Commit | Line | Data |
---|---|---|
a0ed51b3 LW |
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 LW |
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 TC |
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 LW |
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 TC |
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 LW |
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 TC |
27 | * |
28 | * [p.512 of _The Lord of the Rings_, III/vi: "The King of the Golden Hall"] | |
a0ed51b3 LW |
29 | */ |
30 | ||
31 | #include "EXTERN.h" | |
864dbfa3 | 32 | #define PERL_IN_UTF8_C |
a0ed51b3 LW |
33 | #include "perl.h" |
34 | ||
a0c21aa1 | 35 | #ifndef EBCDIC |
970ea3cb | 36 | /* Separate prototypes needed because in ASCII systems these are |
a0c21aa1 JH |
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 JH |
42 | static const char unees[] = |
43 | "Malformed UTF-8 character (unexpected end of string)"; | |
901b21bf | 44 | |
48ef279e | 45 | /* |
ccfc67b7 | 46 | =head1 Unicode Support |
a0ed51b3 | 47 | |
166f8a29 DM |
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 DM |
51 | characters in the ASCII range are unmodified, and a zero byte never appears |
52 | within non-zero characters. | |
166f8a29 | 53 | |
eaf7a4d2 CS |
54 | =cut |
55 | */ | |
56 | ||
57 | /* | |
58 | =for apidoc is_ascii_string | |
59 | ||
970ea3cb KW |
60 | Returns true if the first C<len> bytes of the given string are the same whether |
61 | or not the string is encoded in UTF-8 (or UTF-EBCDIC on EBCDIC machines). That | |
62 | is, if they are invariant. On ASCII-ish machines, only ASCII characters | |
63 | fit this definition, hence the function's name. | |
eaf7a4d2 | 64 | |
9f7e3d64 MH |
65 | If C<len> is 0, it will be calculated using C<strlen(s)>. |
66 | ||
eaf7a4d2 CS |
67 | See also is_utf8_string(), is_utf8_string_loclen(), and is_utf8_string_loc(). |
68 | ||
69 | =cut | |
70 | */ | |
71 | ||
72 | bool | |
668b6d8d | 73 | Perl_is_ascii_string(const U8 *s, STRLEN len) |
eaf7a4d2 CS |
74 | { |
75 | const U8* const send = s + (len ? len : strlen((const char *)s)); | |
76 | const U8* x = s; | |
77 | ||
78 | PERL_ARGS_ASSERT_IS_ASCII_STRING; | |
eaf7a4d2 CS |
79 | |
80 | for (; x < send; ++x) { | |
81 | if (!UTF8_IS_INVARIANT(*x)) | |
82 | break; | |
83 | } | |
84 | ||
85 | return x == send; | |
86 | } | |
87 | ||
88 | /* | |
87cea99e | 89 | =for apidoc uvuni_to_utf8_flags |
eebe1485 | 90 | |
1e54db1a | 91 | Adds the UTF-8 representation of the Unicode codepoint C<uv> to the end |
89ebb4a3 | 92 | of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free |
eebe1485 | 93 | bytes available. The return value is the pointer to the byte after the |
9041c2e3 | 94 | end of the new character. In other words, |
eebe1485 | 95 | |
b851fbc1 JH |
96 | d = uvuni_to_utf8_flags(d, uv, flags); |
97 | ||
98 | or, in most cases, | |
99 | ||
9041c2e3 | 100 | d = uvuni_to_utf8(d, uv); |
eebe1485 | 101 | |
b851fbc1 JH |
102 | (which is equivalent to) |
103 | ||
104 | d = uvuni_to_utf8_flags(d, uv, 0); | |
105 | ||
eebe1485 SC |
106 | is the recommended Unicode-aware way of saying |
107 | ||
108 | *(d++) = uv; | |
109 | ||
110 | =cut | |
111 | */ | |
112 | ||
dfe13c55 | 113 | U8 * |
b851fbc1 | 114 | Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) |
a0ed51b3 | 115 | { |
7918f24d NC |
116 | PERL_ARGS_ASSERT_UVUNI_TO_UTF8_FLAGS; |
117 | ||
62961d2e | 118 | if (ckWARN(WARN_UTF8)) { |
b851fbc1 JH |
119 | if (UNICODE_IS_SURROGATE(uv) && |
120 | !(flags & UNICODE_ALLOW_SURROGATE)) | |
9014280d | 121 | Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv); |
b851fbc1 JH |
122 | else if ( |
123 | ((uv >= 0xFDD0 && uv <= 0xFDEF && | |
124 | !(flags & UNICODE_ALLOW_FDD0)) | |
125 | || | |
c867b360 | 126 | ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */ |
b851fbc1 JH |
127 | !(flags & UNICODE_ALLOW_FFFF))) && |
128 | /* UNICODE_ALLOW_SUPER includes | |
2a20b9da | 129 | * FFFEs and FFFFs beyond 0x10FFFF. */ |
b851fbc1 JH |
130 | ((uv <= PERL_UNICODE_MAX) || |
131 | !(flags & UNICODE_ALLOW_SUPER)) | |
132 | ) | |
9014280d | 133 | Perl_warner(aTHX_ packWARN(WARN_UTF8), |
6f6ac1de | 134 | "Unicode non-character 0x%04"UVxf" is illegal for interchange", uv); |
507b9800 | 135 | } |
c4d5f83a | 136 | if (UNI_IS_INVARIANT(uv)) { |
eb160463 | 137 | *d++ = (U8)UTF_TO_NATIVE(uv); |
a0ed51b3 LW |
138 | return d; |
139 | } | |
2d331972 | 140 | #if defined(EBCDIC) |
1d72bdf6 NIS |
141 | else { |
142 | STRLEN len = UNISKIP(uv); | |
143 | U8 *p = d+len-1; | |
144 | while (p > d) { | |
eb160463 | 145 | *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK); |
1d72bdf6 NIS |
146 | uv >>= UTF_ACCUMULATION_SHIFT; |
147 | } | |
eb160463 | 148 | *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len)); |
1d72bdf6 NIS |
149 | return d+len; |
150 | } | |
151 | #else /* Non loop style */ | |
a0ed51b3 | 152 | if (uv < 0x800) { |
eb160463 GS |
153 | *d++ = (U8)(( uv >> 6) | 0xc0); |
154 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
155 | return d; |
156 | } | |
157 | if (uv < 0x10000) { | |
eb160463 GS |
158 | *d++ = (U8)(( uv >> 12) | 0xe0); |
159 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
160 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
161 | return d; |
162 | } | |
163 | if (uv < 0x200000) { | |
eb160463 GS |
164 | *d++ = (U8)(( uv >> 18) | 0xf0); |
165 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
166 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
167 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
168 | return d; |
169 | } | |
170 | if (uv < 0x4000000) { | |
eb160463 GS |
171 | *d++ = (U8)(( uv >> 24) | 0xf8); |
172 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); | |
173 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
174 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
175 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
176 | return d; |
177 | } | |
178 | if (uv < 0x80000000) { | |
eb160463 GS |
179 | *d++ = (U8)(( uv >> 30) | 0xfc); |
180 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); | |
181 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); | |
182 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
183 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
184 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
185 | return d; |
186 | } | |
6b8eaf93 | 187 | #ifdef HAS_QUAD |
d7578b48 | 188 | if (uv < UTF8_QUAD_MAX) |
a0ed51b3 LW |
189 | #endif |
190 | { | |
eb160463 GS |
191 | *d++ = 0xfe; /* Can't match U+FEFF! */ |
192 | *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); | |
193 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); | |
194 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); | |
195 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
196 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
197 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
198 | return d; |
199 | } | |
6b8eaf93 | 200 | #ifdef HAS_QUAD |
a0ed51b3 | 201 | { |
eb160463 GS |
202 | *d++ = 0xff; /* Can't match U+FFFE! */ |
203 | *d++ = 0x80; /* 6 Reserved bits */ | |
204 | *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */ | |
205 | *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80); | |
206 | *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80); | |
207 | *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80); | |
208 | *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80); | |
209 | *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); | |
210 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); | |
211 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); | |
212 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
213 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
214 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
215 | return d; |
216 | } | |
217 | #endif | |
1d72bdf6 | 218 | #endif /* Loop style */ |
a0ed51b3 | 219 | } |
9041c2e3 | 220 | |
646ca15d JH |
221 | /* |
222 | ||
223 | Tests if some arbitrary number of bytes begins in a valid UTF-8 | |
224 | character. Note that an INVARIANT (i.e. ASCII) character is a valid | |
225 | UTF-8 character. The actual number of bytes in the UTF-8 character | |
226 | will be returned if it is valid, otherwise 0. | |
227 | ||
228 | This is the "slow" version as opposed to the "fast" version which is | |
229 | the "unrolled" IS_UTF8_CHAR(). E.g. for t/uni/class.t the speed | |
230 | difference is a factor of 2 to 3. For lengths (UTF8SKIP(s)) of four | |
231 | or less you should use the IS_UTF8_CHAR(), for lengths of five or more | |
232 | you should use the _slow(). In practice this means that the _slow() | |
233 | will be used very rarely, since the maximum Unicode code point (as of | |
234 | Unicode 4.1) is U+10FFFF, which encodes in UTF-8 to four bytes. Only | |
235 | the "Perl extended UTF-8" (the infamous 'v-strings') will encode into | |
236 | five bytes or more. | |
237 | ||
238 | =cut */ | |
c053b435 | 239 | STATIC STRLEN |
5f66b61c | 240 | S_is_utf8_char_slow(const U8 *s, const STRLEN len) |
646ca15d JH |
241 | { |
242 | U8 u = *s; | |
243 | STRLEN slen; | |
244 | UV uv, ouv; | |
245 | ||
7918f24d NC |
246 | PERL_ARGS_ASSERT_IS_UTF8_CHAR_SLOW; |
247 | ||
646ca15d JH |
248 | if (UTF8_IS_INVARIANT(u)) |
249 | return 1; | |
250 | ||
251 | if (!UTF8_IS_START(u)) | |
252 | return 0; | |
253 | ||
254 | if (len < 2 || !UTF8_IS_CONTINUATION(s[1])) | |
255 | return 0; | |
256 | ||
257 | slen = len - 1; | |
258 | s++; | |
77263263 TS |
259 | #ifdef EBCDIC |
260 | u = NATIVE_TO_UTF(u); | |
261 | #endif | |
646ca15d JH |
262 | u &= UTF_START_MASK(len); |
263 | uv = u; | |
264 | ouv = uv; | |
265 | while (slen--) { | |
266 | if (!UTF8_IS_CONTINUATION(*s)) | |
267 | return 0; | |
268 | uv = UTF8_ACCUMULATE(uv, *s); | |
48ef279e | 269 | if (uv < ouv) |
646ca15d JH |
270 | return 0; |
271 | ouv = uv; | |
272 | s++; | |
273 | } | |
274 | ||
275 | if ((STRLEN)UNISKIP(uv) < len) | |
276 | return 0; | |
277 | ||
278 | return len; | |
279 | } | |
9041c2e3 NIS |
280 | |
281 | /* | |
87cea99e | 282 | =for apidoc is_utf8_char |
eebe1485 | 283 | |
5da9da9e | 284 | Tests if some arbitrary number of bytes begins in a valid UTF-8 |
2bbc8d55 SP |
285 | character. Note that an INVARIANT (i.e. ASCII on non-EBCDIC machines) |
286 | character is a valid UTF-8 character. The actual number of bytes in the UTF-8 | |
287 | character will be returned if it is valid, otherwise 0. | |
9041c2e3 | 288 | |
82686b01 | 289 | =cut */ |
067a85ef | 290 | STRLEN |
668b6d8d | 291 | Perl_is_utf8_char(const U8 *s) |
386d01d6 | 292 | { |
44f8325f | 293 | const STRLEN len = UTF8SKIP(s); |
7918f24d NC |
294 | |
295 | PERL_ARGS_ASSERT_IS_UTF8_CHAR; | |
3b0fc154 | 296 | #ifdef IS_UTF8_CHAR |
768c67ee | 297 | if (IS_UTF8_CHAR_FAST(len)) |
3b0fc154 JH |
298 | return IS_UTF8_CHAR(s, len) ? len : 0; |
299 | #endif /* #ifdef IS_UTF8_CHAR */ | |
2c0c5f92 | 300 | return is_utf8_char_slow(s, len); |
386d01d6 GS |
301 | } |
302 | ||
eaf7a4d2 | 303 | |
6662521e | 304 | /* |
87cea99e | 305 | =for apidoc is_utf8_string |
6662521e | 306 | |
c9ada85f | 307 | Returns true if first C<len> bytes of the given string form a valid |
9f7e3d64 MH |
308 | UTF-8 string, false otherwise. If C<len> is 0, it will be calculated |
309 | using C<strlen(s)>. Note that 'a valid UTF-8 string' does not mean 'a | |
310 | string that contains code points above 0x7F encoded in UTF-8' because a | |
311 | valid ASCII string is a valid UTF-8 string. | |
6662521e | 312 | |
eaf7a4d2 | 313 | See also is_ascii_string(), is_utf8_string_loclen(), and is_utf8_string_loc(). |
768c67ee | 314 | |
6662521e GS |
315 | =cut |
316 | */ | |
317 | ||
8e84507e | 318 | bool |
668b6d8d | 319 | Perl_is_utf8_string(const U8 *s, STRLEN len) |
6662521e | 320 | { |
35da51f7 | 321 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
7fc63493 | 322 | const U8* x = s; |
067a85ef | 323 | |
7918f24d | 324 | PERL_ARGS_ASSERT_IS_UTF8_STRING; |
1aa99e6b | 325 | |
6662521e | 326 | while (x < send) { |
a3b680e6 | 327 | STRLEN c; |
1acdb0da JH |
328 | /* Inline the easy bits of is_utf8_char() here for speed... */ |
329 | if (UTF8_IS_INVARIANT(*x)) | |
330 | c = 1; | |
331 | else if (!UTF8_IS_START(*x)) | |
768c67ee | 332 | goto out; |
1acdb0da JH |
333 | else { |
334 | /* ... and call is_utf8_char() only if really needed. */ | |
646ca15d JH |
335 | #ifdef IS_UTF8_CHAR |
336 | c = UTF8SKIP(x); | |
768c67ee JH |
337 | if (IS_UTF8_CHAR_FAST(c)) { |
338 | if (!IS_UTF8_CHAR(x, c)) | |
3c614e38 AD |
339 | c = 0; |
340 | } | |
341 | else | |
342 | c = is_utf8_char_slow(x, c); | |
646ca15d JH |
343 | #else |
344 | c = is_utf8_char(x); | |
345 | #endif /* #ifdef IS_UTF8_CHAR */ | |
1acdb0da | 346 | if (!c) |
768c67ee | 347 | goto out; |
1acdb0da | 348 | } |
6662521e | 349 | x += c; |
6662521e | 350 | } |
768c67ee JH |
351 | |
352 | out: | |
60006e79 JH |
353 | if (x != send) |
354 | return FALSE; | |
067a85ef A |
355 | |
356 | return TRUE; | |
6662521e GS |
357 | } |
358 | ||
67e989fb | 359 | /* |
814fafa7 NC |
360 | Implemented as a macro in utf8.h |
361 | ||
87cea99e | 362 | =for apidoc is_utf8_string_loc |
814fafa7 NC |
363 | |
364 | Like is_utf8_string() but stores the location of the failure (in the | |
365 | case of "utf8ness failure") or the location s+len (in the case of | |
366 | "utf8ness success") in the C<ep>. | |
367 | ||
368 | See also is_utf8_string_loclen() and is_utf8_string(). | |
369 | ||
87cea99e | 370 | =for apidoc is_utf8_string_loclen |
81cd54e3 | 371 | |
e3e4599f | 372 | Like is_utf8_string() but stores the location of the failure (in the |
768c67ee JH |
373 | case of "utf8ness failure") or the location s+len (in the case of |
374 | "utf8ness success") in the C<ep>, and the number of UTF-8 | |
375 | encoded characters in the C<el>. | |
376 | ||
377 | See also is_utf8_string_loc() and is_utf8_string(). | |
81cd54e3 JH |
378 | |
379 | =cut | |
380 | */ | |
381 | ||
382 | bool | |
668b6d8d | 383 | Perl_is_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el) |
81cd54e3 | 384 | { |
35da51f7 | 385 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
7fc63493 | 386 | const U8* x = s; |
81cd54e3 | 387 | STRLEN c; |
3ebfea28 | 388 | STRLEN outlen = 0; |
7918f24d NC |
389 | |
390 | PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN; | |
81cd54e3 | 391 | |
81cd54e3 JH |
392 | while (x < send) { |
393 | /* Inline the easy bits of is_utf8_char() here for speed... */ | |
394 | if (UTF8_IS_INVARIANT(*x)) | |
768c67ee JH |
395 | c = 1; |
396 | else if (!UTF8_IS_START(*x)) | |
397 | goto out; | |
81cd54e3 | 398 | else { |
768c67ee JH |
399 | /* ... and call is_utf8_char() only if really needed. */ |
400 | #ifdef IS_UTF8_CHAR | |
401 | c = UTF8SKIP(x); | |
402 | if (IS_UTF8_CHAR_FAST(c)) { | |
403 | if (!IS_UTF8_CHAR(x, c)) | |
404 | c = 0; | |
405 | } else | |
406 | c = is_utf8_char_slow(x, c); | |
407 | #else | |
408 | c = is_utf8_char(x); | |
409 | #endif /* #ifdef IS_UTF8_CHAR */ | |
410 | if (!c) | |
411 | goto out; | |
81cd54e3 | 412 | } |
768c67ee | 413 | x += c; |
3ebfea28 | 414 | outlen++; |
81cd54e3 | 415 | } |
768c67ee JH |
416 | |
417 | out: | |
3ebfea28 AL |
418 | if (el) |
419 | *el = outlen; | |
420 | ||
768c67ee JH |
421 | if (ep) |
422 | *ep = x; | |
3ebfea28 | 423 | return (x == send); |
81cd54e3 JH |
424 | } |
425 | ||
426 | /* | |
768c67ee | 427 | |
87cea99e | 428 | =for apidoc utf8n_to_uvuni |
67e989fb | 429 | |
9041c2e3 | 430 | Bottom level UTF-8 decode routine. |
38a44b82 | 431 | Returns the Unicode code point value of the first character in the string C<s> |
1e54db1a | 432 | which is assumed to be in UTF-8 encoding and no longer than C<curlen>; |
7df053ec | 433 | C<retlen> will be set to the length, in bytes, of that character. |
67e989fb | 434 | |
1e54db1a | 435 | If C<s> does not point to a well-formed UTF-8 character, the behaviour |
dcad2880 JH |
436 | is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY, |
437 | it is assumed that the caller will raise a warning, and this function | |
28d3d195 JH |
438 | will silently just set C<retlen> to C<-1> and return zero. If the |
439 | C<flags> does not contain UTF8_CHECK_ONLY, warnings about | |
440 | malformations will be given, C<retlen> will be set to the expected | |
441 | length of the UTF-8 character in bytes, and zero will be returned. | |
442 | ||
443 | The C<flags> can also contain various flags to allow deviations from | |
444 | the strict UTF-8 encoding (see F<utf8.h>). | |
67e989fb | 445 | |
9041c2e3 NIS |
446 | Most code should use utf8_to_uvchr() rather than call this directly. |
447 | ||
37607a96 PK |
448 | =cut |
449 | */ | |
67e989fb | 450 | |
a0ed51b3 | 451 | UV |
7fc63493 | 452 | Perl_utf8n_to_uvuni(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) |
a0ed51b3 | 453 | { |
97aff369 | 454 | dVAR; |
d4c19fe8 | 455 | const U8 * const s0 = s; |
9c5ffd7c | 456 | UV uv = *s, ouv = 0; |
ba210ebe | 457 | STRLEN len = 1; |
7fc63493 AL |
458 | const bool dowarn = ckWARN_d(WARN_UTF8); |
459 | const UV startbyte = *s; | |
ba210ebe | 460 | STRLEN expectlen = 0; |
a0dbb045 | 461 | U32 warning = 0; |
5b311467 | 462 | SV* sv; |
a0dbb045 | 463 | |
7918f24d NC |
464 | PERL_ARGS_ASSERT_UTF8N_TO_UVUNI; |
465 | ||
5b311467 | 466 | /* This list is a superset of the UTF8_ALLOW_XXX. BUT it isn't, eg SUPER missing XXX */ |
a0dbb045 JH |
467 | |
468 | #define UTF8_WARN_EMPTY 1 | |
469 | #define UTF8_WARN_CONTINUATION 2 | |
470 | #define UTF8_WARN_NON_CONTINUATION 3 | |
471 | #define UTF8_WARN_FE_FF 4 | |
472 | #define UTF8_WARN_SHORT 5 | |
473 | #define UTF8_WARN_OVERFLOW 6 | |
474 | #define UTF8_WARN_SURROGATE 7 | |
c867b360 JH |
475 | #define UTF8_WARN_LONG 8 |
476 | #define UTF8_WARN_FFFF 9 /* Also FFFE. */ | |
a0dbb045 JH |
477 | |
478 | if (curlen == 0 && | |
479 | !(flags & UTF8_ALLOW_EMPTY)) { | |
480 | warning = UTF8_WARN_EMPTY; | |
0c443dc2 JH |
481 | goto malformed; |
482 | } | |
483 | ||
1d72bdf6 | 484 | if (UTF8_IS_INVARIANT(uv)) { |
a0ed51b3 LW |
485 | if (retlen) |
486 | *retlen = 1; | |
c4d5f83a | 487 | return (UV) (NATIVE_TO_UTF(*s)); |
a0ed51b3 | 488 | } |
67e989fb | 489 | |
421a8bf2 | 490 | if (UTF8_IS_CONTINUATION(uv) && |
fcc8fcf6 | 491 | !(flags & UTF8_ALLOW_CONTINUATION)) { |
a0dbb045 | 492 | warning = UTF8_WARN_CONTINUATION; |
ba210ebe JH |
493 | goto malformed; |
494 | } | |
495 | ||
421a8bf2 | 496 | if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) && |
fcc8fcf6 | 497 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { |
a0dbb045 | 498 | warning = UTF8_WARN_NON_CONTINUATION; |
ba210ebe JH |
499 | goto malformed; |
500 | } | |
9041c2e3 | 501 | |
1d72bdf6 | 502 | #ifdef EBCDIC |
75383841 | 503 | uv = NATIVE_TO_UTF(uv); |
1d72bdf6 | 504 | #else |
fcc8fcf6 JH |
505 | if ((uv == 0xfe || uv == 0xff) && |
506 | !(flags & UTF8_ALLOW_FE_FF)) { | |
a0dbb045 | 507 | warning = UTF8_WARN_FE_FF; |
ba210ebe | 508 | goto malformed; |
a0ed51b3 | 509 | } |
1d72bdf6 NIS |
510 | #endif |
511 | ||
ba210ebe JH |
512 | if (!(uv & 0x20)) { len = 2; uv &= 0x1f; } |
513 | else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; } | |
514 | else if (!(uv & 0x08)) { len = 4; uv &= 0x07; } | |
515 | else if (!(uv & 0x04)) { len = 5; uv &= 0x03; } | |
1d72bdf6 NIS |
516 | #ifdef EBCDIC |
517 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } | |
518 | else { len = 7; uv &= 0x01; } | |
519 | #else | |
ba210ebe JH |
520 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } |
521 | else if (!(uv & 0x01)) { len = 7; uv = 0; } | |
1d72bdf6 NIS |
522 | else { len = 13; uv = 0; } /* whoa! */ |
523 | #endif | |
524 | ||
a0ed51b3 LW |
525 | if (retlen) |
526 | *retlen = len; | |
9041c2e3 | 527 | |
ba210ebe JH |
528 | expectlen = len; |
529 | ||
fcc8fcf6 JH |
530 | if ((curlen < expectlen) && |
531 | !(flags & UTF8_ALLOW_SHORT)) { | |
a0dbb045 | 532 | warning = UTF8_WARN_SHORT; |
ba210ebe JH |
533 | goto malformed; |
534 | } | |
535 | ||
536 | len--; | |
a0ed51b3 | 537 | s++; |
ba210ebe JH |
538 | ouv = uv; |
539 | ||
a0ed51b3 | 540 | while (len--) { |
421a8bf2 JH |
541 | if (!UTF8_IS_CONTINUATION(*s) && |
542 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { | |
a0dbb045 JH |
543 | s--; |
544 | warning = UTF8_WARN_NON_CONTINUATION; | |
ba210ebe | 545 | goto malformed; |
a0ed51b3 LW |
546 | } |
547 | else | |
8850bf83 | 548 | uv = UTF8_ACCUMULATE(uv, *s); |
a0dbb045 JH |
549 | if (!(uv > ouv)) { |
550 | /* These cannot be allowed. */ | |
551 | if (uv == ouv) { | |
75dbc644 | 552 | if (expectlen != 13 && !(flags & UTF8_ALLOW_LONG)) { |
a0dbb045 JH |
553 | warning = UTF8_WARN_LONG; |
554 | goto malformed; | |
555 | } | |
556 | } | |
557 | else { /* uv < ouv */ | |
558 | /* This cannot be allowed. */ | |
559 | warning = UTF8_WARN_OVERFLOW; | |
560 | goto malformed; | |
561 | } | |
ba210ebe JH |
562 | } |
563 | s++; | |
564 | ouv = uv; | |
565 | } | |
566 | ||
421a8bf2 | 567 | if (UNICODE_IS_SURROGATE(uv) && |
fcc8fcf6 | 568 | !(flags & UTF8_ALLOW_SURROGATE)) { |
a0dbb045 | 569 | warning = UTF8_WARN_SURROGATE; |
ba210ebe | 570 | goto malformed; |
eb160463 | 571 | } else if ((expectlen > (STRLEN)UNISKIP(uv)) && |
fcc8fcf6 | 572 | !(flags & UTF8_ALLOW_LONG)) { |
a0dbb045 | 573 | warning = UTF8_WARN_LONG; |
ba210ebe | 574 | goto malformed; |
421a8bf2 | 575 | } else if (UNICODE_IS_ILLEGAL(uv) && |
a9917092 | 576 | !(flags & UTF8_ALLOW_FFFF)) { |
a0dbb045 | 577 | warning = UTF8_WARN_FFFF; |
a9917092 | 578 | goto malformed; |
a0ed51b3 | 579 | } |
ba210ebe | 580 | |
a0ed51b3 | 581 | return uv; |
ba210ebe JH |
582 | |
583 | malformed: | |
584 | ||
fcc8fcf6 | 585 | if (flags & UTF8_CHECK_ONLY) { |
ba210ebe | 586 | if (retlen) |
10edeb5d | 587 | *retlen = ((STRLEN) -1); |
ba210ebe JH |
588 | return 0; |
589 | } | |
590 | ||
a0dbb045 | 591 | if (dowarn) { |
5b311467 KW |
592 | if (warning == UTF8_WARN_FFFF) { |
593 | sv = newSVpvs_flags("Unicode non-character ", SVs_TEMP); | |
594 | Perl_sv_catpvf(aTHX_ sv, "0x%04"UVxf" is illegal for interchange", uv); | |
595 | } | |
596 | else { | |
597 | sv = newSVpvs_flags("Malformed UTF-8 character ", SVs_TEMP); | |
598 | ||
599 | switch (warning) { | |
600 | case 0: /* Intentionally empty. */ break; | |
601 | case UTF8_WARN_EMPTY: | |
602 | sv_catpvs(sv, "(empty string)"); | |
603 | break; | |
604 | case UTF8_WARN_CONTINUATION: | |
605 | Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv); | |
606 | break; | |
607 | case UTF8_WARN_NON_CONTINUATION: | |
608 | if (s == s0) | |
609 | Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")", | |
610 | (UV)s[1], startbyte); | |
611 | else { | |
612 | const int len = (int)(s-s0); | |
613 | Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)", | |
614 | (UV)s[1], len, len > 1 ? "s" : "", startbyte, (int)expectlen); | |
615 | } | |
616 | ||
617 | break; | |
618 | case UTF8_WARN_FE_FF: | |
619 | Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv); | |
620 | break; | |
621 | case UTF8_WARN_SHORT: | |
622 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", | |
623 | (int)curlen, curlen == 1 ? "" : "s", (int)expectlen, startbyte); | |
624 | expectlen = curlen; /* distance for caller to skip */ | |
625 | break; | |
626 | case UTF8_WARN_OVERFLOW: | |
627 | Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")", | |
628 | ouv, *s, startbyte); | |
629 | break; | |
630 | case UTF8_WARN_SURROGATE: | |
631 | Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv); | |
632 | break; | |
633 | case UTF8_WARN_LONG: | |
634 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", | |
635 | (int)expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte); | |
636 | break; | |
637 | default: | |
638 | sv_catpvs(sv, "(unknown reason)"); | |
639 | break; | |
551405c4 | 640 | } |
a0dbb045 JH |
641 | } |
642 | ||
643 | if (warning) { | |
44f8325f | 644 | const char * const s = SvPVX_const(sv); |
a0dbb045 JH |
645 | |
646 | if (PL_op) | |
9014280d | 647 | Perl_warner(aTHX_ packWARN(WARN_UTF8), |
53e06cf0 | 648 | "%s in %s", s, OP_DESC(PL_op)); |
a0dbb045 | 649 | else |
9014280d | 650 | Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s); |
a0dbb045 JH |
651 | } |
652 | } | |
653 | ||
ba210ebe | 654 | if (retlen) |
28d3d195 | 655 | *retlen = expectlen ? expectlen : len; |
ba210ebe | 656 | |
28d3d195 | 657 | return 0; |
a0ed51b3 LW |
658 | } |
659 | ||
8e84507e | 660 | /* |
87cea99e | 661 | =for apidoc utf8_to_uvchr |
9041c2e3 NIS |
662 | |
663 | Returns the native character value of the first character in the string C<s> | |
1e54db1a | 664 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
9041c2e3 NIS |
665 | length, in bytes, of that character. |
666 | ||
1e54db1a | 667 | If C<s> does not point to a well-formed UTF-8 character, zero is |
9041c2e3 NIS |
668 | returned and retlen is set, if possible, to -1. |
669 | ||
670 | =cut | |
671 | */ | |
672 | ||
673 | UV | |
7fc63493 | 674 | Perl_utf8_to_uvchr(pTHX_ const U8 *s, STRLEN *retlen) |
9041c2e3 | 675 | { |
7918f24d NC |
676 | PERL_ARGS_ASSERT_UTF8_TO_UVCHR; |
677 | ||
1754c1a1 NC |
678 | return utf8n_to_uvchr(s, UTF8_MAXBYTES, retlen, |
679 | ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); | |
9041c2e3 NIS |
680 | } |
681 | ||
682 | /* | |
87cea99e | 683 | =for apidoc utf8_to_uvuni |
9041c2e3 NIS |
684 | |
685 | Returns the Unicode code point of the first character in the string C<s> | |
1e54db1a | 686 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
9041c2e3 NIS |
687 | length, in bytes, of that character. |
688 | ||
2bbc8d55 | 689 | This function should only be used when the returned UV is considered |
9041c2e3 NIS |
690 | an index into the Unicode semantic tables (e.g. swashes). |
691 | ||
1e54db1a | 692 | If C<s> does not point to a well-formed UTF-8 character, zero is |
ba210ebe | 693 | returned and retlen is set, if possible, to -1. |
8e84507e NIS |
694 | |
695 | =cut | |
696 | */ | |
697 | ||
698 | UV | |
7fc63493 | 699 | Perl_utf8_to_uvuni(pTHX_ const U8 *s, STRLEN *retlen) |
8e84507e | 700 | { |
7918f24d NC |
701 | PERL_ARGS_ASSERT_UTF8_TO_UVUNI; |
702 | ||
9041c2e3 | 703 | /* Call the low level routine asking for checks */ |
89ebb4a3 | 704 | return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXBYTES, retlen, |
872c91ae | 705 | ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); |
8e84507e NIS |
706 | } |
707 | ||
b76347f2 | 708 | /* |
87cea99e | 709 | =for apidoc utf8_length |
b76347f2 JH |
710 | |
711 | Return the length of the UTF-8 char encoded string C<s> in characters. | |
02eb7b47 JH |
712 | Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end |
713 | up past C<e>, croaks. | |
b76347f2 JH |
714 | |
715 | =cut | |
716 | */ | |
717 | ||
718 | STRLEN | |
35a4481c | 719 | Perl_utf8_length(pTHX_ const U8 *s, const U8 *e) |
b76347f2 | 720 | { |
97aff369 | 721 | dVAR; |
b76347f2 JH |
722 | STRLEN len = 0; |
723 | ||
7918f24d NC |
724 | PERL_ARGS_ASSERT_UTF8_LENGTH; |
725 | ||
8850bf83 JH |
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 | ||
a3b680e6 AL |
730 | if (e < s) |
731 | goto warn_and_return; | |
b76347f2 | 732 | while (s < e) { |
8e91ec7f AV |
733 | if (!UTF8_IS_INVARIANT(*s)) |
734 | s += UTF8SKIP(s); | |
735 | else | |
736 | s++; | |
737 | len++; | |
738 | } | |
739 | ||
740 | if (e != s) { | |
741 | len--; | |
742 | warn_and_return: | |
9b387841 NC |
743 | if (PL_op) |
744 | Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), | |
745 | "%s in %s", unees, OP_DESC(PL_op)); | |
746 | else | |
747 | Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), unees); | |
b76347f2 JH |
748 | } |
749 | ||
750 | return len; | |
751 | } | |
752 | ||
b06226ff | 753 | /* |
87cea99e | 754 | =for apidoc utf8_distance |
b06226ff | 755 | |
1e54db1a | 756 | Returns the number of UTF-8 characters between the UTF-8 pointers C<a> |
b06226ff JH |
757 | and C<b>. |
758 | ||
759 | WARNING: use only if you *know* that the pointers point inside the | |
760 | same UTF-8 buffer. | |
761 | ||
37607a96 PK |
762 | =cut |
763 | */ | |
a0ed51b3 | 764 | |
02eb7b47 | 765 | IV |
35a4481c | 766 | Perl_utf8_distance(pTHX_ const U8 *a, const U8 *b) |
a0ed51b3 | 767 | { |
7918f24d NC |
768 | PERL_ARGS_ASSERT_UTF8_DISTANCE; |
769 | ||
bf1665bc | 770 | return (a < b) ? -1 * (IV) utf8_length(a, b) : (IV) utf8_length(b, a); |
a0ed51b3 LW |
771 | } |
772 | ||
b06226ff | 773 | /* |
87cea99e | 774 | =for apidoc utf8_hop |
b06226ff | 775 | |
8850bf83 JH |
776 | Return the UTF-8 pointer C<s> displaced by C<off> characters, either |
777 | forward or backward. | |
b06226ff JH |
778 | |
779 | WARNING: do not use the following unless you *know* C<off> is within | |
8850bf83 JH |
780 | the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned |
781 | on the first byte of character or just after the last byte of a character. | |
b06226ff | 782 | |
37607a96 PK |
783 | =cut |
784 | */ | |
a0ed51b3 LW |
785 | |
786 | U8 * | |
4373e329 | 787 | Perl_utf8_hop(pTHX_ const U8 *s, I32 off) |
a0ed51b3 | 788 | { |
7918f24d NC |
789 | PERL_ARGS_ASSERT_UTF8_HOP; |
790 | ||
96a5add6 | 791 | PERL_UNUSED_CONTEXT; |
8850bf83 JH |
792 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g |
793 | * the bitops (especially ~) can create illegal UTF-8. | |
794 | * In other words: in Perl UTF-8 is not just for Unicode. */ | |
795 | ||
a0ed51b3 LW |
796 | if (off >= 0) { |
797 | while (off--) | |
798 | s += UTF8SKIP(s); | |
799 | } | |
800 | else { | |
801 | while (off++) { | |
802 | s--; | |
8850bf83 JH |
803 | while (UTF8_IS_CONTINUATION(*s)) |
804 | s--; | |
a0ed51b3 LW |
805 | } |
806 | } | |
4373e329 | 807 | return (U8 *)s; |
a0ed51b3 LW |
808 | } |
809 | ||
6940069f | 810 | /* |
fed3ba5d NC |
811 | =for apidoc bytes_cmp_utf8 |
812 | ||
813 | Compares the sequence of characters (stored as octets) in b, blen with the | |
814 | sequence of characters (stored as UTF-8) in u, ulen. Returns 0 if they are | |
815 | equal, -1 or -2 if the first string is less than the second string, +1 or +2 | |
816 | if the first string is greater than the second string. | |
817 | ||
818 | -1 or +1 is returned if the shorter string was identical to the start of the | |
819 | longer string. -2 or +2 is returned if the was a difference between characters | |
820 | within the strings. | |
821 | ||
822 | =cut | |
823 | */ | |
824 | ||
825 | int | |
826 | Perl_bytes_cmp_utf8(pTHX_ const U8 *b, STRLEN blen, const U8 *u, STRLEN ulen) | |
827 | { | |
828 | const U8 *const bend = b + blen; | |
829 | const U8 *const uend = u + ulen; | |
830 | ||
831 | PERL_ARGS_ASSERT_BYTES_CMP_UTF8; | |
832 | ||
833 | PERL_UNUSED_CONTEXT; | |
834 | ||
835 | while (b < bend && u < uend) { | |
836 | U8 c = *u++; | |
837 | if (!UTF8_IS_INVARIANT(c)) { | |
838 | if (UTF8_IS_DOWNGRADEABLE_START(c)) { | |
839 | if (u < uend) { | |
840 | U8 c1 = *u++; | |
841 | if (UTF8_IS_CONTINUATION(c1)) { | |
356979f4 | 842 | c = UNI_TO_NATIVE(TWO_BYTE_UTF8_TO_UNI(c, c1)); |
fed3ba5d NC |
843 | } else { |
844 | Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), | |
845 | "Malformed UTF-8 character " | |
846 | "(unexpected non-continuation byte 0x%02x" | |
847 | ", immediately after start byte 0x%02x)" | |
848 | /* Dear diag.t, it's in the pod. */ | |
849 | "%s%s", c1, c, | |
850 | PL_op ? " in " : "", | |
851 | PL_op ? OP_DESC(PL_op) : ""); | |
852 | return -2; | |
853 | } | |
854 | } else { | |
855 | if (PL_op) | |
856 | Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), | |
857 | "%s in %s", unees, OP_DESC(PL_op)); | |
858 | else | |
859 | Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), unees); | |
860 | return -2; /* Really want to return undef :-) */ | |
861 | } | |
862 | } else { | |
863 | return -2; | |
864 | } | |
865 | } | |
866 | if (*b != c) { | |
867 | return *b < c ? -2 : +2; | |
868 | } | |
869 | ++b; | |
870 | } | |
871 | ||
872 | if (b == bend && u == uend) | |
873 | return 0; | |
874 | ||
875 | return b < bend ? +1 : -1; | |
876 | } | |
877 | ||
878 | /* | |
87cea99e | 879 | =for apidoc utf8_to_bytes |
6940069f | 880 | |
2bbc8d55 | 881 | Converts a string C<s> of length C<len> from UTF-8 into native byte encoding. |
246fae53 MG |
882 | Unlike C<bytes_to_utf8>, this over-writes the original string, and |
883 | updates len to contain the new length. | |
67e989fb | 884 | Returns zero on failure, setting C<len> to -1. |
6940069f | 885 | |
95be277c NC |
886 | If you need a copy of the string, see C<bytes_from_utf8>. |
887 | ||
6940069f GS |
888 | =cut |
889 | */ | |
890 | ||
891 | U8 * | |
37607a96 | 892 | Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len) |
6940069f | 893 | { |
d4c19fe8 AL |
894 | U8 * const save = s; |
895 | U8 * const send = s + *len; | |
6940069f | 896 | U8 *d; |
246fae53 | 897 | |
7918f24d NC |
898 | PERL_ARGS_ASSERT_UTF8_TO_BYTES; |
899 | ||
1e54db1a | 900 | /* ensure valid UTF-8 and chars < 256 before updating string */ |
d4c19fe8 | 901 | while (s < send) { |
dcad2880 JH |
902 | U8 c = *s++; |
903 | ||
1d72bdf6 NIS |
904 | if (!UTF8_IS_INVARIANT(c) && |
905 | (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send) | |
906 | || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) { | |
10edeb5d | 907 | *len = ((STRLEN) -1); |
dcad2880 JH |
908 | return 0; |
909 | } | |
246fae53 | 910 | } |
dcad2880 JH |
911 | |
912 | d = s = save; | |
6940069f | 913 | while (s < send) { |
ed646e6e | 914 | STRLEN ulen; |
9041c2e3 | 915 | *d++ = (U8)utf8_to_uvchr(s, &ulen); |
ed646e6e | 916 | s += ulen; |
6940069f GS |
917 | } |
918 | *d = '\0'; | |
246fae53 | 919 | *len = d - save; |
6940069f GS |
920 | return save; |
921 | } | |
922 | ||
923 | /* | |
87cea99e | 924 | =for apidoc bytes_from_utf8 |
f9a63242 | 925 | |
2bbc8d55 | 926 | Converts a string C<s> of length C<len> from UTF-8 into native byte encoding. |
35a4481c | 927 | Unlike C<utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to |
ef9edfd0 JH |
928 | the newly-created string, and updates C<len> to contain the new |
929 | length. Returns the original string if no conversion occurs, C<len> | |
930 | is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to | |
2bbc8d55 SP |
931 | 0 if C<s> is converted or consisted entirely of characters that are invariant |
932 | in utf8 (i.e., US-ASCII on non-EBCDIC machines). | |
f9a63242 | 933 | |
37607a96 PK |
934 | =cut |
935 | */ | |
f9a63242 JH |
936 | |
937 | U8 * | |
e1ec3a88 | 938 | Perl_bytes_from_utf8(pTHX_ const U8 *s, STRLEN *len, bool *is_utf8) |
f9a63242 | 939 | { |
f9a63242 | 940 | U8 *d; |
e1ec3a88 AL |
941 | const U8 *start = s; |
942 | const U8 *send; | |
f9a63242 JH |
943 | I32 count = 0; |
944 | ||
7918f24d NC |
945 | PERL_ARGS_ASSERT_BYTES_FROM_UTF8; |
946 | ||
96a5add6 | 947 | PERL_UNUSED_CONTEXT; |
f9a63242 | 948 | if (!*is_utf8) |
73d840c0 | 949 | return (U8 *)start; |
f9a63242 | 950 | |
1e54db1a | 951 | /* ensure valid UTF-8 and chars < 256 before converting string */ |
f9a63242 | 952 | for (send = s + *len; s < send;) { |
e1ec3a88 | 953 | U8 c = *s++; |
1d72bdf6 | 954 | if (!UTF8_IS_INVARIANT(c)) { |
db42d148 NIS |
955 | if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send && |
956 | (c = *s++) && UTF8_IS_CONTINUATION(c)) | |
957 | count++; | |
958 | else | |
73d840c0 | 959 | return (U8 *)start; |
db42d148 | 960 | } |
f9a63242 JH |
961 | } |
962 | ||
35da51f7 | 963 | *is_utf8 = FALSE; |
f9a63242 | 964 | |
212542aa | 965 | Newx(d, (*len) - count + 1, U8); |
ef9edfd0 | 966 | s = start; start = d; |
f9a63242 JH |
967 | while (s < send) { |
968 | U8 c = *s++; | |
c4d5f83a NIS |
969 | if (!UTF8_IS_INVARIANT(c)) { |
970 | /* Then it is two-byte encoded */ | |
356979f4 | 971 | c = UNI_TO_NATIVE(TWO_BYTE_UTF8_TO_UNI(c, *s++)); |
c4d5f83a NIS |
972 | } |
973 | *d++ = c; | |
f9a63242 JH |
974 | } |
975 | *d = '\0'; | |
976 | *len = d - start; | |
73d840c0 | 977 | return (U8 *)start; |
f9a63242 JH |
978 | } |
979 | ||
980 | /* | |
87cea99e | 981 | =for apidoc bytes_to_utf8 |
6940069f | 982 | |
2bbc8d55 | 983 | Converts a string C<s> of length C<len> from the native encoding into UTF-8. |
6662521e GS |
984 | Returns a pointer to the newly-created string, and sets C<len> to |
985 | reflect the new length. | |
6940069f | 986 | |
2bbc8d55 SP |
987 | A NUL character will be written after the end of the string. |
988 | ||
989 | If you want to convert to UTF-8 from encodings other than | |
990 | the native (Latin1 or EBCDIC), | |
c9ada85f JH |
991 | see sv_recode_to_utf8(). |
992 | ||
497711e7 | 993 | =cut |
6940069f GS |
994 | */ |
995 | ||
996 | U8* | |
35a4481c | 997 | Perl_bytes_to_utf8(pTHX_ const U8 *s, STRLEN *len) |
6940069f | 998 | { |
35a4481c | 999 | const U8 * const send = s + (*len); |
6940069f GS |
1000 | U8 *d; |
1001 | U8 *dst; | |
7918f24d NC |
1002 | |
1003 | PERL_ARGS_ASSERT_BYTES_TO_UTF8; | |
96a5add6 | 1004 | PERL_UNUSED_CONTEXT; |
6940069f | 1005 | |
212542aa | 1006 | Newx(d, (*len) * 2 + 1, U8); |
6940069f GS |
1007 | dst = d; |
1008 | ||
1009 | while (s < send) { | |
35a4481c | 1010 | const UV uv = NATIVE_TO_ASCII(*s++); |
c4d5f83a | 1011 | if (UNI_IS_INVARIANT(uv)) |
eb160463 | 1012 | *d++ = (U8)UTF_TO_NATIVE(uv); |
6940069f | 1013 | else { |
eb160463 GS |
1014 | *d++ = (U8)UTF8_EIGHT_BIT_HI(uv); |
1015 | *d++ = (U8)UTF8_EIGHT_BIT_LO(uv); | |
6940069f GS |
1016 | } |
1017 | } | |
1018 | *d = '\0'; | |
6662521e | 1019 | *len = d-dst; |
6940069f GS |
1020 | return dst; |
1021 | } | |
1022 | ||
a0ed51b3 | 1023 | /* |
dea0fc0b | 1024 | * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8. |
a0ed51b3 LW |
1025 | * |
1026 | * Destination must be pre-extended to 3/2 source. Do not use in-place. | |
1027 | * We optimize for native, for obvious reasons. */ | |
1028 | ||
1029 | U8* | |
dea0fc0b | 1030 | Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 | 1031 | { |
dea0fc0b JH |
1032 | U8* pend; |
1033 | U8* dstart = d; | |
1034 | ||
7918f24d NC |
1035 | PERL_ARGS_ASSERT_UTF16_TO_UTF8; |
1036 | ||
dea0fc0b | 1037 | if (bytelen & 1) |
f5992bc4 | 1038 | Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen %"UVuf, (UV)bytelen); |
dea0fc0b JH |
1039 | |
1040 | pend = p + bytelen; | |
1041 | ||
a0ed51b3 | 1042 | while (p < pend) { |
dea0fc0b JH |
1043 | UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */ |
1044 | p += 2; | |
a0ed51b3 | 1045 | if (uv < 0x80) { |
e294cc5d JH |
1046 | #ifdef EBCDIC |
1047 | *d++ = UNI_TO_NATIVE(uv); | |
1048 | #else | |
eb160463 | 1049 | *d++ = (U8)uv; |
e294cc5d | 1050 | #endif |
a0ed51b3 LW |
1051 | continue; |
1052 | } | |
1053 | if (uv < 0x800) { | |
eb160463 GS |
1054 | *d++ = (U8)(( uv >> 6) | 0xc0); |
1055 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
1056 | continue; |
1057 | } | |
52b9aa85 | 1058 | if (uv >= 0xd800 && uv <= 0xdbff) { /* surrogates */ |
01ea242b | 1059 | if (p >= pend) { |
dea0fc0b | 1060 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
01ea242b NC |
1061 | } else { |
1062 | UV low = (p[0] << 8) + p[1]; | |
1063 | p += 2; | |
52b9aa85 | 1064 | if (low < 0xdc00 || low > 0xdfff) |
01ea242b NC |
1065 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
1066 | uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000; | |
1067 | } | |
dbde1951 NC |
1068 | } else if (uv >= 0xdc00 && uv <= 0xdfff) { |
1069 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); | |
a0ed51b3 LW |
1070 | } |
1071 | if (uv < 0x10000) { | |
eb160463 GS |
1072 | *d++ = (U8)(( uv >> 12) | 0xe0); |
1073 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
1074 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
1075 | continue; |
1076 | } | |
1077 | else { | |
eb160463 GS |
1078 | *d++ = (U8)(( uv >> 18) | 0xf0); |
1079 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
1080 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
1081 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
1082 | continue; |
1083 | } | |
1084 | } | |
dea0fc0b | 1085 | *newlen = d - dstart; |
a0ed51b3 LW |
1086 | return d; |
1087 | } | |
1088 | ||
1089 | /* Note: this one is slightly destructive of the source. */ | |
1090 | ||
1091 | U8* | |
dea0fc0b | 1092 | Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 LW |
1093 | { |
1094 | U8* s = (U8*)p; | |
d4c19fe8 | 1095 | U8* const send = s + bytelen; |
7918f24d NC |
1096 | |
1097 | PERL_ARGS_ASSERT_UTF16_TO_UTF8_REVERSED; | |
1098 | ||
e0ea5e2d NC |
1099 | if (bytelen & 1) |
1100 | Perl_croak(aTHX_ "panic: utf16_to_utf8_reversed: odd bytelen %"UVuf, | |
1101 | (UV)bytelen); | |
1102 | ||
a0ed51b3 | 1103 | while (s < send) { |
d4c19fe8 | 1104 | const U8 tmp = s[0]; |
a0ed51b3 LW |
1105 | s[0] = s[1]; |
1106 | s[1] = tmp; | |
1107 | s += 2; | |
1108 | } | |
dea0fc0b | 1109 | return utf16_to_utf8(p, d, bytelen, newlen); |
a0ed51b3 LW |
1110 | } |
1111 | ||
1112 | /* for now these are all defined (inefficiently) in terms of the utf8 versions */ | |
1113 | ||
1114 | bool | |
84afefe6 | 1115 | Perl_is_uni_alnum(pTHX_ UV c) |
a0ed51b3 | 1116 | { |
89ebb4a3 | 1117 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1118 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1119 | return is_utf8_alnum(tmpbuf); |
1120 | } | |
1121 | ||
1122 | bool | |
84afefe6 | 1123 | Perl_is_uni_idfirst(pTHX_ UV c) |
a0ed51b3 | 1124 | { |
89ebb4a3 | 1125 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1126 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1127 | return is_utf8_idfirst(tmpbuf); |
1128 | } | |
1129 | ||
1130 | bool | |
84afefe6 | 1131 | Perl_is_uni_alpha(pTHX_ UV c) |
a0ed51b3 | 1132 | { |
89ebb4a3 | 1133 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1134 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1135 | return is_utf8_alpha(tmpbuf); |
1136 | } | |
1137 | ||
1138 | bool | |
84afefe6 | 1139 | Perl_is_uni_ascii(pTHX_ UV c) |
4d61ec05 | 1140 | { |
89ebb4a3 | 1141 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1142 | uvchr_to_utf8(tmpbuf, c); |
4d61ec05 GS |
1143 | return is_utf8_ascii(tmpbuf); |
1144 | } | |
1145 | ||
1146 | bool | |
84afefe6 | 1147 | Perl_is_uni_space(pTHX_ UV c) |
a0ed51b3 | 1148 | { |
89ebb4a3 | 1149 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1150 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1151 | return is_utf8_space(tmpbuf); |
1152 | } | |
1153 | ||
1154 | bool | |
84afefe6 | 1155 | Perl_is_uni_digit(pTHX_ UV c) |
a0ed51b3 | 1156 | { |
89ebb4a3 | 1157 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1158 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1159 | return is_utf8_digit(tmpbuf); |
1160 | } | |
1161 | ||
1162 | bool | |
84afefe6 | 1163 | Perl_is_uni_upper(pTHX_ UV c) |
a0ed51b3 | 1164 | { |
89ebb4a3 | 1165 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1166 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1167 | return is_utf8_upper(tmpbuf); |
1168 | } | |
1169 | ||
1170 | bool | |
84afefe6 | 1171 | Perl_is_uni_lower(pTHX_ UV c) |
a0ed51b3 | 1172 | { |
89ebb4a3 | 1173 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1174 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1175 | return is_utf8_lower(tmpbuf); |
1176 | } | |
1177 | ||
1178 | bool | |
84afefe6 | 1179 | Perl_is_uni_cntrl(pTHX_ UV c) |
b8c5462f | 1180 | { |
89ebb4a3 | 1181 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1182 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f JH |
1183 | return is_utf8_cntrl(tmpbuf); |
1184 | } | |
1185 | ||
1186 | bool | |
84afefe6 | 1187 | Perl_is_uni_graph(pTHX_ UV c) |
b8c5462f | 1188 | { |
89ebb4a3 | 1189 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1190 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f JH |
1191 | return is_utf8_graph(tmpbuf); |
1192 | } | |
1193 | ||
1194 | bool | |
84afefe6 | 1195 | Perl_is_uni_print(pTHX_ UV c) |
a0ed51b3 | 1196 | { |
89ebb4a3 | 1197 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1198 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1199 | return is_utf8_print(tmpbuf); |
1200 | } | |
1201 | ||
b8c5462f | 1202 | bool |
84afefe6 | 1203 | Perl_is_uni_punct(pTHX_ UV c) |
b8c5462f | 1204 | { |
89ebb4a3 | 1205 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1206 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f JH |
1207 | return is_utf8_punct(tmpbuf); |
1208 | } | |
1209 | ||
4d61ec05 | 1210 | bool |
84afefe6 | 1211 | Perl_is_uni_xdigit(pTHX_ UV c) |
4d61ec05 | 1212 | { |
89ebb4a3 | 1213 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
230880c1 | 1214 | uvchr_to_utf8(tmpbuf, c); |
4d61ec05 GS |
1215 | return is_utf8_xdigit(tmpbuf); |
1216 | } | |
1217 | ||
84afefe6 JH |
1218 | UV |
1219 | Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp) | |
a0ed51b3 | 1220 | { |
7918f24d NC |
1221 | PERL_ARGS_ASSERT_TO_UNI_UPPER; |
1222 | ||
0ebc6274 JH |
1223 | uvchr_to_utf8(p, c); |
1224 | return to_utf8_upper(p, p, lenp); | |
a0ed51b3 LW |
1225 | } |
1226 | ||
84afefe6 JH |
1227 | UV |
1228 | Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp) | |
a0ed51b3 | 1229 | { |
7918f24d NC |
1230 | PERL_ARGS_ASSERT_TO_UNI_TITLE; |
1231 | ||
0ebc6274 JH |
1232 | uvchr_to_utf8(p, c); |
1233 | return to_utf8_title(p, p, lenp); | |
a0ed51b3 LW |
1234 | } |
1235 | ||
84afefe6 JH |
1236 | UV |
1237 | Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp) | |
a0ed51b3 | 1238 | { |
7918f24d NC |
1239 | PERL_ARGS_ASSERT_TO_UNI_LOWER; |
1240 | ||
0ebc6274 JH |
1241 | uvchr_to_utf8(p, c); |
1242 | return to_utf8_lower(p, p, lenp); | |
a0ed51b3 LW |
1243 | } |
1244 | ||
84afefe6 JH |
1245 | UV |
1246 | Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp) | |
1247 | { | |
7918f24d NC |
1248 | PERL_ARGS_ASSERT_TO_UNI_FOLD; |
1249 | ||
0ebc6274 JH |
1250 | uvchr_to_utf8(p, c); |
1251 | return to_utf8_fold(p, p, lenp); | |
84afefe6 JH |
1252 | } |
1253 | ||
a0ed51b3 LW |
1254 | /* for now these all assume no locale info available for Unicode > 255 */ |
1255 | ||
1256 | bool | |
84afefe6 | 1257 | Perl_is_uni_alnum_lc(pTHX_ UV c) |
a0ed51b3 LW |
1258 | { |
1259 | return is_uni_alnum(c); /* XXX no locale support yet */ | |
1260 | } | |
1261 | ||
1262 | bool | |
84afefe6 | 1263 | Perl_is_uni_idfirst_lc(pTHX_ UV c) |
a0ed51b3 LW |
1264 | { |
1265 | return is_uni_idfirst(c); /* XXX no locale support yet */ | |
1266 | } | |
1267 | ||
1268 | bool | |
84afefe6 | 1269 | Perl_is_uni_alpha_lc(pTHX_ UV c) |
a0ed51b3 LW |
1270 | { |
1271 | return is_uni_alpha(c); /* XXX no locale support yet */ | |
1272 | } | |
1273 | ||
1274 | bool | |
84afefe6 | 1275 | Perl_is_uni_ascii_lc(pTHX_ UV c) |
4d61ec05 GS |
1276 | { |
1277 | return is_uni_ascii(c); /* XXX no locale support yet */ | |
1278 | } | |
1279 | ||
1280 | bool | |
84afefe6 | 1281 | Perl_is_uni_space_lc(pTHX_ UV c) |
a0ed51b3 LW |
1282 | { |
1283 | return is_uni_space(c); /* XXX no locale support yet */ | |
1284 | } | |
1285 | ||
1286 | bool | |
84afefe6 | 1287 | Perl_is_uni_digit_lc(pTHX_ UV c) |
a0ed51b3 LW |
1288 | { |
1289 | return is_uni_digit(c); /* XXX no locale support yet */ | |
1290 | } | |
1291 | ||
1292 | bool | |
84afefe6 | 1293 | Perl_is_uni_upper_lc(pTHX_ UV c) |
a0ed51b3 LW |
1294 | { |
1295 | return is_uni_upper(c); /* XXX no locale support yet */ | |
1296 | } | |
1297 | ||
1298 | bool | |
84afefe6 | 1299 | Perl_is_uni_lower_lc(pTHX_ UV c) |
a0ed51b3 LW |
1300 | { |
1301 | return is_uni_lower(c); /* XXX no locale support yet */ | |
1302 | } | |
1303 | ||
1304 | bool | |
84afefe6 | 1305 | Perl_is_uni_cntrl_lc(pTHX_ UV c) |
b8c5462f JH |
1306 | { |
1307 | return is_uni_cntrl(c); /* XXX no locale support yet */ | |
1308 | } | |
1309 | ||
1310 | bool | |
84afefe6 | 1311 | Perl_is_uni_graph_lc(pTHX_ UV c) |
b8c5462f JH |
1312 | { |
1313 | return is_uni_graph(c); /* XXX no locale support yet */ | |
1314 | } | |
1315 | ||
1316 | bool | |
84afefe6 | 1317 | Perl_is_uni_print_lc(pTHX_ UV c) |
a0ed51b3 LW |
1318 | { |
1319 | return is_uni_print(c); /* XXX no locale support yet */ | |
1320 | } | |
1321 | ||
b8c5462f | 1322 | bool |
84afefe6 | 1323 | Perl_is_uni_punct_lc(pTHX_ UV c) |
b8c5462f JH |
1324 | { |
1325 | return is_uni_punct(c); /* XXX no locale support yet */ | |
1326 | } | |
1327 | ||
4d61ec05 | 1328 | bool |
84afefe6 | 1329 | Perl_is_uni_xdigit_lc(pTHX_ UV c) |
4d61ec05 GS |
1330 | { |
1331 | return is_uni_xdigit(c); /* XXX no locale support yet */ | |
1332 | } | |
1333 | ||
b7ac61fa JH |
1334 | U32 |
1335 | Perl_to_uni_upper_lc(pTHX_ U32 c) | |
1336 | { | |
ee099d14 JH |
1337 | /* XXX returns only the first character -- do not use XXX */ |
1338 | /* XXX no locale support yet */ | |
1339 | STRLEN len; | |
89ebb4a3 | 1340 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 | 1341 | return (U32)to_uni_upper(c, tmpbuf, &len); |
b7ac61fa JH |
1342 | } |
1343 | ||
1344 | U32 | |
1345 | Perl_to_uni_title_lc(pTHX_ U32 c) | |
1346 | { | |
ee099d14 JH |
1347 | /* XXX returns only the first character XXX -- do not use XXX */ |
1348 | /* XXX no locale support yet */ | |
1349 | STRLEN len; | |
89ebb4a3 | 1350 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 | 1351 | return (U32)to_uni_title(c, tmpbuf, &len); |
b7ac61fa JH |
1352 | } |
1353 | ||
1354 | U32 | |
1355 | Perl_to_uni_lower_lc(pTHX_ U32 c) | |
1356 | { | |
ee099d14 JH |
1357 | /* XXX returns only the first character -- do not use XXX */ |
1358 | /* XXX no locale support yet */ | |
1359 | STRLEN len; | |
89ebb4a3 | 1360 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 | 1361 | return (U32)to_uni_lower(c, tmpbuf, &len); |
b7ac61fa JH |
1362 | } |
1363 | ||
7452cf6a | 1364 | static bool |
5141f98e | 1365 | S_is_utf8_common(pTHX_ const U8 *const p, SV **swash, |
bde6a22d NC |
1366 | const char *const swashname) |
1367 | { | |
97aff369 | 1368 | dVAR; |
7918f24d NC |
1369 | |
1370 | PERL_ARGS_ASSERT_IS_UTF8_COMMON; | |
1371 | ||
bde6a22d NC |
1372 | if (!is_utf8_char(p)) |
1373 | return FALSE; | |
1374 | if (!*swash) | |
711a919c | 1375 | *swash = swash_init("utf8", swashname, &PL_sv_undef, 1, 0); |
bde6a22d NC |
1376 | return swash_fetch(*swash, p, TRUE) != 0; |
1377 | } | |
1378 | ||
1379 | bool | |
7fc63493 | 1380 | Perl_is_utf8_alnum(pTHX_ const U8 *p) |
a0ed51b3 | 1381 | { |
97aff369 | 1382 | dVAR; |
7918f24d NC |
1383 | |
1384 | PERL_ARGS_ASSERT_IS_UTF8_ALNUM; | |
1385 | ||
671c33bf NC |
1386 | /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true |
1387 | * descendant of isalnum(3), in other words, it doesn't | |
1388 | * contain the '_'. --jhi */ | |
d4c19fe8 | 1389 | return is_utf8_common(p, &PL_utf8_alnum, "IsWord"); |
a0ed51b3 LW |
1390 | } |
1391 | ||
1392 | bool | |
7fc63493 | 1393 | Perl_is_utf8_idfirst(pTHX_ const U8 *p) /* The naming is historical. */ |
a0ed51b3 | 1394 | { |
97aff369 | 1395 | dVAR; |
7918f24d NC |
1396 | |
1397 | PERL_ARGS_ASSERT_IS_UTF8_IDFIRST; | |
1398 | ||
82686b01 JH |
1399 | if (*p == '_') |
1400 | return TRUE; | |
bde6a22d | 1401 | /* is_utf8_idstart would be more logical. */ |
d4c19fe8 | 1402 | return is_utf8_common(p, &PL_utf8_idstart, "IdStart"); |
82686b01 JH |
1403 | } |
1404 | ||
1405 | bool | |
7fc63493 | 1406 | Perl_is_utf8_idcont(pTHX_ const U8 *p) |
82686b01 | 1407 | { |
97aff369 | 1408 | dVAR; |
7918f24d NC |
1409 | |
1410 | PERL_ARGS_ASSERT_IS_UTF8_IDCONT; | |
1411 | ||
82686b01 JH |
1412 | if (*p == '_') |
1413 | return TRUE; | |
d4c19fe8 | 1414 | return is_utf8_common(p, &PL_utf8_idcont, "IdContinue"); |
a0ed51b3 LW |
1415 | } |
1416 | ||
1417 | bool | |
7fc63493 | 1418 | Perl_is_utf8_alpha(pTHX_ const U8 *p) |
a0ed51b3 | 1419 | { |
97aff369 | 1420 | dVAR; |
7918f24d NC |
1421 | |
1422 | PERL_ARGS_ASSERT_IS_UTF8_ALPHA; | |
1423 | ||
d4c19fe8 | 1424 | return is_utf8_common(p, &PL_utf8_alpha, "IsAlpha"); |
a0ed51b3 LW |
1425 | } |
1426 | ||
1427 | bool | |
7fc63493 | 1428 | Perl_is_utf8_ascii(pTHX_ const U8 *p) |
b8c5462f | 1429 | { |
97aff369 | 1430 | dVAR; |
7918f24d NC |
1431 | |
1432 | PERL_ARGS_ASSERT_IS_UTF8_ASCII; | |
1433 | ||
d4c19fe8 | 1434 | return is_utf8_common(p, &PL_utf8_ascii, "IsAscii"); |
b8c5462f JH |
1435 | } |
1436 | ||
1437 | bool | |
7fc63493 | 1438 | Perl_is_utf8_space(pTHX_ const U8 *p) |
a0ed51b3 | 1439 | { |
97aff369 | 1440 | dVAR; |
7918f24d NC |
1441 | |
1442 | PERL_ARGS_ASSERT_IS_UTF8_SPACE; | |
1443 | ||
d4c19fe8 | 1444 | return is_utf8_common(p, &PL_utf8_space, "IsSpacePerl"); |
a0ed51b3 LW |
1445 | } |
1446 | ||
1447 | bool | |
d1eb3177 YO |
1448 | Perl_is_utf8_perl_space(pTHX_ const U8 *p) |
1449 | { | |
1450 | dVAR; | |
1451 | ||
1452 | PERL_ARGS_ASSERT_IS_UTF8_PERL_SPACE; | |
1453 | ||
1454 | return is_utf8_common(p, &PL_utf8_perl_space, "IsPerlSpace"); | |
1455 | } | |
1456 | ||
1457 | bool | |
1458 | Perl_is_utf8_perl_word(pTHX_ const U8 *p) | |
1459 | { | |
1460 | dVAR; | |
1461 | ||
1462 | PERL_ARGS_ASSERT_IS_UTF8_PERL_WORD; | |
1463 | ||
1464 | return is_utf8_common(p, &PL_utf8_perl_word, "IsPerlWord"); | |
1465 | } | |
1466 | ||
1467 | bool | |
7fc63493 | 1468 | Perl_is_utf8_digit(pTHX_ const U8 *p) |
a0ed51b3 | 1469 | { |
97aff369 | 1470 | dVAR; |
7918f24d NC |
1471 | |
1472 | PERL_ARGS_ASSERT_IS_UTF8_DIGIT; | |
1473 | ||
d4c19fe8 | 1474 | return is_utf8_common(p, &PL_utf8_digit, "IsDigit"); |
a0ed51b3 LW |
1475 | } |
1476 | ||
1477 | bool | |
d1eb3177 YO |
1478 | Perl_is_utf8_posix_digit(pTHX_ const U8 *p) |
1479 | { | |
1480 | dVAR; | |
1481 | ||
1482 | PERL_ARGS_ASSERT_IS_UTF8_POSIX_DIGIT; | |
1483 | ||
1484 | return is_utf8_common(p, &PL_utf8_posix_digit, "IsPosixDigit"); | |
1485 | } | |
1486 | ||
1487 | bool | |
7fc63493 | 1488 | Perl_is_utf8_upper(pTHX_ const U8 *p) |
a0ed51b3 | 1489 | { |
97aff369 | 1490 | dVAR; |
7918f24d NC |
1491 | |
1492 | PERL_ARGS_ASSERT_IS_UTF8_UPPER; | |
1493 | ||
d4c19fe8 | 1494 | return is_utf8_common(p, &PL_utf8_upper, "IsUppercase"); |
a0ed51b3 LW |
1495 | } |
1496 | ||
1497 | bool | |
7fc63493 | 1498 | Perl_is_utf8_lower(pTHX_ const U8 *p) |
a0ed51b3 | 1499 | { |
97aff369 | 1500 | dVAR; |
7918f24d NC |
1501 | |
1502 | PERL_ARGS_ASSERT_IS_UTF8_LOWER; | |
1503 | ||
d4c19fe8 | 1504 | return is_utf8_common(p, &PL_utf8_lower, "IsLowercase"); |
a0ed51b3 LW |
1505 | } |
1506 | ||
1507 | bool | |
7fc63493 | 1508 | Perl_is_utf8_cntrl(pTHX_ const U8 *p) |
b8c5462f | 1509 | { |
97aff369 | 1510 | dVAR; |
7918f24d NC |
1511 | |
1512 | PERL_ARGS_ASSERT_IS_UTF8_CNTRL; | |
1513 | ||
d4c19fe8 | 1514 | return is_utf8_common(p, &PL_utf8_cntrl, "IsCntrl"); |
b8c5462f JH |
1515 | } |
1516 | ||
1517 | bool | |
7fc63493 | 1518 | Perl_is_utf8_graph(pTHX_ const U8 *p) |
b8c5462f | 1519 | { |
97aff369 | 1520 | dVAR; |
7918f24d NC |
1521 | |
1522 | PERL_ARGS_ASSERT_IS_UTF8_GRAPH; | |
1523 | ||
d4c19fe8 | 1524 | return is_utf8_common(p, &PL_utf8_graph, "IsGraph"); |
b8c5462f JH |
1525 | } |
1526 | ||
1527 | bool | |
7fc63493 | 1528 | Perl_is_utf8_print(pTHX_ const U8 *p) |
a0ed51b3 | 1529 | { |
97aff369 | 1530 | dVAR; |
7918f24d NC |
1531 | |
1532 | PERL_ARGS_ASSERT_IS_UTF8_PRINT; | |
1533 | ||
d4c19fe8 | 1534 | return is_utf8_common(p, &PL_utf8_print, "IsPrint"); |
a0ed51b3 LW |
1535 | } |
1536 | ||
1537 | bool | |
7fc63493 | 1538 | Perl_is_utf8_punct(pTHX_ const U8 *p) |
b8c5462f | 1539 | { |
97aff369 | 1540 | dVAR; |
7918f24d NC |
1541 | |
1542 | PERL_ARGS_ASSERT_IS_UTF8_PUNCT; | |
1543 | ||
d4c19fe8 | 1544 | return is_utf8_common(p, &PL_utf8_punct, "IsPunct"); |
b8c5462f JH |
1545 | } |
1546 | ||
1547 | bool | |
7fc63493 | 1548 | Perl_is_utf8_xdigit(pTHX_ const U8 *p) |
b8c5462f | 1549 | { |
97aff369 | 1550 | dVAR; |
7918f24d NC |
1551 | |
1552 | PERL_ARGS_ASSERT_IS_UTF8_XDIGIT; | |
1553 | ||
d1eb3177 | 1554 | return is_utf8_common(p, &PL_utf8_xdigit, "IsXDigit"); |
b8c5462f JH |
1555 | } |
1556 | ||
1557 | bool | |
7fc63493 | 1558 | Perl_is_utf8_mark(pTHX_ const U8 *p) |
a0ed51b3 | 1559 | { |
97aff369 | 1560 | dVAR; |
7918f24d NC |
1561 | |
1562 | PERL_ARGS_ASSERT_IS_UTF8_MARK; | |
1563 | ||
d4c19fe8 | 1564 | return is_utf8_common(p, &PL_utf8_mark, "IsM"); |
a0ed51b3 LW |
1565 | } |
1566 | ||
37e2e78e KW |
1567 | bool |
1568 | Perl_is_utf8_X_begin(pTHX_ const U8 *p) | |
1569 | { | |
1570 | dVAR; | |
1571 | ||
1572 | PERL_ARGS_ASSERT_IS_UTF8_X_BEGIN; | |
1573 | ||
1574 | return is_utf8_common(p, &PL_utf8_X_begin, "_X_Begin"); | |
1575 | } | |
1576 | ||
1577 | bool | |
1578 | Perl_is_utf8_X_extend(pTHX_ const U8 *p) | |
1579 | { | |
1580 | dVAR; | |
1581 | ||
1582 | PERL_ARGS_ASSERT_IS_UTF8_X_EXTEND; | |
1583 | ||
1584 | return is_utf8_common(p, &PL_utf8_X_extend, "_X_Extend"); | |
1585 | } | |
1586 | ||
1587 | bool | |
1588 | Perl_is_utf8_X_prepend(pTHX_ const U8 *p) | |
1589 | { | |
1590 | dVAR; | |
1591 | ||
1592 | PERL_ARGS_ASSERT_IS_UTF8_X_PREPEND; | |
1593 | ||
1594 | return is_utf8_common(p, &PL_utf8_X_prepend, "GCB=Prepend"); | |
1595 | } | |
1596 | ||
1597 | bool | |
1598 | Perl_is_utf8_X_non_hangul(pTHX_ const U8 *p) | |
1599 | { | |
1600 | dVAR; | |
1601 | ||
1602 | PERL_ARGS_ASSERT_IS_UTF8_X_NON_HANGUL; | |
1603 | ||
1604 | return is_utf8_common(p, &PL_utf8_X_non_hangul, "HST=Not_Applicable"); | |
1605 | } | |
1606 | ||
1607 | bool | |
1608 | Perl_is_utf8_X_L(pTHX_ const U8 *p) | |
1609 | { | |
1610 | dVAR; | |
1611 | ||
1612 | PERL_ARGS_ASSERT_IS_UTF8_X_L; | |
1613 | ||
1614 | return is_utf8_common(p, &PL_utf8_X_L, "GCB=L"); | |
1615 | } | |
1616 | ||
1617 | bool | |
1618 | Perl_is_utf8_X_LV(pTHX_ const U8 *p) | |
1619 | { | |
1620 | dVAR; | |
1621 | ||
1622 | PERL_ARGS_ASSERT_IS_UTF8_X_LV; | |
1623 | ||
1624 | return is_utf8_common(p, &PL_utf8_X_LV, "GCB=LV"); | |
1625 | } | |
1626 | ||
1627 | bool | |
1628 | Perl_is_utf8_X_LVT(pTHX_ const U8 *p) | |
1629 | { | |
1630 | dVAR; | |
1631 | ||
1632 | PERL_ARGS_ASSERT_IS_UTF8_X_LVT; | |
1633 | ||
1634 | return is_utf8_common(p, &PL_utf8_X_LVT, "GCB=LVT"); | |
1635 | } | |
1636 | ||
1637 | bool | |
1638 | Perl_is_utf8_X_T(pTHX_ const U8 *p) | |
1639 | { | |
1640 | dVAR; | |
1641 | ||
1642 | PERL_ARGS_ASSERT_IS_UTF8_X_T; | |
1643 | ||
1644 | return is_utf8_common(p, &PL_utf8_X_T, "GCB=T"); | |
1645 | } | |
1646 | ||
1647 | bool | |
1648 | Perl_is_utf8_X_V(pTHX_ const U8 *p) | |
1649 | { | |
1650 | dVAR; | |
1651 | ||
1652 | PERL_ARGS_ASSERT_IS_UTF8_X_V; | |
1653 | ||
1654 | return is_utf8_common(p, &PL_utf8_X_V, "GCB=V"); | |
1655 | } | |
1656 | ||
1657 | bool | |
1658 | Perl_is_utf8_X_LV_LVT_V(pTHX_ const U8 *p) | |
1659 | { | |
1660 | dVAR; | |
1661 | ||
1662 | PERL_ARGS_ASSERT_IS_UTF8_X_LV_LVT_V; | |
1663 | ||
1664 | return is_utf8_common(p, &PL_utf8_X_LV_LVT_V, "_X_LV_LVT_V"); | |
1665 | } | |
1666 | ||
6b5c0936 | 1667 | /* |
87cea99e | 1668 | =for apidoc to_utf8_case |
6b5c0936 JH |
1669 | |
1670 | The "p" contains the pointer to the UTF-8 string encoding | |
1671 | the character that is being converted. | |
1672 | ||
1673 | The "ustrp" is a pointer to the character buffer to put the | |
1674 | conversion result to. The "lenp" is a pointer to the length | |
1675 | of the result. | |
1676 | ||
0134edef | 1677 | The "swashp" is a pointer to the swash to use. |
6b5c0936 | 1678 | |
0134edef | 1679 | Both the special and normal mappings are stored lib/unicore/To/Foo.pl, |
8fe4d5b2 | 1680 | and loaded by SWASHNEW, using lib/utf8_heavy.pl. The special (usually, |
0134edef | 1681 | but not always, a multicharacter mapping), is tried first. |
6b5c0936 | 1682 | |
0134edef JH |
1683 | The "special" is a string like "utf8::ToSpecLower", which means the |
1684 | hash %utf8::ToSpecLower. The access to the hash is through | |
1685 | Perl_to_utf8_case(). | |
6b5c0936 | 1686 | |
0134edef JH |
1687 | The "normal" is a string like "ToLower" which means the swash |
1688 | %utf8::ToLower. | |
1689 | ||
1690 | =cut */ | |
6b5c0936 | 1691 | |
2104c8d9 | 1692 | UV |
9a957fbc AL |
1693 | Perl_to_utf8_case(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp, |
1694 | SV **swashp, const char *normal, const char *special) | |
a0ed51b3 | 1695 | { |
97aff369 | 1696 | dVAR; |
89ebb4a3 | 1697 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
0134edef | 1698 | STRLEN len = 0; |
aec46f14 | 1699 | const UV uv0 = utf8_to_uvchr(p, NULL); |
1feea2c7 JH |
1700 | /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings |
1701 | * are necessary in EBCDIC, they are redundant no-ops | |
1702 | * in ASCII-ish platforms, and hopefully optimized away. */ | |
f54cb97a | 1703 | const UV uv1 = NATIVE_TO_UNI(uv0); |
7918f24d NC |
1704 | |
1705 | PERL_ARGS_ASSERT_TO_UTF8_CASE; | |
1706 | ||
1feea2c7 | 1707 | uvuni_to_utf8(tmpbuf, uv1); |
0134edef JH |
1708 | |
1709 | if (!*swashp) /* load on-demand */ | |
1710 | *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0); | |
37e2e78e KW |
1711 | /* This is the beginnings of a skeleton of code to read the info section |
1712 | * that is in all the swashes in case we ever want to do that, so one can | |
1713 | * read things whose maps aren't code points, and whose default if missing | |
1714 | * is not to the code point itself. This was just to see if it actually | |
1715 | * worked. Details on what the possibilities are are in perluniprops.pod | |
1716 | HV * const hv = get_hv("utf8::SwashInfo", 0); | |
1717 | if (hv) { | |
1718 | SV **svp; | |
1719 | svp = hv_fetch(hv, (const char*)normal, strlen(normal), FALSE); | |
1720 | const char *s; | |
1721 | ||
1722 | HV * const this_hash = SvRV(*svp); | |
1723 | svp = hv_fetch(this_hash, "type", strlen("type"), FALSE); | |
1724 | s = SvPV_const(*svp, len); | |
1725 | } | |
1726 | }*/ | |
0134edef | 1727 | |
a6f87d8c | 1728 | if (special) { |
0134edef | 1729 | /* It might be "special" (sometimes, but not always, |
2a37f04d | 1730 | * a multicharacter mapping) */ |
6673a63c | 1731 | HV * const hv = get_hv(special, 0); |
b08cf34e JH |
1732 | SV **svp; |
1733 | ||
35da51f7 | 1734 | if (hv && |
b08cf34e JH |
1735 | (svp = hv_fetch(hv, (const char*)tmpbuf, UNISKIP(uv1), FALSE)) && |
1736 | (*svp)) { | |
cfd0369c | 1737 | const char *s; |
47654450 | 1738 | |
cfd0369c | 1739 | s = SvPV_const(*svp, len); |
47654450 JH |
1740 | if (len == 1) |
1741 | len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp; | |
2a37f04d | 1742 | else { |
2f9475ad JH |
1743 | #ifdef EBCDIC |
1744 | /* If we have EBCDIC we need to remap the characters | |
1745 | * since any characters in the low 256 are Unicode | |
1746 | * code points, not EBCDIC. */ | |
7cda7a3d | 1747 | U8 *t = (U8*)s, *tend = t + len, *d; |
2f9475ad JH |
1748 | |
1749 | d = tmpbuf; | |
b08cf34e | 1750 | if (SvUTF8(*svp)) { |
2f9475ad JH |
1751 | STRLEN tlen = 0; |
1752 | ||
1753 | while (t < tend) { | |
d4c19fe8 | 1754 | const UV c = utf8_to_uvchr(t, &tlen); |
2f9475ad JH |
1755 | if (tlen > 0) { |
1756 | d = uvchr_to_utf8(d, UNI_TO_NATIVE(c)); | |
1757 | t += tlen; | |
1758 | } | |
1759 | else | |
1760 | break; | |
1761 | } | |
1762 | } | |
1763 | else { | |
36fec512 JH |
1764 | while (t < tend) { |
1765 | d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t)); | |
1766 | t++; | |
1767 | } | |
2f9475ad JH |
1768 | } |
1769 | len = d - tmpbuf; | |
1770 | Copy(tmpbuf, ustrp, len, U8); | |
1771 | #else | |
d2dcd0fb | 1772 | Copy(s, ustrp, len, U8); |
2f9475ad | 1773 | #endif |
29e98929 | 1774 | } |
983ffd37 | 1775 | } |
0134edef JH |
1776 | } |
1777 | ||
1778 | if (!len && *swashp) { | |
d4c19fe8 AL |
1779 | const UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE); |
1780 | ||
0134edef JH |
1781 | if (uv2) { |
1782 | /* It was "normal" (a single character mapping). */ | |
d4c19fe8 | 1783 | const UV uv3 = UNI_TO_NATIVE(uv2); |
e9101d72 | 1784 | len = uvchr_to_utf8(ustrp, uv3) - ustrp; |
2a37f04d JH |
1785 | } |
1786 | } | |
1feea2c7 | 1787 | |
37e2e78e KW |
1788 | if (!len) /* Neither: just copy. In other words, there was no mapping |
1789 | defined, which means that the code point maps to itself */ | |
0134edef JH |
1790 | len = uvchr_to_utf8(ustrp, uv0) - ustrp; |
1791 | ||
2a37f04d JH |
1792 | if (lenp) |
1793 | *lenp = len; | |
1794 | ||
0134edef | 1795 | return len ? utf8_to_uvchr(ustrp, 0) : 0; |
a0ed51b3 LW |
1796 | } |
1797 | ||
d3e79532 | 1798 | /* |
87cea99e | 1799 | =for apidoc to_utf8_upper |
d3e79532 JH |
1800 | |
1801 | Convert the UTF-8 encoded character at p to its uppercase version and | |
1802 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note | |
89ebb4a3 JH |
1803 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since |
1804 | the uppercase version may be longer than the original character. | |
d3e79532 JH |
1805 | |
1806 | The first character of the uppercased version is returned | |
1807 | (but note, as explained above, that there may be more.) | |
1808 | ||
1809 | =cut */ | |
1810 | ||
2104c8d9 | 1811 | UV |
7fc63493 | 1812 | Perl_to_utf8_upper(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 | 1813 | { |
97aff369 | 1814 | dVAR; |
7918f24d NC |
1815 | |
1816 | PERL_ARGS_ASSERT_TO_UTF8_UPPER; | |
1817 | ||
983ffd37 | 1818 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 | 1819 | &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper"); |
983ffd37 | 1820 | } |
a0ed51b3 | 1821 | |
d3e79532 | 1822 | /* |
87cea99e | 1823 | =for apidoc to_utf8_title |
d3e79532 JH |
1824 | |
1825 | Convert the UTF-8 encoded character at p to its titlecase version and | |
1826 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note | |
89ebb4a3 JH |
1827 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
1828 | titlecase version may be longer than the original character. | |
d3e79532 JH |
1829 | |
1830 | The first character of the titlecased version is returned | |
1831 | (but note, as explained above, that there may be more.) | |
1832 | ||
1833 | =cut */ | |
1834 | ||
983ffd37 | 1835 | UV |
7fc63493 | 1836 | Perl_to_utf8_title(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
983ffd37 | 1837 | { |
97aff369 | 1838 | dVAR; |
7918f24d NC |
1839 | |
1840 | PERL_ARGS_ASSERT_TO_UTF8_TITLE; | |
1841 | ||
983ffd37 | 1842 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 | 1843 | &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle"); |
a0ed51b3 LW |
1844 | } |
1845 | ||
d3e79532 | 1846 | /* |
87cea99e | 1847 | =for apidoc to_utf8_lower |
d3e79532 JH |
1848 | |
1849 | Convert the UTF-8 encoded character at p to its lowercase version and | |
1850 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note | |
89ebb4a3 JH |
1851 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
1852 | lowercase version may be longer than the original character. | |
d3e79532 JH |
1853 | |
1854 | The first character of the lowercased version is returned | |
1855 | (but note, as explained above, that there may be more.) | |
1856 | ||
1857 | =cut */ | |
1858 | ||
2104c8d9 | 1859 | UV |
7fc63493 | 1860 | Perl_to_utf8_lower(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 | 1861 | { |
97aff369 | 1862 | dVAR; |
7918f24d NC |
1863 | |
1864 | PERL_ARGS_ASSERT_TO_UTF8_LOWER; | |
1865 | ||
983ffd37 | 1866 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 JH |
1867 | &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower"); |
1868 | } | |
1869 | ||
d3e79532 | 1870 | /* |
87cea99e | 1871 | =for apidoc to_utf8_fold |
d3e79532 JH |
1872 | |
1873 | Convert the UTF-8 encoded character at p to its foldcase version and | |
1874 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note | |
89ebb4a3 | 1875 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
d3e79532 JH |
1876 | foldcase version may be longer than the original character (up to |
1877 | three characters). | |
1878 | ||
1879 | The first character of the foldcased version is returned | |
1880 | (but note, as explained above, that there may be more.) | |
1881 | ||
1882 | =cut */ | |
1883 | ||
b4e400f9 | 1884 | UV |
7fc63493 | 1885 | Perl_to_utf8_fold(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
b4e400f9 | 1886 | { |
97aff369 | 1887 | dVAR; |
7918f24d NC |
1888 | |
1889 | PERL_ARGS_ASSERT_TO_UTF8_FOLD; | |
1890 | ||
b4e400f9 JH |
1891 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
1892 | &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold"); | |
a0ed51b3 LW |
1893 | } |
1894 | ||
711a919c TS |
1895 | /* Note: |
1896 | * A "swash" is a swatch hash. | |
1897 | * A "swatch" is a bit vector generated by utf8.c:S_swash_get(). | |
1898 | * C<pkg> is a pointer to a package name for SWASHNEW, should be "utf8". | |
1899 | * For other parameters, see utf8::SWASHNEW in lib/utf8_heavy.pl. | |
1900 | */ | |
a0ed51b3 | 1901 | SV* |
7fc63493 | 1902 | Perl_swash_init(pTHX_ const char* pkg, const char* name, SV *listsv, I32 minbits, I32 none) |
a0ed51b3 | 1903 | { |
27da23d5 | 1904 | dVAR; |
a0ed51b3 | 1905 | SV* retval; |
8e84507e | 1906 | dSP; |
7fc63493 AL |
1907 | const size_t pkg_len = strlen(pkg); |
1908 | const size_t name_len = strlen(name); | |
da51bb9b | 1909 | HV * const stash = gv_stashpvn(pkg, pkg_len, 0); |
f8be5cf0 | 1910 | SV* errsv_save; |
ce3b816e | 1911 | |
7918f24d NC |
1912 | PERL_ARGS_ASSERT_SWASH_INIT; |
1913 | ||
96ca9f55 DM |
1914 | PUSHSTACKi(PERLSI_MAGIC); |
1915 | ENTER; | |
ec34a119 | 1916 | SAVEHINTS(); |
96ca9f55 | 1917 | save_re_context(); |
1b026014 | 1918 | if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */ |
ce3b816e | 1919 | ENTER; |
f8be5cf0 | 1920 | errsv_save = newSVsv(ERRSV); |
dc0c6abb NC |
1921 | /* It is assumed that callers of this routine are not passing in any |
1922 | user derived data. */ | |
1923 | /* Need to do this after save_re_context() as it will set PL_tainted to | |
1924 | 1 while saving $1 etc (see the code after getrx: in Perl_magic_get). | |
1925 | Even line to create errsv_save can turn on PL_tainted. */ | |
1926 | SAVEBOOL(PL_tainted); | |
1927 | PL_tainted = 0; | |
71bed85a | 1928 | Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpvn(pkg,pkg_len), |
a0714e2c | 1929 | NULL); |
f8be5cf0 JH |
1930 | if (!SvTRUE(ERRSV)) |
1931 | sv_setsv(ERRSV, errsv_save); | |
1932 | SvREFCNT_dec(errsv_save); | |
ce3b816e GS |
1933 | LEAVE; |
1934 | } | |
1935 | SPAGAIN; | |
a0ed51b3 LW |
1936 | PUSHMARK(SP); |
1937 | EXTEND(SP,5); | |
6e449a3a MHM |
1938 | mPUSHp(pkg, pkg_len); |
1939 | mPUSHp(name, name_len); | |
a0ed51b3 | 1940 | PUSHs(listsv); |
6e449a3a MHM |
1941 | mPUSHi(minbits); |
1942 | mPUSHi(none); | |
a0ed51b3 | 1943 | PUTBACK; |
f8be5cf0 | 1944 | errsv_save = newSVsv(ERRSV); |
864dbfa3 | 1945 | if (call_method("SWASHNEW", G_SCALAR)) |
8e84507e | 1946 | retval = newSVsv(*PL_stack_sp--); |
a0ed51b3 | 1947 | else |
e24b16f9 | 1948 | retval = &PL_sv_undef; |
f8be5cf0 JH |
1949 | if (!SvTRUE(ERRSV)) |
1950 | sv_setsv(ERRSV, errsv_save); | |
1951 | SvREFCNT_dec(errsv_save); | |
a0ed51b3 LW |
1952 | LEAVE; |
1953 | POPSTACK; | |
923e4eb5 | 1954 | if (IN_PERL_COMPILETIME) { |
623e6609 | 1955 | CopHINTS_set(PL_curcop, PL_hints); |
a0ed51b3 | 1956 | } |
bc45ce41 JH |
1957 | if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) { |
1958 | if (SvPOK(retval)) | |
35c1215d | 1959 | Perl_croak(aTHX_ "Can't find Unicode property definition \"%"SVf"\"", |
be2597df | 1960 | SVfARG(retval)); |
cea2e8a9 | 1961 | Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref"); |
bc45ce41 | 1962 | } |
a0ed51b3 LW |
1963 | return retval; |
1964 | } | |
1965 | ||
035d37be JH |
1966 | |
1967 | /* This API is wrong for special case conversions since we may need to | |
1968 | * return several Unicode characters for a single Unicode character | |
1969 | * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is | |
1970 | * the lower-level routine, and it is similarly broken for returning | |
1971 | * multiple values. --jhi */ | |
979f2922 | 1972 | /* Now SWASHGET is recasted into S_swash_get in this file. */ |
680c470c TS |
1973 | |
1974 | /* Note: | |
1975 | * Returns the value of property/mapping C<swash> for the first character | |
1976 | * of the string C<ptr>. If C<do_utf8> is true, the string C<ptr> is | |
1977 | * assumed to be in utf8. If C<do_utf8> is false, the string C<ptr> is | |
1978 | * assumed to be in native 8-bit encoding. Caches the swatch in C<swash>. | |
1979 | */ | |
a0ed51b3 | 1980 | UV |
680c470c | 1981 | Perl_swash_fetch(pTHX_ SV *swash, const U8 *ptr, bool do_utf8) |
a0ed51b3 | 1982 | { |
27da23d5 | 1983 | dVAR; |
ef8f7699 | 1984 | HV *const hv = MUTABLE_HV(SvRV(swash)); |
3568d838 JH |
1985 | U32 klen; |
1986 | U32 off; | |
a0ed51b3 | 1987 | STRLEN slen; |
7d85a32c | 1988 | STRLEN needents; |
cfd0369c | 1989 | const U8 *tmps = NULL; |
a0ed51b3 | 1990 | U32 bit; |
979f2922 | 1991 | SV *swatch; |
3568d838 | 1992 | U8 tmputf8[2]; |
35da51f7 | 1993 | const UV c = NATIVE_TO_ASCII(*ptr); |
3568d838 | 1994 | |
7918f24d NC |
1995 | PERL_ARGS_ASSERT_SWASH_FETCH; |
1996 | ||
3568d838 | 1997 | if (!do_utf8 && !UNI_IS_INVARIANT(c)) { |
979f2922 TS |
1998 | tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c); |
1999 | tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c); | |
2000 | ptr = tmputf8; | |
3568d838 JH |
2001 | } |
2002 | /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ | |
37e2e78e | 2003 | * then the "swatch" is a vec() for all the chars which start |
3568d838 JH |
2004 | * with 0xAA..0xYY |
2005 | * So the key in the hash (klen) is length of encoded char -1 | |
2006 | */ | |
2007 | klen = UTF8SKIP(ptr) - 1; | |
2008 | off = ptr[klen]; | |
a0ed51b3 | 2009 | |
979f2922 | 2010 | if (klen == 0) { |
37e2e78e | 2011 | /* If char is invariant then swatch is for all the invariant chars |
1e54db1a | 2012 | * In both UTF-8 and UTF-8-MOD that happens to be UTF_CONTINUATION_MARK |
7d85a32c | 2013 | */ |
979f2922 TS |
2014 | needents = UTF_CONTINUATION_MARK; |
2015 | off = NATIVE_TO_UTF(ptr[klen]); | |
2016 | } | |
2017 | else { | |
7d85a32c | 2018 | /* If char is encoded then swatch is for the prefix */ |
979f2922 TS |
2019 | needents = (1 << UTF_ACCUMULATION_SHIFT); |
2020 | off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK; | |
2021 | } | |
7d85a32c | 2022 | |
a0ed51b3 LW |
2023 | /* |
2024 | * This single-entry cache saves about 1/3 of the utf8 overhead in test | |
2025 | * suite. (That is, only 7-8% overall over just a hash cache. Still, | |
2026 | * it's nothing to sniff at.) Pity we usually come through at least | |
2027 | * two function calls to get here... | |
2028 | * | |
2029 | * NB: this code assumes that swatches are never modified, once generated! | |
2030 | */ | |
2031 | ||
3568d838 | 2032 | if (hv == PL_last_swash_hv && |
a0ed51b3 | 2033 | klen == PL_last_swash_klen && |
27da23d5 | 2034 | (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) ) |
a0ed51b3 LW |
2035 | { |
2036 | tmps = PL_last_swash_tmps; | |
2037 | slen = PL_last_swash_slen; | |
2038 | } | |
2039 | else { | |
2040 | /* Try our second-level swatch cache, kept in a hash. */ | |
e1ec3a88 | 2041 | SV** svp = hv_fetch(hv, (const char*)ptr, klen, FALSE); |
a0ed51b3 | 2042 | |
979f2922 TS |
2043 | /* If not cached, generate it via swash_get */ |
2044 | if (!svp || !SvPOK(*svp) | |
2045 | || !(tmps = (const U8*)SvPV_const(*svp, slen))) { | |
2b9d42f0 NIS |
2046 | /* We use utf8n_to_uvuni() as we want an index into |
2047 | Unicode tables, not a native character number. | |
2048 | */ | |
aec46f14 | 2049 | const UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXBYTES, 0, |
872c91ae JH |
2050 | ckWARN(WARN_UTF8) ? |
2051 | 0 : UTF8_ALLOW_ANY); | |
680c470c | 2052 | swatch = swash_get(swash, |
979f2922 TS |
2053 | /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */ |
2054 | (klen) ? (code_point & ~(needents - 1)) : 0, | |
2055 | needents); | |
2056 | ||
923e4eb5 | 2057 | if (IN_PERL_COMPILETIME) |
623e6609 | 2058 | CopHINTS_set(PL_curcop, PL_hints); |
a0ed51b3 | 2059 | |
979f2922 | 2060 | svp = hv_store(hv, (const char *)ptr, klen, swatch, 0); |
a0ed51b3 | 2061 | |
979f2922 TS |
2062 | if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) |
2063 | || (slen << 3) < needents) | |
660a4616 | 2064 | Perl_croak(aTHX_ "panic: swash_fetch got improper swatch"); |
a0ed51b3 LW |
2065 | } |
2066 | ||
2067 | PL_last_swash_hv = hv; | |
16d8f38a | 2068 | assert(klen <= sizeof(PL_last_swash_key)); |
eac04b2e | 2069 | PL_last_swash_klen = (U8)klen; |
cfd0369c NC |
2070 | /* FIXME change interpvar.h? */ |
2071 | PL_last_swash_tmps = (U8 *) tmps; | |
a0ed51b3 LW |
2072 | PL_last_swash_slen = slen; |
2073 | if (klen) | |
2074 | Copy(ptr, PL_last_swash_key, klen, U8); | |
2075 | } | |
2076 | ||
9faf8d75 | 2077 | switch ((int)((slen << 3) / needents)) { |
a0ed51b3 LW |
2078 | case 1: |
2079 | bit = 1 << (off & 7); | |
2080 | off >>= 3; | |
2081 | return (tmps[off] & bit) != 0; | |
2082 | case 8: | |
2083 | return tmps[off]; | |
2084 | case 16: | |
2085 | off <<= 1; | |
2086 | return (tmps[off] << 8) + tmps[off + 1] ; | |
2087 | case 32: | |
2088 | off <<= 2; | |
2089 | return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ; | |
2090 | } | |
660a4616 | 2091 | Perl_croak(aTHX_ "panic: swash_fetch got swatch of unexpected bit width"); |
670f1322 | 2092 | NORETURN_FUNCTION_END; |
a0ed51b3 | 2093 | } |
2b9d42f0 | 2094 | |
319009ee KW |
2095 | /* Read a single line of the main body of the swash input text. These are of |
2096 | * the form: | |
2097 | * 0053 0056 0073 | |
2098 | * where each number is hex. The first two numbers form the minimum and | |
2099 | * maximum of a range, and the third is the value associated with the range. | |
2100 | * Not all swashes should have a third number | |
2101 | * | |
2102 | * On input: l points to the beginning of the line to be examined; it points | |
2103 | * to somewhere in the string of the whole input text, and is | |
2104 | * terminated by a \n or the null string terminator. | |
2105 | * lend points to the null terminator of that string | |
2106 | * wants_value is non-zero if the swash expects a third number | |
2107 | * typestr is the name of the swash's mapping, like 'ToLower' | |
2108 | * On output: *min, *max, and *val are set to the values read from the line. | |
2109 | * returns a pointer just beyond the line examined. If there was no | |
2110 | * valid min number on the line, returns lend+1 | |
2111 | */ | |
2112 | ||
2113 | STATIC U8* | |
2114 | S_swash_scan_list_line(pTHX_ U8* l, U8* const lend, UV* min, UV* max, UV* val, | |
2115 | const bool wants_value, const U8* const typestr) | |
2116 | { | |
2117 | const int typeto = typestr[0] == 'T' && typestr[1] == 'o'; | |
2118 | STRLEN numlen; /* Length of the number */ | |
2119 | I32 flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX; | |
2120 | ||
2121 | /* nl points to the next \n in the scan */ | |
2122 | U8* const nl = (U8*)memchr(l, '\n', lend - l); | |
2123 | ||
2124 | /* Get the first number on the line: the range minimum */ | |
2125 | numlen = lend - l; | |
2126 | *min = grok_hex((char *)l, &numlen, &flags, NULL); | |
2127 | if (numlen) /* If found a hex number, position past it */ | |
2128 | l += numlen; | |
2129 | else if (nl) { /* Else, go handle next line, if any */ | |
2130 | return nl + 1; /* 1 is length of "\n" */ | |
2131 | } | |
2132 | else { /* Else, no next line */ | |
2133 | return lend + 1; /* to LIST's end at which \n is not found */ | |
2134 | } | |
2135 | ||
2136 | /* The max range value follows, separated by a BLANK */ | |
2137 | if (isBLANK(*l)) { | |
2138 | ++l; | |
2139 | flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX; | |
2140 | numlen = lend - l; | |
2141 | *max = grok_hex((char *)l, &numlen, &flags, NULL); | |
2142 | if (numlen) | |
2143 | l += numlen; | |
2144 | else /* If no value here, it is a single element range */ | |
2145 | *max = *min; | |
2146 | ||
2147 | /* Non-binary tables have a third entry: what the first element of the | |
2148 | * range maps to */ | |
2149 | if (wants_value) { | |
2150 | if (isBLANK(*l)) { | |
2151 | ++l; | |
2152 | flags = PERL_SCAN_SILENT_ILLDIGIT | | |
2153 | PERL_SCAN_DISALLOW_PREFIX; | |
2154 | numlen = lend - l; | |
2155 | *val = grok_hex((char *)l, &numlen, &flags, NULL); | |
2156 | if (numlen) | |
2157 | l += numlen; | |
2158 | else | |
2159 | *val = 0; | |
2160 | } | |
2161 | else { | |
2162 | *val = 0; | |
2163 | if (typeto) { | |
2164 | Perl_croak(aTHX_ "%s: illegal mapping '%s'", | |
2165 | typestr, l); | |
2166 | } | |
2167 | } | |
2168 | } | |
2169 | else | |
2170 | *val = 0; /* bits == 1, then any val should be ignored */ | |
2171 | } | |
2172 | else { /* Nothing following range min, should be single element with no | |
2173 | mapping expected */ | |
2174 | *max = *min; | |
2175 | if (wants_value) { | |
2176 | *val = 0; | |
2177 | if (typeto) { | |
2178 | Perl_croak(aTHX_ "%s: illegal mapping '%s'", typestr, l); | |
2179 | } | |
2180 | } | |
2181 | else | |
2182 | *val = 0; /* bits == 1, then val should be ignored */ | |
2183 | } | |
2184 | ||
2185 | /* Position to next line if any, or EOF */ | |
2186 | if (nl) | |
2187 | l = nl + 1; | |
2188 | else | |
2189 | l = lend; | |
2190 | ||
2191 | return l; | |
2192 | } | |
2193 | ||
979f2922 TS |
2194 | /* Note: |
2195 | * Returns a swatch (a bit vector string) for a code point sequence | |
2196 | * that starts from the value C<start> and comprises the number C<span>. | |
2197 | * A C<swash> must be an object created by SWASHNEW (see lib/utf8_heavy.pl). | |
2198 | * Should be used via swash_fetch, which will cache the swatch in C<swash>. | |
2199 | */ | |
2200 | STATIC SV* | |
2201 | S_swash_get(pTHX_ SV* swash, UV start, UV span) | |
2202 | { | |
2203 | SV *swatch; | |
711a919c | 2204 | U8 *l, *lend, *x, *xend, *s; |
979f2922 | 2205 | STRLEN lcur, xcur, scur; |
ef8f7699 | 2206 | HV *const hv = MUTABLE_HV(SvRV(swash)); |
972dd592 KW |
2207 | |
2208 | /* The string containing the main body of the table */ | |
017a3ce5 | 2209 | SV** const listsvp = hv_fetchs(hv, "LIST", FALSE); |
972dd592 | 2210 | |
017a3ce5 GA |
2211 | SV** const typesvp = hv_fetchs(hv, "TYPE", FALSE); |
2212 | SV** const bitssvp = hv_fetchs(hv, "BITS", FALSE); | |
2213 | SV** const nonesvp = hv_fetchs(hv, "NONE", FALSE); | |
2214 | SV** const extssvp = hv_fetchs(hv, "EXTRAS", FALSE); | |
0bd48802 | 2215 | const U8* const typestr = (U8*)SvPV_nolen(*typesvp); |
0bd48802 AL |
2216 | const STRLEN bits = SvUV(*bitssvp); |
2217 | const STRLEN octets = bits >> 3; /* if bits == 1, then octets == 0 */ | |
2218 | const UV none = SvUV(*nonesvp); | |
2219 | const UV end = start + span; | |
979f2922 | 2220 | |
7918f24d NC |
2221 | PERL_ARGS_ASSERT_SWASH_GET; |
2222 | ||
979f2922 | 2223 | if (bits != 1 && bits != 8 && bits != 16 && bits != 32) { |
660a4616 TS |
2224 | Perl_croak(aTHX_ "panic: swash_get doesn't expect bits %"UVuf, |
2225 | (UV)bits); | |
979f2922 TS |
2226 | } |
2227 | ||
2228 | /* create and initialize $swatch */ | |
979f2922 | 2229 | scur = octets ? (span * octets) : (span + 7) / 8; |
e524fe40 NC |
2230 | swatch = newSV(scur); |
2231 | SvPOK_on(swatch); | |
979f2922 TS |
2232 | s = (U8*)SvPVX(swatch); |
2233 | if (octets && none) { | |
0bd48802 | 2234 | const U8* const e = s + scur; |
979f2922 TS |
2235 | while (s < e) { |
2236 | if (bits == 8) | |
2237 | *s++ = (U8)(none & 0xff); | |
2238 | else if (bits == 16) { | |
2239 | *s++ = (U8)((none >> 8) & 0xff); | |
2240 | *s++ = (U8)( none & 0xff); | |
2241 | } | |
2242 | else if (bits == 32) { | |
2243 | *s++ = (U8)((none >> 24) & 0xff); | |
2244 | *s++ = (U8)((none >> 16) & 0xff); | |
2245 | *s++ = (U8)((none >> 8) & 0xff); | |
2246 | *s++ = (U8)( none & 0xff); | |
2247 | } | |
2248 | } | |
2249 | *s = '\0'; | |
2250 | } | |
2251 | else { | |
2252 | (void)memzero((U8*)s, scur + 1); | |
2253 | } | |
2254 | SvCUR_set(swatch, scur); | |
2255 | s = (U8*)SvPVX(swatch); | |
2256 | ||
2257 | /* read $swash->{LIST} */ | |
2258 | l = (U8*)SvPV(*listsvp, lcur); | |
2259 | lend = l + lcur; | |
2260 | while (l < lend) { | |
35da51f7 | 2261 | UV min, max, val; |
319009ee KW |
2262 | l = S_swash_scan_list_line(aTHX_ l, lend, &min, &max, &val, |
2263 | cBOOL(octets), typestr); | |
2264 | if (l > lend) { | |
979f2922 TS |
2265 | break; |
2266 | } | |
2267 | ||
972dd592 | 2268 | /* If looking for something beyond this range, go try the next one */ |
979f2922 TS |
2269 | if (max < start) |
2270 | continue; | |
2271 | ||
2272 | if (octets) { | |
35da51f7 | 2273 | UV key; |
979f2922 TS |
2274 | if (min < start) { |
2275 | if (!none || val < none) { | |
2276 | val += start - min; | |
2277 | } | |
2278 | min = start; | |
2279 | } | |
2280 | for (key = min; key <= max; key++) { | |
2281 | STRLEN offset; | |
2282 | if (key >= end) | |
2283 | goto go_out_list; | |
2284 | /* offset must be non-negative (start <= min <= key < end) */ | |
2285 | offset = octets * (key - start); | |
2286 | if (bits == 8) | |
2287 | s[offset] = (U8)(val & 0xff); | |
2288 | else if (bits == 16) { | |
2289 | s[offset ] = (U8)((val >> 8) & 0xff); | |
2290 | s[offset + 1] = (U8)( val & 0xff); | |
2291 | } | |
2292 | else if (bits == 32) { | |
2293 | s[offset ] = (U8)((val >> 24) & 0xff); | |
2294 | s[offset + 1] = (U8)((val >> 16) & 0xff); | |
2295 | s[offset + 2] = (U8)((val >> 8) & 0xff); | |
2296 | s[offset + 3] = (U8)( val & 0xff); | |
2297 | } | |
2298 | ||
2299 | if (!none || val < none) | |
2300 | ++val; | |
2301 | } | |
2302 | } | |
711a919c | 2303 | else { /* bits == 1, then val should be ignored */ |
35da51f7 | 2304 | UV key; |
979f2922 TS |
2305 | if (min < start) |
2306 | min = start; | |
2307 | for (key = min; key <= max; key++) { | |
0bd48802 | 2308 | const STRLEN offset = (STRLEN)(key - start); |
979f2922 TS |
2309 | if (key >= end) |
2310 | goto go_out_list; | |
2311 | s[offset >> 3] |= 1 << (offset & 7); | |
2312 | } | |
2313 | } | |
2314 | } /* while */ | |
2315 | go_out_list: | |
2316 | ||
2317 | /* read $swash->{EXTRAS} */ | |
2318 | x = (U8*)SvPV(*extssvp, xcur); | |
2319 | xend = x + xcur; | |
2320 | while (x < xend) { | |
2321 | STRLEN namelen; | |
2322 | U8 *namestr; | |
2323 | SV** othersvp; | |
2324 | HV* otherhv; | |
2325 | STRLEN otherbits; | |
2326 | SV **otherbitssvp, *other; | |
711a919c | 2327 | U8 *s, *o, *nl; |
979f2922 TS |
2328 | STRLEN slen, olen; |
2329 | ||
35da51f7 | 2330 | const U8 opc = *x++; |
979f2922 TS |
2331 | if (opc == '\n') |
2332 | continue; | |
2333 | ||
2334 | nl = (U8*)memchr(x, '\n', xend - x); | |
2335 | ||
2336 | if (opc != '-' && opc != '+' && opc != '!' && opc != '&') { | |
2337 | if (nl) { | |
2338 | x = nl + 1; /* 1 is length of "\n" */ | |
2339 | continue; | |
2340 | } | |
2341 | else { | |
2342 | x = xend; /* to EXTRAS' end at which \n is not found */ | |
2343 | break; | |
2344 | } | |
2345 | } | |
2346 | ||
2347 | namestr = x; | |
2348 | if (nl) { | |
2349 | namelen = nl - namestr; | |
2350 | x = nl + 1; | |
2351 | } | |
2352 | else { | |
2353 | namelen = xend - namestr; | |
2354 | x = xend; | |
2355 | } | |
2356 | ||
2357 | othersvp = hv_fetch(hv, (char *)namestr, namelen, FALSE); | |
ef8f7699 | 2358 | otherhv = MUTABLE_HV(SvRV(*othersvp)); |
017a3ce5 | 2359 | otherbitssvp = hv_fetchs(otherhv, "BITS", FALSE); |
979f2922 TS |
2360 | otherbits = (STRLEN)SvUV(*otherbitssvp); |
2361 | if (bits < otherbits) | |
660a4616 | 2362 | Perl_croak(aTHX_ "panic: swash_get found swatch size mismatch"); |
979f2922 TS |
2363 | |
2364 | /* The "other" swatch must be destroyed after. */ | |
2365 | other = swash_get(*othersvp, start, span); | |
2366 | o = (U8*)SvPV(other, olen); | |
2367 | ||
2368 | if (!olen) | |
660a4616 | 2369 | Perl_croak(aTHX_ "panic: swash_get got improper swatch"); |
979f2922 TS |
2370 | |
2371 | s = (U8*)SvPV(swatch, slen); | |
2372 | if (bits == 1 && otherbits == 1) { | |
2373 | if (slen != olen) | |
660a4616 | 2374 | Perl_croak(aTHX_ "panic: swash_get found swatch length mismatch"); |
979f2922 TS |
2375 | |
2376 | switch (opc) { | |
2377 | case '+': | |
2378 | while (slen--) | |
2379 | *s++ |= *o++; | |
2380 | break; | |
2381 | case '!': | |
2382 | while (slen--) | |
2383 | *s++ |= ~*o++; | |
2384 | break; | |
2385 | case '-': | |
2386 | while (slen--) | |
2387 | *s++ &= ~*o++; | |
2388 | break; | |
2389 | case '&': | |
2390 | while (slen--) | |
2391 | *s++ &= *o++; | |
2392 | break; | |
2393 | default: | |
2394 | break; | |
2395 | } | |
2396 | } | |
711a919c | 2397 | else { |
979f2922 TS |
2398 | STRLEN otheroctets = otherbits >> 3; |
2399 | STRLEN offset = 0; | |
35da51f7 | 2400 | U8* const send = s + slen; |
979f2922 TS |
2401 | |
2402 | while (s < send) { | |
2403 | UV otherval = 0; | |
2404 | ||
2405 | if (otherbits == 1) { | |
2406 | otherval = (o[offset >> 3] >> (offset & 7)) & 1; | |
2407 | ++offset; | |
2408 | } | |
2409 | else { | |
2410 | STRLEN vlen = otheroctets; | |
2411 | otherval = *o++; | |
2412 | while (--vlen) { | |
2413 | otherval <<= 8; | |
2414 | otherval |= *o++; | |
2415 | } | |
2416 | } | |
2417 | ||
711a919c | 2418 | if (opc == '+' && otherval) |
6f207bd3 | 2419 | NOOP; /* replace with otherval */ |
979f2922 TS |
2420 | else if (opc == '!' && !otherval) |
2421 | otherval = 1; | |
2422 | else if (opc == '-' && otherval) | |
2423 | otherval = 0; | |
2424 | else if (opc == '&' && !otherval) | |
2425 | otherval = 0; | |
2426 | else { | |
711a919c | 2427 | s += octets; /* no replacement */ |
979f2922 TS |
2428 | continue; |
2429 | } | |
2430 | ||
2431 | if (bits == 8) | |
2432 | *s++ = (U8)( otherval & 0xff); | |
2433 | else if (bits == 16) { | |
2434 | *s++ = (U8)((otherval >> 8) & 0xff); | |
2435 | *s++ = (U8)( otherval & 0xff); | |
2436 | } | |
2437 | else if (bits == 32) { | |
2438 | *s++ = (U8)((otherval >> 24) & 0xff); | |
2439 | *s++ = (U8)((otherval >> 16) & 0xff); | |
2440 | *s++ = (U8)((otherval >> 8) & 0xff); | |
2441 | *s++ = (U8)( otherval & 0xff); | |
2442 | } | |
2443 | } | |
2444 | } | |
2445 | sv_free(other); /* through with it! */ | |
2446 | } /* while */ | |
2447 | return swatch; | |
2448 | } | |
2449 | ||
064c021d KW |
2450 | HV* |
2451 | Perl__swash_inversion_hash(pTHX_ SV* swash) | |
2452 | { | |
2453 | ||
2454 | /* Subject to change or removal. For use only in one place in regexec.c | |
2455 | * | |
2456 | * Returns a hash which is the inversion and closure of a swash mapping. | |
2457 | * For example, consider the input lines: | |
2458 | * 004B 006B | |
2459 | * 004C 006C | |
2460 | * 212A 006B | |
2461 | * | |
2462 | * The returned hash would have two keys, the utf8 for 006B and the utf8 for | |
2463 | * 006C. The value for each key is an array. For 006C, the array would | |
2464 | * have a two elements, the utf8 for itself, and for 004C. For 006B, there | |
2465 | * would be three elements in its array, the utf8 for 006B, 004B and 212A. | |
2466 | * | |
2467 | * Essentially, for any code point, it gives all the code points that map to | |
2468 | * it, or the list of 'froms' for that point. | |
2469 | * | |
2470 | * Currently it only looks at the main body of the swash, and ignores any | |
2471 | * additions or deletions from other swashes */ | |
2472 | ||
2473 | U8 *l, *lend; | |
2474 | STRLEN lcur; | |
2475 | HV *const hv = MUTABLE_HV(SvRV(swash)); | |
2476 | ||
2477 | /* The string containing the main body of the table */ | |
2478 | SV** const listsvp = hv_fetchs(hv, "LIST", FALSE); | |
2479 | ||
2480 | SV** const typesvp = hv_fetchs(hv, "TYPE", FALSE); | |
2481 | SV** const bitssvp = hv_fetchs(hv, "BITS", FALSE); | |
2482 | SV** const nonesvp = hv_fetchs(hv, "NONE", FALSE); | |
2483 | /*SV** const extssvp = hv_fetchs(hv, "EXTRAS", FALSE);*/ | |
2484 | const U8* const typestr = (U8*)SvPV_nolen(*typesvp); | |
2485 | const STRLEN bits = SvUV(*bitssvp); | |
2486 | const STRLEN octets = bits >> 3; /* if bits == 1, then octets == 0 */ | |
2487 | const UV none = SvUV(*nonesvp); | |
2488 | ||
2489 | HV* ret = newHV(); | |
2490 | ||
2491 | PERL_ARGS_ASSERT__SWASH_INVERSION_HASH; | |
2492 | ||
2493 | /* Must have at least 8 bits to get the mappings */ | |
2494 | if (bits != 8 && bits != 16 && bits != 32) { | |
2495 | Perl_croak(aTHX_ "panic: swash_inversion_hash doesn't expect bits %"UVuf, | |
2496 | (UV)bits); | |
2497 | } | |
2498 | ||
2499 | /* read $swash->{LIST} */ | |
2500 | l = (U8*)SvPV(*listsvp, lcur); | |
2501 | lend = l + lcur; | |
2502 | ||
2503 | /* Go through each input line */ | |
2504 | while (l < lend) { | |
2505 | UV min, max, val; | |
2506 | UV inverse; | |
2507 | l = S_swash_scan_list_line(aTHX_ l, lend, &min, &max, &val, | |
2508 | cBOOL(octets), typestr); | |
2509 | if (l > lend) { | |
2510 | break; | |
2511 | } | |
2512 | ||
2513 | /* Each element in the range is to be inverted */ | |
2514 | for (inverse = min; inverse <= max; inverse++) { | |
2515 | AV* list; | |
2516 | SV* element; | |
2517 | SV** listp; | |
2518 | IV i; | |
2519 | bool found_key = FALSE; | |
2520 | ||
2521 | /* The key is the inverse mapping */ | |
2522 | char key[UTF8_MAXBYTES+1]; | |
2523 | char* key_end = (char *) uvuni_to_utf8((U8*) key, val); | |
2524 | STRLEN key_len = key_end - key; | |
2525 | ||
2526 | /* And the value is what the forward mapping is from. */ | |
2527 | char utf8_inverse[UTF8_MAXBYTES+1]; | |
2528 | char *utf8_inverse_end = (char *) uvuni_to_utf8((U8*) utf8_inverse, inverse); | |
2529 | ||
2530 | /* Get the list for the map */ | |
2531 | if ((listp = hv_fetch(ret, key, key_len, FALSE))) { | |
2532 | list = (AV*) *listp; | |
2533 | } | |
2534 | else { /* No entry yet for it: create one */ | |
2535 | list = newAV(); | |
2536 | if (! hv_store(ret, key, key_len, (SV*) list, FALSE)) { | |
2537 | Perl_croak(aTHX_ "panic: hv_store() unexpectedly failed"); | |
2538 | } | |
2539 | } | |
2540 | ||
2541 | for (i = 0; i < av_len(list); i++) { | |
2542 | SV** entryp = av_fetch(list, i, FALSE); | |
2543 | SV* entry; | |
2544 | if (entryp == NULL) { | |
2545 | Perl_croak(aTHX_ "panic: av_fetch() unexpectedly failed"); | |
2546 | } | |
2547 | entry = *entryp; | |
2548 | if (SvCUR(entry) != key_len) { | |
2549 | continue; | |
2550 | } | |
2551 | if (memEQ(key, SvPVX(entry), key_len)) { | |
2552 | found_key = TRUE; | |
2553 | break; | |
2554 | } | |
2555 | } | |
2556 | if (! found_key) { | |
2557 | element = newSVpvn_flags(key, key_len, SVf_UTF8); | |
2558 | av_push(list, element); | |
2559 | } | |
2560 | ||
2561 | ||
2562 | /* Simply add the value to the list */ | |
2563 | element = newSVpvn_flags(utf8_inverse, utf8_inverse_end - utf8_inverse, SVf_UTF8); | |
2564 | av_push(list, element); | |
2565 | ||
2566 | /* swash_get() increments the value of val for each element in the | |
2567 | * range. That makes more compact tables possible. You can | |
2568 | * express the capitalization, for example, of all consecutive | |
2569 | * letters with a single line: 0061\t007A\t0041 This maps 0061 to | |
2570 | * 0041, 0062 to 0042, etc. I (khw) have never understood 'none', | |
2571 | * and it's not documented, and perhaps not even currently used, | |
2572 | * but I copied the semantics from swash_get(), just in case */ | |
2573 | if (!none || val < none) { | |
2574 | ++val; | |
2575 | } | |
2576 | } | |
2577 | } | |
2578 | ||
2579 | return ret; | |
2580 | } | |
2581 | ||
0f830e0b | 2582 | /* |
87cea99e | 2583 | =for apidoc uvchr_to_utf8 |
0f830e0b NC |
2584 | |
2585 | Adds the UTF-8 representation of the Native codepoint C<uv> to the end | |
2586 | of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free | |
2587 | bytes available. The return value is the pointer to the byte after the | |
2588 | end of the new character. In other words, | |
2589 | ||
2590 | d = uvchr_to_utf8(d, uv); | |
2591 | ||
2592 | is the recommended wide native character-aware way of saying | |
2593 | ||
2594 | *(d++) = uv; | |
2595 | ||
2596 | =cut | |
2597 | */ | |
2598 | ||
2599 | /* On ASCII machines this is normally a macro but we want a | |
2600 | real function in case XS code wants it | |
2601 | */ | |
0f830e0b NC |
2602 | U8 * |
2603 | Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv) | |
2604 | { | |
7918f24d NC |
2605 | PERL_ARGS_ASSERT_UVCHR_TO_UTF8; |
2606 | ||
0f830e0b NC |
2607 | return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0); |
2608 | } | |
2609 | ||
b851fbc1 JH |
2610 | U8 * |
2611 | Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) | |
2612 | { | |
7918f24d NC |
2613 | PERL_ARGS_ASSERT_UVCHR_TO_UTF8_FLAGS; |
2614 | ||
b851fbc1 JH |
2615 | return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags); |
2616 | } | |
2b9d42f0 NIS |
2617 | |
2618 | /* | |
87cea99e | 2619 | =for apidoc utf8n_to_uvchr |
0f830e0b NC |
2620 | flags |
2621 | ||
48ef279e | 2622 | Returns the native character value of the first character in the string |
0f830e0b NC |
2623 | C<s> |
2624 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the | |
2625 | length, in bytes, of that character. | |
2626 | ||
2627 | Allows length and flags to be passed to low level routine. | |
2628 | ||
2629 | =cut | |
2630 | */ | |
2631 | /* On ASCII machines this is normally a macro but we want | |
2632 | a real function in case XS code wants it | |
2633 | */ | |
0f830e0b | 2634 | UV |
48ef279e | 2635 | Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, |
0f830e0b NC |
2636 | U32 flags) |
2637 | { | |
2638 | const UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags); | |
7918f24d NC |
2639 | |
2640 | PERL_ARGS_ASSERT_UTF8N_TO_UVCHR; | |
2641 | ||
0f830e0b NC |
2642 | return UNI_TO_NATIVE(uv); |
2643 | } | |
2644 | ||
2645 | /* | |
87cea99e | 2646 | =for apidoc pv_uni_display |
d2cc3551 JH |
2647 | |
2648 | Build to the scalar dsv a displayable version of the string spv, | |
2649 | length len, the displayable version being at most pvlim bytes long | |
2650 | (if longer, the rest is truncated and "..." will be appended). | |
0a2ef054 | 2651 | |
9e55ce06 | 2652 | The flags argument can have UNI_DISPLAY_ISPRINT set to display |
00e86452 | 2653 | isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH |
0a2ef054 JH |
2654 | to display the \\[nrfta\\] as the backslashed versions (like '\n') |
2655 | (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\). | |
2656 | UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both | |
2657 | UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on. | |
2658 | ||
d2cc3551 JH |
2659 | The pointer to the PV of the dsv is returned. |
2660 | ||
2661 | =cut */ | |
e6b2e755 | 2662 | char * |
e1ec3a88 | 2663 | Perl_pv_uni_display(pTHX_ SV *dsv, const U8 *spv, STRLEN len, STRLEN pvlim, UV flags) |
e6b2e755 JH |
2664 | { |
2665 | int truncated = 0; | |
e1ec3a88 | 2666 | const char *s, *e; |
e6b2e755 | 2667 | |
7918f24d NC |
2668 | PERL_ARGS_ASSERT_PV_UNI_DISPLAY; |
2669 | ||
76f68e9b | 2670 | sv_setpvs(dsv, ""); |
7fddd944 | 2671 | SvUTF8_off(dsv); |
e1ec3a88 | 2672 | for (s = (const char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) { |
e6b2e755 | 2673 | UV u; |
a49f32c6 NC |
2674 | /* This serves double duty as a flag and a character to print after |
2675 | a \ when flags & UNI_DISPLAY_BACKSLASH is true. | |
2676 | */ | |
2677 | char ok = 0; | |
c728cb41 | 2678 | |
e6b2e755 JH |
2679 | if (pvlim && SvCUR(dsv) >= pvlim) { |
2680 | truncated++; | |
2681 | break; | |
2682 | } | |
2683 | u = utf8_to_uvchr((U8*)s, 0); | |
c728cb41 | 2684 | if (u < 256) { |
a3b680e6 | 2685 | const unsigned char c = (unsigned char)u & 0xFF; |
0bd48802 | 2686 | if (flags & UNI_DISPLAY_BACKSLASH) { |
a49f32c6 | 2687 | switch (c) { |
c728cb41 | 2688 | case '\n': |
a49f32c6 | 2689 | ok = 'n'; break; |
c728cb41 | 2690 | case '\r': |
a49f32c6 | 2691 | ok = 'r'; break; |
c728cb41 | 2692 | case '\t': |
a49f32c6 | 2693 | ok = 't'; break; |
c728cb41 | 2694 | case '\f': |
a49f32c6 | 2695 | ok = 'f'; break; |
c728cb41 | 2696 | case '\a': |
a49f32c6 | 2697 | ok = 'a'; break; |
c728cb41 | 2698 | case '\\': |
a49f32c6 | 2699 | ok = '\\'; break; |
c728cb41 JH |
2700 | default: break; |
2701 | } | |
a49f32c6 | 2702 | if (ok) { |
88c9ea1e | 2703 | const char string = ok; |
76f68e9b | 2704 | sv_catpvs(dsv, "\\"); |
5e7aa789 | 2705 | sv_catpvn(dsv, &string, 1); |
a49f32c6 | 2706 | } |
c728cb41 | 2707 | } |
00e86452 | 2708 | /* isPRINT() is the locale-blind version. */ |
a49f32c6 | 2709 | if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(c)) { |
88c9ea1e | 2710 | const char string = c; |
5e7aa789 | 2711 | sv_catpvn(dsv, &string, 1); |
a49f32c6 | 2712 | ok = 1; |
0a2ef054 | 2713 | } |
c728cb41 JH |
2714 | } |
2715 | if (!ok) | |
9e55ce06 | 2716 | Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u); |
e6b2e755 JH |
2717 | } |
2718 | if (truncated) | |
396482e1 | 2719 | sv_catpvs(dsv, "..."); |
48ef279e | 2720 | |
e6b2e755 JH |
2721 | return SvPVX(dsv); |
2722 | } | |
2b9d42f0 | 2723 | |
d2cc3551 | 2724 | /* |
87cea99e | 2725 | =for apidoc sv_uni_display |
d2cc3551 JH |
2726 | |
2727 | Build to the scalar dsv a displayable version of the scalar sv, | |
0a2ef054 | 2728 | the displayable version being at most pvlim bytes long |
d2cc3551 | 2729 | (if longer, the rest is truncated and "..." will be appended). |
0a2ef054 JH |
2730 | |
2731 | The flags argument is as in pv_uni_display(). | |
2732 | ||
d2cc3551 JH |
2733 | The pointer to the PV of the dsv is returned. |
2734 | ||
d4c19fe8 AL |
2735 | =cut |
2736 | */ | |
e6b2e755 JH |
2737 | char * |
2738 | Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags) | |
2739 | { | |
7918f24d NC |
2740 | PERL_ARGS_ASSERT_SV_UNI_DISPLAY; |
2741 | ||
cfd0369c NC |
2742 | return Perl_pv_uni_display(aTHX_ dsv, (const U8*)SvPVX_const(ssv), |
2743 | SvCUR(ssv), pvlim, flags); | |
701a277b JH |
2744 | } |
2745 | ||
d2cc3551 | 2746 | /* |
e6226b18 | 2747 | =for apidoc foldEQ_utf8 |
d2cc3551 | 2748 | |
d51c1b21 | 2749 | Returns true if the leading portions of the strings s1 and s2 (either or both |
e6226b18 | 2750 | of which may be in UTF-8) are the same case-insensitively; false otherwise. |
d51c1b21 | 2751 | How far into the strings to compare is determined by other input parameters. |
8b35872c KW |
2752 | |
2753 | If u1 is true, the string s1 is assumed to be in UTF-8-encoded Unicode; | |
2754 | otherwise it is assumed to be in native 8-bit encoding. Correspondingly for u2 | |
2755 | with respect to s2. | |
2756 | ||
d51c1b21 KW |
2757 | If the byte length l1 is non-zero, it says how far into s1 to check for fold |
2758 | equality. In other words, s1+l1 will be used as a goal to reach. The | |
8b35872c KW |
2759 | scan will not be considered to be a match unless the goal is reached, and |
2760 | scanning won't continue past that goal. Correspondingly for l2 with respect to | |
2761 | s2. | |
2762 | ||
2763 | If pe1 is non-NULL and the pointer it points to is not NULL, that pointer is | |
2764 | considered an end pointer beyond which scanning of s1 will not continue under | |
2765 | any circumstances. This means that if both l1 and pe1 are specified, and pe1 | |
2766 | is less than s1+l1, the match will never be successful because it can never | |
d51c1b21 KW |
2767 | get as far as its goal (and in fact is asserted against). Correspondingly for |
2768 | pe2 with respect to s2. | |
8b35872c | 2769 | |
d51c1b21 KW |
2770 | At least one of s1 and s2 must have a goal (at least one of l1 and l2 must be |
2771 | non-zero), and if both do, both have to be | |
8b35872c KW |
2772 | reached for a successful match. Also, if the fold of a character is multiple |
2773 | characters, all of them must be matched (see tr21 reference below for | |
2774 | 'folding'). | |
2775 | ||
e6226b18 | 2776 | Upon a successful match, if pe1 is non-NULL, |
8b35872c KW |
2777 | it will be set to point to the beginning of the I<next> character of s1 beyond |
2778 | what was matched. Correspondingly for pe2 and s2. | |
d2cc3551 JH |
2779 | |
2780 | For case-insensitiveness, the "casefolding" of Unicode is used | |
2781 | instead of upper/lowercasing both the characters, see | |
2782 | http://www.unicode.org/unicode/reports/tr21/ (Case Mappings). | |
2783 | ||
2784 | =cut */ | |
701a277b | 2785 | I32 |
e6226b18 | 2786 | Perl_foldEQ_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2) |
332ddc25 | 2787 | { |
8b35872c KW |
2788 | dVAR; |
2789 | register const U8 *p1 = (const U8*)s1; /* Point to current char */ | |
2790 | register const U8 *p2 = (const U8*)s2; | |
48ef279e | 2791 | register const U8 *g1 = NULL; /* goal for s1 */ |
8b35872c | 2792 | register const U8 *g2 = NULL; |
48ef279e KW |
2793 | register const U8 *e1 = NULL; /* Don't scan s1 past this */ |
2794 | register U8 *f1 = NULL; /* Point to current folded */ | |
8b35872c KW |
2795 | register const U8 *e2 = NULL; |
2796 | register U8 *f2 = NULL; | |
48ef279e | 2797 | STRLEN n1 = 0, n2 = 0; /* Number of bytes in current char */ |
8b35872c KW |
2798 | U8 foldbuf1[UTF8_MAXBYTES_CASE+1]; |
2799 | U8 foldbuf2[UTF8_MAXBYTES_CASE+1]; | |
48ef279e KW |
2800 | U8 natbuf[2]; /* Holds native 8-bit char converted to utf8; |
2801 | these always fit in 2 bytes */ | |
8b35872c | 2802 | |
e6226b18 | 2803 | PERL_ARGS_ASSERT_FOLDEQ_UTF8; |
8b35872c KW |
2804 | |
2805 | if (pe1) { | |
48ef279e | 2806 | e1 = *(U8**)pe1; |
8b35872c KW |
2807 | } |
2808 | ||
2809 | if (l1) { | |
48ef279e | 2810 | g1 = (const U8*)s1 + l1; |
8b35872c KW |
2811 | } |
2812 | ||
2813 | if (pe2) { | |
48ef279e | 2814 | e2 = *(U8**)pe2; |
8b35872c KW |
2815 | } |
2816 | ||
2817 | if (l2) { | |
48ef279e | 2818 | g2 = (const U8*)s2 + l2; |
8b35872c KW |
2819 | } |
2820 | ||
2821 | /* Must have at least one goal */ | |
2822 | assert(g1 || g2); | |
2823 | ||
2824 | if (g1) { | |
2825 | ||
48ef279e KW |
2826 | /* Will never match if goal is out-of-bounds */ |
2827 | assert(! e1 || e1 >= g1); | |
8b35872c | 2828 | |
48ef279e KW |
2829 | /* Here, there isn't an end pointer, or it is beyond the goal. We |
2830 | * only go as far as the goal */ | |
2831 | e1 = g1; | |
8b35872c | 2832 | } |
313b38e5 NC |
2833 | else { |
2834 | assert(e1); /* Must have an end for looking at s1 */ | |
2835 | } | |
8b35872c KW |
2836 | |
2837 | /* Same for goal for s2 */ | |
2838 | if (g2) { | |
48ef279e KW |
2839 | assert(! e2 || e2 >= g2); |
2840 | e2 = g2; | |
8b35872c | 2841 | } |
313b38e5 NC |
2842 | else { |
2843 | assert(e2); | |
2844 | } | |
8b35872c KW |
2845 | |
2846 | /* Look through both strings, a character at a time */ | |
2847 | while (p1 < e1 && p2 < e2) { | |
2848 | ||
d51c1b21 KW |
2849 | /* If at the beginning of a new character in s1, get its fold to use |
2850 | * and the length of the fold */ | |
48ef279e KW |
2851 | if (n1 == 0) { |
2852 | if (u1) { | |
2853 | to_utf8_fold(p1, foldbuf1, &n1); | |
2854 | } | |
2855 | else { /* Not utf8, convert to it first and then get fold */ | |
2856 | uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p1))); | |
2857 | to_utf8_fold(natbuf, foldbuf1, &n1); | |
2858 | } | |
2859 | f1 = foldbuf1; | |
2860 | } | |
8b35872c | 2861 | |
48ef279e KW |
2862 | if (n2 == 0) { /* Same for s2 */ |
2863 | if (u2) { | |
2864 | to_utf8_fold(p2, foldbuf2, &n2); | |
2865 | } | |
2866 | else { | |
2867 | uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p2))); | |
2868 | to_utf8_fold(natbuf, foldbuf2, &n2); | |
2869 | } | |
2870 | f2 = foldbuf2; | |
2871 | } | |
8b35872c | 2872 | |
48ef279e KW |
2873 | /* While there is more to look for in both folds, see if they |
2874 | * continue to match */ | |
2875 | while (n1 && n2) { | |
2876 | U8 fold_length = UTF8SKIP(f1); | |
2877 | if (fold_length != UTF8SKIP(f2) | |
2878 | || (fold_length == 1 && *f1 != *f2) /* Short circuit memNE | |
2879 | function call for single | |
2880 | character */ | |
2881 | || memNE((char*)f1, (char*)f2, fold_length)) | |
2882 | { | |
e6226b18 | 2883 | return 0; /* mismatch */ |
48ef279e KW |
2884 | } |
2885 | ||
2886 | /* Here, they matched, advance past them */ | |
2887 | n1 -= fold_length; | |
2888 | f1 += fold_length; | |
2889 | n2 -= fold_length; | |
2890 | f2 += fold_length; | |
2891 | } | |
8b35872c | 2892 | |
48ef279e KW |
2893 | /* When reach the end of any fold, advance the input past it */ |
2894 | if (n1 == 0) { | |
2895 | p1 += u1 ? UTF8SKIP(p1) : 1; | |
2896 | } | |
2897 | if (n2 == 0) { | |
2898 | p2 += u2 ? UTF8SKIP(p2) : 1; | |
2899 | } | |
8b35872c KW |
2900 | } /* End of loop through both strings */ |
2901 | ||
2902 | /* A match is defined by each scan that specified an explicit length | |
2903 | * reaching its final goal, and the other not having matched a partial | |
2904 | * character (which can happen when the fold of a character is more than one | |
2905 | * character). */ | |
2906 | if (! ((g1 == 0 || p1 == g1) && (g2 == 0 || p2 == g2)) || n1 || n2) { | |
e6226b18 | 2907 | return 0; |
8b35872c KW |
2908 | } |
2909 | ||
2910 | /* Successful match. Set output pointers */ | |
2911 | if (pe1) { | |
48ef279e | 2912 | *pe1 = (char*)p1; |
8b35872c KW |
2913 | } |
2914 | if (pe2) { | |
48ef279e | 2915 | *pe2 = (char*)p2; |
8b35872c | 2916 | } |
e6226b18 | 2917 | return 1; |
e6b2e755 | 2918 | } |
701a277b | 2919 | |
a49f32c6 NC |
2920 | /* |
2921 | * Local variables: | |
2922 | * c-indentation-style: bsd | |
2923 | * c-basic-offset: 4 | |
2924 | * indent-tabs-mode: t | |
2925 | * End: | |
2926 | * | |
37442d52 RGS |
2927 | * ex: set ts=8 sts=4 sw=4 noet: |
2928 | */ |