3 * Copyright (c) 1998-2001, Larry Wall
5 * You may distribute under the terms of either the GNU General Public
6 * License or the Artistic License, as specified in the README file.
11 * 'What a fix!' said Sam. 'That's the one place in all the lands we've ever
12 * heard of that we don't want to see any closer; and that's the one place
13 * we're trying to get to! And that's just where we can't get, nohow.'
15 * 'Well do I understand your speech,' he answered in the same language;
16 * 'yet few strangers do so. Why then do you not speak in the Common Tongue,
17 * as is the custom in the West, if you wish to be answered?'
19 * ...the travellers perceived that the floor was paved with stones of many
20 * hues; branching runes and strange devices intertwined beneath their feet.
24 #define PERL_IN_UTF8_C
30 =for apidoc A|U8*|uvuni_to_utf8|U8 *d|UV uv
32 Adds the UTF8 representation of the Unicode codepoint C<uv> to the end
33 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
34 bytes available. The return value is the pointer to the byte after the
35 end of the new character. In other words,
37 d = uvuni_to_utf8(d, uv);
39 is the recommended Unicode-aware way of saying
47 Perl_uvuni_to_utf8(pTHX_ U8 *d, UV uv)
49 if (UNI_IS_INVARIANT(uv)) {
50 *d++ = UTF_TO_NATIVE(uv);
53 #if defined(EBCDIC) || 1 /* always for testing */
55 STRLEN len = UNISKIP(uv);
58 *p-- = UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK);
59 uv >>= UTF_ACCUMULATION_SHIFT;
61 *p = UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len));
64 #else /* Non loop style */
66 *d++ = (( uv >> 6) | 0xc0);
67 *d++ = (( uv & 0x3f) | 0x80);
71 *d++ = (( uv >> 12) | 0xe0);
72 *d++ = (((uv >> 6) & 0x3f) | 0x80);
73 *d++ = (( uv & 0x3f) | 0x80);
77 *d++ = (( uv >> 18) | 0xf0);
78 *d++ = (((uv >> 12) & 0x3f) | 0x80);
79 *d++ = (((uv >> 6) & 0x3f) | 0x80);
80 *d++ = (( uv & 0x3f) | 0x80);
84 *d++ = (( uv >> 24) | 0xf8);
85 *d++ = (((uv >> 18) & 0x3f) | 0x80);
86 *d++ = (((uv >> 12) & 0x3f) | 0x80);
87 *d++ = (((uv >> 6) & 0x3f) | 0x80);
88 *d++ = (( uv & 0x3f) | 0x80);
91 if (uv < 0x80000000) {
92 *d++ = (( uv >> 30) | 0xfc);
93 *d++ = (((uv >> 24) & 0x3f) | 0x80);
94 *d++ = (((uv >> 18) & 0x3f) | 0x80);
95 *d++ = (((uv >> 12) & 0x3f) | 0x80);
96 *d++ = (((uv >> 6) & 0x3f) | 0x80);
97 *d++ = (( uv & 0x3f) | 0x80);
101 if (uv < UTF8_QUAD_MAX)
104 *d++ = 0xfe; /* Can't match U+FEFF! */
105 *d++ = (((uv >> 30) & 0x3f) | 0x80);
106 *d++ = (((uv >> 24) & 0x3f) | 0x80);
107 *d++ = (((uv >> 18) & 0x3f) | 0x80);
108 *d++ = (((uv >> 12) & 0x3f) | 0x80);
109 *d++ = (((uv >> 6) & 0x3f) | 0x80);
110 *d++ = (( uv & 0x3f) | 0x80);
115 *d++ = 0xff; /* Can't match U+FFFE! */
116 *d++ = 0x80; /* 6 Reserved bits */
117 *d++ = (((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */
118 *d++ = (((uv >> 54) & 0x3f) | 0x80);
119 *d++ = (((uv >> 48) & 0x3f) | 0x80);
120 *d++ = (((uv >> 42) & 0x3f) | 0x80);
121 *d++ = (((uv >> 36) & 0x3f) | 0x80);
122 *d++ = (((uv >> 30) & 0x3f) | 0x80);
123 *d++ = (((uv >> 24) & 0x3f) | 0x80);
124 *d++ = (((uv >> 18) & 0x3f) | 0x80);
125 *d++ = (((uv >> 12) & 0x3f) | 0x80);
126 *d++ = (((uv >> 6) & 0x3f) | 0x80);
127 *d++ = (( uv & 0x3f) | 0x80);
131 #endif /* Loop style */
135 =for apidoc A|U8*|uvchr_to_utf8|U8 *d|UV uv
137 Adds the UTF8 representation of the Native codepoint C<uv> to the end
138 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
139 bytes available. The return value is the pointer to the byte after the
140 end of the new character. In other words,
142 d = uvchr_to_utf8(d, uv);
144 is the recommended wide native character-aware way of saying
152 Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv)
154 return Perl_uvuni_to_utf8(aTHX_ d, NATIVE_TO_UNI(uv));
159 =for apidoc A|STRLEN|is_utf8_char|U8 *s
161 Tests if some arbitrary number of bytes begins in a valid UTF-8
162 character. Note that an INVARIANT (i.e. ASCII) character is a valid UTF-8 character.
163 The actual number of bytes in the UTF-8 character will be returned if
164 it is valid, otherwise 0.
168 Perl_is_utf8_char(pTHX_ U8 *s)
174 if (UTF8_IS_INVARIANT(u))
177 if (!UTF8_IS_START(u))
182 if (len < 2 || !UTF8_IS_CONTINUATION(s[1]))
187 /* The initial value is dubious */
191 if (!UTF8_IS_CONTINUATION(*s))
193 uv = UTF8_ACCUMULATE(uv, *s);
200 if (UNISKIP(uv) < len)
207 =for apidoc A|bool|is_utf8_string|U8 *s|STRLEN len
209 Returns true if first C<len> bytes of the given string form a valid UTF8
210 string, false otherwise. Note that 'a valid UTF8 string' does not mean
211 'a string that contains UTF8' because a valid ASCII string is a valid
218 Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len)
225 len = strlen((char *)s);
241 =for apidoc A|UV|utf8n_to_uvuni|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
243 Bottom level UTF-8 decode routine.
244 Returns the unicode code point value of the first character in the string C<s>
245 which is assumed to be in UTF8 encoding and no longer than C<curlen>;
246 C<retlen> will be set to the length, in bytes, of that character.
248 If C<s> does not point to a well-formed UTF8 character, the behaviour
249 is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY,
250 it is assumed that the caller will raise a warning, and this function
251 will silently just set C<retlen> to C<-1> and return zero. If the
252 C<flags> does not contain UTF8_CHECK_ONLY, warnings about
253 malformations will be given, C<retlen> will be set to the expected
254 length of the UTF-8 character in bytes, and zero will be returned.
256 The C<flags> can also contain various flags to allow deviations from
257 the strict UTF-8 encoding (see F<utf8.h>).
259 Most code should use utf8_to_uvchr() rather than call this directly.
264 Perl_utf8n_to_uvuni(pTHX_ U8* s, STRLEN curlen, STRLEN* retlen, U32 flags)
268 bool dowarn = ckWARN_d(WARN_UTF8);
269 STRLEN expectlen = 0;
272 /* This list is a superset of the UTF8_ALLOW_XXX. */
274 #define UTF8_WARN_EMPTY 1
275 #define UTF8_WARN_CONTINUATION 2
276 #define UTF8_WARN_NON_CONTINUATION 3
277 #define UTF8_WARN_FE_FF 4
278 #define UTF8_WARN_SHORT 5
279 #define UTF8_WARN_OVERFLOW 6
280 #define UTF8_WARN_SURROGATE 7
281 #define UTF8_WARN_BOM 8
282 #define UTF8_WARN_LONG 9
283 #define UTF8_WARN_FFFF 10
286 !(flags & UTF8_ALLOW_EMPTY)) {
287 warning = UTF8_WARN_EMPTY;
291 if (UTF8_IS_INVARIANT(uv)) {
294 return (UV) (NATIVE_TO_UTF(*s));
297 if (UTF8_IS_CONTINUATION(uv) &&
298 !(flags & UTF8_ALLOW_CONTINUATION)) {
299 warning = UTF8_WARN_CONTINUATION;
303 if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) &&
304 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
305 warning = UTF8_WARN_NON_CONTINUATION;
310 uv = NATIVE_TO_UTF(uv);
312 if ((uv == 0xfe || uv == 0xff) &&
313 !(flags & UTF8_ALLOW_FE_FF)) {
314 warning = UTF8_WARN_FE_FF;
319 if (!(uv & 0x20)) { len = 2; uv &= 0x1f; }
320 else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; }
321 else if (!(uv & 0x08)) { len = 4; uv &= 0x07; }
322 else if (!(uv & 0x04)) { len = 5; uv &= 0x03; }
324 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
325 else { len = 7; uv &= 0x01; }
327 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
328 else if (!(uv & 0x01)) { len = 7; uv = 0; }
329 else { len = 13; uv = 0; } /* whoa! */
337 if ((curlen < expectlen) &&
338 !(flags & UTF8_ALLOW_SHORT)) {
339 warning = UTF8_WARN_SHORT;
348 if (!UTF8_IS_CONTINUATION(*s) &&
349 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
351 warning = UTF8_WARN_NON_CONTINUATION;
355 uv = UTF8_ACCUMULATE(uv, *s);
357 /* These cannot be allowed. */
359 if (!(flags & UTF8_ALLOW_LONG)) {
360 warning = UTF8_WARN_LONG;
364 else { /* uv < ouv */
365 /* This cannot be allowed. */
366 warning = UTF8_WARN_OVERFLOW;
374 if (UNICODE_IS_SURROGATE(uv) &&
375 !(flags & UTF8_ALLOW_SURROGATE)) {
376 warning = UTF8_WARN_SURROGATE;
378 } else if (UNICODE_IS_BYTE_ORDER_MARK(uv) &&
379 !(flags & UTF8_ALLOW_BOM)) {
380 warning = UTF8_WARN_BOM;
382 } else if ((expectlen > UNISKIP(uv)) &&
383 !(flags & UTF8_ALLOW_LONG)) {
384 warning = UTF8_WARN_LONG;
386 } else if (UNICODE_IS_ILLEGAL(uv) &&
387 !(flags & UTF8_ALLOW_FFFF)) {
388 warning = UTF8_WARN_FFFF;
396 if (flags & UTF8_CHECK_ONLY) {
403 SV* sv = sv_2mortal(newSVpv("Malformed UTF-8 character ", 0));
406 case 0: /* Intentionally empty. */ break;
407 case UTF8_WARN_EMPTY:
408 Perl_sv_catpvf(aTHX_ sv, "(empty string)");
410 case UTF8_WARN_CONTINUATION:
411 Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf")", uv);
413 case UTF8_WARN_NON_CONTINUATION:
414 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf" after start byte 0x%02"UVxf")",
417 case UTF8_WARN_FE_FF:
418 Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv);
420 case UTF8_WARN_SHORT:
421 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d)",
422 curlen, curlen == 1 ? "" : "s", expectlen);
424 case UTF8_WARN_OVERFLOW:
425 Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x)",
428 case UTF8_WARN_SURROGATE:
429 Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv);
432 Perl_sv_catpvf(aTHX_ sv, "(byte order mark 0x%04"UVxf")", uv);
435 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d)",
436 expectlen, expectlen == 1 ? "": "s", UNISKIP(uv));
439 Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv);
442 Perl_sv_catpvf(aTHX_ sv, "(unknown reason)");
450 Perl_warner(aTHX_ WARN_UTF8,
451 "%s in %s", s, PL_op_desc[PL_op->op_type]);
453 Perl_warner(aTHX_ WARN_UTF8, "%s", s);
458 *retlen = expectlen ? expectlen : len;
464 =for apidoc A|U8* s|utf8n_to_uvchr|STRLEN curlen, STRLEN *retlen, U32 flags
466 Returns the native character value of the first character in the string C<s>
467 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
468 length, in bytes, of that character.
470 Allows length and flags to be passed to low level routine.
476 Perl_utf8n_to_uvchr(pTHX_ U8* s, STRLEN curlen, STRLEN* retlen, U32 flags)
478 UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags);
479 return UNI_TO_NATIVE(uv);
483 =for apidoc A|U8* s|utf8_to_uvchr|STRLEN *retlen
485 Returns the native character value of the first character in the string C<s>
486 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
487 length, in bytes, of that character.
489 If C<s> does not point to a well-formed UTF8 character, zero is
490 returned and retlen is set, if possible, to -1.
496 Perl_utf8_to_uvchr(pTHX_ U8* s, STRLEN* retlen)
498 return Perl_utf8n_to_uvchr(aTHX_ s, UTF8_MAXLEN, retlen, 0);
502 =for apidoc A|U8* s|utf8_to_uvuni|STRLEN *retlen
504 Returns the Unicode code point of the first character in the string C<s>
505 which is assumed to be in UTF8 encoding; C<retlen> will be set to the
506 length, in bytes, of that character.
508 This function should only be used when returned UV is considered
509 an index into the Unicode semantic tables (e.g. swashes).
511 If C<s> does not point to a well-formed UTF8 character, zero is
512 returned and retlen is set, if possible, to -1.
518 Perl_utf8_to_uvuni(pTHX_ U8* s, STRLEN* retlen)
520 /* Call the low level routine asking for checks */
521 return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXLEN, retlen, 0);
525 =for apidoc A|STRLEN|utf8_length|U8* s|U8 *e
527 Return the length of the UTF-8 char encoded string C<s> in characters.
528 Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end
529 up past C<e>, croaks.
535 Perl_utf8_length(pTHX_ U8* s, U8* e)
539 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
540 * the bitops (especially ~) can create illegal UTF-8.
541 * In other words: in Perl UTF-8 is not just for Unicode. */
544 Perl_croak(aTHX_ "panic: utf8_length: unexpected end");
549 Perl_croak(aTHX_ "panic: utf8_length: s=%p (%02X) e=%p l=%d - unaligned end",s,*s,e,t);
558 =for apidoc A|IV|utf8_distance|U8 *a|U8 *b
560 Returns the number of UTF8 characters between the UTF-8 pointers C<a>
563 WARNING: use only if you *know* that the pointers point inside the
569 Perl_utf8_distance(pTHX_ U8 *a, U8 *b)
573 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
574 * the bitops (especially ~) can create illegal UTF-8.
575 * In other words: in Perl UTF-8 is not just for Unicode. */
582 Perl_croak(aTHX_ "panic: utf8_distance: unaligned end");
592 Perl_croak(aTHX_ "panic: utf8_distance: unaligned end");
602 =for apidoc A|U8*|utf8_hop|U8 *s|I32 off
604 Return the UTF-8 pointer C<s> displaced by C<off> characters, either
607 WARNING: do not use the following unless you *know* C<off> is within
608 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
609 on the first byte of character or just after the last byte of a character.
614 Perl_utf8_hop(pTHX_ U8 *s, I32 off)
616 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
617 * the bitops (especially ~) can create illegal UTF-8.
618 * In other words: in Perl UTF-8 is not just for Unicode. */
627 while (UTF8_IS_CONTINUATION(*s))
635 =for apidoc A|U8 *|utf8_to_bytes|U8 *s|STRLEN *len
637 Converts a string C<s> of length C<len> from UTF8 into byte encoding.
638 Unlike C<bytes_to_utf8>, this over-writes the original string, and
639 updates len to contain the new length.
640 Returns zero on failure, setting C<len> to -1.
646 Perl_utf8_to_bytes(pTHX_ U8* s, STRLEN *len)
652 /* ensure valid UTF8 and chars < 256 before updating string */
653 for (send = s + *len; s < send; ) {
656 if (!UTF8_IS_INVARIANT(c) &&
657 (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send)
658 || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) {
667 *d++ = (U8)utf8_to_uvchr(s, &ulen);
676 =for apidoc A|U8 *|bytes_from_utf8|U8 *s|STRLEN *len|bool *is_utf8
678 Converts a string C<s> of length C<len> from UTF8 into byte encoding.
679 Unlike <utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to
680 the newly-created string, and updates C<len> to contain the new
681 length. Returns the original string if no conversion occurs, C<len>
682 is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to
683 0 if C<s> is converted or contains all 7bit characters.
688 Perl_bytes_from_utf8(pTHX_ U8* s, STRLEN *len, bool *is_utf8)
698 /* ensure valid UTF8 and chars < 256 before converting string */
699 for (send = s + *len; s < send;) {
701 if (!UTF8_IS_INVARIANT(c)) {
702 if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send &&
703 (c = *s++) && UTF8_IS_CONTINUATION(c))
712 Newz(801, d, (*len) - count + 1, U8);
713 s = start; start = d;
716 if (!UTF8_IS_INVARIANT(c)) {
717 /* Then it is two-byte encoded */
718 c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++);
719 c = ASCII_TO_NATIVE(c);
729 =for apidoc A|U8 *|bytes_to_utf8|U8 *s|STRLEN *len
731 Converts a string C<s> of length C<len> from ASCII into UTF8 encoding.
732 Returns a pointer to the newly-created string, and sets C<len> to
733 reflect the new length.
739 Perl_bytes_to_utf8(pTHX_ U8* s, STRLEN *len)
746 Newz(801, d, (*len) * 2 + 1, U8);
750 UV uv = NATIVE_TO_ASCII(*s++);
751 if (UNI_IS_INVARIANT(uv))
752 *d++ = UTF_TO_NATIVE(uv);
754 *d++ = UTF8_EIGHT_BIT_HI(uv);
755 *d++ = UTF8_EIGHT_BIT_LO(uv);
764 * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8.
766 * Destination must be pre-extended to 3/2 source. Do not use in-place.
767 * We optimize for native, for obvious reasons. */
770 Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
776 Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen");
781 UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */
788 *d++ = (( uv >> 6) | 0xc0);
789 *d++ = (( uv & 0x3f) | 0x80);
792 if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */
794 if (low < 0xdc00 || low >= 0xdfff)
795 Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
796 uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000;
799 *d++ = (( uv >> 12) | 0xe0);
800 *d++ = (((uv >> 6) & 0x3f) | 0x80);
801 *d++ = (( uv & 0x3f) | 0x80);
805 *d++ = (( uv >> 18) | 0xf0);
806 *d++ = (((uv >> 12) & 0x3f) | 0x80);
807 *d++ = (((uv >> 6) & 0x3f) | 0x80);
808 *d++ = (( uv & 0x3f) | 0x80);
812 *newlen = d - dstart;
816 /* Note: this one is slightly destructive of the source. */
819 Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
822 U8* send = s + bytelen;
829 return utf16_to_utf8(p, d, bytelen, newlen);
832 /* for now these are all defined (inefficiently) in terms of the utf8 versions */
835 Perl_is_uni_alnum(pTHX_ U32 c)
837 U8 tmpbuf[UTF8_MAXLEN+1];
838 uvuni_to_utf8(tmpbuf, (UV)c);
839 return is_utf8_alnum(tmpbuf);
843 Perl_is_uni_alnumc(pTHX_ U32 c)
845 U8 tmpbuf[UTF8_MAXLEN+1];
846 uvuni_to_utf8(tmpbuf, (UV)c);
847 return is_utf8_alnumc(tmpbuf);
851 Perl_is_uni_idfirst(pTHX_ U32 c)
853 U8 tmpbuf[UTF8_MAXLEN+1];
854 uvuni_to_utf8(tmpbuf, (UV)c);
855 return is_utf8_idfirst(tmpbuf);
859 Perl_is_uni_alpha(pTHX_ U32 c)
861 U8 tmpbuf[UTF8_MAXLEN+1];
862 uvuni_to_utf8(tmpbuf, (UV)c);
863 return is_utf8_alpha(tmpbuf);
867 Perl_is_uni_ascii(pTHX_ U32 c)
869 U8 tmpbuf[UTF8_MAXLEN+1];
870 uvuni_to_utf8(tmpbuf, (UV)c);
871 return is_utf8_ascii(tmpbuf);
875 Perl_is_uni_space(pTHX_ U32 c)
877 U8 tmpbuf[UTF8_MAXLEN+1];
878 uvuni_to_utf8(tmpbuf, (UV)c);
879 return is_utf8_space(tmpbuf);
883 Perl_is_uni_digit(pTHX_ U32 c)
885 U8 tmpbuf[UTF8_MAXLEN+1];
886 uvuni_to_utf8(tmpbuf, (UV)c);
887 return is_utf8_digit(tmpbuf);
891 Perl_is_uni_upper(pTHX_ U32 c)
893 U8 tmpbuf[UTF8_MAXLEN+1];
894 uvuni_to_utf8(tmpbuf, (UV)c);
895 return is_utf8_upper(tmpbuf);
899 Perl_is_uni_lower(pTHX_ U32 c)
901 U8 tmpbuf[UTF8_MAXLEN+1];
902 uvuni_to_utf8(tmpbuf, (UV)c);
903 return is_utf8_lower(tmpbuf);
907 Perl_is_uni_cntrl(pTHX_ U32 c)
909 U8 tmpbuf[UTF8_MAXLEN+1];
910 uvuni_to_utf8(tmpbuf, (UV)c);
911 return is_utf8_cntrl(tmpbuf);
915 Perl_is_uni_graph(pTHX_ U32 c)
917 U8 tmpbuf[UTF8_MAXLEN+1];
918 uvuni_to_utf8(tmpbuf, (UV)c);
919 return is_utf8_graph(tmpbuf);
923 Perl_is_uni_print(pTHX_ U32 c)
925 U8 tmpbuf[UTF8_MAXLEN+1];
926 uvuni_to_utf8(tmpbuf, (UV)c);
927 return is_utf8_print(tmpbuf);
931 Perl_is_uni_punct(pTHX_ U32 c)
933 U8 tmpbuf[UTF8_MAXLEN+1];
934 uvuni_to_utf8(tmpbuf, (UV)c);
935 return is_utf8_punct(tmpbuf);
939 Perl_is_uni_xdigit(pTHX_ U32 c)
941 U8 tmpbuf[UTF8_MAXLEN+1];
942 uvuni_to_utf8(tmpbuf, (UV)c);
943 return is_utf8_xdigit(tmpbuf);
947 Perl_to_uni_upper(pTHX_ U32 c)
949 U8 tmpbuf[UTF8_MAXLEN+1];
950 uvuni_to_utf8(tmpbuf, (UV)c);
951 return to_utf8_upper(tmpbuf);
955 Perl_to_uni_title(pTHX_ U32 c)
957 U8 tmpbuf[UTF8_MAXLEN+1];
958 uvuni_to_utf8(tmpbuf, (UV)c);
959 return to_utf8_title(tmpbuf);
963 Perl_to_uni_lower(pTHX_ U32 c)
965 U8 tmpbuf[UTF8_MAXLEN+1];
966 uvuni_to_utf8(tmpbuf, (UV)c);
967 return to_utf8_lower(tmpbuf);
970 /* for now these all assume no locale info available for Unicode > 255 */
973 Perl_is_uni_alnum_lc(pTHX_ U32 c)
975 return is_uni_alnum(c); /* XXX no locale support yet */
979 Perl_is_uni_alnumc_lc(pTHX_ U32 c)
981 return is_uni_alnumc(c); /* XXX no locale support yet */
985 Perl_is_uni_idfirst_lc(pTHX_ U32 c)
987 return is_uni_idfirst(c); /* XXX no locale support yet */
991 Perl_is_uni_alpha_lc(pTHX_ U32 c)
993 return is_uni_alpha(c); /* XXX no locale support yet */
997 Perl_is_uni_ascii_lc(pTHX_ U32 c)
999 return is_uni_ascii(c); /* XXX no locale support yet */
1003 Perl_is_uni_space_lc(pTHX_ U32 c)
1005 return is_uni_space(c); /* XXX no locale support yet */
1009 Perl_is_uni_digit_lc(pTHX_ U32 c)
1011 return is_uni_digit(c); /* XXX no locale support yet */
1015 Perl_is_uni_upper_lc(pTHX_ U32 c)
1017 return is_uni_upper(c); /* XXX no locale support yet */
1021 Perl_is_uni_lower_lc(pTHX_ U32 c)
1023 return is_uni_lower(c); /* XXX no locale support yet */
1027 Perl_is_uni_cntrl_lc(pTHX_ U32 c)
1029 return is_uni_cntrl(c); /* XXX no locale support yet */
1033 Perl_is_uni_graph_lc(pTHX_ U32 c)
1035 return is_uni_graph(c); /* XXX no locale support yet */
1039 Perl_is_uni_print_lc(pTHX_ U32 c)
1041 return is_uni_print(c); /* XXX no locale support yet */
1045 Perl_is_uni_punct_lc(pTHX_ U32 c)
1047 return is_uni_punct(c); /* XXX no locale support yet */
1051 Perl_is_uni_xdigit_lc(pTHX_ U32 c)
1053 return is_uni_xdigit(c); /* XXX no locale support yet */
1057 Perl_to_uni_upper_lc(pTHX_ U32 c)
1059 return to_uni_upper(c); /* XXX no locale support yet */
1063 Perl_to_uni_title_lc(pTHX_ U32 c)
1065 return to_uni_title(c); /* XXX no locale support yet */
1069 Perl_to_uni_lower_lc(pTHX_ U32 c)
1071 return to_uni_lower(c); /* XXX no locale support yet */
1075 Perl_is_utf8_alnum(pTHX_ U8 *p)
1077 if (!is_utf8_char(p))
1080 /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true
1081 * descendant of isalnum(3), in other words, it doesn't
1082 * contain the '_'. --jhi */
1083 PL_utf8_alnum = swash_init("utf8", "IsWord", &PL_sv_undef, 0, 0);
1084 return swash_fetch(PL_utf8_alnum, p);
1085 /* return *p == '_' || is_utf8_alpha(p) || is_utf8_digit(p); */
1086 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1088 PL_utf8_alnum = swash_init("utf8", "",
1089 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1090 return swash_fetch(PL_utf8_alnum, p);
1095 Perl_is_utf8_alnumc(pTHX_ U8 *p)
1097 if (!is_utf8_char(p))
1100 PL_utf8_alnum = swash_init("utf8", "IsAlnumC", &PL_sv_undef, 0, 0);
1101 return swash_fetch(PL_utf8_alnum, p);
1102 /* return is_utf8_alpha(p) || is_utf8_digit(p); */
1103 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1105 PL_utf8_alnum = swash_init("utf8", "",
1106 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1107 return swash_fetch(PL_utf8_alnum, p);
1112 Perl_is_utf8_idfirst(pTHX_ U8 *p)
1114 return *p == '_' || is_utf8_alpha(p);
1118 Perl_is_utf8_alpha(pTHX_ U8 *p)
1120 if (!is_utf8_char(p))
1123 PL_utf8_alpha = swash_init("utf8", "IsAlpha", &PL_sv_undef, 0, 0);
1124 return swash_fetch(PL_utf8_alpha, p);
1128 Perl_is_utf8_ascii(pTHX_ U8 *p)
1130 if (!is_utf8_char(p))
1133 PL_utf8_ascii = swash_init("utf8", "IsAscii", &PL_sv_undef, 0, 0);
1134 return swash_fetch(PL_utf8_ascii, p);
1138 Perl_is_utf8_space(pTHX_ U8 *p)
1140 if (!is_utf8_char(p))
1143 PL_utf8_space = swash_init("utf8", "IsSpacePerl", &PL_sv_undef, 0, 0);
1144 return swash_fetch(PL_utf8_space, p);
1148 Perl_is_utf8_digit(pTHX_ U8 *p)
1150 if (!is_utf8_char(p))
1153 PL_utf8_digit = swash_init("utf8", "IsDigit", &PL_sv_undef, 0, 0);
1154 return swash_fetch(PL_utf8_digit, p);
1158 Perl_is_utf8_upper(pTHX_ U8 *p)
1160 if (!is_utf8_char(p))
1163 PL_utf8_upper = swash_init("utf8", "IsUpper", &PL_sv_undef, 0, 0);
1164 return swash_fetch(PL_utf8_upper, p);
1168 Perl_is_utf8_lower(pTHX_ U8 *p)
1170 if (!is_utf8_char(p))
1173 PL_utf8_lower = swash_init("utf8", "IsLower", &PL_sv_undef, 0, 0);
1174 return swash_fetch(PL_utf8_lower, p);
1178 Perl_is_utf8_cntrl(pTHX_ U8 *p)
1180 if (!is_utf8_char(p))
1183 PL_utf8_cntrl = swash_init("utf8", "IsCntrl", &PL_sv_undef, 0, 0);
1184 return swash_fetch(PL_utf8_cntrl, p);
1188 Perl_is_utf8_graph(pTHX_ U8 *p)
1190 if (!is_utf8_char(p))
1193 PL_utf8_graph = swash_init("utf8", "IsGraph", &PL_sv_undef, 0, 0);
1194 return swash_fetch(PL_utf8_graph, p);
1198 Perl_is_utf8_print(pTHX_ U8 *p)
1200 if (!is_utf8_char(p))
1203 PL_utf8_print = swash_init("utf8", "IsPrint", &PL_sv_undef, 0, 0);
1204 return swash_fetch(PL_utf8_print, p);
1208 Perl_is_utf8_punct(pTHX_ U8 *p)
1210 if (!is_utf8_char(p))
1213 PL_utf8_punct = swash_init("utf8", "IsPunct", &PL_sv_undef, 0, 0);
1214 return swash_fetch(PL_utf8_punct, p);
1218 Perl_is_utf8_xdigit(pTHX_ U8 *p)
1220 if (!is_utf8_char(p))
1222 if (!PL_utf8_xdigit)
1223 PL_utf8_xdigit = swash_init("utf8", "IsXDigit", &PL_sv_undef, 0, 0);
1224 return swash_fetch(PL_utf8_xdigit, p);
1228 Perl_is_utf8_mark(pTHX_ U8 *p)
1230 if (!is_utf8_char(p))
1233 PL_utf8_mark = swash_init("utf8", "IsM", &PL_sv_undef, 0, 0);
1234 return swash_fetch(PL_utf8_mark, p);
1238 Perl_to_utf8_upper(pTHX_ U8 *p)
1242 if (!PL_utf8_toupper)
1243 PL_utf8_toupper = swash_init("utf8", "ToUpper", &PL_sv_undef, 4, 0);
1244 uv = swash_fetch(PL_utf8_toupper, p);
1245 return uv ? UNI_TO_NATIVE(uv) : utf8_to_uvchr(p,0);
1249 Perl_to_utf8_title(pTHX_ U8 *p)
1253 if (!PL_utf8_totitle)
1254 PL_utf8_totitle = swash_init("utf8", "ToTitle", &PL_sv_undef, 4, 0);
1255 uv = swash_fetch(PL_utf8_totitle, p);
1256 return uv ? UNI_TO_NATIVE(uv) : utf8_to_uvchr(p,0);
1260 Perl_to_utf8_lower(pTHX_ U8 *p)
1264 if (!PL_utf8_tolower)
1265 PL_utf8_tolower = swash_init("utf8", "ToLower", &PL_sv_undef, 4, 0);
1266 uv = swash_fetch(PL_utf8_tolower, p);
1267 return uv ? UNI_TO_NATIVE(uv) : utf8_to_uvchr(p,0);
1270 /* a "swash" is a swatch hash */
1273 Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none)
1276 SV* tokenbufsv = sv_2mortal(NEWSV(0,0));
1278 HV *stash = gv_stashpvn(pkg, strlen(pkg), FALSE);
1280 if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */
1282 Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpv(pkg,0), Nullsv);
1286 PUSHSTACKi(PERLSI_MAGIC);
1289 PUSHs(sv_2mortal(newSVpvn(pkg, strlen(pkg))));
1290 PUSHs(sv_2mortal(newSVpvn(name, strlen(name))));
1292 PUSHs(sv_2mortal(newSViv(minbits)));
1293 PUSHs(sv_2mortal(newSViv(none)));
1299 if (PL_curcop == &PL_compiling)
1300 /* XXX ought to be handled by lex_start */
1301 sv_setpv(tokenbufsv, PL_tokenbuf);
1302 if (call_method("SWASHNEW", G_SCALAR))
1303 retval = newSVsv(*PL_stack_sp--);
1305 retval = &PL_sv_undef;
1308 if (PL_curcop == &PL_compiling) {
1310 char* pv = SvPV(tokenbufsv, len);
1312 Copy(pv, PL_tokenbuf, len+1, char);
1313 PL_curcop->op_private = PL_hints;
1315 if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV)
1316 Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref");
1321 Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr)
1323 HV* hv = (HV*)SvRV(sv);
1324 U32 klen = UTF8SKIP(ptr) - 1;
1325 U32 off = ptr[klen] & 127; /* NB: 64 bit always 0 when len > 1 */
1327 STRLEN needents = (klen ? 64 : 128);
1333 * This single-entry cache saves about 1/3 of the utf8 overhead in test
1334 * suite. (That is, only 7-8% overall over just a hash cache. Still,
1335 * it's nothing to sniff at.) Pity we usually come through at least
1336 * two function calls to get here...
1338 * NB: this code assumes that swatches are never modified, once generated!
1341 if (hv == PL_last_swash_hv &&
1342 klen == PL_last_swash_klen &&
1343 (!klen || memEQ((char *)ptr,(char *)PL_last_swash_key,klen)) )
1345 tmps = PL_last_swash_tmps;
1346 slen = PL_last_swash_slen;
1349 /* Try our second-level swatch cache, kept in a hash. */
1350 SV** svp = hv_fetch(hv, (char*)ptr, klen, FALSE);
1352 /* If not cached, generate it via utf8::SWASHGET */
1353 if (!svp || !SvPOK(*svp) || !(tmps = (U8*)SvPV(*svp, slen))) {
1358 PUSHSTACKi(PERLSI_MAGIC);
1362 /* We call utf8_to_uni as we want and index into Unicode tables,
1363 not a native character number.
1365 PUSHs(sv_2mortal(newSViv(utf8_to_uvuni(ptr, 0) & ~(needents - 1))));
1366 PUSHs(sv_2mortal(newSViv(needents)));
1368 if (call_method("SWASHGET", G_SCALAR))
1369 retval = newSVsv(*PL_stack_sp--);
1371 retval = &PL_sv_undef;
1375 if (PL_curcop == &PL_compiling)
1376 PL_curcop->op_private = PL_hints;
1378 svp = hv_store(hv, (char*)ptr, klen, retval, 0);
1380 if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || slen < 8)
1381 Perl_croak(aTHX_ "SWASHGET didn't return result of proper length");
1384 PL_last_swash_hv = hv;
1385 PL_last_swash_klen = klen;
1386 PL_last_swash_tmps = tmps;
1387 PL_last_swash_slen = slen;
1389 Copy(ptr, PL_last_swash_key, klen, U8);
1392 switch ((int)((slen << 3) / needents)) {
1394 bit = 1 << (off & 7);
1396 return (tmps[off] & bit) != 0;
1401 return (tmps[off] << 8) + tmps[off + 1] ;
1404 return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ;
1406 Perl_croak(aTHX_ "panic: swash_fetch");