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 JH |
35 | #ifndef EBCDIC |
36 | /* Separate prototypes needed because in ASCII systems these | |
37 | * usually macros but they still are compiled as code, too. */ | |
38 | PERL_CALLCONV UV Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags); | |
39 | PERL_CALLCONV U8* Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv); | |
40 | #endif | |
41 | ||
27da23d5 JH |
42 | static const char unees[] = |
43 | "Malformed UTF-8 character (unexpected end of string)"; | |
901b21bf | 44 | |
ccfc67b7 JH |
45 | /* |
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 | ||
60 | Returns true if first C<len> bytes of the given string are ASCII (i.e. none | |
61 | of them even raise the question of UTF-8-ness). | |
62 | ||
63 | See also is_utf8_string(), is_utf8_string_loclen(), and is_utf8_string_loc(). | |
64 | ||
65 | =cut | |
66 | */ | |
67 | ||
68 | bool | |
668b6d8d | 69 | Perl_is_ascii_string(const U8 *s, STRLEN len) |
eaf7a4d2 CS |
70 | { |
71 | const U8* const send = s + (len ? len : strlen((const char *)s)); | |
72 | const U8* x = s; | |
73 | ||
74 | PERL_ARGS_ASSERT_IS_ASCII_STRING; | |
eaf7a4d2 CS |
75 | |
76 | for (; x < send; ++x) { | |
77 | if (!UTF8_IS_INVARIANT(*x)) | |
78 | break; | |
79 | } | |
80 | ||
81 | return x == send; | |
82 | } | |
83 | ||
84 | /* | |
87cea99e | 85 | =for apidoc uvuni_to_utf8_flags |
eebe1485 | 86 | |
1e54db1a | 87 | Adds the UTF-8 representation of the Unicode codepoint C<uv> to the end |
89ebb4a3 | 88 | of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free |
eebe1485 | 89 | bytes available. The return value is the pointer to the byte after the |
9041c2e3 | 90 | end of the new character. In other words, |
eebe1485 | 91 | |
b851fbc1 JH |
92 | d = uvuni_to_utf8_flags(d, uv, flags); |
93 | ||
94 | or, in most cases, | |
95 | ||
9041c2e3 | 96 | d = uvuni_to_utf8(d, uv); |
eebe1485 | 97 | |
b851fbc1 JH |
98 | (which is equivalent to) |
99 | ||
100 | d = uvuni_to_utf8_flags(d, uv, 0); | |
101 | ||
eebe1485 SC |
102 | is the recommended Unicode-aware way of saying |
103 | ||
104 | *(d++) = uv; | |
105 | ||
106 | =cut | |
107 | */ | |
108 | ||
dfe13c55 | 109 | U8 * |
b851fbc1 | 110 | Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) |
a0ed51b3 | 111 | { |
7918f24d NC |
112 | PERL_ARGS_ASSERT_UVUNI_TO_UTF8_FLAGS; |
113 | ||
62961d2e | 114 | if (ckWARN(WARN_UTF8)) { |
b851fbc1 JH |
115 | if (UNICODE_IS_SURROGATE(uv) && |
116 | !(flags & UNICODE_ALLOW_SURROGATE)) | |
9014280d | 117 | Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv); |
b851fbc1 JH |
118 | else if ( |
119 | ((uv >= 0xFDD0 && uv <= 0xFDEF && | |
120 | !(flags & UNICODE_ALLOW_FDD0)) | |
121 | || | |
c867b360 | 122 | ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */ |
b851fbc1 JH |
123 | !(flags & UNICODE_ALLOW_FFFF))) && |
124 | /* UNICODE_ALLOW_SUPER includes | |
2a20b9da | 125 | * FFFEs and FFFFs beyond 0x10FFFF. */ |
b851fbc1 JH |
126 | ((uv <= PERL_UNICODE_MAX) || |
127 | !(flags & UNICODE_ALLOW_SUPER)) | |
128 | ) | |
9014280d | 129 | Perl_warner(aTHX_ packWARN(WARN_UTF8), |
6f6ac1de | 130 | "Unicode non-character 0x%04"UVxf" is illegal for interchange", uv); |
507b9800 | 131 | } |
c4d5f83a | 132 | if (UNI_IS_INVARIANT(uv)) { |
eb160463 | 133 | *d++ = (U8)UTF_TO_NATIVE(uv); |
a0ed51b3 LW |
134 | return d; |
135 | } | |
2d331972 | 136 | #if defined(EBCDIC) |
1d72bdf6 NIS |
137 | else { |
138 | STRLEN len = UNISKIP(uv); | |
139 | U8 *p = d+len-1; | |
140 | while (p > d) { | |
eb160463 | 141 | *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK); |
1d72bdf6 NIS |
142 | uv >>= UTF_ACCUMULATION_SHIFT; |
143 | } | |
eb160463 | 144 | *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len)); |
1d72bdf6 NIS |
145 | return d+len; |
146 | } | |
147 | #else /* Non loop style */ | |
a0ed51b3 | 148 | if (uv < 0x800) { |
eb160463 GS |
149 | *d++ = (U8)(( uv >> 6) | 0xc0); |
150 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
151 | return d; |
152 | } | |
153 | if (uv < 0x10000) { | |
eb160463 GS |
154 | *d++ = (U8)(( uv >> 12) | 0xe0); |
155 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
156 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
157 | return d; |
158 | } | |
159 | if (uv < 0x200000) { | |
eb160463 GS |
160 | *d++ = (U8)(( uv >> 18) | 0xf0); |
161 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
162 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
163 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
164 | return d; |
165 | } | |
166 | if (uv < 0x4000000) { | |
eb160463 GS |
167 | *d++ = (U8)(( uv >> 24) | 0xf8); |
168 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); | |
169 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
170 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
171 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
172 | return d; |
173 | } | |
174 | if (uv < 0x80000000) { | |
eb160463 GS |
175 | *d++ = (U8)(( uv >> 30) | 0xfc); |
176 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); | |
177 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); | |
178 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
179 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
180 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
181 | return d; |
182 | } | |
6b8eaf93 | 183 | #ifdef HAS_QUAD |
d7578b48 | 184 | if (uv < UTF8_QUAD_MAX) |
a0ed51b3 LW |
185 | #endif |
186 | { | |
eb160463 GS |
187 | *d++ = 0xfe; /* Can't match U+FEFF! */ |
188 | *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); | |
189 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); | |
190 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); | |
191 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
192 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
193 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
194 | return d; |
195 | } | |
6b8eaf93 | 196 | #ifdef HAS_QUAD |
a0ed51b3 | 197 | { |
eb160463 GS |
198 | *d++ = 0xff; /* Can't match U+FFFE! */ |
199 | *d++ = 0x80; /* 6 Reserved bits */ | |
200 | *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */ | |
201 | *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80); | |
202 | *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80); | |
203 | *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80); | |
204 | *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80); | |
205 | *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); | |
206 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); | |
207 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); | |
208 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
209 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
210 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
211 | return d; |
212 | } | |
213 | #endif | |
1d72bdf6 | 214 | #endif /* Loop style */ |
a0ed51b3 | 215 | } |
9041c2e3 | 216 | |
646ca15d JH |
217 | /* |
218 | ||
219 | Tests if some arbitrary number of bytes begins in a valid UTF-8 | |
220 | character. Note that an INVARIANT (i.e. ASCII) character is a valid | |
221 | UTF-8 character. The actual number of bytes in the UTF-8 character | |
222 | will be returned if it is valid, otherwise 0. | |
223 | ||
224 | This is the "slow" version as opposed to the "fast" version which is | |
225 | the "unrolled" IS_UTF8_CHAR(). E.g. for t/uni/class.t the speed | |
226 | difference is a factor of 2 to 3. For lengths (UTF8SKIP(s)) of four | |
227 | or less you should use the IS_UTF8_CHAR(), for lengths of five or more | |
228 | you should use the _slow(). In practice this means that the _slow() | |
229 | will be used very rarely, since the maximum Unicode code point (as of | |
230 | Unicode 4.1) is U+10FFFF, which encodes in UTF-8 to four bytes. Only | |
231 | the "Perl extended UTF-8" (the infamous 'v-strings') will encode into | |
232 | five bytes or more. | |
233 | ||
234 | =cut */ | |
c053b435 | 235 | STATIC STRLEN |
5f66b61c | 236 | S_is_utf8_char_slow(const U8 *s, const STRLEN len) |
646ca15d JH |
237 | { |
238 | U8 u = *s; | |
239 | STRLEN slen; | |
240 | UV uv, ouv; | |
241 | ||
7918f24d NC |
242 | PERL_ARGS_ASSERT_IS_UTF8_CHAR_SLOW; |
243 | ||
646ca15d JH |
244 | if (UTF8_IS_INVARIANT(u)) |
245 | return 1; | |
246 | ||
247 | if (!UTF8_IS_START(u)) | |
248 | return 0; | |
249 | ||
250 | if (len < 2 || !UTF8_IS_CONTINUATION(s[1])) | |
251 | return 0; | |
252 | ||
253 | slen = len - 1; | |
254 | s++; | |
77263263 TS |
255 | #ifdef EBCDIC |
256 | u = NATIVE_TO_UTF(u); | |
257 | #endif | |
646ca15d JH |
258 | u &= UTF_START_MASK(len); |
259 | uv = u; | |
260 | ouv = uv; | |
261 | while (slen--) { | |
262 | if (!UTF8_IS_CONTINUATION(*s)) | |
263 | return 0; | |
264 | uv = UTF8_ACCUMULATE(uv, *s); | |
265 | if (uv < ouv) | |
266 | return 0; | |
267 | ouv = uv; | |
268 | s++; | |
269 | } | |
270 | ||
271 | if ((STRLEN)UNISKIP(uv) < len) | |
272 | return 0; | |
273 | ||
274 | return len; | |
275 | } | |
9041c2e3 NIS |
276 | |
277 | /* | |
87cea99e | 278 | =for apidoc is_utf8_char |
eebe1485 | 279 | |
5da9da9e | 280 | Tests if some arbitrary number of bytes begins in a valid UTF-8 |
2bbc8d55 SP |
281 | character. Note that an INVARIANT (i.e. ASCII on non-EBCDIC machines) |
282 | character is a valid UTF-8 character. The actual number of bytes in the UTF-8 | |
283 | character will be returned if it is valid, otherwise 0. | |
9041c2e3 | 284 | |
82686b01 | 285 | =cut */ |
067a85ef | 286 | STRLEN |
668b6d8d | 287 | Perl_is_utf8_char(const U8 *s) |
386d01d6 | 288 | { |
44f8325f | 289 | const STRLEN len = UTF8SKIP(s); |
7918f24d NC |
290 | |
291 | PERL_ARGS_ASSERT_IS_UTF8_CHAR; | |
3b0fc154 | 292 | #ifdef IS_UTF8_CHAR |
768c67ee | 293 | if (IS_UTF8_CHAR_FAST(len)) |
3b0fc154 JH |
294 | return IS_UTF8_CHAR(s, len) ? len : 0; |
295 | #endif /* #ifdef IS_UTF8_CHAR */ | |
2c0c5f92 | 296 | return is_utf8_char_slow(s, len); |
386d01d6 GS |
297 | } |
298 | ||
eaf7a4d2 | 299 | |
6662521e | 300 | /* |
87cea99e | 301 | =for apidoc is_utf8_string |
6662521e | 302 | |
c9ada85f | 303 | Returns true if first C<len> bytes of the given string form a valid |
1e54db1a JH |
304 | UTF-8 string, false otherwise. Note that 'a valid UTF-8 string' does |
305 | not mean 'a string that contains code points above 0x7F encoded in UTF-8' | |
306 | because a valid ASCII string is a valid UTF-8 string. | |
6662521e | 307 | |
eaf7a4d2 | 308 | See also is_ascii_string(), is_utf8_string_loclen(), and is_utf8_string_loc(). |
768c67ee | 309 | |
6662521e GS |
310 | =cut |
311 | */ | |
312 | ||
8e84507e | 313 | bool |
668b6d8d | 314 | Perl_is_utf8_string(const U8 *s, STRLEN len) |
6662521e | 315 | { |
35da51f7 | 316 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
7fc63493 | 317 | const U8* x = s; |
067a85ef | 318 | |
7918f24d | 319 | PERL_ARGS_ASSERT_IS_UTF8_STRING; |
1aa99e6b | 320 | |
6662521e | 321 | while (x < send) { |
a3b680e6 | 322 | STRLEN c; |
1acdb0da JH |
323 | /* Inline the easy bits of is_utf8_char() here for speed... */ |
324 | if (UTF8_IS_INVARIANT(*x)) | |
325 | c = 1; | |
326 | else if (!UTF8_IS_START(*x)) | |
768c67ee | 327 | goto out; |
1acdb0da JH |
328 | else { |
329 | /* ... and call is_utf8_char() only if really needed. */ | |
646ca15d JH |
330 | #ifdef IS_UTF8_CHAR |
331 | c = UTF8SKIP(x); | |
768c67ee JH |
332 | if (IS_UTF8_CHAR_FAST(c)) { |
333 | if (!IS_UTF8_CHAR(x, c)) | |
3c614e38 AD |
334 | c = 0; |
335 | } | |
336 | else | |
337 | c = is_utf8_char_slow(x, c); | |
646ca15d JH |
338 | #else |
339 | c = is_utf8_char(x); | |
340 | #endif /* #ifdef IS_UTF8_CHAR */ | |
1acdb0da | 341 | if (!c) |
768c67ee | 342 | goto out; |
1acdb0da | 343 | } |
6662521e | 344 | x += c; |
6662521e | 345 | } |
768c67ee JH |
346 | |
347 | out: | |
60006e79 JH |
348 | if (x != send) |
349 | return FALSE; | |
067a85ef A |
350 | |
351 | return TRUE; | |
6662521e GS |
352 | } |
353 | ||
67e989fb | 354 | /* |
814fafa7 NC |
355 | Implemented as a macro in utf8.h |
356 | ||
87cea99e | 357 | =for apidoc is_utf8_string_loc |
814fafa7 NC |
358 | |
359 | Like is_utf8_string() but stores the location of the failure (in the | |
360 | case of "utf8ness failure") or the location s+len (in the case of | |
361 | "utf8ness success") in the C<ep>. | |
362 | ||
363 | See also is_utf8_string_loclen() and is_utf8_string(). | |
364 | ||
87cea99e | 365 | =for apidoc is_utf8_string_loclen |
81cd54e3 | 366 | |
e3e4599f | 367 | Like is_utf8_string() but stores the location of the failure (in the |
768c67ee JH |
368 | case of "utf8ness failure") or the location s+len (in the case of |
369 | "utf8ness success") in the C<ep>, and the number of UTF-8 | |
370 | encoded characters in the C<el>. | |
371 | ||
372 | See also is_utf8_string_loc() and is_utf8_string(). | |
81cd54e3 JH |
373 | |
374 | =cut | |
375 | */ | |
376 | ||
377 | bool | |
668b6d8d | 378 | Perl_is_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el) |
81cd54e3 | 379 | { |
35da51f7 | 380 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
7fc63493 | 381 | const U8* x = s; |
81cd54e3 | 382 | STRLEN c; |
3ebfea28 | 383 | STRLEN outlen = 0; |
7918f24d NC |
384 | |
385 | PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN; | |
81cd54e3 | 386 | |
81cd54e3 JH |
387 | while (x < send) { |
388 | /* Inline the easy bits of is_utf8_char() here for speed... */ | |
389 | if (UTF8_IS_INVARIANT(*x)) | |
768c67ee JH |
390 | c = 1; |
391 | else if (!UTF8_IS_START(*x)) | |
392 | goto out; | |
81cd54e3 | 393 | else { |
768c67ee JH |
394 | /* ... and call is_utf8_char() only if really needed. */ |
395 | #ifdef IS_UTF8_CHAR | |
396 | c = UTF8SKIP(x); | |
397 | if (IS_UTF8_CHAR_FAST(c)) { | |
398 | if (!IS_UTF8_CHAR(x, c)) | |
399 | c = 0; | |
400 | } else | |
401 | c = is_utf8_char_slow(x, c); | |
402 | #else | |
403 | c = is_utf8_char(x); | |
404 | #endif /* #ifdef IS_UTF8_CHAR */ | |
405 | if (!c) | |
406 | goto out; | |
81cd54e3 | 407 | } |
768c67ee | 408 | x += c; |
3ebfea28 | 409 | outlen++; |
81cd54e3 | 410 | } |
768c67ee JH |
411 | |
412 | out: | |
3ebfea28 AL |
413 | if (el) |
414 | *el = outlen; | |
415 | ||
768c67ee JH |
416 | if (ep) |
417 | *ep = x; | |
3ebfea28 | 418 | return (x == send); |
81cd54e3 JH |
419 | } |
420 | ||
421 | /* | |
768c67ee | 422 | |
87cea99e | 423 | =for apidoc utf8n_to_uvuni |
67e989fb | 424 | |
9041c2e3 | 425 | Bottom level UTF-8 decode routine. |
38a44b82 | 426 | Returns the Unicode code point value of the first character in the string C<s> |
1e54db1a | 427 | which is assumed to be in UTF-8 encoding and no longer than C<curlen>; |
7df053ec | 428 | C<retlen> will be set to the length, in bytes, of that character. |
67e989fb | 429 | |
1e54db1a | 430 | If C<s> does not point to a well-formed UTF-8 character, the behaviour |
dcad2880 JH |
431 | is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY, |
432 | it is assumed that the caller will raise a warning, and this function | |
28d3d195 JH |
433 | will silently just set C<retlen> to C<-1> and return zero. If the |
434 | C<flags> does not contain UTF8_CHECK_ONLY, warnings about | |
435 | malformations will be given, C<retlen> will be set to the expected | |
436 | length of the UTF-8 character in bytes, and zero will be returned. | |
437 | ||
438 | The C<flags> can also contain various flags to allow deviations from | |
439 | the strict UTF-8 encoding (see F<utf8.h>). | |
67e989fb | 440 | |
9041c2e3 NIS |
441 | Most code should use utf8_to_uvchr() rather than call this directly. |
442 | ||
37607a96 PK |
443 | =cut |
444 | */ | |
67e989fb | 445 | |
a0ed51b3 | 446 | UV |
7fc63493 | 447 | Perl_utf8n_to_uvuni(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) |
a0ed51b3 | 448 | { |
97aff369 | 449 | dVAR; |
d4c19fe8 | 450 | const U8 * const s0 = s; |
9c5ffd7c | 451 | UV uv = *s, ouv = 0; |
ba210ebe | 452 | STRLEN len = 1; |
7fc63493 AL |
453 | const bool dowarn = ckWARN_d(WARN_UTF8); |
454 | const UV startbyte = *s; | |
ba210ebe | 455 | STRLEN expectlen = 0; |
a0dbb045 | 456 | U32 warning = 0; |
5b311467 | 457 | SV* sv; |
a0dbb045 | 458 | |
7918f24d NC |
459 | PERL_ARGS_ASSERT_UTF8N_TO_UVUNI; |
460 | ||
5b311467 | 461 | /* This list is a superset of the UTF8_ALLOW_XXX. BUT it isn't, eg SUPER missing XXX */ |
a0dbb045 JH |
462 | |
463 | #define UTF8_WARN_EMPTY 1 | |
464 | #define UTF8_WARN_CONTINUATION 2 | |
465 | #define UTF8_WARN_NON_CONTINUATION 3 | |
466 | #define UTF8_WARN_FE_FF 4 | |
467 | #define UTF8_WARN_SHORT 5 | |
468 | #define UTF8_WARN_OVERFLOW 6 | |
469 | #define UTF8_WARN_SURROGATE 7 | |
c867b360 JH |
470 | #define UTF8_WARN_LONG 8 |
471 | #define UTF8_WARN_FFFF 9 /* Also FFFE. */ | |
a0dbb045 JH |
472 | |
473 | if (curlen == 0 && | |
474 | !(flags & UTF8_ALLOW_EMPTY)) { | |
475 | warning = UTF8_WARN_EMPTY; | |
0c443dc2 JH |
476 | goto malformed; |
477 | } | |
478 | ||
1d72bdf6 | 479 | if (UTF8_IS_INVARIANT(uv)) { |
a0ed51b3 LW |
480 | if (retlen) |
481 | *retlen = 1; | |
c4d5f83a | 482 | return (UV) (NATIVE_TO_UTF(*s)); |
a0ed51b3 | 483 | } |
67e989fb | 484 | |
421a8bf2 | 485 | if (UTF8_IS_CONTINUATION(uv) && |
fcc8fcf6 | 486 | !(flags & UTF8_ALLOW_CONTINUATION)) { |
a0dbb045 | 487 | warning = UTF8_WARN_CONTINUATION; |
ba210ebe JH |
488 | goto malformed; |
489 | } | |
490 | ||
421a8bf2 | 491 | if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) && |
fcc8fcf6 | 492 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { |
a0dbb045 | 493 | warning = UTF8_WARN_NON_CONTINUATION; |
ba210ebe JH |
494 | goto malformed; |
495 | } | |
9041c2e3 | 496 | |
1d72bdf6 | 497 | #ifdef EBCDIC |
75383841 | 498 | uv = NATIVE_TO_UTF(uv); |
1d72bdf6 | 499 | #else |
fcc8fcf6 JH |
500 | if ((uv == 0xfe || uv == 0xff) && |
501 | !(flags & UTF8_ALLOW_FE_FF)) { | |
a0dbb045 | 502 | warning = UTF8_WARN_FE_FF; |
ba210ebe | 503 | goto malformed; |
a0ed51b3 | 504 | } |
1d72bdf6 NIS |
505 | #endif |
506 | ||
ba210ebe JH |
507 | if (!(uv & 0x20)) { len = 2; uv &= 0x1f; } |
508 | else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; } | |
509 | else if (!(uv & 0x08)) { len = 4; uv &= 0x07; } | |
510 | else if (!(uv & 0x04)) { len = 5; uv &= 0x03; } | |
1d72bdf6 NIS |
511 | #ifdef EBCDIC |
512 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } | |
513 | else { len = 7; uv &= 0x01; } | |
514 | #else | |
ba210ebe JH |
515 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } |
516 | else if (!(uv & 0x01)) { len = 7; uv = 0; } | |
1d72bdf6 NIS |
517 | else { len = 13; uv = 0; } /* whoa! */ |
518 | #endif | |
519 | ||
a0ed51b3 LW |
520 | if (retlen) |
521 | *retlen = len; | |
9041c2e3 | 522 | |
ba210ebe JH |
523 | expectlen = len; |
524 | ||
fcc8fcf6 JH |
525 | if ((curlen < expectlen) && |
526 | !(flags & UTF8_ALLOW_SHORT)) { | |
a0dbb045 | 527 | warning = UTF8_WARN_SHORT; |
ba210ebe JH |
528 | goto malformed; |
529 | } | |
530 | ||
531 | len--; | |
a0ed51b3 | 532 | s++; |
ba210ebe JH |
533 | ouv = uv; |
534 | ||
a0ed51b3 | 535 | while (len--) { |
421a8bf2 JH |
536 | if (!UTF8_IS_CONTINUATION(*s) && |
537 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { | |
a0dbb045 JH |
538 | s--; |
539 | warning = UTF8_WARN_NON_CONTINUATION; | |
ba210ebe | 540 | goto malformed; |
a0ed51b3 LW |
541 | } |
542 | else | |
8850bf83 | 543 | uv = UTF8_ACCUMULATE(uv, *s); |
a0dbb045 JH |
544 | if (!(uv > ouv)) { |
545 | /* These cannot be allowed. */ | |
546 | if (uv == ouv) { | |
75dbc644 | 547 | if (expectlen != 13 && !(flags & UTF8_ALLOW_LONG)) { |
a0dbb045 JH |
548 | warning = UTF8_WARN_LONG; |
549 | goto malformed; | |
550 | } | |
551 | } | |
552 | else { /* uv < ouv */ | |
553 | /* This cannot be allowed. */ | |
554 | warning = UTF8_WARN_OVERFLOW; | |
555 | goto malformed; | |
556 | } | |
ba210ebe JH |
557 | } |
558 | s++; | |
559 | ouv = uv; | |
560 | } | |
561 | ||
421a8bf2 | 562 | if (UNICODE_IS_SURROGATE(uv) && |
fcc8fcf6 | 563 | !(flags & UTF8_ALLOW_SURROGATE)) { |
a0dbb045 | 564 | warning = UTF8_WARN_SURROGATE; |
ba210ebe | 565 | goto malformed; |
eb160463 | 566 | } else if ((expectlen > (STRLEN)UNISKIP(uv)) && |
fcc8fcf6 | 567 | !(flags & UTF8_ALLOW_LONG)) { |
a0dbb045 | 568 | warning = UTF8_WARN_LONG; |
ba210ebe | 569 | goto malformed; |
421a8bf2 | 570 | } else if (UNICODE_IS_ILLEGAL(uv) && |
a9917092 | 571 | !(flags & UTF8_ALLOW_FFFF)) { |
a0dbb045 | 572 | warning = UTF8_WARN_FFFF; |
a9917092 | 573 | goto malformed; |
a0ed51b3 | 574 | } |
ba210ebe | 575 | |
a0ed51b3 | 576 | return uv; |
ba210ebe JH |
577 | |
578 | malformed: | |
579 | ||
fcc8fcf6 | 580 | if (flags & UTF8_CHECK_ONLY) { |
ba210ebe | 581 | if (retlen) |
10edeb5d | 582 | *retlen = ((STRLEN) -1); |
ba210ebe JH |
583 | return 0; |
584 | } | |
585 | ||
a0dbb045 | 586 | if (dowarn) { |
5b311467 KW |
587 | if (warning == UTF8_WARN_FFFF) { |
588 | sv = newSVpvs_flags("Unicode non-character ", SVs_TEMP); | |
589 | Perl_sv_catpvf(aTHX_ sv, "0x%04"UVxf" is illegal for interchange", uv); | |
590 | } | |
591 | else { | |
592 | sv = newSVpvs_flags("Malformed UTF-8 character ", SVs_TEMP); | |
593 | ||
594 | switch (warning) { | |
595 | case 0: /* Intentionally empty. */ break; | |
596 | case UTF8_WARN_EMPTY: | |
597 | sv_catpvs(sv, "(empty string)"); | |
598 | break; | |
599 | case UTF8_WARN_CONTINUATION: | |
600 | Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv); | |
601 | break; | |
602 | case UTF8_WARN_NON_CONTINUATION: | |
603 | if (s == s0) | |
604 | Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")", | |
605 | (UV)s[1], startbyte); | |
606 | else { | |
607 | const int len = (int)(s-s0); | |
608 | Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)", | |
609 | (UV)s[1], len, len > 1 ? "s" : "", startbyte, (int)expectlen); | |
610 | } | |
611 | ||
612 | break; | |
613 | case UTF8_WARN_FE_FF: | |
614 | Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv); | |
615 | break; | |
616 | case UTF8_WARN_SHORT: | |
617 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", | |
618 | (int)curlen, curlen == 1 ? "" : "s", (int)expectlen, startbyte); | |
619 | expectlen = curlen; /* distance for caller to skip */ | |
620 | break; | |
621 | case UTF8_WARN_OVERFLOW: | |
622 | Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")", | |
623 | ouv, *s, startbyte); | |
624 | break; | |
625 | case UTF8_WARN_SURROGATE: | |
626 | Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv); | |
627 | break; | |
628 | case UTF8_WARN_LONG: | |
629 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", | |
630 | (int)expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte); | |
631 | break; | |
632 | default: | |
633 | sv_catpvs(sv, "(unknown reason)"); | |
634 | break; | |
551405c4 | 635 | } |
a0dbb045 JH |
636 | } |
637 | ||
638 | if (warning) { | |
44f8325f | 639 | const char * const s = SvPVX_const(sv); |
a0dbb045 JH |
640 | |
641 | if (PL_op) | |
9014280d | 642 | Perl_warner(aTHX_ packWARN(WARN_UTF8), |
53e06cf0 | 643 | "%s in %s", s, OP_DESC(PL_op)); |
a0dbb045 | 644 | else |
9014280d | 645 | Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s); |
a0dbb045 JH |
646 | } |
647 | } | |
648 | ||
ba210ebe | 649 | if (retlen) |
28d3d195 | 650 | *retlen = expectlen ? expectlen : len; |
ba210ebe | 651 | |
28d3d195 | 652 | return 0; |
a0ed51b3 LW |
653 | } |
654 | ||
8e84507e | 655 | /* |
87cea99e | 656 | =for apidoc utf8_to_uvchr |
9041c2e3 NIS |
657 | |
658 | Returns the native character value of the first character in the string C<s> | |
1e54db1a | 659 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
9041c2e3 NIS |
660 | length, in bytes, of that character. |
661 | ||
1e54db1a | 662 | If C<s> does not point to a well-formed UTF-8 character, zero is |
9041c2e3 NIS |
663 | returned and retlen is set, if possible, to -1. |
664 | ||
665 | =cut | |
666 | */ | |
667 | ||
668 | UV | |
7fc63493 | 669 | Perl_utf8_to_uvchr(pTHX_ const U8 *s, STRLEN *retlen) |
9041c2e3 | 670 | { |
7918f24d NC |
671 | PERL_ARGS_ASSERT_UTF8_TO_UVCHR; |
672 | ||
1754c1a1 NC |
673 | return utf8n_to_uvchr(s, UTF8_MAXBYTES, retlen, |
674 | ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); | |
9041c2e3 NIS |
675 | } |
676 | ||
677 | /* | |
87cea99e | 678 | =for apidoc utf8_to_uvuni |
9041c2e3 NIS |
679 | |
680 | Returns the Unicode code point of the first character in the string C<s> | |
1e54db1a | 681 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
9041c2e3 NIS |
682 | length, in bytes, of that character. |
683 | ||
2bbc8d55 | 684 | This function should only be used when the returned UV is considered |
9041c2e3 NIS |
685 | an index into the Unicode semantic tables (e.g. swashes). |
686 | ||
1e54db1a | 687 | If C<s> does not point to a well-formed UTF-8 character, zero is |
ba210ebe | 688 | returned and retlen is set, if possible, to -1. |
8e84507e NIS |
689 | |
690 | =cut | |
691 | */ | |
692 | ||
693 | UV | |
7fc63493 | 694 | Perl_utf8_to_uvuni(pTHX_ const U8 *s, STRLEN *retlen) |
8e84507e | 695 | { |
7918f24d NC |
696 | PERL_ARGS_ASSERT_UTF8_TO_UVUNI; |
697 | ||
9041c2e3 | 698 | /* Call the low level routine asking for checks */ |
89ebb4a3 | 699 | return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXBYTES, retlen, |
872c91ae | 700 | ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); |
8e84507e NIS |
701 | } |
702 | ||
b76347f2 | 703 | /* |
87cea99e | 704 | =for apidoc utf8_length |
b76347f2 JH |
705 | |
706 | Return the length of the UTF-8 char encoded string C<s> in characters. | |
02eb7b47 JH |
707 | Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end |
708 | up past C<e>, croaks. | |
b76347f2 JH |
709 | |
710 | =cut | |
711 | */ | |
712 | ||
713 | STRLEN | |
35a4481c | 714 | Perl_utf8_length(pTHX_ const U8 *s, const U8 *e) |
b76347f2 | 715 | { |
97aff369 | 716 | dVAR; |
b76347f2 JH |
717 | STRLEN len = 0; |
718 | ||
7918f24d NC |
719 | PERL_ARGS_ASSERT_UTF8_LENGTH; |
720 | ||
8850bf83 JH |
721 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g. |
722 | * the bitops (especially ~) can create illegal UTF-8. | |
723 | * In other words: in Perl UTF-8 is not just for Unicode. */ | |
724 | ||
a3b680e6 AL |
725 | if (e < s) |
726 | goto warn_and_return; | |
b76347f2 | 727 | while (s < e) { |
8e91ec7f AV |
728 | if (!UTF8_IS_INVARIANT(*s)) |
729 | s += UTF8SKIP(s); | |
730 | else | |
731 | s++; | |
732 | len++; | |
733 | } | |
734 | ||
735 | if (e != s) { | |
736 | len--; | |
737 | warn_and_return: | |
9b387841 NC |
738 | if (PL_op) |
739 | Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), | |
740 | "%s in %s", unees, OP_DESC(PL_op)); | |
741 | else | |
742 | Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), unees); | |
b76347f2 JH |
743 | } |
744 | ||
745 | return len; | |
746 | } | |
747 | ||
b06226ff | 748 | /* |
87cea99e | 749 | =for apidoc utf8_distance |
b06226ff | 750 | |
1e54db1a | 751 | Returns the number of UTF-8 characters between the UTF-8 pointers C<a> |
b06226ff JH |
752 | and C<b>. |
753 | ||
754 | WARNING: use only if you *know* that the pointers point inside the | |
755 | same UTF-8 buffer. | |
756 | ||
37607a96 PK |
757 | =cut |
758 | */ | |
a0ed51b3 | 759 | |
02eb7b47 | 760 | IV |
35a4481c | 761 | Perl_utf8_distance(pTHX_ const U8 *a, const U8 *b) |
a0ed51b3 | 762 | { |
7918f24d NC |
763 | PERL_ARGS_ASSERT_UTF8_DISTANCE; |
764 | ||
bf1665bc | 765 | return (a < b) ? -1 * (IV) utf8_length(a, b) : (IV) utf8_length(b, a); |
a0ed51b3 LW |
766 | } |
767 | ||
b06226ff | 768 | /* |
87cea99e | 769 | =for apidoc utf8_hop |
b06226ff | 770 | |
8850bf83 JH |
771 | Return the UTF-8 pointer C<s> displaced by C<off> characters, either |
772 | forward or backward. | |
b06226ff JH |
773 | |
774 | WARNING: do not use the following unless you *know* C<off> is within | |
8850bf83 JH |
775 | the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned |
776 | on the first byte of character or just after the last byte of a character. | |
b06226ff | 777 | |
37607a96 PK |
778 | =cut |
779 | */ | |
a0ed51b3 LW |
780 | |
781 | U8 * | |
4373e329 | 782 | Perl_utf8_hop(pTHX_ const U8 *s, I32 off) |
a0ed51b3 | 783 | { |
7918f24d NC |
784 | PERL_ARGS_ASSERT_UTF8_HOP; |
785 | ||
96a5add6 | 786 | PERL_UNUSED_CONTEXT; |
8850bf83 JH |
787 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g |
788 | * the bitops (especially ~) can create illegal UTF-8. | |
789 | * In other words: in Perl UTF-8 is not just for Unicode. */ | |
790 | ||
a0ed51b3 LW |
791 | if (off >= 0) { |
792 | while (off--) | |
793 | s += UTF8SKIP(s); | |
794 | } | |
795 | else { | |
796 | while (off++) { | |
797 | s--; | |
8850bf83 JH |
798 | while (UTF8_IS_CONTINUATION(*s)) |
799 | s--; | |
a0ed51b3 LW |
800 | } |
801 | } | |
4373e329 | 802 | return (U8 *)s; |
a0ed51b3 LW |
803 | } |
804 | ||
6940069f | 805 | /* |
87cea99e | 806 | =for apidoc utf8_to_bytes |
6940069f | 807 | |
2bbc8d55 | 808 | Converts a string C<s> of length C<len> from UTF-8 into native byte encoding. |
246fae53 MG |
809 | Unlike C<bytes_to_utf8>, this over-writes the original string, and |
810 | updates len to contain the new length. | |
67e989fb | 811 | Returns zero on failure, setting C<len> to -1. |
6940069f | 812 | |
95be277c NC |
813 | If you need a copy of the string, see C<bytes_from_utf8>. |
814 | ||
6940069f GS |
815 | =cut |
816 | */ | |
817 | ||
818 | U8 * | |
37607a96 | 819 | Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len) |
6940069f | 820 | { |
d4c19fe8 AL |
821 | U8 * const save = s; |
822 | U8 * const send = s + *len; | |
6940069f | 823 | U8 *d; |
246fae53 | 824 | |
7918f24d NC |
825 | PERL_ARGS_ASSERT_UTF8_TO_BYTES; |
826 | ||
1e54db1a | 827 | /* ensure valid UTF-8 and chars < 256 before updating string */ |
d4c19fe8 | 828 | while (s < send) { |
dcad2880 JH |
829 | U8 c = *s++; |
830 | ||
1d72bdf6 NIS |
831 | if (!UTF8_IS_INVARIANT(c) && |
832 | (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send) | |
833 | || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) { | |
10edeb5d | 834 | *len = ((STRLEN) -1); |
dcad2880 JH |
835 | return 0; |
836 | } | |
246fae53 | 837 | } |
dcad2880 JH |
838 | |
839 | d = s = save; | |
6940069f | 840 | while (s < send) { |
ed646e6e | 841 | STRLEN ulen; |
9041c2e3 | 842 | *d++ = (U8)utf8_to_uvchr(s, &ulen); |
ed646e6e | 843 | s += ulen; |
6940069f GS |
844 | } |
845 | *d = '\0'; | |
246fae53 | 846 | *len = d - save; |
6940069f GS |
847 | return save; |
848 | } | |
849 | ||
850 | /* | |
87cea99e | 851 | =for apidoc bytes_from_utf8 |
f9a63242 | 852 | |
2bbc8d55 | 853 | Converts a string C<s> of length C<len> from UTF-8 into native byte encoding. |
35a4481c | 854 | Unlike C<utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to |
ef9edfd0 JH |
855 | the newly-created string, and updates C<len> to contain the new |
856 | length. Returns the original string if no conversion occurs, C<len> | |
857 | is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to | |
2bbc8d55 SP |
858 | 0 if C<s> is converted or consisted entirely of characters that are invariant |
859 | in utf8 (i.e., US-ASCII on non-EBCDIC machines). | |
f9a63242 | 860 | |
37607a96 PK |
861 | =cut |
862 | */ | |
f9a63242 JH |
863 | |
864 | U8 * | |
e1ec3a88 | 865 | Perl_bytes_from_utf8(pTHX_ const U8 *s, STRLEN *len, bool *is_utf8) |
f9a63242 | 866 | { |
f9a63242 | 867 | U8 *d; |
e1ec3a88 AL |
868 | const U8 *start = s; |
869 | const U8 *send; | |
f9a63242 JH |
870 | I32 count = 0; |
871 | ||
7918f24d NC |
872 | PERL_ARGS_ASSERT_BYTES_FROM_UTF8; |
873 | ||
96a5add6 | 874 | PERL_UNUSED_CONTEXT; |
f9a63242 | 875 | if (!*is_utf8) |
73d840c0 | 876 | return (U8 *)start; |
f9a63242 | 877 | |
1e54db1a | 878 | /* ensure valid UTF-8 and chars < 256 before converting string */ |
f9a63242 | 879 | for (send = s + *len; s < send;) { |
e1ec3a88 | 880 | U8 c = *s++; |
1d72bdf6 | 881 | if (!UTF8_IS_INVARIANT(c)) { |
db42d148 NIS |
882 | if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send && |
883 | (c = *s++) && UTF8_IS_CONTINUATION(c)) | |
884 | count++; | |
885 | else | |
73d840c0 | 886 | return (U8 *)start; |
db42d148 | 887 | } |
f9a63242 JH |
888 | } |
889 | ||
35da51f7 | 890 | *is_utf8 = FALSE; |
f9a63242 | 891 | |
212542aa | 892 | Newx(d, (*len) - count + 1, U8); |
ef9edfd0 | 893 | s = start; start = d; |
f9a63242 JH |
894 | while (s < send) { |
895 | U8 c = *s++; | |
c4d5f83a NIS |
896 | if (!UTF8_IS_INVARIANT(c)) { |
897 | /* Then it is two-byte encoded */ | |
898 | c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++); | |
899 | c = ASCII_TO_NATIVE(c); | |
900 | } | |
901 | *d++ = c; | |
f9a63242 JH |
902 | } |
903 | *d = '\0'; | |
904 | *len = d - start; | |
73d840c0 | 905 | return (U8 *)start; |
f9a63242 JH |
906 | } |
907 | ||
908 | /* | |
87cea99e | 909 | =for apidoc bytes_to_utf8 |
6940069f | 910 | |
2bbc8d55 | 911 | Converts a string C<s> of length C<len> from the native encoding into UTF-8. |
6662521e GS |
912 | Returns a pointer to the newly-created string, and sets C<len> to |
913 | reflect the new length. | |
6940069f | 914 | |
2bbc8d55 SP |
915 | A NUL character will be written after the end of the string. |
916 | ||
917 | If you want to convert to UTF-8 from encodings other than | |
918 | the native (Latin1 or EBCDIC), | |
c9ada85f JH |
919 | see sv_recode_to_utf8(). |
920 | ||
497711e7 | 921 | =cut |
6940069f GS |
922 | */ |
923 | ||
924 | U8* | |
35a4481c | 925 | Perl_bytes_to_utf8(pTHX_ const U8 *s, STRLEN *len) |
6940069f | 926 | { |
35a4481c | 927 | const U8 * const send = s + (*len); |
6940069f GS |
928 | U8 *d; |
929 | U8 *dst; | |
7918f24d NC |
930 | |
931 | PERL_ARGS_ASSERT_BYTES_TO_UTF8; | |
96a5add6 | 932 | PERL_UNUSED_CONTEXT; |
6940069f | 933 | |
212542aa | 934 | Newx(d, (*len) * 2 + 1, U8); |
6940069f GS |
935 | dst = d; |
936 | ||
937 | while (s < send) { | |
35a4481c | 938 | const UV uv = NATIVE_TO_ASCII(*s++); |
c4d5f83a | 939 | if (UNI_IS_INVARIANT(uv)) |
eb160463 | 940 | *d++ = (U8)UTF_TO_NATIVE(uv); |
6940069f | 941 | else { |
eb160463 GS |
942 | *d++ = (U8)UTF8_EIGHT_BIT_HI(uv); |
943 | *d++ = (U8)UTF8_EIGHT_BIT_LO(uv); | |
6940069f GS |
944 | } |
945 | } | |
946 | *d = '\0'; | |
6662521e | 947 | *len = d-dst; |
6940069f GS |
948 | return dst; |
949 | } | |
950 | ||
a0ed51b3 | 951 | /* |
dea0fc0b | 952 | * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8. |
a0ed51b3 LW |
953 | * |
954 | * Destination must be pre-extended to 3/2 source. Do not use in-place. | |
955 | * We optimize for native, for obvious reasons. */ | |
956 | ||
957 | U8* | |
dea0fc0b | 958 | Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 | 959 | { |
dea0fc0b JH |
960 | U8* pend; |
961 | U8* dstart = d; | |
962 | ||
7918f24d NC |
963 | PERL_ARGS_ASSERT_UTF16_TO_UTF8; |
964 | ||
dea0fc0b | 965 | if (bytelen & 1) |
f5992bc4 | 966 | Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen %"UVuf, (UV)bytelen); |
dea0fc0b JH |
967 | |
968 | pend = p + bytelen; | |
969 | ||
a0ed51b3 | 970 | while (p < pend) { |
dea0fc0b JH |
971 | UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */ |
972 | p += 2; | |
a0ed51b3 | 973 | if (uv < 0x80) { |
e294cc5d JH |
974 | #ifdef EBCDIC |
975 | *d++ = UNI_TO_NATIVE(uv); | |
976 | #else | |
eb160463 | 977 | *d++ = (U8)uv; |
e294cc5d | 978 | #endif |
a0ed51b3 LW |
979 | continue; |
980 | } | |
981 | if (uv < 0x800) { | |
eb160463 GS |
982 | *d++ = (U8)(( uv >> 6) | 0xc0); |
983 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
984 | continue; |
985 | } | |
52b9aa85 | 986 | if (uv >= 0xd800 && uv <= 0xdbff) { /* surrogates */ |
01ea242b | 987 | if (p >= pend) { |
dea0fc0b | 988 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
01ea242b NC |
989 | } else { |
990 | UV low = (p[0] << 8) + p[1]; | |
991 | p += 2; | |
52b9aa85 | 992 | if (low < 0xdc00 || low > 0xdfff) |
01ea242b NC |
993 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
994 | uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000; | |
995 | } | |
dbde1951 NC |
996 | } else if (uv >= 0xdc00 && uv <= 0xdfff) { |
997 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); | |
a0ed51b3 LW |
998 | } |
999 | if (uv < 0x10000) { | |
eb160463 GS |
1000 | *d++ = (U8)(( uv >> 12) | 0xe0); |
1001 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
1002 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
1003 | continue; |
1004 | } | |
1005 | else { | |
eb160463 GS |
1006 | *d++ = (U8)(( uv >> 18) | 0xf0); |
1007 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); | |
1008 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); | |
1009 | *d++ = (U8)(( uv & 0x3f) | 0x80); | |
a0ed51b3 LW |
1010 | continue; |
1011 | } | |
1012 | } | |
dea0fc0b | 1013 | *newlen = d - dstart; |
a0ed51b3 LW |
1014 | return d; |
1015 | } | |
1016 | ||
1017 | /* Note: this one is slightly destructive of the source. */ | |
1018 | ||
1019 | U8* | |
dea0fc0b | 1020 | Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 LW |
1021 | { |
1022 | U8* s = (U8*)p; | |
d4c19fe8 | 1023 | U8* const send = s + bytelen; |
7918f24d NC |
1024 | |
1025 | PERL_ARGS_ASSERT_UTF16_TO_UTF8_REVERSED; | |
1026 | ||
e0ea5e2d NC |
1027 | if (bytelen & 1) |
1028 | Perl_croak(aTHX_ "panic: utf16_to_utf8_reversed: odd bytelen %"UVuf, | |
1029 | (UV)bytelen); | |
1030 | ||
a0ed51b3 | 1031 | while (s < send) { |
d4c19fe8 | 1032 | const U8 tmp = s[0]; |
a0ed51b3 LW |
1033 | s[0] = s[1]; |
1034 | s[1] = tmp; | |
1035 | s += 2; | |
1036 | } | |
dea0fc0b | 1037 | return utf16_to_utf8(p, d, bytelen, newlen); |
a0ed51b3 LW |
1038 | } |
1039 | ||
1040 | /* for now these are all defined (inefficiently) in terms of the utf8 versions */ | |
1041 | ||
1042 | bool | |
84afefe6 | 1043 | Perl_is_uni_alnum(pTHX_ UV c) |
a0ed51b3 | 1044 | { |
89ebb4a3 | 1045 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1046 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1047 | return is_utf8_alnum(tmpbuf); |
1048 | } | |
1049 | ||
1050 | bool | |
84afefe6 | 1051 | Perl_is_uni_idfirst(pTHX_ UV c) |
a0ed51b3 | 1052 | { |
89ebb4a3 | 1053 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1054 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1055 | return is_utf8_idfirst(tmpbuf); |
1056 | } | |
1057 | ||
1058 | bool | |
84afefe6 | 1059 | Perl_is_uni_alpha(pTHX_ UV c) |
a0ed51b3 | 1060 | { |
89ebb4a3 | 1061 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1062 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1063 | return is_utf8_alpha(tmpbuf); |
1064 | } | |
1065 | ||
1066 | bool | |
84afefe6 | 1067 | Perl_is_uni_ascii(pTHX_ UV c) |
4d61ec05 | 1068 | { |
89ebb4a3 | 1069 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1070 | uvchr_to_utf8(tmpbuf, c); |
4d61ec05 GS |
1071 | return is_utf8_ascii(tmpbuf); |
1072 | } | |
1073 | ||
1074 | bool | |
84afefe6 | 1075 | Perl_is_uni_space(pTHX_ UV c) |
a0ed51b3 | 1076 | { |
89ebb4a3 | 1077 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1078 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1079 | return is_utf8_space(tmpbuf); |
1080 | } | |
1081 | ||
1082 | bool | |
84afefe6 | 1083 | Perl_is_uni_digit(pTHX_ UV c) |
a0ed51b3 | 1084 | { |
89ebb4a3 | 1085 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1086 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1087 | return is_utf8_digit(tmpbuf); |
1088 | } | |
1089 | ||
1090 | bool | |
84afefe6 | 1091 | Perl_is_uni_upper(pTHX_ UV c) |
a0ed51b3 | 1092 | { |
89ebb4a3 | 1093 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1094 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1095 | return is_utf8_upper(tmpbuf); |
1096 | } | |
1097 | ||
1098 | bool | |
84afefe6 | 1099 | Perl_is_uni_lower(pTHX_ UV c) |
a0ed51b3 | 1100 | { |
89ebb4a3 | 1101 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1102 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 LW |
1103 | return is_utf8_lower(tmpbuf); |
1104 | } | |
1105 | ||
1106 | bool | |
84afefe6 | 1107 | Perl_is_uni_cntrl(pTHX_ UV c) |
b8c5462f | 1108 | { |
89ebb4a3 | 1109 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1110 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f JH |
1111 | return is_utf8_cntrl(tmpbuf); |
1112 | } | |
1113 | ||
1114 | bool | |
84afefe6 | 1115 | Perl_is_uni_graph(pTHX_ UV c) |
b8c5462f | 1116 | { |
89ebb4a3 | 1117 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1118 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f JH |
1119 | return is_utf8_graph(tmpbuf); |
1120 | } | |
1121 | ||
1122 | bool | |
84afefe6 | 1123 | Perl_is_uni_print(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_print(tmpbuf); |
1128 | } | |
1129 | ||
b8c5462f | 1130 | bool |
84afefe6 | 1131 | Perl_is_uni_punct(pTHX_ UV c) |
b8c5462f | 1132 | { |
89ebb4a3 | 1133 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 | 1134 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f JH |
1135 | return is_utf8_punct(tmpbuf); |
1136 | } | |
1137 | ||
4d61ec05 | 1138 | bool |
84afefe6 | 1139 | Perl_is_uni_xdigit(pTHX_ UV c) |
4d61ec05 | 1140 | { |
89ebb4a3 | 1141 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
230880c1 | 1142 | uvchr_to_utf8(tmpbuf, c); |
4d61ec05 GS |
1143 | return is_utf8_xdigit(tmpbuf); |
1144 | } | |
1145 | ||
84afefe6 JH |
1146 | UV |
1147 | Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp) | |
a0ed51b3 | 1148 | { |
7918f24d NC |
1149 | PERL_ARGS_ASSERT_TO_UNI_UPPER; |
1150 | ||
0ebc6274 JH |
1151 | uvchr_to_utf8(p, c); |
1152 | return to_utf8_upper(p, p, lenp); | |
a0ed51b3 LW |
1153 | } |
1154 | ||
84afefe6 JH |
1155 | UV |
1156 | Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp) | |
a0ed51b3 | 1157 | { |
7918f24d NC |
1158 | PERL_ARGS_ASSERT_TO_UNI_TITLE; |
1159 | ||
0ebc6274 JH |
1160 | uvchr_to_utf8(p, c); |
1161 | return to_utf8_title(p, p, lenp); | |
a0ed51b3 LW |
1162 | } |
1163 | ||
84afefe6 JH |
1164 | UV |
1165 | Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp) | |
a0ed51b3 | 1166 | { |
7918f24d NC |
1167 | PERL_ARGS_ASSERT_TO_UNI_LOWER; |
1168 | ||
0ebc6274 JH |
1169 | uvchr_to_utf8(p, c); |
1170 | return to_utf8_lower(p, p, lenp); | |
a0ed51b3 LW |
1171 | } |
1172 | ||
84afefe6 JH |
1173 | UV |
1174 | Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp) | |
1175 | { | |
7918f24d NC |
1176 | PERL_ARGS_ASSERT_TO_UNI_FOLD; |
1177 | ||
0ebc6274 JH |
1178 | uvchr_to_utf8(p, c); |
1179 | return to_utf8_fold(p, p, lenp); | |
84afefe6 JH |
1180 | } |
1181 | ||
a0ed51b3 LW |
1182 | /* for now these all assume no locale info available for Unicode > 255 */ |
1183 | ||
1184 | bool | |
84afefe6 | 1185 | Perl_is_uni_alnum_lc(pTHX_ UV c) |
a0ed51b3 LW |
1186 | { |
1187 | return is_uni_alnum(c); /* XXX no locale support yet */ | |
1188 | } | |
1189 | ||
1190 | bool | |
84afefe6 | 1191 | Perl_is_uni_idfirst_lc(pTHX_ UV c) |
a0ed51b3 LW |
1192 | { |
1193 | return is_uni_idfirst(c); /* XXX no locale support yet */ | |
1194 | } | |
1195 | ||
1196 | bool | |
84afefe6 | 1197 | Perl_is_uni_alpha_lc(pTHX_ UV c) |
a0ed51b3 LW |
1198 | { |
1199 | return is_uni_alpha(c); /* XXX no locale support yet */ | |
1200 | } | |
1201 | ||
1202 | bool | |
84afefe6 | 1203 | Perl_is_uni_ascii_lc(pTHX_ UV c) |
4d61ec05 GS |
1204 | { |
1205 | return is_uni_ascii(c); /* XXX no locale support yet */ | |
1206 | } | |
1207 | ||
1208 | bool | |
84afefe6 | 1209 | Perl_is_uni_space_lc(pTHX_ UV c) |
a0ed51b3 LW |
1210 | { |
1211 | return is_uni_space(c); /* XXX no locale support yet */ | |
1212 | } | |
1213 | ||
1214 | bool | |
84afefe6 | 1215 | Perl_is_uni_digit_lc(pTHX_ UV c) |
a0ed51b3 LW |
1216 | { |
1217 | return is_uni_digit(c); /* XXX no locale support yet */ | |
1218 | } | |
1219 | ||
1220 | bool | |
84afefe6 | 1221 | Perl_is_uni_upper_lc(pTHX_ UV c) |
a0ed51b3 LW |
1222 | { |
1223 | return is_uni_upper(c); /* XXX no locale support yet */ | |
1224 | } | |
1225 | ||
1226 | bool | |
84afefe6 | 1227 | Perl_is_uni_lower_lc(pTHX_ UV c) |
a0ed51b3 LW |
1228 | { |
1229 | return is_uni_lower(c); /* XXX no locale support yet */ | |
1230 | } | |
1231 | ||
1232 | bool | |
84afefe6 | 1233 | Perl_is_uni_cntrl_lc(pTHX_ UV c) |
b8c5462f JH |
1234 | { |
1235 | return is_uni_cntrl(c); /* XXX no locale support yet */ | |
1236 | } | |
1237 | ||
1238 | bool | |
84afefe6 | 1239 | Perl_is_uni_graph_lc(pTHX_ UV c) |
b8c5462f JH |
1240 | { |
1241 | return is_uni_graph(c); /* XXX no locale support yet */ | |
1242 | } | |
1243 | ||
1244 | bool | |
84afefe6 | 1245 | Perl_is_uni_print_lc(pTHX_ UV c) |
a0ed51b3 LW |
1246 | { |
1247 | return is_uni_print(c); /* XXX no locale support yet */ | |
1248 | } | |
1249 | ||
b8c5462f | 1250 | bool |
84afefe6 | 1251 | Perl_is_uni_punct_lc(pTHX_ UV c) |
b8c5462f JH |
1252 | { |
1253 | return is_uni_punct(c); /* XXX no locale support yet */ | |
1254 | } | |
1255 | ||
4d61ec05 | 1256 | bool |
84afefe6 | 1257 | Perl_is_uni_xdigit_lc(pTHX_ UV c) |
4d61ec05 GS |
1258 | { |
1259 | return is_uni_xdigit(c); /* XXX no locale support yet */ | |
1260 | } | |
1261 | ||
b7ac61fa JH |
1262 | U32 |
1263 | Perl_to_uni_upper_lc(pTHX_ U32 c) | |
1264 | { | |
ee099d14 JH |
1265 | /* XXX returns only the first character -- do not use XXX */ |
1266 | /* XXX no locale support yet */ | |
1267 | STRLEN len; | |
89ebb4a3 | 1268 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 | 1269 | return (U32)to_uni_upper(c, tmpbuf, &len); |
b7ac61fa JH |
1270 | } |
1271 | ||
1272 | U32 | |
1273 | Perl_to_uni_title_lc(pTHX_ U32 c) | |
1274 | { | |
ee099d14 JH |
1275 | /* XXX returns only the first character XXX -- do not use XXX */ |
1276 | /* XXX no locale support yet */ | |
1277 | STRLEN len; | |
89ebb4a3 | 1278 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 | 1279 | return (U32)to_uni_title(c, tmpbuf, &len); |
b7ac61fa JH |
1280 | } |
1281 | ||
1282 | U32 | |
1283 | Perl_to_uni_lower_lc(pTHX_ U32 c) | |
1284 | { | |
ee099d14 JH |
1285 | /* XXX returns only the first character -- do not use XXX */ |
1286 | /* XXX no locale support yet */ | |
1287 | STRLEN len; | |
89ebb4a3 | 1288 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 | 1289 | return (U32)to_uni_lower(c, tmpbuf, &len); |
b7ac61fa JH |
1290 | } |
1291 | ||
7452cf6a | 1292 | static bool |
5141f98e | 1293 | S_is_utf8_common(pTHX_ const U8 *const p, SV **swash, |
bde6a22d NC |
1294 | const char *const swashname) |
1295 | { | |
97aff369 | 1296 | dVAR; |
7918f24d NC |
1297 | |
1298 | PERL_ARGS_ASSERT_IS_UTF8_COMMON; | |
1299 | ||
bde6a22d NC |
1300 | if (!is_utf8_char(p)) |
1301 | return FALSE; | |
1302 | if (!*swash) | |
711a919c | 1303 | *swash = swash_init("utf8", swashname, &PL_sv_undef, 1, 0); |
bde6a22d NC |
1304 | return swash_fetch(*swash, p, TRUE) != 0; |
1305 | } | |
1306 | ||
1307 | bool | |
7fc63493 | 1308 | Perl_is_utf8_alnum(pTHX_ const U8 *p) |
a0ed51b3 | 1309 | { |
97aff369 | 1310 | dVAR; |
7918f24d NC |
1311 | |
1312 | PERL_ARGS_ASSERT_IS_UTF8_ALNUM; | |
1313 | ||
671c33bf NC |
1314 | /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true |
1315 | * descendant of isalnum(3), in other words, it doesn't | |
1316 | * contain the '_'. --jhi */ | |
d4c19fe8 | 1317 | return is_utf8_common(p, &PL_utf8_alnum, "IsWord"); |
a0ed51b3 LW |
1318 | } |
1319 | ||
1320 | bool | |
7fc63493 | 1321 | Perl_is_utf8_idfirst(pTHX_ const U8 *p) /* The naming is historical. */ |
a0ed51b3 | 1322 | { |
97aff369 | 1323 | dVAR; |
7918f24d NC |
1324 | |
1325 | PERL_ARGS_ASSERT_IS_UTF8_IDFIRST; | |
1326 | ||
82686b01 JH |
1327 | if (*p == '_') |
1328 | return TRUE; | |
bde6a22d | 1329 | /* is_utf8_idstart would be more logical. */ |
d4c19fe8 | 1330 | return is_utf8_common(p, &PL_utf8_idstart, "IdStart"); |
82686b01 JH |
1331 | } |
1332 | ||
1333 | bool | |
7fc63493 | 1334 | Perl_is_utf8_idcont(pTHX_ const U8 *p) |
82686b01 | 1335 | { |
97aff369 | 1336 | dVAR; |
7918f24d NC |
1337 | |
1338 | PERL_ARGS_ASSERT_IS_UTF8_IDCONT; | |
1339 | ||
82686b01 JH |
1340 | if (*p == '_') |
1341 | return TRUE; | |
d4c19fe8 | 1342 | return is_utf8_common(p, &PL_utf8_idcont, "IdContinue"); |
a0ed51b3 LW |
1343 | } |
1344 | ||
1345 | bool | |
7fc63493 | 1346 | Perl_is_utf8_alpha(pTHX_ const U8 *p) |
a0ed51b3 | 1347 | { |
97aff369 | 1348 | dVAR; |
7918f24d NC |
1349 | |
1350 | PERL_ARGS_ASSERT_IS_UTF8_ALPHA; | |
1351 | ||
d4c19fe8 | 1352 | return is_utf8_common(p, &PL_utf8_alpha, "IsAlpha"); |
a0ed51b3 LW |
1353 | } |
1354 | ||
1355 | bool | |
7fc63493 | 1356 | Perl_is_utf8_ascii(pTHX_ const U8 *p) |
b8c5462f | 1357 | { |
97aff369 | 1358 | dVAR; |
7918f24d NC |
1359 | |
1360 | PERL_ARGS_ASSERT_IS_UTF8_ASCII; | |
1361 | ||
d4c19fe8 | 1362 | return is_utf8_common(p, &PL_utf8_ascii, "IsAscii"); |
b8c5462f JH |
1363 | } |
1364 | ||
1365 | bool | |
7fc63493 | 1366 | Perl_is_utf8_space(pTHX_ const U8 *p) |
a0ed51b3 | 1367 | { |
97aff369 | 1368 | dVAR; |
7918f24d NC |
1369 | |
1370 | PERL_ARGS_ASSERT_IS_UTF8_SPACE; | |
1371 | ||
d4c19fe8 | 1372 | return is_utf8_common(p, &PL_utf8_space, "IsSpacePerl"); |
a0ed51b3 LW |
1373 | } |
1374 | ||
1375 | bool | |
d1eb3177 YO |
1376 | Perl_is_utf8_perl_space(pTHX_ const U8 *p) |
1377 | { | |
1378 | dVAR; | |
1379 | ||
1380 | PERL_ARGS_ASSERT_IS_UTF8_PERL_SPACE; | |
1381 | ||
1382 | return is_utf8_common(p, &PL_utf8_perl_space, "IsPerlSpace"); | |
1383 | } | |
1384 | ||
1385 | bool | |
1386 | Perl_is_utf8_perl_word(pTHX_ const U8 *p) | |
1387 | { | |
1388 | dVAR; | |
1389 | ||
1390 | PERL_ARGS_ASSERT_IS_UTF8_PERL_WORD; | |
1391 | ||
1392 | return is_utf8_common(p, &PL_utf8_perl_word, "IsPerlWord"); | |
1393 | } | |
1394 | ||
1395 | bool | |
7fc63493 | 1396 | Perl_is_utf8_digit(pTHX_ const U8 *p) |
a0ed51b3 | 1397 | { |
97aff369 | 1398 | dVAR; |
7918f24d NC |
1399 | |
1400 | PERL_ARGS_ASSERT_IS_UTF8_DIGIT; | |
1401 | ||
d4c19fe8 | 1402 | return is_utf8_common(p, &PL_utf8_digit, "IsDigit"); |
a0ed51b3 LW |
1403 | } |
1404 | ||
1405 | bool | |
d1eb3177 YO |
1406 | Perl_is_utf8_posix_digit(pTHX_ const U8 *p) |
1407 | { | |
1408 | dVAR; | |
1409 | ||
1410 | PERL_ARGS_ASSERT_IS_UTF8_POSIX_DIGIT; | |
1411 | ||
1412 | return is_utf8_common(p, &PL_utf8_posix_digit, "IsPosixDigit"); | |
1413 | } | |
1414 | ||
1415 | bool | |
7fc63493 | 1416 | Perl_is_utf8_upper(pTHX_ const U8 *p) |
a0ed51b3 | 1417 | { |
97aff369 | 1418 | dVAR; |
7918f24d NC |
1419 | |
1420 | PERL_ARGS_ASSERT_IS_UTF8_UPPER; | |
1421 | ||
d4c19fe8 | 1422 | return is_utf8_common(p, &PL_utf8_upper, "IsUppercase"); |
a0ed51b3 LW |
1423 | } |
1424 | ||
1425 | bool | |
7fc63493 | 1426 | Perl_is_utf8_lower(pTHX_ const U8 *p) |
a0ed51b3 | 1427 | { |
97aff369 | 1428 | dVAR; |
7918f24d NC |
1429 | |
1430 | PERL_ARGS_ASSERT_IS_UTF8_LOWER; | |
1431 | ||
d4c19fe8 | 1432 | return is_utf8_common(p, &PL_utf8_lower, "IsLowercase"); |
a0ed51b3 LW |
1433 | } |
1434 | ||
1435 | bool | |
7fc63493 | 1436 | Perl_is_utf8_cntrl(pTHX_ const U8 *p) |
b8c5462f | 1437 | { |
97aff369 | 1438 | dVAR; |
7918f24d NC |
1439 | |
1440 | PERL_ARGS_ASSERT_IS_UTF8_CNTRL; | |
1441 | ||
d4c19fe8 | 1442 | return is_utf8_common(p, &PL_utf8_cntrl, "IsCntrl"); |
b8c5462f JH |
1443 | } |
1444 | ||
1445 | bool | |
7fc63493 | 1446 | Perl_is_utf8_graph(pTHX_ const U8 *p) |
b8c5462f | 1447 | { |
97aff369 | 1448 | dVAR; |
7918f24d NC |
1449 | |
1450 | PERL_ARGS_ASSERT_IS_UTF8_GRAPH; | |
1451 | ||
d4c19fe8 | 1452 | return is_utf8_common(p, &PL_utf8_graph, "IsGraph"); |
b8c5462f JH |
1453 | } |
1454 | ||
1455 | bool | |
7fc63493 | 1456 | Perl_is_utf8_print(pTHX_ const U8 *p) |
a0ed51b3 | 1457 | { |
97aff369 | 1458 | dVAR; |
7918f24d NC |
1459 | |
1460 | PERL_ARGS_ASSERT_IS_UTF8_PRINT; | |
1461 | ||
d4c19fe8 | 1462 | return is_utf8_common(p, &PL_utf8_print, "IsPrint"); |
a0ed51b3 LW |
1463 | } |
1464 | ||
1465 | bool | |
7fc63493 | 1466 | Perl_is_utf8_punct(pTHX_ const U8 *p) |
b8c5462f | 1467 | { |
97aff369 | 1468 | dVAR; |
7918f24d NC |
1469 | |
1470 | PERL_ARGS_ASSERT_IS_UTF8_PUNCT; | |
1471 | ||
d4c19fe8 | 1472 | return is_utf8_common(p, &PL_utf8_punct, "IsPunct"); |
b8c5462f JH |
1473 | } |
1474 | ||
1475 | bool | |
7fc63493 | 1476 | Perl_is_utf8_xdigit(pTHX_ const U8 *p) |
b8c5462f | 1477 | { |
97aff369 | 1478 | dVAR; |
7918f24d NC |
1479 | |
1480 | PERL_ARGS_ASSERT_IS_UTF8_XDIGIT; | |
1481 | ||
d1eb3177 | 1482 | return is_utf8_common(p, &PL_utf8_xdigit, "IsXDigit"); |
b8c5462f JH |
1483 | } |
1484 | ||
1485 | bool | |
7fc63493 | 1486 | Perl_is_utf8_mark(pTHX_ const U8 *p) |
a0ed51b3 | 1487 | { |
97aff369 | 1488 | dVAR; |
7918f24d NC |
1489 | |
1490 | PERL_ARGS_ASSERT_IS_UTF8_MARK; | |
1491 | ||
d4c19fe8 | 1492 | return is_utf8_common(p, &PL_utf8_mark, "IsM"); |
a0ed51b3 LW |
1493 | } |
1494 | ||
37e2e78e KW |
1495 | bool |
1496 | Perl_is_utf8_X_begin(pTHX_ const U8 *p) | |
1497 | { | |
1498 | dVAR; | |
1499 | ||
1500 | PERL_ARGS_ASSERT_IS_UTF8_X_BEGIN; | |
1501 | ||
1502 | return is_utf8_common(p, &PL_utf8_X_begin, "_X_Begin"); | |
1503 | } | |
1504 | ||
1505 | bool | |
1506 | Perl_is_utf8_X_extend(pTHX_ const U8 *p) | |
1507 | { | |
1508 | dVAR; | |
1509 | ||
1510 | PERL_ARGS_ASSERT_IS_UTF8_X_EXTEND; | |
1511 | ||
1512 | return is_utf8_common(p, &PL_utf8_X_extend, "_X_Extend"); | |
1513 | } | |
1514 | ||
1515 | bool | |
1516 | Perl_is_utf8_X_prepend(pTHX_ const U8 *p) | |
1517 | { | |
1518 | dVAR; | |
1519 | ||
1520 | PERL_ARGS_ASSERT_IS_UTF8_X_PREPEND; | |
1521 | ||
1522 | return is_utf8_common(p, &PL_utf8_X_prepend, "GCB=Prepend"); | |
1523 | } | |
1524 | ||
1525 | bool | |
1526 | Perl_is_utf8_X_non_hangul(pTHX_ const U8 *p) | |
1527 | { | |
1528 | dVAR; | |
1529 | ||
1530 | PERL_ARGS_ASSERT_IS_UTF8_X_NON_HANGUL; | |
1531 | ||
1532 | return is_utf8_common(p, &PL_utf8_X_non_hangul, "HST=Not_Applicable"); | |
1533 | } | |
1534 | ||
1535 | bool | |
1536 | Perl_is_utf8_X_L(pTHX_ const U8 *p) | |
1537 | { | |
1538 | dVAR; | |
1539 | ||
1540 | PERL_ARGS_ASSERT_IS_UTF8_X_L; | |
1541 | ||
1542 | return is_utf8_common(p, &PL_utf8_X_L, "GCB=L"); | |
1543 | } | |
1544 | ||
1545 | bool | |
1546 | Perl_is_utf8_X_LV(pTHX_ const U8 *p) | |
1547 | { | |
1548 | dVAR; | |
1549 | ||
1550 | PERL_ARGS_ASSERT_IS_UTF8_X_LV; | |
1551 | ||
1552 | return is_utf8_common(p, &PL_utf8_X_LV, "GCB=LV"); | |
1553 | } | |
1554 | ||
1555 | bool | |
1556 | Perl_is_utf8_X_LVT(pTHX_ const U8 *p) | |
1557 | { | |
1558 | dVAR; | |
1559 | ||
1560 | PERL_ARGS_ASSERT_IS_UTF8_X_LVT; | |
1561 | ||
1562 | return is_utf8_common(p, &PL_utf8_X_LVT, "GCB=LVT"); | |
1563 | } | |
1564 | ||
1565 | bool | |
1566 | Perl_is_utf8_X_T(pTHX_ const U8 *p) | |
1567 | { | |
1568 | dVAR; | |
1569 | ||
1570 | PERL_ARGS_ASSERT_IS_UTF8_X_T; | |
1571 | ||
1572 | return is_utf8_common(p, &PL_utf8_X_T, "GCB=T"); | |
1573 | } | |
1574 | ||
1575 | bool | |
1576 | Perl_is_utf8_X_V(pTHX_ const U8 *p) | |
1577 | { | |
1578 | dVAR; | |
1579 | ||
1580 | PERL_ARGS_ASSERT_IS_UTF8_X_V; | |
1581 | ||
1582 | return is_utf8_common(p, &PL_utf8_X_V, "GCB=V"); | |
1583 | } | |
1584 | ||
1585 | bool | |
1586 | Perl_is_utf8_X_LV_LVT_V(pTHX_ const U8 *p) | |
1587 | { | |
1588 | dVAR; | |
1589 | ||
1590 | PERL_ARGS_ASSERT_IS_UTF8_X_LV_LVT_V; | |
1591 | ||
1592 | return is_utf8_common(p, &PL_utf8_X_LV_LVT_V, "_X_LV_LVT_V"); | |
1593 | } | |
1594 | ||
6b5c0936 | 1595 | /* |
87cea99e | 1596 | =for apidoc to_utf8_case |
6b5c0936 JH |
1597 | |
1598 | The "p" contains the pointer to the UTF-8 string encoding | |
1599 | the character that is being converted. | |
1600 | ||
1601 | The "ustrp" is a pointer to the character buffer to put the | |
1602 | conversion result to. The "lenp" is a pointer to the length | |
1603 | of the result. | |
1604 | ||
0134edef | 1605 | The "swashp" is a pointer to the swash to use. |
6b5c0936 | 1606 | |
0134edef | 1607 | Both the special and normal mappings are stored lib/unicore/To/Foo.pl, |
8fe4d5b2 | 1608 | and loaded by SWASHNEW, using lib/utf8_heavy.pl. The special (usually, |
0134edef | 1609 | but not always, a multicharacter mapping), is tried first. |
6b5c0936 | 1610 | |
0134edef JH |
1611 | The "special" is a string like "utf8::ToSpecLower", which means the |
1612 | hash %utf8::ToSpecLower. The access to the hash is through | |
1613 | Perl_to_utf8_case(). | |
6b5c0936 | 1614 | |
0134edef JH |
1615 | The "normal" is a string like "ToLower" which means the swash |
1616 | %utf8::ToLower. | |
1617 | ||
1618 | =cut */ | |
6b5c0936 | 1619 | |
2104c8d9 | 1620 | UV |
9a957fbc AL |
1621 | Perl_to_utf8_case(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp, |
1622 | SV **swashp, const char *normal, const char *special) | |
a0ed51b3 | 1623 | { |
97aff369 | 1624 | dVAR; |
89ebb4a3 | 1625 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
0134edef | 1626 | STRLEN len = 0; |
aec46f14 | 1627 | const UV uv0 = utf8_to_uvchr(p, NULL); |
1feea2c7 JH |
1628 | /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings |
1629 | * are necessary in EBCDIC, they are redundant no-ops | |
1630 | * in ASCII-ish platforms, and hopefully optimized away. */ | |
f54cb97a | 1631 | const UV uv1 = NATIVE_TO_UNI(uv0); |
7918f24d NC |
1632 | |
1633 | PERL_ARGS_ASSERT_TO_UTF8_CASE; | |
1634 | ||
1feea2c7 | 1635 | uvuni_to_utf8(tmpbuf, uv1); |
0134edef JH |
1636 | |
1637 | if (!*swashp) /* load on-demand */ | |
1638 | *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0); | |
37e2e78e KW |
1639 | /* This is the beginnings of a skeleton of code to read the info section |
1640 | * that is in all the swashes in case we ever want to do that, so one can | |
1641 | * read things whose maps aren't code points, and whose default if missing | |
1642 | * is not to the code point itself. This was just to see if it actually | |
1643 | * worked. Details on what the possibilities are are in perluniprops.pod | |
1644 | HV * const hv = get_hv("utf8::SwashInfo", 0); | |
1645 | if (hv) { | |
1646 | SV **svp; | |
1647 | svp = hv_fetch(hv, (const char*)normal, strlen(normal), FALSE); | |
1648 | const char *s; | |
1649 | ||
1650 | HV * const this_hash = SvRV(*svp); | |
1651 | svp = hv_fetch(this_hash, "type", strlen("type"), FALSE); | |
1652 | s = SvPV_const(*svp, len); | |
1653 | } | |
1654 | }*/ | |
0134edef | 1655 | |
b08cf34e JH |
1656 | /* The 0xDF is the only special casing Unicode code point below 0x100. */ |
1657 | if (special && (uv1 == 0xDF || uv1 > 0xFF)) { | |
0134edef | 1658 | /* It might be "special" (sometimes, but not always, |
2a37f04d | 1659 | * a multicharacter mapping) */ |
6673a63c | 1660 | HV * const hv = get_hv(special, 0); |
b08cf34e JH |
1661 | SV **svp; |
1662 | ||
35da51f7 | 1663 | if (hv && |
b08cf34e JH |
1664 | (svp = hv_fetch(hv, (const char*)tmpbuf, UNISKIP(uv1), FALSE)) && |
1665 | (*svp)) { | |
cfd0369c | 1666 | const char *s; |
47654450 | 1667 | |
cfd0369c | 1668 | s = SvPV_const(*svp, len); |
47654450 JH |
1669 | if (len == 1) |
1670 | len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp; | |
2a37f04d | 1671 | else { |
2f9475ad JH |
1672 | #ifdef EBCDIC |
1673 | /* If we have EBCDIC we need to remap the characters | |
1674 | * since any characters in the low 256 are Unicode | |
1675 | * code points, not EBCDIC. */ | |
7cda7a3d | 1676 | U8 *t = (U8*)s, *tend = t + len, *d; |
2f9475ad JH |
1677 | |
1678 | d = tmpbuf; | |
b08cf34e | 1679 | if (SvUTF8(*svp)) { |
2f9475ad JH |
1680 | STRLEN tlen = 0; |
1681 | ||
1682 | while (t < tend) { | |
d4c19fe8 | 1683 | const UV c = utf8_to_uvchr(t, &tlen); |
2f9475ad JH |
1684 | if (tlen > 0) { |
1685 | d = uvchr_to_utf8(d, UNI_TO_NATIVE(c)); | |
1686 | t += tlen; | |
1687 | } | |
1688 | else | |
1689 | break; | |
1690 | } | |
1691 | } | |
1692 | else { | |
36fec512 JH |
1693 | while (t < tend) { |
1694 | d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t)); | |
1695 | t++; | |
1696 | } | |
2f9475ad JH |
1697 | } |
1698 | len = d - tmpbuf; | |
1699 | Copy(tmpbuf, ustrp, len, U8); | |
1700 | #else | |
d2dcd0fb | 1701 | Copy(s, ustrp, len, U8); |
2f9475ad | 1702 | #endif |
29e98929 | 1703 | } |
983ffd37 | 1704 | } |
0134edef JH |
1705 | } |
1706 | ||
1707 | if (!len && *swashp) { | |
d4c19fe8 AL |
1708 | const UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE); |
1709 | ||
0134edef JH |
1710 | if (uv2) { |
1711 | /* It was "normal" (a single character mapping). */ | |
d4c19fe8 | 1712 | const UV uv3 = UNI_TO_NATIVE(uv2); |
e9101d72 | 1713 | len = uvchr_to_utf8(ustrp, uv3) - ustrp; |
2a37f04d JH |
1714 | } |
1715 | } | |
1feea2c7 | 1716 | |
37e2e78e KW |
1717 | if (!len) /* Neither: just copy. In other words, there was no mapping |
1718 | defined, which means that the code point maps to itself */ | |
0134edef JH |
1719 | len = uvchr_to_utf8(ustrp, uv0) - ustrp; |
1720 | ||
2a37f04d JH |
1721 | if (lenp) |
1722 | *lenp = len; | |
1723 | ||
0134edef | 1724 | return len ? utf8_to_uvchr(ustrp, 0) : 0; |
a0ed51b3 LW |
1725 | } |
1726 | ||
d3e79532 | 1727 | /* |
87cea99e | 1728 | =for apidoc to_utf8_upper |
d3e79532 JH |
1729 | |
1730 | Convert the UTF-8 encoded character at p to its uppercase version and | |
1731 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note | |
89ebb4a3 JH |
1732 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since |
1733 | the uppercase version may be longer than the original character. | |
d3e79532 JH |
1734 | |
1735 | The first character of the uppercased version is returned | |
1736 | (but note, as explained above, that there may be more.) | |
1737 | ||
1738 | =cut */ | |
1739 | ||
2104c8d9 | 1740 | UV |
7fc63493 | 1741 | Perl_to_utf8_upper(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 | 1742 | { |
97aff369 | 1743 | dVAR; |
7918f24d NC |
1744 | |
1745 | PERL_ARGS_ASSERT_TO_UTF8_UPPER; | |
1746 | ||
983ffd37 | 1747 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 | 1748 | &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper"); |
983ffd37 | 1749 | } |
a0ed51b3 | 1750 | |
d3e79532 | 1751 | /* |
87cea99e | 1752 | =for apidoc to_utf8_title |
d3e79532 JH |
1753 | |
1754 | Convert the UTF-8 encoded character at p to its titlecase version and | |
1755 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note | |
89ebb4a3 JH |
1756 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
1757 | titlecase version may be longer than the original character. | |
d3e79532 JH |
1758 | |
1759 | The first character of the titlecased version is returned | |
1760 | (but note, as explained above, that there may be more.) | |
1761 | ||
1762 | =cut */ | |
1763 | ||
983ffd37 | 1764 | UV |
7fc63493 | 1765 | Perl_to_utf8_title(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
983ffd37 | 1766 | { |
97aff369 | 1767 | dVAR; |
7918f24d NC |
1768 | |
1769 | PERL_ARGS_ASSERT_TO_UTF8_TITLE; | |
1770 | ||
983ffd37 | 1771 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 | 1772 | &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle"); |
a0ed51b3 LW |
1773 | } |
1774 | ||
d3e79532 | 1775 | /* |
87cea99e | 1776 | =for apidoc to_utf8_lower |
d3e79532 JH |
1777 | |
1778 | Convert the UTF-8 encoded character at p to its lowercase version and | |
1779 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note | |
89ebb4a3 JH |
1780 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
1781 | lowercase version may be longer than the original character. | |
d3e79532 JH |
1782 | |
1783 | The first character of the lowercased version is returned | |
1784 | (but note, as explained above, that there may be more.) | |
1785 | ||
1786 | =cut */ | |
1787 | ||
2104c8d9 | 1788 | UV |
7fc63493 | 1789 | Perl_to_utf8_lower(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 | 1790 | { |
97aff369 | 1791 | dVAR; |
7918f24d NC |
1792 | |
1793 | PERL_ARGS_ASSERT_TO_UTF8_LOWER; | |
1794 | ||
983ffd37 | 1795 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 JH |
1796 | &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower"); |
1797 | } | |
1798 | ||
d3e79532 | 1799 | /* |
87cea99e | 1800 | =for apidoc to_utf8_fold |
d3e79532 JH |
1801 | |
1802 | Convert the UTF-8 encoded character at p to its foldcase version and | |
1803 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note | |
89ebb4a3 | 1804 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
d3e79532 JH |
1805 | foldcase version may be longer than the original character (up to |
1806 | three characters). | |
1807 | ||
1808 | The first character of the foldcased version is returned | |
1809 | (but note, as explained above, that there may be more.) | |
1810 | ||
1811 | =cut */ | |
1812 | ||
b4e400f9 | 1813 | UV |
7fc63493 | 1814 | Perl_to_utf8_fold(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
b4e400f9 | 1815 | { |
97aff369 | 1816 | dVAR; |
7918f24d NC |
1817 | |
1818 | PERL_ARGS_ASSERT_TO_UTF8_FOLD; | |
1819 | ||
b4e400f9 JH |
1820 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
1821 | &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold"); | |
a0ed51b3 LW |
1822 | } |
1823 | ||
711a919c TS |
1824 | /* Note: |
1825 | * A "swash" is a swatch hash. | |
1826 | * A "swatch" is a bit vector generated by utf8.c:S_swash_get(). | |
1827 | * C<pkg> is a pointer to a package name for SWASHNEW, should be "utf8". | |
1828 | * For other parameters, see utf8::SWASHNEW in lib/utf8_heavy.pl. | |
1829 | */ | |
a0ed51b3 | 1830 | SV* |
7fc63493 | 1831 | Perl_swash_init(pTHX_ const char* pkg, const char* name, SV *listsv, I32 minbits, I32 none) |
a0ed51b3 | 1832 | { |
27da23d5 | 1833 | dVAR; |
a0ed51b3 | 1834 | SV* retval; |
8e84507e | 1835 | dSP; |
7fc63493 AL |
1836 | const size_t pkg_len = strlen(pkg); |
1837 | const size_t name_len = strlen(name); | |
da51bb9b | 1838 | HV * const stash = gv_stashpvn(pkg, pkg_len, 0); |
f8be5cf0 | 1839 | SV* errsv_save; |
ce3b816e | 1840 | |
7918f24d NC |
1841 | PERL_ARGS_ASSERT_SWASH_INIT; |
1842 | ||
96ca9f55 DM |
1843 | PUSHSTACKi(PERLSI_MAGIC); |
1844 | ENTER; | |
ec34a119 | 1845 | SAVEHINTS(); |
96ca9f55 | 1846 | save_re_context(); |
1b026014 | 1847 | if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */ |
ce3b816e | 1848 | ENTER; |
f8be5cf0 | 1849 | errsv_save = newSVsv(ERRSV); |
dc0c6abb NC |
1850 | /* It is assumed that callers of this routine are not passing in any |
1851 | user derived data. */ | |
1852 | /* Need to do this after save_re_context() as it will set PL_tainted to | |
1853 | 1 while saving $1 etc (see the code after getrx: in Perl_magic_get). | |
1854 | Even line to create errsv_save can turn on PL_tainted. */ | |
1855 | SAVEBOOL(PL_tainted); | |
1856 | PL_tainted = 0; | |
71bed85a | 1857 | Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpvn(pkg,pkg_len), |
a0714e2c | 1858 | NULL); |
f8be5cf0 JH |
1859 | if (!SvTRUE(ERRSV)) |
1860 | sv_setsv(ERRSV, errsv_save); | |
1861 | SvREFCNT_dec(errsv_save); | |
ce3b816e GS |
1862 | LEAVE; |
1863 | } | |
1864 | SPAGAIN; | |
a0ed51b3 LW |
1865 | PUSHMARK(SP); |
1866 | EXTEND(SP,5); | |
6e449a3a MHM |
1867 | mPUSHp(pkg, pkg_len); |
1868 | mPUSHp(name, name_len); | |
a0ed51b3 | 1869 | PUSHs(listsv); |
6e449a3a MHM |
1870 | mPUSHi(minbits); |
1871 | mPUSHi(none); | |
a0ed51b3 | 1872 | PUTBACK; |
f8be5cf0 | 1873 | errsv_save = newSVsv(ERRSV); |
864dbfa3 | 1874 | if (call_method("SWASHNEW", G_SCALAR)) |
8e84507e | 1875 | retval = newSVsv(*PL_stack_sp--); |
a0ed51b3 | 1876 | else |
e24b16f9 | 1877 | retval = &PL_sv_undef; |
f8be5cf0 JH |
1878 | if (!SvTRUE(ERRSV)) |
1879 | sv_setsv(ERRSV, errsv_save); | |
1880 | SvREFCNT_dec(errsv_save); | |
a0ed51b3 LW |
1881 | LEAVE; |
1882 | POPSTACK; | |
923e4eb5 | 1883 | if (IN_PERL_COMPILETIME) { |
623e6609 | 1884 | CopHINTS_set(PL_curcop, PL_hints); |
a0ed51b3 | 1885 | } |
bc45ce41 JH |
1886 | if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) { |
1887 | if (SvPOK(retval)) | |
35c1215d | 1888 | Perl_croak(aTHX_ "Can't find Unicode property definition \"%"SVf"\"", |
be2597df | 1889 | SVfARG(retval)); |
cea2e8a9 | 1890 | Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref"); |
bc45ce41 | 1891 | } |
a0ed51b3 LW |
1892 | return retval; |
1893 | } | |
1894 | ||
035d37be JH |
1895 | |
1896 | /* This API is wrong for special case conversions since we may need to | |
1897 | * return several Unicode characters for a single Unicode character | |
1898 | * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is | |
1899 | * the lower-level routine, and it is similarly broken for returning | |
1900 | * multiple values. --jhi */ | |
979f2922 | 1901 | /* Now SWASHGET is recasted into S_swash_get in this file. */ |
680c470c TS |
1902 | |
1903 | /* Note: | |
1904 | * Returns the value of property/mapping C<swash> for the first character | |
1905 | * of the string C<ptr>. If C<do_utf8> is true, the string C<ptr> is | |
1906 | * assumed to be in utf8. If C<do_utf8> is false, the string C<ptr> is | |
1907 | * assumed to be in native 8-bit encoding. Caches the swatch in C<swash>. | |
1908 | */ | |
a0ed51b3 | 1909 | UV |
680c470c | 1910 | Perl_swash_fetch(pTHX_ SV *swash, const U8 *ptr, bool do_utf8) |
a0ed51b3 | 1911 | { |
27da23d5 | 1912 | dVAR; |
ef8f7699 | 1913 | HV *const hv = MUTABLE_HV(SvRV(swash)); |
3568d838 JH |
1914 | U32 klen; |
1915 | U32 off; | |
a0ed51b3 | 1916 | STRLEN slen; |
7d85a32c | 1917 | STRLEN needents; |
cfd0369c | 1918 | const U8 *tmps = NULL; |
a0ed51b3 | 1919 | U32 bit; |
979f2922 | 1920 | SV *swatch; |
3568d838 | 1921 | U8 tmputf8[2]; |
35da51f7 | 1922 | const UV c = NATIVE_TO_ASCII(*ptr); |
3568d838 | 1923 | |
7918f24d NC |
1924 | PERL_ARGS_ASSERT_SWASH_FETCH; |
1925 | ||
3568d838 | 1926 | if (!do_utf8 && !UNI_IS_INVARIANT(c)) { |
979f2922 TS |
1927 | tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c); |
1928 | tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c); | |
1929 | ptr = tmputf8; | |
3568d838 JH |
1930 | } |
1931 | /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ | |
37e2e78e | 1932 | * then the "swatch" is a vec() for all the chars which start |
3568d838 JH |
1933 | * with 0xAA..0xYY |
1934 | * So the key in the hash (klen) is length of encoded char -1 | |
1935 | */ | |
1936 | klen = UTF8SKIP(ptr) - 1; | |
1937 | off = ptr[klen]; | |
a0ed51b3 | 1938 | |
979f2922 | 1939 | if (klen == 0) { |
37e2e78e | 1940 | /* If char is invariant then swatch is for all the invariant chars |
1e54db1a | 1941 | * In both UTF-8 and UTF-8-MOD that happens to be UTF_CONTINUATION_MARK |
7d85a32c | 1942 | */ |
979f2922 TS |
1943 | needents = UTF_CONTINUATION_MARK; |
1944 | off = NATIVE_TO_UTF(ptr[klen]); | |
1945 | } | |
1946 | else { | |
7d85a32c | 1947 | /* If char is encoded then swatch is for the prefix */ |
979f2922 TS |
1948 | needents = (1 << UTF_ACCUMULATION_SHIFT); |
1949 | off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK; | |
1950 | } | |
7d85a32c | 1951 | |
a0ed51b3 LW |
1952 | /* |
1953 | * This single-entry cache saves about 1/3 of the utf8 overhead in test | |
1954 | * suite. (That is, only 7-8% overall over just a hash cache. Still, | |
1955 | * it's nothing to sniff at.) Pity we usually come through at least | |
1956 | * two function calls to get here... | |
1957 | * | |
1958 | * NB: this code assumes that swatches are never modified, once generated! | |
1959 | */ | |
1960 | ||
3568d838 | 1961 | if (hv == PL_last_swash_hv && |
a0ed51b3 | 1962 | klen == PL_last_swash_klen && |
27da23d5 | 1963 | (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) ) |
a0ed51b3 LW |
1964 | { |
1965 | tmps = PL_last_swash_tmps; | |
1966 | slen = PL_last_swash_slen; | |
1967 | } | |
1968 | else { | |
1969 | /* Try our second-level swatch cache, kept in a hash. */ | |
e1ec3a88 | 1970 | SV** svp = hv_fetch(hv, (const char*)ptr, klen, FALSE); |
a0ed51b3 | 1971 | |
979f2922 TS |
1972 | /* If not cached, generate it via swash_get */ |
1973 | if (!svp || !SvPOK(*svp) | |
1974 | || !(tmps = (const U8*)SvPV_const(*svp, slen))) { | |
2b9d42f0 NIS |
1975 | /* We use utf8n_to_uvuni() as we want an index into |
1976 | Unicode tables, not a native character number. | |
1977 | */ | |
aec46f14 | 1978 | const UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXBYTES, 0, |
872c91ae JH |
1979 | ckWARN(WARN_UTF8) ? |
1980 | 0 : UTF8_ALLOW_ANY); | |
680c470c | 1981 | swatch = swash_get(swash, |
979f2922 TS |
1982 | /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */ |
1983 | (klen) ? (code_point & ~(needents - 1)) : 0, | |
1984 | needents); | |
1985 | ||
923e4eb5 | 1986 | if (IN_PERL_COMPILETIME) |
623e6609 | 1987 | CopHINTS_set(PL_curcop, PL_hints); |
a0ed51b3 | 1988 | |
979f2922 | 1989 | svp = hv_store(hv, (const char *)ptr, klen, swatch, 0); |
a0ed51b3 | 1990 | |
979f2922 TS |
1991 | if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) |
1992 | || (slen << 3) < needents) | |
660a4616 | 1993 | Perl_croak(aTHX_ "panic: swash_fetch got improper swatch"); |
a0ed51b3 LW |
1994 | } |
1995 | ||
1996 | PL_last_swash_hv = hv; | |
16d8f38a | 1997 | assert(klen <= sizeof(PL_last_swash_key)); |
eac04b2e | 1998 | PL_last_swash_klen = (U8)klen; |
cfd0369c NC |
1999 | /* FIXME change interpvar.h? */ |
2000 | PL_last_swash_tmps = (U8 *) tmps; | |
a0ed51b3 LW |
2001 | PL_last_swash_slen = slen; |
2002 | if (klen) | |
2003 | Copy(ptr, PL_last_swash_key, klen, U8); | |
2004 | } | |
2005 | ||
9faf8d75 | 2006 | switch ((int)((slen << 3) / needents)) { |
a0ed51b3 LW |
2007 | case 1: |
2008 | bit = 1 << (off & 7); | |
2009 | off >>= 3; | |
2010 | return (tmps[off] & bit) != 0; | |
2011 | case 8: | |
2012 | return tmps[off]; | |
2013 | case 16: | |
2014 | off <<= 1; | |
2015 | return (tmps[off] << 8) + tmps[off + 1] ; | |
2016 | case 32: | |
2017 | off <<= 2; | |
2018 | return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ; | |
2019 | } | |
660a4616 | 2020 | Perl_croak(aTHX_ "panic: swash_fetch got swatch of unexpected bit width"); |
670f1322 | 2021 | NORETURN_FUNCTION_END; |
a0ed51b3 | 2022 | } |
2b9d42f0 | 2023 | |
979f2922 TS |
2024 | /* Note: |
2025 | * Returns a swatch (a bit vector string) for a code point sequence | |
2026 | * that starts from the value C<start> and comprises the number C<span>. | |
2027 | * A C<swash> must be an object created by SWASHNEW (see lib/utf8_heavy.pl). | |
2028 | * Should be used via swash_fetch, which will cache the swatch in C<swash>. | |
2029 | */ | |
2030 | STATIC SV* | |
2031 | S_swash_get(pTHX_ SV* swash, UV start, UV span) | |
2032 | { | |
2033 | SV *swatch; | |
711a919c | 2034 | U8 *l, *lend, *x, *xend, *s; |
979f2922 | 2035 | STRLEN lcur, xcur, scur; |
ef8f7699 | 2036 | HV *const hv = MUTABLE_HV(SvRV(swash)); |
017a3ce5 GA |
2037 | SV** const listsvp = hv_fetchs(hv, "LIST", FALSE); |
2038 | SV** const typesvp = hv_fetchs(hv, "TYPE", FALSE); | |
2039 | SV** const bitssvp = hv_fetchs(hv, "BITS", FALSE); | |
2040 | SV** const nonesvp = hv_fetchs(hv, "NONE", FALSE); | |
2041 | SV** const extssvp = hv_fetchs(hv, "EXTRAS", FALSE); | |
0bd48802 AL |
2042 | const U8* const typestr = (U8*)SvPV_nolen(*typesvp); |
2043 | const int typeto = typestr[0] == 'T' && typestr[1] == 'o'; | |
2044 | const STRLEN bits = SvUV(*bitssvp); | |
2045 | const STRLEN octets = bits >> 3; /* if bits == 1, then octets == 0 */ | |
2046 | const UV none = SvUV(*nonesvp); | |
2047 | const UV end = start + span; | |
979f2922 | 2048 | |
7918f24d NC |
2049 | PERL_ARGS_ASSERT_SWASH_GET; |
2050 | ||
979f2922 | 2051 | if (bits != 1 && bits != 8 && bits != 16 && bits != 32) { |
660a4616 TS |
2052 | Perl_croak(aTHX_ "panic: swash_get doesn't expect bits %"UVuf, |
2053 | (UV)bits); | |
979f2922 TS |
2054 | } |
2055 | ||
2056 | /* create and initialize $swatch */ | |
979f2922 | 2057 | scur = octets ? (span * octets) : (span + 7) / 8; |
e524fe40 NC |
2058 | swatch = newSV(scur); |
2059 | SvPOK_on(swatch); | |
979f2922 TS |
2060 | s = (U8*)SvPVX(swatch); |
2061 | if (octets && none) { | |
0bd48802 | 2062 | const U8* const e = s + scur; |
979f2922 TS |
2063 | while (s < e) { |
2064 | if (bits == 8) | |
2065 | *s++ = (U8)(none & 0xff); | |
2066 | else if (bits == 16) { | |
2067 | *s++ = (U8)((none >> 8) & 0xff); | |
2068 | *s++ = (U8)( none & 0xff); | |
2069 | } | |
2070 | else if (bits == 32) { | |
2071 | *s++ = (U8)((none >> 24) & 0xff); | |
2072 | *s++ = (U8)((none >> 16) & 0xff); | |
2073 | *s++ = (U8)((none >> 8) & 0xff); | |
2074 | *s++ = (U8)( none & 0xff); | |
2075 | } | |
2076 | } | |
2077 | *s = '\0'; | |
2078 | } | |
2079 | else { | |
2080 | (void)memzero((U8*)s, scur + 1); | |
2081 | } | |
2082 | SvCUR_set(swatch, scur); | |
2083 | s = (U8*)SvPVX(swatch); | |
2084 | ||
2085 | /* read $swash->{LIST} */ | |
2086 | l = (U8*)SvPV(*listsvp, lcur); | |
2087 | lend = l + lcur; | |
2088 | while (l < lend) { | |
35da51f7 | 2089 | UV min, max, val; |
979f2922 TS |
2090 | STRLEN numlen; |
2091 | I32 flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX; | |
2092 | ||
0bd48802 | 2093 | U8* const nl = (U8*)memchr(l, '\n', lend - l); |
979f2922 TS |
2094 | |
2095 | numlen = lend - l; | |
2096 | min = grok_hex((char *)l, &numlen, &flags, NULL); | |
2097 | if (numlen) | |
2098 | l += numlen; | |
2099 | else if (nl) { | |
2100 | l = nl + 1; /* 1 is length of "\n" */ | |
2101 | continue; | |
2102 | } | |
2103 | else { | |
2104 | l = lend; /* to LIST's end at which \n is not found */ | |
2105 | break; | |
2106 | } | |
2107 | ||
2108 | if (isBLANK(*l)) { | |
2109 | ++l; | |
2110 | flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX; | |
2111 | numlen = lend - l; | |
2112 | max = grok_hex((char *)l, &numlen, &flags, NULL); | |
2113 | if (numlen) | |
2114 | l += numlen; | |
2115 | else | |
2116 | max = min; | |
2117 | ||
2118 | if (octets) { | |
2119 | if (isBLANK(*l)) { | |
2120 | ++l; | |
2121 | flags = PERL_SCAN_SILENT_ILLDIGIT | | |
2122 | PERL_SCAN_DISALLOW_PREFIX; | |
2123 | numlen = lend - l; | |
2124 | val = grok_hex((char *)l, &numlen, &flags, NULL); | |
2125 | if (numlen) | |
2126 | l += numlen; | |
2127 | else | |
2128 | val = 0; | |
2129 | } | |
2130 | else { | |
2131 | val = 0; | |
2132 | if (typeto) { | |
2133 | Perl_croak(aTHX_ "%s: illegal mapping '%s'", | |
2134 | typestr, l); | |
2135 | } | |
2136 | } | |
2137 | } | |
711a919c TS |
2138 | else |
2139 | val = 0; /* bits == 1, then val should be ignored */ | |
979f2922 TS |
2140 | } |
2141 | else { | |
2142 | max = min; | |
2143 | if (octets) { | |
2144 | val = 0; | |
2145 | if (typeto) { | |
2146 | Perl_croak(aTHX_ "%s: illegal mapping '%s'", typestr, l); | |
2147 | } | |
2148 | } | |
711a919c TS |
2149 | else |
2150 | val = 0; /* bits == 1, then val should be ignored */ | |
979f2922 TS |
2151 | } |
2152 | ||
2153 | if (nl) | |
2154 | l = nl + 1; | |
2155 | else | |
2156 | l = lend; | |
2157 | ||
2158 | if (max < start) | |
2159 | continue; | |
2160 | ||
2161 | if (octets) { | |
35da51f7 | 2162 | UV key; |
979f2922 TS |
2163 | if (min < start) { |
2164 | if (!none || val < none) { | |
2165 | val += start - min; | |
2166 | } | |
2167 | min = start; | |
2168 | } | |
2169 | for (key = min; key <= max; key++) { | |
2170 | STRLEN offset; | |
2171 | if (key >= end) | |
2172 | goto go_out_list; | |
2173 | /* offset must be non-negative (start <= min <= key < end) */ | |
2174 | offset = octets * (key - start); | |
2175 | if (bits == 8) | |
2176 | s[offset] = (U8)(val & 0xff); | |
2177 | else if (bits == 16) { | |
2178 | s[offset ] = (U8)((val >> 8) & 0xff); | |
2179 | s[offset + 1] = (U8)( val & 0xff); | |
2180 | } | |
2181 | else if (bits == 32) { | |
2182 | s[offset ] = (U8)((val >> 24) & 0xff); | |
2183 | s[offset + 1] = (U8)((val >> 16) & 0xff); | |
2184 | s[offset + 2] = (U8)((val >> 8) & 0xff); | |
2185 | s[offset + 3] = (U8)( val & 0xff); | |
2186 | } | |
2187 | ||
2188 | if (!none || val < none) | |
2189 | ++val; | |
2190 | } | |
2191 | } | |
711a919c | 2192 | else { /* bits == 1, then val should be ignored */ |
35da51f7 | 2193 | UV key; |
979f2922 TS |
2194 | if (min < start) |
2195 | min = start; | |
2196 | for (key = min; key <= max; key++) { | |
0bd48802 | 2197 | const STRLEN offset = (STRLEN)(key - start); |
979f2922 TS |
2198 | if (key >= end) |
2199 | goto go_out_list; | |
2200 | s[offset >> 3] |= 1 << (offset & 7); | |
2201 | } | |
2202 | } | |
2203 | } /* while */ | |
2204 | go_out_list: | |
2205 | ||
2206 | /* read $swash->{EXTRAS} */ | |
2207 | x = (U8*)SvPV(*extssvp, xcur); | |
2208 | xend = x + xcur; | |
2209 | while (x < xend) { | |
2210 | STRLEN namelen; | |
2211 | U8 *namestr; | |
2212 | SV** othersvp; | |
2213 | HV* otherhv; | |
2214 | STRLEN otherbits; | |
2215 | SV **otherbitssvp, *other; | |
711a919c | 2216 | U8 *s, *o, *nl; |
979f2922 TS |
2217 | STRLEN slen, olen; |
2218 | ||
35da51f7 | 2219 | const U8 opc = *x++; |
979f2922 TS |
2220 | if (opc == '\n') |
2221 | continue; | |
2222 | ||
2223 | nl = (U8*)memchr(x, '\n', xend - x); | |
2224 | ||
2225 | if (opc != '-' && opc != '+' && opc != '!' && opc != '&') { | |
2226 | if (nl) { | |
2227 | x = nl + 1; /* 1 is length of "\n" */ | |
2228 | continue; | |
2229 | } | |
2230 | else { | |
2231 | x = xend; /* to EXTRAS' end at which \n is not found */ | |
2232 | break; | |
2233 | } | |
2234 | } | |
2235 | ||
2236 | namestr = x; | |
2237 | if (nl) { | |
2238 | namelen = nl - namestr; | |
2239 | x = nl + 1; | |
2240 | } | |
2241 | else { | |
2242 | namelen = xend - namestr; | |
2243 | x = xend; | |
2244 | } | |
2245 | ||
2246 | othersvp = hv_fetch(hv, (char *)namestr, namelen, FALSE); | |
ef8f7699 | 2247 | otherhv = MUTABLE_HV(SvRV(*othersvp)); |
017a3ce5 | 2248 | otherbitssvp = hv_fetchs(otherhv, "BITS", FALSE); |
979f2922 TS |
2249 | otherbits = (STRLEN)SvUV(*otherbitssvp); |
2250 | if (bits < otherbits) | |
660a4616 | 2251 | Perl_croak(aTHX_ "panic: swash_get found swatch size mismatch"); |
979f2922 TS |
2252 | |
2253 | /* The "other" swatch must be destroyed after. */ | |
2254 | other = swash_get(*othersvp, start, span); | |
2255 | o = (U8*)SvPV(other, olen); | |
2256 | ||
2257 | if (!olen) | |
660a4616 | 2258 | Perl_croak(aTHX_ "panic: swash_get got improper swatch"); |
979f2922 TS |
2259 | |
2260 | s = (U8*)SvPV(swatch, slen); | |
2261 | if (bits == 1 && otherbits == 1) { | |
2262 | if (slen != olen) | |
660a4616 | 2263 | Perl_croak(aTHX_ "panic: swash_get found swatch length mismatch"); |
979f2922 TS |
2264 | |
2265 | switch (opc) { | |
2266 | case '+': | |
2267 | while (slen--) | |
2268 | *s++ |= *o++; | |
2269 | break; | |
2270 | case '!': | |
2271 | while (slen--) | |
2272 | *s++ |= ~*o++; | |
2273 | break; | |
2274 | case '-': | |
2275 | while (slen--) | |
2276 | *s++ &= ~*o++; | |
2277 | break; | |
2278 | case '&': | |
2279 | while (slen--) | |
2280 | *s++ &= *o++; | |
2281 | break; | |
2282 | default: | |
2283 | break; | |
2284 | } | |
2285 | } | |
711a919c | 2286 | else { |
979f2922 TS |
2287 | STRLEN otheroctets = otherbits >> 3; |
2288 | STRLEN offset = 0; | |
35da51f7 | 2289 | U8* const send = s + slen; |
979f2922 TS |
2290 | |
2291 | while (s < send) { | |
2292 | UV otherval = 0; | |
2293 | ||
2294 | if (otherbits == 1) { | |
2295 | otherval = (o[offset >> 3] >> (offset & 7)) & 1; | |
2296 | ++offset; | |
2297 | } | |
2298 | else { | |
2299 | STRLEN vlen = otheroctets; | |
2300 | otherval = *o++; | |
2301 | while (--vlen) { | |
2302 | otherval <<= 8; | |
2303 | otherval |= *o++; | |
2304 | } | |
2305 | } | |
2306 | ||
711a919c | 2307 | if (opc == '+' && otherval) |
6f207bd3 | 2308 | NOOP; /* replace with otherval */ |
979f2922 TS |
2309 | else if (opc == '!' && !otherval) |
2310 | otherval = 1; | |
2311 | else if (opc == '-' && otherval) | |
2312 | otherval = 0; | |
2313 | else if (opc == '&' && !otherval) | |
2314 | otherval = 0; | |
2315 | else { | |
711a919c | 2316 | s += octets; /* no replacement */ |
979f2922 TS |
2317 | continue; |
2318 | } | |
2319 | ||
2320 | if (bits == 8) | |
2321 | *s++ = (U8)( otherval & 0xff); | |
2322 | else if (bits == 16) { | |
2323 | *s++ = (U8)((otherval >> 8) & 0xff); | |
2324 | *s++ = (U8)( otherval & 0xff); | |
2325 | } | |
2326 | else if (bits == 32) { | |
2327 | *s++ = (U8)((otherval >> 24) & 0xff); | |
2328 | *s++ = (U8)((otherval >> 16) & 0xff); | |
2329 | *s++ = (U8)((otherval >> 8) & 0xff); | |
2330 | *s++ = (U8)( otherval & 0xff); | |
2331 | } | |
2332 | } | |
2333 | } | |
2334 | sv_free(other); /* through with it! */ | |
2335 | } /* while */ | |
2336 | return swatch; | |
2337 | } | |
2338 | ||
0f830e0b | 2339 | /* |
87cea99e | 2340 | =for apidoc uvchr_to_utf8 |
0f830e0b NC |
2341 | |
2342 | Adds the UTF-8 representation of the Native codepoint C<uv> to the end | |
2343 | of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free | |
2344 | bytes available. The return value is the pointer to the byte after the | |
2345 | end of the new character. In other words, | |
2346 | ||
2347 | d = uvchr_to_utf8(d, uv); | |
2348 | ||
2349 | is the recommended wide native character-aware way of saying | |
2350 | ||
2351 | *(d++) = uv; | |
2352 | ||
2353 | =cut | |
2354 | */ | |
2355 | ||
2356 | /* On ASCII machines this is normally a macro but we want a | |
2357 | real function in case XS code wants it | |
2358 | */ | |
0f830e0b NC |
2359 | U8 * |
2360 | Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv) | |
2361 | { | |
7918f24d NC |
2362 | PERL_ARGS_ASSERT_UVCHR_TO_UTF8; |
2363 | ||
0f830e0b NC |
2364 | return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0); |
2365 | } | |
2366 | ||
b851fbc1 JH |
2367 | U8 * |
2368 | Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) | |
2369 | { | |
7918f24d NC |
2370 | PERL_ARGS_ASSERT_UVCHR_TO_UTF8_FLAGS; |
2371 | ||
b851fbc1 JH |
2372 | return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags); |
2373 | } | |
2b9d42f0 NIS |
2374 | |
2375 | /* | |
87cea99e | 2376 | =for apidoc utf8n_to_uvchr |
0f830e0b NC |
2377 | flags |
2378 | ||
2379 | Returns the native character value of the first character in the string | |
2380 | C<s> | |
2381 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the | |
2382 | length, in bytes, of that character. | |
2383 | ||
2384 | Allows length and flags to be passed to low level routine. | |
2385 | ||
2386 | =cut | |
2387 | */ | |
2388 | /* On ASCII machines this is normally a macro but we want | |
2389 | a real function in case XS code wants it | |
2390 | */ | |
0f830e0b NC |
2391 | UV |
2392 | Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, | |
2393 | U32 flags) | |
2394 | { | |
2395 | const UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags); | |
7918f24d NC |
2396 | |
2397 | PERL_ARGS_ASSERT_UTF8N_TO_UVCHR; | |
2398 | ||
0f830e0b NC |
2399 | return UNI_TO_NATIVE(uv); |
2400 | } | |
2401 | ||
2402 | /* | |
87cea99e | 2403 | =for apidoc pv_uni_display |
d2cc3551 JH |
2404 | |
2405 | Build to the scalar dsv a displayable version of the string spv, | |
2406 | length len, the displayable version being at most pvlim bytes long | |
2407 | (if longer, the rest is truncated and "..." will be appended). | |
0a2ef054 | 2408 | |
9e55ce06 | 2409 | The flags argument can have UNI_DISPLAY_ISPRINT set to display |
00e86452 | 2410 | isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH |
0a2ef054 JH |
2411 | to display the \\[nrfta\\] as the backslashed versions (like '\n') |
2412 | (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\). | |
2413 | UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both | |
2414 | UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on. | |
2415 | ||
d2cc3551 JH |
2416 | The pointer to the PV of the dsv is returned. |
2417 | ||
2418 | =cut */ | |
e6b2e755 | 2419 | char * |
e1ec3a88 | 2420 | Perl_pv_uni_display(pTHX_ SV *dsv, const U8 *spv, STRLEN len, STRLEN pvlim, UV flags) |
e6b2e755 JH |
2421 | { |
2422 | int truncated = 0; | |
e1ec3a88 | 2423 | const char *s, *e; |
e6b2e755 | 2424 | |
7918f24d NC |
2425 | PERL_ARGS_ASSERT_PV_UNI_DISPLAY; |
2426 | ||
76f68e9b | 2427 | sv_setpvs(dsv, ""); |
7fddd944 | 2428 | SvUTF8_off(dsv); |
e1ec3a88 | 2429 | for (s = (const char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) { |
e6b2e755 | 2430 | UV u; |
a49f32c6 NC |
2431 | /* This serves double duty as a flag and a character to print after |
2432 | a \ when flags & UNI_DISPLAY_BACKSLASH is true. | |
2433 | */ | |
2434 | char ok = 0; | |
c728cb41 | 2435 | |
e6b2e755 JH |
2436 | if (pvlim && SvCUR(dsv) >= pvlim) { |
2437 | truncated++; | |
2438 | break; | |
2439 | } | |
2440 | u = utf8_to_uvchr((U8*)s, 0); | |
c728cb41 | 2441 | if (u < 256) { |
a3b680e6 | 2442 | const unsigned char c = (unsigned char)u & 0xFF; |
0bd48802 | 2443 | if (flags & UNI_DISPLAY_BACKSLASH) { |
a49f32c6 | 2444 | switch (c) { |
c728cb41 | 2445 | case '\n': |
a49f32c6 | 2446 | ok = 'n'; break; |
c728cb41 | 2447 | case '\r': |
a49f32c6 | 2448 | ok = 'r'; break; |
c728cb41 | 2449 | case '\t': |
a49f32c6 | 2450 | ok = 't'; break; |
c728cb41 | 2451 | case '\f': |
a49f32c6 | 2452 | ok = 'f'; break; |
c728cb41 | 2453 | case '\a': |
a49f32c6 | 2454 | ok = 'a'; break; |
c728cb41 | 2455 | case '\\': |
a49f32c6 | 2456 | ok = '\\'; break; |
c728cb41 JH |
2457 | default: break; |
2458 | } | |
a49f32c6 | 2459 | if (ok) { |
88c9ea1e | 2460 | const char string = ok; |
76f68e9b | 2461 | sv_catpvs(dsv, "\\"); |
5e7aa789 | 2462 | sv_catpvn(dsv, &string, 1); |
a49f32c6 | 2463 | } |
c728cb41 | 2464 | } |
00e86452 | 2465 | /* isPRINT() is the locale-blind version. */ |
a49f32c6 | 2466 | if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(c)) { |
88c9ea1e | 2467 | const char string = c; |
5e7aa789 | 2468 | sv_catpvn(dsv, &string, 1); |
a49f32c6 | 2469 | ok = 1; |
0a2ef054 | 2470 | } |
c728cb41 JH |
2471 | } |
2472 | if (!ok) | |
9e55ce06 | 2473 | Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u); |
e6b2e755 JH |
2474 | } |
2475 | if (truncated) | |
396482e1 | 2476 | sv_catpvs(dsv, "..."); |
e6b2e755 JH |
2477 | |
2478 | return SvPVX(dsv); | |
2479 | } | |
2b9d42f0 | 2480 | |
d2cc3551 | 2481 | /* |
87cea99e | 2482 | =for apidoc sv_uni_display |
d2cc3551 JH |
2483 | |
2484 | Build to the scalar dsv a displayable version of the scalar sv, | |
0a2ef054 | 2485 | the displayable version being at most pvlim bytes long |
d2cc3551 | 2486 | (if longer, the rest is truncated and "..." will be appended). |
0a2ef054 JH |
2487 | |
2488 | The flags argument is as in pv_uni_display(). | |
2489 | ||
d2cc3551 JH |
2490 | The pointer to the PV of the dsv is returned. |
2491 | ||
d4c19fe8 AL |
2492 | =cut |
2493 | */ | |
e6b2e755 JH |
2494 | char * |
2495 | Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags) | |
2496 | { | |
7918f24d NC |
2497 | PERL_ARGS_ASSERT_SV_UNI_DISPLAY; |
2498 | ||
cfd0369c NC |
2499 | return Perl_pv_uni_display(aTHX_ dsv, (const U8*)SvPVX_const(ssv), |
2500 | SvCUR(ssv), pvlim, flags); | |
701a277b JH |
2501 | } |
2502 | ||
d2cc3551 | 2503 | /* |
87cea99e | 2504 | =for apidoc ibcmp_utf8 |
d2cc3551 JH |
2505 | |
2506 | Return true if the strings s1 and s2 differ case-insensitively, false | |
2507 | if not (if they are equal case-insensitively). If u1 is true, the | |
2508 | string s1 is assumed to be in UTF-8-encoded Unicode. If u2 is true, | |
d07ddd77 JH |
2509 | the string s2 is assumed to be in UTF-8-encoded Unicode. If u1 or u2 |
2510 | are false, the respective string is assumed to be in native 8-bit | |
2511 | encoding. | |
2512 | ||
2513 | If the pe1 and pe2 are non-NULL, the scanning pointers will be copied | |
2514 | in there (they will point at the beginning of the I<next> character). | |
2515 | If the pointers behind pe1 or pe2 are non-NULL, they are the end | |
2516 | pointers beyond which scanning will not continue under any | |
4cdaeff7 | 2517 | circumstances. If the byte lengths l1 and l2 are non-zero, s1+l1 and |
d07ddd77 JH |
2518 | s2+l2 will be used as goal end pointers that will also stop the scan, |
2519 | and which qualify towards defining a successful match: all the scans | |
2520 | that define an explicit length must reach their goal pointers for | |
2521 | a match to succeed). | |
d2cc3551 JH |
2522 | |
2523 | For case-insensitiveness, the "casefolding" of Unicode is used | |
2524 | instead of upper/lowercasing both the characters, see | |
2525 | http://www.unicode.org/unicode/reports/tr21/ (Case Mappings). | |
2526 | ||
2527 | =cut */ | |
701a277b | 2528 | I32 |
d07ddd77 | 2529 | Perl_ibcmp_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2) |
332ddc25 | 2530 | { |
97aff369 | 2531 | dVAR; |
e1ec3a88 AL |
2532 | register const U8 *p1 = (const U8*)s1; |
2533 | register const U8 *p2 = (const U8*)s2; | |
cbbf8932 | 2534 | register const U8 *f1 = NULL; |
2f73348c | 2535 | register const U8 *f2 = NULL; |
cbbf8932 AL |
2536 | register U8 *e1 = NULL; |
2537 | register U8 *q1 = NULL; | |
2538 | register U8 *e2 = NULL; | |
2539 | register U8 *q2 = NULL; | |
d07ddd77 | 2540 | STRLEN n1 = 0, n2 = 0; |
89ebb4a3 JH |
2541 | U8 foldbuf1[UTF8_MAXBYTES_CASE+1]; |
2542 | U8 foldbuf2[UTF8_MAXBYTES_CASE+1]; | |
d7f013c8 JH |
2543 | U8 natbuf[1+1]; |
2544 | STRLEN foldlen1, foldlen2; | |
d07ddd77 | 2545 | bool match; |
7918f24d NC |
2546 | |
2547 | PERL_ARGS_ASSERT_IBCMP_UTF8; | |
332ddc25 | 2548 | |
d07ddd77 JH |
2549 | if (pe1) |
2550 | e1 = *(U8**)pe1; | |
a0a388a1 | 2551 | /* assert(e1 || l1); */ |
e1ec3a88 AL |
2552 | if (e1 == 0 || (l1 && l1 < (UV)(e1 - (const U8*)s1))) |
2553 | f1 = (const U8*)s1 + l1; | |
d07ddd77 JH |
2554 | if (pe2) |
2555 | e2 = *(U8**)pe2; | |
a0a388a1 | 2556 | /* assert(e2 || l2); */ |
e1ec3a88 AL |
2557 | if (e2 == 0 || (l2 && l2 < (UV)(e2 - (const U8*)s2))) |
2558 | f2 = (const U8*)s2 + l2; | |
d07ddd77 | 2559 | |
a0a388a1 YO |
2560 | /* This shouldn't happen. However, putting an assert() there makes some |
2561 | * tests fail. */ | |
2562 | /* assert((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0)); */ | |
d07ddd77 JH |
2563 | if ((e1 == 0 && f1 == 0) || (e2 == 0 && f2 == 0) || (f1 == 0 && f2 == 0)) |
2564 | return 1; /* mismatch; possible infinite loop or false positive */ | |
2565 | ||
a6872d42 JH |
2566 | if (!u1 || !u2) |
2567 | natbuf[1] = 0; /* Need to terminate the buffer. */ | |
2568 | ||
d07ddd77 JH |
2569 | while ((e1 == 0 || p1 < e1) && |
2570 | (f1 == 0 || p1 < f1) && | |
2571 | (e2 == 0 || p2 < e2) && | |
2572 | (f2 == 0 || p2 < f2)) { | |
2573 | if (n1 == 0) { | |
d7f013c8 JH |
2574 | if (u1) |
2575 | to_utf8_fold(p1, foldbuf1, &foldlen1); | |
2576 | else { | |
809e8e66 | 2577 | uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p1))); |
d7f013c8 JH |
2578 | to_utf8_fold(natbuf, foldbuf1, &foldlen1); |
2579 | } | |
2580 | q1 = foldbuf1; | |
d07ddd77 | 2581 | n1 = foldlen1; |
332ddc25 | 2582 | } |
d07ddd77 | 2583 | if (n2 == 0) { |
d7f013c8 JH |
2584 | if (u2) |
2585 | to_utf8_fold(p2, foldbuf2, &foldlen2); | |
2586 | else { | |
809e8e66 | 2587 | uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p2))); |
d7f013c8 JH |
2588 | to_utf8_fold(natbuf, foldbuf2, &foldlen2); |
2589 | } | |
2590 | q2 = foldbuf2; | |
d07ddd77 | 2591 | n2 = foldlen2; |
332ddc25 | 2592 | } |
d07ddd77 JH |
2593 | while (n1 && n2) { |
2594 | if ( UTF8SKIP(q1) != UTF8SKIP(q2) || | |
2595 | (UTF8SKIP(q1) == 1 && *q1 != *q2) || | |
2596 | memNE((char*)q1, (char*)q2, UTF8SKIP(q1)) ) | |
d7f013c8 | 2597 | return 1; /* mismatch */ |
d07ddd77 | 2598 | n1 -= UTF8SKIP(q1); |
d7f013c8 | 2599 | q1 += UTF8SKIP(q1); |
d07ddd77 | 2600 | n2 -= UTF8SKIP(q2); |
d7f013c8 | 2601 | q2 += UTF8SKIP(q2); |
701a277b | 2602 | } |
d07ddd77 | 2603 | if (n1 == 0) |
d7f013c8 | 2604 | p1 += u1 ? UTF8SKIP(p1) : 1; |
d07ddd77 | 2605 | if (n2 == 0) |
d7f013c8 JH |
2606 | p2 += u2 ? UTF8SKIP(p2) : 1; |
2607 | ||
d2cc3551 | 2608 | } |
5469e704 | 2609 | |
d07ddd77 JH |
2610 | /* A match is defined by all the scans that specified |
2611 | * an explicit length reaching their final goals. */ | |
7dcb3b25 KW |
2612 | match = (n1 == 0 && n2 == 0 /* Must not match partial char; Bug #72998 */ |
2613 | && (f1 == 0 || p1 == f1) && (f2 == 0 || p2 == f2)); | |
5469e704 JH |
2614 | |
2615 | if (match) { | |
d07ddd77 JH |
2616 | if (pe1) |
2617 | *pe1 = (char*)p1; | |
2618 | if (pe2) | |
2619 | *pe2 = (char*)p2; | |
5469e704 JH |
2620 | } |
2621 | ||
2622 | return match ? 0 : 1; /* 0 match, 1 mismatch */ | |
e6b2e755 | 2623 | } |
701a277b | 2624 | |
a49f32c6 NC |
2625 | /* |
2626 | * Local variables: | |
2627 | * c-indentation-style: bsd | |
2628 | * c-basic-offset: 4 | |
2629 | * indent-tabs-mode: t | |
2630 | * End: | |
2631 | * | |
37442d52 RGS |
2632 | * ex: set ts=8 sts=4 sw=4 noet: |
2633 | */ |