Commit | Line | Data |
a0ed51b3 |
1 | /* utf8.c |
2 | * |
1129b882 |
3 | * Copyright (C) 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008 |
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 | /* |
4ac71550 |
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 | * [p.603 of _The Lord of the Rings_, IV/I: "The Taming of Sméagol"] |
a0ed51b3 |
17 | * |
18 | * 'Well do I understand your speech,' he answered in the same language; |
19 | * 'yet few strangers do so. Why then do you not speak in the Common Tongue, |
4ac71550 |
20 | * as is the custom in the West, if you wish to be answered?' |
21 | * --Gandalf, addressing Théoden's door wardens |
22 | * |
23 | * [p.508 of _The Lord of the Rings_, III/vi: "The King of the Golden Hall"] |
a0ed51b3 |
24 | * |
25 | * ...the travellers perceived that the floor was paved with stones of many |
26 | * hues; branching runes and strange devices intertwined beneath their feet. |
4ac71550 |
27 | * |
28 | * [p.512 of _The Lord of the Rings_, III/vi: "The King of the Golden Hall"] |
a0ed51b3 |
29 | */ |
30 | |
31 | #include "EXTERN.h" |
864dbfa3 |
32 | #define PERL_IN_UTF8_C |
a0ed51b3 |
33 | #include "perl.h" |
34 | |
a0c21aa1 |
35 | #ifndef EBCDIC |
36 | /* Separate prototypes needed because in ASCII systems these |
37 | * usually macros but they still are compiled as code, too. */ |
38 | PERL_CALLCONV UV Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags); |
39 | PERL_CALLCONV U8* Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv); |
40 | #endif |
41 | |
27da23d5 |
42 | static const char unees[] = |
43 | "Malformed UTF-8 character (unexpected end of string)"; |
901b21bf |
44 | |
ccfc67b7 |
45 | /* |
46 | =head1 Unicode Support |
a0ed51b3 |
47 | |
166f8a29 |
48 | This file contains various utility functions for manipulating UTF8-encoded |
49 | strings. For the uninitiated, this is a method of representing arbitrary |
61296642 |
50 | Unicode characters as a variable number of bytes, in such a way that |
56da48f7 |
51 | characters in the ASCII range are unmodified, and a zero byte never appears |
52 | within non-zero characters. |
166f8a29 |
53 | |
eaf7a4d2 |
54 | =cut |
55 | */ |
56 | |
57 | /* |
58 | =for apidoc is_ascii_string |
59 | |
60 | Returns true if first C<len> bytes of the given string are ASCII (i.e. none |
61 | of them even raise the question of UTF-8-ness). |
62 | |
63 | See also is_utf8_string(), is_utf8_string_loclen(), and is_utf8_string_loc(). |
64 | |
65 | =cut |
66 | */ |
67 | |
68 | bool |
668b6d8d |
69 | Perl_is_ascii_string(const U8 *s, STRLEN len) |
eaf7a4d2 |
70 | { |
71 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
72 | const U8* x = s; |
73 | |
74 | PERL_ARGS_ASSERT_IS_ASCII_STRING; |
eaf7a4d2 |
75 | |
76 | for (; x < send; ++x) { |
77 | if (!UTF8_IS_INVARIANT(*x)) |
78 | break; |
79 | } |
80 | |
81 | return x == send; |
82 | } |
83 | |
84 | /* |
87cea99e |
85 | =for apidoc uvuni_to_utf8_flags |
eebe1485 |
86 | |
1e54db1a |
87 | Adds the UTF-8 representation of the Unicode codepoint C<uv> to the end |
89ebb4a3 |
88 | of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free |
eebe1485 |
89 | bytes available. The return value is the pointer to the byte after the |
9041c2e3 |
90 | end of the new character. In other words, |
eebe1485 |
91 | |
b851fbc1 |
92 | d = uvuni_to_utf8_flags(d, uv, flags); |
93 | |
94 | or, in most cases, |
95 | |
9041c2e3 |
96 | d = uvuni_to_utf8(d, uv); |
eebe1485 |
97 | |
b851fbc1 |
98 | (which is equivalent to) |
99 | |
100 | d = uvuni_to_utf8_flags(d, uv, 0); |
101 | |
eebe1485 |
102 | is the recommended Unicode-aware way of saying |
103 | |
104 | *(d++) = uv; |
105 | |
106 | =cut |
107 | */ |
108 | |
dfe13c55 |
109 | U8 * |
b851fbc1 |
110 | Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) |
a0ed51b3 |
111 | { |
7918f24d |
112 | PERL_ARGS_ASSERT_UVUNI_TO_UTF8_FLAGS; |
113 | |
62961d2e |
114 | if (ckWARN(WARN_UTF8)) { |
b851fbc1 |
115 | if (UNICODE_IS_SURROGATE(uv) && |
116 | !(flags & UNICODE_ALLOW_SURROGATE)) |
9014280d |
117 | Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv); |
b851fbc1 |
118 | else if ( |
119 | ((uv >= 0xFDD0 && uv <= 0xFDEF && |
120 | !(flags & UNICODE_ALLOW_FDD0)) |
121 | || |
c867b360 |
122 | ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */ |
b851fbc1 |
123 | !(flags & UNICODE_ALLOW_FFFF))) && |
124 | /* UNICODE_ALLOW_SUPER includes |
2a20b9da |
125 | * FFFEs and FFFFs beyond 0x10FFFF. */ |
b851fbc1 |
126 | ((uv <= PERL_UNICODE_MAX) || |
127 | !(flags & UNICODE_ALLOW_SUPER)) |
128 | ) |
9014280d |
129 | Perl_warner(aTHX_ packWARN(WARN_UTF8), |
6f6ac1de |
130 | "Unicode non-character 0x%04"UVxf" is illegal for interchange", uv); |
507b9800 |
131 | } |
c4d5f83a |
132 | if (UNI_IS_INVARIANT(uv)) { |
eb160463 |
133 | *d++ = (U8)UTF_TO_NATIVE(uv); |
a0ed51b3 |
134 | return d; |
135 | } |
2d331972 |
136 | #if defined(EBCDIC) |
1d72bdf6 |
137 | else { |
138 | STRLEN len = UNISKIP(uv); |
139 | U8 *p = d+len-1; |
140 | while (p > d) { |
eb160463 |
141 | *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK); |
1d72bdf6 |
142 | uv >>= UTF_ACCUMULATION_SHIFT; |
143 | } |
eb160463 |
144 | *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len)); |
1d72bdf6 |
145 | return d+len; |
146 | } |
147 | #else /* Non loop style */ |
a0ed51b3 |
148 | if (uv < 0x800) { |
eb160463 |
149 | *d++ = (U8)(( uv >> 6) | 0xc0); |
150 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
151 | return d; |
152 | } |
153 | if (uv < 0x10000) { |
eb160463 |
154 | *d++ = (U8)(( uv >> 12) | 0xe0); |
155 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
156 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
157 | return d; |
158 | } |
159 | if (uv < 0x200000) { |
eb160463 |
160 | *d++ = (U8)(( uv >> 18) | 0xf0); |
161 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
162 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
163 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
164 | return d; |
165 | } |
166 | if (uv < 0x4000000) { |
eb160463 |
167 | *d++ = (U8)(( uv >> 24) | 0xf8); |
168 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
169 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
170 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
171 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
172 | return d; |
173 | } |
174 | if (uv < 0x80000000) { |
eb160463 |
175 | *d++ = (U8)(( uv >> 30) | 0xfc); |
176 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); |
177 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
178 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
179 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
180 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
181 | return d; |
182 | } |
6b8eaf93 |
183 | #ifdef HAS_QUAD |
d7578b48 |
184 | if (uv < UTF8_QUAD_MAX) |
a0ed51b3 |
185 | #endif |
186 | { |
eb160463 |
187 | *d++ = 0xfe; /* Can't match U+FEFF! */ |
188 | *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); |
189 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); |
190 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
191 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
192 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
193 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
194 | return d; |
195 | } |
6b8eaf93 |
196 | #ifdef HAS_QUAD |
a0ed51b3 |
197 | { |
eb160463 |
198 | *d++ = 0xff; /* Can't match U+FFFE! */ |
199 | *d++ = 0x80; /* 6 Reserved bits */ |
200 | *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */ |
201 | *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80); |
202 | *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80); |
203 | *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80); |
204 | *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80); |
205 | *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); |
206 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); |
207 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
208 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
209 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
210 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
211 | return d; |
212 | } |
213 | #endif |
1d72bdf6 |
214 | #endif /* Loop style */ |
a0ed51b3 |
215 | } |
9041c2e3 |
216 | |
646ca15d |
217 | /* |
218 | |
219 | Tests if some arbitrary number of bytes begins in a valid UTF-8 |
220 | character. Note that an INVARIANT (i.e. ASCII) character is a valid |
221 | UTF-8 character. The actual number of bytes in the UTF-8 character |
222 | will be returned if it is valid, otherwise 0. |
223 | |
224 | This is the "slow" version as opposed to the "fast" version which is |
225 | the "unrolled" IS_UTF8_CHAR(). E.g. for t/uni/class.t the speed |
226 | difference is a factor of 2 to 3. For lengths (UTF8SKIP(s)) of four |
227 | or less you should use the IS_UTF8_CHAR(), for lengths of five or more |
228 | you should use the _slow(). In practice this means that the _slow() |
229 | will be used very rarely, since the maximum Unicode code point (as of |
230 | Unicode 4.1) is U+10FFFF, which encodes in UTF-8 to four bytes. Only |
231 | the "Perl extended UTF-8" (the infamous 'v-strings') will encode into |
232 | five bytes or more. |
233 | |
234 | =cut */ |
c053b435 |
235 | STATIC STRLEN |
5f66b61c |
236 | S_is_utf8_char_slow(const U8 *s, const STRLEN len) |
646ca15d |
237 | { |
238 | U8 u = *s; |
239 | STRLEN slen; |
240 | UV uv, ouv; |
241 | |
7918f24d |
242 | PERL_ARGS_ASSERT_IS_UTF8_CHAR_SLOW; |
243 | |
646ca15d |
244 | if (UTF8_IS_INVARIANT(u)) |
245 | return 1; |
246 | |
247 | if (!UTF8_IS_START(u)) |
248 | return 0; |
249 | |
250 | if (len < 2 || !UTF8_IS_CONTINUATION(s[1])) |
251 | return 0; |
252 | |
253 | slen = len - 1; |
254 | s++; |
77263263 |
255 | #ifdef EBCDIC |
256 | u = NATIVE_TO_UTF(u); |
257 | #endif |
646ca15d |
258 | u &= UTF_START_MASK(len); |
259 | uv = u; |
260 | ouv = uv; |
261 | while (slen--) { |
262 | if (!UTF8_IS_CONTINUATION(*s)) |
263 | return 0; |
264 | uv = UTF8_ACCUMULATE(uv, *s); |
265 | if (uv < ouv) |
266 | return 0; |
267 | ouv = uv; |
268 | s++; |
269 | } |
270 | |
271 | if ((STRLEN)UNISKIP(uv) < len) |
272 | return 0; |
273 | |
274 | return len; |
275 | } |
9041c2e3 |
276 | |
277 | /* |
87cea99e |
278 | =for apidoc is_utf8_char |
eebe1485 |
279 | |
5da9da9e |
280 | Tests if some arbitrary number of bytes begins in a valid UTF-8 |
2bbc8d55 |
281 | character. Note that an INVARIANT (i.e. ASCII on non-EBCDIC machines) |
282 | character is a valid UTF-8 character. The actual number of bytes in the UTF-8 |
283 | character will be returned if it is valid, otherwise 0. |
9041c2e3 |
284 | |
82686b01 |
285 | =cut */ |
067a85ef |
286 | STRLEN |
668b6d8d |
287 | Perl_is_utf8_char(const U8 *s) |
386d01d6 |
288 | { |
44f8325f |
289 | const STRLEN len = UTF8SKIP(s); |
7918f24d |
290 | |
291 | PERL_ARGS_ASSERT_IS_UTF8_CHAR; |
3b0fc154 |
292 | #ifdef IS_UTF8_CHAR |
768c67ee |
293 | if (IS_UTF8_CHAR_FAST(len)) |
3b0fc154 |
294 | return IS_UTF8_CHAR(s, len) ? len : 0; |
295 | #endif /* #ifdef IS_UTF8_CHAR */ |
2c0c5f92 |
296 | return is_utf8_char_slow(s, len); |
386d01d6 |
297 | } |
298 | |
eaf7a4d2 |
299 | |
6662521e |
300 | /* |
87cea99e |
301 | =for apidoc is_utf8_string |
6662521e |
302 | |
c9ada85f |
303 | Returns true if first C<len> bytes of the given string form a valid |
1e54db1a |
304 | UTF-8 string, false otherwise. Note that 'a valid UTF-8 string' does |
305 | not mean 'a string that contains code points above 0x7F encoded in UTF-8' |
306 | because a valid ASCII string is a valid UTF-8 string. |
6662521e |
307 | |
eaf7a4d2 |
308 | See also is_ascii_string(), is_utf8_string_loclen(), and is_utf8_string_loc(). |
768c67ee |
309 | |
6662521e |
310 | =cut |
311 | */ |
312 | |
8e84507e |
313 | bool |
668b6d8d |
314 | Perl_is_utf8_string(const U8 *s, STRLEN len) |
6662521e |
315 | { |
35da51f7 |
316 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
7fc63493 |
317 | const U8* x = s; |
067a85ef |
318 | |
7918f24d |
319 | PERL_ARGS_ASSERT_IS_UTF8_STRING; |
1aa99e6b |
320 | |
6662521e |
321 | while (x < send) { |
a3b680e6 |
322 | STRLEN c; |
1acdb0da |
323 | /* Inline the easy bits of is_utf8_char() here for speed... */ |
324 | if (UTF8_IS_INVARIANT(*x)) |
325 | c = 1; |
326 | else if (!UTF8_IS_START(*x)) |
768c67ee |
327 | goto out; |
1acdb0da |
328 | else { |
329 | /* ... and call is_utf8_char() only if really needed. */ |
646ca15d |
330 | #ifdef IS_UTF8_CHAR |
331 | c = UTF8SKIP(x); |
768c67ee |
332 | if (IS_UTF8_CHAR_FAST(c)) { |
333 | if (!IS_UTF8_CHAR(x, c)) |
3c614e38 |
334 | c = 0; |
335 | } |
336 | else |
337 | c = is_utf8_char_slow(x, c); |
646ca15d |
338 | #else |
339 | c = is_utf8_char(x); |
340 | #endif /* #ifdef IS_UTF8_CHAR */ |
1acdb0da |
341 | if (!c) |
768c67ee |
342 | goto out; |
1acdb0da |
343 | } |
6662521e |
344 | x += c; |
6662521e |
345 | } |
768c67ee |
346 | |
347 | out: |
60006e79 |
348 | if (x != send) |
349 | return FALSE; |
067a85ef |
350 | |
351 | return TRUE; |
6662521e |
352 | } |
353 | |
67e989fb |
354 | /* |
814fafa7 |
355 | Implemented as a macro in utf8.h |
356 | |
87cea99e |
357 | =for apidoc is_utf8_string_loc |
814fafa7 |
358 | |
359 | Like is_utf8_string() but stores the location of the failure (in the |
360 | case of "utf8ness failure") or the location s+len (in the case of |
361 | "utf8ness success") in the C<ep>. |
362 | |
363 | See also is_utf8_string_loclen() and is_utf8_string(). |
364 | |
87cea99e |
365 | =for apidoc is_utf8_string_loclen |
81cd54e3 |
366 | |
e3e4599f |
367 | Like is_utf8_string() but stores the location of the failure (in the |
768c67ee |
368 | case of "utf8ness failure") or the location s+len (in the case of |
369 | "utf8ness success") in the C<ep>, and the number of UTF-8 |
370 | encoded characters in the C<el>. |
371 | |
372 | See also is_utf8_string_loc() and is_utf8_string(). |
81cd54e3 |
373 | |
374 | =cut |
375 | */ |
376 | |
377 | bool |
668b6d8d |
378 | Perl_is_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el) |
81cd54e3 |
379 | { |
35da51f7 |
380 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
7fc63493 |
381 | const U8* x = s; |
81cd54e3 |
382 | STRLEN c; |
3ebfea28 |
383 | STRLEN outlen = 0; |
7918f24d |
384 | |
385 | PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN; |
81cd54e3 |
386 | |
81cd54e3 |
387 | while (x < send) { |
388 | /* Inline the easy bits of is_utf8_char() here for speed... */ |
389 | if (UTF8_IS_INVARIANT(*x)) |
768c67ee |
390 | c = 1; |
391 | else if (!UTF8_IS_START(*x)) |
392 | goto out; |
81cd54e3 |
393 | else { |
768c67ee |
394 | /* ... and call is_utf8_char() only if really needed. */ |
395 | #ifdef IS_UTF8_CHAR |
396 | c = UTF8SKIP(x); |
397 | if (IS_UTF8_CHAR_FAST(c)) { |
398 | if (!IS_UTF8_CHAR(x, c)) |
399 | c = 0; |
400 | } else |
401 | c = is_utf8_char_slow(x, c); |
402 | #else |
403 | c = is_utf8_char(x); |
404 | #endif /* #ifdef IS_UTF8_CHAR */ |
405 | if (!c) |
406 | goto out; |
81cd54e3 |
407 | } |
768c67ee |
408 | x += c; |
3ebfea28 |
409 | outlen++; |
81cd54e3 |
410 | } |
768c67ee |
411 | |
412 | out: |
3ebfea28 |
413 | if (el) |
414 | *el = outlen; |
415 | |
768c67ee |
416 | if (ep) |
417 | *ep = x; |
3ebfea28 |
418 | return (x == send); |
81cd54e3 |
419 | } |
420 | |
421 | /* |
768c67ee |
422 | |
87cea99e |
423 | =for apidoc utf8n_to_uvuni |
67e989fb |
424 | |
9041c2e3 |
425 | Bottom level UTF-8 decode routine. |
38a44b82 |
426 | Returns the Unicode code point value of the first character in the string C<s> |
1e54db1a |
427 | which is assumed to be in UTF-8 encoding and no longer than C<curlen>; |
7df053ec |
428 | C<retlen> will be set to the length, in bytes, of that character. |
67e989fb |
429 | |
1e54db1a |
430 | If C<s> does not point to a well-formed UTF-8 character, the behaviour |
dcad2880 |
431 | is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY, |
432 | it is assumed that the caller will raise a warning, and this function |
28d3d195 |
433 | will silently just set C<retlen> to C<-1> and return zero. If the |
434 | C<flags> does not contain UTF8_CHECK_ONLY, warnings about |
435 | malformations will be given, C<retlen> will be set to the expected |
436 | length of the UTF-8 character in bytes, and zero will be returned. |
437 | |
438 | The C<flags> can also contain various flags to allow deviations from |
439 | the strict UTF-8 encoding (see F<utf8.h>). |
67e989fb |
440 | |
9041c2e3 |
441 | Most code should use utf8_to_uvchr() rather than call this directly. |
442 | |
37607a96 |
443 | =cut |
444 | */ |
67e989fb |
445 | |
a0ed51b3 |
446 | UV |
7fc63493 |
447 | Perl_utf8n_to_uvuni(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) |
a0ed51b3 |
448 | { |
97aff369 |
449 | dVAR; |
d4c19fe8 |
450 | const U8 * const s0 = s; |
9c5ffd7c |
451 | UV uv = *s, ouv = 0; |
ba210ebe |
452 | STRLEN len = 1; |
7fc63493 |
453 | const bool dowarn = ckWARN_d(WARN_UTF8); |
454 | const UV startbyte = *s; |
ba210ebe |
455 | STRLEN expectlen = 0; |
a0dbb045 |
456 | U32 warning = 0; |
457 | |
7918f24d |
458 | PERL_ARGS_ASSERT_UTF8N_TO_UVUNI; |
459 | |
a0dbb045 |
460 | /* This list is a superset of the UTF8_ALLOW_XXX. */ |
461 | |
462 | #define UTF8_WARN_EMPTY 1 |
463 | #define UTF8_WARN_CONTINUATION 2 |
464 | #define UTF8_WARN_NON_CONTINUATION 3 |
465 | #define UTF8_WARN_FE_FF 4 |
466 | #define UTF8_WARN_SHORT 5 |
467 | #define UTF8_WARN_OVERFLOW 6 |
468 | #define UTF8_WARN_SURROGATE 7 |
c867b360 |
469 | #define UTF8_WARN_LONG 8 |
470 | #define UTF8_WARN_FFFF 9 /* Also FFFE. */ |
a0dbb045 |
471 | |
472 | if (curlen == 0 && |
473 | !(flags & UTF8_ALLOW_EMPTY)) { |
474 | warning = UTF8_WARN_EMPTY; |
0c443dc2 |
475 | goto malformed; |
476 | } |
477 | |
1d72bdf6 |
478 | if (UTF8_IS_INVARIANT(uv)) { |
a0ed51b3 |
479 | if (retlen) |
480 | *retlen = 1; |
c4d5f83a |
481 | return (UV) (NATIVE_TO_UTF(*s)); |
a0ed51b3 |
482 | } |
67e989fb |
483 | |
421a8bf2 |
484 | if (UTF8_IS_CONTINUATION(uv) && |
fcc8fcf6 |
485 | !(flags & UTF8_ALLOW_CONTINUATION)) { |
a0dbb045 |
486 | warning = UTF8_WARN_CONTINUATION; |
ba210ebe |
487 | goto malformed; |
488 | } |
489 | |
421a8bf2 |
490 | if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) && |
fcc8fcf6 |
491 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { |
a0dbb045 |
492 | warning = UTF8_WARN_NON_CONTINUATION; |
ba210ebe |
493 | goto malformed; |
494 | } |
9041c2e3 |
495 | |
1d72bdf6 |
496 | #ifdef EBCDIC |
75383841 |
497 | uv = NATIVE_TO_UTF(uv); |
1d72bdf6 |
498 | #else |
fcc8fcf6 |
499 | if ((uv == 0xfe || uv == 0xff) && |
500 | !(flags & UTF8_ALLOW_FE_FF)) { |
a0dbb045 |
501 | warning = UTF8_WARN_FE_FF; |
ba210ebe |
502 | goto malformed; |
a0ed51b3 |
503 | } |
1d72bdf6 |
504 | #endif |
505 | |
ba210ebe |
506 | if (!(uv & 0x20)) { len = 2; uv &= 0x1f; } |
507 | else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; } |
508 | else if (!(uv & 0x08)) { len = 4; uv &= 0x07; } |
509 | else if (!(uv & 0x04)) { len = 5; uv &= 0x03; } |
1d72bdf6 |
510 | #ifdef EBCDIC |
511 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } |
512 | else { len = 7; uv &= 0x01; } |
513 | #else |
ba210ebe |
514 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } |
515 | else if (!(uv & 0x01)) { len = 7; uv = 0; } |
1d72bdf6 |
516 | else { len = 13; uv = 0; } /* whoa! */ |
517 | #endif |
518 | |
a0ed51b3 |
519 | if (retlen) |
520 | *retlen = len; |
9041c2e3 |
521 | |
ba210ebe |
522 | expectlen = len; |
523 | |
fcc8fcf6 |
524 | if ((curlen < expectlen) && |
525 | !(flags & UTF8_ALLOW_SHORT)) { |
a0dbb045 |
526 | warning = UTF8_WARN_SHORT; |
ba210ebe |
527 | goto malformed; |
528 | } |
529 | |
530 | len--; |
a0ed51b3 |
531 | s++; |
ba210ebe |
532 | ouv = uv; |
533 | |
a0ed51b3 |
534 | while (len--) { |
421a8bf2 |
535 | if (!UTF8_IS_CONTINUATION(*s) && |
536 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { |
a0dbb045 |
537 | s--; |
538 | warning = UTF8_WARN_NON_CONTINUATION; |
ba210ebe |
539 | goto malformed; |
a0ed51b3 |
540 | } |
541 | else |
8850bf83 |
542 | uv = UTF8_ACCUMULATE(uv, *s); |
a0dbb045 |
543 | if (!(uv > ouv)) { |
544 | /* These cannot be allowed. */ |
545 | if (uv == ouv) { |
75dbc644 |
546 | if (expectlen != 13 && !(flags & UTF8_ALLOW_LONG)) { |
a0dbb045 |
547 | warning = UTF8_WARN_LONG; |
548 | goto malformed; |
549 | } |
550 | } |
551 | else { /* uv < ouv */ |
552 | /* This cannot be allowed. */ |
553 | warning = UTF8_WARN_OVERFLOW; |
554 | goto malformed; |
555 | } |
ba210ebe |
556 | } |
557 | s++; |
558 | ouv = uv; |
559 | } |
560 | |
421a8bf2 |
561 | if (UNICODE_IS_SURROGATE(uv) && |
fcc8fcf6 |
562 | !(flags & UTF8_ALLOW_SURROGATE)) { |
a0dbb045 |
563 | warning = UTF8_WARN_SURROGATE; |
ba210ebe |
564 | goto malformed; |
eb160463 |
565 | } else if ((expectlen > (STRLEN)UNISKIP(uv)) && |
fcc8fcf6 |
566 | !(flags & UTF8_ALLOW_LONG)) { |
a0dbb045 |
567 | warning = UTF8_WARN_LONG; |
ba210ebe |
568 | goto malformed; |
421a8bf2 |
569 | } else if (UNICODE_IS_ILLEGAL(uv) && |
a9917092 |
570 | !(flags & UTF8_ALLOW_FFFF)) { |
a0dbb045 |
571 | warning = UTF8_WARN_FFFF; |
a9917092 |
572 | goto malformed; |
a0ed51b3 |
573 | } |
ba210ebe |
574 | |
a0ed51b3 |
575 | return uv; |
ba210ebe |
576 | |
577 | malformed: |
578 | |
fcc8fcf6 |
579 | if (flags & UTF8_CHECK_ONLY) { |
ba210ebe |
580 | if (retlen) |
10edeb5d |
581 | *retlen = ((STRLEN) -1); |
ba210ebe |
582 | return 0; |
583 | } |
584 | |
a0dbb045 |
585 | if (dowarn) { |
84bafc02 |
586 | SV* const sv = newSVpvs_flags("Malformed UTF-8 character ", SVs_TEMP); |
a0dbb045 |
587 | |
588 | switch (warning) { |
589 | case 0: /* Intentionally empty. */ break; |
590 | case UTF8_WARN_EMPTY: |
396482e1 |
591 | sv_catpvs(sv, "(empty string)"); |
a0dbb045 |
592 | break; |
593 | case UTF8_WARN_CONTINUATION: |
097fb8e2 |
594 | Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv); |
a0dbb045 |
595 | break; |
596 | case UTF8_WARN_NON_CONTINUATION: |
097fb8e2 |
597 | if (s == s0) |
598 | Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")", |
599 | (UV)s[1], startbyte); |
551405c4 |
600 | else { |
601 | const int len = (int)(s-s0); |
097fb8e2 |
602 | 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 |
603 | (UV)s[1], len, len > 1 ? "s" : "", startbyte, (int)expectlen); |
604 | } |
605 | |
a0dbb045 |
606 | break; |
607 | case UTF8_WARN_FE_FF: |
608 | Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv); |
609 | break; |
610 | case UTF8_WARN_SHORT: |
097fb8e2 |
611 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", |
5d7488b2 |
612 | (int)curlen, curlen == 1 ? "" : "s", (int)expectlen, startbyte); |
b31f83c2 |
613 | expectlen = curlen; /* distance for caller to skip */ |
a0dbb045 |
614 | break; |
615 | case UTF8_WARN_OVERFLOW: |
097fb8e2 |
616 | Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")", |
617 | ouv, *s, startbyte); |
a0dbb045 |
618 | break; |
619 | case UTF8_WARN_SURROGATE: |
620 | Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv); |
621 | break; |
a0dbb045 |
622 | case UTF8_WARN_LONG: |
097fb8e2 |
623 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", |
5d7488b2 |
624 | (int)expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte); |
a0dbb045 |
625 | break; |
626 | case UTF8_WARN_FFFF: |
627 | Perl_sv_catpvf(aTHX_ sv, "(character 0x%04"UVxf")", uv); |
628 | break; |
629 | default: |
396482e1 |
630 | sv_catpvs(sv, "(unknown reason)"); |
a0dbb045 |
631 | break; |
632 | } |
633 | |
634 | if (warning) { |
44f8325f |
635 | const char * const s = SvPVX_const(sv); |
a0dbb045 |
636 | |
637 | if (PL_op) |
9014280d |
638 | Perl_warner(aTHX_ packWARN(WARN_UTF8), |
53e06cf0 |
639 | "%s in %s", s, OP_DESC(PL_op)); |
a0dbb045 |
640 | else |
9014280d |
641 | Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s); |
a0dbb045 |
642 | } |
643 | } |
644 | |
ba210ebe |
645 | if (retlen) |
28d3d195 |
646 | *retlen = expectlen ? expectlen : len; |
ba210ebe |
647 | |
28d3d195 |
648 | return 0; |
a0ed51b3 |
649 | } |
650 | |
8e84507e |
651 | /* |
87cea99e |
652 | =for apidoc utf8_to_uvchr |
9041c2e3 |
653 | |
654 | Returns the native character value of the first character in the string C<s> |
1e54db1a |
655 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
9041c2e3 |
656 | length, in bytes, of that character. |
657 | |
1e54db1a |
658 | If C<s> does not point to a well-formed UTF-8 character, zero is |
9041c2e3 |
659 | returned and retlen is set, if possible, to -1. |
660 | |
661 | =cut |
662 | */ |
663 | |
664 | UV |
7fc63493 |
665 | Perl_utf8_to_uvchr(pTHX_ const U8 *s, STRLEN *retlen) |
9041c2e3 |
666 | { |
7918f24d |
667 | PERL_ARGS_ASSERT_UTF8_TO_UVCHR; |
668 | |
1754c1a1 |
669 | return utf8n_to_uvchr(s, UTF8_MAXBYTES, retlen, |
670 | ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); |
9041c2e3 |
671 | } |
672 | |
673 | /* |
87cea99e |
674 | =for apidoc utf8_to_uvuni |
9041c2e3 |
675 | |
676 | Returns the Unicode code point of the first character in the string C<s> |
1e54db1a |
677 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
9041c2e3 |
678 | length, in bytes, of that character. |
679 | |
2bbc8d55 |
680 | This function should only be used when the returned UV is considered |
9041c2e3 |
681 | an index into the Unicode semantic tables (e.g. swashes). |
682 | |
1e54db1a |
683 | If C<s> does not point to a well-formed UTF-8 character, zero is |
ba210ebe |
684 | returned and retlen is set, if possible, to -1. |
8e84507e |
685 | |
686 | =cut |
687 | */ |
688 | |
689 | UV |
7fc63493 |
690 | Perl_utf8_to_uvuni(pTHX_ const U8 *s, STRLEN *retlen) |
8e84507e |
691 | { |
7918f24d |
692 | PERL_ARGS_ASSERT_UTF8_TO_UVUNI; |
693 | |
9041c2e3 |
694 | /* Call the low level routine asking for checks */ |
89ebb4a3 |
695 | return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXBYTES, retlen, |
872c91ae |
696 | ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); |
8e84507e |
697 | } |
698 | |
b76347f2 |
699 | /* |
87cea99e |
700 | =for apidoc utf8_length |
b76347f2 |
701 | |
702 | Return the length of the UTF-8 char encoded string C<s> in characters. |
02eb7b47 |
703 | Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end |
704 | up past C<e>, croaks. |
b76347f2 |
705 | |
706 | =cut |
707 | */ |
708 | |
709 | STRLEN |
35a4481c |
710 | Perl_utf8_length(pTHX_ const U8 *s, const U8 *e) |
b76347f2 |
711 | { |
97aff369 |
712 | dVAR; |
b76347f2 |
713 | STRLEN len = 0; |
714 | |
7918f24d |
715 | PERL_ARGS_ASSERT_UTF8_LENGTH; |
716 | |
8850bf83 |
717 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g. |
718 | * the bitops (especially ~) can create illegal UTF-8. |
719 | * In other words: in Perl UTF-8 is not just for Unicode. */ |
720 | |
a3b680e6 |
721 | if (e < s) |
722 | goto warn_and_return; |
b76347f2 |
723 | while (s < e) { |
8e91ec7f |
724 | if (!UTF8_IS_INVARIANT(*s)) |
725 | s += UTF8SKIP(s); |
726 | else |
727 | s++; |
728 | len++; |
729 | } |
730 | |
731 | if (e != s) { |
732 | len--; |
733 | warn_and_return: |
9b387841 |
734 | if (PL_op) |
735 | Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), |
736 | "%s in %s", unees, OP_DESC(PL_op)); |
737 | else |
738 | Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), unees); |
b76347f2 |
739 | } |
740 | |
741 | return len; |
742 | } |
743 | |
b06226ff |
744 | /* |
87cea99e |
745 | =for apidoc utf8_distance |
b06226ff |
746 | |
1e54db1a |
747 | Returns the number of UTF-8 characters between the UTF-8 pointers C<a> |
b06226ff |
748 | and C<b>. |
749 | |
750 | WARNING: use only if you *know* that the pointers point inside the |
751 | same UTF-8 buffer. |
752 | |
37607a96 |
753 | =cut |
754 | */ |
a0ed51b3 |
755 | |
02eb7b47 |
756 | IV |
35a4481c |
757 | Perl_utf8_distance(pTHX_ const U8 *a, const U8 *b) |
a0ed51b3 |
758 | { |
7918f24d |
759 | PERL_ARGS_ASSERT_UTF8_DISTANCE; |
760 | |
bf1665bc |
761 | return (a < b) ? -1 * (IV) utf8_length(a, b) : (IV) utf8_length(b, a); |
a0ed51b3 |
762 | } |
763 | |
b06226ff |
764 | /* |
87cea99e |
765 | =for apidoc utf8_hop |
b06226ff |
766 | |
8850bf83 |
767 | Return the UTF-8 pointer C<s> displaced by C<off> characters, either |
768 | forward or backward. |
b06226ff |
769 | |
770 | WARNING: do not use the following unless you *know* C<off> is within |
8850bf83 |
771 | the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned |
772 | on the first byte of character or just after the last byte of a character. |
b06226ff |
773 | |
37607a96 |
774 | =cut |
775 | */ |
a0ed51b3 |
776 | |
777 | U8 * |
4373e329 |
778 | Perl_utf8_hop(pTHX_ const U8 *s, I32 off) |
a0ed51b3 |
779 | { |
7918f24d |
780 | PERL_ARGS_ASSERT_UTF8_HOP; |
781 | |
96a5add6 |
782 | PERL_UNUSED_CONTEXT; |
8850bf83 |
783 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g |
784 | * the bitops (especially ~) can create illegal UTF-8. |
785 | * In other words: in Perl UTF-8 is not just for Unicode. */ |
786 | |
a0ed51b3 |
787 | if (off >= 0) { |
788 | while (off--) |
789 | s += UTF8SKIP(s); |
790 | } |
791 | else { |
792 | while (off++) { |
793 | s--; |
8850bf83 |
794 | while (UTF8_IS_CONTINUATION(*s)) |
795 | s--; |
a0ed51b3 |
796 | } |
797 | } |
4373e329 |
798 | return (U8 *)s; |
a0ed51b3 |
799 | } |
800 | |
6940069f |
801 | /* |
87cea99e |
802 | =for apidoc utf8_to_bytes |
6940069f |
803 | |
2bbc8d55 |
804 | Converts a string C<s> of length C<len> from UTF-8 into native byte encoding. |
246fae53 |
805 | Unlike C<bytes_to_utf8>, this over-writes the original string, and |
806 | updates len to contain the new length. |
67e989fb |
807 | Returns zero on failure, setting C<len> to -1. |
6940069f |
808 | |
95be277c |
809 | If you need a copy of the string, see C<bytes_from_utf8>. |
810 | |
6940069f |
811 | =cut |
812 | */ |
813 | |
814 | U8 * |
37607a96 |
815 | Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len) |
6940069f |
816 | { |
d4c19fe8 |
817 | U8 * const save = s; |
818 | U8 * const send = s + *len; |
6940069f |
819 | U8 *d; |
246fae53 |
820 | |
7918f24d |
821 | PERL_ARGS_ASSERT_UTF8_TO_BYTES; |
822 | |
1e54db1a |
823 | /* ensure valid UTF-8 and chars < 256 before updating string */ |
d4c19fe8 |
824 | while (s < send) { |
dcad2880 |
825 | U8 c = *s++; |
826 | |
1d72bdf6 |
827 | if (!UTF8_IS_INVARIANT(c) && |
828 | (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send) |
829 | || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) { |
10edeb5d |
830 | *len = ((STRLEN) -1); |
dcad2880 |
831 | return 0; |
832 | } |
246fae53 |
833 | } |
dcad2880 |
834 | |
835 | d = s = save; |
6940069f |
836 | while (s < send) { |
ed646e6e |
837 | STRLEN ulen; |
9041c2e3 |
838 | *d++ = (U8)utf8_to_uvchr(s, &ulen); |
ed646e6e |
839 | s += ulen; |
6940069f |
840 | } |
841 | *d = '\0'; |
246fae53 |
842 | *len = d - save; |
6940069f |
843 | return save; |
844 | } |
845 | |
846 | /* |
87cea99e |
847 | =for apidoc bytes_from_utf8 |
f9a63242 |
848 | |
2bbc8d55 |
849 | Converts a string C<s> of length C<len> from UTF-8 into native byte encoding. |
35a4481c |
850 | Unlike C<utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to |
ef9edfd0 |
851 | the newly-created string, and updates C<len> to contain the new |
852 | length. Returns the original string if no conversion occurs, C<len> |
853 | is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to |
2bbc8d55 |
854 | 0 if C<s> is converted or consisted entirely of characters that are invariant |
855 | in utf8 (i.e., US-ASCII on non-EBCDIC machines). |
f9a63242 |
856 | |
37607a96 |
857 | =cut |
858 | */ |
f9a63242 |
859 | |
860 | U8 * |
e1ec3a88 |
861 | Perl_bytes_from_utf8(pTHX_ const U8 *s, STRLEN *len, bool *is_utf8) |
f9a63242 |
862 | { |
f9a63242 |
863 | U8 *d; |
e1ec3a88 |
864 | const U8 *start = s; |
865 | const U8 *send; |
f9a63242 |
866 | I32 count = 0; |
867 | |
7918f24d |
868 | PERL_ARGS_ASSERT_BYTES_FROM_UTF8; |
869 | |
96a5add6 |
870 | PERL_UNUSED_CONTEXT; |
f9a63242 |
871 | if (!*is_utf8) |
73d840c0 |
872 | return (U8 *)start; |
f9a63242 |
873 | |
1e54db1a |
874 | /* ensure valid UTF-8 and chars < 256 before converting string */ |
f9a63242 |
875 | for (send = s + *len; s < send;) { |
e1ec3a88 |
876 | U8 c = *s++; |
1d72bdf6 |
877 | if (!UTF8_IS_INVARIANT(c)) { |
db42d148 |
878 | if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send && |
879 | (c = *s++) && UTF8_IS_CONTINUATION(c)) |
880 | count++; |
881 | else |
73d840c0 |
882 | return (U8 *)start; |
db42d148 |
883 | } |
f9a63242 |
884 | } |
885 | |
35da51f7 |
886 | *is_utf8 = FALSE; |
f9a63242 |
887 | |
212542aa |
888 | Newx(d, (*len) - count + 1, U8); |
ef9edfd0 |
889 | s = start; start = d; |
f9a63242 |
890 | while (s < send) { |
891 | U8 c = *s++; |
c4d5f83a |
892 | if (!UTF8_IS_INVARIANT(c)) { |
893 | /* Then it is two-byte encoded */ |
894 | c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++); |
895 | c = ASCII_TO_NATIVE(c); |
896 | } |
897 | *d++ = c; |
f9a63242 |
898 | } |
899 | *d = '\0'; |
900 | *len = d - start; |
73d840c0 |
901 | return (U8 *)start; |
f9a63242 |
902 | } |
903 | |
904 | /* |
87cea99e |
905 | =for apidoc bytes_to_utf8 |
6940069f |
906 | |
2bbc8d55 |
907 | Converts a string C<s> of length C<len> from the native encoding into UTF-8. |
6662521e |
908 | Returns a pointer to the newly-created string, and sets C<len> to |
909 | reflect the new length. |
6940069f |
910 | |
2bbc8d55 |
911 | A NUL character will be written after the end of the string. |
912 | |
913 | If you want to convert to UTF-8 from encodings other than |
914 | the native (Latin1 or EBCDIC), |
c9ada85f |
915 | see sv_recode_to_utf8(). |
916 | |
497711e7 |
917 | =cut |
6940069f |
918 | */ |
919 | |
920 | U8* |
35a4481c |
921 | Perl_bytes_to_utf8(pTHX_ const U8 *s, STRLEN *len) |
6940069f |
922 | { |
35a4481c |
923 | const U8 * const send = s + (*len); |
6940069f |
924 | U8 *d; |
925 | U8 *dst; |
7918f24d |
926 | |
927 | PERL_ARGS_ASSERT_BYTES_TO_UTF8; |
96a5add6 |
928 | PERL_UNUSED_CONTEXT; |
6940069f |
929 | |
212542aa |
930 | Newx(d, (*len) * 2 + 1, U8); |
6940069f |
931 | dst = d; |
932 | |
933 | while (s < send) { |
35a4481c |
934 | const UV uv = NATIVE_TO_ASCII(*s++); |
c4d5f83a |
935 | if (UNI_IS_INVARIANT(uv)) |
eb160463 |
936 | *d++ = (U8)UTF_TO_NATIVE(uv); |
6940069f |
937 | else { |
eb160463 |
938 | *d++ = (U8)UTF8_EIGHT_BIT_HI(uv); |
939 | *d++ = (U8)UTF8_EIGHT_BIT_LO(uv); |
6940069f |
940 | } |
941 | } |
942 | *d = '\0'; |
6662521e |
943 | *len = d-dst; |
6940069f |
944 | return dst; |
945 | } |
946 | |
a0ed51b3 |
947 | /* |
dea0fc0b |
948 | * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8. |
a0ed51b3 |
949 | * |
950 | * Destination must be pre-extended to 3/2 source. Do not use in-place. |
951 | * We optimize for native, for obvious reasons. */ |
952 | |
953 | U8* |
dea0fc0b |
954 | Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 |
955 | { |
dea0fc0b |
956 | U8* pend; |
957 | U8* dstart = d; |
958 | |
7918f24d |
959 | PERL_ARGS_ASSERT_UTF16_TO_UTF8; |
960 | |
dea0fc0b |
961 | if (bytelen & 1) |
f5992bc4 |
962 | Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen %"UVuf, (UV)bytelen); |
dea0fc0b |
963 | |
964 | pend = p + bytelen; |
965 | |
a0ed51b3 |
966 | while (p < pend) { |
dea0fc0b |
967 | UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */ |
968 | p += 2; |
a0ed51b3 |
969 | if (uv < 0x80) { |
e294cc5d |
970 | #ifdef EBCDIC |
971 | *d++ = UNI_TO_NATIVE(uv); |
972 | #else |
eb160463 |
973 | *d++ = (U8)uv; |
e294cc5d |
974 | #endif |
a0ed51b3 |
975 | continue; |
976 | } |
977 | if (uv < 0x800) { |
eb160463 |
978 | *d++ = (U8)(( uv >> 6) | 0xc0); |
979 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
980 | continue; |
981 | } |
52b9aa85 |
982 | if (uv >= 0xd800 && uv <= 0xdbff) { /* surrogates */ |
01ea242b |
983 | if (p >= pend) { |
dea0fc0b |
984 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
01ea242b |
985 | } else { |
986 | UV low = (p[0] << 8) + p[1]; |
987 | p += 2; |
52b9aa85 |
988 | if (low < 0xdc00 || low > 0xdfff) |
01ea242b |
989 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
990 | uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000; |
991 | } |
dbde1951 |
992 | } else if (uv >= 0xdc00 && uv <= 0xdfff) { |
993 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
a0ed51b3 |
994 | } |
995 | if (uv < 0x10000) { |
eb160463 |
996 | *d++ = (U8)(( uv >> 12) | 0xe0); |
997 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
998 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
999 | continue; |
1000 | } |
1001 | else { |
eb160463 |
1002 | *d++ = (U8)(( uv >> 18) | 0xf0); |
1003 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
1004 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
1005 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
1006 | continue; |
1007 | } |
1008 | } |
dea0fc0b |
1009 | *newlen = d - dstart; |
a0ed51b3 |
1010 | return d; |
1011 | } |
1012 | |
1013 | /* Note: this one is slightly destructive of the source. */ |
1014 | |
1015 | U8* |
dea0fc0b |
1016 | Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 |
1017 | { |
1018 | U8* s = (U8*)p; |
d4c19fe8 |
1019 | U8* const send = s + bytelen; |
7918f24d |
1020 | |
1021 | PERL_ARGS_ASSERT_UTF16_TO_UTF8_REVERSED; |
1022 | |
e0ea5e2d |
1023 | if (bytelen & 1) |
1024 | Perl_croak(aTHX_ "panic: utf16_to_utf8_reversed: odd bytelen %"UVuf, |
1025 | (UV)bytelen); |
1026 | |
a0ed51b3 |
1027 | while (s < send) { |
d4c19fe8 |
1028 | const U8 tmp = s[0]; |
a0ed51b3 |
1029 | s[0] = s[1]; |
1030 | s[1] = tmp; |
1031 | s += 2; |
1032 | } |
dea0fc0b |
1033 | return utf16_to_utf8(p, d, bytelen, newlen); |
a0ed51b3 |
1034 | } |
1035 | |
1036 | /* for now these are all defined (inefficiently) in terms of the utf8 versions */ |
1037 | |
1038 | bool |
84afefe6 |
1039 | Perl_is_uni_alnum(pTHX_ UV c) |
a0ed51b3 |
1040 | { |
89ebb4a3 |
1041 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1042 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1043 | return is_utf8_alnum(tmpbuf); |
1044 | } |
1045 | |
1046 | bool |
84afefe6 |
1047 | Perl_is_uni_idfirst(pTHX_ UV c) |
a0ed51b3 |
1048 | { |
89ebb4a3 |
1049 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1050 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1051 | return is_utf8_idfirst(tmpbuf); |
1052 | } |
1053 | |
1054 | bool |
84afefe6 |
1055 | Perl_is_uni_alpha(pTHX_ UV c) |
a0ed51b3 |
1056 | { |
89ebb4a3 |
1057 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1058 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1059 | return is_utf8_alpha(tmpbuf); |
1060 | } |
1061 | |
1062 | bool |
84afefe6 |
1063 | Perl_is_uni_ascii(pTHX_ UV c) |
4d61ec05 |
1064 | { |
89ebb4a3 |
1065 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1066 | uvchr_to_utf8(tmpbuf, c); |
4d61ec05 |
1067 | return is_utf8_ascii(tmpbuf); |
1068 | } |
1069 | |
1070 | bool |
84afefe6 |
1071 | Perl_is_uni_space(pTHX_ UV c) |
a0ed51b3 |
1072 | { |
89ebb4a3 |
1073 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1074 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1075 | return is_utf8_space(tmpbuf); |
1076 | } |
1077 | |
1078 | bool |
84afefe6 |
1079 | Perl_is_uni_digit(pTHX_ UV c) |
a0ed51b3 |
1080 | { |
89ebb4a3 |
1081 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1082 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1083 | return is_utf8_digit(tmpbuf); |
1084 | } |
1085 | |
1086 | bool |
84afefe6 |
1087 | Perl_is_uni_upper(pTHX_ UV c) |
a0ed51b3 |
1088 | { |
89ebb4a3 |
1089 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1090 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1091 | return is_utf8_upper(tmpbuf); |
1092 | } |
1093 | |
1094 | bool |
84afefe6 |
1095 | Perl_is_uni_lower(pTHX_ UV c) |
a0ed51b3 |
1096 | { |
89ebb4a3 |
1097 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1098 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1099 | return is_utf8_lower(tmpbuf); |
1100 | } |
1101 | |
1102 | bool |
84afefe6 |
1103 | Perl_is_uni_cntrl(pTHX_ UV c) |
b8c5462f |
1104 | { |
89ebb4a3 |
1105 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1106 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f |
1107 | return is_utf8_cntrl(tmpbuf); |
1108 | } |
1109 | |
1110 | bool |
84afefe6 |
1111 | Perl_is_uni_graph(pTHX_ UV c) |
b8c5462f |
1112 | { |
89ebb4a3 |
1113 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1114 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f |
1115 | return is_utf8_graph(tmpbuf); |
1116 | } |
1117 | |
1118 | bool |
84afefe6 |
1119 | Perl_is_uni_print(pTHX_ UV c) |
a0ed51b3 |
1120 | { |
89ebb4a3 |
1121 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1122 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1123 | return is_utf8_print(tmpbuf); |
1124 | } |
1125 | |
b8c5462f |
1126 | bool |
84afefe6 |
1127 | Perl_is_uni_punct(pTHX_ UV c) |
b8c5462f |
1128 | { |
89ebb4a3 |
1129 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1130 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f |
1131 | return is_utf8_punct(tmpbuf); |
1132 | } |
1133 | |
4d61ec05 |
1134 | bool |
84afefe6 |
1135 | Perl_is_uni_xdigit(pTHX_ UV c) |
4d61ec05 |
1136 | { |
89ebb4a3 |
1137 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
230880c1 |
1138 | uvchr_to_utf8(tmpbuf, c); |
4d61ec05 |
1139 | return is_utf8_xdigit(tmpbuf); |
1140 | } |
1141 | |
84afefe6 |
1142 | UV |
1143 | Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
1144 | { |
7918f24d |
1145 | PERL_ARGS_ASSERT_TO_UNI_UPPER; |
1146 | |
0ebc6274 |
1147 | uvchr_to_utf8(p, c); |
1148 | return to_utf8_upper(p, p, lenp); |
a0ed51b3 |
1149 | } |
1150 | |
84afefe6 |
1151 | UV |
1152 | Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
1153 | { |
7918f24d |
1154 | PERL_ARGS_ASSERT_TO_UNI_TITLE; |
1155 | |
0ebc6274 |
1156 | uvchr_to_utf8(p, c); |
1157 | return to_utf8_title(p, p, lenp); |
a0ed51b3 |
1158 | } |
1159 | |
84afefe6 |
1160 | UV |
1161 | Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
1162 | { |
7918f24d |
1163 | PERL_ARGS_ASSERT_TO_UNI_LOWER; |
1164 | |
0ebc6274 |
1165 | uvchr_to_utf8(p, c); |
1166 | return to_utf8_lower(p, p, lenp); |
a0ed51b3 |
1167 | } |
1168 | |
84afefe6 |
1169 | UV |
1170 | Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp) |
1171 | { |
7918f24d |
1172 | PERL_ARGS_ASSERT_TO_UNI_FOLD; |
1173 | |
0ebc6274 |
1174 | uvchr_to_utf8(p, c); |
1175 | return to_utf8_fold(p, p, lenp); |
84afefe6 |
1176 | } |
1177 | |
a0ed51b3 |
1178 | /* for now these all assume no locale info available for Unicode > 255 */ |
1179 | |
1180 | bool |
84afefe6 |
1181 | Perl_is_uni_alnum_lc(pTHX_ UV c) |
a0ed51b3 |
1182 | { |
1183 | return is_uni_alnum(c); /* XXX no locale support yet */ |
1184 | } |
1185 | |
1186 | bool |
84afefe6 |
1187 | Perl_is_uni_idfirst_lc(pTHX_ UV c) |
a0ed51b3 |
1188 | { |
1189 | return is_uni_idfirst(c); /* XXX no locale support yet */ |
1190 | } |
1191 | |
1192 | bool |
84afefe6 |
1193 | Perl_is_uni_alpha_lc(pTHX_ UV c) |
a0ed51b3 |
1194 | { |
1195 | return is_uni_alpha(c); /* XXX no locale support yet */ |
1196 | } |
1197 | |
1198 | bool |
84afefe6 |
1199 | Perl_is_uni_ascii_lc(pTHX_ UV c) |
4d61ec05 |
1200 | { |
1201 | return is_uni_ascii(c); /* XXX no locale support yet */ |
1202 | } |
1203 | |
1204 | bool |
84afefe6 |
1205 | Perl_is_uni_space_lc(pTHX_ UV c) |
a0ed51b3 |
1206 | { |
1207 | return is_uni_space(c); /* XXX no locale support yet */ |
1208 | } |
1209 | |
1210 | bool |
84afefe6 |
1211 | Perl_is_uni_digit_lc(pTHX_ UV c) |
a0ed51b3 |
1212 | { |
1213 | return is_uni_digit(c); /* XXX no locale support yet */ |
1214 | } |
1215 | |
1216 | bool |
84afefe6 |
1217 | Perl_is_uni_upper_lc(pTHX_ UV c) |
a0ed51b3 |
1218 | { |
1219 | return is_uni_upper(c); /* XXX no locale support yet */ |
1220 | } |
1221 | |
1222 | bool |
84afefe6 |
1223 | Perl_is_uni_lower_lc(pTHX_ UV c) |
a0ed51b3 |
1224 | { |
1225 | return is_uni_lower(c); /* XXX no locale support yet */ |
1226 | } |
1227 | |
1228 | bool |
84afefe6 |
1229 | Perl_is_uni_cntrl_lc(pTHX_ UV c) |
b8c5462f |
1230 | { |
1231 | return is_uni_cntrl(c); /* XXX no locale support yet */ |
1232 | } |
1233 | |
1234 | bool |
84afefe6 |
1235 | Perl_is_uni_graph_lc(pTHX_ UV c) |
b8c5462f |
1236 | { |
1237 | return is_uni_graph(c); /* XXX no locale support yet */ |
1238 | } |
1239 | |
1240 | bool |
84afefe6 |
1241 | Perl_is_uni_print_lc(pTHX_ UV c) |
a0ed51b3 |
1242 | { |
1243 | return is_uni_print(c); /* XXX no locale support yet */ |
1244 | } |
1245 | |
b8c5462f |
1246 | bool |
84afefe6 |
1247 | Perl_is_uni_punct_lc(pTHX_ UV c) |
b8c5462f |
1248 | { |
1249 | return is_uni_punct(c); /* XXX no locale support yet */ |
1250 | } |
1251 | |
4d61ec05 |
1252 | bool |
84afefe6 |
1253 | Perl_is_uni_xdigit_lc(pTHX_ UV c) |
4d61ec05 |
1254 | { |
1255 | return is_uni_xdigit(c); /* XXX no locale support yet */ |
1256 | } |
1257 | |
b7ac61fa |
1258 | U32 |
1259 | Perl_to_uni_upper_lc(pTHX_ U32 c) |
1260 | { |
ee099d14 |
1261 | /* XXX returns only the first character -- do not use XXX */ |
1262 | /* XXX no locale support yet */ |
1263 | STRLEN len; |
89ebb4a3 |
1264 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 |
1265 | return (U32)to_uni_upper(c, tmpbuf, &len); |
b7ac61fa |
1266 | } |
1267 | |
1268 | U32 |
1269 | Perl_to_uni_title_lc(pTHX_ U32 c) |
1270 | { |
ee099d14 |
1271 | /* XXX returns only the first character XXX -- do not use XXX */ |
1272 | /* XXX no locale support yet */ |
1273 | STRLEN len; |
89ebb4a3 |
1274 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 |
1275 | return (U32)to_uni_title(c, tmpbuf, &len); |
b7ac61fa |
1276 | } |
1277 | |
1278 | U32 |
1279 | Perl_to_uni_lower_lc(pTHX_ U32 c) |
1280 | { |
ee099d14 |
1281 | /* XXX returns only the first character -- do not use XXX */ |
1282 | /* XXX no locale support yet */ |
1283 | STRLEN len; |
89ebb4a3 |
1284 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 |
1285 | return (U32)to_uni_lower(c, tmpbuf, &len); |
b7ac61fa |
1286 | } |
1287 | |
7452cf6a |
1288 | static bool |
5141f98e |
1289 | S_is_utf8_common(pTHX_ const U8 *const p, SV **swash, |
bde6a22d |
1290 | const char *const swashname) |
1291 | { |
97aff369 |
1292 | dVAR; |
7918f24d |
1293 | |
1294 | PERL_ARGS_ASSERT_IS_UTF8_COMMON; |
1295 | |
bde6a22d |
1296 | if (!is_utf8_char(p)) |
1297 | return FALSE; |
1298 | if (!*swash) |
711a919c |
1299 | *swash = swash_init("utf8", swashname, &PL_sv_undef, 1, 0); |
bde6a22d |
1300 | return swash_fetch(*swash, p, TRUE) != 0; |
1301 | } |
1302 | |
1303 | bool |
7fc63493 |
1304 | Perl_is_utf8_alnum(pTHX_ const U8 *p) |
a0ed51b3 |
1305 | { |
97aff369 |
1306 | dVAR; |
7918f24d |
1307 | |
1308 | PERL_ARGS_ASSERT_IS_UTF8_ALNUM; |
1309 | |
671c33bf |
1310 | /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true |
1311 | * descendant of isalnum(3), in other words, it doesn't |
1312 | * contain the '_'. --jhi */ |
d4c19fe8 |
1313 | return is_utf8_common(p, &PL_utf8_alnum, "IsWord"); |
a0ed51b3 |
1314 | } |
1315 | |
1316 | bool |
7fc63493 |
1317 | Perl_is_utf8_idfirst(pTHX_ const U8 *p) /* The naming is historical. */ |
a0ed51b3 |
1318 | { |
97aff369 |
1319 | dVAR; |
7918f24d |
1320 | |
1321 | PERL_ARGS_ASSERT_IS_UTF8_IDFIRST; |
1322 | |
82686b01 |
1323 | if (*p == '_') |
1324 | return TRUE; |
bde6a22d |
1325 | /* is_utf8_idstart would be more logical. */ |
d4c19fe8 |
1326 | return is_utf8_common(p, &PL_utf8_idstart, "IdStart"); |
82686b01 |
1327 | } |
1328 | |
1329 | bool |
7fc63493 |
1330 | Perl_is_utf8_idcont(pTHX_ const U8 *p) |
82686b01 |
1331 | { |
97aff369 |
1332 | dVAR; |
7918f24d |
1333 | |
1334 | PERL_ARGS_ASSERT_IS_UTF8_IDCONT; |
1335 | |
82686b01 |
1336 | if (*p == '_') |
1337 | return TRUE; |
d4c19fe8 |
1338 | return is_utf8_common(p, &PL_utf8_idcont, "IdContinue"); |
a0ed51b3 |
1339 | } |
1340 | |
1341 | bool |
7fc63493 |
1342 | Perl_is_utf8_alpha(pTHX_ const U8 *p) |
a0ed51b3 |
1343 | { |
97aff369 |
1344 | dVAR; |
7918f24d |
1345 | |
1346 | PERL_ARGS_ASSERT_IS_UTF8_ALPHA; |
1347 | |
d4c19fe8 |
1348 | return is_utf8_common(p, &PL_utf8_alpha, "IsAlpha"); |
a0ed51b3 |
1349 | } |
1350 | |
1351 | bool |
7fc63493 |
1352 | Perl_is_utf8_ascii(pTHX_ const U8 *p) |
b8c5462f |
1353 | { |
97aff369 |
1354 | dVAR; |
7918f24d |
1355 | |
1356 | PERL_ARGS_ASSERT_IS_UTF8_ASCII; |
1357 | |
d4c19fe8 |
1358 | return is_utf8_common(p, &PL_utf8_ascii, "IsAscii"); |
b8c5462f |
1359 | } |
1360 | |
1361 | bool |
7fc63493 |
1362 | Perl_is_utf8_space(pTHX_ const U8 *p) |
a0ed51b3 |
1363 | { |
97aff369 |
1364 | dVAR; |
7918f24d |
1365 | |
1366 | PERL_ARGS_ASSERT_IS_UTF8_SPACE; |
1367 | |
d4c19fe8 |
1368 | return is_utf8_common(p, &PL_utf8_space, "IsSpacePerl"); |
a0ed51b3 |
1369 | } |
1370 | |
1371 | bool |
d1eb3177 |
1372 | Perl_is_utf8_perl_space(pTHX_ const U8 *p) |
1373 | { |
1374 | dVAR; |
1375 | |
1376 | PERL_ARGS_ASSERT_IS_UTF8_PERL_SPACE; |
1377 | |
1378 | return is_utf8_common(p, &PL_utf8_perl_space, "IsPerlSpace"); |
1379 | } |
1380 | |
1381 | bool |
1382 | Perl_is_utf8_perl_word(pTHX_ const U8 *p) |
1383 | { |
1384 | dVAR; |
1385 | |
1386 | PERL_ARGS_ASSERT_IS_UTF8_PERL_WORD; |
1387 | |
1388 | return is_utf8_common(p, &PL_utf8_perl_word, "IsPerlWord"); |
1389 | } |
1390 | |
1391 | bool |
7fc63493 |
1392 | Perl_is_utf8_digit(pTHX_ const U8 *p) |
a0ed51b3 |
1393 | { |
97aff369 |
1394 | dVAR; |
7918f24d |
1395 | |
1396 | PERL_ARGS_ASSERT_IS_UTF8_DIGIT; |
1397 | |
d4c19fe8 |
1398 | return is_utf8_common(p, &PL_utf8_digit, "IsDigit"); |
a0ed51b3 |
1399 | } |
1400 | |
1401 | bool |
d1eb3177 |
1402 | Perl_is_utf8_posix_digit(pTHX_ const U8 *p) |
1403 | { |
1404 | dVAR; |
1405 | |
1406 | PERL_ARGS_ASSERT_IS_UTF8_POSIX_DIGIT; |
1407 | |
1408 | return is_utf8_common(p, &PL_utf8_posix_digit, "IsPosixDigit"); |
1409 | } |
1410 | |
1411 | bool |
7fc63493 |
1412 | Perl_is_utf8_upper(pTHX_ const U8 *p) |
a0ed51b3 |
1413 | { |
97aff369 |
1414 | dVAR; |
7918f24d |
1415 | |
1416 | PERL_ARGS_ASSERT_IS_UTF8_UPPER; |
1417 | |
d4c19fe8 |
1418 | return is_utf8_common(p, &PL_utf8_upper, "IsUppercase"); |
a0ed51b3 |
1419 | } |
1420 | |
1421 | bool |
7fc63493 |
1422 | Perl_is_utf8_lower(pTHX_ const U8 *p) |
a0ed51b3 |
1423 | { |
97aff369 |
1424 | dVAR; |
7918f24d |
1425 | |
1426 | PERL_ARGS_ASSERT_IS_UTF8_LOWER; |
1427 | |
d4c19fe8 |
1428 | return is_utf8_common(p, &PL_utf8_lower, "IsLowercase"); |
a0ed51b3 |
1429 | } |
1430 | |
1431 | bool |
7fc63493 |
1432 | Perl_is_utf8_cntrl(pTHX_ const U8 *p) |
b8c5462f |
1433 | { |
97aff369 |
1434 | dVAR; |
7918f24d |
1435 | |
1436 | PERL_ARGS_ASSERT_IS_UTF8_CNTRL; |
1437 | |
d4c19fe8 |
1438 | return is_utf8_common(p, &PL_utf8_cntrl, "IsCntrl"); |
b8c5462f |
1439 | } |
1440 | |
1441 | bool |
7fc63493 |
1442 | Perl_is_utf8_graph(pTHX_ const U8 *p) |
b8c5462f |
1443 | { |
97aff369 |
1444 | dVAR; |
7918f24d |
1445 | |
1446 | PERL_ARGS_ASSERT_IS_UTF8_GRAPH; |
1447 | |
d4c19fe8 |
1448 | return is_utf8_common(p, &PL_utf8_graph, "IsGraph"); |
b8c5462f |
1449 | } |
1450 | |
1451 | bool |
7fc63493 |
1452 | Perl_is_utf8_print(pTHX_ const U8 *p) |
a0ed51b3 |
1453 | { |
97aff369 |
1454 | dVAR; |
7918f24d |
1455 | |
1456 | PERL_ARGS_ASSERT_IS_UTF8_PRINT; |
1457 | |
d4c19fe8 |
1458 | return is_utf8_common(p, &PL_utf8_print, "IsPrint"); |
a0ed51b3 |
1459 | } |
1460 | |
1461 | bool |
7fc63493 |
1462 | Perl_is_utf8_punct(pTHX_ const U8 *p) |
b8c5462f |
1463 | { |
97aff369 |
1464 | dVAR; |
7918f24d |
1465 | |
1466 | PERL_ARGS_ASSERT_IS_UTF8_PUNCT; |
1467 | |
d4c19fe8 |
1468 | return is_utf8_common(p, &PL_utf8_punct, "IsPunct"); |
b8c5462f |
1469 | } |
1470 | |
1471 | bool |
7fc63493 |
1472 | Perl_is_utf8_xdigit(pTHX_ const U8 *p) |
b8c5462f |
1473 | { |
97aff369 |
1474 | dVAR; |
7918f24d |
1475 | |
1476 | PERL_ARGS_ASSERT_IS_UTF8_XDIGIT; |
1477 | |
d1eb3177 |
1478 | return is_utf8_common(p, &PL_utf8_xdigit, "IsXDigit"); |
b8c5462f |
1479 | } |
1480 | |
1481 | bool |
7fc63493 |
1482 | Perl_is_utf8_mark(pTHX_ const U8 *p) |
a0ed51b3 |
1483 | { |
97aff369 |
1484 | dVAR; |
7918f24d |
1485 | |
1486 | PERL_ARGS_ASSERT_IS_UTF8_MARK; |
1487 | |
d4c19fe8 |
1488 | return is_utf8_common(p, &PL_utf8_mark, "IsM"); |
a0ed51b3 |
1489 | } |
1490 | |
6b5c0936 |
1491 | /* |
87cea99e |
1492 | =for apidoc to_utf8_case |
6b5c0936 |
1493 | |
1494 | The "p" contains the pointer to the UTF-8 string encoding |
1495 | the character that is being converted. |
1496 | |
1497 | The "ustrp" is a pointer to the character buffer to put the |
1498 | conversion result to. The "lenp" is a pointer to the length |
1499 | of the result. |
1500 | |
0134edef |
1501 | The "swashp" is a pointer to the swash to use. |
6b5c0936 |
1502 | |
0134edef |
1503 | Both the special and normal mappings are stored lib/unicore/To/Foo.pl, |
8fe4d5b2 |
1504 | and loaded by SWASHNEW, using lib/utf8_heavy.pl. The special (usually, |
0134edef |
1505 | but not always, a multicharacter mapping), is tried first. |
6b5c0936 |
1506 | |
0134edef |
1507 | The "special" is a string like "utf8::ToSpecLower", which means the |
1508 | hash %utf8::ToSpecLower. The access to the hash is through |
1509 | Perl_to_utf8_case(). |
6b5c0936 |
1510 | |
0134edef |
1511 | The "normal" is a string like "ToLower" which means the swash |
1512 | %utf8::ToLower. |
1513 | |
1514 | =cut */ |
6b5c0936 |
1515 | |
2104c8d9 |
1516 | UV |
9a957fbc |
1517 | Perl_to_utf8_case(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp, |
1518 | SV **swashp, const char *normal, const char *special) |
a0ed51b3 |
1519 | { |
97aff369 |
1520 | dVAR; |
89ebb4a3 |
1521 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
0134edef |
1522 | STRLEN len = 0; |
aec46f14 |
1523 | const UV uv0 = utf8_to_uvchr(p, NULL); |
1feea2c7 |
1524 | /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings |
1525 | * are necessary in EBCDIC, they are redundant no-ops |
1526 | * in ASCII-ish platforms, and hopefully optimized away. */ |
f54cb97a |
1527 | const UV uv1 = NATIVE_TO_UNI(uv0); |
7918f24d |
1528 | |
1529 | PERL_ARGS_ASSERT_TO_UTF8_CASE; |
1530 | |
1feea2c7 |
1531 | uvuni_to_utf8(tmpbuf, uv1); |
0134edef |
1532 | |
1533 | if (!*swashp) /* load on-demand */ |
1534 | *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0); |
1535 | |
b08cf34e |
1536 | /* The 0xDF is the only special casing Unicode code point below 0x100. */ |
1537 | if (special && (uv1 == 0xDF || uv1 > 0xFF)) { |
0134edef |
1538 | /* It might be "special" (sometimes, but not always, |
2a37f04d |
1539 | * a multicharacter mapping) */ |
6673a63c |
1540 | HV * const hv = get_hv(special, 0); |
b08cf34e |
1541 | SV **svp; |
1542 | |
35da51f7 |
1543 | if (hv && |
b08cf34e |
1544 | (svp = hv_fetch(hv, (const char*)tmpbuf, UNISKIP(uv1), FALSE)) && |
1545 | (*svp)) { |
cfd0369c |
1546 | const char *s; |
47654450 |
1547 | |
cfd0369c |
1548 | s = SvPV_const(*svp, len); |
47654450 |
1549 | if (len == 1) |
1550 | len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp; |
2a37f04d |
1551 | else { |
2f9475ad |
1552 | #ifdef EBCDIC |
1553 | /* If we have EBCDIC we need to remap the characters |
1554 | * since any characters in the low 256 are Unicode |
1555 | * code points, not EBCDIC. */ |
7cda7a3d |
1556 | U8 *t = (U8*)s, *tend = t + len, *d; |
2f9475ad |
1557 | |
1558 | d = tmpbuf; |
b08cf34e |
1559 | if (SvUTF8(*svp)) { |
2f9475ad |
1560 | STRLEN tlen = 0; |
1561 | |
1562 | while (t < tend) { |
d4c19fe8 |
1563 | const UV c = utf8_to_uvchr(t, &tlen); |
2f9475ad |
1564 | if (tlen > 0) { |
1565 | d = uvchr_to_utf8(d, UNI_TO_NATIVE(c)); |
1566 | t += tlen; |
1567 | } |
1568 | else |
1569 | break; |
1570 | } |
1571 | } |
1572 | else { |
36fec512 |
1573 | while (t < tend) { |
1574 | d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t)); |
1575 | t++; |
1576 | } |
2f9475ad |
1577 | } |
1578 | len = d - tmpbuf; |
1579 | Copy(tmpbuf, ustrp, len, U8); |
1580 | #else |
d2dcd0fb |
1581 | Copy(s, ustrp, len, U8); |
2f9475ad |
1582 | #endif |
29e98929 |
1583 | } |
983ffd37 |
1584 | } |
0134edef |
1585 | } |
1586 | |
1587 | if (!len && *swashp) { |
d4c19fe8 |
1588 | const UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE); |
1589 | |
0134edef |
1590 | if (uv2) { |
1591 | /* It was "normal" (a single character mapping). */ |
d4c19fe8 |
1592 | const UV uv3 = UNI_TO_NATIVE(uv2); |
e9101d72 |
1593 | len = uvchr_to_utf8(ustrp, uv3) - ustrp; |
2a37f04d |
1594 | } |
1595 | } |
1feea2c7 |
1596 | |
0134edef |
1597 | if (!len) /* Neither: just copy. */ |
1598 | len = uvchr_to_utf8(ustrp, uv0) - ustrp; |
1599 | |
2a37f04d |
1600 | if (lenp) |
1601 | *lenp = len; |
1602 | |
0134edef |
1603 | return len ? utf8_to_uvchr(ustrp, 0) : 0; |
a0ed51b3 |
1604 | } |
1605 | |
d3e79532 |
1606 | /* |
87cea99e |
1607 | =for apidoc to_utf8_upper |
d3e79532 |
1608 | |
1609 | Convert the UTF-8 encoded character at p to its uppercase version and |
1610 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1611 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since |
1612 | the uppercase version may be longer than the original character. |
d3e79532 |
1613 | |
1614 | The first character of the uppercased version is returned |
1615 | (but note, as explained above, that there may be more.) |
1616 | |
1617 | =cut */ |
1618 | |
2104c8d9 |
1619 | UV |
7fc63493 |
1620 | Perl_to_utf8_upper(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 |
1621 | { |
97aff369 |
1622 | dVAR; |
7918f24d |
1623 | |
1624 | PERL_ARGS_ASSERT_TO_UTF8_UPPER; |
1625 | |
983ffd37 |
1626 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1627 | &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper"); |
983ffd37 |
1628 | } |
a0ed51b3 |
1629 | |
d3e79532 |
1630 | /* |
87cea99e |
1631 | =for apidoc to_utf8_title |
d3e79532 |
1632 | |
1633 | Convert the UTF-8 encoded character at p to its titlecase version and |
1634 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1635 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
1636 | titlecase version may be longer than the original character. |
d3e79532 |
1637 | |
1638 | The first character of the titlecased version is returned |
1639 | (but note, as explained above, that there may be more.) |
1640 | |
1641 | =cut */ |
1642 | |
983ffd37 |
1643 | UV |
7fc63493 |
1644 | Perl_to_utf8_title(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
983ffd37 |
1645 | { |
97aff369 |
1646 | dVAR; |
7918f24d |
1647 | |
1648 | PERL_ARGS_ASSERT_TO_UTF8_TITLE; |
1649 | |
983ffd37 |
1650 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1651 | &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle"); |
a0ed51b3 |
1652 | } |
1653 | |
d3e79532 |
1654 | /* |
87cea99e |
1655 | =for apidoc to_utf8_lower |
d3e79532 |
1656 | |
1657 | Convert the UTF-8 encoded character at p to its lowercase version and |
1658 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1659 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
1660 | lowercase version may be longer than the original character. |
d3e79532 |
1661 | |
1662 | The first character of the lowercased version is returned |
1663 | (but note, as explained above, that there may be more.) |
1664 | |
1665 | =cut */ |
1666 | |
2104c8d9 |
1667 | UV |
7fc63493 |
1668 | Perl_to_utf8_lower(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 |
1669 | { |
97aff369 |
1670 | dVAR; |
7918f24d |
1671 | |
1672 | PERL_ARGS_ASSERT_TO_UTF8_LOWER; |
1673 | |
983ffd37 |
1674 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1675 | &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower"); |
1676 | } |
1677 | |
d3e79532 |
1678 | /* |
87cea99e |
1679 | =for apidoc to_utf8_fold |
d3e79532 |
1680 | |
1681 | Convert the UTF-8 encoded character at p to its foldcase version and |
1682 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1683 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
d3e79532 |
1684 | foldcase version may be longer than the original character (up to |
1685 | three characters). |
1686 | |
1687 | The first character of the foldcased version is returned |
1688 | (but note, as explained above, that there may be more.) |
1689 | |
1690 | =cut */ |
1691 | |
b4e400f9 |
1692 | UV |
7fc63493 |
1693 | Perl_to_utf8_fold(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
b4e400f9 |
1694 | { |
97aff369 |
1695 | dVAR; |
7918f24d |
1696 | |
1697 | PERL_ARGS_ASSERT_TO_UTF8_FOLD; |
1698 | |
b4e400f9 |
1699 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
1700 | &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold"); |
a0ed51b3 |
1701 | } |
1702 | |
711a919c |
1703 | /* Note: |
1704 | * A "swash" is a swatch hash. |
1705 | * A "swatch" is a bit vector generated by utf8.c:S_swash_get(). |
1706 | * C<pkg> is a pointer to a package name for SWASHNEW, should be "utf8". |
1707 | * For other parameters, see utf8::SWASHNEW in lib/utf8_heavy.pl. |
1708 | */ |
a0ed51b3 |
1709 | SV* |
7fc63493 |
1710 | Perl_swash_init(pTHX_ const char* pkg, const char* name, SV *listsv, I32 minbits, I32 none) |
a0ed51b3 |
1711 | { |
27da23d5 |
1712 | dVAR; |
a0ed51b3 |
1713 | SV* retval; |
8e84507e |
1714 | dSP; |
7fc63493 |
1715 | const size_t pkg_len = strlen(pkg); |
1716 | const size_t name_len = strlen(name); |
da51bb9b |
1717 | HV * const stash = gv_stashpvn(pkg, pkg_len, 0); |
f8be5cf0 |
1718 | SV* errsv_save; |
ce3b816e |
1719 | |
7918f24d |
1720 | PERL_ARGS_ASSERT_SWASH_INIT; |
1721 | |
96ca9f55 |
1722 | PUSHSTACKi(PERLSI_MAGIC); |
1723 | ENTER; |
1724 | SAVEI32(PL_hints); |
1725 | PL_hints = 0; |
1726 | save_re_context(); |
1b026014 |
1727 | if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */ |
ce3b816e |
1728 | ENTER; |
f8be5cf0 |
1729 | errsv_save = newSVsv(ERRSV); |
dc0c6abb |
1730 | /* It is assumed that callers of this routine are not passing in any |
1731 | user derived data. */ |
1732 | /* Need to do this after save_re_context() as it will set PL_tainted to |
1733 | 1 while saving $1 etc (see the code after getrx: in Perl_magic_get). |
1734 | Even line to create errsv_save can turn on PL_tainted. */ |
1735 | SAVEBOOL(PL_tainted); |
1736 | PL_tainted = 0; |
71bed85a |
1737 | Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpvn(pkg,pkg_len), |
a0714e2c |
1738 | NULL); |
f8be5cf0 |
1739 | if (!SvTRUE(ERRSV)) |
1740 | sv_setsv(ERRSV, errsv_save); |
1741 | SvREFCNT_dec(errsv_save); |
ce3b816e |
1742 | LEAVE; |
1743 | } |
1744 | SPAGAIN; |
a0ed51b3 |
1745 | PUSHMARK(SP); |
1746 | EXTEND(SP,5); |
6e449a3a |
1747 | mPUSHp(pkg, pkg_len); |
1748 | mPUSHp(name, name_len); |
a0ed51b3 |
1749 | PUSHs(listsv); |
6e449a3a |
1750 | mPUSHi(minbits); |
1751 | mPUSHi(none); |
a0ed51b3 |
1752 | PUTBACK; |
f8be5cf0 |
1753 | errsv_save = newSVsv(ERRSV); |
864dbfa3 |
1754 | if (call_method("SWASHNEW", G_SCALAR)) |
8e84507e |
1755 | retval = newSVsv(*PL_stack_sp--); |
a0ed51b3 |
1756 | else |
e24b16f9 |
1757 | retval = &PL_sv_undef; |
f8be5cf0 |
1758 | if (!SvTRUE(ERRSV)) |
1759 | sv_setsv(ERRSV, errsv_save); |
1760 | SvREFCNT_dec(errsv_save); |
a0ed51b3 |
1761 | LEAVE; |
1762 | POPSTACK; |
923e4eb5 |
1763 | if (IN_PERL_COMPILETIME) { |
623e6609 |
1764 | CopHINTS_set(PL_curcop, PL_hints); |
a0ed51b3 |
1765 | } |
bc45ce41 |
1766 | if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) { |
1767 | if (SvPOK(retval)) |
35c1215d |
1768 | Perl_croak(aTHX_ "Can't find Unicode property definition \"%"SVf"\"", |
be2597df |
1769 | SVfARG(retval)); |
cea2e8a9 |
1770 | Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref"); |
bc45ce41 |
1771 | } |
a0ed51b3 |
1772 | return retval; |
1773 | } |
1774 | |
035d37be |
1775 | |
1776 | /* This API is wrong for special case conversions since we may need to |
1777 | * return several Unicode characters for a single Unicode character |
1778 | * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is |
1779 | * the lower-level routine, and it is similarly broken for returning |
1780 | * multiple values. --jhi */ |
979f2922 |
1781 | /* Now SWASHGET is recasted into S_swash_get in this file. */ |
680c470c |
1782 | |
1783 | /* Note: |
1784 | * Returns the value of property/mapping C<swash> for the first character |
1785 | * of the string C<ptr>. If C<do_utf8> is true, the string C<ptr> is |
1786 | * assumed to be in utf8. If C<do_utf8> is false, the string C<ptr> is |
1787 | * assumed to be in native 8-bit encoding. Caches the swatch in C<swash>. |
1788 | */ |
a0ed51b3 |
1789 | UV |
680c470c |
1790 | Perl_swash_fetch(pTHX_ SV *swash, const U8 *ptr, bool do_utf8) |
a0ed51b3 |
1791 | { |
27da23d5 |
1792 | dVAR; |
ef8f7699 |
1793 | HV *const hv = MUTABLE_HV(SvRV(swash)); |
3568d838 |
1794 | U32 klen; |
1795 | U32 off; |
a0ed51b3 |
1796 | STRLEN slen; |
7d85a32c |
1797 | STRLEN needents; |
cfd0369c |
1798 | const U8 *tmps = NULL; |
a0ed51b3 |
1799 | U32 bit; |
979f2922 |
1800 | SV *swatch; |
3568d838 |
1801 | U8 tmputf8[2]; |
35da51f7 |
1802 | const UV c = NATIVE_TO_ASCII(*ptr); |
3568d838 |
1803 | |
7918f24d |
1804 | PERL_ARGS_ASSERT_SWASH_FETCH; |
1805 | |
3568d838 |
1806 | if (!do_utf8 && !UNI_IS_INVARIANT(c)) { |
979f2922 |
1807 | tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c); |
1808 | tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c); |
1809 | ptr = tmputf8; |
3568d838 |
1810 | } |
1811 | /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ |
1812 | * then the "swatch" is a vec() for al the chars which start |
1813 | * with 0xAA..0xYY |
1814 | * So the key in the hash (klen) is length of encoded char -1 |
1815 | */ |
1816 | klen = UTF8SKIP(ptr) - 1; |
1817 | off = ptr[klen]; |
a0ed51b3 |
1818 | |
979f2922 |
1819 | if (klen == 0) { |
7d85a32c |
1820 | /* If char in invariant then swatch is for all the invariant chars |
1e54db1a |
1821 | * In both UTF-8 and UTF-8-MOD that happens to be UTF_CONTINUATION_MARK |
7d85a32c |
1822 | */ |
979f2922 |
1823 | needents = UTF_CONTINUATION_MARK; |
1824 | off = NATIVE_TO_UTF(ptr[klen]); |
1825 | } |
1826 | else { |
7d85a32c |
1827 | /* If char is encoded then swatch is for the prefix */ |
979f2922 |
1828 | needents = (1 << UTF_ACCUMULATION_SHIFT); |
1829 | off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK; |
1830 | } |
7d85a32c |
1831 | |
a0ed51b3 |
1832 | /* |
1833 | * This single-entry cache saves about 1/3 of the utf8 overhead in test |
1834 | * suite. (That is, only 7-8% overall over just a hash cache. Still, |
1835 | * it's nothing to sniff at.) Pity we usually come through at least |
1836 | * two function calls to get here... |
1837 | * |
1838 | * NB: this code assumes that swatches are never modified, once generated! |
1839 | */ |
1840 | |
3568d838 |
1841 | if (hv == PL_last_swash_hv && |
a0ed51b3 |
1842 | klen == PL_last_swash_klen && |
27da23d5 |
1843 | (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) ) |
a0ed51b3 |
1844 | { |
1845 | tmps = PL_last_swash_tmps; |
1846 | slen = PL_last_swash_slen; |
1847 | } |
1848 | else { |
1849 | /* Try our second-level swatch cache, kept in a hash. */ |
e1ec3a88 |
1850 | SV** svp = hv_fetch(hv, (const char*)ptr, klen, FALSE); |
a0ed51b3 |
1851 | |
979f2922 |
1852 | /* If not cached, generate it via swash_get */ |
1853 | if (!svp || !SvPOK(*svp) |
1854 | || !(tmps = (const U8*)SvPV_const(*svp, slen))) { |
2b9d42f0 |
1855 | /* We use utf8n_to_uvuni() as we want an index into |
1856 | Unicode tables, not a native character number. |
1857 | */ |
aec46f14 |
1858 | const UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXBYTES, 0, |
872c91ae |
1859 | ckWARN(WARN_UTF8) ? |
1860 | 0 : UTF8_ALLOW_ANY); |
680c470c |
1861 | swatch = swash_get(swash, |
979f2922 |
1862 | /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */ |
1863 | (klen) ? (code_point & ~(needents - 1)) : 0, |
1864 | needents); |
1865 | |
923e4eb5 |
1866 | if (IN_PERL_COMPILETIME) |
623e6609 |
1867 | CopHINTS_set(PL_curcop, PL_hints); |
a0ed51b3 |
1868 | |
979f2922 |
1869 | svp = hv_store(hv, (const char *)ptr, klen, swatch, 0); |
a0ed51b3 |
1870 | |
979f2922 |
1871 | if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) |
1872 | || (slen << 3) < needents) |
660a4616 |
1873 | Perl_croak(aTHX_ "panic: swash_fetch got improper swatch"); |
a0ed51b3 |
1874 | } |
1875 | |
1876 | PL_last_swash_hv = hv; |
16d8f38a |
1877 | assert(klen <= sizeof(PL_last_swash_key)); |
eac04b2e |
1878 | PL_last_swash_klen = (U8)klen; |
cfd0369c |
1879 | /* FIXME change interpvar.h? */ |
1880 | PL_last_swash_tmps = (U8 *) tmps; |
a0ed51b3 |
1881 | PL_last_swash_slen = slen; |
1882 | if (klen) |
1883 | Copy(ptr, PL_last_swash_key, klen, U8); |
1884 | } |
1885 | |
9faf8d75 |
1886 | switch ((int)((slen << 3) / needents)) { |
a0ed51b3 |
1887 | case 1: |
1888 | bit = 1 << (off & 7); |
1889 | off >>= 3; |
1890 | return (tmps[off] & bit) != 0; |
1891 | case 8: |
1892 | return tmps[off]; |
1893 | case 16: |
1894 | off <<= 1; |
1895 | return (tmps[off] << 8) + tmps[off + 1] ; |
1896 | case 32: |
1897 | off <<= 2; |
1898 | return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ; |
1899 | } |
660a4616 |
1900 | Perl_croak(aTHX_ "panic: swash_fetch got swatch of unexpected bit width"); |
670f1322 |
1901 | NORETURN_FUNCTION_END; |
a0ed51b3 |
1902 | } |
2b9d42f0 |
1903 | |
979f2922 |
1904 | /* Note: |
1905 | * Returns a swatch (a bit vector string) for a code point sequence |
1906 | * that starts from the value C<start> and comprises the number C<span>. |
1907 | * A C<swash> must be an object created by SWASHNEW (see lib/utf8_heavy.pl). |
1908 | * Should be used via swash_fetch, which will cache the swatch in C<swash>. |
1909 | */ |
1910 | STATIC SV* |
1911 | S_swash_get(pTHX_ SV* swash, UV start, UV span) |
1912 | { |
1913 | SV *swatch; |
711a919c |
1914 | U8 *l, *lend, *x, *xend, *s; |
979f2922 |
1915 | STRLEN lcur, xcur, scur; |
ef8f7699 |
1916 | HV *const hv = MUTABLE_HV(SvRV(swash)); |
017a3ce5 |
1917 | SV** const listsvp = hv_fetchs(hv, "LIST", FALSE); |
1918 | SV** const typesvp = hv_fetchs(hv, "TYPE", FALSE); |
1919 | SV** const bitssvp = hv_fetchs(hv, "BITS", FALSE); |
1920 | SV** const nonesvp = hv_fetchs(hv, "NONE", FALSE); |
1921 | SV** const extssvp = hv_fetchs(hv, "EXTRAS", FALSE); |
0bd48802 |
1922 | const U8* const typestr = (U8*)SvPV_nolen(*typesvp); |
1923 | const int typeto = typestr[0] == 'T' && typestr[1] == 'o'; |
1924 | const STRLEN bits = SvUV(*bitssvp); |
1925 | const STRLEN octets = bits >> 3; /* if bits == 1, then octets == 0 */ |
1926 | const UV none = SvUV(*nonesvp); |
1927 | const UV end = start + span; |
979f2922 |
1928 | |
7918f24d |
1929 | PERL_ARGS_ASSERT_SWASH_GET; |
1930 | |
979f2922 |
1931 | if (bits != 1 && bits != 8 && bits != 16 && bits != 32) { |
660a4616 |
1932 | Perl_croak(aTHX_ "panic: swash_get doesn't expect bits %"UVuf, |
1933 | (UV)bits); |
979f2922 |
1934 | } |
1935 | |
1936 | /* create and initialize $swatch */ |
979f2922 |
1937 | scur = octets ? (span * octets) : (span + 7) / 8; |
e524fe40 |
1938 | swatch = newSV(scur); |
1939 | SvPOK_on(swatch); |
979f2922 |
1940 | s = (U8*)SvPVX(swatch); |
1941 | if (octets && none) { |
0bd48802 |
1942 | const U8* const e = s + scur; |
979f2922 |
1943 | while (s < e) { |
1944 | if (bits == 8) |
1945 | *s++ = (U8)(none & 0xff); |
1946 | else if (bits == 16) { |
1947 | *s++ = (U8)((none >> 8) & 0xff); |
1948 | *s++ = (U8)( none & 0xff); |
1949 | } |
1950 | else if (bits == 32) { |
1951 | *s++ = (U8)((none >> 24) & 0xff); |
1952 | *s++ = (U8)((none >> 16) & 0xff); |
1953 | *s++ = (U8)((none >> 8) & 0xff); |
1954 | *s++ = (U8)( none & 0xff); |
1955 | } |
1956 | } |
1957 | *s = '\0'; |
1958 | } |
1959 | else { |
1960 | (void)memzero((U8*)s, scur + 1); |
1961 | } |
1962 | SvCUR_set(swatch, scur); |
1963 | s = (U8*)SvPVX(swatch); |
1964 | |
1965 | /* read $swash->{LIST} */ |
1966 | l = (U8*)SvPV(*listsvp, lcur); |
1967 | lend = l + lcur; |
1968 | while (l < lend) { |
35da51f7 |
1969 | UV min, max, val; |
979f2922 |
1970 | STRLEN numlen; |
1971 | I32 flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX; |
1972 | |
0bd48802 |
1973 | U8* const nl = (U8*)memchr(l, '\n', lend - l); |
979f2922 |
1974 | |
1975 | numlen = lend - l; |
1976 | min = grok_hex((char *)l, &numlen, &flags, NULL); |
1977 | if (numlen) |
1978 | l += numlen; |
1979 | else if (nl) { |
1980 | l = nl + 1; /* 1 is length of "\n" */ |
1981 | continue; |
1982 | } |
1983 | else { |
1984 | l = lend; /* to LIST's end at which \n is not found */ |
1985 | break; |
1986 | } |
1987 | |
1988 | if (isBLANK(*l)) { |
1989 | ++l; |
1990 | flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX; |
1991 | numlen = lend - l; |
1992 | max = grok_hex((char *)l, &numlen, &flags, NULL); |
1993 | if (numlen) |
1994 | l += numlen; |
1995 | else |
1996 | max = min; |
1997 | |
1998 | if (octets) { |
1999 | if (isBLANK(*l)) { |
2000 | ++l; |
2001 | flags = PERL_SCAN_SILENT_ILLDIGIT | |
2002 | PERL_SCAN_DISALLOW_PREFIX; |
2003 | numlen = lend - l; |
2004 | val = grok_hex((char *)l, &numlen, &flags, NULL); |
2005 | if (numlen) |
2006 | l += numlen; |
2007 | else |
2008 | val = 0; |
2009 | } |
2010 | else { |
2011 | val = 0; |
2012 | if (typeto) { |
2013 | Perl_croak(aTHX_ "%s: illegal mapping '%s'", |
2014 | typestr, l); |
2015 | } |
2016 | } |
2017 | } |
711a919c |
2018 | else |
2019 | val = 0; /* bits == 1, then val should be ignored */ |
979f2922 |
2020 | } |
2021 | else { |
2022 | max = min; |
2023 | if (octets) { |
2024 | val = 0; |
2025 | if (typeto) { |
2026 | Perl_croak(aTHX_ "%s: illegal mapping '%s'", typestr, l); |
2027 | } |
2028 | } |
711a919c |
2029 | else |
2030 | val = 0; /* bits == 1, then val should be ignored */ |
979f2922 |
2031 | } |
2032 | |
2033 | if (nl) |
2034 | l = nl + 1; |
2035 | else |
2036 | l = lend; |
2037 | |
2038 | if (max < start) |
2039 | continue; |
2040 | |
2041 | if (octets) { |
35da51f7 |
2042 | UV key; |
979f2922 |
2043 | if (min < start) { |
2044 | if (!none || val < none) { |
2045 | val += start - min; |
2046 | } |
2047 | min = start; |
2048 | } |
2049 | for (key = min; key <= max; key++) { |
2050 | STRLEN offset; |
2051 | if (key >= end) |
2052 | goto go_out_list; |
2053 | /* offset must be non-negative (start <= min <= key < end) */ |
2054 | offset = octets * (key - start); |
2055 | if (bits == 8) |
2056 | s[offset] = (U8)(val & 0xff); |
2057 | else if (bits == 16) { |
2058 | s[offset ] = (U8)((val >> 8) & 0xff); |
2059 | s[offset + 1] = (U8)( val & 0xff); |
2060 | } |
2061 | else if (bits == 32) { |
2062 | s[offset ] = (U8)((val >> 24) & 0xff); |
2063 | s[offset + 1] = (U8)((val >> 16) & 0xff); |
2064 | s[offset + 2] = (U8)((val >> 8) & 0xff); |
2065 | s[offset + 3] = (U8)( val & 0xff); |
2066 | } |
2067 | |
2068 | if (!none || val < none) |
2069 | ++val; |
2070 | } |
2071 | } |
711a919c |
2072 | else { /* bits == 1, then val should be ignored */ |
35da51f7 |
2073 | UV key; |
979f2922 |
2074 | if (min < start) |
2075 | min = start; |
2076 | for (key = min; key <= max; key++) { |
0bd48802 |
2077 | const STRLEN offset = (STRLEN)(key - start); |
979f2922 |
2078 | if (key >= end) |
2079 | goto go_out_list; |
2080 | s[offset >> 3] |= 1 << (offset & 7); |
2081 | } |
2082 | } |
2083 | } /* while */ |
2084 | go_out_list: |
2085 | |
2086 | /* read $swash->{EXTRAS} */ |
2087 | x = (U8*)SvPV(*extssvp, xcur); |
2088 | xend = x + xcur; |
2089 | while (x < xend) { |
2090 | STRLEN namelen; |
2091 | U8 *namestr; |
2092 | SV** othersvp; |
2093 | HV* otherhv; |
2094 | STRLEN otherbits; |
2095 | SV **otherbitssvp, *other; |
711a919c |
2096 | U8 *s, *o, *nl; |
979f2922 |
2097 | STRLEN slen, olen; |
2098 | |
35da51f7 |
2099 | const U8 opc = *x++; |
979f2922 |
2100 | if (opc == '\n') |
2101 | continue; |
2102 | |
2103 | nl = (U8*)memchr(x, '\n', xend - x); |
2104 | |
2105 | if (opc != '-' && opc != '+' && opc != '!' && opc != '&') { |
2106 | if (nl) { |
2107 | x = nl + 1; /* 1 is length of "\n" */ |
2108 | continue; |
2109 | } |
2110 | else { |
2111 | x = xend; /* to EXTRAS' end at which \n is not found */ |
2112 | break; |
2113 | } |
2114 | } |
2115 | |
2116 | namestr = x; |
2117 | if (nl) { |
2118 | namelen = nl - namestr; |
2119 | x = nl + 1; |
2120 | } |
2121 | else { |
2122 | namelen = xend - namestr; |
2123 | x = xend; |
2124 | } |
2125 | |
2126 | othersvp = hv_fetch(hv, (char *)namestr, namelen, FALSE); |
ef8f7699 |
2127 | otherhv = MUTABLE_HV(SvRV(*othersvp)); |
017a3ce5 |
2128 | otherbitssvp = hv_fetchs(otherhv, "BITS", FALSE); |
979f2922 |
2129 | otherbits = (STRLEN)SvUV(*otherbitssvp); |
2130 | if (bits < otherbits) |
660a4616 |
2131 | Perl_croak(aTHX_ "panic: swash_get found swatch size mismatch"); |
979f2922 |
2132 | |
2133 | /* The "other" swatch must be destroyed after. */ |
2134 | other = swash_get(*othersvp, start, span); |
2135 | o = (U8*)SvPV(other, olen); |
2136 | |
2137 | if (!olen) |
660a4616 |
2138 | Perl_croak(aTHX_ "panic: swash_get got improper swatch"); |
979f2922 |
2139 | |
2140 | s = (U8*)SvPV(swatch, slen); |
2141 | if (bits == 1 && otherbits == 1) { |
2142 | if (slen != olen) |
660a4616 |
2143 | Perl_croak(aTHX_ "panic: swash_get found swatch length mismatch"); |
979f2922 |
2144 | |
2145 | switch (opc) { |
2146 | case '+': |
2147 | while (slen--) |
2148 | *s++ |= *o++; |
2149 | break; |
2150 | case '!': |
2151 | while (slen--) |
2152 | *s++ |= ~*o++; |
2153 | break; |
2154 | case '-': |
2155 | while (slen--) |
2156 | *s++ &= ~*o++; |
2157 | break; |
2158 | case '&': |
2159 | while (slen--) |
2160 | *s++ &= *o++; |
2161 | break; |
2162 | default: |
2163 | break; |
2164 | } |
2165 | } |
711a919c |
2166 | else { |
979f2922 |
2167 | STRLEN otheroctets = otherbits >> 3; |
2168 | STRLEN offset = 0; |
35da51f7 |
2169 | U8* const send = s + slen; |
979f2922 |
2170 | |
2171 | while (s < send) { |
2172 | UV otherval = 0; |
2173 | |
2174 | if (otherbits == 1) { |
2175 | otherval = (o[offset >> 3] >> (offset & 7)) & 1; |
2176 | ++offset; |
2177 | } |
2178 | else { |
2179 | STRLEN vlen = otheroctets; |
2180 | otherval = *o++; |
2181 | while (--vlen) { |
2182 | otherval <<= 8; |
2183 | otherval |= *o++; |
2184 | } |
2185 | } |
2186 | |
711a919c |
2187 | if (opc == '+' && otherval) |
6f207bd3 |
2188 | NOOP; /* replace with otherval */ |
979f2922 |
2189 | else if (opc == '!' && !otherval) |
2190 | otherval = 1; |
2191 | else if (opc == '-' && otherval) |
2192 | otherval = 0; |
2193 | else if (opc == '&' && !otherval) |
2194 | otherval = 0; |
2195 | else { |
711a919c |
2196 | s += octets; /* no replacement */ |
979f2922 |
2197 | continue; |
2198 | } |
2199 | |
2200 | if (bits == 8) |
2201 | *s++ = (U8)( otherval & 0xff); |
2202 | else if (bits == 16) { |
2203 | *s++ = (U8)((otherval >> 8) & 0xff); |
2204 | *s++ = (U8)( otherval & 0xff); |
2205 | } |
2206 | else if (bits == 32) { |
2207 | *s++ = (U8)((otherval >> 24) & 0xff); |
2208 | *s++ = (U8)((otherval >> 16) & 0xff); |
2209 | *s++ = (U8)((otherval >> 8) & 0xff); |
2210 | *s++ = (U8)( otherval & 0xff); |
2211 | } |
2212 | } |
2213 | } |
2214 | sv_free(other); /* through with it! */ |
2215 | } /* while */ |
2216 | return swatch; |
2217 | } |
2218 | |
0f830e0b |
2219 | /* |
87cea99e |
2220 | =for apidoc uvchr_to_utf8 |
0f830e0b |
2221 | |
2222 | Adds the UTF-8 representation of the Native codepoint C<uv> to the end |
2223 | of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free |
2224 | bytes available. The return value is the pointer to the byte after the |
2225 | end of the new character. In other words, |
2226 | |
2227 | d = uvchr_to_utf8(d, uv); |
2228 | |
2229 | is the recommended wide native character-aware way of saying |
2230 | |
2231 | *(d++) = uv; |
2232 | |
2233 | =cut |
2234 | */ |
2235 | |
2236 | /* On ASCII machines this is normally a macro but we want a |
2237 | real function in case XS code wants it |
2238 | */ |
0f830e0b |
2239 | U8 * |
2240 | Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv) |
2241 | { |
7918f24d |
2242 | PERL_ARGS_ASSERT_UVCHR_TO_UTF8; |
2243 | |
0f830e0b |
2244 | return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0); |
2245 | } |
2246 | |
b851fbc1 |
2247 | U8 * |
2248 | Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) |
2249 | { |
7918f24d |
2250 | PERL_ARGS_ASSERT_UVCHR_TO_UTF8_FLAGS; |
2251 | |
b851fbc1 |
2252 | return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags); |
2253 | } |
2b9d42f0 |
2254 | |
2255 | /* |
87cea99e |
2256 | =for apidoc utf8n_to_uvchr |
0f830e0b |
2257 | flags |
2258 | |
2259 | Returns the native character value of the first character in the string |
2260 | C<s> |
2261 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
2262 | length, in bytes, of that character. |
2263 | |
2264 | Allows length and flags to be passed to low level routine. |
2265 | |
2266 | =cut |
2267 | */ |
2268 | /* On ASCII machines this is normally a macro but we want |
2269 | a real function in case XS code wants it |
2270 | */ |
0f830e0b |
2271 | UV |
2272 | Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, |
2273 | U32 flags) |
2274 | { |
2275 | const UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags); |
7918f24d |
2276 | |
2277 | PERL_ARGS_ASSERT_UTF8N_TO_UVCHR; |
2278 | |
0f830e0b |
2279 | return UNI_TO_NATIVE(uv); |
2280 | } |
2281 | |
2282 | /* |
87cea99e |
2283 | =for apidoc pv_uni_display |
d2cc3551 |
2284 | |
2285 | Build to the scalar dsv a displayable version of the string spv, |
2286 | length len, the displayable version being at most pvlim bytes long |
2287 | (if longer, the rest is truncated and "..." will be appended). |
0a2ef054 |
2288 | |
9e55ce06 |
2289 | The flags argument can have UNI_DISPLAY_ISPRINT set to display |
00e86452 |
2290 | isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH |
0a2ef054 |
2291 | to display the \\[nrfta\\] as the backslashed versions (like '\n') |
2292 | (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\). |
2293 | UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both |
2294 | UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on. |
2295 | |
d2cc3551 |
2296 | The pointer to the PV of the dsv is returned. |
2297 | |
2298 | =cut */ |
e6b2e755 |
2299 | char * |
e1ec3a88 |
2300 | Perl_pv_uni_display(pTHX_ SV *dsv, const U8 *spv, STRLEN len, STRLEN pvlim, UV flags) |
e6b2e755 |
2301 | { |
2302 | int truncated = 0; |
e1ec3a88 |
2303 | const char *s, *e; |
e6b2e755 |
2304 | |
7918f24d |
2305 | PERL_ARGS_ASSERT_PV_UNI_DISPLAY; |
2306 | |
76f68e9b |
2307 | sv_setpvs(dsv, ""); |
7fddd944 |
2308 | SvUTF8_off(dsv); |
e1ec3a88 |
2309 | for (s = (const char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) { |
e6b2e755 |
2310 | UV u; |
a49f32c6 |
2311 | /* This serves double duty as a flag and a character to print after |
2312 | a \ when flags & UNI_DISPLAY_BACKSLASH is true. |
2313 | */ |
2314 | char ok = 0; |
c728cb41 |
2315 | |
e6b2e755 |
2316 | if (pvlim && SvCUR(dsv) >= pvlim) { |
2317 | truncated++; |
2318 | break; |
2319 | } |
2320 | u = utf8_to_uvchr((U8*)s, 0); |
c728cb41 |
2321 | if (u < 256) { |
a3b680e6 |
2322 | const unsigned char c = (unsigned char)u & 0xFF; |
0bd48802 |
2323 | if (flags & UNI_DISPLAY_BACKSLASH) { |
a49f32c6 |
2324 | switch (c) { |
c728cb41 |
2325 | case '\n': |
a49f32c6 |
2326 | ok = 'n'; break; |
c728cb41 |
2327 | case '\r': |
a49f32c6 |
2328 | ok = 'r'; break; |
c728cb41 |
2329 | case '\t': |
a49f32c6 |
2330 | ok = 't'; break; |
c728cb41 |
2331 | case '\f': |
a49f32c6 |
2332 | ok = 'f'; break; |
c728cb41 |
2333 | case '\a': |
a49f32c6 |
2334 | ok = 'a'; break; |
c728cb41 |
2335 | case '\\': |
a49f32c6 |
2336 | ok = '\\'; break; |
c728cb41 |
2337 | default: break; |
2338 | } |
a49f32c6 |
2339 | if (ok) { |
88c9ea1e |
2340 | const char string = ok; |
76f68e9b |
2341 | sv_catpvs(dsv, "\\"); |
5e7aa789 |
2342 | sv_catpvn(dsv, &string, 1); |
a49f32c6 |
2343 | } |
c728cb41 |
2344 | } |
00e86452 |
2345 | /* isPRINT() is the locale-blind version. */ |
a49f32c6 |
2346 | if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(c)) { |
88c9ea1e |
2347 | const char string = c; |
5e7aa789 |
2348 | sv_catpvn(dsv, &string, 1); |
a49f32c6 |
2349 | ok = 1; |
0a2ef054 |
2350 | } |
c728cb41 |
2351 | } |
2352 | if (!ok) |
9e55ce06 |
2353 | Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u); |
e6b2e755 |
2354 | } |
2355 | if (truncated) |
396482e1 |
2356 | sv_catpvs(dsv, "..."); |
e6b2e755 |
2357 | |
2358 | return SvPVX(dsv); |
2359 | } |
2b9d42f0 |
2360 | |
d2cc3551 |
2361 | /* |
87cea99e |
2362 | =for apidoc sv_uni_display |
d2cc3551 |
2363 | |
2364 | Build to the scalar dsv a displayable version of the scalar sv, |
0a2ef054 |
2365 | the displayable version being at most pvlim bytes long |
d2cc3551 |
2366 | (if longer, the rest is truncated and "..." will be appended). |
0a2ef054 |
2367 | |
2368 | The flags argument is as in pv_uni_display(). |
2369 | |
d2cc3551 |
2370 | The pointer to the PV of the dsv is returned. |
2371 | |
d4c19fe8 |
2372 | =cut |
2373 | */ |
e6b2e755 |
2374 | char * |
2375 | Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags) |
2376 | { |
7918f24d |
2377 | PERL_ARGS_ASSERT_SV_UNI_DISPLAY; |
2378 | |
cfd0369c |
2379 | return Perl_pv_uni_display(aTHX_ dsv, (const U8*)SvPVX_const(ssv), |
2380 | SvCUR(ssv), pvlim, flags); |
701a277b |
2381 | } |
2382 | |
d2cc3551 |
2383 | /* |
87cea99e |
2384 | =for apidoc ibcmp_utf8 |
d2cc3551 |
2385 | |
2386 | Return true if the strings s1 and s2 differ case-insensitively, false |
2387 | if not (if they are equal case-insensitively). If u1 is true, the |
2388 | string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true, |
d07ddd77 |
2389 | the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2 |
2390 | are false, the respective string is assumed to be in native 8-bit |
2391 | encoding. |
2392 | |
2393 | If the pe1 and pe2 are non-NULL, the scanning pointers will be copied |
2394 | in there (they will point at the beginning of the I<next> character). |
2395 | If the pointers behind pe1 or pe2 are non-NULL, they are the end |
2396 | pointers beyond which scanning will not continue under any |
4cdaeff7 |
2397 | circumstances. If the byte lengths l1 and l2 are non-zero, s1+l1 and |
d07ddd77 |
2398 | s2+l2 will be used as goal end pointers that will also stop the scan, |
2399 | and which qualify towards defining a successful match: all the scans |
2400 | that define an explicit length must reach their goal pointers for |
2401 | a match to succeed). |
d2cc3551 |
2402 | |
2403 | For case-insensitiveness, the "casefolding" of Unicode is used |
2404 | instead of upper/lowercasing both the characters, see |
2405 | http://www.unicode.org/unicode/reports/tr21/ (Case Mappings). |
2406 | |
2407 | =cut */ |
701a277b |
2408 | I32 |
d07ddd77 |
2409 | 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 |
2410 | { |
97aff369 |
2411 | dVAR; |
e1ec3a88 |
2412 | register const U8 *p1 = (const U8*)s1; |
2413 | register const U8 *p2 = (const U8*)s2; |
cbbf8932 |
2414 | register const U8 *f1 = NULL; |
2f73348c |
2415 | register const U8 *f2 = NULL; |
cbbf8932 |
2416 | register U8 *e1 = NULL; |
2417 | register U8 *q1 = NULL; |
2418 | register U8 *e2 = NULL; |
2419 | register U8 *q2 = NULL; |
d07ddd77 |
2420 | STRLEN n1 = 0, n2 = 0; |
89ebb4a3 |
2421 | U8 foldbuf1[UTF8_MAXBYTES_CASE+1]; |
2422 | U8 foldbuf2[UTF8_MAXBYTES_CASE+1]; |
d7f013c8 |
2423 | U8 natbuf[1+1]; |
2424 | STRLEN foldlen1, foldlen2; |
d07ddd77 |
2425 | bool match; |
7918f24d |
2426 | |
2427 | PERL_ARGS_ASSERT_IBCMP_UTF8; |
332ddc25 |
2428 | |
d07ddd77 |
2429 | if (pe1) |
2430 | e1 = *(U8**)pe1; |
a0a388a1 |
2431 | /* assert(e1 || l1); */ |
e1ec3a88 |
2432 | if (e1 == 0 || (l1 && l1 < (UV)(e1 - (const U8*)s1))) |
2433 | f1 = (const U8*)s1 + l1; |
d07ddd77 |
2434 | if (pe2) |
2435 | e2 = *(U8**)pe2; |
a0a388a1 |
2436 | /* assert(e2 || l2); */ |
e1ec3a88 |
2437 | if (e2 == 0 || (l2 && l2 < (UV)(e2 - (const U8*)s2))) |
2438 | f2 = (const U8*)s2 + l2; |
d07ddd77 |
2439 | |
a0a388a1 |
2440 | /* This shouldn't happen. However, putting an assert() there makes some |
2441 | * tests fail. */ |
2442 | /* assert((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0)); */ |
d07ddd77 |
2443 | if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0)) |
2444 | return 1; /* mismatch; possible infinite loop or false positive */ |
2445 | |
a6872d42 |
2446 | if (!u1 || !u2) |
2447 | natbuf[1] = 0; /* Need to terminate the buffer. */ |
2448 | |
d07ddd77 |
2449 | while ((e1 == 0 || p1 < e1) && |
2450 | (f1 == 0 || p1 < f1) && |
2451 | (e2 == 0 || p2 < e2) && |
2452 | (f2 == 0 || p2 < f2)) { |
2453 | if (n1 == 0) { |
d7f013c8 |
2454 | if (u1) |
2455 | to_utf8_fold(p1, foldbuf1, &foldlen1); |
2456 | else { |
809e8e66 |
2457 | uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p1))); |
d7f013c8 |
2458 | to_utf8_fold(natbuf, foldbuf1, &foldlen1); |
2459 | } |
2460 | q1 = foldbuf1; |
d07ddd77 |
2461 | n1 = foldlen1; |
332ddc25 |
2462 | } |
d07ddd77 |
2463 | if (n2 == 0) { |
d7f013c8 |
2464 | if (u2) |
2465 | to_utf8_fold(p2, foldbuf2, &foldlen2); |
2466 | else { |
809e8e66 |
2467 | uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p2))); |
d7f013c8 |
2468 | to_utf8_fold(natbuf, foldbuf2, &foldlen2); |
2469 | } |
2470 | q2 = foldbuf2; |
d07ddd77 |
2471 | n2 = foldlen2; |
332ddc25 |
2472 | } |
d07ddd77 |
2473 | while (n1 && n2) { |
2474 | if ( UTF8SKIP(q1) != UTF8SKIP(q2) || |
2475 | (UTF8SKIP(q1) == 1 && *q1 != *q2) || |
2476 | memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) ) |
d7f013c8 |
2477 | return 1; /* mismatch */ |
d07ddd77 |
2478 | n1 -= UTF8SKIP(q1); |
d7f013c8 |
2479 | q1 += UTF8SKIP(q1); |
d07ddd77 |
2480 | n2 -= UTF8SKIP(q2); |
d7f013c8 |
2481 | q2 += UTF8SKIP(q2); |
701a277b |
2482 | } |
d07ddd77 |
2483 | if (n1 == 0) |
d7f013c8 |
2484 | p1 += u1 ? UTF8SKIP(p1) : 1; |
d07ddd77 |
2485 | if (n2 == 0) |
d7f013c8 |
2486 | p2 += u2 ? UTF8SKIP(p2) : 1; |
2487 | |
d2cc3551 |
2488 | } |
5469e704 |
2489 | |
d07ddd77 |
2490 | /* A match is defined by all the scans that specified |
2491 | * an explicit length reaching their final goals. */ |
2492 | match = (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2); |
5469e704 |
2493 | |
2494 | if (match) { |
d07ddd77 |
2495 | if (pe1) |
2496 | *pe1 = (char*)p1; |
2497 | if (pe2) |
2498 | *pe2 = (char*)p2; |
5469e704 |
2499 | } |
2500 | |
2501 | return match ? 0 : 1; /* 0 match, 1 mismatch */ |
e6b2e755 |
2502 | } |
701a277b |
2503 | |
a49f32c6 |
2504 | /* |
2505 | * Local variables: |
2506 | * c-indentation-style: bsd |
2507 | * c-basic-offset: 4 |
2508 | * indent-tabs-mode: t |
2509 | * End: |
2510 | * |
37442d52 |
2511 | * ex: set ts=8 sts=4 sw=4 noet: |
2512 | */ |