Commit | Line | Data |
a0ed51b3 |
1 | /* utf8.c |
2 | * |
bc89e66f |
3 | * Copyright (c) 1998-2001, Larry Wall |
a0ed51b3 |
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" |
864dbfa3 |
24 | #define PERL_IN_UTF8_C |
a0ed51b3 |
25 | #include "perl.h" |
26 | |
27 | /* Unicode support */ |
28 | |
eebe1485 |
29 | /* |
37607a96 |
30 | =for apidoc A|U8 *|uvuni_to_utf8|U8 *d|UV uv |
eebe1485 |
31 | |
32 | Adds the UTF8 representation of the Unicode codepoint C<uv> to the end |
33 | of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free |
34 | bytes available. The return value is the pointer to the byte after the |
9041c2e3 |
35 | end of the new character. In other words, |
eebe1485 |
36 | |
9041c2e3 |
37 | d = uvuni_to_utf8(d, uv); |
eebe1485 |
38 | |
39 | is the recommended Unicode-aware way of saying |
40 | |
41 | *(d++) = uv; |
42 | |
43 | =cut |
44 | */ |
45 | |
dfe13c55 |
46 | U8 * |
9041c2e3 |
47 | Perl_uvuni_to_utf8(pTHX_ U8 *d, UV uv) |
a0ed51b3 |
48 | { |
c4d5f83a |
49 | if (UNI_IS_INVARIANT(uv)) { |
50 | *d++ = UTF_TO_NATIVE(uv); |
a0ed51b3 |
51 | return d; |
52 | } |
2d331972 |
53 | #if defined(EBCDIC) |
1d72bdf6 |
54 | else { |
55 | STRLEN len = UNISKIP(uv); |
56 | U8 *p = d+len-1; |
57 | while (p > d) { |
58 | *p-- = UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK); |
59 | uv >>= UTF_ACCUMULATION_SHIFT; |
60 | } |
61 | *p = UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len)); |
62 | return d+len; |
63 | } |
64 | #else /* Non loop style */ |
a0ed51b3 |
65 | if (uv < 0x800) { |
66 | *d++ = (( uv >> 6) | 0xc0); |
67 | *d++ = (( uv & 0x3f) | 0x80); |
68 | return d; |
69 | } |
70 | if (uv < 0x10000) { |
71 | *d++ = (( uv >> 12) | 0xe0); |
72 | *d++ = (((uv >> 6) & 0x3f) | 0x80); |
73 | *d++ = (( uv & 0x3f) | 0x80); |
74 | return d; |
75 | } |
76 | if (uv < 0x200000) { |
77 | *d++ = (( uv >> 18) | 0xf0); |
78 | *d++ = (((uv >> 12) & 0x3f) | 0x80); |
79 | *d++ = (((uv >> 6) & 0x3f) | 0x80); |
80 | *d++ = (( uv & 0x3f) | 0x80); |
81 | return d; |
82 | } |
83 | if (uv < 0x4000000) { |
84 | *d++ = (( uv >> 24) | 0xf8); |
85 | *d++ = (((uv >> 18) & 0x3f) | 0x80); |
86 | *d++ = (((uv >> 12) & 0x3f) | 0x80); |
87 | *d++ = (((uv >> 6) & 0x3f) | 0x80); |
88 | *d++ = (( uv & 0x3f) | 0x80); |
89 | return d; |
90 | } |
91 | if (uv < 0x80000000) { |
92 | *d++ = (( uv >> 30) | 0xfc); |
93 | *d++ = (((uv >> 24) & 0x3f) | 0x80); |
94 | *d++ = (((uv >> 18) & 0x3f) | 0x80); |
95 | *d++ = (((uv >> 12) & 0x3f) | 0x80); |
96 | *d++ = (((uv >> 6) & 0x3f) | 0x80); |
97 | *d++ = (( uv & 0x3f) | 0x80); |
98 | return d; |
99 | } |
6b8eaf93 |
100 | #ifdef HAS_QUAD |
d7578b48 |
101 | if (uv < UTF8_QUAD_MAX) |
a0ed51b3 |
102 | #endif |
103 | { |
104 | *d++ = 0xfe; /* Can't match U+FEFF! */ |
105 | *d++ = (((uv >> 30) & 0x3f) | 0x80); |
106 | *d++ = (((uv >> 24) & 0x3f) | 0x80); |
107 | *d++ = (((uv >> 18) & 0x3f) | 0x80); |
108 | *d++ = (((uv >> 12) & 0x3f) | 0x80); |
109 | *d++ = (((uv >> 6) & 0x3f) | 0x80); |
110 | *d++ = (( uv & 0x3f) | 0x80); |
111 | return d; |
112 | } |
6b8eaf93 |
113 | #ifdef HAS_QUAD |
a0ed51b3 |
114 | { |
115 | *d++ = 0xff; /* Can't match U+FFFE! */ |
3c77ea2b |
116 | *d++ = 0x80; /* 6 Reserved bits */ |
117 | *d++ = (((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */ |
118 | *d++ = (((uv >> 54) & 0x3f) | 0x80); |
119 | *d++ = (((uv >> 48) & 0x3f) | 0x80); |
120 | *d++ = (((uv >> 42) & 0x3f) | 0x80); |
a0ed51b3 |
121 | *d++ = (((uv >> 36) & 0x3f) | 0x80); |
122 | *d++ = (((uv >> 30) & 0x3f) | 0x80); |
123 | *d++ = (((uv >> 24) & 0x3f) | 0x80); |
124 | *d++ = (((uv >> 18) & 0x3f) | 0x80); |
125 | *d++ = (((uv >> 12) & 0x3f) | 0x80); |
126 | *d++ = (((uv >> 6) & 0x3f) | 0x80); |
127 | *d++ = (( uv & 0x3f) | 0x80); |
128 | return d; |
129 | } |
130 | #endif |
1d72bdf6 |
131 | #endif /* Loop style */ |
a0ed51b3 |
132 | } |
133 | |
9041c2e3 |
134 | |
135 | |
136 | /* |
eebe1485 |
137 | =for apidoc A|STRLEN|is_utf8_char|U8 *s |
138 | |
5da9da9e |
139 | Tests if some arbitrary number of bytes begins in a valid UTF-8 |
1d72bdf6 |
140 | character. Note that an INVARIANT (i.e. ASCII) character is a valid UTF-8 character. |
5da9da9e |
141 | The actual number of bytes in the UTF-8 character will be returned if |
142 | it is valid, otherwise 0. |
9041c2e3 |
143 | |
37607a96 |
144 | =cut |
145 | */ |
067a85ef |
146 | STRLEN |
386d01d6 |
147 | Perl_is_utf8_char(pTHX_ U8 *s) |
148 | { |
149 | U8 u = *s; |
067a85ef |
150 | STRLEN slen, len; |
151 | UV uv, ouv; |
386d01d6 |
152 | |
1d72bdf6 |
153 | if (UTF8_IS_INVARIANT(u)) |
386d01d6 |
154 | return 1; |
155 | |
60006e79 |
156 | if (!UTF8_IS_START(u)) |
386d01d6 |
157 | return 0; |
158 | |
9f07fdcd |
159 | len = UTF8SKIP(s); |
386d01d6 |
160 | |
60006e79 |
161 | if (len < 2 || !UTF8_IS_CONTINUATION(s[1])) |
067a85ef |
162 | return 0; |
163 | |
386d01d6 |
164 | slen = len - 1; |
165 | s++; |
6eb6869e |
166 | u &= UTF_START_MASK(len); |
1d72bdf6 |
167 | uv = u; |
067a85ef |
168 | ouv = uv; |
386d01d6 |
169 | while (slen--) { |
60006e79 |
170 | if (!UTF8_IS_CONTINUATION(*s)) |
386d01d6 |
171 | return 0; |
8850bf83 |
172 | uv = UTF8_ACCUMULATE(uv, *s); |
209a9bc1 |
173 | if (uv < ouv) |
067a85ef |
174 | return 0; |
175 | ouv = uv; |
386d01d6 |
176 | s++; |
177 | } |
067a85ef |
178 | |
5bbb0b5a |
179 | if (UNISKIP(uv) < len) |
067a85ef |
180 | return 0; |
181 | |
386d01d6 |
182 | return len; |
183 | } |
184 | |
6662521e |
185 | /* |
eebe1485 |
186 | =for apidoc A|bool|is_utf8_string|U8 *s|STRLEN len |
6662521e |
187 | |
5da9da9e |
188 | Returns true if first C<len> bytes of the given string form a valid UTF8 |
189 | string, false otherwise. Note that 'a valid UTF8 string' does not mean |
190 | 'a string that contains UTF8' because a valid ASCII string is a valid |
191 | UTF8 string. |
6662521e |
192 | |
193 | =cut |
194 | */ |
195 | |
8e84507e |
196 | bool |
6662521e |
197 | Perl_is_utf8_string(pTHX_ U8 *s, STRLEN len) |
198 | { |
067a85ef |
199 | U8* x = s; |
1aa99e6b |
200 | U8* send; |
067a85ef |
201 | STRLEN c; |
202 | |
1aa99e6b |
203 | if (!len) |
6cd5fe39 |
204 | len = strlen((char *)s); |
1aa99e6b |
205 | send = s + len; |
206 | |
6662521e |
207 | while (x < send) { |
208 | c = is_utf8_char(x); |
067a85ef |
209 | if (!c) |
210 | return FALSE; |
6662521e |
211 | x += c; |
6662521e |
212 | } |
60006e79 |
213 | if (x != send) |
214 | return FALSE; |
067a85ef |
215 | |
216 | return TRUE; |
6662521e |
217 | } |
218 | |
67e989fb |
219 | /* |
9041c2e3 |
220 | =for apidoc A|UV|utf8n_to_uvuni|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags |
67e989fb |
221 | |
9041c2e3 |
222 | Bottom level UTF-8 decode routine. |
223 | Returns the unicode code point value of the first character in the string C<s> |
ba210ebe |
224 | which is assumed to be in UTF8 encoding and no longer than C<curlen>; |
7df053ec |
225 | C<retlen> will be set to the length, in bytes, of that character. |
67e989fb |
226 | |
227 | If C<s> does not point to a well-formed UTF8 character, the behaviour |
dcad2880 |
228 | is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY, |
229 | it is assumed that the caller will raise a warning, and this function |
28d3d195 |
230 | will silently just set C<retlen> to C<-1> and return zero. If the |
231 | C<flags> does not contain UTF8_CHECK_ONLY, warnings about |
232 | malformations will be given, C<retlen> will be set to the expected |
233 | length of the UTF-8 character in bytes, and zero will be returned. |
234 | |
235 | The C<flags> can also contain various flags to allow deviations from |
236 | the strict UTF-8 encoding (see F<utf8.h>). |
67e989fb |
237 | |
9041c2e3 |
238 | Most code should use utf8_to_uvchr() rather than call this directly. |
239 | |
37607a96 |
240 | =cut |
241 | */ |
67e989fb |
242 | |
a0ed51b3 |
243 | UV |
37607a96 |
244 | Perl_utf8n_to_uvuni(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) |
a0ed51b3 |
245 | { |
9c5ffd7c |
246 | UV uv = *s, ouv = 0; |
ba210ebe |
247 | STRLEN len = 1; |
248 | bool dowarn = ckWARN_d(WARN_UTF8); |
249 | STRLEN expectlen = 0; |
a0dbb045 |
250 | U32 warning = 0; |
251 | |
252 | /* This list is a superset of the UTF8_ALLOW_XXX. */ |
253 | |
254 | #define UTF8_WARN_EMPTY 1 |
255 | #define UTF8_WARN_CONTINUATION 2 |
256 | #define UTF8_WARN_NON_CONTINUATION 3 |
257 | #define UTF8_WARN_FE_FF 4 |
258 | #define UTF8_WARN_SHORT 5 |
259 | #define UTF8_WARN_OVERFLOW 6 |
260 | #define UTF8_WARN_SURROGATE 7 |
261 | #define UTF8_WARN_BOM 8 |
262 | #define UTF8_WARN_LONG 9 |
263 | #define UTF8_WARN_FFFF 10 |
264 | |
265 | if (curlen == 0 && |
266 | !(flags & UTF8_ALLOW_EMPTY)) { |
267 | warning = UTF8_WARN_EMPTY; |
0c443dc2 |
268 | goto malformed; |
269 | } |
270 | |
1d72bdf6 |
271 | if (UTF8_IS_INVARIANT(uv)) { |
a0ed51b3 |
272 | if (retlen) |
273 | *retlen = 1; |
c4d5f83a |
274 | return (UV) (NATIVE_TO_UTF(*s)); |
a0ed51b3 |
275 | } |
67e989fb |
276 | |
421a8bf2 |
277 | if (UTF8_IS_CONTINUATION(uv) && |
fcc8fcf6 |
278 | !(flags & UTF8_ALLOW_CONTINUATION)) { |
a0dbb045 |
279 | warning = UTF8_WARN_CONTINUATION; |
ba210ebe |
280 | goto malformed; |
281 | } |
282 | |
421a8bf2 |
283 | if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) && |
fcc8fcf6 |
284 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { |
a0dbb045 |
285 | warning = UTF8_WARN_NON_CONTINUATION; |
ba210ebe |
286 | goto malformed; |
287 | } |
9041c2e3 |
288 | |
1d72bdf6 |
289 | #ifdef EBCDIC |
75383841 |
290 | uv = NATIVE_TO_UTF(uv); |
1d72bdf6 |
291 | #else |
fcc8fcf6 |
292 | if ((uv == 0xfe || uv == 0xff) && |
293 | !(flags & UTF8_ALLOW_FE_FF)) { |
a0dbb045 |
294 | warning = UTF8_WARN_FE_FF; |
ba210ebe |
295 | goto malformed; |
a0ed51b3 |
296 | } |
1d72bdf6 |
297 | #endif |
298 | |
ba210ebe |
299 | if (!(uv & 0x20)) { len = 2; uv &= 0x1f; } |
300 | else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; } |
301 | else if (!(uv & 0x08)) { len = 4; uv &= 0x07; } |
302 | else if (!(uv & 0x04)) { len = 5; uv &= 0x03; } |
1d72bdf6 |
303 | #ifdef EBCDIC |
304 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } |
305 | else { len = 7; uv &= 0x01; } |
306 | #else |
ba210ebe |
307 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } |
308 | else if (!(uv & 0x01)) { len = 7; uv = 0; } |
1d72bdf6 |
309 | else { len = 13; uv = 0; } /* whoa! */ |
310 | #endif |
311 | |
a0ed51b3 |
312 | if (retlen) |
313 | *retlen = len; |
9041c2e3 |
314 | |
ba210ebe |
315 | expectlen = len; |
316 | |
fcc8fcf6 |
317 | if ((curlen < expectlen) && |
318 | !(flags & UTF8_ALLOW_SHORT)) { |
a0dbb045 |
319 | warning = UTF8_WARN_SHORT; |
ba210ebe |
320 | goto malformed; |
321 | } |
322 | |
323 | len--; |
a0ed51b3 |
324 | s++; |
ba210ebe |
325 | ouv = uv; |
326 | |
a0ed51b3 |
327 | while (len--) { |
421a8bf2 |
328 | if (!UTF8_IS_CONTINUATION(*s) && |
329 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { |
a0dbb045 |
330 | s--; |
331 | warning = UTF8_WARN_NON_CONTINUATION; |
ba210ebe |
332 | goto malformed; |
a0ed51b3 |
333 | } |
334 | else |
8850bf83 |
335 | uv = UTF8_ACCUMULATE(uv, *s); |
a0dbb045 |
336 | if (!(uv > ouv)) { |
337 | /* These cannot be allowed. */ |
338 | if (uv == ouv) { |
339 | if (!(flags & UTF8_ALLOW_LONG)) { |
340 | warning = UTF8_WARN_LONG; |
341 | goto malformed; |
342 | } |
343 | } |
344 | else { /* uv < ouv */ |
345 | /* This cannot be allowed. */ |
346 | warning = UTF8_WARN_OVERFLOW; |
347 | goto malformed; |
348 | } |
ba210ebe |
349 | } |
350 | s++; |
351 | ouv = uv; |
352 | } |
353 | |
421a8bf2 |
354 | if (UNICODE_IS_SURROGATE(uv) && |
fcc8fcf6 |
355 | !(flags & UTF8_ALLOW_SURROGATE)) { |
a0dbb045 |
356 | warning = UTF8_WARN_SURROGATE; |
ba210ebe |
357 | goto malformed; |
421a8bf2 |
358 | } else if (UNICODE_IS_BYTE_ORDER_MARK(uv) && |
fcc8fcf6 |
359 | !(flags & UTF8_ALLOW_BOM)) { |
a0dbb045 |
360 | warning = UTF8_WARN_BOM; |
ba210ebe |
361 | goto malformed; |
fcc8fcf6 |
362 | } else if ((expectlen > UNISKIP(uv)) && |
363 | !(flags & UTF8_ALLOW_LONG)) { |
a0dbb045 |
364 | warning = UTF8_WARN_LONG; |
ba210ebe |
365 | goto malformed; |
421a8bf2 |
366 | } else if (UNICODE_IS_ILLEGAL(uv) && |
a9917092 |
367 | !(flags & UTF8_ALLOW_FFFF)) { |
a0dbb045 |
368 | warning = UTF8_WARN_FFFF; |
a9917092 |
369 | goto malformed; |
a0ed51b3 |
370 | } |
ba210ebe |
371 | |
a0ed51b3 |
372 | return uv; |
ba210ebe |
373 | |
374 | malformed: |
375 | |
fcc8fcf6 |
376 | if (flags & UTF8_CHECK_ONLY) { |
ba210ebe |
377 | if (retlen) |
cc366d4b |
378 | *retlen = -1; |
ba210ebe |
379 | return 0; |
380 | } |
381 | |
a0dbb045 |
382 | if (dowarn) { |
383 | SV* sv = sv_2mortal(newSVpv("Malformed UTF-8 character ", 0)); |
384 | |
385 | switch (warning) { |
386 | case 0: /* Intentionally empty. */ break; |
387 | case UTF8_WARN_EMPTY: |
388 | Perl_sv_catpvf(aTHX_ sv, "(empty string)"); |
389 | break; |
390 | case UTF8_WARN_CONTINUATION: |
391 | Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf")", uv); |
392 | break; |
393 | case UTF8_WARN_NON_CONTINUATION: |
394 | Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf" after start byte 0x%02"UVxf")", |
395 | (UV)s[1], uv); |
396 | break; |
397 | case UTF8_WARN_FE_FF: |
398 | Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv); |
399 | break; |
400 | case UTF8_WARN_SHORT: |
401 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d)", |
402 | curlen, curlen == 1 ? "" : "s", expectlen); |
403 | break; |
404 | case UTF8_WARN_OVERFLOW: |
405 | Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x)", |
406 | ouv, *s); |
407 | break; |
408 | case UTF8_WARN_SURROGATE: |
409 | Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv); |
410 | break; |
411 | case UTF8_WARN_BOM: |
412 | Perl_sv_catpvf(aTHX_ sv, "(byte order mark 0x%04"UVxf")", uv); |
413 | break; |
414 | case UTF8_WARN_LONG: |
415 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d)", |
416 | expectlen, expectlen == 1 ? "": "s", UNISKIP(uv)); |
417 | break; |
418 | case UTF8_WARN_FFFF: |
419 | Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv); |
420 | break; |
421 | default: |
422 | Perl_sv_catpvf(aTHX_ sv, "(unknown reason)"); |
423 | break; |
424 | } |
425 | |
426 | if (warning) { |
427 | char *s = SvPVX(sv); |
428 | |
429 | if (PL_op) |
430 | Perl_warner(aTHX_ WARN_UTF8, |
53e06cf0 |
431 | "%s in %s", s, OP_DESC(PL_op)); |
a0dbb045 |
432 | else |
433 | Perl_warner(aTHX_ WARN_UTF8, "%s", s); |
434 | } |
435 | } |
436 | |
ba210ebe |
437 | if (retlen) |
28d3d195 |
438 | *retlen = expectlen ? expectlen : len; |
ba210ebe |
439 | |
28d3d195 |
440 | return 0; |
a0ed51b3 |
441 | } |
442 | |
8e84507e |
443 | /* |
37607a96 |
444 | =for apidoc A|UV|utf8_to_uvchr|U8 *s|STRLEN *retlen |
9041c2e3 |
445 | |
446 | Returns the native character value of the first character in the string C<s> |
447 | which is assumed to be in UTF8 encoding; C<retlen> will be set to the |
448 | length, in bytes, of that character. |
449 | |
450 | If C<s> does not point to a well-formed UTF8 character, zero is |
451 | returned and retlen is set, if possible, to -1. |
452 | |
453 | =cut |
454 | */ |
455 | |
456 | UV |
37607a96 |
457 | Perl_utf8_to_uvchr(pTHX_ U8 *s, STRLEN *retlen) |
9041c2e3 |
458 | { |
459 | return Perl_utf8n_to_uvchr(aTHX_ s, UTF8_MAXLEN, retlen, 0); |
460 | } |
461 | |
462 | /* |
37607a96 |
463 | =for apidoc A|UV|utf8_to_uvuni|U8 *s|STRLEN *retlen |
9041c2e3 |
464 | |
465 | Returns the Unicode code point of the first character in the string C<s> |
466 | which is assumed to be in UTF8 encoding; C<retlen> will be set to the |
467 | length, in bytes, of that character. |
468 | |
469 | This function should only be used when returned UV is considered |
470 | an index into the Unicode semantic tables (e.g. swashes). |
471 | |
ba210ebe |
472 | If C<s> does not point to a well-formed UTF8 character, zero is |
473 | returned and retlen is set, if possible, to -1. |
8e84507e |
474 | |
475 | =cut |
476 | */ |
477 | |
478 | UV |
37607a96 |
479 | Perl_utf8_to_uvuni(pTHX_ U8 *s, STRLEN *retlen) |
8e84507e |
480 | { |
9041c2e3 |
481 | /* Call the low level routine asking for checks */ |
482 | return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXLEN, retlen, 0); |
8e84507e |
483 | } |
484 | |
b76347f2 |
485 | /* |
37607a96 |
486 | =for apidoc A|STRLEN|utf8_length|U8 *s|U8 *e |
b76347f2 |
487 | |
488 | Return the length of the UTF-8 char encoded string C<s> in characters. |
02eb7b47 |
489 | Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end |
490 | up past C<e>, croaks. |
b76347f2 |
491 | |
492 | =cut |
493 | */ |
494 | |
495 | STRLEN |
37607a96 |
496 | Perl_utf8_length(pTHX_ U8 *s, U8 *e) |
b76347f2 |
497 | { |
498 | STRLEN len = 0; |
499 | |
8850bf83 |
500 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g. |
501 | * the bitops (especially ~) can create illegal UTF-8. |
502 | * In other words: in Perl UTF-8 is not just for Unicode. */ |
503 | |
b76347f2 |
504 | if (e < s) |
02eb7b47 |
505 | Perl_croak(aTHX_ "panic: utf8_length: unexpected end"); |
b76347f2 |
506 | while (s < e) { |
02eb7b47 |
507 | U8 t = UTF8SKIP(s); |
b76347f2 |
508 | |
509 | if (e - s < t) |
0064a8a9 |
510 | Perl_croak(aTHX_ "panic: utf8_length: unaligned end"); |
b76347f2 |
511 | s += t; |
512 | len++; |
513 | } |
514 | |
515 | return len; |
516 | } |
517 | |
b06226ff |
518 | /* |
eebe1485 |
519 | =for apidoc A|IV|utf8_distance|U8 *a|U8 *b |
b06226ff |
520 | |
521 | Returns the number of UTF8 characters between the UTF-8 pointers C<a> |
522 | and C<b>. |
523 | |
524 | WARNING: use only if you *know* that the pointers point inside the |
525 | same UTF-8 buffer. |
526 | |
37607a96 |
527 | =cut |
528 | */ |
a0ed51b3 |
529 | |
02eb7b47 |
530 | IV |
864dbfa3 |
531 | Perl_utf8_distance(pTHX_ U8 *a, U8 *b) |
a0ed51b3 |
532 | { |
02eb7b47 |
533 | IV off = 0; |
534 | |
8850bf83 |
535 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g. |
536 | * the bitops (especially ~) can create illegal UTF-8. |
537 | * In other words: in Perl UTF-8 is not just for Unicode. */ |
538 | |
a0ed51b3 |
539 | if (a < b) { |
540 | while (a < b) { |
02eb7b47 |
541 | U8 c = UTF8SKIP(a); |
542 | |
543 | if (b - a < c) |
544 | Perl_croak(aTHX_ "panic: utf8_distance: unaligned end"); |
545 | a += c; |
a0ed51b3 |
546 | off--; |
547 | } |
548 | } |
549 | else { |
550 | while (b < a) { |
02eb7b47 |
551 | U8 c = UTF8SKIP(b); |
552 | |
553 | if (a - b < c) |
554 | Perl_croak(aTHX_ "panic: utf8_distance: unaligned end"); |
555 | b += c; |
a0ed51b3 |
556 | off++; |
557 | } |
558 | } |
02eb7b47 |
559 | |
a0ed51b3 |
560 | return off; |
561 | } |
562 | |
b06226ff |
563 | /* |
37607a96 |
564 | =for apidoc A|U8 *|utf8_hop|U8 *s|I32 off |
b06226ff |
565 | |
8850bf83 |
566 | Return the UTF-8 pointer C<s> displaced by C<off> characters, either |
567 | forward or backward. |
b06226ff |
568 | |
569 | WARNING: do not use the following unless you *know* C<off> is within |
8850bf83 |
570 | the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned |
571 | on the first byte of character or just after the last byte of a character. |
b06226ff |
572 | |
37607a96 |
573 | =cut |
574 | */ |
a0ed51b3 |
575 | |
576 | U8 * |
864dbfa3 |
577 | Perl_utf8_hop(pTHX_ U8 *s, I32 off) |
a0ed51b3 |
578 | { |
8850bf83 |
579 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g |
580 | * the bitops (especially ~) can create illegal UTF-8. |
581 | * In other words: in Perl UTF-8 is not just for Unicode. */ |
582 | |
a0ed51b3 |
583 | if (off >= 0) { |
584 | while (off--) |
585 | s += UTF8SKIP(s); |
586 | } |
587 | else { |
588 | while (off++) { |
589 | s--; |
8850bf83 |
590 | while (UTF8_IS_CONTINUATION(*s)) |
591 | s--; |
a0ed51b3 |
592 | } |
593 | } |
594 | return s; |
595 | } |
596 | |
6940069f |
597 | /* |
eebe1485 |
598 | =for apidoc A|U8 *|utf8_to_bytes|U8 *s|STRLEN *len |
6940069f |
599 | |
246fae53 |
600 | Converts a string C<s> of length C<len> from UTF8 into byte encoding. |
601 | Unlike C<bytes_to_utf8>, this over-writes the original string, and |
602 | updates len to contain the new length. |
67e989fb |
603 | Returns zero on failure, setting C<len> to -1. |
6940069f |
604 | |
605 | =cut |
606 | */ |
607 | |
608 | U8 * |
37607a96 |
609 | Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len) |
6940069f |
610 | { |
6940069f |
611 | U8 *send; |
612 | U8 *d; |
dcad2880 |
613 | U8 *save = s; |
246fae53 |
614 | |
615 | /* ensure valid UTF8 and chars < 256 before updating string */ |
dcad2880 |
616 | for (send = s + *len; s < send; ) { |
617 | U8 c = *s++; |
618 | |
1d72bdf6 |
619 | if (!UTF8_IS_INVARIANT(c) && |
620 | (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send) |
621 | || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) { |
dcad2880 |
622 | *len = -1; |
623 | return 0; |
624 | } |
246fae53 |
625 | } |
dcad2880 |
626 | |
627 | d = s = save; |
6940069f |
628 | while (s < send) { |
ed646e6e |
629 | STRLEN ulen; |
9041c2e3 |
630 | *d++ = (U8)utf8_to_uvchr(s, &ulen); |
ed646e6e |
631 | s += ulen; |
6940069f |
632 | } |
633 | *d = '\0'; |
246fae53 |
634 | *len = d - save; |
6940069f |
635 | return save; |
636 | } |
637 | |
638 | /* |
f9a63242 |
639 | =for apidoc A|U8 *|bytes_from_utf8|U8 *s|STRLEN *len|bool *is_utf8 |
640 | |
641 | Converts a string C<s> of length C<len> from UTF8 into byte encoding. |
642 | Unlike <utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to |
ef9edfd0 |
643 | the newly-created string, and updates C<len> to contain the new |
644 | length. Returns the original string if no conversion occurs, C<len> |
645 | is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to |
646 | 0 if C<s> is converted or contains all 7bit characters. |
f9a63242 |
647 | |
37607a96 |
648 | =cut |
649 | */ |
f9a63242 |
650 | |
651 | U8 * |
37607a96 |
652 | Perl_bytes_from_utf8(pTHX_ U8 *s, STRLEN *len, bool *is_utf8) |
f9a63242 |
653 | { |
f9a63242 |
654 | U8 *d; |
655 | U8 *start = s; |
db42d148 |
656 | U8 *send; |
f9a63242 |
657 | I32 count = 0; |
658 | |
659 | if (!*is_utf8) |
660 | return start; |
661 | |
ef9edfd0 |
662 | /* ensure valid UTF8 and chars < 256 before converting string */ |
f9a63242 |
663 | for (send = s + *len; s < send;) { |
664 | U8 c = *s++; |
1d72bdf6 |
665 | if (!UTF8_IS_INVARIANT(c)) { |
db42d148 |
666 | if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send && |
667 | (c = *s++) && UTF8_IS_CONTINUATION(c)) |
668 | count++; |
669 | else |
f9a63242 |
670 | return start; |
db42d148 |
671 | } |
f9a63242 |
672 | } |
673 | |
674 | *is_utf8 = 0; |
675 | |
f9a63242 |
676 | Newz(801, d, (*len) - count + 1, U8); |
ef9edfd0 |
677 | s = start; start = d; |
f9a63242 |
678 | while (s < send) { |
679 | U8 c = *s++; |
c4d5f83a |
680 | if (!UTF8_IS_INVARIANT(c)) { |
681 | /* Then it is two-byte encoded */ |
682 | c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++); |
683 | c = ASCII_TO_NATIVE(c); |
684 | } |
685 | *d++ = c; |
f9a63242 |
686 | } |
687 | *d = '\0'; |
688 | *len = d - start; |
689 | return start; |
690 | } |
691 | |
692 | /* |
eebe1485 |
693 | =for apidoc A|U8 *|bytes_to_utf8|U8 *s|STRLEN *len |
6940069f |
694 | |
695 | Converts a string C<s> of length C<len> from ASCII into UTF8 encoding. |
6662521e |
696 | Returns a pointer to the newly-created string, and sets C<len> to |
697 | reflect the new length. |
6940069f |
698 | |
497711e7 |
699 | =cut |
6940069f |
700 | */ |
701 | |
702 | U8* |
37607a96 |
703 | Perl_bytes_to_utf8(pTHX_ U8 *s, STRLEN *len) |
6940069f |
704 | { |
6940069f |
705 | U8 *send; |
706 | U8 *d; |
707 | U8 *dst; |
6662521e |
708 | send = s + (*len); |
6940069f |
709 | |
6662521e |
710 | Newz(801, d, (*len) * 2 + 1, U8); |
6940069f |
711 | dst = d; |
712 | |
713 | while (s < send) { |
db42d148 |
714 | UV uv = NATIVE_TO_ASCII(*s++); |
c4d5f83a |
715 | if (UNI_IS_INVARIANT(uv)) |
716 | *d++ = UTF_TO_NATIVE(uv); |
6940069f |
717 | else { |
90f44359 |
718 | *d++ = UTF8_EIGHT_BIT_HI(uv); |
719 | *d++ = UTF8_EIGHT_BIT_LO(uv); |
6940069f |
720 | } |
721 | } |
722 | *d = '\0'; |
6662521e |
723 | *len = d-dst; |
6940069f |
724 | return dst; |
725 | } |
726 | |
a0ed51b3 |
727 | /* |
dea0fc0b |
728 | * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8. |
a0ed51b3 |
729 | * |
730 | * Destination must be pre-extended to 3/2 source. Do not use in-place. |
731 | * We optimize for native, for obvious reasons. */ |
732 | |
733 | U8* |
dea0fc0b |
734 | Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 |
735 | { |
dea0fc0b |
736 | U8* pend; |
737 | U8* dstart = d; |
738 | |
739 | if (bytelen & 1) |
a7867d0a |
740 | Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen"); |
dea0fc0b |
741 | |
742 | pend = p + bytelen; |
743 | |
a0ed51b3 |
744 | while (p < pend) { |
dea0fc0b |
745 | UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */ |
746 | p += 2; |
a0ed51b3 |
747 | if (uv < 0x80) { |
748 | *d++ = uv; |
749 | continue; |
750 | } |
751 | if (uv < 0x800) { |
752 | *d++ = (( uv >> 6) | 0xc0); |
753 | *d++ = (( uv & 0x3f) | 0x80); |
754 | continue; |
755 | } |
756 | if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */ |
dea0fc0b |
757 | UV low = *p++; |
758 | if (low < 0xdc00 || low >= 0xdfff) |
759 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
a0ed51b3 |
760 | uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000; |
761 | } |
762 | if (uv < 0x10000) { |
763 | *d++ = (( uv >> 12) | 0xe0); |
764 | *d++ = (((uv >> 6) & 0x3f) | 0x80); |
765 | *d++ = (( uv & 0x3f) | 0x80); |
766 | continue; |
767 | } |
768 | else { |
769 | *d++ = (( uv >> 18) | 0xf0); |
770 | *d++ = (((uv >> 12) & 0x3f) | 0x80); |
771 | *d++ = (((uv >> 6) & 0x3f) | 0x80); |
772 | *d++ = (( uv & 0x3f) | 0x80); |
773 | continue; |
774 | } |
775 | } |
dea0fc0b |
776 | *newlen = d - dstart; |
a0ed51b3 |
777 | return d; |
778 | } |
779 | |
780 | /* Note: this one is slightly destructive of the source. */ |
781 | |
782 | U8* |
dea0fc0b |
783 | Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 |
784 | { |
785 | U8* s = (U8*)p; |
786 | U8* send = s + bytelen; |
787 | while (s < send) { |
788 | U8 tmp = s[0]; |
789 | s[0] = s[1]; |
790 | s[1] = tmp; |
791 | s += 2; |
792 | } |
dea0fc0b |
793 | return utf16_to_utf8(p, d, bytelen, newlen); |
a0ed51b3 |
794 | } |
795 | |
796 | /* for now these are all defined (inefficiently) in terms of the utf8 versions */ |
797 | |
798 | bool |
84afefe6 |
799 | Perl_is_uni_alnum(pTHX_ UV c) |
a0ed51b3 |
800 | { |
ad391ad9 |
801 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
802 | uvchr_to_utf8(tmpbuf, (UV)c); |
a0ed51b3 |
803 | return is_utf8_alnum(tmpbuf); |
804 | } |
805 | |
806 | bool |
84afefe6 |
807 | Perl_is_uni_alnumc(pTHX_ UV c) |
b8c5462f |
808 | { |
ad391ad9 |
809 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
810 | uvchr_to_utf8(tmpbuf, (UV)c); |
b8c5462f |
811 | return is_utf8_alnumc(tmpbuf); |
812 | } |
813 | |
814 | bool |
84afefe6 |
815 | Perl_is_uni_idfirst(pTHX_ UV c) |
a0ed51b3 |
816 | { |
ad391ad9 |
817 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
818 | uvchr_to_utf8(tmpbuf, (UV)c); |
a0ed51b3 |
819 | return is_utf8_idfirst(tmpbuf); |
820 | } |
821 | |
822 | bool |
84afefe6 |
823 | Perl_is_uni_alpha(pTHX_ UV c) |
a0ed51b3 |
824 | { |
ad391ad9 |
825 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
826 | uvchr_to_utf8(tmpbuf, (UV)c); |
a0ed51b3 |
827 | return is_utf8_alpha(tmpbuf); |
828 | } |
829 | |
830 | bool |
84afefe6 |
831 | Perl_is_uni_ascii(pTHX_ UV c) |
4d61ec05 |
832 | { |
ad391ad9 |
833 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
834 | uvchr_to_utf8(tmpbuf, (UV)c); |
4d61ec05 |
835 | return is_utf8_ascii(tmpbuf); |
836 | } |
837 | |
838 | bool |
84afefe6 |
839 | Perl_is_uni_space(pTHX_ UV c) |
a0ed51b3 |
840 | { |
ad391ad9 |
841 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
842 | uvchr_to_utf8(tmpbuf, (UV)c); |
a0ed51b3 |
843 | return is_utf8_space(tmpbuf); |
844 | } |
845 | |
846 | bool |
84afefe6 |
847 | Perl_is_uni_digit(pTHX_ UV c) |
a0ed51b3 |
848 | { |
ad391ad9 |
849 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
850 | uvchr_to_utf8(tmpbuf, (UV)c); |
a0ed51b3 |
851 | return is_utf8_digit(tmpbuf); |
852 | } |
853 | |
854 | bool |
84afefe6 |
855 | Perl_is_uni_upper(pTHX_ UV c) |
a0ed51b3 |
856 | { |
ad391ad9 |
857 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
858 | uvchr_to_utf8(tmpbuf, (UV)c); |
a0ed51b3 |
859 | return is_utf8_upper(tmpbuf); |
860 | } |
861 | |
862 | bool |
84afefe6 |
863 | Perl_is_uni_lower(pTHX_ UV c) |
a0ed51b3 |
864 | { |
ad391ad9 |
865 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
866 | uvchr_to_utf8(tmpbuf, (UV)c); |
a0ed51b3 |
867 | return is_utf8_lower(tmpbuf); |
868 | } |
869 | |
870 | bool |
84afefe6 |
871 | Perl_is_uni_cntrl(pTHX_ UV c) |
b8c5462f |
872 | { |
ad391ad9 |
873 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
874 | uvchr_to_utf8(tmpbuf, (UV)c); |
b8c5462f |
875 | return is_utf8_cntrl(tmpbuf); |
876 | } |
877 | |
878 | bool |
84afefe6 |
879 | Perl_is_uni_graph(pTHX_ UV c) |
b8c5462f |
880 | { |
ad391ad9 |
881 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
882 | uvchr_to_utf8(tmpbuf, (UV)c); |
b8c5462f |
883 | return is_utf8_graph(tmpbuf); |
884 | } |
885 | |
886 | bool |
84afefe6 |
887 | Perl_is_uni_print(pTHX_ UV c) |
a0ed51b3 |
888 | { |
ad391ad9 |
889 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
890 | uvchr_to_utf8(tmpbuf, (UV)c); |
a0ed51b3 |
891 | return is_utf8_print(tmpbuf); |
892 | } |
893 | |
b8c5462f |
894 | bool |
84afefe6 |
895 | Perl_is_uni_punct(pTHX_ UV c) |
b8c5462f |
896 | { |
ad391ad9 |
897 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
898 | uvchr_to_utf8(tmpbuf, (UV)c); |
b8c5462f |
899 | return is_utf8_punct(tmpbuf); |
900 | } |
901 | |
4d61ec05 |
902 | bool |
84afefe6 |
903 | Perl_is_uni_xdigit(pTHX_ UV c) |
4d61ec05 |
904 | { |
a2a2844f |
905 | U8 tmpbuf[UTF8_MAXLEN*2+1]; |
2b9d42f0 |
906 | uvchr_to_utf8(tmpbuf, (UV)c); |
4d61ec05 |
907 | return is_utf8_xdigit(tmpbuf); |
908 | } |
909 | |
84afefe6 |
910 | UV |
911 | Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
912 | { |
a2a2844f |
913 | U8 tmpbuf[UTF8_MAXLEN*2+1]; |
2b9d42f0 |
914 | uvchr_to_utf8(tmpbuf, (UV)c); |
a2a2844f |
915 | return to_utf8_upper(tmpbuf, p, lenp); |
a0ed51b3 |
916 | } |
917 | |
84afefe6 |
918 | UV |
919 | Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
920 | { |
a2a2844f |
921 | U8 tmpbuf[UTF8_MAXLEN*2+1]; |
2b9d42f0 |
922 | uvchr_to_utf8(tmpbuf, (UV)c); |
a2a2844f |
923 | return to_utf8_title(tmpbuf, p, lenp); |
a0ed51b3 |
924 | } |
925 | |
84afefe6 |
926 | UV |
927 | Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
928 | { |
ad391ad9 |
929 | U8 tmpbuf[UTF8_MAXLEN+1]; |
2b9d42f0 |
930 | uvchr_to_utf8(tmpbuf, (UV)c); |
a2a2844f |
931 | return to_utf8_lower(tmpbuf, p, lenp); |
a0ed51b3 |
932 | } |
933 | |
84afefe6 |
934 | UV |
935 | Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp) |
936 | { |
937 | U8 tmpbuf[UTF8_MAXLEN+1]; |
938 | uvchr_to_utf8(tmpbuf, (UV)c); |
939 | return to_utf8_fold(tmpbuf, p, lenp); |
940 | } |
941 | |
a0ed51b3 |
942 | /* for now these all assume no locale info available for Unicode > 255 */ |
943 | |
944 | bool |
84afefe6 |
945 | Perl_is_uni_alnum_lc(pTHX_ UV c) |
a0ed51b3 |
946 | { |
947 | return is_uni_alnum(c); /* XXX no locale support yet */ |
948 | } |
949 | |
950 | bool |
84afefe6 |
951 | Perl_is_uni_alnumc_lc(pTHX_ UV c) |
b8c5462f |
952 | { |
953 | return is_uni_alnumc(c); /* XXX no locale support yet */ |
954 | } |
955 | |
956 | bool |
84afefe6 |
957 | Perl_is_uni_idfirst_lc(pTHX_ UV c) |
a0ed51b3 |
958 | { |
959 | return is_uni_idfirst(c); /* XXX no locale support yet */ |
960 | } |
961 | |
962 | bool |
84afefe6 |
963 | Perl_is_uni_alpha_lc(pTHX_ UV c) |
a0ed51b3 |
964 | { |
965 | return is_uni_alpha(c); /* XXX no locale support yet */ |
966 | } |
967 | |
968 | bool |
84afefe6 |
969 | Perl_is_uni_ascii_lc(pTHX_ UV c) |
4d61ec05 |
970 | { |
971 | return is_uni_ascii(c); /* XXX no locale support yet */ |
972 | } |
973 | |
974 | bool |
84afefe6 |
975 | Perl_is_uni_space_lc(pTHX_ UV c) |
a0ed51b3 |
976 | { |
977 | return is_uni_space(c); /* XXX no locale support yet */ |
978 | } |
979 | |
980 | bool |
84afefe6 |
981 | Perl_is_uni_digit_lc(pTHX_ UV c) |
a0ed51b3 |
982 | { |
983 | return is_uni_digit(c); /* XXX no locale support yet */ |
984 | } |
985 | |
986 | bool |
84afefe6 |
987 | Perl_is_uni_upper_lc(pTHX_ UV c) |
a0ed51b3 |
988 | { |
989 | return is_uni_upper(c); /* XXX no locale support yet */ |
990 | } |
991 | |
992 | bool |
84afefe6 |
993 | Perl_is_uni_lower_lc(pTHX_ UV c) |
a0ed51b3 |
994 | { |
995 | return is_uni_lower(c); /* XXX no locale support yet */ |
996 | } |
997 | |
998 | bool |
84afefe6 |
999 | Perl_is_uni_cntrl_lc(pTHX_ UV c) |
b8c5462f |
1000 | { |
1001 | return is_uni_cntrl(c); /* XXX no locale support yet */ |
1002 | } |
1003 | |
1004 | bool |
84afefe6 |
1005 | Perl_is_uni_graph_lc(pTHX_ UV c) |
b8c5462f |
1006 | { |
1007 | return is_uni_graph(c); /* XXX no locale support yet */ |
1008 | } |
1009 | |
1010 | bool |
84afefe6 |
1011 | Perl_is_uni_print_lc(pTHX_ UV c) |
a0ed51b3 |
1012 | { |
1013 | return is_uni_print(c); /* XXX no locale support yet */ |
1014 | } |
1015 | |
b8c5462f |
1016 | bool |
84afefe6 |
1017 | Perl_is_uni_punct_lc(pTHX_ UV c) |
b8c5462f |
1018 | { |
1019 | return is_uni_punct(c); /* XXX no locale support yet */ |
1020 | } |
1021 | |
4d61ec05 |
1022 | bool |
84afefe6 |
1023 | Perl_is_uni_xdigit_lc(pTHX_ UV c) |
4d61ec05 |
1024 | { |
1025 | return is_uni_xdigit(c); /* XXX no locale support yet */ |
1026 | } |
1027 | |
a0ed51b3 |
1028 | bool |
864dbfa3 |
1029 | Perl_is_utf8_alnum(pTHX_ U8 *p) |
a0ed51b3 |
1030 | { |
386d01d6 |
1031 | if (!is_utf8_char(p)) |
1032 | return FALSE; |
a0ed51b3 |
1033 | if (!PL_utf8_alnum) |
289d4f09 |
1034 | /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true |
1035 | * descendant of isalnum(3), in other words, it doesn't |
1036 | * contain the '_'. --jhi */ |
1037 | PL_utf8_alnum = swash_init("utf8", "IsWord", &PL_sv_undef, 0, 0); |
3568d838 |
1038 | return swash_fetch(PL_utf8_alnum, p, TRUE); |
a0ed51b3 |
1039 | /* return *p == '_' || is_utf8_alpha(p) || is_utf8_digit(p); */ |
1040 | #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */ |
1041 | if (!PL_utf8_alnum) |
1042 | PL_utf8_alnum = swash_init("utf8", "", |
1043 | sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0); |
3568d838 |
1044 | return swash_fetch(PL_utf8_alnum, p, TRUE); |
a0ed51b3 |
1045 | #endif |
1046 | } |
1047 | |
1048 | bool |
b8c5462f |
1049 | Perl_is_utf8_alnumc(pTHX_ U8 *p) |
1050 | { |
386d01d6 |
1051 | if (!is_utf8_char(p)) |
1052 | return FALSE; |
b8c5462f |
1053 | if (!PL_utf8_alnum) |
1054 | PL_utf8_alnum = swash_init("utf8", "IsAlnumC", &PL_sv_undef, 0, 0); |
3568d838 |
1055 | return swash_fetch(PL_utf8_alnum, p, TRUE); |
b8c5462f |
1056 | /* return is_utf8_alpha(p) || is_utf8_digit(p); */ |
1057 | #ifdef SURPRISINGLY_SLOWER /* probably because alpha is usually true */ |
1058 | if (!PL_utf8_alnum) |
1059 | PL_utf8_alnum = swash_init("utf8", "", |
1060 | sv_2mortal(newSVpv("+utf8::IsAlpha\n+utf8::IsDigit\n005F\n",0)), 0, 0); |
3568d838 |
1061 | return swash_fetch(PL_utf8_alnum, p, TRUE); |
b8c5462f |
1062 | #endif |
1063 | } |
1064 | |
1065 | bool |
864dbfa3 |
1066 | Perl_is_utf8_idfirst(pTHX_ U8 *p) |
a0ed51b3 |
1067 | { |
1068 | return *p == '_' || is_utf8_alpha(p); |
1069 | } |
1070 | |
1071 | bool |
864dbfa3 |
1072 | Perl_is_utf8_alpha(pTHX_ U8 *p) |
a0ed51b3 |
1073 | { |
386d01d6 |
1074 | if (!is_utf8_char(p)) |
1075 | return FALSE; |
a0ed51b3 |
1076 | if (!PL_utf8_alpha) |
e24b16f9 |
1077 | PL_utf8_alpha = swash_init("utf8", "IsAlpha", &PL_sv_undef, 0, 0); |
3568d838 |
1078 | return swash_fetch(PL_utf8_alpha, p, TRUE); |
a0ed51b3 |
1079 | } |
1080 | |
1081 | bool |
b8c5462f |
1082 | Perl_is_utf8_ascii(pTHX_ U8 *p) |
1083 | { |
386d01d6 |
1084 | if (!is_utf8_char(p)) |
1085 | return FALSE; |
b8c5462f |
1086 | if (!PL_utf8_ascii) |
1087 | PL_utf8_ascii = swash_init("utf8", "IsAscii", &PL_sv_undef, 0, 0); |
3568d838 |
1088 | return swash_fetch(PL_utf8_ascii, p, TRUE); |
b8c5462f |
1089 | } |
1090 | |
1091 | bool |
864dbfa3 |
1092 | Perl_is_utf8_space(pTHX_ U8 *p) |
a0ed51b3 |
1093 | { |
386d01d6 |
1094 | if (!is_utf8_char(p)) |
1095 | return FALSE; |
a0ed51b3 |
1096 | if (!PL_utf8_space) |
3bec3564 |
1097 | PL_utf8_space = swash_init("utf8", "IsSpacePerl", &PL_sv_undef, 0, 0); |
3568d838 |
1098 | return swash_fetch(PL_utf8_space, p, TRUE); |
a0ed51b3 |
1099 | } |
1100 | |
1101 | bool |
864dbfa3 |
1102 | Perl_is_utf8_digit(pTHX_ U8 *p) |
a0ed51b3 |
1103 | { |
386d01d6 |
1104 | if (!is_utf8_char(p)) |
1105 | return FALSE; |
a0ed51b3 |
1106 | if (!PL_utf8_digit) |
e24b16f9 |
1107 | PL_utf8_digit = swash_init("utf8", "IsDigit", &PL_sv_undef, 0, 0); |
3568d838 |
1108 | return swash_fetch(PL_utf8_digit, p, TRUE); |
a0ed51b3 |
1109 | } |
1110 | |
1111 | bool |
864dbfa3 |
1112 | Perl_is_utf8_upper(pTHX_ U8 *p) |
a0ed51b3 |
1113 | { |
386d01d6 |
1114 | if (!is_utf8_char(p)) |
1115 | return FALSE; |
a0ed51b3 |
1116 | if (!PL_utf8_upper) |
e24b16f9 |
1117 | PL_utf8_upper = swash_init("utf8", "IsUpper", &PL_sv_undef, 0, 0); |
3568d838 |
1118 | return swash_fetch(PL_utf8_upper, p, TRUE); |
a0ed51b3 |
1119 | } |
1120 | |
1121 | bool |
864dbfa3 |
1122 | Perl_is_utf8_lower(pTHX_ U8 *p) |
a0ed51b3 |
1123 | { |
386d01d6 |
1124 | if (!is_utf8_char(p)) |
1125 | return FALSE; |
a0ed51b3 |
1126 | if (!PL_utf8_lower) |
e24b16f9 |
1127 | PL_utf8_lower = swash_init("utf8", "IsLower", &PL_sv_undef, 0, 0); |
3568d838 |
1128 | return swash_fetch(PL_utf8_lower, p, TRUE); |
a0ed51b3 |
1129 | } |
1130 | |
1131 | bool |
b8c5462f |
1132 | Perl_is_utf8_cntrl(pTHX_ U8 *p) |
1133 | { |
386d01d6 |
1134 | if (!is_utf8_char(p)) |
1135 | return FALSE; |
b8c5462f |
1136 | if (!PL_utf8_cntrl) |
1137 | PL_utf8_cntrl = swash_init("utf8", "IsCntrl", &PL_sv_undef, 0, 0); |
3568d838 |
1138 | return swash_fetch(PL_utf8_cntrl, p, TRUE); |
b8c5462f |
1139 | } |
1140 | |
1141 | bool |
1142 | Perl_is_utf8_graph(pTHX_ U8 *p) |
1143 | { |
386d01d6 |
1144 | if (!is_utf8_char(p)) |
1145 | return FALSE; |
b8c5462f |
1146 | if (!PL_utf8_graph) |
1147 | PL_utf8_graph = swash_init("utf8", "IsGraph", &PL_sv_undef, 0, 0); |
3568d838 |
1148 | return swash_fetch(PL_utf8_graph, p, TRUE); |
b8c5462f |
1149 | } |
1150 | |
1151 | bool |
864dbfa3 |
1152 | Perl_is_utf8_print(pTHX_ U8 *p) |
a0ed51b3 |
1153 | { |
386d01d6 |
1154 | if (!is_utf8_char(p)) |
1155 | return FALSE; |
a0ed51b3 |
1156 | if (!PL_utf8_print) |
e24b16f9 |
1157 | PL_utf8_print = swash_init("utf8", "IsPrint", &PL_sv_undef, 0, 0); |
3568d838 |
1158 | return swash_fetch(PL_utf8_print, p, TRUE); |
a0ed51b3 |
1159 | } |
1160 | |
1161 | bool |
b8c5462f |
1162 | Perl_is_utf8_punct(pTHX_ U8 *p) |
1163 | { |
386d01d6 |
1164 | if (!is_utf8_char(p)) |
1165 | return FALSE; |
b8c5462f |
1166 | if (!PL_utf8_punct) |
1167 | PL_utf8_punct = swash_init("utf8", "IsPunct", &PL_sv_undef, 0, 0); |
3568d838 |
1168 | return swash_fetch(PL_utf8_punct, p, TRUE); |
b8c5462f |
1169 | } |
1170 | |
1171 | bool |
1172 | Perl_is_utf8_xdigit(pTHX_ U8 *p) |
1173 | { |
386d01d6 |
1174 | if (!is_utf8_char(p)) |
1175 | return FALSE; |
b8c5462f |
1176 | if (!PL_utf8_xdigit) |
1177 | PL_utf8_xdigit = swash_init("utf8", "IsXDigit", &PL_sv_undef, 0, 0); |
3568d838 |
1178 | return swash_fetch(PL_utf8_xdigit, p, TRUE); |
b8c5462f |
1179 | } |
1180 | |
1181 | bool |
864dbfa3 |
1182 | Perl_is_utf8_mark(pTHX_ U8 *p) |
a0ed51b3 |
1183 | { |
386d01d6 |
1184 | if (!is_utf8_char(p)) |
1185 | return FALSE; |
a0ed51b3 |
1186 | if (!PL_utf8_mark) |
e24b16f9 |
1187 | PL_utf8_mark = swash_init("utf8", "IsM", &PL_sv_undef, 0, 0); |
3568d838 |
1188 | return swash_fetch(PL_utf8_mark, p, TRUE); |
a0ed51b3 |
1189 | } |
1190 | |
6b5c0936 |
1191 | /* |
1192 | =for apidoc A|UV|to_utf8_case|U8 *p|U8* ustrp|STRLEN *lenp|SV **swash|char *normal|char *special |
1193 | |
1194 | The "p" contains the pointer to the UTF-8 string encoding |
1195 | the character that is being converted. |
1196 | |
1197 | The "ustrp" is a pointer to the character buffer to put the |
1198 | conversion result to. The "lenp" is a pointer to the length |
1199 | of the result. |
1200 | |
1201 | The "swash" is a pointer to the swash to use. |
1202 | |
1203 | The "normal" is a string like "ToLower" which means the swash |
1204 | $utf8::ToLower, which is stored in lib/unicore/To/Lower.pl, |
1205 | and loaded by SWASHGET, using lib/utf8_heavy.pl. |
1206 | |
1207 | The "special" is a string like "utf8::ToSpecLower", which means |
d1be9408 |
1208 | the hash %utf8::ToSpecLower, which is stored in the same file, |
6b5c0936 |
1209 | lib/unicore/To/Lower.pl, and also loaded by SWASHGET. The access |
1210 | to the hash is by Perl_to_utf8_case(). |
1211 | |
1212 | =cut |
1213 | */ |
1214 | |
2104c8d9 |
1215 | UV |
983ffd37 |
1216 | Perl_to_utf8_case(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp, SV **swashp,char *normal, char *special) |
a0ed51b3 |
1217 | { |
1218 | UV uv; |
1219 | |
983ffd37 |
1220 | if (!*swashp) |
1221 | *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0); |
1222 | uv = swash_fetch(*swashp, p, TRUE); |
1223 | if (uv) |
1224 | uv = UNI_TO_NATIVE(uv); |
1225 | else { |
1226 | HV *hv; |
1227 | SV *keysv; |
1228 | HE *he; |
1229 | |
1230 | uv = utf8_to_uvchr(p, 0); |
1231 | |
1232 | if ((hv = get_hv(special, FALSE)) && |
6d47b937 |
1233 | (keysv = sv_2mortal(Perl_newSVpvf(aTHX_ "%04"UVXf, uv))) && |
983ffd37 |
1234 | (he = hv_fetch_ent(hv, keysv, FALSE, 0))) { |
1235 | SV *val = HeVAL(he); |
1236 | char *s = SvPV(val, *lenp); |
1237 | U8 c = *(U8*)s; |
1238 | if (*lenp > 1 || UNI_IS_INVARIANT(c)) |
1239 | Copy(s, ustrp, *lenp, U8); |
1240 | else { |
1241 | /* something in the 0x80..0xFF range */ |
1242 | ustrp[0] = UTF8_EIGHT_BIT_HI(c); |
1243 | ustrp[1] = UTF8_EIGHT_BIT_LO(c); |
1244 | *lenp = 2; |
1245 | } |
1246 | return 0; |
1247 | } |
1248 | } |
a2a2844f |
1249 | *lenp = UNISKIP(uv); |
1250 | uvuni_to_utf8(ustrp, uv); |
1251 | return uv; |
a0ed51b3 |
1252 | } |
1253 | |
2104c8d9 |
1254 | UV |
983ffd37 |
1255 | Perl_to_utf8_upper(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 |
1256 | { |
983ffd37 |
1257 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1258 | &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper"); |
983ffd37 |
1259 | } |
a0ed51b3 |
1260 | |
983ffd37 |
1261 | UV |
1262 | Perl_to_utf8_title(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) |
1263 | { |
1264 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1265 | &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle"); |
a0ed51b3 |
1266 | } |
1267 | |
2104c8d9 |
1268 | UV |
a2a2844f |
1269 | Perl_to_utf8_lower(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 |
1270 | { |
983ffd37 |
1271 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1272 | &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower"); |
1273 | } |
1274 | |
1275 | UV |
1276 | Perl_to_utf8_fold(pTHX_ U8 *p, U8* ustrp, STRLEN *lenp) |
1277 | { |
1278 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
1279 | &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold"); |
a0ed51b3 |
1280 | } |
1281 | |
1282 | /* a "swash" is a swatch hash */ |
1283 | |
1284 | SV* |
864dbfa3 |
1285 | Perl_swash_init(pTHX_ char* pkg, char* name, SV *listsv, I32 minbits, I32 none) |
a0ed51b3 |
1286 | { |
1287 | SV* retval; |
bf1fed83 |
1288 | SV* tokenbufsv = sv_2mortal(NEWSV(0,0)); |
8e84507e |
1289 | dSP; |
1b026014 |
1290 | HV *stash = gv_stashpvn(pkg, strlen(pkg), FALSE); |
f8be5cf0 |
1291 | SV* errsv_save; |
ce3b816e |
1292 | |
1b026014 |
1293 | if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */ |
ce3b816e |
1294 | ENTER; |
f8be5cf0 |
1295 | errsv_save = newSVsv(ERRSV); |
ce3b816e |
1296 | Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpv(pkg,0), Nullsv); |
f8be5cf0 |
1297 | if (!SvTRUE(ERRSV)) |
1298 | sv_setsv(ERRSV, errsv_save); |
1299 | SvREFCNT_dec(errsv_save); |
ce3b816e |
1300 | LEAVE; |
1301 | } |
1302 | SPAGAIN; |
a0ed51b3 |
1303 | PUSHSTACKi(PERLSI_MAGIC); |
1304 | PUSHMARK(SP); |
1305 | EXTEND(SP,5); |
1306 | PUSHs(sv_2mortal(newSVpvn(pkg, strlen(pkg)))); |
1307 | PUSHs(sv_2mortal(newSVpvn(name, strlen(name)))); |
1308 | PUSHs(listsv); |
1309 | PUSHs(sv_2mortal(newSViv(minbits))); |
1310 | PUSHs(sv_2mortal(newSViv(none))); |
1311 | PUTBACK; |
1312 | ENTER; |
1313 | SAVEI32(PL_hints); |
1314 | PL_hints = 0; |
1315 | save_re_context(); |
bf1fed83 |
1316 | if (PL_curcop == &PL_compiling) |
1317 | /* XXX ought to be handled by lex_start */ |
1318 | sv_setpv(tokenbufsv, PL_tokenbuf); |
f8be5cf0 |
1319 | errsv_save = newSVsv(ERRSV); |
864dbfa3 |
1320 | if (call_method("SWASHNEW", G_SCALAR)) |
8e84507e |
1321 | retval = newSVsv(*PL_stack_sp--); |
a0ed51b3 |
1322 | else |
e24b16f9 |
1323 | retval = &PL_sv_undef; |
f8be5cf0 |
1324 | if (!SvTRUE(ERRSV)) |
1325 | sv_setsv(ERRSV, errsv_save); |
1326 | SvREFCNT_dec(errsv_save); |
a0ed51b3 |
1327 | LEAVE; |
1328 | POPSTACK; |
e24b16f9 |
1329 | if (PL_curcop == &PL_compiling) { |
bf1fed83 |
1330 | STRLEN len; |
1331 | char* pv = SvPV(tokenbufsv, len); |
1332 | |
1333 | Copy(pv, PL_tokenbuf, len+1, char); |
e24b16f9 |
1334 | PL_curcop->op_private = PL_hints; |
a0ed51b3 |
1335 | } |
1336 | if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) |
cea2e8a9 |
1337 | Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref"); |
a0ed51b3 |
1338 | return retval; |
1339 | } |
1340 | |
035d37be |
1341 | |
1342 | /* This API is wrong for special case conversions since we may need to |
1343 | * return several Unicode characters for a single Unicode character |
1344 | * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is |
1345 | * the lower-level routine, and it is similarly broken for returning |
1346 | * multiple values. --jhi */ |
a0ed51b3 |
1347 | UV |
3568d838 |
1348 | Perl_swash_fetch(pTHX_ SV *sv, U8 *ptr, bool do_utf8) |
a0ed51b3 |
1349 | { |
1350 | HV* hv = (HV*)SvRV(sv); |
3568d838 |
1351 | U32 klen; |
1352 | U32 off; |
a0ed51b3 |
1353 | STRLEN slen; |
7d85a32c |
1354 | STRLEN needents; |
4ea42e7f |
1355 | U8 *tmps = NULL; |
a0ed51b3 |
1356 | U32 bit; |
1357 | SV *retval; |
3568d838 |
1358 | U8 tmputf8[2]; |
1359 | UV c = NATIVE_TO_ASCII(*ptr); |
1360 | |
1361 | if (!do_utf8 && !UNI_IS_INVARIANT(c)) { |
1362 | tmputf8[0] = UTF8_EIGHT_BIT_HI(c); |
1363 | tmputf8[1] = UTF8_EIGHT_BIT_LO(c); |
1364 | ptr = tmputf8; |
1365 | } |
1366 | /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ |
1367 | * then the "swatch" is a vec() for al the chars which start |
1368 | * with 0xAA..0xYY |
1369 | * So the key in the hash (klen) is length of encoded char -1 |
1370 | */ |
1371 | klen = UTF8SKIP(ptr) - 1; |
1372 | off = ptr[klen]; |
a0ed51b3 |
1373 | |
7d85a32c |
1374 | if (klen == 0) |
1375 | { |
1376 | /* If char in invariant then swatch is for all the invariant chars |
1377 | * In both UTF-8 and UTF8-MOD that happens to be UTF_CONTINUATION_MARK |
1378 | */ |
1379 | needents = UTF_CONTINUATION_MARK; |
1380 | off = NATIVE_TO_UTF(ptr[klen]); |
1381 | } |
1382 | else |
1383 | { |
1384 | /* If char is encoded then swatch is for the prefix */ |
1385 | needents = (1 << UTF_ACCUMULATION_SHIFT); |
1386 | off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK; |
1387 | } |
1388 | |
a0ed51b3 |
1389 | /* |
1390 | * This single-entry cache saves about 1/3 of the utf8 overhead in test |
1391 | * suite. (That is, only 7-8% overall over just a hash cache. Still, |
1392 | * it's nothing to sniff at.) Pity we usually come through at least |
1393 | * two function calls to get here... |
1394 | * |
1395 | * NB: this code assumes that swatches are never modified, once generated! |
1396 | */ |
1397 | |
3568d838 |
1398 | if (hv == PL_last_swash_hv && |
a0ed51b3 |
1399 | klen == PL_last_swash_klen && |
3568d838 |
1400 | (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) ) |
a0ed51b3 |
1401 | { |
1402 | tmps = PL_last_swash_tmps; |
1403 | slen = PL_last_swash_slen; |
1404 | } |
1405 | else { |
1406 | /* Try our second-level swatch cache, kept in a hash. */ |
dfe13c55 |
1407 | SV** svp = hv_fetch(hv, (char*)ptr, klen, FALSE); |
a0ed51b3 |
1408 | |
1409 | /* If not cached, generate it via utf8::SWASHGET */ |
dfe13c55 |
1410 | if (!svp || !SvPOK(*svp) || !(tmps = (U8*)SvPV(*svp, slen))) { |
a0ed51b3 |
1411 | dSP; |
2b9d42f0 |
1412 | /* We use utf8n_to_uvuni() as we want an index into |
1413 | Unicode tables, not a native character number. |
1414 | */ |
1415 | UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXLEN, NULL, 0); |
f8be5cf0 |
1416 | SV *errsv_save; |
a0ed51b3 |
1417 | ENTER; |
1418 | SAVETMPS; |
1419 | save_re_context(); |
1420 | PUSHSTACKi(PERLSI_MAGIC); |
1421 | PUSHMARK(SP); |
1422 | EXTEND(SP,3); |
1423 | PUSHs((SV*)sv); |
ffbc6a93 |
1424 | /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */ |
3568d838 |
1425 | PUSHs(sv_2mortal(newSViv((klen) ? |
1426 | (code_point & ~(needents - 1)) : 0))); |
a0ed51b3 |
1427 | PUSHs(sv_2mortal(newSViv(needents))); |
1428 | PUTBACK; |
f8be5cf0 |
1429 | errsv_save = newSVsv(ERRSV); |
864dbfa3 |
1430 | if (call_method("SWASHGET", G_SCALAR)) |
8e84507e |
1431 | retval = newSVsv(*PL_stack_sp--); |
a0ed51b3 |
1432 | else |
e24b16f9 |
1433 | retval = &PL_sv_undef; |
f8be5cf0 |
1434 | if (!SvTRUE(ERRSV)) |
1435 | sv_setsv(ERRSV, errsv_save); |
1436 | SvREFCNT_dec(errsv_save); |
a0ed51b3 |
1437 | POPSTACK; |
1438 | FREETMPS; |
1439 | LEAVE; |
e24b16f9 |
1440 | if (PL_curcop == &PL_compiling) |
1441 | PL_curcop->op_private = PL_hints; |
a0ed51b3 |
1442 | |
dfe13c55 |
1443 | svp = hv_store(hv, (char*)ptr, klen, retval, 0); |
a0ed51b3 |
1444 | |
7d85a32c |
1445 | if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) || (slen << 3) < needents) |
cea2e8a9 |
1446 | Perl_croak(aTHX_ "SWASHGET didn't return result of proper length"); |
a0ed51b3 |
1447 | } |
1448 | |
1449 | PL_last_swash_hv = hv; |
1450 | PL_last_swash_klen = klen; |
1451 | PL_last_swash_tmps = tmps; |
1452 | PL_last_swash_slen = slen; |
1453 | if (klen) |
1454 | Copy(ptr, PL_last_swash_key, klen, U8); |
1455 | } |
1456 | |
9faf8d75 |
1457 | switch ((int)((slen << 3) / needents)) { |
a0ed51b3 |
1458 | case 1: |
1459 | bit = 1 << (off & 7); |
1460 | off >>= 3; |
1461 | return (tmps[off] & bit) != 0; |
1462 | case 8: |
1463 | return tmps[off]; |
1464 | case 16: |
1465 | off <<= 1; |
1466 | return (tmps[off] << 8) + tmps[off + 1] ; |
1467 | case 32: |
1468 | off <<= 2; |
1469 | return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ; |
1470 | } |
cea2e8a9 |
1471 | Perl_croak(aTHX_ "panic: swash_fetch"); |
a0ed51b3 |
1472 | return 0; |
1473 | } |
2b9d42f0 |
1474 | |
1475 | |
1476 | /* |
37607a96 |
1477 | =for apidoc A|U8 *|uvchr_to_utf8|U8 *d|UV uv |
2b9d42f0 |
1478 | |
1479 | Adds the UTF8 representation of the Native codepoint C<uv> to the end |
1480 | of the string C<d>; C<d> should be have at least C<UTF8_MAXLEN+1> free |
1481 | bytes available. The return value is the pointer to the byte after the |
1482 | end of the new character. In other words, |
1483 | |
1484 | d = uvchr_to_utf8(d, uv); |
1485 | |
1486 | is the recommended wide native character-aware way of saying |
1487 | |
1488 | *(d++) = uv; |
1489 | |
1490 | =cut |
1491 | */ |
1492 | |
1493 | /* On ASCII machines this is normally a macro but we want a |
1494 | real function in case XS code wants it |
1495 | */ |
1496 | #undef Perl_uvchr_to_utf8 |
1497 | U8 * |
1498 | Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv) |
1499 | { |
1500 | return Perl_uvuni_to_utf8(aTHX_ d, NATIVE_TO_UNI(uv)); |
1501 | } |
1502 | |
1503 | |
1504 | /* |
37607a96 |
1505 | =for apidoc A|UV|utf8n_to_uvchr|U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags |
2b9d42f0 |
1506 | |
1507 | Returns the native character value of the first character in the string C<s> |
1508 | which is assumed to be in UTF8 encoding; C<retlen> will be set to the |
1509 | length, in bytes, of that character. |
1510 | |
1511 | Allows length and flags to be passed to low level routine. |
1512 | |
1513 | =cut |
1514 | */ |
1515 | /* On ASCII machines this is normally a macro but we want a |
1516 | real function in case XS code wants it |
1517 | */ |
1518 | #undef Perl_utf8n_to_uvchr |
1519 | UV |
37607a96 |
1520 | Perl_utf8n_to_uvchr(pTHX_ U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) |
2b9d42f0 |
1521 | { |
1522 | UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags); |
1523 | return UNI_TO_NATIVE(uv); |
1524 | } |
1525 | |
e6b2e755 |
1526 | char * |
1527 | Perl_pv_uni_display(pTHX_ SV *dsv, U8 *spv, STRLEN len, STRLEN pvlim, UV flags) |
1528 | { |
1529 | int truncated = 0; |
1530 | char *s, *e; |
1531 | |
1532 | sv_setpvn(dsv, "", 0); |
1533 | for (s = (char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) { |
1534 | UV u; |
1535 | if (pvlim && SvCUR(dsv) >= pvlim) { |
1536 | truncated++; |
1537 | break; |
1538 | } |
1539 | u = utf8_to_uvchr((U8*)s, 0); |
1540 | Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u); |
1541 | } |
1542 | if (truncated) |
1543 | sv_catpvn(dsv, "...", 3); |
1544 | |
1545 | return SvPVX(dsv); |
1546 | } |
2b9d42f0 |
1547 | |
e6b2e755 |
1548 | char * |
1549 | Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags) |
1550 | { |
701a277b |
1551 | return Perl_pv_uni_display(aTHX_ dsv, (U8*)SvPVX(ssv), SvCUR(ssv), |
1552 | pvlim, flags); |
1553 | } |
1554 | |
1555 | I32 |
1556 | Perl_ibcmp_utf8(pTHX_ const char *s1, bool u1, const char *s2, bool u2, register I32 len) |
1557 | { |
1558 | register U8 *a = (U8*)s1; |
1559 | register U8 *b = (U8*)s2; |
1560 | STRLEN la, lb; |
1561 | UV ca, cb; |
1562 | STRLEN ulen1, ulen2; |
1563 | U8 tmpbuf1[UTF8_MAXLEN*3+1]; |
1564 | U8 tmpbuf2[UTF8_MAXLEN*3+1]; |
1565 | |
1566 | while (len) { |
1567 | if (u1) |
1568 | ca = utf8_to_uvchr((U8*)a, &la); |
1569 | else { |
1570 | ca = *a; |
1571 | la = 1; |
1572 | } |
1573 | if (u2) |
1574 | cb = utf8_to_uvchr((U8*)b, &lb); |
1575 | else { |
1576 | cb = *b; |
1577 | lb = 1; |
1578 | } |
1579 | if (ca != cb) { |
1580 | if (u1) |
596717ce |
1581 | to_uni_fold(NATIVE_TO_UNI(ca), tmpbuf1, &ulen1); |
701a277b |
1582 | else |
1583 | ulen1 = 1; |
1584 | if (u2) |
596717ce |
1585 | to_uni_fold(NATIVE_TO_UNI(cb), tmpbuf2, &ulen2); |
701a277b |
1586 | else |
1587 | ulen2 = 1; |
1588 | if (ulen1 != ulen2 |
1589 | || (ulen1 == 1 && PL_fold[ca] != PL_fold[cb]) |
726d7ae4 |
1590 | || memNE((char *)tmpbuf1, (char *)tmpbuf2, ulen1)) |
701a277b |
1591 | return 1; |
1592 | } |
1593 | a += la; |
1594 | b += lb; |
1595 | } |
1596 | return 0; |
e6b2e755 |
1597 | } |
701a277b |
1598 | |