Commit | Line | Data |
a0ed51b3 |
1 | /* utf8.c |
2 | * |
fdf8c088 |
3 | * Copyright (C) 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007 |
b94e2f88 |
4 | * by Larry Wall and others |
a0ed51b3 |
5 | * |
6 | * You may distribute under the terms of either the GNU General Public |
7 | * License or the Artistic License, as specified in the README file. |
8 | * |
9 | */ |
10 | |
11 | /* |
12 | * 'What a fix!' said Sam. 'That's the one place in all the lands we've ever |
13 | * heard of that we don't want to see any closer; and that's the one place |
14 | * we're trying to get to! And that's just where we can't get, nohow.' |
15 | * |
16 | * 'Well do I understand your speech,' he answered in the same language; |
17 | * 'yet few strangers do so. Why then do you not speak in the Common Tongue, |
18 | * as is the custom in the West, if you wish to be answered?' |
19 | * |
20 | * ...the travellers perceived that the floor was paved with stones of many |
21 | * hues; branching runes and strange devices intertwined beneath their feet. |
22 | */ |
23 | |
24 | #include "EXTERN.h" |
864dbfa3 |
25 | #define PERL_IN_UTF8_C |
a0ed51b3 |
26 | #include "perl.h" |
27 | |
a0c21aa1 |
28 | #ifndef EBCDIC |
29 | /* Separate prototypes needed because in ASCII systems these |
30 | * usually macros but they still are compiled as code, too. */ |
31 | PERL_CALLCONV UV Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags); |
32 | PERL_CALLCONV U8* Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv); |
33 | #endif |
34 | |
27da23d5 |
35 | static const char unees[] = |
36 | "Malformed UTF-8 character (unexpected end of string)"; |
901b21bf |
37 | |
ccfc67b7 |
38 | /* |
39 | =head1 Unicode Support |
a0ed51b3 |
40 | |
166f8a29 |
41 | This file contains various utility functions for manipulating UTF8-encoded |
42 | strings. For the uninitiated, this is a method of representing arbitrary |
61296642 |
43 | Unicode characters as a variable number of bytes, in such a way that |
56da48f7 |
44 | characters in the ASCII range are unmodified, and a zero byte never appears |
45 | within non-zero characters. |
166f8a29 |
46 | |
b851fbc1 |
47 | =for apidoc A|U8 *|uvuni_to_utf8_flags|U8 *d|UV uv|UV flags |
eebe1485 |
48 | |
1e54db1a |
49 | Adds the UTF-8 representation of the Unicode codepoint C<uv> to the end |
89ebb4a3 |
50 | of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free |
eebe1485 |
51 | bytes available. The return value is the pointer to the byte after the |
9041c2e3 |
52 | end of the new character. In other words, |
eebe1485 |
53 | |
b851fbc1 |
54 | d = uvuni_to_utf8_flags(d, uv, flags); |
55 | |
56 | or, in most cases, |
57 | |
9041c2e3 |
58 | d = uvuni_to_utf8(d, uv); |
eebe1485 |
59 | |
b851fbc1 |
60 | (which is equivalent to) |
61 | |
62 | d = uvuni_to_utf8_flags(d, uv, 0); |
63 | |
eebe1485 |
64 | is the recommended Unicode-aware way of saying |
65 | |
66 | *(d++) = uv; |
67 | |
68 | =cut |
69 | */ |
70 | |
dfe13c55 |
71 | U8 * |
b851fbc1 |
72 | Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) |
a0ed51b3 |
73 | { |
62961d2e |
74 | if (ckWARN(WARN_UTF8)) { |
b851fbc1 |
75 | if (UNICODE_IS_SURROGATE(uv) && |
76 | !(flags & UNICODE_ALLOW_SURROGATE)) |
9014280d |
77 | Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv); |
b851fbc1 |
78 | else if ( |
79 | ((uv >= 0xFDD0 && uv <= 0xFDEF && |
80 | !(flags & UNICODE_ALLOW_FDD0)) |
81 | || |
c867b360 |
82 | ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */ |
b851fbc1 |
83 | !(flags & UNICODE_ALLOW_FFFF))) && |
84 | /* UNICODE_ALLOW_SUPER includes |
2a20b9da |
85 | * FFFEs and FFFFs beyond 0x10FFFF. */ |
b851fbc1 |
86 | ((uv <= PERL_UNICODE_MAX) || |
87 | !(flags & UNICODE_ALLOW_SUPER)) |
88 | ) |
9014280d |
89 | Perl_warner(aTHX_ packWARN(WARN_UTF8), |
507b9800 |
90 | "Unicode character 0x%04"UVxf" is illegal", uv); |
91 | } |
c4d5f83a |
92 | if (UNI_IS_INVARIANT(uv)) { |
eb160463 |
93 | *d++ = (U8)UTF_TO_NATIVE(uv); |
a0ed51b3 |
94 | return d; |
95 | } |
2d331972 |
96 | #if defined(EBCDIC) |
1d72bdf6 |
97 | else { |
98 | STRLEN len = UNISKIP(uv); |
99 | U8 *p = d+len-1; |
100 | while (p > d) { |
eb160463 |
101 | *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK); |
1d72bdf6 |
102 | uv >>= UTF_ACCUMULATION_SHIFT; |
103 | } |
eb160463 |
104 | *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len)); |
1d72bdf6 |
105 | return d+len; |
106 | } |
107 | #else /* Non loop style */ |
a0ed51b3 |
108 | if (uv < 0x800) { |
eb160463 |
109 | *d++ = (U8)(( uv >> 6) | 0xc0); |
110 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
111 | return d; |
112 | } |
113 | if (uv < 0x10000) { |
eb160463 |
114 | *d++ = (U8)(( uv >> 12) | 0xe0); |
115 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
116 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
117 | return d; |
118 | } |
119 | if (uv < 0x200000) { |
eb160463 |
120 | *d++ = (U8)(( uv >> 18) | 0xf0); |
121 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
122 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
123 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
124 | return d; |
125 | } |
126 | if (uv < 0x4000000) { |
eb160463 |
127 | *d++ = (U8)(( uv >> 24) | 0xf8); |
128 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
129 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
130 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
131 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
132 | return d; |
133 | } |
134 | if (uv < 0x80000000) { |
eb160463 |
135 | *d++ = (U8)(( uv >> 30) | 0xfc); |
136 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); |
137 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
138 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
139 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
140 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
141 | return d; |
142 | } |
6b8eaf93 |
143 | #ifdef HAS_QUAD |
d7578b48 |
144 | if (uv < UTF8_QUAD_MAX) |
a0ed51b3 |
145 | #endif |
146 | { |
eb160463 |
147 | *d++ = 0xfe; /* Can't match U+FEFF! */ |
148 | *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); |
149 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); |
150 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
151 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
152 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
153 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
154 | return d; |
155 | } |
6b8eaf93 |
156 | #ifdef HAS_QUAD |
a0ed51b3 |
157 | { |
eb160463 |
158 | *d++ = 0xff; /* Can't match U+FFFE! */ |
159 | *d++ = 0x80; /* 6 Reserved bits */ |
160 | *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */ |
161 | *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80); |
162 | *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80); |
163 | *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80); |
164 | *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80); |
165 | *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); |
166 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); |
167 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
168 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
169 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
170 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
171 | return d; |
172 | } |
173 | #endif |
1d72bdf6 |
174 | #endif /* Loop style */ |
a0ed51b3 |
175 | } |
9041c2e3 |
176 | |
646ca15d |
177 | /* |
178 | |
179 | Tests if some arbitrary number of bytes begins in a valid UTF-8 |
180 | character. Note that an INVARIANT (i.e. ASCII) character is a valid |
181 | UTF-8 character. The actual number of bytes in the UTF-8 character |
182 | will be returned if it is valid, otherwise 0. |
183 | |
184 | This is the "slow" version as opposed to the "fast" version which is |
185 | the "unrolled" IS_UTF8_CHAR(). E.g. for t/uni/class.t the speed |
186 | difference is a factor of 2 to 3. For lengths (UTF8SKIP(s)) of four |
187 | or less you should use the IS_UTF8_CHAR(), for lengths of five or more |
188 | you should use the _slow(). In practice this means that the _slow() |
189 | will be used very rarely, since the maximum Unicode code point (as of |
190 | Unicode 4.1) is U+10FFFF, which encodes in UTF-8 to four bytes. Only |
191 | the "Perl extended UTF-8" (the infamous 'v-strings') will encode into |
192 | five bytes or more. |
193 | |
194 | =cut */ |
c053b435 |
195 | STATIC STRLEN |
5f66b61c |
196 | S_is_utf8_char_slow(const U8 *s, const STRLEN len) |
646ca15d |
197 | { |
198 | U8 u = *s; |
199 | STRLEN slen; |
200 | UV uv, ouv; |
201 | |
202 | if (UTF8_IS_INVARIANT(u)) |
203 | return 1; |
204 | |
205 | if (!UTF8_IS_START(u)) |
206 | return 0; |
207 | |
208 | if (len < 2 || !UTF8_IS_CONTINUATION(s[1])) |
209 | return 0; |
210 | |
211 | slen = len - 1; |
212 | s++; |
77263263 |
213 | #ifdef EBCDIC |
214 | u = NATIVE_TO_UTF(u); |
215 | #endif |
646ca15d |
216 | u &= UTF_START_MASK(len); |
217 | uv = u; |
218 | ouv = uv; |
219 | while (slen--) { |
220 | if (!UTF8_IS_CONTINUATION(*s)) |
221 | return 0; |
222 | uv = UTF8_ACCUMULATE(uv, *s); |
223 | if (uv < ouv) |
224 | return 0; |
225 | ouv = uv; |
226 | s++; |
227 | } |
228 | |
229 | if ((STRLEN)UNISKIP(uv) < len) |
230 | return 0; |
231 | |
232 | return len; |
233 | } |
9041c2e3 |
234 | |
235 | /* |
7fc63493 |
236 | =for apidoc A|STRLEN|is_utf8_char|const U8 *s |
eebe1485 |
237 | |
5da9da9e |
238 | Tests if some arbitrary number of bytes begins in a valid UTF-8 |
82686b01 |
239 | character. Note that an INVARIANT (i.e. ASCII) character is a valid |
240 | UTF-8 character. The actual number of bytes in the UTF-8 character |
241 | will be returned if it is valid, otherwise 0. |
9041c2e3 |
242 | |
82686b01 |
243 | =cut */ |
067a85ef |
244 | STRLEN |
7fc63493 |
245 | Perl_is_utf8_char(pTHX_ const U8 *s) |
386d01d6 |
246 | { |
44f8325f |
247 | const STRLEN len = UTF8SKIP(s); |
96a5add6 |
248 | PERL_UNUSED_CONTEXT; |
3b0fc154 |
249 | #ifdef IS_UTF8_CHAR |
768c67ee |
250 | if (IS_UTF8_CHAR_FAST(len)) |
3b0fc154 |
251 | return IS_UTF8_CHAR(s, len) ? len : 0; |
252 | #endif /* #ifdef IS_UTF8_CHAR */ |
2c0c5f92 |
253 | return is_utf8_char_slow(s, len); |
386d01d6 |
254 | } |
255 | |
6662521e |
256 | /* |
7fc63493 |
257 | =for apidoc A|bool|is_utf8_string|const U8 *s|STRLEN len |
6662521e |
258 | |
c9ada85f |
259 | Returns true if first C<len> bytes of the given string form a valid |
1e54db1a |
260 | UTF-8 string, false otherwise. Note that 'a valid UTF-8 string' does |
261 | not mean 'a string that contains code points above 0x7F encoded in UTF-8' |
262 | because a valid ASCII string is a valid UTF-8 string. |
6662521e |
263 | |
768c67ee |
264 | See also is_utf8_string_loclen() and is_utf8_string_loc(). |
265 | |
6662521e |
266 | =cut |
267 | */ |
268 | |
8e84507e |
269 | bool |
7fc63493 |
270 | Perl_is_utf8_string(pTHX_ const U8 *s, STRLEN len) |
6662521e |
271 | { |
35da51f7 |
272 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
7fc63493 |
273 | const U8* x = s; |
067a85ef |
274 | |
96a5add6 |
275 | PERL_UNUSED_CONTEXT; |
1aa99e6b |
276 | |
6662521e |
277 | while (x < send) { |
a3b680e6 |
278 | STRLEN c; |
1acdb0da |
279 | /* Inline the easy bits of is_utf8_char() here for speed... */ |
280 | if (UTF8_IS_INVARIANT(*x)) |
281 | c = 1; |
282 | else if (!UTF8_IS_START(*x)) |
768c67ee |
283 | goto out; |
1acdb0da |
284 | else { |
285 | /* ... and call is_utf8_char() only if really needed. */ |
646ca15d |
286 | #ifdef IS_UTF8_CHAR |
287 | c = UTF8SKIP(x); |
768c67ee |
288 | if (IS_UTF8_CHAR_FAST(c)) { |
289 | if (!IS_UTF8_CHAR(x, c)) |
3c614e38 |
290 | c = 0; |
291 | } |
292 | else |
293 | c = is_utf8_char_slow(x, c); |
646ca15d |
294 | #else |
295 | c = is_utf8_char(x); |
296 | #endif /* #ifdef IS_UTF8_CHAR */ |
1acdb0da |
297 | if (!c) |
768c67ee |
298 | goto out; |
1acdb0da |
299 | } |
6662521e |
300 | x += c; |
6662521e |
301 | } |
768c67ee |
302 | |
303 | out: |
60006e79 |
304 | if (x != send) |
305 | return FALSE; |
067a85ef |
306 | |
307 | return TRUE; |
6662521e |
308 | } |
309 | |
67e989fb |
310 | /* |
814fafa7 |
311 | Implemented as a macro in utf8.h |
312 | |
313 | =for apidoc A|bool|is_utf8_string_loc|const U8 *s|STRLEN len|const U8 **ep |
314 | |
315 | Like is_utf8_string() but stores the location of the failure (in the |
316 | case of "utf8ness failure") or the location s+len (in the case of |
317 | "utf8ness success") in the C<ep>. |
318 | |
319 | See also is_utf8_string_loclen() and is_utf8_string(). |
320 | |
768c67ee |
321 | =for apidoc A|bool|is_utf8_string_loclen|const U8 *s|STRLEN len|const U8 **ep|const STRLEN *el |
81cd54e3 |
322 | |
e3e4599f |
323 | Like is_utf8_string() but stores the location of the failure (in the |
768c67ee |
324 | case of "utf8ness failure") or the location s+len (in the case of |
325 | "utf8ness success") in the C<ep>, and the number of UTF-8 |
326 | encoded characters in the C<el>. |
327 | |
328 | See also is_utf8_string_loc() and is_utf8_string(). |
81cd54e3 |
329 | |
330 | =cut |
331 | */ |
332 | |
333 | bool |
768c67ee |
334 | Perl_is_utf8_string_loclen(pTHX_ const U8 *s, STRLEN len, const U8 **ep, STRLEN *el) |
81cd54e3 |
335 | { |
35da51f7 |
336 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
7fc63493 |
337 | const U8* x = s; |
81cd54e3 |
338 | STRLEN c; |
3ebfea28 |
339 | STRLEN outlen = 0; |
96a5add6 |
340 | PERL_UNUSED_CONTEXT; |
81cd54e3 |
341 | |
81cd54e3 |
342 | while (x < send) { |
343 | /* Inline the easy bits of is_utf8_char() here for speed... */ |
344 | if (UTF8_IS_INVARIANT(*x)) |
768c67ee |
345 | c = 1; |
346 | else if (!UTF8_IS_START(*x)) |
347 | goto out; |
81cd54e3 |
348 | else { |
768c67ee |
349 | /* ... and call is_utf8_char() only if really needed. */ |
350 | #ifdef IS_UTF8_CHAR |
351 | c = UTF8SKIP(x); |
352 | if (IS_UTF8_CHAR_FAST(c)) { |
353 | if (!IS_UTF8_CHAR(x, c)) |
354 | c = 0; |
355 | } else |
356 | c = is_utf8_char_slow(x, c); |
357 | #else |
358 | c = is_utf8_char(x); |
359 | #endif /* #ifdef IS_UTF8_CHAR */ |
360 | if (!c) |
361 | goto out; |
81cd54e3 |
362 | } |
768c67ee |
363 | x += c; |
3ebfea28 |
364 | outlen++; |
81cd54e3 |
365 | } |
768c67ee |
366 | |
367 | out: |
3ebfea28 |
368 | if (el) |
369 | *el = outlen; |
370 | |
768c67ee |
371 | if (ep) |
372 | *ep = x; |
3ebfea28 |
373 | return (x == send); |
81cd54e3 |
374 | } |
375 | |
376 | /* |
768c67ee |
377 | |
7fc63493 |
378 | =for apidoc A|UV|utf8n_to_uvuni|const U8 *s|STRLEN curlen|STRLEN *retlen|U32 flags |
67e989fb |
379 | |
9041c2e3 |
380 | Bottom level UTF-8 decode routine. |
38a44b82 |
381 | Returns the Unicode code point value of the first character in the string C<s> |
1e54db1a |
382 | which is assumed to be in UTF-8 encoding and no longer than C<curlen>; |
7df053ec |
383 | C<retlen> will be set to the length, in bytes, of that character. |
67e989fb |
384 | |
1e54db1a |
385 | If C<s> does not point to a well-formed UTF-8 character, the behaviour |
dcad2880 |
386 | is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY, |
387 | it is assumed that the caller will raise a warning, and this function |
28d3d195 |
388 | will silently just set C<retlen> to C<-1> and return zero. If the |
389 | C<flags> does not contain UTF8_CHECK_ONLY, warnings about |
390 | malformations will be given, C<retlen> will be set to the expected |
391 | length of the UTF-8 character in bytes, and zero will be returned. |
392 | |
393 | The C<flags> can also contain various flags to allow deviations from |
394 | the strict UTF-8 encoding (see F<utf8.h>). |
67e989fb |
395 | |
9041c2e3 |
396 | Most code should use utf8_to_uvchr() rather than call this directly. |
397 | |
37607a96 |
398 | =cut |
399 | */ |
67e989fb |
400 | |
a0ed51b3 |
401 | UV |
7fc63493 |
402 | Perl_utf8n_to_uvuni(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) |
a0ed51b3 |
403 | { |
97aff369 |
404 | dVAR; |
d4c19fe8 |
405 | const U8 * const s0 = s; |
9c5ffd7c |
406 | UV uv = *s, ouv = 0; |
ba210ebe |
407 | STRLEN len = 1; |
7fc63493 |
408 | const bool dowarn = ckWARN_d(WARN_UTF8); |
409 | const UV startbyte = *s; |
ba210ebe |
410 | STRLEN expectlen = 0; |
a0dbb045 |
411 | U32 warning = 0; |
412 | |
413 | /* This list is a superset of the UTF8_ALLOW_XXX. */ |
414 | |
415 | #define UTF8_WARN_EMPTY 1 |
416 | #define UTF8_WARN_CONTINUATION 2 |
417 | #define UTF8_WARN_NON_CONTINUATION 3 |
418 | #define UTF8_WARN_FE_FF 4 |
419 | #define UTF8_WARN_SHORT 5 |
420 | #define UTF8_WARN_OVERFLOW 6 |
421 | #define UTF8_WARN_SURROGATE 7 |
c867b360 |
422 | #define UTF8_WARN_LONG 8 |
423 | #define UTF8_WARN_FFFF 9 /* Also FFFE. */ |
a0dbb045 |
424 | |
425 | if (curlen == 0 && |
426 | !(flags & UTF8_ALLOW_EMPTY)) { |
427 | warning = UTF8_WARN_EMPTY; |
0c443dc2 |
428 | goto malformed; |
429 | } |
430 | |
1d72bdf6 |
431 | if (UTF8_IS_INVARIANT(uv)) { |
a0ed51b3 |
432 | if (retlen) |
433 | *retlen = 1; |
c4d5f83a |
434 | return (UV) (NATIVE_TO_UTF(*s)); |
a0ed51b3 |
435 | } |
67e989fb |
436 | |
421a8bf2 |
437 | if (UTF8_IS_CONTINUATION(uv) && |
fcc8fcf6 |
438 | !(flags & UTF8_ALLOW_CONTINUATION)) { |
a0dbb045 |
439 | warning = UTF8_WARN_CONTINUATION; |
ba210ebe |
440 | goto malformed; |
441 | } |
442 | |
421a8bf2 |
443 | if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) && |
fcc8fcf6 |
444 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { |
a0dbb045 |
445 | warning = UTF8_WARN_NON_CONTINUATION; |
ba210ebe |
446 | goto malformed; |
447 | } |
9041c2e3 |
448 | |
1d72bdf6 |
449 | #ifdef EBCDIC |
75383841 |
450 | uv = NATIVE_TO_UTF(uv); |
1d72bdf6 |
451 | #else |
fcc8fcf6 |
452 | if ((uv == 0xfe || uv == 0xff) && |
453 | !(flags & UTF8_ALLOW_FE_FF)) { |
a0dbb045 |
454 | warning = UTF8_WARN_FE_FF; |
ba210ebe |
455 | goto malformed; |
a0ed51b3 |
456 | } |
1d72bdf6 |
457 | #endif |
458 | |
ba210ebe |
459 | if (!(uv & 0x20)) { len = 2; uv &= 0x1f; } |
460 | else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; } |
461 | else if (!(uv & 0x08)) { len = 4; uv &= 0x07; } |
462 | else if (!(uv & 0x04)) { len = 5; uv &= 0x03; } |
1d72bdf6 |
463 | #ifdef EBCDIC |
464 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } |
465 | else { len = 7; uv &= 0x01; } |
466 | #else |
ba210ebe |
467 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } |
468 | else if (!(uv & 0x01)) { len = 7; uv = 0; } |
1d72bdf6 |
469 | else { len = 13; uv = 0; } /* whoa! */ |
470 | #endif |
471 | |
a0ed51b3 |
472 | if (retlen) |
473 | *retlen = len; |
9041c2e3 |
474 | |
ba210ebe |
475 | expectlen = len; |
476 | |
fcc8fcf6 |
477 | if ((curlen < expectlen) && |
478 | !(flags & UTF8_ALLOW_SHORT)) { |
a0dbb045 |
479 | warning = UTF8_WARN_SHORT; |
ba210ebe |
480 | goto malformed; |
481 | } |
482 | |
483 | len--; |
a0ed51b3 |
484 | s++; |
ba210ebe |
485 | ouv = uv; |
486 | |
a0ed51b3 |
487 | while (len--) { |
421a8bf2 |
488 | if (!UTF8_IS_CONTINUATION(*s) && |
489 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { |
a0dbb045 |
490 | s--; |
491 | warning = UTF8_WARN_NON_CONTINUATION; |
ba210ebe |
492 | goto malformed; |
a0ed51b3 |
493 | } |
494 | else |
8850bf83 |
495 | uv = UTF8_ACCUMULATE(uv, *s); |
a0dbb045 |
496 | if (!(uv > ouv)) { |
497 | /* These cannot be allowed. */ |
498 | if (uv == ouv) { |
75dbc644 |
499 | if (expectlen != 13 && !(flags & UTF8_ALLOW_LONG)) { |
a0dbb045 |
500 | warning = UTF8_WARN_LONG; |
501 | goto malformed; |
502 | } |
503 | } |
504 | else { /* uv < ouv */ |
505 | /* This cannot be allowed. */ |
506 | warning = UTF8_WARN_OVERFLOW; |
507 | goto malformed; |
508 | } |
ba210ebe |
509 | } |
510 | s++; |
511 | ouv = uv; |
512 | } |
513 | |
421a8bf2 |
514 | if (UNICODE_IS_SURROGATE(uv) && |
fcc8fcf6 |
515 | !(flags & UTF8_ALLOW_SURROGATE)) { |
a0dbb045 |
516 | warning = UTF8_WARN_SURROGATE; |
ba210ebe |
517 | goto malformed; |
eb160463 |
518 | } else if ((expectlen > (STRLEN)UNISKIP(uv)) && |
fcc8fcf6 |
519 | !(flags & UTF8_ALLOW_LONG)) { |
a0dbb045 |
520 | warning = UTF8_WARN_LONG; |
ba210ebe |
521 | goto malformed; |
421a8bf2 |
522 | } else if (UNICODE_IS_ILLEGAL(uv) && |
a9917092 |
523 | !(flags & UTF8_ALLOW_FFFF)) { |
a0dbb045 |
524 | warning = UTF8_WARN_FFFF; |
a9917092 |
525 | goto malformed; |
a0ed51b3 |
526 | } |
ba210ebe |
527 | |
a0ed51b3 |
528 | return uv; |
ba210ebe |
529 | |
530 | malformed: |
531 | |
fcc8fcf6 |
532 | if (flags & UTF8_CHECK_ONLY) { |
ba210ebe |
533 | if (retlen) |
10edeb5d |
534 | *retlen = ((STRLEN) -1); |
ba210ebe |
535 | return 0; |
536 | } |
537 | |
a0dbb045 |
538 | if (dowarn) { |
396482e1 |
539 | SV* const sv = sv_2mortal(newSVpvs("Malformed UTF-8 character ")); |
a0dbb045 |
540 | |
541 | switch (warning) { |
542 | case 0: /* Intentionally empty. */ break; |
543 | case UTF8_WARN_EMPTY: |
396482e1 |
544 | sv_catpvs(sv, "(empty string)"); |
a0dbb045 |
545 | break; |
546 | case UTF8_WARN_CONTINUATION: |
097fb8e2 |
547 | Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv); |
a0dbb045 |
548 | break; |
549 | case UTF8_WARN_NON_CONTINUATION: |
097fb8e2 |
550 | if (s == s0) |
551 | Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")", |
552 | (UV)s[1], startbyte); |
551405c4 |
553 | else { |
554 | const int len = (int)(s-s0); |
097fb8e2 |
555 | Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)", |
551405c4 |
556 | (UV)s[1], len, len > 1 ? "s" : "", startbyte, (int)expectlen); |
557 | } |
558 | |
a0dbb045 |
559 | break; |
560 | case UTF8_WARN_FE_FF: |
561 | Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv); |
562 | break; |
563 | case UTF8_WARN_SHORT: |
097fb8e2 |
564 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", |
5d7488b2 |
565 | (int)curlen, curlen == 1 ? "" : "s", (int)expectlen, startbyte); |
b31f83c2 |
566 | expectlen = curlen; /* distance for caller to skip */ |
a0dbb045 |
567 | break; |
568 | case UTF8_WARN_OVERFLOW: |
097fb8e2 |
569 | Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")", |
570 | ouv, *s, startbyte); |
a0dbb045 |
571 | break; |
572 | case UTF8_WARN_SURROGATE: |
573 | Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv); |
574 | break; |
a0dbb045 |
575 | case UTF8_WARN_LONG: |
097fb8e2 |
576 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", |
5d7488b2 |
577 | (int)expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte); |
a0dbb045 |
578 | break; |
579 | case UTF8_WARN_FFFF: |
580 | Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv); |
581 | break; |
582 | default: |
396482e1 |
583 | sv_catpvs(sv, "(unknown reason)"); |
a0dbb045 |
584 | break; |
585 | } |
586 | |
587 | if (warning) { |
44f8325f |
588 | const char * const s = SvPVX_const(sv); |
a0dbb045 |
589 | |
590 | if (PL_op) |
9014280d |
591 | Perl_warner(aTHX_ packWARN(WARN_UTF8), |
53e06cf0 |
592 | "%s in %s", s, OP_DESC(PL_op)); |
a0dbb045 |
593 | else |
9014280d |
594 | Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s); |
a0dbb045 |
595 | } |
596 | } |
597 | |
ba210ebe |
598 | if (retlen) |
28d3d195 |
599 | *retlen = expectlen ? expectlen : len; |
ba210ebe |
600 | |
28d3d195 |
601 | return 0; |
a0ed51b3 |
602 | } |
603 | |
8e84507e |
604 | /* |
7fc63493 |
605 | =for apidoc A|UV|utf8_to_uvchr|const U8 *s|STRLEN *retlen |
9041c2e3 |
606 | |
607 | Returns the native character value of the first character in the string C<s> |
1e54db1a |
608 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
9041c2e3 |
609 | length, in bytes, of that character. |
610 | |
1e54db1a |
611 | If C<s> does not point to a well-formed UTF-8 character, zero is |
9041c2e3 |
612 | returned and retlen is set, if possible, to -1. |
613 | |
614 | =cut |
615 | */ |
616 | |
617 | UV |
7fc63493 |
618 | Perl_utf8_to_uvchr(pTHX_ const U8 *s, STRLEN *retlen) |
9041c2e3 |
619 | { |
1754c1a1 |
620 | return utf8n_to_uvchr(s, UTF8_MAXBYTES, retlen, |
621 | ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); |
9041c2e3 |
622 | } |
623 | |
624 | /* |
7fc63493 |
625 | =for apidoc A|UV|utf8_to_uvuni|const U8 *s|STRLEN *retlen |
9041c2e3 |
626 | |
627 | Returns the Unicode code point of the first character in the string C<s> |
1e54db1a |
628 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
9041c2e3 |
629 | length, in bytes, of that character. |
630 | |
631 | This function should only be used when returned UV is considered |
632 | an index into the Unicode semantic tables (e.g. swashes). |
633 | |
1e54db1a |
634 | If C<s> does not point to a well-formed UTF-8 character, zero is |
ba210ebe |
635 | returned and retlen is set, if possible, to -1. |
8e84507e |
636 | |
637 | =cut |
638 | */ |
639 | |
640 | UV |
7fc63493 |
641 | Perl_utf8_to_uvuni(pTHX_ const U8 *s, STRLEN *retlen) |
8e84507e |
642 | { |
9041c2e3 |
643 | /* Call the low level routine asking for checks */ |
89ebb4a3 |
644 | return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXBYTES, retlen, |
872c91ae |
645 | ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); |
8e84507e |
646 | } |
647 | |
b76347f2 |
648 | /* |
35a4481c |
649 | =for apidoc A|STRLEN|utf8_length|const U8 *s|const U8 *e |
b76347f2 |
650 | |
651 | Return the length of the UTF-8 char encoded string C<s> in characters. |
02eb7b47 |
652 | Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end |
653 | up past C<e>, croaks. |
b76347f2 |
654 | |
655 | =cut |
656 | */ |
657 | |
658 | STRLEN |
35a4481c |
659 | Perl_utf8_length(pTHX_ const U8 *s, const U8 *e) |
b76347f2 |
660 | { |
97aff369 |
661 | dVAR; |
b76347f2 |
662 | STRLEN len = 0; |
10edeb5d |
663 | U8 t = 0; |
b76347f2 |
664 | |
8850bf83 |
665 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g. |
666 | * the bitops (especially ~) can create illegal UTF-8. |
667 | * In other words: in Perl UTF-8 is not just for Unicode. */ |
668 | |
a3b680e6 |
669 | if (e < s) |
670 | goto warn_and_return; |
b76347f2 |
671 | while (s < e) { |
10edeb5d |
672 | t = UTF8SKIP(s); |
901b21bf |
673 | if (e - s < t) { |
a3b680e6 |
674 | warn_and_return: |
901b21bf |
675 | if (ckWARN_d(WARN_UTF8)) { |
676 | if (PL_op) |
677 | Perl_warner(aTHX_ packWARN(WARN_UTF8), |
a3b680e6 |
678 | "%s in %s", unees, OP_DESC(PL_op)); |
901b21bf |
679 | else |
680 | Perl_warner(aTHX_ packWARN(WARN_UTF8), unees); |
681 | } |
682 | return len; |
683 | } |
b76347f2 |
684 | s += t; |
685 | len++; |
686 | } |
687 | |
688 | return len; |
689 | } |
690 | |
b06226ff |
691 | /* |
35a4481c |
692 | =for apidoc A|IV|utf8_distance|const U8 *a|const U8 *b |
b06226ff |
693 | |
1e54db1a |
694 | Returns the number of UTF-8 characters between the UTF-8 pointers C<a> |
b06226ff |
695 | and C<b>. |
696 | |
697 | WARNING: use only if you *know* that the pointers point inside the |
698 | same UTF-8 buffer. |
699 | |
37607a96 |
700 | =cut |
701 | */ |
a0ed51b3 |
702 | |
02eb7b47 |
703 | IV |
35a4481c |
704 | Perl_utf8_distance(pTHX_ const U8 *a, const U8 *b) |
a0ed51b3 |
705 | { |
bf1665bc |
706 | return (a < b) ? -1 * (IV) utf8_length(a, b) : (IV) utf8_length(b, a); |
a0ed51b3 |
707 | } |
708 | |
b06226ff |
709 | /* |
37607a96 |
710 | =for apidoc A|U8 *|utf8_hop|U8 *s|I32 off |
b06226ff |
711 | |
8850bf83 |
712 | Return the UTF-8 pointer C<s> displaced by C<off> characters, either |
713 | forward or backward. |
b06226ff |
714 | |
715 | WARNING: do not use the following unless you *know* C<off> is within |
8850bf83 |
716 | the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned |
717 | on the first byte of character or just after the last byte of a character. |
b06226ff |
718 | |
37607a96 |
719 | =cut |
720 | */ |
a0ed51b3 |
721 | |
722 | U8 * |
4373e329 |
723 | Perl_utf8_hop(pTHX_ const U8 *s, I32 off) |
a0ed51b3 |
724 | { |
96a5add6 |
725 | PERL_UNUSED_CONTEXT; |
8850bf83 |
726 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g |
727 | * the bitops (especially ~) can create illegal UTF-8. |
728 | * In other words: in Perl UTF-8 is not just for Unicode. */ |
729 | |
a0ed51b3 |
730 | if (off >= 0) { |
731 | while (off--) |
732 | s += UTF8SKIP(s); |
733 | } |
734 | else { |
735 | while (off++) { |
736 | s--; |
8850bf83 |
737 | while (UTF8_IS_CONTINUATION(*s)) |
738 | s--; |
a0ed51b3 |
739 | } |
740 | } |
4373e329 |
741 | return (U8 *)s; |
a0ed51b3 |
742 | } |
743 | |
6940069f |
744 | /* |
eebe1485 |
745 | =for apidoc A|U8 *|utf8_to_bytes|U8 *s|STRLEN *len |
6940069f |
746 | |
1e54db1a |
747 | Converts a string C<s> of length C<len> from UTF-8 into byte encoding. |
246fae53 |
748 | Unlike C<bytes_to_utf8>, this over-writes the original string, and |
749 | updates len to contain the new length. |
67e989fb |
750 | Returns zero on failure, setting C<len> to -1. |
6940069f |
751 | |
95be277c |
752 | If you need a copy of the string, see C<bytes_from_utf8>. |
753 | |
6940069f |
754 | =cut |
755 | */ |
756 | |
757 | U8 * |
37607a96 |
758 | Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len) |
6940069f |
759 | { |
d4c19fe8 |
760 | U8 * const save = s; |
761 | U8 * const send = s + *len; |
6940069f |
762 | U8 *d; |
246fae53 |
763 | |
1e54db1a |
764 | /* ensure valid UTF-8 and chars < 256 before updating string */ |
d4c19fe8 |
765 | while (s < send) { |
dcad2880 |
766 | U8 c = *s++; |
767 | |
1d72bdf6 |
768 | if (!UTF8_IS_INVARIANT(c) && |
769 | (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send) |
770 | || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) { |
10edeb5d |
771 | *len = ((STRLEN) -1); |
dcad2880 |
772 | return 0; |
773 | } |
246fae53 |
774 | } |
dcad2880 |
775 | |
776 | d = s = save; |
6940069f |
777 | while (s < send) { |
ed646e6e |
778 | STRLEN ulen; |
9041c2e3 |
779 | *d++ = (U8)utf8_to_uvchr(s, &ulen); |
ed646e6e |
780 | s += ulen; |
6940069f |
781 | } |
782 | *d = '\0'; |
246fae53 |
783 | *len = d - save; |
6940069f |
784 | return save; |
785 | } |
786 | |
787 | /* |
e1ec3a88 |
788 | =for apidoc A|U8 *|bytes_from_utf8|const U8 *s|STRLEN *len|bool *is_utf8 |
f9a63242 |
789 | |
1e54db1a |
790 | Converts a string C<s> of length C<len> from UTF-8 into byte encoding. |
35a4481c |
791 | Unlike C<utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to |
ef9edfd0 |
792 | the newly-created string, and updates C<len> to contain the new |
793 | length. Returns the original string if no conversion occurs, C<len> |
794 | is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to |
795 | 0 if C<s> is converted or contains all 7bit characters. |
f9a63242 |
796 | |
37607a96 |
797 | =cut |
798 | */ |
f9a63242 |
799 | |
800 | U8 * |
e1ec3a88 |
801 | Perl_bytes_from_utf8(pTHX_ const U8 *s, STRLEN *len, bool *is_utf8) |
f9a63242 |
802 | { |
f9a63242 |
803 | U8 *d; |
e1ec3a88 |
804 | const U8 *start = s; |
805 | const U8 *send; |
f9a63242 |
806 | I32 count = 0; |
807 | |
96a5add6 |
808 | PERL_UNUSED_CONTEXT; |
f9a63242 |
809 | if (!*is_utf8) |
73d840c0 |
810 | return (U8 *)start; |
f9a63242 |
811 | |
1e54db1a |
812 | /* ensure valid UTF-8 and chars < 256 before converting string */ |
f9a63242 |
813 | for (send = s + *len; s < send;) { |
e1ec3a88 |
814 | U8 c = *s++; |
1d72bdf6 |
815 | if (!UTF8_IS_INVARIANT(c)) { |
db42d148 |
816 | if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send && |
817 | (c = *s++) && UTF8_IS_CONTINUATION(c)) |
818 | count++; |
819 | else |
73d840c0 |
820 | return (U8 *)start; |
db42d148 |
821 | } |
f9a63242 |
822 | } |
823 | |
35da51f7 |
824 | *is_utf8 = FALSE; |
f9a63242 |
825 | |
212542aa |
826 | Newx(d, (*len) - count + 1, U8); |
ef9edfd0 |
827 | s = start; start = d; |
f9a63242 |
828 | while (s < send) { |
829 | U8 c = *s++; |
c4d5f83a |
830 | if (!UTF8_IS_INVARIANT(c)) { |
831 | /* Then it is two-byte encoded */ |
832 | c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++); |
833 | c = ASCII_TO_NATIVE(c); |
834 | } |
835 | *d++ = c; |
f9a63242 |
836 | } |
837 | *d = '\0'; |
838 | *len = d - start; |
73d840c0 |
839 | return (U8 *)start; |
f9a63242 |
840 | } |
841 | |
842 | /* |
35a4481c |
843 | =for apidoc A|U8 *|bytes_to_utf8|const U8 *s|STRLEN *len |
6940069f |
844 | |
1e54db1a |
845 | Converts a string C<s> of length C<len> from ASCII into UTF-8 encoding. |
6662521e |
846 | Returns a pointer to the newly-created string, and sets C<len> to |
847 | reflect the new length. |
6940069f |
848 | |
1e54db1a |
849 | If you want to convert to UTF-8 from other encodings than ASCII, |
c9ada85f |
850 | see sv_recode_to_utf8(). |
851 | |
497711e7 |
852 | =cut |
6940069f |
853 | */ |
854 | |
855 | U8* |
35a4481c |
856 | Perl_bytes_to_utf8(pTHX_ const U8 *s, STRLEN *len) |
6940069f |
857 | { |
35a4481c |
858 | const U8 * const send = s + (*len); |
6940069f |
859 | U8 *d; |
860 | U8 *dst; |
96a5add6 |
861 | PERL_UNUSED_CONTEXT; |
6940069f |
862 | |
212542aa |
863 | Newx(d, (*len) * 2 + 1, U8); |
6940069f |
864 | dst = d; |
865 | |
866 | while (s < send) { |
35a4481c |
867 | const UV uv = NATIVE_TO_ASCII(*s++); |
c4d5f83a |
868 | if (UNI_IS_INVARIANT(uv)) |
eb160463 |
869 | *d++ = (U8)UTF_TO_NATIVE(uv); |
6940069f |
870 | else { |
eb160463 |
871 | *d++ = (U8)UTF8_EIGHT_BIT_HI(uv); |
872 | *d++ = (U8)UTF8_EIGHT_BIT_LO(uv); |
6940069f |
873 | } |
874 | } |
875 | *d = '\0'; |
6662521e |
876 | *len = d-dst; |
6940069f |
877 | return dst; |
878 | } |
879 | |
a0ed51b3 |
880 | /* |
dea0fc0b |
881 | * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8. |
a0ed51b3 |
882 | * |
883 | * Destination must be pre-extended to 3/2 source. Do not use in-place. |
884 | * We optimize for native, for obvious reasons. */ |
885 | |
886 | U8* |
dea0fc0b |
887 | Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 |
888 | { |
dea0fc0b |
889 | U8* pend; |
890 | U8* dstart = d; |
891 | |
1de9afcd |
892 | if (bytelen == 1 && p[0] == 0) { /* Be understanding. */ |
893 | d[0] = 0; |
894 | *newlen = 1; |
895 | return d; |
896 | } |
897 | |
dea0fc0b |
898 | if (bytelen & 1) |
f5992bc4 |
899 | Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen %"UVuf, (UV)bytelen); |
dea0fc0b |
900 | |
901 | pend = p + bytelen; |
902 | |
a0ed51b3 |
903 | while (p < pend) { |
dea0fc0b |
904 | UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */ |
905 | p += 2; |
a0ed51b3 |
906 | if (uv < 0x80) { |
e294cc5d |
907 | #ifdef EBCDIC |
908 | *d++ = UNI_TO_NATIVE(uv); |
909 | #else |
eb160463 |
910 | *d++ = (U8)uv; |
e294cc5d |
911 | #endif |
a0ed51b3 |
912 | continue; |
913 | } |
914 | if (uv < 0x800) { |
eb160463 |
915 | *d++ = (U8)(( uv >> 6) | 0xc0); |
916 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
917 | continue; |
918 | } |
919 | if (uv >= 0xd800 && uv < 0xdbff) { /* surrogates */ |
30f84f9e |
920 | UV low = (p[0] << 8) + p[1]; |
921 | p += 2; |
dea0fc0b |
922 | if (low < 0xdc00 || low >= 0xdfff) |
923 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
a0ed51b3 |
924 | uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000; |
925 | } |
926 | if (uv < 0x10000) { |
eb160463 |
927 | *d++ = (U8)(( uv >> 12) | 0xe0); |
928 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
929 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
930 | continue; |
931 | } |
932 | else { |
eb160463 |
933 | *d++ = (U8)(( uv >> 18) | 0xf0); |
934 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
935 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
936 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
937 | continue; |
938 | } |
939 | } |
dea0fc0b |
940 | *newlen = d - dstart; |
a0ed51b3 |
941 | return d; |
942 | } |
943 | |
944 | /* Note: this one is slightly destructive of the source. */ |
945 | |
946 | U8* |
dea0fc0b |
947 | Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 |
948 | { |
949 | U8* s = (U8*)p; |
d4c19fe8 |
950 | U8* const send = s + bytelen; |
a0ed51b3 |
951 | while (s < send) { |
d4c19fe8 |
952 | const U8 tmp = s[0]; |
a0ed51b3 |
953 | s[0] = s[1]; |
954 | s[1] = tmp; |
955 | s += 2; |
956 | } |
dea0fc0b |
957 | return utf16_to_utf8(p, d, bytelen, newlen); |
a0ed51b3 |
958 | } |
959 | |
960 | /* for now these are all defined (inefficiently) in terms of the utf8 versions */ |
961 | |
962 | bool |
84afefe6 |
963 | Perl_is_uni_alnum(pTHX_ UV c) |
a0ed51b3 |
964 | { |
89ebb4a3 |
965 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
966 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
967 | return is_utf8_alnum(tmpbuf); |
968 | } |
969 | |
970 | bool |
84afefe6 |
971 | Perl_is_uni_alnumc(pTHX_ UV c) |
b8c5462f |
972 | { |
89ebb4a3 |
973 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
974 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f |
975 | return is_utf8_alnumc(tmpbuf); |
976 | } |
977 | |
978 | bool |
84afefe6 |
979 | Perl_is_uni_idfirst(pTHX_ UV c) |
a0ed51b3 |
980 | { |
89ebb4a3 |
981 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
982 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
983 | return is_utf8_idfirst(tmpbuf); |
984 | } |
985 | |
986 | bool |
84afefe6 |
987 | Perl_is_uni_alpha(pTHX_ UV c) |
a0ed51b3 |
988 | { |
89ebb4a3 |
989 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
990 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
991 | return is_utf8_alpha(tmpbuf); |
992 | } |
993 | |
994 | bool |
84afefe6 |
995 | Perl_is_uni_ascii(pTHX_ UV c) |
4d61ec05 |
996 | { |
89ebb4a3 |
997 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
998 | uvchr_to_utf8(tmpbuf, c); |
4d61ec05 |
999 | return is_utf8_ascii(tmpbuf); |
1000 | } |
1001 | |
1002 | bool |
84afefe6 |
1003 | Perl_is_uni_space(pTHX_ UV c) |
a0ed51b3 |
1004 | { |
89ebb4a3 |
1005 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1006 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1007 | return is_utf8_space(tmpbuf); |
1008 | } |
1009 | |
1010 | bool |
84afefe6 |
1011 | Perl_is_uni_digit(pTHX_ UV c) |
a0ed51b3 |
1012 | { |
89ebb4a3 |
1013 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1014 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1015 | return is_utf8_digit(tmpbuf); |
1016 | } |
1017 | |
1018 | bool |
84afefe6 |
1019 | Perl_is_uni_upper(pTHX_ UV c) |
a0ed51b3 |
1020 | { |
89ebb4a3 |
1021 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1022 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1023 | return is_utf8_upper(tmpbuf); |
1024 | } |
1025 | |
1026 | bool |
84afefe6 |
1027 | Perl_is_uni_lower(pTHX_ UV c) |
a0ed51b3 |
1028 | { |
89ebb4a3 |
1029 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1030 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1031 | return is_utf8_lower(tmpbuf); |
1032 | } |
1033 | |
1034 | bool |
84afefe6 |
1035 | Perl_is_uni_cntrl(pTHX_ UV c) |
b8c5462f |
1036 | { |
89ebb4a3 |
1037 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1038 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f |
1039 | return is_utf8_cntrl(tmpbuf); |
1040 | } |
1041 | |
1042 | bool |
84afefe6 |
1043 | Perl_is_uni_graph(pTHX_ UV c) |
b8c5462f |
1044 | { |
89ebb4a3 |
1045 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1046 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f |
1047 | return is_utf8_graph(tmpbuf); |
1048 | } |
1049 | |
1050 | bool |
84afefe6 |
1051 | Perl_is_uni_print(pTHX_ UV c) |
a0ed51b3 |
1052 | { |
89ebb4a3 |
1053 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1054 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1055 | return is_utf8_print(tmpbuf); |
1056 | } |
1057 | |
b8c5462f |
1058 | bool |
84afefe6 |
1059 | Perl_is_uni_punct(pTHX_ UV c) |
b8c5462f |
1060 | { |
89ebb4a3 |
1061 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1062 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f |
1063 | return is_utf8_punct(tmpbuf); |
1064 | } |
1065 | |
4d61ec05 |
1066 | bool |
84afefe6 |
1067 | Perl_is_uni_xdigit(pTHX_ UV c) |
4d61ec05 |
1068 | { |
89ebb4a3 |
1069 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
230880c1 |
1070 | uvchr_to_utf8(tmpbuf, c); |
4d61ec05 |
1071 | return is_utf8_xdigit(tmpbuf); |
1072 | } |
1073 | |
84afefe6 |
1074 | UV |
1075 | Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
1076 | { |
0ebc6274 |
1077 | uvchr_to_utf8(p, c); |
1078 | return to_utf8_upper(p, p, lenp); |
a0ed51b3 |
1079 | } |
1080 | |
84afefe6 |
1081 | UV |
1082 | Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
1083 | { |
0ebc6274 |
1084 | uvchr_to_utf8(p, c); |
1085 | return to_utf8_title(p, p, lenp); |
a0ed51b3 |
1086 | } |
1087 | |
84afefe6 |
1088 | UV |
1089 | Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
1090 | { |
0ebc6274 |
1091 | uvchr_to_utf8(p, c); |
1092 | return to_utf8_lower(p, p, lenp); |
a0ed51b3 |
1093 | } |
1094 | |
84afefe6 |
1095 | UV |
1096 | Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp) |
1097 | { |
0ebc6274 |
1098 | uvchr_to_utf8(p, c); |
1099 | return to_utf8_fold(p, p, lenp); |
84afefe6 |
1100 | } |
1101 | |
a0ed51b3 |
1102 | /* for now these all assume no locale info available for Unicode > 255 */ |
1103 | |
1104 | bool |
84afefe6 |
1105 | Perl_is_uni_alnum_lc(pTHX_ UV c) |
a0ed51b3 |
1106 | { |
1107 | return is_uni_alnum(c); /* XXX no locale support yet */ |
1108 | } |
1109 | |
1110 | bool |
84afefe6 |
1111 | Perl_is_uni_alnumc_lc(pTHX_ UV c) |
b8c5462f |
1112 | { |
1113 | return is_uni_alnumc(c); /* XXX no locale support yet */ |
1114 | } |
1115 | |
1116 | bool |
84afefe6 |
1117 | Perl_is_uni_idfirst_lc(pTHX_ UV c) |
a0ed51b3 |
1118 | { |
1119 | return is_uni_idfirst(c); /* XXX no locale support yet */ |
1120 | } |
1121 | |
1122 | bool |
84afefe6 |
1123 | Perl_is_uni_alpha_lc(pTHX_ UV c) |
a0ed51b3 |
1124 | { |
1125 | return is_uni_alpha(c); /* XXX no locale support yet */ |
1126 | } |
1127 | |
1128 | bool |
84afefe6 |
1129 | Perl_is_uni_ascii_lc(pTHX_ UV c) |
4d61ec05 |
1130 | { |
1131 | return is_uni_ascii(c); /* XXX no locale support yet */ |
1132 | } |
1133 | |
1134 | bool |
84afefe6 |
1135 | Perl_is_uni_space_lc(pTHX_ UV c) |
a0ed51b3 |
1136 | { |
1137 | return is_uni_space(c); /* XXX no locale support yet */ |
1138 | } |
1139 | |
1140 | bool |
84afefe6 |
1141 | Perl_is_uni_digit_lc(pTHX_ UV c) |
a0ed51b3 |
1142 | { |
1143 | return is_uni_digit(c); /* XXX no locale support yet */ |
1144 | } |
1145 | |
1146 | bool |
84afefe6 |
1147 | Perl_is_uni_upper_lc(pTHX_ UV c) |
a0ed51b3 |
1148 | { |
1149 | return is_uni_upper(c); /* XXX no locale support yet */ |
1150 | } |
1151 | |
1152 | bool |
84afefe6 |
1153 | Perl_is_uni_lower_lc(pTHX_ UV c) |
a0ed51b3 |
1154 | { |
1155 | return is_uni_lower(c); /* XXX no locale support yet */ |
1156 | } |
1157 | |
1158 | bool |
84afefe6 |
1159 | Perl_is_uni_cntrl_lc(pTHX_ UV c) |
b8c5462f |
1160 | { |
1161 | return is_uni_cntrl(c); /* XXX no locale support yet */ |
1162 | } |
1163 | |
1164 | bool |
84afefe6 |
1165 | Perl_is_uni_graph_lc(pTHX_ UV c) |
b8c5462f |
1166 | { |
1167 | return is_uni_graph(c); /* XXX no locale support yet */ |
1168 | } |
1169 | |
1170 | bool |
84afefe6 |
1171 | Perl_is_uni_print_lc(pTHX_ UV c) |
a0ed51b3 |
1172 | { |
1173 | return is_uni_print(c); /* XXX no locale support yet */ |
1174 | } |
1175 | |
b8c5462f |
1176 | bool |
84afefe6 |
1177 | Perl_is_uni_punct_lc(pTHX_ UV c) |
b8c5462f |
1178 | { |
1179 | return is_uni_punct(c); /* XXX no locale support yet */ |
1180 | } |
1181 | |
4d61ec05 |
1182 | bool |
84afefe6 |
1183 | Perl_is_uni_xdigit_lc(pTHX_ UV c) |
4d61ec05 |
1184 | { |
1185 | return is_uni_xdigit(c); /* XXX no locale support yet */ |
1186 | } |
1187 | |
b7ac61fa |
1188 | U32 |
1189 | Perl_to_uni_upper_lc(pTHX_ U32 c) |
1190 | { |
ee099d14 |
1191 | /* XXX returns only the first character -- do not use XXX */ |
1192 | /* XXX no locale support yet */ |
1193 | STRLEN len; |
89ebb4a3 |
1194 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 |
1195 | return (U32)to_uni_upper(c, tmpbuf, &len); |
b7ac61fa |
1196 | } |
1197 | |
1198 | U32 |
1199 | Perl_to_uni_title_lc(pTHX_ U32 c) |
1200 | { |
ee099d14 |
1201 | /* XXX returns only the first character XXX -- do not use XXX */ |
1202 | /* XXX no locale support yet */ |
1203 | STRLEN len; |
89ebb4a3 |
1204 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 |
1205 | return (U32)to_uni_title(c, tmpbuf, &len); |
b7ac61fa |
1206 | } |
1207 | |
1208 | U32 |
1209 | Perl_to_uni_lower_lc(pTHX_ U32 c) |
1210 | { |
ee099d14 |
1211 | /* XXX returns only the first character -- do not use XXX */ |
1212 | /* XXX no locale support yet */ |
1213 | STRLEN len; |
89ebb4a3 |
1214 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 |
1215 | return (U32)to_uni_lower(c, tmpbuf, &len); |
b7ac61fa |
1216 | } |
1217 | |
7452cf6a |
1218 | static bool |
5141f98e |
1219 | S_is_utf8_common(pTHX_ const U8 *const p, SV **swash, |
bde6a22d |
1220 | const char *const swashname) |
1221 | { |
97aff369 |
1222 | dVAR; |
bde6a22d |
1223 | if (!is_utf8_char(p)) |
1224 | return FALSE; |
1225 | if (!*swash) |
711a919c |
1226 | *swash = swash_init("utf8", swashname, &PL_sv_undef, 1, 0); |
bde6a22d |
1227 | return swash_fetch(*swash, p, TRUE) != 0; |
1228 | } |
1229 | |
1230 | bool |
7fc63493 |
1231 | Perl_is_utf8_alnum(pTHX_ const U8 *p) |
a0ed51b3 |
1232 | { |
97aff369 |
1233 | dVAR; |
671c33bf |
1234 | /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true |
1235 | * descendant of isalnum(3), in other words, it doesn't |
1236 | * contain the '_'. --jhi */ |
d4c19fe8 |
1237 | return is_utf8_common(p, &PL_utf8_alnum, "IsWord"); |
a0ed51b3 |
1238 | } |
1239 | |
1240 | bool |
7fc63493 |
1241 | Perl_is_utf8_alnumc(pTHX_ const U8 *p) |
b8c5462f |
1242 | { |
97aff369 |
1243 | dVAR; |
d4c19fe8 |
1244 | return is_utf8_common(p, &PL_utf8_alnumc, "IsAlnumC"); |
b8c5462f |
1245 | } |
1246 | |
1247 | bool |
7fc63493 |
1248 | Perl_is_utf8_idfirst(pTHX_ const U8 *p) /* The naming is historical. */ |
a0ed51b3 |
1249 | { |
97aff369 |
1250 | dVAR; |
82686b01 |
1251 | if (*p == '_') |
1252 | return TRUE; |
bde6a22d |
1253 | /* is_utf8_idstart would be more logical. */ |
d4c19fe8 |
1254 | return is_utf8_common(p, &PL_utf8_idstart, "IdStart"); |
82686b01 |
1255 | } |
1256 | |
1257 | bool |
7fc63493 |
1258 | Perl_is_utf8_idcont(pTHX_ const U8 *p) |
82686b01 |
1259 | { |
97aff369 |
1260 | dVAR; |
82686b01 |
1261 | if (*p == '_') |
1262 | return TRUE; |
d4c19fe8 |
1263 | return is_utf8_common(p, &PL_utf8_idcont, "IdContinue"); |
a0ed51b3 |
1264 | } |
1265 | |
1266 | bool |
7fc63493 |
1267 | Perl_is_utf8_alpha(pTHX_ const U8 *p) |
a0ed51b3 |
1268 | { |
97aff369 |
1269 | dVAR; |
d4c19fe8 |
1270 | return is_utf8_common(p, &PL_utf8_alpha, "IsAlpha"); |
a0ed51b3 |
1271 | } |
1272 | |
1273 | bool |
7fc63493 |
1274 | Perl_is_utf8_ascii(pTHX_ const U8 *p) |
b8c5462f |
1275 | { |
97aff369 |
1276 | dVAR; |
d4c19fe8 |
1277 | return is_utf8_common(p, &PL_utf8_ascii, "IsAscii"); |
b8c5462f |
1278 | } |
1279 | |
1280 | bool |
7fc63493 |
1281 | Perl_is_utf8_space(pTHX_ const U8 *p) |
a0ed51b3 |
1282 | { |
97aff369 |
1283 | dVAR; |
d4c19fe8 |
1284 | return is_utf8_common(p, &PL_utf8_space, "IsSpacePerl"); |
a0ed51b3 |
1285 | } |
1286 | |
1287 | bool |
7fc63493 |
1288 | Perl_is_utf8_digit(pTHX_ const U8 *p) |
a0ed51b3 |
1289 | { |
97aff369 |
1290 | dVAR; |
d4c19fe8 |
1291 | return is_utf8_common(p, &PL_utf8_digit, "IsDigit"); |
a0ed51b3 |
1292 | } |
1293 | |
1294 | bool |
7fc63493 |
1295 | Perl_is_utf8_upper(pTHX_ const U8 *p) |
a0ed51b3 |
1296 | { |
97aff369 |
1297 | dVAR; |
d4c19fe8 |
1298 | return is_utf8_common(p, &PL_utf8_upper, "IsUppercase"); |
a0ed51b3 |
1299 | } |
1300 | |
1301 | bool |
7fc63493 |
1302 | Perl_is_utf8_lower(pTHX_ const U8 *p) |
a0ed51b3 |
1303 | { |
97aff369 |
1304 | dVAR; |
d4c19fe8 |
1305 | return is_utf8_common(p, &PL_utf8_lower, "IsLowercase"); |
a0ed51b3 |
1306 | } |
1307 | |
1308 | bool |
7fc63493 |
1309 | Perl_is_utf8_cntrl(pTHX_ const U8 *p) |
b8c5462f |
1310 | { |
97aff369 |
1311 | dVAR; |
d4c19fe8 |
1312 | return is_utf8_common(p, &PL_utf8_cntrl, "IsCntrl"); |
b8c5462f |
1313 | } |
1314 | |
1315 | bool |
7fc63493 |
1316 | Perl_is_utf8_graph(pTHX_ const U8 *p) |
b8c5462f |
1317 | { |
97aff369 |
1318 | dVAR; |
d4c19fe8 |
1319 | return is_utf8_common(p, &PL_utf8_graph, "IsGraph"); |
b8c5462f |
1320 | } |
1321 | |
1322 | bool |
7fc63493 |
1323 | Perl_is_utf8_print(pTHX_ const U8 *p) |
a0ed51b3 |
1324 | { |
97aff369 |
1325 | dVAR; |
d4c19fe8 |
1326 | return is_utf8_common(p, &PL_utf8_print, "IsPrint"); |
a0ed51b3 |
1327 | } |
1328 | |
1329 | bool |
7fc63493 |
1330 | Perl_is_utf8_punct(pTHX_ const U8 *p) |
b8c5462f |
1331 | { |
97aff369 |
1332 | dVAR; |
d4c19fe8 |
1333 | return is_utf8_common(p, &PL_utf8_punct, "IsPunct"); |
b8c5462f |
1334 | } |
1335 | |
1336 | bool |
7fc63493 |
1337 | Perl_is_utf8_xdigit(pTHX_ const U8 *p) |
b8c5462f |
1338 | { |
97aff369 |
1339 | dVAR; |
d4c19fe8 |
1340 | return is_utf8_common(p, &PL_utf8_xdigit, "Isxdigit"); |
b8c5462f |
1341 | } |
1342 | |
1343 | bool |
7fc63493 |
1344 | Perl_is_utf8_mark(pTHX_ const U8 *p) |
a0ed51b3 |
1345 | { |
97aff369 |
1346 | dVAR; |
d4c19fe8 |
1347 | return is_utf8_common(p, &PL_utf8_mark, "IsM"); |
a0ed51b3 |
1348 | } |
1349 | |
6b5c0936 |
1350 | /* |
1351 | =for apidoc A|UV|to_utf8_case|U8 *p|U8* ustrp|STRLEN *lenp|SV **swash|char *normal|char *special |
1352 | |
1353 | The "p" contains the pointer to the UTF-8 string encoding |
1354 | the character that is being converted. |
1355 | |
1356 | The "ustrp" is a pointer to the character buffer to put the |
1357 | conversion result to. The "lenp" is a pointer to the length |
1358 | of the result. |
1359 | |
0134edef |
1360 | The "swashp" is a pointer to the swash to use. |
6b5c0936 |
1361 | |
0134edef |
1362 | Both the special and normal mappings are stored lib/unicore/To/Foo.pl, |
8fe4d5b2 |
1363 | and loaded by SWASHNEW, using lib/utf8_heavy.pl. The special (usually, |
0134edef |
1364 | but not always, a multicharacter mapping), is tried first. |
6b5c0936 |
1365 | |
0134edef |
1366 | The "special" is a string like "utf8::ToSpecLower", which means the |
1367 | hash %utf8::ToSpecLower. The access to the hash is through |
1368 | Perl_to_utf8_case(). |
6b5c0936 |
1369 | |
0134edef |
1370 | The "normal" is a string like "ToLower" which means the swash |
1371 | %utf8::ToLower. |
1372 | |
1373 | =cut */ |
6b5c0936 |
1374 | |
2104c8d9 |
1375 | UV |
9a957fbc |
1376 | Perl_to_utf8_case(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp, |
1377 | SV **swashp, const char *normal, const char *special) |
a0ed51b3 |
1378 | { |
97aff369 |
1379 | dVAR; |
89ebb4a3 |
1380 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
0134edef |
1381 | STRLEN len = 0; |
a0ed51b3 |
1382 | |
aec46f14 |
1383 | const UV uv0 = utf8_to_uvchr(p, NULL); |
1feea2c7 |
1384 | /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings |
1385 | * are necessary in EBCDIC, they are redundant no-ops |
1386 | * in ASCII-ish platforms, and hopefully optimized away. */ |
f54cb97a |
1387 | const UV uv1 = NATIVE_TO_UNI(uv0); |
1feea2c7 |
1388 | uvuni_to_utf8(tmpbuf, uv1); |
0134edef |
1389 | |
1390 | if (!*swashp) /* load on-demand */ |
1391 | *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0); |
1392 | |
b08cf34e |
1393 | /* The 0xDF is the only special casing Unicode code point below 0x100. */ |
1394 | if (special && (uv1 == 0xDF || uv1 > 0xFF)) { |
0134edef |
1395 | /* It might be "special" (sometimes, but not always, |
2a37f04d |
1396 | * a multicharacter mapping) */ |
35da51f7 |
1397 | HV * const hv = get_hv(special, FALSE); |
b08cf34e |
1398 | SV **svp; |
1399 | |
35da51f7 |
1400 | if (hv && |
b08cf34e |
1401 | (svp = hv_fetch(hv, (const char*)tmpbuf, UNISKIP(uv1), FALSE)) && |
1402 | (*svp)) { |
cfd0369c |
1403 | const char *s; |
47654450 |
1404 | |
cfd0369c |
1405 | s = SvPV_const(*svp, len); |
47654450 |
1406 | if (len == 1) |
1407 | len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp; |
2a37f04d |
1408 | else { |
2f9475ad |
1409 | #ifdef EBCDIC |
1410 | /* If we have EBCDIC we need to remap the characters |
1411 | * since any characters in the low 256 are Unicode |
1412 | * code points, not EBCDIC. */ |
7cda7a3d |
1413 | U8 *t = (U8*)s, *tend = t + len, *d; |
2f9475ad |
1414 | |
1415 | d = tmpbuf; |
b08cf34e |
1416 | if (SvUTF8(*svp)) { |
2f9475ad |
1417 | STRLEN tlen = 0; |
1418 | |
1419 | while (t < tend) { |
d4c19fe8 |
1420 | const UV c = utf8_to_uvchr(t, &tlen); |
2f9475ad |
1421 | if (tlen > 0) { |
1422 | d = uvchr_to_utf8(d, UNI_TO_NATIVE(c)); |
1423 | t += tlen; |
1424 | } |
1425 | else |
1426 | break; |
1427 | } |
1428 | } |
1429 | else { |
36fec512 |
1430 | while (t < tend) { |
1431 | d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t)); |
1432 | t++; |
1433 | } |
2f9475ad |
1434 | } |
1435 | len = d - tmpbuf; |
1436 | Copy(tmpbuf, ustrp, len, U8); |
1437 | #else |
d2dcd0fb |
1438 | Copy(s, ustrp, len, U8); |
2f9475ad |
1439 | #endif |
29e98929 |
1440 | } |
983ffd37 |
1441 | } |
0134edef |
1442 | } |
1443 | |
1444 | if (!len && *swashp) { |
d4c19fe8 |
1445 | const UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE); |
1446 | |
0134edef |
1447 | if (uv2) { |
1448 | /* It was "normal" (a single character mapping). */ |
d4c19fe8 |
1449 | const UV uv3 = UNI_TO_NATIVE(uv2); |
e9101d72 |
1450 | len = uvchr_to_utf8(ustrp, uv3) - ustrp; |
2a37f04d |
1451 | } |
1452 | } |
1feea2c7 |
1453 | |
0134edef |
1454 | if (!len) /* Neither: just copy. */ |
1455 | len = uvchr_to_utf8(ustrp, uv0) - ustrp; |
1456 | |
2a37f04d |
1457 | if (lenp) |
1458 | *lenp = len; |
1459 | |
0134edef |
1460 | return len ? utf8_to_uvchr(ustrp, 0) : 0; |
a0ed51b3 |
1461 | } |
1462 | |
d3e79532 |
1463 | /* |
7fc63493 |
1464 | =for apidoc A|UV|to_utf8_upper|const U8 *p|U8 *ustrp|STRLEN *lenp |
d3e79532 |
1465 | |
1466 | Convert the UTF-8 encoded character at p to its uppercase version and |
1467 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1468 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since |
1469 | the uppercase version may be longer than the original character. |
d3e79532 |
1470 | |
1471 | The first character of the uppercased version is returned |
1472 | (but note, as explained above, that there may be more.) |
1473 | |
1474 | =cut */ |
1475 | |
2104c8d9 |
1476 | UV |
7fc63493 |
1477 | Perl_to_utf8_upper(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 |
1478 | { |
97aff369 |
1479 | dVAR; |
983ffd37 |
1480 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1481 | &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper"); |
983ffd37 |
1482 | } |
a0ed51b3 |
1483 | |
d3e79532 |
1484 | /* |
7fc63493 |
1485 | =for apidoc A|UV|to_utf8_title|const U8 *p|U8 *ustrp|STRLEN *lenp |
d3e79532 |
1486 | |
1487 | Convert the UTF-8 encoded character at p to its titlecase version and |
1488 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1489 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
1490 | titlecase version may be longer than the original character. |
d3e79532 |
1491 | |
1492 | The first character of the titlecased version is returned |
1493 | (but note, as explained above, that there may be more.) |
1494 | |
1495 | =cut */ |
1496 | |
983ffd37 |
1497 | UV |
7fc63493 |
1498 | Perl_to_utf8_title(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
983ffd37 |
1499 | { |
97aff369 |
1500 | dVAR; |
983ffd37 |
1501 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1502 | &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle"); |
a0ed51b3 |
1503 | } |
1504 | |
d3e79532 |
1505 | /* |
7fc63493 |
1506 | =for apidoc A|UV|to_utf8_lower|const U8 *p|U8 *ustrp|STRLEN *lenp |
d3e79532 |
1507 | |
1508 | Convert the UTF-8 encoded character at p to its lowercase version and |
1509 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1510 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
1511 | lowercase version may be longer than the original character. |
d3e79532 |
1512 | |
1513 | The first character of the lowercased version is returned |
1514 | (but note, as explained above, that there may be more.) |
1515 | |
1516 | =cut */ |
1517 | |
2104c8d9 |
1518 | UV |
7fc63493 |
1519 | Perl_to_utf8_lower(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 |
1520 | { |
97aff369 |
1521 | dVAR; |
983ffd37 |
1522 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1523 | &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower"); |
1524 | } |
1525 | |
d3e79532 |
1526 | /* |
7fc63493 |
1527 | =for apidoc A|UV|to_utf8_fold|const U8 *p|U8 *ustrp|STRLEN *lenp |
d3e79532 |
1528 | |
1529 | Convert the UTF-8 encoded character at p to its foldcase version and |
1530 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1531 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
d3e79532 |
1532 | foldcase version may be longer than the original character (up to |
1533 | three characters). |
1534 | |
1535 | The first character of the foldcased version is returned |
1536 | (but note, as explained above, that there may be more.) |
1537 | |
1538 | =cut */ |
1539 | |
b4e400f9 |
1540 | UV |
7fc63493 |
1541 | Perl_to_utf8_fold(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
b4e400f9 |
1542 | { |
97aff369 |
1543 | dVAR; |
b4e400f9 |
1544 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
1545 | &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold"); |
a0ed51b3 |
1546 | } |
1547 | |
711a919c |
1548 | /* Note: |
1549 | * A "swash" is a swatch hash. |
1550 | * A "swatch" is a bit vector generated by utf8.c:S_swash_get(). |
1551 | * C<pkg> is a pointer to a package name for SWASHNEW, should be "utf8". |
1552 | * For other parameters, see utf8::SWASHNEW in lib/utf8_heavy.pl. |
1553 | */ |
a0ed51b3 |
1554 | SV* |
7fc63493 |
1555 | Perl_swash_init(pTHX_ const char* pkg, const char* name, SV *listsv, I32 minbits, I32 none) |
a0ed51b3 |
1556 | { |
27da23d5 |
1557 | dVAR; |
a0ed51b3 |
1558 | SV* retval; |
8e84507e |
1559 | dSP; |
7fc63493 |
1560 | const size_t pkg_len = strlen(pkg); |
1561 | const size_t name_len = strlen(name); |
da51bb9b |
1562 | HV * const stash = gv_stashpvn(pkg, pkg_len, 0); |
f8be5cf0 |
1563 | SV* errsv_save; |
ce3b816e |
1564 | |
96ca9f55 |
1565 | PUSHSTACKi(PERLSI_MAGIC); |
1566 | ENTER; |
1567 | SAVEI32(PL_hints); |
1568 | PL_hints = 0; |
1569 | save_re_context(); |
1b026014 |
1570 | if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */ |
ce3b816e |
1571 | ENTER; |
f8be5cf0 |
1572 | errsv_save = newSVsv(ERRSV); |
dc0c6abb |
1573 | /* It is assumed that callers of this routine are not passing in any |
1574 | user derived data. */ |
1575 | /* Need to do this after save_re_context() as it will set PL_tainted to |
1576 | 1 while saving $1 etc (see the code after getrx: in Perl_magic_get). |
1577 | Even line to create errsv_save can turn on PL_tainted. */ |
1578 | SAVEBOOL(PL_tainted); |
1579 | PL_tainted = 0; |
71bed85a |
1580 | Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpvn(pkg,pkg_len), |
a0714e2c |
1581 | NULL); |
f8be5cf0 |
1582 | if (!SvTRUE(ERRSV)) |
1583 | sv_setsv(ERRSV, errsv_save); |
1584 | SvREFCNT_dec(errsv_save); |
ce3b816e |
1585 | LEAVE; |
1586 | } |
1587 | SPAGAIN; |
a0ed51b3 |
1588 | PUSHMARK(SP); |
1589 | EXTEND(SP,5); |
71bed85a |
1590 | PUSHs(sv_2mortal(newSVpvn(pkg, pkg_len))); |
1591 | PUSHs(sv_2mortal(newSVpvn(name, name_len))); |
a0ed51b3 |
1592 | PUSHs(listsv); |
1593 | PUSHs(sv_2mortal(newSViv(minbits))); |
1594 | PUSHs(sv_2mortal(newSViv(none))); |
1595 | PUTBACK; |
f8be5cf0 |
1596 | errsv_save = newSVsv(ERRSV); |
864dbfa3 |
1597 | if (call_method("SWASHNEW", G_SCALAR)) |
8e84507e |
1598 | retval = newSVsv(*PL_stack_sp--); |
a0ed51b3 |
1599 | else |
e24b16f9 |
1600 | retval = &PL_sv_undef; |
f8be5cf0 |
1601 | if (!SvTRUE(ERRSV)) |
1602 | sv_setsv(ERRSV, errsv_save); |
1603 | SvREFCNT_dec(errsv_save); |
a0ed51b3 |
1604 | LEAVE; |
1605 | POPSTACK; |
923e4eb5 |
1606 | if (IN_PERL_COMPILETIME) { |
623e6609 |
1607 | CopHINTS_set(PL_curcop, PL_hints); |
a0ed51b3 |
1608 | } |
bc45ce41 |
1609 | if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) { |
1610 | if (SvPOK(retval)) |
35c1215d |
1611 | Perl_croak(aTHX_ "Can't find Unicode property definition \"%"SVf"\"", |
be2597df |
1612 | SVfARG(retval)); |
cea2e8a9 |
1613 | Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref"); |
bc45ce41 |
1614 | } |
a0ed51b3 |
1615 | return retval; |
1616 | } |
1617 | |
035d37be |
1618 | |
1619 | /* This API is wrong for special case conversions since we may need to |
1620 | * return several Unicode characters for a single Unicode character |
1621 | * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is |
1622 | * the lower-level routine, and it is similarly broken for returning |
1623 | * multiple values. --jhi */ |
979f2922 |
1624 | /* Now SWASHGET is recasted into S_swash_get in this file. */ |
680c470c |
1625 | |
1626 | /* Note: |
1627 | * Returns the value of property/mapping C<swash> for the first character |
1628 | * of the string C<ptr>. If C<do_utf8> is true, the string C<ptr> is |
1629 | * assumed to be in utf8. If C<do_utf8> is false, the string C<ptr> is |
1630 | * assumed to be in native 8-bit encoding. Caches the swatch in C<swash>. |
1631 | */ |
a0ed51b3 |
1632 | UV |
680c470c |
1633 | Perl_swash_fetch(pTHX_ SV *swash, const U8 *ptr, bool do_utf8) |
a0ed51b3 |
1634 | { |
27da23d5 |
1635 | dVAR; |
680c470c |
1636 | HV* const hv = (HV*)SvRV(swash); |
3568d838 |
1637 | U32 klen; |
1638 | U32 off; |
a0ed51b3 |
1639 | STRLEN slen; |
7d85a32c |
1640 | STRLEN needents; |
cfd0369c |
1641 | const U8 *tmps = NULL; |
a0ed51b3 |
1642 | U32 bit; |
979f2922 |
1643 | SV *swatch; |
3568d838 |
1644 | U8 tmputf8[2]; |
35da51f7 |
1645 | const UV c = NATIVE_TO_ASCII(*ptr); |
3568d838 |
1646 | |
1647 | if (!do_utf8 && !UNI_IS_INVARIANT(c)) { |
979f2922 |
1648 | tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c); |
1649 | tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c); |
1650 | ptr = tmputf8; |
3568d838 |
1651 | } |
1652 | /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ |
1653 | * then the "swatch" is a vec() for al the chars which start |
1654 | * with 0xAA..0xYY |
1655 | * So the key in the hash (klen) is length of encoded char -1 |
1656 | */ |
1657 | klen = UTF8SKIP(ptr) - 1; |
1658 | off = ptr[klen]; |
a0ed51b3 |
1659 | |
979f2922 |
1660 | if (klen == 0) { |
7d85a32c |
1661 | /* If char in invariant then swatch is for all the invariant chars |
1e54db1a |
1662 | * In both UTF-8 and UTF-8-MOD that happens to be UTF_CONTINUATION_MARK |
7d85a32c |
1663 | */ |
979f2922 |
1664 | needents = UTF_CONTINUATION_MARK; |
1665 | off = NATIVE_TO_UTF(ptr[klen]); |
1666 | } |
1667 | else { |
7d85a32c |
1668 | /* If char is encoded then swatch is for the prefix */ |
979f2922 |
1669 | needents = (1 << UTF_ACCUMULATION_SHIFT); |
1670 | off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK; |
1671 | } |
7d85a32c |
1672 | |
a0ed51b3 |
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 | |
3568d838 |
1682 | if (hv == PL_last_swash_hv && |
a0ed51b3 |
1683 | klen == PL_last_swash_klen && |
27da23d5 |
1684 | (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) ) |
a0ed51b3 |
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. */ |
e1ec3a88 |
1691 | SV** svp = hv_fetch(hv, (const char*)ptr, klen, FALSE); |
a0ed51b3 |
1692 | |
979f2922 |
1693 | /* If not cached, generate it via swash_get */ |
1694 | if (!svp || !SvPOK(*svp) |
1695 | || !(tmps = (const U8*)SvPV_const(*svp, slen))) { |
2b9d42f0 |
1696 | /* We use utf8n_to_uvuni() as we want an index into |
1697 | Unicode tables, not a native character number. |
1698 | */ |
aec46f14 |
1699 | const UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXBYTES, 0, |
872c91ae |
1700 | ckWARN(WARN_UTF8) ? |
1701 | 0 : UTF8_ALLOW_ANY); |
680c470c |
1702 | swatch = swash_get(swash, |
979f2922 |
1703 | /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */ |
1704 | (klen) ? (code_point & ~(needents - 1)) : 0, |
1705 | needents); |
1706 | |
923e4eb5 |
1707 | if (IN_PERL_COMPILETIME) |
623e6609 |
1708 | CopHINTS_set(PL_curcop, PL_hints); |
a0ed51b3 |
1709 | |
979f2922 |
1710 | svp = hv_store(hv, (const char *)ptr, klen, swatch, 0); |
a0ed51b3 |
1711 | |
979f2922 |
1712 | if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) |
1713 | || (slen << 3) < needents) |
660a4616 |
1714 | Perl_croak(aTHX_ "panic: swash_fetch got improper swatch"); |
a0ed51b3 |
1715 | } |
1716 | |
1717 | PL_last_swash_hv = hv; |
16d8f38a |
1718 | assert(klen <= sizeof(PL_last_swash_key)); |
eac04b2e |
1719 | PL_last_swash_klen = (U8)klen; |
cfd0369c |
1720 | /* FIXME change interpvar.h? */ |
1721 | PL_last_swash_tmps = (U8 *) tmps; |
a0ed51b3 |
1722 | PL_last_swash_slen = slen; |
1723 | if (klen) |
1724 | Copy(ptr, PL_last_swash_key, klen, U8); |
1725 | } |
1726 | |
9faf8d75 |
1727 | switch ((int)((slen << 3) / needents)) { |
a0ed51b3 |
1728 | case 1: |
1729 | bit = 1 << (off & 7); |
1730 | off >>= 3; |
1731 | return (tmps[off] & bit) != 0; |
1732 | case 8: |
1733 | return tmps[off]; |
1734 | case 16: |
1735 | off <<= 1; |
1736 | return (tmps[off] << 8) + tmps[off + 1] ; |
1737 | case 32: |
1738 | off <<= 2; |
1739 | return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ; |
1740 | } |
660a4616 |
1741 | Perl_croak(aTHX_ "panic: swash_fetch got swatch of unexpected bit width"); |
670f1322 |
1742 | NORETURN_FUNCTION_END; |
a0ed51b3 |
1743 | } |
2b9d42f0 |
1744 | |
979f2922 |
1745 | /* Note: |
1746 | * Returns a swatch (a bit vector string) for a code point sequence |
1747 | * that starts from the value C<start> and comprises the number C<span>. |
1748 | * A C<swash> must be an object created by SWASHNEW (see lib/utf8_heavy.pl). |
1749 | * Should be used via swash_fetch, which will cache the swatch in C<swash>. |
1750 | */ |
1751 | STATIC SV* |
1752 | S_swash_get(pTHX_ SV* swash, UV start, UV span) |
1753 | { |
1754 | SV *swatch; |
711a919c |
1755 | U8 *l, *lend, *x, *xend, *s; |
979f2922 |
1756 | STRLEN lcur, xcur, scur; |
1757 | |
1758 | HV* const hv = (HV*)SvRV(swash); |
017a3ce5 |
1759 | SV** const listsvp = hv_fetchs(hv, "LIST", FALSE); |
1760 | SV** const typesvp = hv_fetchs(hv, "TYPE", FALSE); |
1761 | SV** const bitssvp = hv_fetchs(hv, "BITS", FALSE); |
1762 | SV** const nonesvp = hv_fetchs(hv, "NONE", FALSE); |
1763 | SV** const extssvp = hv_fetchs(hv, "EXTRAS", FALSE); |
0bd48802 |
1764 | const U8* const typestr = (U8*)SvPV_nolen(*typesvp); |
1765 | const int typeto = typestr[0] == 'T' && typestr[1] == 'o'; |
1766 | const STRLEN bits = SvUV(*bitssvp); |
1767 | const STRLEN octets = bits >> 3; /* if bits == 1, then octets == 0 */ |
1768 | const UV none = SvUV(*nonesvp); |
1769 | const UV end = start + span; |
979f2922 |
1770 | |
1771 | if (bits != 1 && bits != 8 && bits != 16 && bits != 32) { |
660a4616 |
1772 | Perl_croak(aTHX_ "panic: swash_get doesn't expect bits %"UVuf, |
1773 | (UV)bits); |
979f2922 |
1774 | } |
1775 | |
1776 | /* create and initialize $swatch */ |
979f2922 |
1777 | scur = octets ? (span * octets) : (span + 7) / 8; |
e524fe40 |
1778 | swatch = newSV(scur); |
1779 | SvPOK_on(swatch); |
979f2922 |
1780 | s = (U8*)SvPVX(swatch); |
1781 | if (octets && none) { |
0bd48802 |
1782 | const U8* const e = s + scur; |
979f2922 |
1783 | while (s < e) { |
1784 | if (bits == 8) |
1785 | *s++ = (U8)(none & 0xff); |
1786 | else if (bits == 16) { |
1787 | *s++ = (U8)((none >> 8) & 0xff); |
1788 | *s++ = (U8)( none & 0xff); |
1789 | } |
1790 | else if (bits == 32) { |
1791 | *s++ = (U8)((none >> 24) & 0xff); |
1792 | *s++ = (U8)((none >> 16) & 0xff); |
1793 | *s++ = (U8)((none >> 8) & 0xff); |
1794 | *s++ = (U8)( none & 0xff); |
1795 | } |
1796 | } |
1797 | *s = '\0'; |
1798 | } |
1799 | else { |
1800 | (void)memzero((U8*)s, scur + 1); |
1801 | } |
1802 | SvCUR_set(swatch, scur); |
1803 | s = (U8*)SvPVX(swatch); |
1804 | |
1805 | /* read $swash->{LIST} */ |
1806 | l = (U8*)SvPV(*listsvp, lcur); |
1807 | lend = l + lcur; |
1808 | while (l < lend) { |
35da51f7 |
1809 | UV min, max, val; |
979f2922 |
1810 | STRLEN numlen; |
1811 | I32 flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX; |
1812 | |
0bd48802 |
1813 | U8* const nl = (U8*)memchr(l, '\n', lend - l); |
979f2922 |
1814 | |
1815 | numlen = lend - l; |
1816 | min = grok_hex((char *)l, &numlen, &flags, NULL); |
1817 | if (numlen) |
1818 | l += numlen; |
1819 | else if (nl) { |
1820 | l = nl + 1; /* 1 is length of "\n" */ |
1821 | continue; |
1822 | } |
1823 | else { |
1824 | l = lend; /* to LIST's end at which \n is not found */ |
1825 | break; |
1826 | } |
1827 | |
1828 | if (isBLANK(*l)) { |
1829 | ++l; |
1830 | flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX; |
1831 | numlen = lend - l; |
1832 | max = grok_hex((char *)l, &numlen, &flags, NULL); |
1833 | if (numlen) |
1834 | l += numlen; |
1835 | else |
1836 | max = min; |
1837 | |
1838 | if (octets) { |
1839 | if (isBLANK(*l)) { |
1840 | ++l; |
1841 | flags = PERL_SCAN_SILENT_ILLDIGIT | |
1842 | PERL_SCAN_DISALLOW_PREFIX; |
1843 | numlen = lend - l; |
1844 | val = grok_hex((char *)l, &numlen, &flags, NULL); |
1845 | if (numlen) |
1846 | l += numlen; |
1847 | else |
1848 | val = 0; |
1849 | } |
1850 | else { |
1851 | val = 0; |
1852 | if (typeto) { |
1853 | Perl_croak(aTHX_ "%s: illegal mapping '%s'", |
1854 | typestr, l); |
1855 | } |
1856 | } |
1857 | } |
711a919c |
1858 | else |
1859 | val = 0; /* bits == 1, then val should be ignored */ |
979f2922 |
1860 | } |
1861 | else { |
1862 | max = min; |
1863 | if (octets) { |
1864 | val = 0; |
1865 | if (typeto) { |
1866 | Perl_croak(aTHX_ "%s: illegal mapping '%s'", typestr, l); |
1867 | } |
1868 | } |
711a919c |
1869 | else |
1870 | val = 0; /* bits == 1, then val should be ignored */ |
979f2922 |
1871 | } |
1872 | |
1873 | if (nl) |
1874 | l = nl + 1; |
1875 | else |
1876 | l = lend; |
1877 | |
1878 | if (max < start) |
1879 | continue; |
1880 | |
1881 | if (octets) { |
35da51f7 |
1882 | UV key; |
979f2922 |
1883 | if (min < start) { |
1884 | if (!none || val < none) { |
1885 | val += start - min; |
1886 | } |
1887 | min = start; |
1888 | } |
1889 | for (key = min; key <= max; key++) { |
1890 | STRLEN offset; |
1891 | if (key >= end) |
1892 | goto go_out_list; |
1893 | /* offset must be non-negative (start <= min <= key < end) */ |
1894 | offset = octets * (key - start); |
1895 | if (bits == 8) |
1896 | s[offset] = (U8)(val & 0xff); |
1897 | else if (bits == 16) { |
1898 | s[offset ] = (U8)((val >> 8) & 0xff); |
1899 | s[offset + 1] = (U8)( val & 0xff); |
1900 | } |
1901 | else if (bits == 32) { |
1902 | s[offset ] = (U8)((val >> 24) & 0xff); |
1903 | s[offset + 1] = (U8)((val >> 16) & 0xff); |
1904 | s[offset + 2] = (U8)((val >> 8) & 0xff); |
1905 | s[offset + 3] = (U8)( val & 0xff); |
1906 | } |
1907 | |
1908 | if (!none || val < none) |
1909 | ++val; |
1910 | } |
1911 | } |
711a919c |
1912 | else { /* bits == 1, then val should be ignored */ |
35da51f7 |
1913 | UV key; |
979f2922 |
1914 | if (min < start) |
1915 | min = start; |
1916 | for (key = min; key <= max; key++) { |
0bd48802 |
1917 | const STRLEN offset = (STRLEN)(key - start); |
979f2922 |
1918 | if (key >= end) |
1919 | goto go_out_list; |
1920 | s[offset >> 3] |= 1 << (offset & 7); |
1921 | } |
1922 | } |
1923 | } /* while */ |
1924 | go_out_list: |
1925 | |
1926 | /* read $swash->{EXTRAS} */ |
1927 | x = (U8*)SvPV(*extssvp, xcur); |
1928 | xend = x + xcur; |
1929 | while (x < xend) { |
1930 | STRLEN namelen; |
1931 | U8 *namestr; |
1932 | SV** othersvp; |
1933 | HV* otherhv; |
1934 | STRLEN otherbits; |
1935 | SV **otherbitssvp, *other; |
711a919c |
1936 | U8 *s, *o, *nl; |
979f2922 |
1937 | STRLEN slen, olen; |
1938 | |
35da51f7 |
1939 | const U8 opc = *x++; |
979f2922 |
1940 | if (opc == '\n') |
1941 | continue; |
1942 | |
1943 | nl = (U8*)memchr(x, '\n', xend - x); |
1944 | |
1945 | if (opc != '-' && opc != '+' && opc != '!' && opc != '&') { |
1946 | if (nl) { |
1947 | x = nl + 1; /* 1 is length of "\n" */ |
1948 | continue; |
1949 | } |
1950 | else { |
1951 | x = xend; /* to EXTRAS' end at which \n is not found */ |
1952 | break; |
1953 | } |
1954 | } |
1955 | |
1956 | namestr = x; |
1957 | if (nl) { |
1958 | namelen = nl - namestr; |
1959 | x = nl + 1; |
1960 | } |
1961 | else { |
1962 | namelen = xend - namestr; |
1963 | x = xend; |
1964 | } |
1965 | |
1966 | othersvp = hv_fetch(hv, (char *)namestr, namelen, FALSE); |
660a4616 |
1967 | otherhv = (HV*)SvRV(*othersvp); |
017a3ce5 |
1968 | otherbitssvp = hv_fetchs(otherhv, "BITS", FALSE); |
979f2922 |
1969 | otherbits = (STRLEN)SvUV(*otherbitssvp); |
1970 | if (bits < otherbits) |
660a4616 |
1971 | Perl_croak(aTHX_ "panic: swash_get found swatch size mismatch"); |
979f2922 |
1972 | |
1973 | /* The "other" swatch must be destroyed after. */ |
1974 | other = swash_get(*othersvp, start, span); |
1975 | o = (U8*)SvPV(other, olen); |
1976 | |
1977 | if (!olen) |
660a4616 |
1978 | Perl_croak(aTHX_ "panic: swash_get got improper swatch"); |
979f2922 |
1979 | |
1980 | s = (U8*)SvPV(swatch, slen); |
1981 | if (bits == 1 && otherbits == 1) { |
1982 | if (slen != olen) |
660a4616 |
1983 | Perl_croak(aTHX_ "panic: swash_get found swatch length mismatch"); |
979f2922 |
1984 | |
1985 | switch (opc) { |
1986 | case '+': |
1987 | while (slen--) |
1988 | *s++ |= *o++; |
1989 | break; |
1990 | case '!': |
1991 | while (slen--) |
1992 | *s++ |= ~*o++; |
1993 | break; |
1994 | case '-': |
1995 | while (slen--) |
1996 | *s++ &= ~*o++; |
1997 | break; |
1998 | case '&': |
1999 | while (slen--) |
2000 | *s++ &= *o++; |
2001 | break; |
2002 | default: |
2003 | break; |
2004 | } |
2005 | } |
711a919c |
2006 | else { |
979f2922 |
2007 | STRLEN otheroctets = otherbits >> 3; |
2008 | STRLEN offset = 0; |
35da51f7 |
2009 | U8* const send = s + slen; |
979f2922 |
2010 | |
2011 | while (s < send) { |
2012 | UV otherval = 0; |
2013 | |
2014 | if (otherbits == 1) { |
2015 | otherval = (o[offset >> 3] >> (offset & 7)) & 1; |
2016 | ++offset; |
2017 | } |
2018 | else { |
2019 | STRLEN vlen = otheroctets; |
2020 | otherval = *o++; |
2021 | while (--vlen) { |
2022 | otherval <<= 8; |
2023 | otherval |= *o++; |
2024 | } |
2025 | } |
2026 | |
711a919c |
2027 | if (opc == '+' && otherval) |
6f207bd3 |
2028 | NOOP; /* replace with otherval */ |
979f2922 |
2029 | else if (opc == '!' && !otherval) |
2030 | otherval = 1; |
2031 | else if (opc == '-' && otherval) |
2032 | otherval = 0; |
2033 | else if (opc == '&' && !otherval) |
2034 | otherval = 0; |
2035 | else { |
711a919c |
2036 | s += octets; /* no replacement */ |
979f2922 |
2037 | continue; |
2038 | } |
2039 | |
2040 | if (bits == 8) |
2041 | *s++ = (U8)( otherval & 0xff); |
2042 | else if (bits == 16) { |
2043 | *s++ = (U8)((otherval >> 8) & 0xff); |
2044 | *s++ = (U8)( otherval & 0xff); |
2045 | } |
2046 | else if (bits == 32) { |
2047 | *s++ = (U8)((otherval >> 24) & 0xff); |
2048 | *s++ = (U8)((otherval >> 16) & 0xff); |
2049 | *s++ = (U8)((otherval >> 8) & 0xff); |
2050 | *s++ = (U8)( otherval & 0xff); |
2051 | } |
2052 | } |
2053 | } |
2054 | sv_free(other); /* through with it! */ |
2055 | } /* while */ |
2056 | return swatch; |
2057 | } |
2058 | |
0f830e0b |
2059 | /* |
2060 | =for apidoc A|U8 *|uvchr_to_utf8|U8 *d|UV uv |
2061 | |
2062 | Adds the UTF-8 representation of the Native codepoint C<uv> to the end |
2063 | of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free |
2064 | bytes available. The return value is the pointer to the byte after the |
2065 | end of the new character. In other words, |
2066 | |
2067 | d = uvchr_to_utf8(d, uv); |
2068 | |
2069 | is the recommended wide native character-aware way of saying |
2070 | |
2071 | *(d++) = uv; |
2072 | |
2073 | =cut |
2074 | */ |
2075 | |
2076 | /* On ASCII machines this is normally a macro but we want a |
2077 | real function in case XS code wants it |
2078 | */ |
0f830e0b |
2079 | U8 * |
2080 | Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv) |
2081 | { |
2082 | return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0); |
2083 | } |
2084 | |
b851fbc1 |
2085 | U8 * |
2086 | Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) |
2087 | { |
2088 | return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags); |
2089 | } |
2b9d42f0 |
2090 | |
2091 | /* |
0f830e0b |
2092 | =for apidoc A|UV|utf8n_to_uvchr|U8 *s|STRLEN curlen|STRLEN *retlen|U32 |
2093 | flags |
2094 | |
2095 | Returns the native character value of the first character in the string |
2096 | C<s> |
2097 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
2098 | length, in bytes, of that character. |
2099 | |
2100 | Allows length and flags to be passed to low level routine. |
2101 | |
2102 | =cut |
2103 | */ |
2104 | /* On ASCII machines this is normally a macro but we want |
2105 | a real function in case XS code wants it |
2106 | */ |
0f830e0b |
2107 | UV |
2108 | Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, |
2109 | U32 flags) |
2110 | { |
2111 | const UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags); |
2112 | return UNI_TO_NATIVE(uv); |
2113 | } |
2114 | |
2115 | /* |
d2cc3551 |
2116 | =for apidoc A|char *|pv_uni_display|SV *dsv|U8 *spv|STRLEN len|STRLEN pvlim|UV flags |
2117 | |
2118 | Build to the scalar dsv a displayable version of the string spv, |
2119 | length len, the displayable version being at most pvlim bytes long |
2120 | (if longer, the rest is truncated and "..." will be appended). |
0a2ef054 |
2121 | |
9e55ce06 |
2122 | The flags argument can have UNI_DISPLAY_ISPRINT set to display |
00e86452 |
2123 | isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH |
0a2ef054 |
2124 | to display the \\[nrfta\\] as the backslashed versions (like '\n') |
2125 | (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\). |
2126 | UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both |
2127 | UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on. |
2128 | |
d2cc3551 |
2129 | The pointer to the PV of the dsv is returned. |
2130 | |
2131 | =cut */ |
e6b2e755 |
2132 | char * |
e1ec3a88 |
2133 | Perl_pv_uni_display(pTHX_ SV *dsv, const U8 *spv, STRLEN len, STRLEN pvlim, UV flags) |
e6b2e755 |
2134 | { |
2135 | int truncated = 0; |
e1ec3a88 |
2136 | const char *s, *e; |
e6b2e755 |
2137 | |
2138 | sv_setpvn(dsv, "", 0); |
7fddd944 |
2139 | SvUTF8_off(dsv); |
e1ec3a88 |
2140 | for (s = (const char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) { |
e6b2e755 |
2141 | UV u; |
a49f32c6 |
2142 | /* This serves double duty as a flag and a character to print after |
2143 | a \ when flags & UNI_DISPLAY_BACKSLASH is true. |
2144 | */ |
2145 | char ok = 0; |
c728cb41 |
2146 | |
e6b2e755 |
2147 | if (pvlim && SvCUR(dsv) >= pvlim) { |
2148 | truncated++; |
2149 | break; |
2150 | } |
2151 | u = utf8_to_uvchr((U8*)s, 0); |
c728cb41 |
2152 | if (u < 256) { |
a3b680e6 |
2153 | const unsigned char c = (unsigned char)u & 0xFF; |
0bd48802 |
2154 | if (flags & UNI_DISPLAY_BACKSLASH) { |
a49f32c6 |
2155 | switch (c) { |
c728cb41 |
2156 | case '\n': |
a49f32c6 |
2157 | ok = 'n'; break; |
c728cb41 |
2158 | case '\r': |
a49f32c6 |
2159 | ok = 'r'; break; |
c728cb41 |
2160 | case '\t': |
a49f32c6 |
2161 | ok = 't'; break; |
c728cb41 |
2162 | case '\f': |
a49f32c6 |
2163 | ok = 'f'; break; |
c728cb41 |
2164 | case '\a': |
a49f32c6 |
2165 | ok = 'a'; break; |
c728cb41 |
2166 | case '\\': |
a49f32c6 |
2167 | ok = '\\'; break; |
c728cb41 |
2168 | default: break; |
2169 | } |
a49f32c6 |
2170 | if (ok) { |
88c9ea1e |
2171 | const char string = ok; |
5e7aa789 |
2172 | sv_catpvn(dsv, &string, 1); |
a49f32c6 |
2173 | } |
c728cb41 |
2174 | } |
00e86452 |
2175 | /* isPRINT() is the locale-blind version. */ |
a49f32c6 |
2176 | if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(c)) { |
88c9ea1e |
2177 | const char string = c; |
5e7aa789 |
2178 | sv_catpvn(dsv, &string, 1); |
a49f32c6 |
2179 | ok = 1; |
0a2ef054 |
2180 | } |
c728cb41 |
2181 | } |
2182 | if (!ok) |
9e55ce06 |
2183 | Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u); |
e6b2e755 |
2184 | } |
2185 | if (truncated) |
396482e1 |
2186 | sv_catpvs(dsv, "..."); |
e6b2e755 |
2187 | |
2188 | return SvPVX(dsv); |
2189 | } |
2b9d42f0 |
2190 | |
d2cc3551 |
2191 | /* |
2192 | =for apidoc A|char *|sv_uni_display|SV *dsv|SV *ssv|STRLEN pvlim|UV flags |
2193 | |
2194 | Build to the scalar dsv a displayable version of the scalar sv, |
0a2ef054 |
2195 | the displayable version being at most pvlim bytes long |
d2cc3551 |
2196 | (if longer, the rest is truncated and "..." will be appended). |
0a2ef054 |
2197 | |
2198 | The flags argument is as in pv_uni_display(). |
2199 | |
d2cc3551 |
2200 | The pointer to the PV of the dsv is returned. |
2201 | |
d4c19fe8 |
2202 | =cut |
2203 | */ |
e6b2e755 |
2204 | char * |
2205 | Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags) |
2206 | { |
cfd0369c |
2207 | return Perl_pv_uni_display(aTHX_ dsv, (const U8*)SvPVX_const(ssv), |
2208 | SvCUR(ssv), pvlim, flags); |
701a277b |
2209 | } |
2210 | |
d2cc3551 |
2211 | /* |
d07ddd77 |
2212 | =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 |
d2cc3551 |
2213 | |
2214 | Return true if the strings s1 and s2 differ case-insensitively, false |
2215 | if not (if they are equal case-insensitively). If u1 is true, the |
2216 | string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true, |
d07ddd77 |
2217 | the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2 |
2218 | are false, the respective string is assumed to be in native 8-bit |
2219 | encoding. |
2220 | |
2221 | If the pe1 and pe2 are non-NULL, the scanning pointers will be copied |
2222 | in there (they will point at the beginning of the I<next> character). |
2223 | If the pointers behind pe1 or pe2 are non-NULL, they are the end |
2224 | pointers beyond which scanning will not continue under any |
4cdaeff7 |
2225 | circumstances. If the byte lengths l1 and l2 are non-zero, s1+l1 and |
d07ddd77 |
2226 | s2+l2 will be used as goal end pointers that will also stop the scan, |
2227 | and which qualify towards defining a successful match: all the scans |
2228 | that define an explicit length must reach their goal pointers for |
2229 | a match to succeed). |
d2cc3551 |
2230 | |
2231 | For case-insensitiveness, the "casefolding" of Unicode is used |
2232 | instead of upper/lowercasing both the characters, see |
2233 | http://www.unicode.org/unicode/reports/tr21/ (Case Mappings). |
2234 | |
2235 | =cut */ |
701a277b |
2236 | I32 |
d07ddd77 |
2237 | Perl_ibcmp_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2) |
332ddc25 |
2238 | { |
97aff369 |
2239 | dVAR; |
e1ec3a88 |
2240 | register const U8 *p1 = (const U8*)s1; |
2241 | register const U8 *p2 = (const U8*)s2; |
cbbf8932 |
2242 | register const U8 *f1 = NULL; |
2f73348c |
2243 | register const U8 *f2 = NULL; |
cbbf8932 |
2244 | register U8 *e1 = NULL; |
2245 | register U8 *q1 = NULL; |
2246 | register U8 *e2 = NULL; |
2247 | register U8 *q2 = NULL; |
d07ddd77 |
2248 | STRLEN n1 = 0, n2 = 0; |
89ebb4a3 |
2249 | U8 foldbuf1[UTF8_MAXBYTES_CASE+1]; |
2250 | U8 foldbuf2[UTF8_MAXBYTES_CASE+1]; |
d7f013c8 |
2251 | U8 natbuf[1+1]; |
2252 | STRLEN foldlen1, foldlen2; |
d07ddd77 |
2253 | bool match; |
332ddc25 |
2254 | |
d07ddd77 |
2255 | if (pe1) |
2256 | e1 = *(U8**)pe1; |
a0a388a1 |
2257 | /* assert(e1 || l1); */ |
e1ec3a88 |
2258 | if (e1 == 0 || (l1 && l1 < (UV)(e1 - (const U8*)s1))) |
2259 | f1 = (const U8*)s1 + l1; |
d07ddd77 |
2260 | if (pe2) |
2261 | e2 = *(U8**)pe2; |
a0a388a1 |
2262 | /* assert(e2 || l2); */ |
e1ec3a88 |
2263 | if (e2 == 0 || (l2 && l2 < (UV)(e2 - (const U8*)s2))) |
2264 | f2 = (const U8*)s2 + l2; |
d07ddd77 |
2265 | |
a0a388a1 |
2266 | /* This shouldn't happen. However, putting an assert() there makes some |
2267 | * tests fail. */ |
2268 | /* assert((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0)); */ |
d07ddd77 |
2269 | if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0)) |
2270 | return 1; /* mismatch; possible infinite loop or false positive */ |
2271 | |
a6872d42 |
2272 | if (!u1 || !u2) |
2273 | natbuf[1] = 0; /* Need to terminate the buffer. */ |
2274 | |
d07ddd77 |
2275 | while ((e1 == 0 || p1 < e1) && |
2276 | (f1 == 0 || p1 < f1) && |
2277 | (e2 == 0 || p2 < e2) && |
2278 | (f2 == 0 || p2 < f2)) { |
2279 | if (n1 == 0) { |
d7f013c8 |
2280 | if (u1) |
2281 | to_utf8_fold(p1, foldbuf1, &foldlen1); |
2282 | else { |
809e8e66 |
2283 | uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p1))); |
d7f013c8 |
2284 | to_utf8_fold(natbuf, foldbuf1, &foldlen1); |
2285 | } |
2286 | q1 = foldbuf1; |
d07ddd77 |
2287 | n1 = foldlen1; |
332ddc25 |
2288 | } |
d07ddd77 |
2289 | if (n2 == 0) { |
d7f013c8 |
2290 | if (u2) |
2291 | to_utf8_fold(p2, foldbuf2, &foldlen2); |
2292 | else { |
809e8e66 |
2293 | uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p2))); |
d7f013c8 |
2294 | to_utf8_fold(natbuf, foldbuf2, &foldlen2); |
2295 | } |
2296 | q2 = foldbuf2; |
d07ddd77 |
2297 | n2 = foldlen2; |
332ddc25 |
2298 | } |
d07ddd77 |
2299 | while (n1 && n2) { |
2300 | if ( UTF8SKIP(q1) != UTF8SKIP(q2) || |
2301 | (UTF8SKIP(q1) == 1 && *q1 != *q2) || |
2302 | memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) ) |
d7f013c8 |
2303 | return 1; /* mismatch */ |
d07ddd77 |
2304 | n1 -= UTF8SKIP(q1); |
d7f013c8 |
2305 | q1 += UTF8SKIP(q1); |
d07ddd77 |
2306 | n2 -= UTF8SKIP(q2); |
d7f013c8 |
2307 | q2 += UTF8SKIP(q2); |
701a277b |
2308 | } |
d07ddd77 |
2309 | if (n1 == 0) |
d7f013c8 |
2310 | p1 += u1 ? UTF8SKIP(p1) : 1; |
d07ddd77 |
2311 | if (n2 == 0) |
d7f013c8 |
2312 | p2 += u2 ? UTF8SKIP(p2) : 1; |
2313 | |
d2cc3551 |
2314 | } |
5469e704 |
2315 | |
d07ddd77 |
2316 | /* A match is defined by all the scans that specified |
2317 | * an explicit length reaching their final goals. */ |
2318 | match = (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2); |
5469e704 |
2319 | |
2320 | if (match) { |
d07ddd77 |
2321 | if (pe1) |
2322 | *pe1 = (char*)p1; |
2323 | if (pe2) |
2324 | *pe2 = (char*)p2; |
5469e704 |
2325 | } |
2326 | |
2327 | return match ? 0 : 1; /* 0 match, 1 mismatch */ |
e6b2e755 |
2328 | } |
701a277b |
2329 | |
a49f32c6 |
2330 | /* |
2331 | * Local variables: |
2332 | * c-indentation-style: bsd |
2333 | * c-basic-offset: 4 |
2334 | * indent-tabs-mode: t |
2335 | * End: |
2336 | * |
37442d52 |
2337 | * ex: set ts=8 sts=4 sw=4 noet: |
2338 | */ |