Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen %"UVuf, (UV)bytelen);
971 UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */
975 *d++ = UNI_TO_NATIVE(uv);
982 *d++ = (U8)(( uv >> 6) | 0xc0);
983 *d++ = (U8)(( uv & 0x3f) | 0x80);
986 if (uv >= 0xd800 && uv <= 0xdbff) { /* surrogates */
988 Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
990 UV low = (p[0] << 8) + p[1];
992 if (low < 0xdc00 || low > 0xdfff)
993 Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
994 uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000;
996 } else if (uv >= 0xdc00 && uv <= 0xdfff) {
997 Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
1000 *d++ = (U8)(( uv >> 12) | 0xe0);
1001 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
1002 *d++ = (U8)(( uv & 0x3f) | 0x80);
1006 *d++ = (U8)(( uv >> 18) | 0xf0);
1007 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
1008 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
1009 *d++ = (U8)(( uv & 0x3f) | 0x80);
1013 *newlen = d - dstart;
1017 /* Note: this one is slightly destructive of the source. */
1020 Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
1023 U8* const send = s + bytelen;
1025 PERL_ARGS_ASSERT_UTF16_TO_UTF8_REVERSED;
1028 Perl_croak(aTHX_ "panic: utf16_to_utf8_reversed: odd bytelen %"UVuf,
1032 const U8 tmp = s[0];
1037 return utf16_to_utf8(p, d, bytelen, newlen);
1040 /* for now these are all defined (inefficiently) in terms of the utf8 versions */
1043 Perl_is_uni_alnum(pTHX_ UV c)
1045 U8 tmpbuf[UTF8_MAXBYTES+1];
1046 uvchr_to_utf8(tmpbuf, c);
1047 return is_utf8_alnum(tmpbuf);
1051 Perl_is_uni_idfirst(pTHX_ UV c)
1053 U8 tmpbuf[UTF8_MAXBYTES+1];
1054 uvchr_to_utf8(tmpbuf, c);
1055 return is_utf8_idfirst(tmpbuf);
1059 Perl_is_uni_alpha(pTHX_ UV c)
1061 U8 tmpbuf[UTF8_MAXBYTES+1];
1062 uvchr_to_utf8(tmpbuf, c);
1063 return is_utf8_alpha(tmpbuf);
1067 Perl_is_uni_ascii(pTHX_ UV c)
1069 U8 tmpbuf[UTF8_MAXBYTES+1];
1070 uvchr_to_utf8(tmpbuf, c);
1071 return is_utf8_ascii(tmpbuf);
1075 Perl_is_uni_space(pTHX_ UV c)
1077 U8 tmpbuf[UTF8_MAXBYTES+1];
1078 uvchr_to_utf8(tmpbuf, c);
1079 return is_utf8_space(tmpbuf);
1083 Perl_is_uni_digit(pTHX_ UV c)
1085 U8 tmpbuf[UTF8_MAXBYTES+1];
1086 uvchr_to_utf8(tmpbuf, c);
1087 return is_utf8_digit(tmpbuf);
1091 Perl_is_uni_upper(pTHX_ UV c)
1093 U8 tmpbuf[UTF8_MAXBYTES+1];
1094 uvchr_to_utf8(tmpbuf, c);
1095 return is_utf8_upper(tmpbuf);
1099 Perl_is_uni_lower(pTHX_ UV c)
1101 U8 tmpbuf[UTF8_MAXBYTES+1];
1102 uvchr_to_utf8(tmpbuf, c);
1103 return is_utf8_lower(tmpbuf);
1107 Perl_is_uni_cntrl(pTHX_ UV c)
1109 U8 tmpbuf[UTF8_MAXBYTES+1];
1110 uvchr_to_utf8(tmpbuf, c);
1111 return is_utf8_cntrl(tmpbuf);
1115 Perl_is_uni_graph(pTHX_ UV c)
1117 U8 tmpbuf[UTF8_MAXBYTES+1];
1118 uvchr_to_utf8(tmpbuf, c);
1119 return is_utf8_graph(tmpbuf);
1123 Perl_is_uni_print(pTHX_ UV c)
1125 U8 tmpbuf[UTF8_MAXBYTES+1];
1126 uvchr_to_utf8(tmpbuf, c);
1127 return is_utf8_print(tmpbuf);
1131 Perl_is_uni_punct(pTHX_ UV c)
1133 U8 tmpbuf[UTF8_MAXBYTES+1];
1134 uvchr_to_utf8(tmpbuf, c);
1135 return is_utf8_punct(tmpbuf);
1139 Perl_is_uni_xdigit(pTHX_ UV c)
1141 U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
1142 uvchr_to_utf8(tmpbuf, c);
1143 return is_utf8_xdigit(tmpbuf);
1147 Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp)
1149 PERL_ARGS_ASSERT_TO_UNI_UPPER;
1151 uvchr_to_utf8(p, c);
1152 return to_utf8_upper(p, p, lenp);
1156 Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp)
1158 PERL_ARGS_ASSERT_TO_UNI_TITLE;
1160 uvchr_to_utf8(p, c);
1161 return to_utf8_title(p, p, lenp);
1165 Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp)
1167 PERL_ARGS_ASSERT_TO_UNI_LOWER;
1169 uvchr_to_utf8(p, c);
1170 return to_utf8_lower(p, p, lenp);
1174 Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp)
1176 PERL_ARGS_ASSERT_TO_UNI_FOLD;
1178 uvchr_to_utf8(p, c);
1179 return to_utf8_fold(p, p, lenp);
1182 /* for now these all assume no locale info available for Unicode > 255 */
1185 Perl_is_uni_alnum_lc(pTHX_ UV c)
1187 return is_uni_alnum(c); /* XXX no locale support yet */
1191 Perl_is_uni_idfirst_lc(pTHX_ UV c)
1193 return is_uni_idfirst(c); /* XXX no locale support yet */
1197 Perl_is_uni_alpha_lc(pTHX_ UV c)
1199 return is_uni_alpha(c); /* XXX no locale support yet */
1203 Perl_is_uni_ascii_lc(pTHX_ UV c)
1205 return is_uni_ascii(c); /* XXX no locale support yet */
1209 Perl_is_uni_space_lc(pTHX_ UV c)
1211 return is_uni_space(c); /* XXX no locale support yet */
1215 Perl_is_uni_digit_lc(pTHX_ UV c)
1217 return is_uni_digit(c); /* XXX no locale support yet */
1221 Perl_is_uni_upper_lc(pTHX_ UV c)
1223 return is_uni_upper(c); /* XXX no locale support yet */
1227 Perl_is_uni_lower_lc(pTHX_ UV c)
1229 return is_uni_lower(c); /* XXX no locale support yet */
1233 Perl_is_uni_cntrl_lc(pTHX_ UV c)
1235 return is_uni_cntrl(c); /* XXX no locale support yet */
1239 Perl_is_uni_graph_lc(pTHX_ UV c)
1241 return is_uni_graph(c); /* XXX no locale support yet */
1245 Perl_is_uni_print_lc(pTHX_ UV c)
1247 return is_uni_print(c); /* XXX no locale support yet */
1251 Perl_is_uni_punct_lc(pTHX_ UV c)
1253 return is_uni_punct(c); /* XXX no locale support yet */
1257 Perl_is_uni_xdigit_lc(pTHX_ UV c)
1259 return is_uni_xdigit(c); /* XXX no locale support yet */
1263 Perl_to_uni_upper_lc(pTHX_ U32 c)
1265 /* XXX returns only the first character -- do not use XXX */
1266 /* XXX no locale support yet */
1268 U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
1269 return (U32)to_uni_upper(c, tmpbuf, &len);
1273 Perl_to_uni_title_lc(pTHX_ U32 c)
1275 /* XXX returns only the first character XXX -- do not use XXX */
1276 /* XXX no locale support yet */
1278 U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
1279 return (U32)to_uni_title(c, tmpbuf, &len);
1283 Perl_to_uni_lower_lc(pTHX_ U32 c)
1285 /* XXX returns only the first character -- do not use XXX */
1286 /* XXX no locale support yet */
1288 U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
1289 return (U32)to_uni_lower(c, tmpbuf, &len);
1293 S_is_utf8_common(pTHX_ const U8 *const p, SV **swash,
1294 const char *const swashname)
1298 PERL_ARGS_ASSERT_IS_UTF8_COMMON;
1300 if (!is_utf8_char(p))
1303 *swash = swash_init("utf8", swashname, &PL_sv_undef, 1, 0);
1304 return swash_fetch(*swash, p, TRUE) != 0;
1308 Perl_is_utf8_alnum(pTHX_ const U8 *p)
1312 PERL_ARGS_ASSERT_IS_UTF8_ALNUM;
1314 /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true
1315 * descendant of isalnum(3), in other words, it doesn't
1316 * contain the '_'. --jhi */
1317 return is_utf8_common(p, &PL_utf8_alnum, "IsWord");
1321 Perl_is_utf8_idfirst(pTHX_ const U8 *p) /* The naming is historical. */
1325 PERL_ARGS_ASSERT_IS_UTF8_IDFIRST;
1329 /* is_utf8_idstart would be more logical. */
1330 return is_utf8_common(p, &PL_utf8_idstart, "IdStart");
1334 Perl_is_utf8_idcont(pTHX_ const U8 *p)
1338 PERL_ARGS_ASSERT_IS_UTF8_IDCONT;
1342 return is_utf8_common(p, &PL_utf8_idcont, "IdContinue");
1346 Perl_is_utf8_alpha(pTHX_ const U8 *p)
1350 PERL_ARGS_ASSERT_IS_UTF8_ALPHA;
1352 return is_utf8_common(p, &PL_utf8_alpha, "IsAlpha");
1356 Perl_is_utf8_ascii(pTHX_ const U8 *p)
1360 PERL_ARGS_ASSERT_IS_UTF8_ASCII;
1362 return is_utf8_common(p, &PL_utf8_ascii, "IsAscii");
1366 Perl_is_utf8_space(pTHX_ const U8 *p)
1370 PERL_ARGS_ASSERT_IS_UTF8_SPACE;
1372 return is_utf8_common(p, &PL_utf8_space, "IsSpacePerl");
1376 Perl_is_utf8_perl_space(pTHX_ const U8 *p)
1380 PERL_ARGS_ASSERT_IS_UTF8_PERL_SPACE;
1382 return is_utf8_common(p, &PL_utf8_perl_space, "IsPerlSpace");
1386 Perl_is_utf8_perl_word(pTHX_ const U8 *p)
1390 PERL_ARGS_ASSERT_IS_UTF8_PERL_WORD;
1392 return is_utf8_common(p, &PL_utf8_perl_word, "IsPerlWord");
1396 Perl_is_utf8_digit(pTHX_ const U8 *p)
1400 PERL_ARGS_ASSERT_IS_UTF8_DIGIT;
1402 return is_utf8_common(p, &PL_utf8_digit, "IsDigit");
1406 Perl_is_utf8_posix_digit(pTHX_ const U8 *p)
1410 PERL_ARGS_ASSERT_IS_UTF8_POSIX_DIGIT;
1412 return is_utf8_common(p, &PL_utf8_posix_digit, "IsPosixDigit");
1416 Perl_is_utf8_upper(pTHX_ const U8 *p)
1420 PERL_ARGS_ASSERT_IS_UTF8_UPPER;
1422 return is_utf8_common(p, &PL_utf8_upper, "IsUppercase");
1426 Perl_is_utf8_lower(pTHX_ const U8 *p)
1430 PERL_ARGS_ASSERT_IS_UTF8_LOWER;
1432 return is_utf8_common(p, &PL_utf8_lower, "IsLowercase");
1436 Perl_is_utf8_cntrl(pTHX_ const U8 *p)
1440 PERL_ARGS_ASSERT_IS_UTF8_CNTRL;
1442 return is_utf8_common(p, &PL_utf8_cntrl, "IsCntrl");
1446 Perl_is_utf8_graph(pTHX_ const U8 *p)
1450 PERL_ARGS_ASSERT_IS_UTF8_GRAPH;
1452 return is_utf8_common(p, &PL_utf8_graph, "IsGraph");
1456 Perl_is_utf8_print(pTHX_ const U8 *p)
1460 PERL_ARGS_ASSERT_IS_UTF8_PRINT;
1462 return is_utf8_common(p, &PL_utf8_print, "IsPrint");
1466 Perl_is_utf8_punct(pTHX_ const U8 *p)
1470 PERL_ARGS_ASSERT_IS_UTF8_PUNCT;
1472 return is_utf8_common(p, &PL_utf8_punct, "IsPunct");
1476 Perl_is_utf8_xdigit(pTHX_ const U8 *p)
1480 PERL_ARGS_ASSERT_IS_UTF8_XDIGIT;
1482 return is_utf8_common(p, &PL_utf8_xdigit, "IsXDigit");
1486 Perl_is_utf8_mark(pTHX_ const U8 *p)
1490 PERL_ARGS_ASSERT_IS_UTF8_MARK;
1492 return is_utf8_common(p, &PL_utf8_mark, "IsM");
1496 Perl_is_utf8_X_begin(pTHX_ const U8 *p)
1500 PERL_ARGS_ASSERT_IS_UTF8_X_BEGIN;
1502 return is_utf8_common(p, &PL_utf8_X_begin, "_X_Begin");
1506 Perl_is_utf8_X_extend(pTHX_ const U8 *p)
1510 PERL_ARGS_ASSERT_IS_UTF8_X_EXTEND;
1512 return is_utf8_common(p, &PL_utf8_X_extend, "_X_Extend");
1516 Perl_is_utf8_X_prepend(pTHX_ const U8 *p)
1520 PERL_ARGS_ASSERT_IS_UTF8_X_PREPEND;
1522 return is_utf8_common(p, &PL_utf8_X_prepend, "GCB=Prepend");
1526 Perl_is_utf8_X_non_hangul(pTHX_ const U8 *p)
1530 PERL_ARGS_ASSERT_IS_UTF8_X_NON_HANGUL;
1532 return is_utf8_common(p, &PL_utf8_X_non_hangul, "HST=Not_Applicable");
1536 Perl_is_utf8_X_L(pTHX_ const U8 *p)
1540 PERL_ARGS_ASSERT_IS_UTF8_X_L;
1542 return is_utf8_common(p, &PL_utf8_X_L, "GCB=L");
1546 Perl_is_utf8_X_LV(pTHX_ const U8 *p)
1550 PERL_ARGS_ASSERT_IS_UTF8_X_LV;
1552 return is_utf8_common(p, &PL_utf8_X_LV, "GCB=LV");
1556 Perl_is_utf8_X_LVT(pTHX_ const U8 *p)
1560 PERL_ARGS_ASSERT_IS_UTF8_X_LVT;
1562 return is_utf8_common(p, &PL_utf8_X_LVT, "GCB=LVT");
1566 Perl_is_utf8_X_T(pTHX_ const U8 *p)
1570 PERL_ARGS_ASSERT_IS_UTF8_X_T;
1572 return is_utf8_common(p, &PL_utf8_X_T, "GCB=T");
1576 Perl_is_utf8_X_V(pTHX_ const U8 *p)
1580 PERL_ARGS_ASSERT_IS_UTF8_X_V;
1582 return is_utf8_common(p, &PL_utf8_X_V, "GCB=V");
1586 Perl_is_utf8_X_LV_LVT_V(pTHX_ const U8 *p)
1590 PERL_ARGS_ASSERT_IS_UTF8_X_LV_LVT_V;
1592 return is_utf8_common(p, &PL_utf8_X_LV_LVT_V, "_X_LV_LVT_V");
1596 =for apidoc to_utf8_case
1598 The "p" contains the pointer to the UTF-8 string encoding
1599 the character that is being converted.
1601 The "ustrp" is a pointer to the character buffer to put the
1602 conversion result to. The "lenp" is a pointer to the length
1605 The "swashp" is a pointer to the swash to use.
1607 Both the special and normal mappings are stored lib/unicore/To/Foo.pl,
1608 and loaded by SWASHNEW, using lib/utf8_heavy.pl. The special (usually,
1609 but not always, a multicharacter mapping), is tried first.
1611 The "special" is a string like "utf8::ToSpecLower", which means the
1612 hash %utf8::ToSpecLower. The access to the hash is through
1613 Perl_to_utf8_case().
1615 The "normal" is a string like "ToLower" which means the swash
1621 Perl_to_utf8_case(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp,
1622 SV **swashp, const char *normal, const char *special)
1625 U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
1627 const UV uv0 = utf8_to_uvchr(p, NULL);
1628 /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings
1629 * are necessary in EBCDIC, they are redundant no-ops
1630 * in ASCII-ish platforms, and hopefully optimized away. */
1631 const UV uv1 = NATIVE_TO_UNI(uv0);
1633 PERL_ARGS_ASSERT_TO_UTF8_CASE;
1635 uvuni_to_utf8(tmpbuf, uv1);
1637 if (!*swashp) /* load on-demand */
1638 *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0);
1639 /* This is the beginnings of a skeleton of code to read the info section
1640 * that is in all the swashes in case we ever want to do that, so one can
1641 * read things whose maps aren't code points, and whose default if missing
1642 * is not to the code point itself. This was just to see if it actually
1643 * worked. Details on what the possibilities are are in perluniprops.pod
1644 HV * const hv = get_hv("utf8::SwashInfo", 0);
1647 svp = hv_fetch(hv, (const char*)normal, strlen(normal), FALSE);
1650 HV * const this_hash = SvRV(*svp);
1651 svp = hv_fetch(this_hash, "type", strlen("type"), FALSE);
1652 s = SvPV_const(*svp, len);
1656 /* The 0xDF is the only special casing Unicode code point below 0x100. */
1657 if (special && (uv1 == 0xDF || uv1 > 0xFF)) {
1658 /* It might be "special" (sometimes, but not always,
1659 * a multicharacter mapping) */
1660 HV * const hv = get_hv(special, 0);
1664 (svp = hv_fetch(hv, (const char*)tmpbuf, UNISKIP(uv1), FALSE)) &&
1668 s = SvPV_const(*svp, len);
1670 len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp;
1673 /* If we have EBCDIC we need to remap the characters
1674 * since any characters in the low 256 are Unicode
1675 * code points, not EBCDIC. */
1676 U8 *t = (U8*)s, *tend = t + len, *d;
1683 const UV c = utf8_to_uvchr(t, &tlen);
1685 d = uvchr_to_utf8(d, UNI_TO_NATIVE(c));
1694 d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t));
1699 Copy(tmpbuf, ustrp, len, U8);
1701 Copy(s, ustrp, len, U8);
1707 if (!len && *swashp) {
1708 const UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE);
1711 /* It was "normal" (a single character mapping). */
1712 const UV uv3 = UNI_TO_NATIVE(uv2);
1713 len = uvchr_to_utf8(ustrp, uv3) - ustrp;
1717 if (!len) /* Neither: just copy. In other words, there was no mapping
1718 defined, which means that the code point maps to itself */
1719 len = uvchr_to_utf8(ustrp, uv0) - ustrp;
1724 return len ? utf8_to_uvchr(ustrp, 0) : 0;
1728 =for apidoc to_utf8_upper
1730 Convert the UTF-8 encoded character at p to its uppercase version and
1731 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1732 that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since
1733 the uppercase version may be longer than the original character.
1735 The first character of the uppercased version is returned
1736 (but note, as explained above, that there may be more.)
1741 Perl_to_utf8_upper(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp)
1745 PERL_ARGS_ASSERT_TO_UTF8_UPPER;
1747 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1748 &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper");
1752 =for apidoc to_utf8_title
1754 Convert the UTF-8 encoded character at p to its titlecase version and
1755 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1756 that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the
1757 titlecase version may be longer than the original character.
1759 The first character of the titlecased version is returned
1760 (but note, as explained above, that there may be more.)
1765 Perl_to_utf8_title(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp)
1769 PERL_ARGS_ASSERT_TO_UTF8_TITLE;
1771 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1772 &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle");
1776 =for apidoc to_utf8_lower
1778 Convert the UTF-8 encoded character at p to its lowercase version and
1779 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1780 that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the
1781 lowercase version may be longer than the original character.
1783 The first character of the lowercased version is returned
1784 (but note, as explained above, that there may be more.)
1789 Perl_to_utf8_lower(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp)
1793 PERL_ARGS_ASSERT_TO_UTF8_LOWER;
1795 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1796 &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower");
1800 =for apidoc to_utf8_fold
1802 Convert the UTF-8 encoded character at p to its foldcase version and
1803 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1804 that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the
1805 foldcase version may be longer than the original character (up to
1808 The first character of the foldcased version is returned
1809 (but note, as explained above, that there may be more.)
1814 Perl_to_utf8_fold(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp)
1818 PERL_ARGS_ASSERT_TO_UTF8_FOLD;
1820 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1821 &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold");
1825 * A "swash" is a swatch hash.
1826 * A "swatch" is a bit vector generated by utf8.c:S_swash_get().
1827 * C<pkg> is a pointer to a package name for SWASHNEW, should be "utf8".
1828 * For other parameters, see utf8::SWASHNEW in lib/utf8_heavy.pl.
1831 Perl_swash_init(pTHX_ const char* pkg, const char* name, SV *listsv, I32 minbits, I32 none)
1836 const size_t pkg_len = strlen(pkg);
1837 const size_t name_len = strlen(name);
1838 HV * const stash = gv_stashpvn(pkg, pkg_len, 0);
1841 PERL_ARGS_ASSERT_SWASH_INIT;
1843 PUSHSTACKi(PERLSI_MAGIC);
1848 if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */
1850 errsv_save = newSVsv(ERRSV);
1851 /* It is assumed that callers of this routine are not passing in any
1852 user derived data. */
1853 /* Need to do this after save_re_context() as it will set PL_tainted to
1854 1 while saving $1 etc (see the code after getrx: in Perl_magic_get).
1855 Even line to create errsv_save can turn on PL_tainted. */
1856 SAVEBOOL(PL_tainted);
1858 Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpvn(pkg,pkg_len),
1861 sv_setsv(ERRSV, errsv_save);
1862 SvREFCNT_dec(errsv_save);
1868 mPUSHp(pkg, pkg_len);
1869 mPUSHp(name, name_len);
1874 errsv_save = newSVsv(ERRSV);
1875 if (call_method("SWASHNEW", G_SCALAR))
1876 retval = newSVsv(*PL_stack_sp--);
1878 retval = &PL_sv_undef;
1880 sv_setsv(ERRSV, errsv_save);
1881 SvREFCNT_dec(errsv_save);
1884 if (IN_PERL_COMPILETIME) {
1885 CopHINTS_set(PL_curcop, PL_hints);
1887 if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) {
1889 Perl_croak(aTHX_ "Can't find Unicode property definition \"%"SVf"\"",
1891 Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref");
1897 /* This API is wrong for special case conversions since we may need to
1898 * return several Unicode characters for a single Unicode character
1899 * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is
1900 * the lower-level routine, and it is similarly broken for returning
1901 * multiple values. --jhi */
1902 /* Now SWASHGET is recasted into S_swash_get in this file. */
1905 * Returns the value of property/mapping C<swash> for the first character
1906 * of the string C<ptr>. If C<do_utf8> is true, the string C<ptr> is
1907 * assumed to be in utf8. If C<do_utf8> is false, the string C<ptr> is
1908 * assumed to be in native 8-bit encoding. Caches the swatch in C<swash>.
1911 Perl_swash_fetch(pTHX_ SV *swash, const U8 *ptr, bool do_utf8)
1914 HV *const hv = MUTABLE_HV(SvRV(swash));
1919 const U8 *tmps = NULL;
1923 const UV c = NATIVE_TO_ASCII(*ptr);
1925 PERL_ARGS_ASSERT_SWASH_FETCH;
1927 if (!do_utf8 && !UNI_IS_INVARIANT(c)) {
1928 tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c);
1929 tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c);
1932 /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ
1933 * then the "swatch" is a vec() for all the chars which start
1935 * So the key in the hash (klen) is length of encoded char -1
1937 klen = UTF8SKIP(ptr) - 1;
1941 /* If char is invariant then swatch is for all the invariant chars
1942 * In both UTF-8 and UTF-8-MOD that happens to be UTF_CONTINUATION_MARK
1944 needents = UTF_CONTINUATION_MARK;
1945 off = NATIVE_TO_UTF(ptr[klen]);
1948 /* If char is encoded then swatch is for the prefix */
1949 needents = (1 << UTF_ACCUMULATION_SHIFT);
1950 off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK;
1954 * This single-entry cache saves about 1/3 of the utf8 overhead in test
1955 * suite. (That is, only 7-8% overall over just a hash cache. Still,
1956 * it's nothing to sniff at.) Pity we usually come through at least
1957 * two function calls to get here...
1959 * NB: this code assumes that swatches are never modified, once generated!
1962 if (hv == PL_last_swash_hv &&
1963 klen == PL_last_swash_klen &&
1964 (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) )
1966 tmps = PL_last_swash_tmps;
1967 slen = PL_last_swash_slen;
1970 /* Try our second-level swatch cache, kept in a hash. */
1971 SV** svp = hv_fetch(hv, (const char*)ptr, klen, FALSE);
1973 /* If not cached, generate it via swash_get */
1974 if (!svp || !SvPOK(*svp)
1975 || !(tmps = (const U8*)SvPV_const(*svp, slen))) {
1976 /* We use utf8n_to_uvuni() as we want an index into
1977 Unicode tables, not a native character number.
1979 const UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXBYTES, 0,
1981 0 : UTF8_ALLOW_ANY);
1982 swatch = swash_get(swash,
1983 /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */
1984 (klen) ? (code_point & ~(needents - 1)) : 0,
1987 if (IN_PERL_COMPILETIME)
1988 CopHINTS_set(PL_curcop, PL_hints);
1990 svp = hv_store(hv, (const char *)ptr, klen, swatch, 0);
1992 if (!svp || !(tmps = (U8*)SvPV(*svp, slen))
1993 || (slen << 3) < needents)
1994 Perl_croak(aTHX_ "panic: swash_fetch got improper swatch");
1997 PL_last_swash_hv = hv;
1998 assert(klen <= sizeof(PL_last_swash_key));
1999 PL_last_swash_klen = (U8)klen;
2000 /* FIXME change interpvar.h? */
2001 PL_last_swash_tmps = (U8 *) tmps;
2002 PL_last_swash_slen = slen;
2004 Copy(ptr, PL_last_swash_key, klen, U8);
2007 switch ((int)((slen << 3) / needents)) {
2009 bit = 1 << (off & 7);
2011 return (tmps[off] & bit) != 0;
2016 return (tmps[off] << 8) + tmps[off + 1] ;
2019 return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ;
2021 Perl_croak(aTHX_ "panic: swash_fetch got swatch of unexpected bit width");
2022 NORETURN_FUNCTION_END;
2026 * Returns a swatch (a bit vector string) for a code point sequence
2027 * that starts from the value C<start> and comprises the number C<span>.
2028 * A C<swash> must be an object created by SWASHNEW (see lib/utf8_heavy.pl).
2029 * Should be used via swash_fetch, which will cache the swatch in C<swash>.
2032 S_swash_get(pTHX_ SV* swash, UV start, UV span)
2035 U8 *l, *lend, *x, *xend, *s;
2036 STRLEN lcur, xcur, scur;
2037 HV *const hv = MUTABLE_HV(SvRV(swash));
2038 SV** const listsvp = hv_fetchs(hv, "LIST", FALSE);
2039 SV** const typesvp = hv_fetchs(hv, "TYPE", FALSE);
2040 SV** const bitssvp = hv_fetchs(hv, "BITS", FALSE);
2041 SV** const nonesvp = hv_fetchs(hv, "NONE", FALSE);
2042 SV** const extssvp = hv_fetchs(hv, "EXTRAS", FALSE);
2043 const U8* const typestr = (U8*)SvPV_nolen(*typesvp);
2044 const int typeto = typestr[0] == 'T' && typestr[1] == 'o';
2045 const STRLEN bits = SvUV(*bitssvp);
2046 const STRLEN octets = bits >> 3; /* if bits == 1, then octets == 0 */
2047 const UV none = SvUV(*nonesvp);
2048 const UV end = start + span;
2050 PERL_ARGS_ASSERT_SWASH_GET;
2052 if (bits != 1 && bits != 8 && bits != 16 && bits != 32) {
2053 Perl_croak(aTHX_ "panic: swash_get doesn't expect bits %"UVuf,
2057 /* create and initialize $swatch */
2058 scur = octets ? (span * octets) : (span + 7) / 8;
2059 swatch = newSV(scur);
2061 s = (U8*)SvPVX(swatch);
2062 if (octets && none) {
2063 const U8* const e = s + scur;
2066 *s++ = (U8)(none & 0xff);
2067 else if (bits == 16) {
2068 *s++ = (U8)((none >> 8) & 0xff);
2069 *s++ = (U8)( none & 0xff);
2071 else if (bits == 32) {
2072 *s++ = (U8)((none >> 24) & 0xff);
2073 *s++ = (U8)((none >> 16) & 0xff);
2074 *s++ = (U8)((none >> 8) & 0xff);
2075 *s++ = (U8)( none & 0xff);
2081 (void)memzero((U8*)s, scur + 1);
2083 SvCUR_set(swatch, scur);
2084 s = (U8*)SvPVX(swatch);
2086 /* read $swash->{LIST} */
2087 l = (U8*)SvPV(*listsvp, lcur);
2092 I32 flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX;
2094 U8* const nl = (U8*)memchr(l, '\n', lend - l);
2097 min = grok_hex((char *)l, &numlen, &flags, NULL);
2101 l = nl + 1; /* 1 is length of "\n" */
2105 l = lend; /* to LIST's end at which \n is not found */
2111 flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX;
2113 max = grok_hex((char *)l, &numlen, &flags, NULL);
2122 flags = PERL_SCAN_SILENT_ILLDIGIT |
2123 PERL_SCAN_DISALLOW_PREFIX;
2125 val = grok_hex((char *)l, &numlen, &flags, NULL);
2134 Perl_croak(aTHX_ "%s: illegal mapping '%s'",
2140 val = 0; /* bits == 1, then val should be ignored */
2147 Perl_croak(aTHX_ "%s: illegal mapping '%s'", typestr, l);
2151 val = 0; /* bits == 1, then val should be ignored */
2165 if (!none || val < none) {
2170 for (key = min; key <= max; key++) {
2174 /* offset must be non-negative (start <= min <= key < end) */
2175 offset = octets * (key - start);
2177 s[offset] = (U8)(val & 0xff);
2178 else if (bits == 16) {
2179 s[offset ] = (U8)((val >> 8) & 0xff);
2180 s[offset + 1] = (U8)( val & 0xff);
2182 else if (bits == 32) {
2183 s[offset ] = (U8)((val >> 24) & 0xff);
2184 s[offset + 1] = (U8)((val >> 16) & 0xff);
2185 s[offset + 2] = (U8)((val >> 8) & 0xff);
2186 s[offset + 3] = (U8)( val & 0xff);
2189 if (!none || val < none)
2193 else { /* bits == 1, then val should be ignored */
2197 for (key = min; key <= max; key++) {
2198 const STRLEN offset = (STRLEN)(key - start);
2201 s[offset >> 3] |= 1 << (offset & 7);
2207 /* read $swash->{EXTRAS} */
2208 x = (U8*)SvPV(*extssvp, xcur);
2216 SV **otherbitssvp, *other;
2220 const U8 opc = *x++;
2224 nl = (U8*)memchr(x, '\n', xend - x);
2226 if (opc != '-' && opc != '+' && opc != '!' && opc != '&') {
2228 x = nl + 1; /* 1 is length of "\n" */
2232 x = xend; /* to EXTRAS' end at which \n is not found */
2239 namelen = nl - namestr;
2243 namelen = xend - namestr;
2247 othersvp = hv_fetch(hv, (char *)namestr, namelen, FALSE);
2248 otherhv = MUTABLE_HV(SvRV(*othersvp));
2249 otherbitssvp = hv_fetchs(otherhv, "BITS", FALSE);
2250 otherbits = (STRLEN)SvUV(*otherbitssvp);
2251 if (bits < otherbits)
2252 Perl_croak(aTHX_ "panic: swash_get found swatch size mismatch");
2254 /* The "other" swatch must be destroyed after. */
2255 other = swash_get(*othersvp, start, span);
2256 o = (U8*)SvPV(other, olen);
2259 Perl_croak(aTHX_ "panic: swash_get got improper swatch");
2261 s = (U8*)SvPV(swatch, slen);
2262 if (bits == 1 && otherbits == 1) {
2264 Perl_croak(aTHX_ "panic: swash_get found swatch length mismatch");
2288 STRLEN otheroctets = otherbits >> 3;
2290 U8* const send = s + slen;
2295 if (otherbits == 1) {
2296 otherval = (o[offset >> 3] >> (offset & 7)) & 1;
2300 STRLEN vlen = otheroctets;
2308 if (opc == '+' && otherval)
2309 NOOP; /* replace with otherval */
2310 else if (opc == '!' && !otherval)
2312 else if (opc == '-' && otherval)
2314 else if (opc == '&' && !otherval)
2317 s += octets; /* no replacement */
2322 *s++ = (U8)( otherval & 0xff);
2323 else if (bits == 16) {
2324 *s++ = (U8)((otherval >> 8) & 0xff);
2325 *s++ = (U8)( otherval & 0xff);
2327 else if (bits == 32) {
2328 *s++ = (U8)((otherval >> 24) & 0xff);
2329 *s++ = (U8)((otherval >> 16) & 0xff);
2330 *s++ = (U8)((otherval >> 8) & 0xff);
2331 *s++ = (U8)( otherval & 0xff);
2335 sv_free(other); /* through with it! */
2341 =for apidoc uvchr_to_utf8
2343 Adds the UTF-8 representation of the Native codepoint C<uv> to the end
2344 of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free
2345 bytes available. The return value is the pointer to the byte after the
2346 end of the new character. In other words,
2348 d = uvchr_to_utf8(d, uv);
2350 is the recommended wide native character-aware way of saying
2357 /* On ASCII machines this is normally a macro but we want a
2358 real function in case XS code wants it
2361 Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv)
2363 PERL_ARGS_ASSERT_UVCHR_TO_UTF8;
2365 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0);
2369 Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
2371 PERL_ARGS_ASSERT_UVCHR_TO_UTF8_FLAGS;
2373 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags);
2377 =for apidoc utf8n_to_uvchr
2380 Returns the native character value of the first character in the string
2382 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
2383 length, in bytes, of that character.
2385 Allows length and flags to be passed to low level routine.
2389 /* On ASCII machines this is normally a macro but we want
2390 a real function in case XS code wants it
2393 Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen,
2396 const UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags);
2398 PERL_ARGS_ASSERT_UTF8N_TO_UVCHR;
2400 return UNI_TO_NATIVE(uv);
2404 =for apidoc pv_uni_display
2406 Build to the scalar dsv a displayable version of the string spv,
2407 length len, the displayable version being at most pvlim bytes long
2408 (if longer, the rest is truncated and "..." will be appended).
2410 The flags argument can have UNI_DISPLAY_ISPRINT set to display
2411 isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH
2412 to display the \\[nrfta\\] as the backslashed versions (like '\n')
2413 (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\).
2414 UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both
2415 UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on.
2417 The pointer to the PV of the dsv is returned.
2421 Perl_pv_uni_display(pTHX_ SV *dsv, const U8 *spv, STRLEN len, STRLEN pvlim, UV flags)
2426 PERL_ARGS_ASSERT_PV_UNI_DISPLAY;
2430 for (s = (const char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) {
2432 /* This serves double duty as a flag and a character to print after
2433 a \ when flags & UNI_DISPLAY_BACKSLASH is true.
2437 if (pvlim && SvCUR(dsv) >= pvlim) {
2441 u = utf8_to_uvchr((U8*)s, 0);
2443 const unsigned char c = (unsigned char)u & 0xFF;
2444 if (flags & UNI_DISPLAY_BACKSLASH) {
2461 const char string = ok;
2462 sv_catpvs(dsv, "\\");
2463 sv_catpvn(dsv, &string, 1);
2466 /* isPRINT() is the locale-blind version. */
2467 if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(c)) {
2468 const char string = c;
2469 sv_catpvn(dsv, &string, 1);
2474 Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u);
2477 sv_catpvs(dsv, "...");
2483 =for apidoc sv_uni_display
2485 Build to the scalar dsv a displayable version of the scalar sv,
2486 the displayable version being at most pvlim bytes long
2487 (if longer, the rest is truncated and "..." will be appended).
2489 The flags argument is as in pv_uni_display().
2491 The pointer to the PV of the dsv is returned.
2496 Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags)
2498 PERL_ARGS_ASSERT_SV_UNI_DISPLAY;
2500 return Perl_pv_uni_display(aTHX_ dsv, (const U8*)SvPVX_const(ssv),
2501 SvCUR(ssv), pvlim, flags);
2505 =for apidoc ibcmp_utf8
2507 Return true if the strings s1 and s2 differ case-insensitively, false
2508 if not (if they are equal case-insensitively). If u1 is true, the
2509 string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true,
2510 the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2
2511 are false, the respective string is assumed to be in native 8-bit
2514 If the pe1 and pe2 are non-NULL, the scanning pointers will be copied
2515 in there (they will point at the beginning of the I<next> character).
2516 If the pointers behind pe1 or pe2 are non-NULL, they are the end
2517 pointers beyond which scanning will not continue under any
2518 circumstances. If the byte lengths l1 and l2 are non-zero, s1+l1 and
2519 s2+l2 will be used as goal end pointers that will also stop the scan,
2520 and which qualify towards defining a successful match: all the scans
2521 that define an explicit length must reach their goal pointers for
2522 a match to succeed).
2524 For case-insensitiveness, the "casefolding" of Unicode is used
2525 instead of upper/lowercasing both the characters, see
2526 http://www.unicode.org/unicode/reports/tr21/ (Case Mappings).
2530 Perl_ibcmp_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2)
2533 register const U8 *p1 = (const U8*)s1;
2534 register const U8 *p2 = (const U8*)s2;
2535 register const U8 *f1 = NULL;
2536 register const U8 *f2 = NULL;
2537 register U8 *e1 = NULL;
2538 register U8 *q1 = NULL;
2539 register U8 *e2 = NULL;
2540 register U8 *q2 = NULL;
2541 STRLEN n1 = 0, n2 = 0;
2542 U8 foldbuf1[UTF8_MAXBYTES_CASE+1];
2543 U8 foldbuf2[UTF8_MAXBYTES_CASE+1];
2545 STRLEN foldlen1, foldlen2;
2548 PERL_ARGS_ASSERT_IBCMP_UTF8;
2552 /* assert(e1 || l1); */
2553 if (e1 == 0 || (l1 && l1 < (UV)(e1 - (const U8*)s1)))
2554 f1 = (const U8*)s1 + l1;
2557 /* assert(e2 || l2); */
2558 if (e2 == 0 || (l2 && l2 < (UV)(e2 - (const U8*)s2)))
2559 f2 = (const U8*)s2 + l2;
2561 /* This shouldn't happen. However, putting an assert() there makes some
2563 /* assert((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0)); */
2564 if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0))
2565 return 1; /* mismatch; possible infinite loop or false positive */
2568 natbuf[1] = 0; /* Need to terminate the buffer. */
2570 while ((e1 == 0 || p1 < e1) &&
2571 (f1 == 0 || p1 < f1) &&
2572 (e2 == 0 || p2 < e2) &&
2573 (f2 == 0 || p2 < f2)) {
2576 to_utf8_fold(p1, foldbuf1, &foldlen1);
2578 uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p1)));
2579 to_utf8_fold(natbuf, foldbuf1, &foldlen1);
2586 to_utf8_fold(p2, foldbuf2, &foldlen2);
2588 uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p2)));
2589 to_utf8_fold(natbuf, foldbuf2, &foldlen2);
2595 if ( UTF8SKIP(q1) != UTF8SKIP(q2) ||
2596 (UTF8SKIP(q1) == 1 && *q1 != *q2) ||
2597 memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) )
2598 return 1; /* mismatch */
2605 p1 += u1 ? UTF8SKIP(p1) : 1;
2607 p2 += u2 ? UTF8SKIP(p2) : 1;
2611 /* A match is defined by all the scans that specified
2612 * an explicit length reaching their final goals. */
2613 match = (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2);
2622 return match ? 0 : 1; /* 0 match, 1 mismatch */
2627 * c-indentation-style: bsd
2629 * indent-tabs-mode: t
2632 * ex: set ts=8 sts=4 sw=4 noet: