1 /* utf8.c
2 *
3 * Copyright (C) 2000, 2001, 2002, 2003, 2004, by Larry Wall and others
4 *
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.
7 *
8 */
9
10 /*
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.'
14 *
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?'
18 *
19 * ...the travellers perceived that the floor was paved with stones of many
20 * hues; branching runes and strange devices intertwined beneath their feet.
21 */
22
23 #include "EXTERN.h"
24 #define PERL_IN_UTF8_C
25 #include "perl.h"
26
27 static char unees[] = "Malformed UTF-8 character (unexpected end of string)";
28
29 /*
30 =head1 Unicode Support
31
32 =for apidoc A|U8 *|uvuni_to_utf8_flags|U8 *d|UV uv|UV flags
33
34 Adds the UTF-8 representation of the Unicode codepoint C<uv> to the end
35 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
36 bytes available. The return value is the pointer to the byte after the
37 end of the new character. In other words,
38
39 d = uvuni_to_utf8_flags(d, uv, flags);
40
41 or, in most cases,
42
43 d = uvuni_to_utf8(d, uv);
44
45 (which is equivalent to)
46
47 d = uvuni_to_utf8_flags(d, uv, 0);
48
49 is the recommended Unicode-aware way of saying
50
51 *(d++) = uv;
52
53 =cut
54 */
55
56 U8 *
Perl_uvuni_to_utf8_flags(pTHX_ U8 * d,UV uv,UV flags)57 Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
58 {
59 if (ckWARN(WARN_UTF8)) {
60 if (UNICODE_IS_SURROGATE(uv) &&
61 !(flags & UNICODE_ALLOW_SURROGATE))
62 Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv);
63 else if (
64 ((uv >= 0xFDD0 && uv <= 0xFDEF &&
65 !(flags & UNICODE_ALLOW_FDD0))
66 ||
67 ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */
68 !(flags & UNICODE_ALLOW_FFFF))) &&
69 /* UNICODE_ALLOW_SUPER includes
70 * FFFEs and FFFFs beyond 0x10FFFF. */
71 ((uv <= PERL_UNICODE_MAX) ||
72 !(flags & UNICODE_ALLOW_SUPER))
73 )
74 Perl_warner(aTHX_ packWARN(WARN_UTF8),
75 "Unicode character 0x%04"UVxf" is illegal", uv);
76 }
77 if (UNI_IS_INVARIANT(uv)) {
78 *d++ = (U8)UTF_TO_NATIVE(uv);
79 return d;
80 }
81 #if defined(EBCDIC)
82 else {
83 STRLEN len = UNISKIP(uv);
84 U8 *p = d+len-1;
85 while (p > d) {
86 *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK);
87 uv >>= UTF_ACCUMULATION_SHIFT;
88 }
89 *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len));
90 return d+len;
91 }
92 #else /* Non loop style */
93 if (uv < 0x800) {
94 *d++ = (U8)(( uv >> 6) | 0xc0);
95 *d++ = (U8)(( uv & 0x3f) | 0x80);
96 return d;
97 }
98 if (uv < 0x10000) {
99 *d++ = (U8)(( uv >> 12) | 0xe0);
100 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
101 *d++ = (U8)(( uv & 0x3f) | 0x80);
102 return d;
103 }
104 if (uv < 0x200000) {
105 *d++ = (U8)(( uv >> 18) | 0xf0);
106 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
107 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
108 *d++ = (U8)(( uv & 0x3f) | 0x80);
109 return d;
110 }
111 if (uv < 0x4000000) {
112 *d++ = (U8)(( uv >> 24) | 0xf8);
113 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
114 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
115 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
116 *d++ = (U8)(( uv & 0x3f) | 0x80);
117 return d;
118 }
119 if (uv < 0x80000000) {
120 *d++ = (U8)(( uv >> 30) | 0xfc);
121 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
122 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
123 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
124 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
125 *d++ = (U8)(( uv & 0x3f) | 0x80);
126 return d;
127 }
128 #ifdef HAS_QUAD
129 if (uv < UTF8_QUAD_MAX)
130 #endif
131 {
132 *d++ = 0xfe; /* Can't match U+FEFF! */
133 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
134 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
135 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
136 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
137 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
138 *d++ = (U8)(( uv & 0x3f) | 0x80);
139 return d;
140 }
141 #ifdef HAS_QUAD
142 {
143 *d++ = 0xff; /* Can't match U+FFFE! */
144 *d++ = 0x80; /* 6 Reserved bits */
145 *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */
146 *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80);
147 *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80);
148 *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80);
149 *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80);
150 *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80);
151 *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80);
152 *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80);
153 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
154 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
155 *d++ = (U8)(( uv & 0x3f) | 0x80);
156 return d;
157 }
158 #endif
159 #endif /* Loop style */
160 }
161
162 U8 *
Perl_uvuni_to_utf8(pTHX_ U8 * d,UV uv)163 Perl_uvuni_to_utf8(pTHX_ U8 *d, UV uv)
164 {
165 return Perl_uvuni_to_utf8_flags(aTHX_ d, uv, 0);
166 }
167
168
169 /*
170 =for apidoc A|STRLEN|is_utf8_char|U8 *s
171
172 Tests if some arbitrary number of bytes begins in a valid UTF-8
173 character. Note that an INVARIANT (i.e. ASCII) character is a valid
174 UTF-8 character. The actual number of bytes in the UTF-8 character
175 will be returned if it is valid, otherwise 0.
176
177 =cut */
178 STRLEN
Perl_is_utf8_char(pTHX_ U8 * s)179 Perl_is_utf8_char(pTHX_ U8 *s)
180 {
181 U8 u = *s;
182 STRLEN slen, len;
183 UV uv, ouv;
184
185 if (UTF8_IS_INVARIANT(u))
186 return 1;
187
188 if (!UTF8_IS_START(u))
189 return 0;
190
191 len = UTF8SKIP(s);
192
193 if (len < 2 || !UTF8_IS_CONTINUATION(s[1]))
194 return 0;
195
196 slen = len - 1;
197 s++;
198 u &= UTF_START_MASK(len);
199 uv = u;
200 ouv = uv;
201 while (slen--) {
202 if (!UTF8_IS_CONTINUATION(*s))
203 return 0;
204 uv = UTF8_ACCUMULATE(uv, *s);
205 if (uv < ouv)
206 return 0;
207 ouv = uv;
208 s++;
209 }
210
211 if ((STRLEN)UNISKIP(uv) < len)
212 return 0;
213
214 return len;
215 }
216
217 /*
218 =for apidoc A|bool|is_utf8_string|U8 *s|STRLEN len
219
220 Returns true if first C<len> bytes of the given string form a valid
221 UTF-8 string, false otherwise. Note that 'a valid UTF-8 string' does
222 not mean 'a string that contains code points above 0x7F encoded in UTF-8'
223 because a valid ASCII string is a valid UTF-8 string.
224
225 =cut
226 */
227
228 bool
Perl_is_utf8_string(pTHX_ U8 * s,STRLEN len)229 Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len)
230 {
231 U8* x = s;
232 U8* send;
233 STRLEN c;
234
235 if (!len)
236 len = strlen((char *)s);
237 send = s + len;
238
239 while (x < send) {
240 /* Inline the easy bits of is_utf8_char() here for speed... */
241 if (UTF8_IS_INVARIANT(*x))
242 c = 1;
243 else if (!UTF8_IS_START(*x))
244 return FALSE;
245 else {
246 /* ... and call is_utf8_char() only if really needed. */
247 c = is_utf8_char(x);
248 if (!c)
249 return FALSE;
250 }
251 x += c;
252 }
253 if (x != send)
254 return FALSE;
255
256 return TRUE;
257 }
258
259 /*
260 =for apidoc A|bool|is_utf8_string_loc|U8 *s|STRLEN len|U8 **p
261
262 Like is_ut8_string but store the location of the failure in
263 the last argument.
264
265 =cut
266 */
267
268 bool
Perl_is_utf8_string_loc(pTHX_ U8 * s,STRLEN len,U8 ** p)269 Perl_is_utf8_string_loc(pTHX_ U8 *s, STRLEN len, U8 **p)
270 {
271 U8* x = s;
272 U8* send;
273 STRLEN c;
274
275 if (!len)
276 len = strlen((char *)s);
277 send = s + len;
278
279 while (x < send) {
280 /* Inline the easy bits of is_utf8_char() here for speed... */
281 if (UTF8_IS_INVARIANT(*x))
282 c = 1;
283 else if (!UTF8_IS_START(*x)) {
284 if (p)
285 *p = x;
286 return FALSE;
287 }
288 else {
289 /* ... and call is_utf8_char() only if really needed. */
290 c = is_utf8_char(x);
291 if (!c) {
292 if (p)
293 *p = x;
294 return FALSE;
295 }
296 }
297 x += c;
298 }
299 if (x != send) {
300 if (p)
301 *p = x;
302 return FALSE;
303 }
304
305 return TRUE;
306 }
307
308 /*
309 =for apidoc A|UV|utf8n_to_uvuni|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
310
311 Bottom level UTF-8 decode routine.
312 Returns the unicode code point value of the first character in the string C<s>
313 which is assumed to be in UTF-8 encoding and no longer than C<curlen>;
314 C<retlen> will be set to the length, in bytes, of that character.
315
316 If C<s> does not point to a well-formed UTF-8 character, the behaviour
317 is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY,
318 it is assumed that the caller will raise a warning, and this function
319 will silently just set C<retlen> to C<-1> and return zero. If the
320 C<flags> does not contain UTF8_CHECK_ONLY, warnings about
321 malformations will be given, C<retlen> will be set to the expected
322 length of the UTF-8 character in bytes, and zero will be returned.
323
324 The C<flags> can also contain various flags to allow deviations from
325 the strict UTF-8 encoding (see F<utf8.h>).
326
327 Most code should use utf8_to_uvchr() rather than call this directly.
328
329 =cut
330 */
331
332 UV
Perl_utf8n_to_uvuni(pTHX_ U8 * s,STRLEN curlen,STRLEN * retlen,U32 flags)333 Perl_utf8n_to_uvuni(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
334 {
335 U8 *s0 = s;
336 UV uv = *s, ouv = 0;
337 STRLEN len = 1;
338 bool dowarn = ckWARN_d(WARN_UTF8);
339 UV startbyte = *s;
340 STRLEN expectlen = 0;
341 U32 warning = 0;
342
343 /* This list is a superset of the UTF8_ALLOW_XXX. */
344
345 #define UTF8_WARN_EMPTY 1
346 #define UTF8_WARN_CONTINUATION 2
347 #define UTF8_WARN_NON_CONTINUATION 3
348 #define UTF8_WARN_FE_FF 4
349 #define UTF8_WARN_SHORT 5
350 #define UTF8_WARN_OVERFLOW 6
351 #define UTF8_WARN_SURROGATE 7
352 #define UTF8_WARN_LONG 8
353 #define UTF8_WARN_FFFF 9 /* Also FFFE. */
354
355 if (curlen == 0 &&
356 !(flags & UTF8_ALLOW_EMPTY)) {
357 warning = UTF8_WARN_EMPTY;
358 goto malformed;
359 }
360
361 if (UTF8_IS_INVARIANT(uv)) {
362 if (retlen)
363 *retlen = 1;
364 return (UV) (NATIVE_TO_UTF(*s));
365 }
366
367 if (UTF8_IS_CONTINUATION(uv) &&
368 !(flags & UTF8_ALLOW_CONTINUATION)) {
369 warning = UTF8_WARN_CONTINUATION;
370 goto malformed;
371 }
372
373 if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) &&
374 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
375 warning = UTF8_WARN_NON_CONTINUATION;
376 goto malformed;
377 }
378
379 #ifdef EBCDIC
380 uv = NATIVE_TO_UTF(uv);
381 #else
382 if ((uv == 0xfe || uv == 0xff) &&
383 !(flags & UTF8_ALLOW_FE_FF)) {
384 warning = UTF8_WARN_FE_FF;
385 goto malformed;
386 }
387 #endif
388
389 if (!(uv & 0x20)) { len = 2; uv &= 0x1f; }
390 else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; }
391 else if (!(uv & 0x08)) { len = 4; uv &= 0x07; }
392 else if (!(uv & 0x04)) { len = 5; uv &= 0x03; }
393 #ifdef EBCDIC
394 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
395 else { len = 7; uv &= 0x01; }
396 #else
397 else if (!(uv & 0x02)) { len = 6; uv &= 0x01; }
398 else if (!(uv & 0x01)) { len = 7; uv = 0; }
399 else { len = 13; uv = 0; } /* whoa! */
400 #endif
401
402 if (retlen)
403 *retlen = len;
404
405 expectlen = len;
406
407 if ((curlen < expectlen) &&
408 !(flags & UTF8_ALLOW_SHORT)) {
409 warning = UTF8_WARN_SHORT;
410 goto malformed;
411 }
412
413 len--;
414 s++;
415 ouv = uv;
416
417 while (len--) {
418 if (!UTF8_IS_CONTINUATION(*s) &&
419 !(flags & UTF8_ALLOW_NON_CONTINUATION)) {
420 s--;
421 warning = UTF8_WARN_NON_CONTINUATION;
422 goto malformed;
423 }
424 else
425 uv = UTF8_ACCUMULATE(uv, *s);
426 if (!(uv > ouv)) {
427 /* These cannot be allowed. */
428 if (uv == ouv) {
429 if (!(flags & UTF8_ALLOW_LONG)) {
430 warning = UTF8_WARN_LONG;
431 goto malformed;
432 }
433 }
434 else { /* uv < ouv */
435 /* This cannot be allowed. */
436 warning = UTF8_WARN_OVERFLOW;
437 goto malformed;
438 }
439 }
440 s++;
441 ouv = uv;
442 }
443
444 if (UNICODE_IS_SURROGATE(uv) &&
445 !(flags & UTF8_ALLOW_SURROGATE)) {
446 warning = UTF8_WARN_SURROGATE;
447 goto malformed;
448 } else if ((expectlen > (STRLEN)UNISKIP(uv)) &&
449 !(flags & UTF8_ALLOW_LONG)) {
450 warning = UTF8_WARN_LONG;
451 goto malformed;
452 } else if (UNICODE_IS_ILLEGAL(uv) &&
453 !(flags & UTF8_ALLOW_FFFF)) {
454 warning = UTF8_WARN_FFFF;
455 goto malformed;
456 }
457
458 return uv;
459
460 malformed:
461
462 if (flags & UTF8_CHECK_ONLY) {
463 if (retlen)
464 *retlen = -1;
465 return 0;
466 }
467
468 if (dowarn) {
469 SV* sv = sv_2mortal(newSVpv("Malformed UTF-8 character ", 0));
470
471 switch (warning) {
472 case 0: /* Intentionally empty. */ break;
473 case UTF8_WARN_EMPTY:
474 Perl_sv_catpvf(aTHX_ sv, "(empty string)");
475 break;
476 case UTF8_WARN_CONTINUATION:
477 Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv);
478 break;
479 case UTF8_WARN_NON_CONTINUATION:
480 if (s == s0)
481 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")",
482 (UV)s[1], startbyte);
483 else
484 Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)",
485 (UV)s[1], s - s0, s - s0 > 1 ? "s" : "", startbyte, expectlen);
486
487 break;
488 case UTF8_WARN_FE_FF:
489 Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv);
490 break;
491 case UTF8_WARN_SHORT:
492 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
493 curlen, curlen == 1 ? "" : "s", expectlen, startbyte);
494 expectlen = curlen; /* distance for caller to skip */
495 break;
496 case UTF8_WARN_OVERFLOW:
497 Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")",
498 ouv, *s, startbyte);
499 break;
500 case UTF8_WARN_SURROGATE:
501 Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv);
502 break;
503 case UTF8_WARN_LONG:
504 Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")",
505 expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte);
506 break;
507 case UTF8_WARN_FFFF:
508 Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv);
509 break;
510 default:
511 Perl_sv_catpvf(aTHX_ sv, "(unknown reason)");
512 break;
513 }
514
515 if (warning) {
516 char *s = SvPVX(sv);
517
518 if (PL_op)
519 Perl_warner(aTHX_ packWARN(WARN_UTF8),
520 "%s in %s", s, OP_DESC(PL_op));
521 else
522 Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s);
523 }
524 }
525
526 if (retlen)
527 *retlen = expectlen ? expectlen : len;
528
529 return 0;
530 }
531
532 /*
533 =for apidoc A|UV|utf8_to_uvchr|U8 *s|STRLEN *retlen
534
535 Returns the native character value of the first character in the string C<s>
536 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
537 length, in bytes, of that character.
538
539 If C<s> does not point to a well-formed UTF-8 character, zero is
540 returned and retlen is set, if possible, to -1.
541
542 =cut
543 */
544
545 UV
Perl_utf8_to_uvchr(pTHX_ U8 * s,STRLEN * retlen)546 Perl_utf8_to_uvchr(pTHX_ U8 *s, STRLEN *retlen)
547 {
548 return Perl_utf8n_to_uvchr(aTHX_ s, UTF8_MAXLEN, retlen,
549 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
550 }
551
552 /*
553 =for apidoc A|UV|utf8_to_uvuni|U8 *s|STRLEN *retlen
554
555 Returns the Unicode code point of the first character in the string C<s>
556 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
557 length, in bytes, of that character.
558
559 This function should only be used when returned UV is considered
560 an index into the Unicode semantic tables (e.g. swashes).
561
562 If C<s> does not point to a well-formed UTF-8 character, zero is
563 returned and retlen is set, if possible, to -1.
564
565 =cut
566 */
567
568 UV
Perl_utf8_to_uvuni(pTHX_ U8 * s,STRLEN * retlen)569 Perl_utf8_to_uvuni(pTHX_ U8 *s, STRLEN *retlen)
570 {
571 /* Call the low level routine asking for checks */
572 return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXLEN, retlen,
573 ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
574 }
575
576 /*
577 =for apidoc A|STRLEN|utf8_length|U8 *s|U8 *e
578
579 Return the length of the UTF-8 char encoded string C<s> in characters.
580 Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end
581 up past C<e>, croaks.
582
583 =cut
584 */
585
586 STRLEN
Perl_utf8_length(pTHX_ U8 * s,U8 * e)587 Perl_utf8_length(pTHX_ U8 *s, U8 *e)
588 {
589 STRLEN len = 0;
590
591 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
592 * the bitops (especially ~) can create illegal UTF-8.
593 * In other words: in Perl UTF-8 is not just for Unicode. */
594
595 if (e < s) {
596 if (ckWARN_d(WARN_UTF8)) {
597 if (PL_op)
598 Perl_warner(aTHX_ packWARN(WARN_UTF8),
599 "%s in %s", unees, OP_DESC(PL_op));
600 else
601 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
602 }
603 return 0;
604 }
605 while (s < e) {
606 U8 t = UTF8SKIP(s);
607
608 if (e - s < t) {
609 if (ckWARN_d(WARN_UTF8)) {
610 if (PL_op)
611 Perl_warner(aTHX_ packWARN(WARN_UTF8),
612 unees, OP_DESC(PL_op));
613 else
614 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
615 }
616 return len;
617 }
618 s += t;
619 len++;
620 }
621
622 return len;
623 }
624
625 /*
626 =for apidoc A|IV|utf8_distance|U8 *a|U8 *b
627
628 Returns the number of UTF-8 characters between the UTF-8 pointers C<a>
629 and C<b>.
630
631 WARNING: use only if you *know* that the pointers point inside the
632 same UTF-8 buffer.
633
634 =cut
635 */
636
637 IV
Perl_utf8_distance(pTHX_ U8 * a,U8 * b)638 Perl_utf8_distance(pTHX_ U8 *a, U8 *b)
639 {
640 IV off = 0;
641
642 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g.
643 * the bitops (especially ~) can create illegal UTF-8.
644 * In other words: in Perl UTF-8 is not just for Unicode. */
645
646 if (a < b) {
647 while (a < b) {
648 U8 c = UTF8SKIP(a);
649
650 if (b - a < c) {
651 if (ckWARN_d(WARN_UTF8)) {
652 if (PL_op)
653 Perl_warner(aTHX_ packWARN(WARN_UTF8),
654 "%s in %s", unees, OP_DESC(PL_op));
655 else
656 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
657 }
658 return off;
659 }
660 a += c;
661 off--;
662 }
663 }
664 else {
665 while (b < a) {
666 U8 c = UTF8SKIP(b);
667
668 if (a - b < c) {
669 if (ckWARN_d(WARN_UTF8)) {
670 if (PL_op)
671 Perl_warner(aTHX_ packWARN(WARN_UTF8),
672 "%s in %s", unees, OP_DESC(PL_op));
673 else
674 Perl_warner(aTHX_ packWARN(WARN_UTF8), unees);
675 }
676 return off;
677 }
678 b += c;
679 off++;
680 }
681 }
682
683 return off;
684 }
685
686 /*
687 =for apidoc A|U8 *|utf8_hop|U8 *s|I32 off
688
689 Return the UTF-8 pointer C<s> displaced by C<off> characters, either
690 forward or backward.
691
692 WARNING: do not use the following unless you *know* C<off> is within
693 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
694 on the first byte of character or just after the last byte of a character.
695
696 =cut
697 */
698
699 U8 *
Perl_utf8_hop(pTHX_ U8 * s,I32 off)700 Perl_utf8_hop(pTHX_ U8 *s, I32 off)
701 {
702 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
703 * the bitops (especially ~) can create illegal UTF-8.
704 * In other words: in Perl UTF-8 is not just for Unicode. */
705
706 if (off >= 0) {
707 while (off--)
708 s += UTF8SKIP(s);
709 }
710 else {
711 while (off++) {
712 s--;
713 while (UTF8_IS_CONTINUATION(*s))
714 s--;
715 }
716 }
717 return s;
718 }
719
720 /*
721 =for apidoc A|U8 *|utf8_to_bytes|U8 *s|STRLEN *len
722
723 Converts a string C<s> of length C<len> from UTF-8 into byte encoding.
724 Unlike C<bytes_to_utf8>, this over-writes the original string, and
725 updates len to contain the new length.
726 Returns zero on failure, setting C<len> to -1.
727
728 =cut
729 */
730
731 U8 *
Perl_utf8_to_bytes(pTHX_ U8 * s,STRLEN * len)732 Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len)
733 {
734 U8 *send;
735 U8 *d;
736 U8 *save = s;
737
738 /* ensure valid UTF-8 and chars < 256 before updating string */
739 for (send = s + *len; s < send; ) {
740 U8 c = *s++;
741
742 if (!UTF8_IS_INVARIANT(c) &&
743 (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send)
744 || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) {
745 *len = -1;
746 return 0;
747 }
748 }
749
750 d = s = save;
751 while (s < send) {
752 STRLEN ulen;
753 *d++ = (U8)utf8_to_uvchr(s, &ulen);
754 s += ulen;
755 }
756 *d = '\0';
757 *len = d - save;
758 return save;
759 }
760
761 /*
762 =for apidoc A|U8 *|bytes_from_utf8|U8 *s|STRLEN *len|bool *is_utf8
763
764 Converts a string C<s> of length C<len> from UTF-8 into byte encoding.
765 Unlike <utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to
766 the newly-created string, and updates C<len> to contain the new
767 length. Returns the original string if no conversion occurs, C<len>
768 is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to
769 0 if C<s> is converted or contains all 7bit characters.
770
771 =cut
772 */
773
774 U8 *
Perl_bytes_from_utf8(pTHX_ U8 * s,STRLEN * len,bool * is_utf8)775 Perl_bytes_from_utf8(pTHX_ U8 *s, STRLEN *len, bool *is_utf8)
776 {
777 U8 *d;
778 U8 *start = s;
779 U8 *send;
780 I32 count = 0;
781
782 if (!*is_utf8)
783 return start;
784
785 /* ensure valid UTF-8 and chars < 256 before converting string */
786 for (send = s + *len; s < send;) {
787 U8 c = *s++;
788 if (!UTF8_IS_INVARIANT(c)) {
789 if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send &&
790 (c = *s++) && UTF8_IS_CONTINUATION(c))
791 count++;
792 else
793 return start;
794 }
795 }
796
797 *is_utf8 = 0;
798
799 Newz(801, d, (*len) - count + 1, U8);
800 s = start; start = d;
801 while (s < send) {
802 U8 c = *s++;
803 if (!UTF8_IS_INVARIANT(c)) {
804 /* Then it is two-byte encoded */
805 c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++);
806 c = ASCII_TO_NATIVE(c);
807 }
808 *d++ = c;
809 }
810 *d = '\0';
811 *len = d - start;
812 return start;
813 }
814
815 /*
816 =for apidoc A|U8 *|bytes_to_utf8|U8 *s|STRLEN *len
817
818 Converts a string C<s> of length C<len> from ASCII into UTF-8 encoding.
819 Returns a pointer to the newly-created string, and sets C<len> to
820 reflect the new length.
821
822 If you want to convert to UTF-8 from other encodings than ASCII,
823 see sv_recode_to_utf8().
824
825 =cut
826 */
827
828 U8*
Perl_bytes_to_utf8(pTHX_ U8 * s,STRLEN * len)829 Perl_bytes_to_utf8(pTHX_ U8 *s, STRLEN *len)
830 {
831 U8 *send;
832 U8 *d;
833 U8 *dst;
834 send = s + (*len);
835
836 Newz(801, d, (*len) * 2 + 1, U8);
837 dst = d;
838
839 while (s < send) {
840 UV uv = NATIVE_TO_ASCII(*s++);
841 if (UNI_IS_INVARIANT(uv))
842 *d++ = (U8)UTF_TO_NATIVE(uv);
843 else {
844 *d++ = (U8)UTF8_EIGHT_BIT_HI(uv);
845 *d++ = (U8)UTF8_EIGHT_BIT_LO(uv);
846 }
847 }
848 *d = '\0';
849 *len = d-dst;
850 return dst;
851 }
852
853 /*
854 * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8.
855 *
856 * Destination must be pre-extended to 3/2 source. Do not use in-place.
857 * We optimize for native, for obvious reasons. */
858
859 U8*
Perl_utf16_to_utf8(pTHX_ U8 * p,U8 * d,I32 bytelen,I32 * newlen)860 Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
861 {
862 U8* pend;
863 U8* dstart = d;
864
865 if (bytelen & 1)
866 Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen");
867
868 pend = p + bytelen;
869
870 while (p < pend) {
871 UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */
872 p += 2;
873 if (uv < 0x80) {
874 *d++ = (U8)uv;
875 continue;
876 }
877 if (uv < 0x800) {
878 *d++ = (U8)(( uv >> 6) | 0xc0);
879 *d++ = (U8)(( uv & 0x3f) | 0x80);
880 continue;
881 }
882 if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */
883 UV low = (p[0] << 8) + p[1];
884 p += 2;
885 if (low < 0xdc00 || low >= 0xdfff)
886 Perl_croak(aTHX_ "Malformed UTF-16 surrogate");
887 uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000;
888 }
889 if (uv < 0x10000) {
890 *d++ = (U8)(( uv >> 12) | 0xe0);
891 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
892 *d++ = (U8)(( uv & 0x3f) | 0x80);
893 continue;
894 }
895 else {
896 *d++ = (U8)(( uv >> 18) | 0xf0);
897 *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80);
898 *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80);
899 *d++ = (U8)(( uv & 0x3f) | 0x80);
900 continue;
901 }
902 }
903 *newlen = d - dstart;
904 return d;
905 }
906
907 /* Note: this one is slightly destructive of the source. */
908
909 U8*
Perl_utf16_to_utf8_reversed(pTHX_ U8 * p,U8 * d,I32 bytelen,I32 * newlen)910 Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen)
911 {
912 U8* s = (U8*)p;
913 U8* send = s + bytelen;
914 while (s < send) {
915 U8 tmp = s[0];
916 s[0] = s[1];
917 s[1] = tmp;
918 s += 2;
919 }
920 return utf16_to_utf8(p, d, bytelen, newlen);
921 }
922
923 /* for now these are all defined (inefficiently) in terms of the utf8 versions */
924
925 bool
Perl_is_uni_alnum(pTHX_ UV c)926 Perl_is_uni_alnum(pTHX_ UV c)
927 {
928 U8 tmpbuf[UTF8_MAXLEN+1];
929 uvchr_to_utf8(tmpbuf, c);
930 return is_utf8_alnum(tmpbuf);
931 }
932
933 bool
Perl_is_uni_alnumc(pTHX_ UV c)934 Perl_is_uni_alnumc(pTHX_ UV c)
935 {
936 U8 tmpbuf[UTF8_MAXLEN+1];
937 uvchr_to_utf8(tmpbuf, c);
938 return is_utf8_alnumc(tmpbuf);
939 }
940
941 bool
Perl_is_uni_idfirst(pTHX_ UV c)942 Perl_is_uni_idfirst(pTHX_ UV c)
943 {
944 U8 tmpbuf[UTF8_MAXLEN+1];
945 uvchr_to_utf8(tmpbuf, c);
946 return is_utf8_idfirst(tmpbuf);
947 }
948
949 bool
Perl_is_uni_alpha(pTHX_ UV c)950 Perl_is_uni_alpha(pTHX_ UV c)
951 {
952 U8 tmpbuf[UTF8_MAXLEN+1];
953 uvchr_to_utf8(tmpbuf, c);
954 return is_utf8_alpha(tmpbuf);
955 }
956
957 bool
Perl_is_uni_ascii(pTHX_ UV c)958 Perl_is_uni_ascii(pTHX_ UV c)
959 {
960 U8 tmpbuf[UTF8_MAXLEN+1];
961 uvchr_to_utf8(tmpbuf, c);
962 return is_utf8_ascii(tmpbuf);
963 }
964
965 bool
Perl_is_uni_space(pTHX_ UV c)966 Perl_is_uni_space(pTHX_ UV c)
967 {
968 U8 tmpbuf[UTF8_MAXLEN+1];
969 uvchr_to_utf8(tmpbuf, c);
970 return is_utf8_space(tmpbuf);
971 }
972
973 bool
Perl_is_uni_digit(pTHX_ UV c)974 Perl_is_uni_digit(pTHX_ UV c)
975 {
976 U8 tmpbuf[UTF8_MAXLEN+1];
977 uvchr_to_utf8(tmpbuf, c);
978 return is_utf8_digit(tmpbuf);
979 }
980
981 bool
Perl_is_uni_upper(pTHX_ UV c)982 Perl_is_uni_upper(pTHX_ UV c)
983 {
984 U8 tmpbuf[UTF8_MAXLEN+1];
985 uvchr_to_utf8(tmpbuf, c);
986 return is_utf8_upper(tmpbuf);
987 }
988
989 bool
Perl_is_uni_lower(pTHX_ UV c)990 Perl_is_uni_lower(pTHX_ UV c)
991 {
992 U8 tmpbuf[UTF8_MAXLEN+1];
993 uvchr_to_utf8(tmpbuf, c);
994 return is_utf8_lower(tmpbuf);
995 }
996
997 bool
Perl_is_uni_cntrl(pTHX_ UV c)998 Perl_is_uni_cntrl(pTHX_ UV c)
999 {
1000 U8 tmpbuf[UTF8_MAXLEN+1];
1001 uvchr_to_utf8(tmpbuf, c);
1002 return is_utf8_cntrl(tmpbuf);
1003 }
1004
1005 bool
Perl_is_uni_graph(pTHX_ UV c)1006 Perl_is_uni_graph(pTHX_ UV c)
1007 {
1008 U8 tmpbuf[UTF8_MAXLEN+1];
1009 uvchr_to_utf8(tmpbuf, c);
1010 return is_utf8_graph(tmpbuf);
1011 }
1012
1013 bool
Perl_is_uni_print(pTHX_ UV c)1014 Perl_is_uni_print(pTHX_ UV c)
1015 {
1016 U8 tmpbuf[UTF8_MAXLEN+1];
1017 uvchr_to_utf8(tmpbuf, c);
1018 return is_utf8_print(tmpbuf);
1019 }
1020
1021 bool
Perl_is_uni_punct(pTHX_ UV c)1022 Perl_is_uni_punct(pTHX_ UV c)
1023 {
1024 U8 tmpbuf[UTF8_MAXLEN+1];
1025 uvchr_to_utf8(tmpbuf, c);
1026 return is_utf8_punct(tmpbuf);
1027 }
1028
1029 bool
Perl_is_uni_xdigit(pTHX_ UV c)1030 Perl_is_uni_xdigit(pTHX_ UV c)
1031 {
1032 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1033 uvchr_to_utf8(tmpbuf, c);
1034 return is_utf8_xdigit(tmpbuf);
1035 }
1036
1037 UV
Perl_to_uni_upper(pTHX_ UV c,U8 * p,STRLEN * lenp)1038 Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp)
1039 {
1040 uvchr_to_utf8(p, c);
1041 return to_utf8_upper(p, p, lenp);
1042 }
1043
1044 UV
Perl_to_uni_title(pTHX_ UV c,U8 * p,STRLEN * lenp)1045 Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp)
1046 {
1047 uvchr_to_utf8(p, c);
1048 return to_utf8_title(p, p, lenp);
1049 }
1050
1051 UV
Perl_to_uni_lower(pTHX_ UV c,U8 * p,STRLEN * lenp)1052 Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp)
1053 {
1054 uvchr_to_utf8(p, c);
1055 return to_utf8_lower(p, p, lenp);
1056 }
1057
1058 UV
Perl_to_uni_fold(pTHX_ UV c,U8 * p,STRLEN * lenp)1059 Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp)
1060 {
1061 uvchr_to_utf8(p, c);
1062 return to_utf8_fold(p, p, lenp);
1063 }
1064
1065 /* for now these all assume no locale info available for Unicode > 255 */
1066
1067 bool
Perl_is_uni_alnum_lc(pTHX_ UV c)1068 Perl_is_uni_alnum_lc(pTHX_ UV c)
1069 {
1070 return is_uni_alnum(c); /* XXX no locale support yet */
1071 }
1072
1073 bool
Perl_is_uni_alnumc_lc(pTHX_ UV c)1074 Perl_is_uni_alnumc_lc(pTHX_ UV c)
1075 {
1076 return is_uni_alnumc(c); /* XXX no locale support yet */
1077 }
1078
1079 bool
Perl_is_uni_idfirst_lc(pTHX_ UV c)1080 Perl_is_uni_idfirst_lc(pTHX_ UV c)
1081 {
1082 return is_uni_idfirst(c); /* XXX no locale support yet */
1083 }
1084
1085 bool
Perl_is_uni_alpha_lc(pTHX_ UV c)1086 Perl_is_uni_alpha_lc(pTHX_ UV c)
1087 {
1088 return is_uni_alpha(c); /* XXX no locale support yet */
1089 }
1090
1091 bool
Perl_is_uni_ascii_lc(pTHX_ UV c)1092 Perl_is_uni_ascii_lc(pTHX_ UV c)
1093 {
1094 return is_uni_ascii(c); /* XXX no locale support yet */
1095 }
1096
1097 bool
Perl_is_uni_space_lc(pTHX_ UV c)1098 Perl_is_uni_space_lc(pTHX_ UV c)
1099 {
1100 return is_uni_space(c); /* XXX no locale support yet */
1101 }
1102
1103 bool
Perl_is_uni_digit_lc(pTHX_ UV c)1104 Perl_is_uni_digit_lc(pTHX_ UV c)
1105 {
1106 return is_uni_digit(c); /* XXX no locale support yet */
1107 }
1108
1109 bool
Perl_is_uni_upper_lc(pTHX_ UV c)1110 Perl_is_uni_upper_lc(pTHX_ UV c)
1111 {
1112 return is_uni_upper(c); /* XXX no locale support yet */
1113 }
1114
1115 bool
Perl_is_uni_lower_lc(pTHX_ UV c)1116 Perl_is_uni_lower_lc(pTHX_ UV c)
1117 {
1118 return is_uni_lower(c); /* XXX no locale support yet */
1119 }
1120
1121 bool
Perl_is_uni_cntrl_lc(pTHX_ UV c)1122 Perl_is_uni_cntrl_lc(pTHX_ UV c)
1123 {
1124 return is_uni_cntrl(c); /* XXX no locale support yet */
1125 }
1126
1127 bool
Perl_is_uni_graph_lc(pTHX_ UV c)1128 Perl_is_uni_graph_lc(pTHX_ UV c)
1129 {
1130 return is_uni_graph(c); /* XXX no locale support yet */
1131 }
1132
1133 bool
Perl_is_uni_print_lc(pTHX_ UV c)1134 Perl_is_uni_print_lc(pTHX_ UV c)
1135 {
1136 return is_uni_print(c); /* XXX no locale support yet */
1137 }
1138
1139 bool
Perl_is_uni_punct_lc(pTHX_ UV c)1140 Perl_is_uni_punct_lc(pTHX_ UV c)
1141 {
1142 return is_uni_punct(c); /* XXX no locale support yet */
1143 }
1144
1145 bool
Perl_is_uni_xdigit_lc(pTHX_ UV c)1146 Perl_is_uni_xdigit_lc(pTHX_ UV c)
1147 {
1148 return is_uni_xdigit(c); /* XXX no locale support yet */
1149 }
1150
1151 U32
Perl_to_uni_upper_lc(pTHX_ U32 c)1152 Perl_to_uni_upper_lc(pTHX_ U32 c)
1153 {
1154 /* XXX returns only the first character -- do not use XXX */
1155 /* XXX no locale support yet */
1156 STRLEN len;
1157 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1158 return (U32)to_uni_upper(c, tmpbuf, &len);
1159 }
1160
1161 U32
Perl_to_uni_title_lc(pTHX_ U32 c)1162 Perl_to_uni_title_lc(pTHX_ U32 c)
1163 {
1164 /* XXX returns only the first character XXX -- do not use XXX */
1165 /* XXX no locale support yet */
1166 STRLEN len;
1167 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1168 return (U32)to_uni_title(c, tmpbuf, &len);
1169 }
1170
1171 U32
Perl_to_uni_lower_lc(pTHX_ U32 c)1172 Perl_to_uni_lower_lc(pTHX_ U32 c)
1173 {
1174 /* XXX returns only the first character -- do not use XXX */
1175 /* XXX no locale support yet */
1176 STRLEN len;
1177 U8 tmpbuf[UTF8_MAXLEN_UCLC+1];
1178 return (U32)to_uni_lower(c, tmpbuf, &len);
1179 }
1180
1181 bool
Perl_is_utf8_alnum(pTHX_ U8 * p)1182 Perl_is_utf8_alnum(pTHX_ U8 *p)
1183 {
1184 if (!is_utf8_char(p))
1185 return FALSE;
1186 if (!PL_utf8_alnum)
1187 /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true
1188 * descendant of isalnum(3), in other words, it doesn't
1189 * contain the '_'. --jhi */
1190 PL_utf8_alnum = swash_init("utf8", "IsWord", &PL_sv_undef, 0, 0);
1191 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1192 /* return *p == '_' || is_utf8_alpha(p) || is_utf8_digit(p); */
1193 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1194 if (!PL_utf8_alnum)
1195 PL_utf8_alnum = swash_init("utf8", "",
1196 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1197 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1198 #endif
1199 }
1200
1201 bool
Perl_is_utf8_alnumc(pTHX_ U8 * p)1202 Perl_is_utf8_alnumc(pTHX_ U8 *p)
1203 {
1204 if (!is_utf8_char(p))
1205 return FALSE;
1206 if (!PL_utf8_alnum)
1207 PL_utf8_alnum = swash_init("utf8", "IsAlnumC", &PL_sv_undef, 0, 0);
1208 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1209 /* return is_utf8_alpha(p) || is_utf8_digit(p); */
1210 #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */
1211 if (!PL_utf8_alnum)
1212 PL_utf8_alnum = swash_init("utf8", "",
1213 sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0);
1214 return swash_fetch(PL_utf8_alnum, p, TRUE) != 0;
1215 #endif
1216 }
1217
1218 bool
Perl_is_utf8_idfirst(pTHX_ U8 * p)1219 Perl_is_utf8_idfirst(pTHX_ U8 *p) /* The naming is historical. */
1220 {
1221 if (*p == '_')
1222 return TRUE;
1223 if (!is_utf8_char(p))
1224 return FALSE;
1225 if (!PL_utf8_idstart) /* is_utf8_idstart would be more logical. */
1226 PL_utf8_idstart = swash_init("utf8", "IdStart", &PL_sv_undef, 0, 0);
1227 return swash_fetch(PL_utf8_idstart, p, TRUE) != 0;
1228 }
1229
1230 bool
Perl_is_utf8_idcont(pTHX_ U8 * p)1231 Perl_is_utf8_idcont(pTHX_ U8 *p)
1232 {
1233 if (*p == '_')
1234 return TRUE;
1235 if (!is_utf8_char(p))
1236 return FALSE;
1237 if (!PL_utf8_idcont)
1238 PL_utf8_idcont = swash_init("utf8", "IdContinue", &PL_sv_undef, 0, 0);
1239 return swash_fetch(PL_utf8_idcont, p, TRUE) != 0;
1240 }
1241
1242 bool
Perl_is_utf8_alpha(pTHX_ U8 * p)1243 Perl_is_utf8_alpha(pTHX_ U8 *p)
1244 {
1245 if (!is_utf8_char(p))
1246 return FALSE;
1247 if (!PL_utf8_alpha)
1248 PL_utf8_alpha = swash_init("utf8", "IsAlpha", &PL_sv_undef, 0, 0);
1249 return swash_fetch(PL_utf8_alpha, p, TRUE) != 0;
1250 }
1251
1252 bool
Perl_is_utf8_ascii(pTHX_ U8 * p)1253 Perl_is_utf8_ascii(pTHX_ U8 *p)
1254 {
1255 if (!is_utf8_char(p))
1256 return FALSE;
1257 if (!PL_utf8_ascii)
1258 PL_utf8_ascii = swash_init("utf8", "IsAscii", &PL_sv_undef, 0, 0);
1259 return swash_fetch(PL_utf8_ascii, p, TRUE) != 0;
1260 }
1261
1262 bool
Perl_is_utf8_space(pTHX_ U8 * p)1263 Perl_is_utf8_space(pTHX_ U8 *p)
1264 {
1265 if (!is_utf8_char(p))
1266 return FALSE;
1267 if (!PL_utf8_space)
1268 PL_utf8_space = swash_init("utf8", "IsSpacePerl", &PL_sv_undef, 0, 0);
1269 return swash_fetch(PL_utf8_space, p, TRUE) != 0;
1270 }
1271
1272 bool
Perl_is_utf8_digit(pTHX_ U8 * p)1273 Perl_is_utf8_digit(pTHX_ U8 *p)
1274 {
1275 if (!is_utf8_char(p))
1276 return FALSE;
1277 if (!PL_utf8_digit)
1278 PL_utf8_digit = swash_init("utf8", "IsDigit", &PL_sv_undef, 0, 0);
1279 return swash_fetch(PL_utf8_digit, p, TRUE) != 0;
1280 }
1281
1282 bool
Perl_is_utf8_upper(pTHX_ U8 * p)1283 Perl_is_utf8_upper(pTHX_ U8 *p)
1284 {
1285 if (!is_utf8_char(p))
1286 return FALSE;
1287 if (!PL_utf8_upper)
1288 PL_utf8_upper = swash_init("utf8", "IsUppercase", &PL_sv_undef, 0, 0);
1289 return swash_fetch(PL_utf8_upper, p, TRUE) != 0;
1290 }
1291
1292 bool
Perl_is_utf8_lower(pTHX_ U8 * p)1293 Perl_is_utf8_lower(pTHX_ U8 *p)
1294 {
1295 if (!is_utf8_char(p))
1296 return FALSE;
1297 if (!PL_utf8_lower)
1298 PL_utf8_lower = swash_init("utf8", "IsLowercase", &PL_sv_undef, 0, 0);
1299 return swash_fetch(PL_utf8_lower, p, TRUE) != 0;
1300 }
1301
1302 bool
Perl_is_utf8_cntrl(pTHX_ U8 * p)1303 Perl_is_utf8_cntrl(pTHX_ U8 *p)
1304 {
1305 if (!is_utf8_char(p))
1306 return FALSE;
1307 if (!PL_utf8_cntrl)
1308 PL_utf8_cntrl = swash_init("utf8", "IsCntrl", &PL_sv_undef, 0, 0);
1309 return swash_fetch(PL_utf8_cntrl, p, TRUE) != 0;
1310 }
1311
1312 bool
Perl_is_utf8_graph(pTHX_ U8 * p)1313 Perl_is_utf8_graph(pTHX_ U8 *p)
1314 {
1315 if (!is_utf8_char(p))
1316 return FALSE;
1317 if (!PL_utf8_graph)
1318 PL_utf8_graph = swash_init("utf8", "IsGraph", &PL_sv_undef, 0, 0);
1319 return swash_fetch(PL_utf8_graph, p, TRUE) != 0;
1320 }
1321
1322 bool
Perl_is_utf8_print(pTHX_ U8 * p)1323 Perl_is_utf8_print(pTHX_ U8 *p)
1324 {
1325 if (!is_utf8_char(p))
1326 return FALSE;
1327 if (!PL_utf8_print)
1328 PL_utf8_print = swash_init("utf8", "IsPrint", &PL_sv_undef, 0, 0);
1329 return swash_fetch(PL_utf8_print, p, TRUE) != 0;
1330 }
1331
1332 bool
Perl_is_utf8_punct(pTHX_ U8 * p)1333 Perl_is_utf8_punct(pTHX_ U8 *p)
1334 {
1335 if (!is_utf8_char(p))
1336 return FALSE;
1337 if (!PL_utf8_punct)
1338 PL_utf8_punct = swash_init("utf8", "IsPunct", &PL_sv_undef, 0, 0);
1339 return swash_fetch(PL_utf8_punct, p, TRUE) != 0;
1340 }
1341
1342 bool
Perl_is_utf8_xdigit(pTHX_ U8 * p)1343 Perl_is_utf8_xdigit(pTHX_ U8 *p)
1344 {
1345 if (!is_utf8_char(p))
1346 return FALSE;
1347 if (!PL_utf8_xdigit)
1348 PL_utf8_xdigit = swash_init("utf8", "IsXDigit", &PL_sv_undef, 0, 0);
1349 return swash_fetch(PL_utf8_xdigit, p, TRUE) != 0;
1350 }
1351
1352 bool
Perl_is_utf8_mark(pTHX_ U8 * p)1353 Perl_is_utf8_mark(pTHX_ U8 *p)
1354 {
1355 if (!is_utf8_char(p))
1356 return FALSE;
1357 if (!PL_utf8_mark)
1358 PL_utf8_mark = swash_init("utf8", "IsM", &PL_sv_undef, 0, 0);
1359 return swash_fetch(PL_utf8_mark, p, TRUE) != 0;
1360 }
1361
1362 /*
1363 =for apidoc A|UV|to_utf8_case|U8 *p|U8* ustrp|STRLEN *lenp|SV **swash|char *normal|char *special
1364
1365 The "p" contains the pointer to the UTF-8 string encoding
1366 the character that is being converted.
1367
1368 The "ustrp" is a pointer to the character buffer to put the
1369 conversion result to. The "lenp" is a pointer to the length
1370 of the result.
1371
1372 The "swashp" is a pointer to the swash to use.
1373
1374 Both the special and normal mappings are stored lib/unicore/To/Foo.pl,
1375 and loaded by SWASHGET, using lib/utf8_heavy.pl. The special (usually,
1376 but not always, a multicharacter mapping), is tried first.
1377
1378 The "special" is a string like "utf8::ToSpecLower", which means the
1379 hash %utf8::ToSpecLower. The access to the hash is through
1380 Perl_to_utf8_case().
1381
1382 The "normal" is a string like "ToLower" which means the swash
1383 %utf8::ToLower.
1384
1385 =cut */
1386
1387 UV
Perl_to_utf8_case(pTHX_ U8 * p,U8 * ustrp,STRLEN * lenp,SV ** swashp,char * normal,char * special)1388 Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp, char *normal, char *special)
1389 {
1390 UV uv0, uv1;
1391 U8 tmpbuf[UTF8_MAXLEN_FOLD+1];
1392 STRLEN len = 0;
1393
1394 uv0 = utf8_to_uvchr(p, 0);
1395 /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings
1396 * are necessary in EBCDIC, they are redundant no-ops
1397 * in ASCII-ish platforms, and hopefully optimized away. */
1398 uv1 = NATIVE_TO_UNI(uv0);
1399 uvuni_to_utf8(tmpbuf, uv1);
1400
1401 if (!*swashp) /* load on-demand */
1402 *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0);
1403
1404 /* The 0xDF is the only special casing Unicode code point below 0x100. */
1405 if (special && (uv1 == 0xDF || uv1 > 0xFF)) {
1406 /* It might be "special" (sometimes, but not always,
1407 * a multicharacter mapping) */
1408 HV *hv;
1409 SV **svp;
1410
1411 if ((hv = get_hv(special, FALSE)) &&
1412 (svp = hv_fetch(hv, (const char*)tmpbuf, UNISKIP(uv1), FALSE)) &&
1413 (*svp)) {
1414 char *s;
1415
1416 s = SvPV(*svp, len);
1417 if (len == 1)
1418 len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp;
1419 else {
1420 #ifdef EBCDIC
1421 /* If we have EBCDIC we need to remap the characters
1422 * since any characters in the low 256 are Unicode
1423 * code points, not EBCDIC. */
1424 U8 *t = (U8*)s, *tend = t + len, *d;
1425
1426 d = tmpbuf;
1427 if (SvUTF8(*svp)) {
1428 STRLEN tlen = 0;
1429
1430 while (t < tend) {
1431 UV c = utf8_to_uvchr(t, &tlen);
1432 if (tlen > 0) {
1433 d = uvchr_to_utf8(d, UNI_TO_NATIVE(c));
1434 t += tlen;
1435 }
1436 else
1437 break;
1438 }
1439 }
1440 else {
1441 while (t < tend) {
1442 d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t));
1443 t++;
1444 }
1445 }
1446 len = d - tmpbuf;
1447 Copy(tmpbuf, ustrp, len, U8);
1448 #else
1449 Copy(s, ustrp, len, U8);
1450 #endif
1451 }
1452 }
1453 }
1454
1455 if (!len && *swashp) {
1456 UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE);
1457
1458 if (uv2) {
1459 /* It was "normal" (a single character mapping). */
1460 UV uv3 = UNI_TO_NATIVE(uv2);
1461
1462 len = uvchr_to_utf8(ustrp, uv3) - ustrp;
1463 }
1464 }
1465
1466 if (!len) /* Neither: just copy. */
1467 len = uvchr_to_utf8(ustrp, uv0) - ustrp;
1468
1469 if (lenp)
1470 *lenp = len;
1471
1472 return len ? utf8_to_uvchr(ustrp, 0) : 0;
1473 }
1474
1475 /*
1476 =for apidoc A|UV|to_utf8_upper|U8 *p|U8 *ustrp|STRLEN *lenp
1477
1478 Convert the UTF-8 encoded character at p to its uppercase version and
1479 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1480 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1481 uppercase version may be longer than the original character (up to two
1482 characters).
1483
1484 The first character of the uppercased version is returned
1485 (but note, as explained above, that there may be more.)
1486
1487 =cut */
1488
1489 UV
Perl_to_utf8_upper(pTHX_ U8 * p,U8 * ustrp,STRLEN * lenp)1490 Perl_to_utf8_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1491 {
1492 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1493 &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper");
1494 }
1495
1496 /*
1497 =for apidoc A|UV|to_utf8_title|U8 *p|U8 *ustrp|STRLEN *lenp
1498
1499 Convert the UTF-8 encoded character at p to its titlecase version and
1500 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1501 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1502 titlecase version may be longer than the original character (up to two
1503 characters).
1504
1505 The first character of the titlecased version is returned
1506 (but note, as explained above, that there may be more.)
1507
1508 =cut */
1509
1510 UV
Perl_to_utf8_title(pTHX_ U8 * p,U8 * ustrp,STRLEN * lenp)1511 Perl_to_utf8_title(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1512 {
1513 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1514 &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle");
1515 }
1516
1517 /*
1518 =for apidoc A|UV|to_utf8_lower|U8 *p|U8 *ustrp|STRLEN *lenp
1519
1520 Convert the UTF-8 encoded character at p to its lowercase version and
1521 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1522 that the ustrp needs to be at least UTF8_MAXLEN_UCLC+1 bytes since the
1523 lowercase version may be longer than the original character (up to two
1524 characters).
1525
1526 The first character of the lowercased version is returned
1527 (but note, as explained above, that there may be more.)
1528
1529 =cut */
1530
1531 UV
Perl_to_utf8_lower(pTHX_ U8 * p,U8 * ustrp,STRLEN * lenp)1532 Perl_to_utf8_lower(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1533 {
1534 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1535 &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower");
1536 }
1537
1538 /*
1539 =for apidoc A|UV|to_utf8_fold|U8 *p|U8 *ustrp|STRLEN *lenp
1540
1541 Convert the UTF-8 encoded character at p to its foldcase version and
1542 store that in UTF-8 in ustrp and its length in bytes in lenp. Note
1543 that the ustrp needs to be at least UTF8_MAXLEN_FOLD+1 bytes since the
1544 foldcase version may be longer than the original character (up to
1545 three characters).
1546
1547 The first character of the foldcased version is returned
1548 (but note, as explained above, that there may be more.)
1549
1550 =cut */
1551
1552 UV
Perl_to_utf8_fold(pTHX_ U8 * p,U8 * ustrp,STRLEN * lenp)1553 Perl_to_utf8_fold(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp)
1554 {
1555 return Perl_to_utf8_case(aTHX_ p, ustrp, lenp,
1556 &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold");
1557 }
1558
1559 /* a "swash" is a swatch hash */
1560
1561 SV*
Perl_swash_init(pTHX_ char * pkg,char * name,SV * listsv,I32 minbits,I32 none)1562 Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none)
1563 {
1564 SV* retval;
1565 SV* tokenbufsv = sv_2mortal(NEWSV(0,0));
1566 dSP;
1567 HV *stash = gv_stashpvn(pkg, strlen(pkg), FALSE);
1568 SV* errsv_save;
1569
1570 if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */
1571 ENTER;
1572 errsv_save = newSVsv(ERRSV);
1573 Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpv(pkg,0), Nullsv);
1574 if (!SvTRUE(ERRSV))
1575 sv_setsv(ERRSV, errsv_save);
1576 SvREFCNT_dec(errsv_save);
1577 LEAVE;
1578 }
1579 SPAGAIN;
1580 PUSHSTACKi(PERLSI_MAGIC);
1581 PUSHMARK(SP);
1582 EXTEND(SP,5);
1583 PUSHs(sv_2mortal(newSVpvn(pkg, strlen(pkg))));
1584 PUSHs(sv_2mortal(newSVpvn(name, strlen(name))));
1585 PUSHs(listsv);
1586 PUSHs(sv_2mortal(newSViv(minbits)));
1587 PUSHs(sv_2mortal(newSViv(none)));
1588 PUTBACK;
1589 ENTER;
1590 SAVEI32(PL_hints);
1591 PL_hints = 0;
1592 save_re_context();
1593 if (IN_PERL_COMPILETIME) {
1594 /* XXX ought to be handled by lex_start */
1595 SAVEI32(PL_in_my);
1596 PL_in_my = 0;
1597 sv_setpv(tokenbufsv, PL_tokenbuf);
1598 }
1599 errsv_save = newSVsv(ERRSV);
1600 if (call_method("SWASHNEW", G_SCALAR))
1601 retval = newSVsv(*PL_stack_sp--);
1602 else
1603 retval = &PL_sv_undef;
1604 if (!SvTRUE(ERRSV))
1605 sv_setsv(ERRSV, errsv_save);
1606 SvREFCNT_dec(errsv_save);
1607 LEAVE;
1608 POPSTACK;
1609 if (IN_PERL_COMPILETIME) {
1610 STRLEN len;
1611 char* pv = SvPV(tokenbufsv, len);
1612
1613 Copy(pv, PL_tokenbuf, len+1, char);
1614 PL_curcop->op_private = (U8)(PL_hints & HINT_PRIVATE_MASK);
1615 }
1616 if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) {
1617 if (SvPOK(retval))
1618 Perl_croak(aTHX_ "Can't find Unicode property definition \"%"SVf"\"",
1619 retval);
1620 Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref");
1621 }
1622 return retval;
1623 }
1624
1625
1626 /* This API is wrong for special case conversions since we may need to
1627 * return several Unicode characters for a single Unicode character
1628 * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is
1629 * the lower-level routine, and it is similarly broken for returning
1630 * multiple values. --jhi */
1631 UV
Perl_swash_fetch(pTHX_ SV * sv,U8 * ptr,bool do_utf8)1632 Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr, bool do_utf8)
1633 {
1634 HV* hv = (HV*)SvRV(sv);
1635 U32 klen;
1636 U32 off;
1637 STRLEN slen;
1638 STRLEN needents;
1639 U8 *tmps = NULL;
1640 U32 bit;
1641 SV *retval;
1642 U8 tmputf8[2];
1643 UV c = NATIVE_TO_ASCII(*ptr);
1644
1645 if (!do_utf8 && !UNI_IS_INVARIANT(c)) {
1646 tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c);
1647 tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c);
1648 ptr = tmputf8;
1649 }
1650 /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ
1651 * then the "swatch" is a vec() for al the chars which start
1652 * with 0xAA..0xYY
1653 * So the key in the hash (klen) is length of encoded char -1
1654 */
1655 klen = UTF8SKIP(ptr) - 1;
1656 off = ptr[klen];
1657
1658 if (klen == 0)
1659 {
1660 /* If char in invariant then swatch is for all the invariant chars
1661 * In both UTF-8 and UTF-8-MOD that happens to be UTF_CONTINUATION_MARK
1662 */
1663 needents = UTF_CONTINUATION_MARK;
1664 off = NATIVE_TO_UTF(ptr[klen]);
1665 }
1666 else
1667 {
1668 /* If char is encoded then swatch is for the prefix */
1669 needents = (1 << UTF_ACCUMULATION_SHIFT);
1670 off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK;
1671 }
1672
1673 /*
1674 * This single-entry cache saves about 1/3 of the utf8 overhead in test
1675 * suite. (That is, only 7-8% overall over just a hash cache. Still,
1676 * it's nothing to sniff at.) Pity we usually come through at least
1677 * two function calls to get here...
1678 *
1679 * NB: this code assumes that swatches are never modified, once generated!
1680 */
1681
1682 if (hv == PL_last_swash_hv &&
1683 klen == PL_last_swash_klen &&
1684 (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) )
1685 {
1686 tmps = PL_last_swash_tmps;
1687 slen = PL_last_swash_slen;
1688 }
1689 else {
1690 /* Try our second-level swatch cache, kept in a hash. */
1691 SV** svp = hv_fetch(hv, (char*)ptr, klen, FALSE);
1692
1693 /* If not cached, generate it via utf8::SWASHGET */
1694 if (!svp || !SvPOK(*svp) || !(tmps = (U8*)SvPV(*svp, slen))) {
1695 dSP;
1696 /* We use utf8n_to_uvuni() as we want an index into
1697 Unicode tables, not a native character number.
1698 */
1699 UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXLEN, 0,
1700 ckWARN(WARN_UTF8) ?
1701 0 : UTF8_ALLOW_ANY);
1702 SV *errsv_save;
1703 ENTER;
1704 SAVETMPS;
1705 save_re_context();
1706 PUSHSTACKi(PERLSI_MAGIC);
1707 PUSHMARK(SP);
1708 EXTEND(SP,3);
1709 PUSHs((SV*)sv);
1710 /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */
1711 PUSHs(sv_2mortal(newSViv((klen) ?
1712 (code_point & ~(needents - 1)) : 0)));
1713 PUSHs(sv_2mortal(newSViv(needents)));
1714 PUTBACK;
1715 errsv_save = newSVsv(ERRSV);
1716 if (call_method("SWASHGET", G_SCALAR))
1717 retval = newSVsv(*PL_stack_sp--);
1718 else
1719 retval = &PL_sv_undef;
1720 if (!SvTRUE(ERRSV))
1721 sv_setsv(ERRSV, errsv_save);
1722 SvREFCNT_dec(errsv_save);
1723 POPSTACK;
1724 FREETMPS;
1725 LEAVE;
1726 if (IN_PERL_COMPILETIME)
1727 PL_curcop->op_private = (U8)(PL_hints & HINT_PRIVATE_MASK);
1728
1729 svp = hv_store(hv, (char*)ptr, klen, retval, 0);
1730
1731 if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || (slen << 3) < needents)
1732 Perl_croak(aTHX_ "SWASHGET didn't return result of proper length");
1733 }
1734
1735 PL_last_swash_hv = hv;
1736 PL_last_swash_klen = klen;
1737 PL_last_swash_tmps = tmps;
1738 PL_last_swash_slen = slen;
1739 if (klen)
1740 Copy(ptr, PL_last_swash_key, klen, U8);
1741 }
1742
1743 switch ((int)((slen << 3) / needents)) {
1744 case 1:
1745 bit = 1 << (off & 7);
1746 off >>= 3;
1747 return (tmps[off] & bit) != 0;
1748 case 8:
1749 return tmps[off];
1750 case 16:
1751 off <<= 1;
1752 return (tmps[off] << 8) + tmps[off + 1] ;
1753 case 32:
1754 off <<= 2;
1755 return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ;
1756 }
1757 Perl_croak(aTHX_ "panic: swash_fetch");
1758 return 0;
1759 }
1760
1761
1762 /*
1763 =for apidoc A|U8 *|uvchr_to_utf8|U8 *d|UV uv
1764
1765 Adds the UTF-8 representation of the Native codepoint C<uv> to the end
1766 of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free
1767 bytes available. The return value is the pointer to the byte after the
1768 end of the new character. In other words,
1769
1770 d = uvchr_to_utf8(d, uv);
1771
1772 is the recommended wide native character-aware way of saying
1773
1774 *(d++) = uv;
1775
1776 =cut
1777 */
1778
1779 /* On ASCII machines this is normally a macro but we want a
1780 real function in case XS code wants it
1781 */
1782 #undef Perl_uvchr_to_utf8
1783 U8 *
Perl_uvchr_to_utf8(pTHX_ U8 * d,UV uv)1784 Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv)
1785 {
1786 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0);
1787 }
1788
1789 U8 *
Perl_uvchr_to_utf8_flags(pTHX_ U8 * d,UV uv,UV flags)1790 Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags)
1791 {
1792 return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags);
1793 }
1794
1795 /*
1796 =for apidoc A|UV|utf8n_to_uvchr|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags
1797
1798 Returns the native character value of the first character in the string C<s>
1799 which is assumed to be in UTF-8 encoding; C<retlen> will be set to the
1800 length, in bytes, of that character.
1801
1802 Allows length and flags to be passed to low level routine.
1803
1804 =cut
1805 */
1806 /* On ASCII machines this is normally a macro but we want
1807 a real function in case XS code wants it
1808 */
1809 #undef Perl_utf8n_to_uvchr
1810 UV
Perl_utf8n_to_uvchr(pTHX_ U8 * s,STRLEN curlen,STRLEN * retlen,U32 flags)1811 Perl_utf8n_to_uvchr(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags)
1812 {
1813 UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags);
1814 return UNI_TO_NATIVE(uv);
1815 }
1816
1817 /*
1818 =for apidoc A|char *|pv_uni_display|SV *dsv|U8 *spv|STRLEN len|STRLEN pvlim|UV flags
1819
1820 Build to the scalar dsv a displayable version of the string spv,
1821 length len, the displayable version being at most pvlim bytes long
1822 (if longer, the rest is truncated and "..." will be appended).
1823
1824 The flags argument can have UNI_DISPLAY_ISPRINT set to display
1825 isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH
1826 to display the \\[nrfta\\] as the backslashed versions (like '\n')
1827 (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\).
1828 UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both
1829 UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on.
1830
1831 The pointer to the PV of the dsv is returned.
1832
1833 =cut */
1834 char *
Perl_pv_uni_display(pTHX_ SV * dsv,U8 * spv,STRLEN len,STRLEN pvlim,UV flags)1835 Perl_pv_uni_display(pTHX_ SV *dsv, U8 *spv, STRLEN len, STRLEN pvlim, UV flags)
1836 {
1837 int truncated = 0;
1838 char *s, *e;
1839
1840 sv_setpvn(dsv, "", 0);
1841 for (s = (char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) {
1842 UV u;
1843 bool ok = FALSE;
1844
1845 if (pvlim && SvCUR(dsv) >= pvlim) {
1846 truncated++;
1847 break;
1848 }
1849 u = utf8_to_uvchr((U8*)s, 0);
1850 if (u < 256) {
1851 if (!ok && (flags & UNI_DISPLAY_BACKSLASH)) {
1852 switch (u & 0xFF) {
1853 case '\n':
1854 Perl_sv_catpvf(aTHX_ dsv, "\\n"); ok = TRUE; break;
1855 case '\r':
1856 Perl_sv_catpvf(aTHX_ dsv, "\\r"); ok = TRUE; break;
1857 case '\t':
1858 Perl_sv_catpvf(aTHX_ dsv, "\\t"); ok = TRUE; break;
1859 case '\f':
1860 Perl_sv_catpvf(aTHX_ dsv, "\\f"); ok = TRUE; break;
1861 case '\a':
1862 Perl_sv_catpvf(aTHX_ dsv, "\\a"); ok = TRUE; break;
1863 case '\\':
1864 Perl_sv_catpvf(aTHX_ dsv, "\\\\" ); ok = TRUE; break;
1865 default: break;
1866 }
1867 }
1868 /* isPRINT() is the locale-blind version. */
1869 if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(u & 0xFF)) {
1870 Perl_sv_catpvf(aTHX_ dsv, "%c", (char)(u & 0xFF));
1871 ok = TRUE;
1872 }
1873 }
1874 if (!ok)
1875 Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u);
1876 }
1877 if (truncated)
1878 sv_catpvn(dsv, "...", 3);
1879
1880 return SvPVX(dsv);
1881 }
1882
1883 /*
1884 =for apidoc A|char *|sv_uni_display|SV *dsv|SV *ssv|STRLEN pvlim|UV flags
1885
1886 Build to the scalar dsv a displayable version of the scalar sv,
1887 the displayable version being at most pvlim bytes long
1888 (if longer, the rest is truncated and "..." will be appended).
1889
1890 The flags argument is as in pv_uni_display().
1891
1892 The pointer to the PV of the dsv is returned.
1893
1894 =cut */
1895 char *
Perl_sv_uni_display(pTHX_ SV * dsv,SV * ssv,STRLEN pvlim,UV flags)1896 Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags)
1897 {
1898 return Perl_pv_uni_display(aTHX_ dsv, (U8*)SvPVX(ssv), SvCUR(ssv),
1899 pvlim, flags);
1900 }
1901
1902 /*
1903 =for apidoc A|I32|ibcmp_utf8|const char *s1|char **pe1|register UV l1|bool u1|const char *s2|char **pe2|register UV l2|bool u2
1904
1905 Return true if the strings s1 and s2 differ case-insensitively, false
1906 if not (if they are equal case-insensitively). If u1 is true, the
1907 string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true,
1908 the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2
1909 are false, the respective string is assumed to be in native 8-bit
1910 encoding.
1911
1912 If the pe1 and pe2 are non-NULL, the scanning pointers will be copied
1913 in there (they will point at the beginning of the I<next> character).
1914 If the pointers behind pe1 or pe2 are non-NULL, they are the end
1915 pointers beyond which scanning will not continue under any
1916 circustances. If the byte lengths l1 and l2 are non-zero, s1+l1 and
1917 s2+l2 will be used as goal end pointers that will also stop the scan,
1918 and which qualify towards defining a successful match: all the scans
1919 that define an explicit length must reach their goal pointers for
1920 a match to succeed).
1921
1922 For case-insensitiveness, the "casefolding" of Unicode is used
1923 instead of upper/lowercasing both the characters, see
1924 http://www.unicode.org/unicode/reports/tr21/ (Case Mappings).
1925
1926 =cut */
1927 I32
Perl_ibcmp_utf8(pTHX_ const char * s1,char ** pe1,register UV l1,bool u1,const char * s2,char ** pe2,register UV l2,bool u2)1928 Perl_ibcmp_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2)
1929 {
1930 register U8 *p1 = (U8*)s1;
1931 register U8 *p2 = (U8*)s2;
1932 register U8 *e1 = 0, *f1 = 0, *q1 = 0;
1933 register U8 *e2 = 0, *f2 = 0, *q2 = 0;
1934 STRLEN n1 = 0, n2 = 0;
1935 U8 foldbuf1[UTF8_MAXLEN_FOLD+1];
1936 U8 foldbuf2[UTF8_MAXLEN_FOLD+1];
1937 U8 natbuf[1+1];
1938 STRLEN foldlen1, foldlen2;
1939 bool match;
1940
1941 if (pe1)
1942 e1 = *(U8**)pe1;
1943 if (e1 == 0 || (l1 && l1 < (UV)(e1 - (U8*)s1)))
1944 f1 = (U8*)s1 + l1;
1945 if (pe2)
1946 e2 = *(U8**)pe2;
1947 if (e2 == 0 || (l2 && l2 < (UV)(e2 - (U8*)s2)))
1948 f2 = (U8*)s2 + l2;
1949
1950 if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0))
1951 return 1; /* mismatch; possible infinite loop or false positive */
1952
1953 if (!u1 || !u2)
1954 natbuf[1] = 0; /* Need to terminate the buffer. */
1955
1956 while ((e1 == 0 || p1 < e1) &&
1957 (f1 == 0 || p1 < f1) &&
1958 (e2 == 0 || p2 < e2) &&
1959 (f2 == 0 || p2 < f2)) {
1960 if (n1 == 0) {
1961 if (u1)
1962 to_utf8_fold(p1, foldbuf1, &foldlen1);
1963 else {
1964 natbuf[0] = *p1;
1965 to_utf8_fold(natbuf, foldbuf1, &foldlen1);
1966 }
1967 q1 = foldbuf1;
1968 n1 = foldlen1;
1969 }
1970 if (n2 == 0) {
1971 if (u2)
1972 to_utf8_fold(p2, foldbuf2, &foldlen2);
1973 else {
1974 natbuf[0] = *p2;
1975 to_utf8_fold(natbuf, foldbuf2, &foldlen2);
1976 }
1977 q2 = foldbuf2;
1978 n2 = foldlen2;
1979 }
1980 while (n1 && n2) {
1981 if ( UTF8SKIP(q1) != UTF8SKIP(q2) ||
1982 (UTF8SKIP(q1) == 1 && *q1 != *q2) ||
1983 memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) )
1984 return 1; /* mismatch */
1985 n1 -= UTF8SKIP(q1);
1986 q1 += UTF8SKIP(q1);
1987 n2 -= UTF8SKIP(q2);
1988 q2 += UTF8SKIP(q2);
1989 }
1990 if (n1 == 0)
1991 p1 += u1 ? UTF8SKIP(p1) : 1;
1992 if (n2 == 0)
1993 p2 += u2 ? UTF8SKIP(p2) : 1;
1994
1995 }
1996
1997 /* A match is defined by all the scans that specified
1998 * an explicit length reaching their final goals. */
1999 match = (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2);
2000
2001 if (match) {
2002 if (pe1)
2003 *pe1 = (char*)p1;
2004 if (pe2)
2005 *pe2 = (char*)p2;
2006 }
2007
2008 return match ? 0 : 1; /* 0 match, 1 mismatch */
2009 }
2010
2011