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.
54 =for apidoc uvuni_to_utf8_flags
56 Adds the UTF-8 representation of the Unicode codepoint C<uv> to the end
57 of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free
58 bytes available. The return value is the pointer to the byte after the
59 end of the new character. In other words,
61 d = uvuni_to_utf8_flags(d, uv, flags);
65 d = uvuni_to_utf8(d, uv);
67 (which is equivalent to)
69 d = uvuni_to_utf8_flags(d, uv, 0);
71 is the recommended Unicode-aware way of saying
79 Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
81 PERL_ARGS_ASSERT_UVUNI_TO_UTF8_FLAGS;
83 if (ckWARN(WARN_UTF8)) {
84 if (UNICODE_IS_SURROGATE(uv) &&
85 !(flags & UNICODE_ALLOW_SURROGATE))
86 Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv);
88 ((uv >= 0xFDD0 && uv <= 0xFDEF &&
89 !(flags & UNICODE_ALLOW_FDD0))
91 ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */
92 !(flags & UNICODE_ALLOW_FFFF))) &&
93 /* UNICODE_ALLOW_SUPER includes
94 * FFFEs and FFFFs beyond 0x10FFFF. */
95 ((uv <= PERL_UNICODE_MAX) ||
96 !(flags & UNICODE_ALLOW_SUPER))
98 Perl_warner(aTHX_ packWARN(WARN_UTF8),
99 "Unicode character 0x%04"UVxf" is illegal", uv);
101 if (UNI_IS_INVARIANT(uv)) {
102 *d++ = (U8)UTF_TO_NATIVE(uv);
107 STRLEN len = UNISKIP(uv);
110 *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK);
111 uv >>= UTF_ACCUMULATION_SHIFT;
113 *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len));
116 #else /* Non loop style */
118 *d++ = (U8)(( uv >> 6) | 0xc0);
119 *d++ = (U8)(( uv & 0x3f) | 0x80);
123 *d++ = (U8)(( uv >> 12) | 0xe0);
124 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
125 *d++ = (U8)(( uv & 0x3f) | 0x80);
129 *d++ = (U8)(( uv >> 18) | 0xf0);
130 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
131 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
132 *d++ = (U8)(( uv & 0x3f) | 0x80);
135 if (uv < 0x4000000) {
136 *d++ = (U8)(( uv >> 24) | 0xf8);
137 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
138 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
139 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
140 *d++ = (U8)(( uv & 0x3f) | 0x80);
143 if (uv < 0x80000000) {
144 *d++ = (U8)(( uv >> 30) | 0xfc);
145 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
146 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
147 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
148 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
149 *d++ = (U8)(( uv & 0x3f) | 0x80);
153 if (uv < UTF8_QUAD_MAX)
156 *d++ = 0xfe; /* Can't match U+FEFF! */
157 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
158 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
159 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
160 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
161 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
162 *d++ = (U8)(( uv & 0x3f) | 0x80);
167 *d++ = 0xff; /* Can't match U+FFFE! */
168 *d++ = 0x80; /* 6 Reserved bits */
169 *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */
170 *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80);
171 *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80);
172 *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80);
173 *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80);
174 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
175 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
176 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
177 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
178 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
179 *d++ = (U8)(( uv & 0x3f) | 0x80);
183 #endif /* Loop style */
188 Tests if some arbitrary number of bytes begins in a valid UTF-8
189 character. Note that an INVARIANT (i.e. ASCII) character is a valid
190 UTF-8 character. The actual number of bytes in the UTF-8 character
191 will be returned if it is valid, otherwise 0.
193 This is the "slow" version as opposed to the "fast" version which is
194 the "unrolled" IS_UTF8_CHAR(). E.g. for t/uni/class.t the speed
195 difference is a factor of 2 to 3. For lengths (UTF8SKIP(s)) of four
196 or less you should use the IS_UTF8_CHAR(), for lengths of five or more
197 you should use the _slow(). In practice this means that the _slow()
198 will be used very rarely, since the maximum Unicode code point (as of
199 Unicode 4.1) is U+10FFFF, which encodes in UTF-8 to four bytes. Only
200 the "Perl extended UTF-8" (the infamous 'v-strings') will encode into
205 S_is_utf8_char_slow(const U8 *s, const STRLEN len)
211 PERL_ARGS_ASSERT_IS_UTF8_CHAR_SLOW;
213 if (UTF8_IS_INVARIANT(u))
216 if (!UTF8_IS_START(u))
219 if (len < 2 || !UTF8_IS_CONTINUATION(s[1]))
225 u = NATIVE_TO_UTF(u);
227 u &= UTF_START_MASK(len);
231 if (!UTF8_IS_CONTINUATION(*s))
233 uv = UTF8_ACCUMULATE(uv, *s);
240 if ((STRLEN)UNISKIP(uv) < len)
247 =for apidoc is_utf8_char
249 Tests if some arbitrary number of bytes begins in a valid UTF-8
250 character. Note that an INVARIANT (i.e. ASCII on non-EBCDIC machines)
251 character is a valid UTF-8 character. The actual number of bytes in the UTF-8
252 character will be returned if it is valid, otherwise 0.
256 Perl_is_utf8_char(pTHX_ const U8 *s)
258 const STRLEN len = UTF8SKIP(s);
260 PERL_ARGS_ASSERT_IS_UTF8_CHAR;
263 if (IS_UTF8_CHAR_FAST(len))
264 return IS_UTF8_CHAR(s, len) ? len : 0;
265 #endif /* #ifdef IS_UTF8_CHAR */
266 return is_utf8_char_slow(s, len);
270 =for apidoc is_utf8_string
272 Returns true if first C<len> bytes of the given string form a valid
273 UTF-8 string, false otherwise. Note that 'a valid UTF-8 string' does
274 not mean 'a string that contains code points above 0x7F encoded in UTF-8'
275 because a valid ASCII string is a valid UTF-8 string.
277 See also is_utf8_string_loclen() and is_utf8_string_loc().
283 Perl_is_utf8_string(pTHX_ const U8 *s, STRLEN len)
285 const U8* const send = s + (len ? len : strlen((const char *)s));
288 PERL_ARGS_ASSERT_IS_UTF8_STRING;
293 /* Inline the easy bits of is_utf8_char() here for speed... */
294 if (UTF8_IS_INVARIANT(*x))
296 else if (!UTF8_IS_START(*x))
299 /* ... and call is_utf8_char() only if really needed. */
302 if (IS_UTF8_CHAR_FAST(c)) {
303 if (!IS_UTF8_CHAR(x, c))
307 c = is_utf8_char_slow(x, c);
310 #endif /* #ifdef IS_UTF8_CHAR */
325 Implemented as a macro in utf8.h
327 =for apidoc is_utf8_string_loc
329 Like is_utf8_string() but stores the location of the failure (in the
330 case of "utf8ness failure") or the location s+len (in the case of
331 "utf8ness success") in the C<ep>.
333 See also is_utf8_string_loclen() and is_utf8_string().
335 =for apidoc is_utf8_string_loclen
337 Like is_utf8_string() but stores the location of the failure (in the
338 case of "utf8ness failure") or the location s+len (in the case of
339 "utf8ness success") in the C<ep>, and the number of UTF-8
340 encoded characters in the C<el>.
342 See also is_utf8_string_loc() and is_utf8_string().
348 Perl_is_utf8_string_loclen(pTHX_ const U8 *s, STRLEN len, const U8 **ep, STRLEN *el)
350 const U8* const send = s + (len ? len : strlen((const char *)s));
355 PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN;
359 /* Inline the easy bits of is_utf8_char() here for speed... */
360 if (UTF8_IS_INVARIANT(*x))
362 else if (!UTF8_IS_START(*x))
365 /* ... and call is_utf8_char() only if really needed. */
368 if (IS_UTF8_CHAR_FAST(c)) {
369 if (!IS_UTF8_CHAR(x, c))
372 c = is_utf8_char_slow(x, c);
375 #endif /* #ifdef IS_UTF8_CHAR */
394 =for apidoc utf8n_to_uvuni
396 Bottom level UTF-8 decode routine.
397 Returns the Unicode code point value of the first character in the string C<s>
398 which is assumed to be in UTF-8 encoding and no longer than C<curlen>;
399 C<retlen> will be set to the length, in bytes, of that character.
401 If C<s> does not point to a well-formed UTF-8 character, the behaviour
402 is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY,
403 it is assumed that the caller will raise a warning, and this function
404 will silently just set C<retlen> to C<-1> and return zero. If the
405 C<flags> does not contain UTF8_CHECK_ONLY, warnings about
406 malformations will be given, C<retlen> will be set to the expected
407 length of the UTF-8 character in bytes, and zero will be returned.
409 The C<flags> can also contain various flags to allow deviations from
410 the strict UTF-8 encoding (see F<utf8.h>).
412 Most code should use utf8_to_uvchr() rather than call this directly.
418 Perl_utf8n_to_uvuni(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
421 const U8 * const s0 = s;
424 const bool dowarn = ckWARN_d(WARN_UTF8);
425 const UV startbyte = *s;
426 STRLEN expectlen = 0;
429 PERL_ARGS_ASSERT_UTF8N_TO_UVUNI;
431 /* This list is a superset of the UTF8_ALLOW_XXX. */
433 #define UTF8_WARN_EMPTY 1
434 #define UTF8_WARN_CONTINUATION 2
435 #define UTF8_WARN_NON_CONTINUATION 3
436 #define UTF8_WARN_FE_FF 4
437 #define UTF8_WARN_SHORT 5
438 #define UTF8_WARN_OVERFLOW 6
439 #define UTF8_WARN_SURROGATE 7
440 #define UTF8_WARN_LONG 8
441 #define UTF8_WARN_FFFF 9 /* Also FFFE. */
444 !(flags & UTF8_ALLOW_EMPTY)) {
445 warning = UTF8_WARN_EMPTY;
449 if (UTF8_IS_INVARIANT(uv)) {
452 return (UV) (NATIVE_TO_UTF(*s));
455 if (UTF8_IS_CONTINUATION(uv) &&
456 !(flags & UTF8_ALLOW_CONTINUATION)) {
457 warning = UTF8_WARN_CONTINUATION;
461 if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) &&
462 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
463 warning = UTF8_WARN_NON_CONTINUATION;
468 uv = NATIVE_TO_UTF(uv);
470 if ((uv == 0xfe || uv == 0xff) &&
471 !(flags & UTF8_ALLOW_FE_FF)) {
472 warning = UTF8_WARN_FE_FF;
477 if (!(uv & 0x20)) { len = 2; uv &= 0x1f; }
478 else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; }
479 else if (!(uv & 0x08)) { len = 4; uv &= 0x07; }
480 else if (!(uv & 0x04)) { len = 5; uv &= 0x03; }
482 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
483 else { len = 7; uv &= 0x01; }
485 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
486 else if (!(uv & 0x01)) { len = 7; uv = 0; }
487 else { len = 13; uv = 0; } /* whoa! */
495 if ((curlen < expectlen) &&
496 !(flags & UTF8_ALLOW_SHORT)) {
497 warning = UTF8_WARN_SHORT;
506 if (!UTF8_IS_CONTINUATION(*s) &&
507 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
509 warning = UTF8_WARN_NON_CONTINUATION;
513 uv = UTF8_ACCUMULATE(uv, *s);
515 /* These cannot be allowed. */
517 if (expectlen != 13 && !(flags & UTF8_ALLOW_LONG)) {
518 warning = UTF8_WARN_LONG;
522 else { /* uv < ouv */
523 /* This cannot be allowed. */
524 warning = UTF8_WARN_OVERFLOW;
532 if (UNICODE_IS_SURROGATE(uv) &&
533 !(flags & UTF8_ALLOW_SURROGATE)) {
534 warning = UTF8_WARN_SURROGATE;
536 } else if ((expectlen > (STRLEN)UNISKIP(uv)) &&
537 !(flags & UTF8_ALLOW_LONG)) {
538 warning = UTF8_WARN_LONG;
540 } else if (UNICODE_IS_ILLEGAL(uv) &&
541 !(flags & UTF8_ALLOW_FFFF)) {
542 warning = UTF8_WARN_FFFF;
550 if (flags & UTF8_CHECK_ONLY) {
552 *retlen = ((STRLEN) -1);
557 SV* const sv = newSVpvs_flags("Malformed UTF-8 character ", SVs_TEMP);
560 case 0: /* Intentionally empty. */ break;
561 case UTF8_WARN_EMPTY:
562 sv_catpvs(sv, "(empty string)");
564 case UTF8_WARN_CONTINUATION:
565 Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv);
567 case UTF8_WARN_NON_CONTINUATION:
569 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")",
570 (UV)s[1], startbyte);
572 const int len = (int)(s-s0);
573 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)",
574 (UV)s[1], len, len > 1 ? "s" : "", startbyte, (int)expectlen);
578 case UTF8_WARN_FE_FF:
579 Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv);
581 case UTF8_WARN_SHORT:
582 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
583 (int)curlen, curlen == 1 ? "" : "s", (int)expectlen, startbyte);
584 expectlen = curlen; /* distance for caller to skip */
586 case UTF8_WARN_OVERFLOW:
587 Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")",
590 case UTF8_WARN_SURROGATE:
591 Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv);
594 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
595 (int)expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte);
598 Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv);
601 sv_catpvs(sv, "(unknown reason)");
606 const char * const s = SvPVX_const(sv);
609 Perl_warner(aTHX_ packWARN(WARN_UTF8),
610 "%s in %s", s, OP_DESC(PL_op));
612 Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s);
617 *retlen = expectlen ? expectlen : len;
623 =for apidoc utf8_to_uvchr
625 Returns the native character value of the first character in the string C<s>
626 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
627 length, in bytes, of that character.
629 If C<s> does not point to a well-formed UTF-8 character, zero is
630 returned and retlen is set, if possible, to -1.
636 Perl_utf8_to_uvchr(pTHX_ const U8 *s, STRLEN *retlen)
638 PERL_ARGS_ASSERT_UTF8_TO_UVCHR;
640 return utf8n_to_uvchr(s, UTF8_MAXBYTES, retlen,
641 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
645 =for apidoc utf8_to_uvuni
647 Returns the Unicode code point of the first character in the string C<s>
648 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
649 length, in bytes, of that character.
651 This function should only be used when the returned UV is considered
652 an index into the Unicode semantic tables (e.g. swashes).
654 If C<s> does not point to a well-formed UTF-8 character, zero is
655 returned and retlen is set, if possible, to -1.
661 Perl_utf8_to_uvuni(pTHX_ const U8 *s, STRLEN *retlen)
663 PERL_ARGS_ASSERT_UTF8_TO_UVUNI;
665 /* Call the low level routine asking for checks */
666 return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXBYTES, retlen,
667 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
671 =for apidoc utf8_length
673 Return the length of the UTF-8 char encoded string C<s> in characters.
674 Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end
675 up past C<e>, croaks.
681 Perl_utf8_length(pTHX_ const U8 *s, const U8 *e)
686 PERL_ARGS_ASSERT_UTF8_LENGTH;
688 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
689 * the bitops (especially ~) can create illegal UTF-8.
690 * In other words: in Perl UTF-8 is not just for Unicode. */
693 goto warn_and_return;
695 if (!UTF8_IS_INVARIANT(*s))
705 if (ckWARN_d(WARN_UTF8)) {
707 Perl_warner(aTHX_ packWARN(WARN_UTF8),
708 "%s in %s", unees, OP_DESC(PL_op));
710 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
718 =for apidoc utf8_distance
720 Returns the number of UTF-8 characters between the UTF-8 pointers C<a>
723 WARNING: use only if you *know* that the pointers point inside the
730 Perl_utf8_distance(pTHX_ const U8 *a, const U8 *b)
732 PERL_ARGS_ASSERT_UTF8_DISTANCE;
734 return (a < b) ? -1 * (IV) utf8_length(a, b) : (IV) utf8_length(b, a);
740 Return the UTF-8 pointer C<s> displaced by C<off> characters, either
743 WARNING: do not use the following unless you *know* C<off> is within
744 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
745 on the first byte of character or just after the last byte of a character.
751 Perl_utf8_hop(pTHX_ const U8 *s, I32 off)
753 PERL_ARGS_ASSERT_UTF8_HOP;
756 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
757 * the bitops (especially ~) can create illegal UTF-8.
758 * In other words: in Perl UTF-8 is not just for Unicode. */
767 while (UTF8_IS_CONTINUATION(*s))
775 =for apidoc utf8_to_bytes
777 Converts a string C<s> of length C<len> from UTF-8 into native byte encoding.
778 Unlike C<bytes_to_utf8>, this over-writes the original string, and
779 updates len to contain the new length.
780 Returns zero on failure, setting C<len> to -1.
782 If you need a copy of the string, see C<bytes_from_utf8>.
788 Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len)
791 U8 * const send = s + *len;
794 PERL_ARGS_ASSERT_UTF8_TO_BYTES;
796 /* ensure valid UTF-8 and chars < 256 before updating string */
800 if (!UTF8_IS_INVARIANT(c) &&
801 (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send)
802 || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) {
803 *len = ((STRLEN) -1);
811 *d++ = (U8)utf8_to_uvchr(s, &ulen);
820 =for apidoc bytes_from_utf8
822 Converts a string C<s> of length C<len> from UTF-8 into native byte encoding.
823 Unlike C<utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to
824 the newly-created string, and updates C<len> to contain the new
825 length. Returns the original string if no conversion occurs, C<len>
826 is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to
827 0 if C<s> is converted or consisted entirely of characters that are invariant
828 in utf8 (i.e., US-ASCII on non-EBCDIC machines).
834 Perl_bytes_from_utf8(pTHX_ const U8 *s, STRLEN *len, bool *is_utf8)
841 PERL_ARGS_ASSERT_BYTES_FROM_UTF8;
847 /* ensure valid UTF-8 and chars < 256 before converting string */
848 for (send = s + *len; s < send;) {
850 if (!UTF8_IS_INVARIANT(c)) {
851 if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send &&
852 (c = *s++) && UTF8_IS_CONTINUATION(c))
861 Newx(d, (*len) - count + 1, U8);
862 s = start; start = d;
865 if (!UTF8_IS_INVARIANT(c)) {
866 /* Then it is two-byte encoded */
867 c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++);
868 c = ASCII_TO_NATIVE(c);
878 =for apidoc bytes_to_utf8
880 Converts a string C<s> of length C<len> from the native encoding into UTF-8.
881 Returns a pointer to the newly-created string, and sets C<len> to
882 reflect the new length.
884 A NUL character will be written after the end of the string.
886 If you want to convert to UTF-8 from encodings other than
887 the native (Latin1 or EBCDIC),
888 see sv_recode_to_utf8().
894 Perl_bytes_to_utf8(pTHX_ const U8 *s, STRLEN *len)
896 const U8 * const send = s + (*len);
900 PERL_ARGS_ASSERT_BYTES_TO_UTF8;
903 Newx(d, (*len) * 2 + 1, U8);
907 const UV uv = NATIVE_TO_ASCII(*s++);
908 if (UNI_IS_INVARIANT(uv))
909 *d++ = (U8)UTF_TO_NATIVE(uv);
911 *d++ = (U8)UTF8_EIGHT_BIT_HI(uv);
912 *d++ = (U8)UTF8_EIGHT_BIT_LO(uv);
921 * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8.
923 * Destination must be pre-extended to 3/2 source. Do not use in-place.
924 * We optimize for native, for obvious reasons. */
927 Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
932 PERL_ARGS_ASSERT_UTF16_TO_UTF8;
934 if (bytelen == 1 && p[0] == 0) { /* Be understanding. */
941 Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen %"UVuf, (UV)bytelen);
946 UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */
950 *d++ = UNI_TO_NATIVE(uv);
957 *d++ = (U8)(( uv >> 6) | 0xc0);
958 *d++ = (U8)(( uv & 0x3f) | 0x80);
961 if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */
962 UV low = (p[0] << 8) + p[1];
964 if (low < 0xdc00 || low >= 0xdfff)
965 Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
966 uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000;
969 *d++ = (U8)(( uv >> 12) | 0xe0);
970 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
971 *d++ = (U8)(( uv & 0x3f) | 0x80);
975 *d++ = (U8)(( uv >> 18) | 0xf0);
976 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
977 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
978 *d++ = (U8)(( uv & 0x3f) | 0x80);
982 *newlen = d - dstart;
986 /* Note: this one is slightly destructive of the source. */
989 Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
992 U8* const send = s + bytelen;
994 PERL_ARGS_ASSERT_UTF16_TO_UTF8_REVERSED;
1002 return utf16_to_utf8(p, d, bytelen, newlen);
1005 /* for now these are all defined (inefficiently) in terms of the utf8 versions */
1008 Perl_is_uni_alnum(pTHX_ UV c)
1010 U8 tmpbuf[UTF8_MAXBYTES+1];
1011 uvchr_to_utf8(tmpbuf, c);
1012 return is_utf8_alnum(tmpbuf);
1016 Perl_is_uni_alnumc(pTHX_ UV c)
1018 U8 tmpbuf[UTF8_MAXBYTES+1];
1019 uvchr_to_utf8(tmpbuf, c);
1020 return is_utf8_alnumc(tmpbuf);
1024 Perl_is_uni_idfirst(pTHX_ UV c)
1026 U8 tmpbuf[UTF8_MAXBYTES+1];
1027 uvchr_to_utf8(tmpbuf, c);
1028 return is_utf8_idfirst(tmpbuf);
1032 Perl_is_uni_alpha(pTHX_ UV c)
1034 U8 tmpbuf[UTF8_MAXBYTES+1];
1035 uvchr_to_utf8(tmpbuf, c);
1036 return is_utf8_alpha(tmpbuf);
1040 Perl_is_uni_ascii(pTHX_ UV c)
1042 U8 tmpbuf[UTF8_MAXBYTES+1];
1043 uvchr_to_utf8(tmpbuf, c);
1044 return is_utf8_ascii(tmpbuf);
1048 Perl_is_uni_space(pTHX_ UV c)
1050 U8 tmpbuf[UTF8_MAXBYTES+1];
1051 uvchr_to_utf8(tmpbuf, c);
1052 return is_utf8_space(tmpbuf);
1056 Perl_is_uni_digit(pTHX_ UV c)
1058 U8 tmpbuf[UTF8_MAXBYTES+1];
1059 uvchr_to_utf8(tmpbuf, c);
1060 return is_utf8_digit(tmpbuf);
1064 Perl_is_uni_upper(pTHX_ UV c)
1066 U8 tmpbuf[UTF8_MAXBYTES+1];
1067 uvchr_to_utf8(tmpbuf, c);
1068 return is_utf8_upper(tmpbuf);
1072 Perl_is_uni_lower(pTHX_ UV c)
1074 U8 tmpbuf[UTF8_MAXBYTES+1];
1075 uvchr_to_utf8(tmpbuf, c);
1076 return is_utf8_lower(tmpbuf);
1080 Perl_is_uni_cntrl(pTHX_ UV c)
1082 U8 tmpbuf[UTF8_MAXBYTES+1];
1083 uvchr_to_utf8(tmpbuf, c);
1084 return is_utf8_cntrl(tmpbuf);
1088 Perl_is_uni_graph(pTHX_ UV c)
1090 U8 tmpbuf[UTF8_MAXBYTES+1];
1091 uvchr_to_utf8(tmpbuf, c);
1092 return is_utf8_graph(tmpbuf);
1096 Perl_is_uni_print(pTHX_ UV c)
1098 U8 tmpbuf[UTF8_MAXBYTES+1];
1099 uvchr_to_utf8(tmpbuf, c);
1100 return is_utf8_print(tmpbuf);
1104 Perl_is_uni_punct(pTHX_ UV c)
1106 U8 tmpbuf[UTF8_MAXBYTES+1];
1107 uvchr_to_utf8(tmpbuf, c);
1108 return is_utf8_punct(tmpbuf);
1112 Perl_is_uni_xdigit(pTHX_ UV c)
1114 U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
1115 uvchr_to_utf8(tmpbuf, c);
1116 return is_utf8_xdigit(tmpbuf);
1120 Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp)
1122 PERL_ARGS_ASSERT_TO_UNI_UPPER;
1124 uvchr_to_utf8(p, c);
1125 return to_utf8_upper(p, p, lenp);
1129 Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp)
1131 PERL_ARGS_ASSERT_TO_UNI_TITLE;
1133 uvchr_to_utf8(p, c);
1134 return to_utf8_title(p, p, lenp);
1138 Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp)
1140 PERL_ARGS_ASSERT_TO_UNI_LOWER;
1142 uvchr_to_utf8(p, c);
1143 return to_utf8_lower(p, p, lenp);
1147 Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp)
1149 PERL_ARGS_ASSERT_TO_UNI_FOLD;
1151 uvchr_to_utf8(p, c);
1152 return to_utf8_fold(p, p, lenp);
1155 /* for now these all assume no locale info available for Unicode > 255 */
1158 Perl_is_uni_alnum_lc(pTHX_ UV c)
1160 return is_uni_alnum(c); /* XXX no locale support yet */
1164 Perl_is_uni_alnumc_lc(pTHX_ UV c)
1166 return is_uni_alnumc(c); /* XXX no locale support yet */
1170 Perl_is_uni_idfirst_lc(pTHX_ UV c)
1172 return is_uni_idfirst(c); /* XXX no locale support yet */
1176 Perl_is_uni_alpha_lc(pTHX_ UV c)
1178 return is_uni_alpha(c); /* XXX no locale support yet */
1182 Perl_is_uni_ascii_lc(pTHX_ UV c)
1184 return is_uni_ascii(c); /* XXX no locale support yet */
1188 Perl_is_uni_space_lc(pTHX_ UV c)
1190 return is_uni_space(c); /* XXX no locale support yet */
1194 Perl_is_uni_digit_lc(pTHX_ UV c)
1196 return is_uni_digit(c); /* XXX no locale support yet */
1200 Perl_is_uni_upper_lc(pTHX_ UV c)
1202 return is_uni_upper(c); /* XXX no locale support yet */
1206 Perl_is_uni_lower_lc(pTHX_ UV c)
1208 return is_uni_lower(c); /* XXX no locale support yet */
1212 Perl_is_uni_cntrl_lc(pTHX_ UV c)
1214 return is_uni_cntrl(c); /* XXX no locale support yet */
1218 Perl_is_uni_graph_lc(pTHX_ UV c)
1220 return is_uni_graph(c); /* XXX no locale support yet */
1224 Perl_is_uni_print_lc(pTHX_ UV c)
1226 return is_uni_print(c); /* XXX no locale support yet */
1230 Perl_is_uni_punct_lc(pTHX_ UV c)
1232 return is_uni_punct(c); /* XXX no locale support yet */
1236 Perl_is_uni_xdigit_lc(pTHX_ UV c)
1238 return is_uni_xdigit(c); /* XXX no locale support yet */
1242 Perl_to_uni_upper_lc(pTHX_ U32 c)
1244 /* XXX returns only the first character -- do not use XXX */
1245 /* XXX no locale support yet */
1247 U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
1248 return (U32)to_uni_upper(c, tmpbuf, &len);
1252 Perl_to_uni_title_lc(pTHX_ U32 c)
1254 /* XXX returns only the first character XXX -- do not use XXX */
1255 /* XXX no locale support yet */
1257 U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
1258 return (U32)to_uni_title(c, tmpbuf, &len);
1262 Perl_to_uni_lower_lc(pTHX_ U32 c)
1264 /* XXX returns only the first character -- do not use XXX */
1265 /* XXX no locale support yet */
1267 U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
1268 return (U32)to_uni_lower(c, tmpbuf, &len);
1272 S_is_utf8_common(pTHX_ const U8 *const p, SV **swash,
1273 const char *const swashname)
1277 PERL_ARGS_ASSERT_IS_UTF8_COMMON;
1279 if (!is_utf8_char(p))
1282 *swash = swash_init("utf8", swashname, &PL_sv_undef, 1, 0);
1283 return swash_fetch(*swash, p, TRUE) != 0;
1287 Perl_is_utf8_alnum(pTHX_ const U8 *p)
1291 PERL_ARGS_ASSERT_IS_UTF8_ALNUM;
1293 /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true
1294 * descendant of isalnum(3), in other words, it doesn't
1295 * contain the '_'. --jhi */
1296 return is_utf8_common(p, &PL_utf8_alnum, "IsWord");
1300 Perl_is_utf8_alnumc(pTHX_ const U8 *p)
1304 PERL_ARGS_ASSERT_IS_UTF8_ALNUMC;
1306 return is_utf8_common(p, &PL_utf8_alnumc, "IsAlnumC");
1310 Perl_is_utf8_idfirst(pTHX_ const U8 *p) /* The naming is historical. */
1314 PERL_ARGS_ASSERT_IS_UTF8_IDFIRST;
1318 /* is_utf8_idstart would be more logical. */
1319 return is_utf8_common(p, &PL_utf8_idstart, "IdStart");
1323 Perl_is_utf8_idcont(pTHX_ const U8 *p)
1327 PERL_ARGS_ASSERT_IS_UTF8_IDCONT;
1331 return is_utf8_common(p, &PL_utf8_idcont, "IdContinue");
1335 Perl_is_utf8_alpha(pTHX_ const U8 *p)
1339 PERL_ARGS_ASSERT_IS_UTF8_ALPHA;
1341 return is_utf8_common(p, &PL_utf8_alpha, "IsAlpha");
1345 Perl_is_utf8_ascii(pTHX_ const U8 *p)
1349 PERL_ARGS_ASSERT_IS_UTF8_ASCII;
1351 return is_utf8_common(p, &PL_utf8_ascii, "IsAscii");
1355 Perl_is_utf8_space(pTHX_ const U8 *p)
1359 PERL_ARGS_ASSERT_IS_UTF8_SPACE;
1361 return is_utf8_common(p, &PL_utf8_space, "IsSpacePerl");
1365 Perl_is_utf8_digit(pTHX_ const U8 *p)
1369 PERL_ARGS_ASSERT_IS_UTF8_DIGIT;
1371 return is_utf8_common(p, &PL_utf8_digit, "IsDigit");
1375 Perl_is_utf8_upper(pTHX_ const U8 *p)
1379 PERL_ARGS_ASSERT_IS_UTF8_UPPER;
1381 return is_utf8_common(p, &PL_utf8_upper, "IsUppercase");
1385 Perl_is_utf8_lower(pTHX_ const U8 *p)
1389 PERL_ARGS_ASSERT_IS_UTF8_LOWER;
1391 return is_utf8_common(p, &PL_utf8_lower, "IsLowercase");
1395 Perl_is_utf8_cntrl(pTHX_ const U8 *p)
1399 PERL_ARGS_ASSERT_IS_UTF8_CNTRL;
1401 return is_utf8_common(p, &PL_utf8_cntrl, "IsCntrl");
1405 Perl_is_utf8_graph(pTHX_ const U8 *p)
1409 PERL_ARGS_ASSERT_IS_UTF8_GRAPH;
1411 return is_utf8_common(p, &PL_utf8_graph, "IsGraph");
1415 Perl_is_utf8_print(pTHX_ const U8 *p)
1419 PERL_ARGS_ASSERT_IS_UTF8_PRINT;
1421 return is_utf8_common(p, &PL_utf8_print, "IsPrint");
1425 Perl_is_utf8_punct(pTHX_ const U8 *p)
1429 PERL_ARGS_ASSERT_IS_UTF8_PUNCT;
1431 return is_utf8_common(p, &PL_utf8_punct, "IsPunct");
1435 Perl_is_utf8_xdigit(pTHX_ const U8 *p)
1439 PERL_ARGS_ASSERT_IS_UTF8_XDIGIT;
1441 return is_utf8_common(p, &PL_utf8_xdigit, "Isxdigit");
1445 Perl_is_utf8_mark(pTHX_ const U8 *p)
1449 PERL_ARGS_ASSERT_IS_UTF8_MARK;
1451 return is_utf8_common(p, &PL_utf8_mark, "IsM");
1455 =for apidoc to_utf8_case
1457 The "p" contains the pointer to the UTF-8 string encoding
1458 the character that is being converted.
1460 The "ustrp" is a pointer to the character buffer to put the
1461 conversion result to. The "lenp" is a pointer to the length
1464 The "swashp" is a pointer to the swash to use.
1466 Both the special and normal mappings are stored lib/unicore/To/Foo.pl,
1467 and loaded by SWASHNEW, using lib/utf8_heavy.pl. The special (usually,
1468 but not always, a multicharacter mapping), is tried first.
1470 The "special" is a string like "utf8::ToSpecLower", which means the
1471 hash %utf8::ToSpecLower. The access to the hash is through
1472 Perl_to_utf8_case().
1474 The "normal" is a string like "ToLower" which means the swash
1480 Perl_to_utf8_case(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp,
1481 SV **swashp, const char *normal, const char *special)
1484 U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
1486 const UV uv0 = utf8_to_uvchr(p, NULL);
1487 /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings
1488 * are necessary in EBCDIC, they are redundant no-ops
1489 * in ASCII-ish platforms, and hopefully optimized away. */
1490 const UV uv1 = NATIVE_TO_UNI(uv0);
1492 PERL_ARGS_ASSERT_TO_UTF8_CASE;
1494 uvuni_to_utf8(tmpbuf, uv1);
1496 if (!*swashp) /* load on-demand */
1497 *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0);
1499 /* The 0xDF is the only special casing Unicode code point below 0x100. */
1500 if (special && (uv1 == 0xDF || uv1 > 0xFF)) {
1501 /* It might be "special" (sometimes, but not always,
1502 * a multicharacter mapping) */
1503 HV * const hv = get_hv(special, 0);
1507 (svp = hv_fetch(hv, (const char*)tmpbuf, UNISKIP(uv1), FALSE)) &&
1511 s = SvPV_const(*svp, len);
1513 len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp;
1516 /* If we have EBCDIC we need to remap the characters
1517 * since any characters in the low 256 are Unicode
1518 * code points, not EBCDIC. */
1519 U8 *t = (U8*)s, *tend = t + len, *d;
1526 const UV c = utf8_to_uvchr(t, &tlen);
1528 d = uvchr_to_utf8(d, UNI_TO_NATIVE(c));
1537 d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t));
1542 Copy(tmpbuf, ustrp, len, U8);
1544 Copy(s, ustrp, len, U8);
1550 if (!len && *swashp) {
1551 const UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE);
1554 /* It was "normal" (a single character mapping). */
1555 const UV uv3 = UNI_TO_NATIVE(uv2);
1556 len = uvchr_to_utf8(ustrp, uv3) - ustrp;
1560 if (!len) /* Neither: just copy. */
1561 len = uvchr_to_utf8(ustrp, uv0) - ustrp;
1566 return len ? utf8_to_uvchr(ustrp, 0) : 0;
1570 =for apidoc to_utf8_upper
1572 Convert the UTF-8 encoded character at p to its uppercase version and
1573 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1574 that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since
1575 the uppercase version may be longer than the original character.
1577 The first character of the uppercased version is returned
1578 (but note, as explained above, that there may be more.)
1583 Perl_to_utf8_upper(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp)
1587 PERL_ARGS_ASSERT_TO_UTF8_UPPER;
1589 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1590 &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper");
1594 =for apidoc to_utf8_title
1596 Convert the UTF-8 encoded character at p to its titlecase version and
1597 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1598 that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the
1599 titlecase version may be longer than the original character.
1601 The first character of the titlecased version is returned
1602 (but note, as explained above, that there may be more.)
1607 Perl_to_utf8_title(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp)
1611 PERL_ARGS_ASSERT_TO_UTF8_TITLE;
1613 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1614 &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle");
1618 =for apidoc to_utf8_lower
1620 Convert the UTF-8 encoded character at p to its lowercase version and
1621 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1622 that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the
1623 lowercase version may be longer than the original character.
1625 The first character of the lowercased version is returned
1626 (but note, as explained above, that there may be more.)
1631 Perl_to_utf8_lower(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp)
1635 PERL_ARGS_ASSERT_TO_UTF8_LOWER;
1637 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1638 &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower");
1642 =for apidoc to_utf8_fold
1644 Convert the UTF-8 encoded character at p to its foldcase version and
1645 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1646 that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the
1647 foldcase version may be longer than the original character (up to
1650 The first character of the foldcased version is returned
1651 (but note, as explained above, that there may be more.)
1656 Perl_to_utf8_fold(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp)
1660 PERL_ARGS_ASSERT_TO_UTF8_FOLD;
1662 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1663 &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold");
1667 * A "swash" is a swatch hash.
1668 * A "swatch" is a bit vector generated by utf8.c:S_swash_get().
1669 * C<pkg> is a pointer to a package name for SWASHNEW, should be "utf8".
1670 * For other parameters, see utf8::SWASHNEW in lib/utf8_heavy.pl.
1673 Perl_swash_init(pTHX_ const char* pkg, const char* name, SV *listsv, I32 minbits, I32 none)
1678 const size_t pkg_len = strlen(pkg);
1679 const size_t name_len = strlen(name);
1680 HV * const stash = gv_stashpvn(pkg, pkg_len, 0);
1683 PERL_ARGS_ASSERT_SWASH_INIT;
1685 PUSHSTACKi(PERLSI_MAGIC);
1690 if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */
1692 errsv_save = newSVsv(ERRSV);
1693 /* It is assumed that callers of this routine are not passing in any
1694 user derived data. */
1695 /* Need to do this after save_re_context() as it will set PL_tainted to
1696 1 while saving $1 etc (see the code after getrx: in Perl_magic_get).
1697 Even line to create errsv_save can turn on PL_tainted. */
1698 SAVEBOOL(PL_tainted);
1700 Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpvn(pkg,pkg_len),
1703 sv_setsv(ERRSV, errsv_save);
1704 SvREFCNT_dec(errsv_save);
1710 mPUSHp(pkg, pkg_len);
1711 mPUSHp(name, name_len);
1716 errsv_save = newSVsv(ERRSV);
1717 if (call_method("SWASHNEW", G_SCALAR))
1718 retval = newSVsv(*PL_stack_sp--);
1720 retval = &PL_sv_undef;
1722 sv_setsv(ERRSV, errsv_save);
1723 SvREFCNT_dec(errsv_save);
1726 if (IN_PERL_COMPILETIME) {
1727 CopHINTS_set(PL_curcop, PL_hints);
1729 if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) {
1731 Perl_croak(aTHX_ "Can't find Unicode property definition \"%"SVf"\"",
1733 Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref");
1739 /* This API is wrong for special case conversions since we may need to
1740 * return several Unicode characters for a single Unicode character
1741 * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is
1742 * the lower-level routine, and it is similarly broken for returning
1743 * multiple values. --jhi */
1744 /* Now SWASHGET is recasted into S_swash_get in this file. */
1747 * Returns the value of property/mapping C<swash> for the first character
1748 * of the string C<ptr>. If C<do_utf8> is true, the string C<ptr> is
1749 * assumed to be in utf8. If C<do_utf8> is false, the string C<ptr> is
1750 * assumed to be in native 8-bit encoding. Caches the swatch in C<swash>.
1753 Perl_swash_fetch(pTHX_ SV *swash, const U8 *ptr, bool do_utf8)
1756 HV *const hv = MUTABLE_HV(SvRV(swash));
1761 const U8 *tmps = NULL;
1765 const UV c = NATIVE_TO_ASCII(*ptr);
1767 PERL_ARGS_ASSERT_SWASH_FETCH;
1769 if (!do_utf8 && !UNI_IS_INVARIANT(c)) {
1770 tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c);
1771 tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c);
1774 /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ
1775 * then the "swatch" is a vec() for al the chars which start
1777 * So the key in the hash (klen) is length of encoded char -1
1779 klen = UTF8SKIP(ptr) - 1;
1783 /* If char in invariant then swatch is for all the invariant chars
1784 * In both UTF-8 and UTF-8-MOD that happens to be UTF_CONTINUATION_MARK
1786 needents = UTF_CONTINUATION_MARK;
1787 off = NATIVE_TO_UTF(ptr[klen]);
1790 /* If char is encoded then swatch is for the prefix */
1791 needents = (1 << UTF_ACCUMULATION_SHIFT);
1792 off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK;
1796 * This single-entry cache saves about 1/3 of the utf8 overhead in test
1797 * suite. (That is, only 7-8% overall over just a hash cache. Still,
1798 * it's nothing to sniff at.) Pity we usually come through at least
1799 * two function calls to get here...
1801 * NB: this code assumes that swatches are never modified, once generated!
1804 if (hv == PL_last_swash_hv &&
1805 klen == PL_last_swash_klen &&
1806 (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) )
1808 tmps = PL_last_swash_tmps;
1809 slen = PL_last_swash_slen;
1812 /* Try our second-level swatch cache, kept in a hash. */
1813 SV** svp = hv_fetch(hv, (const char*)ptr, klen, FALSE);
1815 /* If not cached, generate it via swash_get */
1816 if (!svp || !SvPOK(*svp)
1817 || !(tmps = (const U8*)SvPV_const(*svp, slen))) {
1818 /* We use utf8n_to_uvuni() as we want an index into
1819 Unicode tables, not a native character number.
1821 const UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXBYTES, 0,
1823 0 : UTF8_ALLOW_ANY);
1824 swatch = swash_get(swash,
1825 /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */
1826 (klen) ? (code_point & ~(needents - 1)) : 0,
1829 if (IN_PERL_COMPILETIME)
1830 CopHINTS_set(PL_curcop, PL_hints);
1832 svp = hv_store(hv, (const char *)ptr, klen, swatch, 0);
1834 if (!svp || !(tmps = (U8*)SvPV(*svp, slen))
1835 || (slen << 3) < needents)
1836 Perl_croak(aTHX_ "panic: swash_fetch got improper swatch");
1839 PL_last_swash_hv = hv;
1840 assert(klen <= sizeof(PL_last_swash_key));
1841 PL_last_swash_klen = (U8)klen;
1842 /* FIXME change interpvar.h? */
1843 PL_last_swash_tmps = (U8 *) tmps;
1844 PL_last_swash_slen = slen;
1846 Copy(ptr, PL_last_swash_key, klen, U8);
1849 switch ((int)((slen << 3) / needents)) {
1851 bit = 1 << (off & 7);
1853 return (tmps[off] & bit) != 0;
1858 return (tmps[off] << 8) + tmps[off + 1] ;
1861 return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ;
1863 Perl_croak(aTHX_ "panic: swash_fetch got swatch of unexpected bit width");
1864 NORETURN_FUNCTION_END;
1868 * Returns a swatch (a bit vector string) for a code point sequence
1869 * that starts from the value C<start> and comprises the number C<span>.
1870 * A C<swash> must be an object created by SWASHNEW (see lib/utf8_heavy.pl).
1871 * Should be used via swash_fetch, which will cache the swatch in C<swash>.
1874 S_swash_get(pTHX_ SV* swash, UV start, UV span)
1877 U8 *l, *lend, *x, *xend, *s;
1878 STRLEN lcur, xcur, scur;
1879 HV *const hv = MUTABLE_HV(SvRV(swash));
1880 SV** const listsvp = hv_fetchs(hv, "LIST", FALSE);
1881 SV** const typesvp = hv_fetchs(hv, "TYPE", FALSE);
1882 SV** const bitssvp = hv_fetchs(hv, "BITS", FALSE);
1883 SV** const nonesvp = hv_fetchs(hv, "NONE", FALSE);
1884 SV** const extssvp = hv_fetchs(hv, "EXTRAS", FALSE);
1885 const U8* const typestr = (U8*)SvPV_nolen(*typesvp);
1886 const int typeto = typestr[0] == 'T' && typestr[1] == 'o';
1887 const STRLEN bits = SvUV(*bitssvp);
1888 const STRLEN octets = bits >> 3; /* if bits == 1, then octets == 0 */
1889 const UV none = SvUV(*nonesvp);
1890 const UV end = start + span;
1892 PERL_ARGS_ASSERT_SWASH_GET;
1894 if (bits != 1 && bits != 8 && bits != 16 && bits != 32) {
1895 Perl_croak(aTHX_ "panic: swash_get doesn't expect bits %"UVuf,
1899 /* create and initialize $swatch */
1900 scur = octets ? (span * octets) : (span + 7) / 8;
1901 swatch = newSV(scur);
1903 s = (U8*)SvPVX(swatch);
1904 if (octets && none) {
1905 const U8* const e = s + scur;
1908 *s++ = (U8)(none & 0xff);
1909 else if (bits == 16) {
1910 *s++ = (U8)((none >> 8) & 0xff);
1911 *s++ = (U8)( none & 0xff);
1913 else if (bits == 32) {
1914 *s++ = (U8)((none >> 24) & 0xff);
1915 *s++ = (U8)((none >> 16) & 0xff);
1916 *s++ = (U8)((none >> 8) & 0xff);
1917 *s++ = (U8)( none & 0xff);
1923 (void)memzero((U8*)s, scur + 1);
1925 SvCUR_set(swatch, scur);
1926 s = (U8*)SvPVX(swatch);
1928 /* read $swash->{LIST} */
1929 l = (U8*)SvPV(*listsvp, lcur);
1934 I32 flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX;
1936 U8* const nl = (U8*)memchr(l, '\n', lend - l);
1939 min = grok_hex((char *)l, &numlen, &flags, NULL);
1943 l = nl + 1; /* 1 is length of "\n" */
1947 l = lend; /* to LIST's end at which \n is not found */
1953 flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX;
1955 max = grok_hex((char *)l, &numlen, &flags, NULL);
1964 flags = PERL_SCAN_SILENT_ILLDIGIT |
1965 PERL_SCAN_DISALLOW_PREFIX;
1967 val = grok_hex((char *)l, &numlen, &flags, NULL);
1976 Perl_croak(aTHX_ "%s: illegal mapping '%s'",
1982 val = 0; /* bits == 1, then val should be ignored */
1989 Perl_croak(aTHX_ "%s: illegal mapping '%s'", typestr, l);
1993 val = 0; /* bits == 1, then val should be ignored */
2007 if (!none || val < none) {
2012 for (key = min; key <= max; key++) {
2016 /* offset must be non-negative (start <= min <= key < end) */
2017 offset = octets * (key - start);
2019 s[offset] = (U8)(val & 0xff);
2020 else if (bits == 16) {
2021 s[offset ] = (U8)((val >> 8) & 0xff);
2022 s[offset + 1] = (U8)( val & 0xff);
2024 else if (bits == 32) {
2025 s[offset ] = (U8)((val >> 24) & 0xff);
2026 s[offset + 1] = (U8)((val >> 16) & 0xff);
2027 s[offset + 2] = (U8)((val >> 8) & 0xff);
2028 s[offset + 3] = (U8)( val & 0xff);
2031 if (!none || val < none)
2035 else { /* bits == 1, then val should be ignored */
2039 for (key = min; key <= max; key++) {
2040 const STRLEN offset = (STRLEN)(key - start);
2043 s[offset >> 3] |= 1 << (offset & 7);
2049 /* read $swash->{EXTRAS} */
2050 x = (U8*)SvPV(*extssvp, xcur);
2058 SV **otherbitssvp, *other;
2062 const U8 opc = *x++;
2066 nl = (U8*)memchr(x, '\n', xend - x);
2068 if (opc != '-' && opc != '+' && opc != '!' && opc != '&') {
2070 x = nl + 1; /* 1 is length of "\n" */
2074 x = xend; /* to EXTRAS' end at which \n is not found */
2081 namelen = nl - namestr;
2085 namelen = xend - namestr;
2089 othersvp = hv_fetch(hv, (char *)namestr, namelen, FALSE);
2090 otherhv = MUTABLE_HV(SvRV(*othersvp));
2091 otherbitssvp = hv_fetchs(otherhv, "BITS", FALSE);
2092 otherbits = (STRLEN)SvUV(*otherbitssvp);
2093 if (bits < otherbits)
2094 Perl_croak(aTHX_ "panic: swash_get found swatch size mismatch");
2096 /* The "other" swatch must be destroyed after. */
2097 other = swash_get(*othersvp, start, span);
2098 o = (U8*)SvPV(other, olen);
2101 Perl_croak(aTHX_ "panic: swash_get got improper swatch");
2103 s = (U8*)SvPV(swatch, slen);
2104 if (bits == 1 && otherbits == 1) {
2106 Perl_croak(aTHX_ "panic: swash_get found swatch length mismatch");
2130 STRLEN otheroctets = otherbits >> 3;
2132 U8* const send = s + slen;
2137 if (otherbits == 1) {
2138 otherval = (o[offset >> 3] >> (offset & 7)) & 1;
2142 STRLEN vlen = otheroctets;
2150 if (opc == '+' && otherval)
2151 NOOP; /* replace with otherval */
2152 else if (opc == '!' && !otherval)
2154 else if (opc == '-' && otherval)
2156 else if (opc == '&' && !otherval)
2159 s += octets; /* no replacement */
2164 *s++ = (U8)( otherval & 0xff);
2165 else if (bits == 16) {
2166 *s++ = (U8)((otherval >> 8) & 0xff);
2167 *s++ = (U8)( otherval & 0xff);
2169 else if (bits == 32) {
2170 *s++ = (U8)((otherval >> 24) & 0xff);
2171 *s++ = (U8)((otherval >> 16) & 0xff);
2172 *s++ = (U8)((otherval >> 8) & 0xff);
2173 *s++ = (U8)( otherval & 0xff);
2177 sv_free(other); /* through with it! */
2183 =for apidoc uvchr_to_utf8
2185 Adds the UTF-8 representation of the Native codepoint C<uv> to the end
2186 of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free
2187 bytes available. The return value is the pointer to the byte after the
2188 end of the new character. In other words,
2190 d = uvchr_to_utf8(d, uv);
2192 is the recommended wide native character-aware way of saying
2199 /* On ASCII machines this is normally a macro but we want a
2200 real function in case XS code wants it
2203 Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv)
2205 PERL_ARGS_ASSERT_UVCHR_TO_UTF8;
2207 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0);
2211 Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
2213 PERL_ARGS_ASSERT_UVCHR_TO_UTF8_FLAGS;
2215 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags);
2219 =for apidoc utf8n_to_uvchr
2222 Returns the native character value of the first character in the string
2224 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
2225 length, in bytes, of that character.
2227 Allows length and flags to be passed to low level routine.
2231 /* On ASCII machines this is normally a macro but we want
2232 a real function in case XS code wants it
2235 Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen,
2238 const UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags);
2240 PERL_ARGS_ASSERT_UTF8N_TO_UVCHR;
2242 return UNI_TO_NATIVE(uv);
2246 =for apidoc pv_uni_display
2248 Build to the scalar dsv a displayable version of the string spv,
2249 length len, the displayable version being at most pvlim bytes long
2250 (if longer, the rest is truncated and "..." will be appended).
2252 The flags argument can have UNI_DISPLAY_ISPRINT set to display
2253 isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH
2254 to display the \\[nrfta\\] as the backslashed versions (like '\n')
2255 (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\).
2256 UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both
2257 UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on.
2259 The pointer to the PV of the dsv is returned.
2263 Perl_pv_uni_display(pTHX_ SV *dsv, const U8 *spv, STRLEN len, STRLEN pvlim, UV flags)
2268 PERL_ARGS_ASSERT_PV_UNI_DISPLAY;
2272 for (s = (const char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) {
2274 /* This serves double duty as a flag and a character to print after
2275 a \ when flags & UNI_DISPLAY_BACKSLASH is true.
2279 if (pvlim && SvCUR(dsv) >= pvlim) {
2283 u = utf8_to_uvchr((U8*)s, 0);
2285 const unsigned char c = (unsigned char)u & 0xFF;
2286 if (flags & UNI_DISPLAY_BACKSLASH) {
2303 const char string = ok;
2304 sv_catpvs(dsv, "\\");
2305 sv_catpvn(dsv, &string, 1);
2308 /* isPRINT() is the locale-blind version. */
2309 if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(c)) {
2310 const char string = c;
2311 sv_catpvn(dsv, &string, 1);
2316 Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u);
2319 sv_catpvs(dsv, "...");
2325 =for apidoc sv_uni_display
2327 Build to the scalar dsv a displayable version of the scalar sv,
2328 the displayable version being at most pvlim bytes long
2329 (if longer, the rest is truncated and "..." will be appended).
2331 The flags argument is as in pv_uni_display().
2333 The pointer to the PV of the dsv is returned.
2338 Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags)
2340 PERL_ARGS_ASSERT_SV_UNI_DISPLAY;
2342 return Perl_pv_uni_display(aTHX_ dsv, (const U8*)SvPVX_const(ssv),
2343 SvCUR(ssv), pvlim, flags);
2347 =for apidoc ibcmp_utf8
2349 Return true if the strings s1 and s2 differ case-insensitively, false
2350 if not (if they are equal case-insensitively). If u1 is true, the
2351 string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true,
2352 the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2
2353 are false, the respective string is assumed to be in native 8-bit
2356 If the pe1 and pe2 are non-NULL, the scanning pointers will be copied
2357 in there (they will point at the beginning of the I<next> character).
2358 If the pointers behind pe1 or pe2 are non-NULL, they are the end
2359 pointers beyond which scanning will not continue under any
2360 circumstances. If the byte lengths l1 and l2 are non-zero, s1+l1 and
2361 s2+l2 will be used as goal end pointers that will also stop the scan,
2362 and which qualify towards defining a successful match: all the scans
2363 that define an explicit length must reach their goal pointers for
2364 a match to succeed).
2366 For case-insensitiveness, the "casefolding" of Unicode is used
2367 instead of upper/lowercasing both the characters, see
2368 http://www.unicode.org/unicode/reports/tr21/ (Case Mappings).
2372 Perl_ibcmp_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2)
2375 register const U8 *p1 = (const U8*)s1;
2376 register const U8 *p2 = (const U8*)s2;
2377 register const U8 *f1 = NULL;
2378 register const U8 *f2 = NULL;
2379 register U8 *e1 = NULL;
2380 register U8 *q1 = NULL;
2381 register U8 *e2 = NULL;
2382 register U8 *q2 = NULL;
2383 STRLEN n1 = 0, n2 = 0;
2384 U8 foldbuf1[UTF8_MAXBYTES_CASE+1];
2385 U8 foldbuf2[UTF8_MAXBYTES_CASE+1];
2387 STRLEN foldlen1, foldlen2;
2390 PERL_ARGS_ASSERT_IBCMP_UTF8;
2394 /* assert(e1 || l1); */
2395 if (e1 == 0 || (l1 && l1 < (UV)(e1 - (const U8*)s1)))
2396 f1 = (const U8*)s1 + l1;
2399 /* assert(e2 || l2); */
2400 if (e2 == 0 || (l2 && l2 < (UV)(e2 - (const U8*)s2)))
2401 f2 = (const U8*)s2 + l2;
2403 /* This shouldn't happen. However, putting an assert() there makes some
2405 /* assert((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0)); */
2406 if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0))
2407 return 1; /* mismatch; possible infinite loop or false positive */
2410 natbuf[1] = 0; /* Need to terminate the buffer. */
2412 while ((e1 == 0 || p1 < e1) &&
2413 (f1 == 0 || p1 < f1) &&
2414 (e2 == 0 || p2 < e2) &&
2415 (f2 == 0 || p2 < f2)) {
2418 to_utf8_fold(p1, foldbuf1, &foldlen1);
2420 uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p1)));
2421 to_utf8_fold(natbuf, foldbuf1, &foldlen1);
2428 to_utf8_fold(p2, foldbuf2, &foldlen2);
2430 uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p2)));
2431 to_utf8_fold(natbuf, foldbuf2, &foldlen2);
2437 if ( UTF8SKIP(q1) != UTF8SKIP(q2) ||
2438 (UTF8SKIP(q1) == 1 && *q1 != *q2) ||
2439 memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) )
2440 return 1; /* mismatch */
2447 p1 += u1 ? UTF8SKIP(p1) : 1;
2449 p2 += u2 ? UTF8SKIP(p2) : 1;
2453 /* A match is defined by all the scans that specified
2454 * an explicit length reaching their final goals. */
2455 match = (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2);
2464 return match ? 0 : 1; /* 0 match, 1 mismatch */
2469 * c-indentation-style: bsd
2471 * indent-tabs-mode: t
2474 * ex: set ts=8 sts=4 sw=4 noet: