Commit | Line | Data |
a0ed51b3 |
1 | /* utf8.c |
2 | * |
1129b882 |
3 | * Copyright (C) 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008 |
b94e2f88 |
4 | * by Larry Wall and others |
a0ed51b3 |
5 | * |
6 | * You may distribute under the terms of either the GNU General Public |
7 | * License or the Artistic License, as specified in the README file. |
8 | * |
9 | */ |
10 | |
11 | /* |
4ac71550 |
12 | * 'What a fix!' said Sam. 'That's the one place in all the lands we've ever |
13 | * heard of that we don't want to see any closer; and that's the one place |
14 | * we're trying to get to! And that's just where we can't get, nohow.' |
15 | * |
16 | * [p.603 of _The Lord of the Rings_, IV/I: "The Taming of Sméagol"] |
a0ed51b3 |
17 | * |
18 | * 'Well do I understand your speech,' he answered in the same language; |
19 | * 'yet few strangers do so. Why then do you not speak in the Common Tongue, |
4ac71550 |
20 | * as is the custom in the West, if you wish to be answered?' |
21 | * --Gandalf, addressing Théoden's door wardens |
22 | * |
23 | * [p.508 of _The Lord of the Rings_, III/vi: "The King of the Golden Hall"] |
a0ed51b3 |
24 | * |
25 | * ...the travellers perceived that the floor was paved with stones of many |
26 | * hues; branching runes and strange devices intertwined beneath their feet. |
4ac71550 |
27 | * |
28 | * [p.512 of _The Lord of the Rings_, III/vi: "The King of the Golden Hall"] |
a0ed51b3 |
29 | */ |
30 | |
31 | #include "EXTERN.h" |
864dbfa3 |
32 | #define PERL_IN_UTF8_C |
a0ed51b3 |
33 | #include "perl.h" |
34 | |
a0c21aa1 |
35 | #ifndef EBCDIC |
970ea3cb |
36 | /* Separate prototypes needed because in ASCII systems these are |
a0c21aa1 |
37 | * usually macros but they still are compiled as code, too. */ |
38 | PERL_CALLCONV UV Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags); |
39 | PERL_CALLCONV U8* Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv); |
40 | #endif |
41 | |
27da23d5 |
42 | static const char unees[] = |
43 | "Malformed UTF-8 character (unexpected end of string)"; |
901b21bf |
44 | |
48ef279e |
45 | /* |
ccfc67b7 |
46 | =head1 Unicode Support |
a0ed51b3 |
47 | |
166f8a29 |
48 | This file contains various utility functions for manipulating UTF8-encoded |
49 | strings. For the uninitiated, this is a method of representing arbitrary |
61296642 |
50 | Unicode characters as a variable number of bytes, in such a way that |
56da48f7 |
51 | characters in the ASCII range are unmodified, and a zero byte never appears |
52 | within non-zero characters. |
166f8a29 |
53 | |
eaf7a4d2 |
54 | =cut |
55 | */ |
56 | |
57 | /* |
58 | =for apidoc is_ascii_string |
59 | |
970ea3cb |
60 | Returns true if the first C<len> bytes of the given string are the same whether |
61 | or not the string is encoded in UTF-8 (or UTF-EBCDIC on EBCDIC machines). That |
62 | is, if they are invariant. On ASCII-ish machines, only ASCII characters |
63 | fit this definition, hence the function's name. |
eaf7a4d2 |
64 | |
65 | See also is_utf8_string(), is_utf8_string_loclen(), and is_utf8_string_loc(). |
66 | |
67 | =cut |
68 | */ |
69 | |
70 | bool |
668b6d8d |
71 | Perl_is_ascii_string(const U8 *s, STRLEN len) |
eaf7a4d2 |
72 | { |
73 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
74 | const U8* x = s; |
75 | |
76 | PERL_ARGS_ASSERT_IS_ASCII_STRING; |
eaf7a4d2 |
77 | |
78 | for (; x < send; ++x) { |
79 | if (!UTF8_IS_INVARIANT(*x)) |
80 | break; |
81 | } |
82 | |
83 | return x == send; |
84 | } |
85 | |
86 | /* |
87cea99e |
87 | =for apidoc uvuni_to_utf8_flags |
eebe1485 |
88 | |
1e54db1a |
89 | Adds the UTF-8 representation of the Unicode codepoint C<uv> to the end |
89ebb4a3 |
90 | of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free |
eebe1485 |
91 | bytes available. The return value is the pointer to the byte after the |
9041c2e3 |
92 | end of the new character. In other words, |
eebe1485 |
93 | |
b851fbc1 |
94 | d = uvuni_to_utf8_flags(d, uv, flags); |
95 | |
96 | or, in most cases, |
97 | |
9041c2e3 |
98 | d = uvuni_to_utf8(d, uv); |
eebe1485 |
99 | |
b851fbc1 |
100 | (which is equivalent to) |
101 | |
102 | d = uvuni_to_utf8_flags(d, uv, 0); |
103 | |
eebe1485 |
104 | is the recommended Unicode-aware way of saying |
105 | |
106 | *(d++) = uv; |
107 | |
108 | =cut |
109 | */ |
110 | |
dfe13c55 |
111 | U8 * |
b851fbc1 |
112 | Perl_uvuni_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) |
a0ed51b3 |
113 | { |
7918f24d |
114 | PERL_ARGS_ASSERT_UVUNI_TO_UTF8_FLAGS; |
115 | |
62961d2e |
116 | if (ckWARN(WARN_UTF8)) { |
b851fbc1 |
117 | if (UNICODE_IS_SURROGATE(uv) && |
118 | !(flags & UNICODE_ALLOW_SURROGATE)) |
9014280d |
119 | Perl_warner(aTHX_ packWARN(WARN_UTF8), "UTF-16 surrogate 0x%04"UVxf, uv); |
b851fbc1 |
120 | else if ( |
121 | ((uv >= 0xFDD0 && uv <= 0xFDEF && |
122 | !(flags & UNICODE_ALLOW_FDD0)) |
123 | || |
c867b360 |
124 | ((uv & 0xFFFE) == 0xFFFE && /* Either FFFE or FFFF. */ |
b851fbc1 |
125 | !(flags & UNICODE_ALLOW_FFFF))) && |
126 | /* UNICODE_ALLOW_SUPER includes |
2a20b9da |
127 | * FFFEs and FFFFs beyond 0x10FFFF. */ |
b851fbc1 |
128 | ((uv <= PERL_UNICODE_MAX) || |
129 | !(flags & UNICODE_ALLOW_SUPER)) |
130 | ) |
9014280d |
131 | Perl_warner(aTHX_ packWARN(WARN_UTF8), |
6f6ac1de |
132 | "Unicode non-character 0x%04"UVxf" is illegal for interchange", uv); |
507b9800 |
133 | } |
c4d5f83a |
134 | if (UNI_IS_INVARIANT(uv)) { |
eb160463 |
135 | *d++ = (U8)UTF_TO_NATIVE(uv); |
a0ed51b3 |
136 | return d; |
137 | } |
2d331972 |
138 | #if defined(EBCDIC) |
1d72bdf6 |
139 | else { |
140 | STRLEN len = UNISKIP(uv); |
141 | U8 *p = d+len-1; |
142 | while (p > d) { |
eb160463 |
143 | *p-- = (U8)UTF_TO_NATIVE((uv & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK); |
1d72bdf6 |
144 | uv >>= UTF_ACCUMULATION_SHIFT; |
145 | } |
eb160463 |
146 | *p = (U8)UTF_TO_NATIVE((uv & UTF_START_MASK(len)) | UTF_START_MARK(len)); |
1d72bdf6 |
147 | return d+len; |
148 | } |
149 | #else /* Non loop style */ |
a0ed51b3 |
150 | if (uv < 0x800) { |
eb160463 |
151 | *d++ = (U8)(( uv >> 6) | 0xc0); |
152 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
153 | return d; |
154 | } |
155 | if (uv < 0x10000) { |
eb160463 |
156 | *d++ = (U8)(( uv >> 12) | 0xe0); |
157 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
158 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
159 | return d; |
160 | } |
161 | if (uv < 0x200000) { |
eb160463 |
162 | *d++ = (U8)(( uv >> 18) | 0xf0); |
163 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
164 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
165 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
166 | return d; |
167 | } |
168 | if (uv < 0x4000000) { |
eb160463 |
169 | *d++ = (U8)(( uv >> 24) | 0xf8); |
170 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
171 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
172 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
173 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
174 | return d; |
175 | } |
176 | if (uv < 0x80000000) { |
eb160463 |
177 | *d++ = (U8)(( uv >> 30) | 0xfc); |
178 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); |
179 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
180 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
181 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
182 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
183 | return d; |
184 | } |
6b8eaf93 |
185 | #ifdef HAS_QUAD |
d7578b48 |
186 | if (uv < UTF8_QUAD_MAX) |
a0ed51b3 |
187 | #endif |
188 | { |
eb160463 |
189 | *d++ = 0xfe; /* Can't match U+FEFF! */ |
190 | *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); |
191 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); |
192 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
193 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
194 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
195 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
196 | return d; |
197 | } |
6b8eaf93 |
198 | #ifdef HAS_QUAD |
a0ed51b3 |
199 | { |
eb160463 |
200 | *d++ = 0xff; /* Can't match U+FFFE! */ |
201 | *d++ = 0x80; /* 6 Reserved bits */ |
202 | *d++ = (U8)(((uv >> 60) & 0x0f) | 0x80); /* 2 Reserved bits */ |
203 | *d++ = (U8)(((uv >> 54) & 0x3f) | 0x80); |
204 | *d++ = (U8)(((uv >> 48) & 0x3f) | 0x80); |
205 | *d++ = (U8)(((uv >> 42) & 0x3f) | 0x80); |
206 | *d++ = (U8)(((uv >> 36) & 0x3f) | 0x80); |
207 | *d++ = (U8)(((uv >> 30) & 0x3f) | 0x80); |
208 | *d++ = (U8)(((uv >> 24) & 0x3f) | 0x80); |
209 | *d++ = (U8)(((uv >> 18) & 0x3f) | 0x80); |
210 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
211 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
212 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
213 | return d; |
214 | } |
215 | #endif |
1d72bdf6 |
216 | #endif /* Loop style */ |
a0ed51b3 |
217 | } |
9041c2e3 |
218 | |
646ca15d |
219 | /* |
220 | |
221 | Tests if some arbitrary number of bytes begins in a valid UTF-8 |
222 | character. Note that an INVARIANT (i.e. ASCII) character is a valid |
223 | UTF-8 character. The actual number of bytes in the UTF-8 character |
224 | will be returned if it is valid, otherwise 0. |
225 | |
226 | This is the "slow" version as opposed to the "fast" version which is |
227 | the "unrolled" IS_UTF8_CHAR(). E.g. for t/uni/class.t the speed |
228 | difference is a factor of 2 to 3. For lengths (UTF8SKIP(s)) of four |
229 | or less you should use the IS_UTF8_CHAR(), for lengths of five or more |
230 | you should use the _slow(). In practice this means that the _slow() |
231 | will be used very rarely, since the maximum Unicode code point (as of |
232 | Unicode 4.1) is U+10FFFF, which encodes in UTF-8 to four bytes. Only |
233 | the "Perl extended UTF-8" (the infamous 'v-strings') will encode into |
234 | five bytes or more. |
235 | |
236 | =cut */ |
c053b435 |
237 | STATIC STRLEN |
5f66b61c |
238 | S_is_utf8_char_slow(const U8 *s, const STRLEN len) |
646ca15d |
239 | { |
240 | U8 u = *s; |
241 | STRLEN slen; |
242 | UV uv, ouv; |
243 | |
7918f24d |
244 | PERL_ARGS_ASSERT_IS_UTF8_CHAR_SLOW; |
245 | |
646ca15d |
246 | if (UTF8_IS_INVARIANT(u)) |
247 | return 1; |
248 | |
249 | if (!UTF8_IS_START(u)) |
250 | return 0; |
251 | |
252 | if (len < 2 || !UTF8_IS_CONTINUATION(s[1])) |
253 | return 0; |
254 | |
255 | slen = len - 1; |
256 | s++; |
77263263 |
257 | #ifdef EBCDIC |
258 | u = NATIVE_TO_UTF(u); |
259 | #endif |
646ca15d |
260 | u &= UTF_START_MASK(len); |
261 | uv = u; |
262 | ouv = uv; |
263 | while (slen--) { |
264 | if (!UTF8_IS_CONTINUATION(*s)) |
265 | return 0; |
266 | uv = UTF8_ACCUMULATE(uv, *s); |
48ef279e |
267 | if (uv < ouv) |
646ca15d |
268 | return 0; |
269 | ouv = uv; |
270 | s++; |
271 | } |
272 | |
273 | if ((STRLEN)UNISKIP(uv) < len) |
274 | return 0; |
275 | |
276 | return len; |
277 | } |
9041c2e3 |
278 | |
279 | /* |
87cea99e |
280 | =for apidoc is_utf8_char |
eebe1485 |
281 | |
5da9da9e |
282 | Tests if some arbitrary number of bytes begins in a valid UTF-8 |
2bbc8d55 |
283 | character. Note that an INVARIANT (i.e. ASCII on non-EBCDIC machines) |
284 | character is a valid UTF-8 character. The actual number of bytes in the UTF-8 |
285 | character will be returned if it is valid, otherwise 0. |
9041c2e3 |
286 | |
82686b01 |
287 | =cut */ |
067a85ef |
288 | STRLEN |
668b6d8d |
289 | Perl_is_utf8_char(const U8 *s) |
386d01d6 |
290 | { |
44f8325f |
291 | const STRLEN len = UTF8SKIP(s); |
7918f24d |
292 | |
293 | PERL_ARGS_ASSERT_IS_UTF8_CHAR; |
3b0fc154 |
294 | #ifdef IS_UTF8_CHAR |
768c67ee |
295 | if (IS_UTF8_CHAR_FAST(len)) |
3b0fc154 |
296 | return IS_UTF8_CHAR(s, len) ? len : 0; |
297 | #endif /* #ifdef IS_UTF8_CHAR */ |
2c0c5f92 |
298 | return is_utf8_char_slow(s, len); |
386d01d6 |
299 | } |
300 | |
eaf7a4d2 |
301 | |
6662521e |
302 | /* |
87cea99e |
303 | =for apidoc is_utf8_string |
6662521e |
304 | |
c9ada85f |
305 | Returns true if first C<len> bytes of the given string form a valid |
1e54db1a |
306 | UTF-8 string, false otherwise. Note that 'a valid UTF-8 string' does |
307 | not mean 'a string that contains code points above 0x7F encoded in UTF-8' |
308 | because a valid ASCII string is a valid UTF-8 string. |
6662521e |
309 | |
eaf7a4d2 |
310 | See also is_ascii_string(), is_utf8_string_loclen(), and is_utf8_string_loc(). |
768c67ee |
311 | |
6662521e |
312 | =cut |
313 | */ |
314 | |
8e84507e |
315 | bool |
668b6d8d |
316 | Perl_is_utf8_string(const U8 *s, STRLEN len) |
6662521e |
317 | { |
35da51f7 |
318 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
7fc63493 |
319 | const U8* x = s; |
067a85ef |
320 | |
7918f24d |
321 | PERL_ARGS_ASSERT_IS_UTF8_STRING; |
1aa99e6b |
322 | |
6662521e |
323 | while (x < send) { |
a3b680e6 |
324 | STRLEN c; |
1acdb0da |
325 | /* Inline the easy bits of is_utf8_char() here for speed... */ |
326 | if (UTF8_IS_INVARIANT(*x)) |
327 | c = 1; |
328 | else if (!UTF8_IS_START(*x)) |
768c67ee |
329 | goto out; |
1acdb0da |
330 | else { |
331 | /* ... and call is_utf8_char() only if really needed. */ |
646ca15d |
332 | #ifdef IS_UTF8_CHAR |
333 | c = UTF8SKIP(x); |
768c67ee |
334 | if (IS_UTF8_CHAR_FAST(c)) { |
335 | if (!IS_UTF8_CHAR(x, c)) |
3c614e38 |
336 | c = 0; |
337 | } |
338 | else |
339 | c = is_utf8_char_slow(x, c); |
646ca15d |
340 | #else |
341 | c = is_utf8_char(x); |
342 | #endif /* #ifdef IS_UTF8_CHAR */ |
1acdb0da |
343 | if (!c) |
768c67ee |
344 | goto out; |
1acdb0da |
345 | } |
6662521e |
346 | x += c; |
6662521e |
347 | } |
768c67ee |
348 | |
349 | out: |
60006e79 |
350 | if (x != send) |
351 | return FALSE; |
067a85ef |
352 | |
353 | return TRUE; |
6662521e |
354 | } |
355 | |
67e989fb |
356 | /* |
814fafa7 |
357 | Implemented as a macro in utf8.h |
358 | |
87cea99e |
359 | =for apidoc is_utf8_string_loc |
814fafa7 |
360 | |
361 | Like is_utf8_string() but stores the location of the failure (in the |
362 | case of "utf8ness failure") or the location s+len (in the case of |
363 | "utf8ness success") in the C<ep>. |
364 | |
365 | See also is_utf8_string_loclen() and is_utf8_string(). |
366 | |
87cea99e |
367 | =for apidoc is_utf8_string_loclen |
81cd54e3 |
368 | |
e3e4599f |
369 | Like is_utf8_string() but stores the location of the failure (in the |
768c67ee |
370 | case of "utf8ness failure") or the location s+len (in the case of |
371 | "utf8ness success") in the C<ep>, and the number of UTF-8 |
372 | encoded characters in the C<el>. |
373 | |
374 | See also is_utf8_string_loc() and is_utf8_string(). |
81cd54e3 |
375 | |
376 | =cut |
377 | */ |
378 | |
379 | bool |
668b6d8d |
380 | Perl_is_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el) |
81cd54e3 |
381 | { |
35da51f7 |
382 | const U8* const send = s + (len ? len : strlen((const char *)s)); |
7fc63493 |
383 | const U8* x = s; |
81cd54e3 |
384 | STRLEN c; |
3ebfea28 |
385 | STRLEN outlen = 0; |
7918f24d |
386 | |
387 | PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN; |
81cd54e3 |
388 | |
81cd54e3 |
389 | while (x < send) { |
390 | /* Inline the easy bits of is_utf8_char() here for speed... */ |
391 | if (UTF8_IS_INVARIANT(*x)) |
768c67ee |
392 | c = 1; |
393 | else if (!UTF8_IS_START(*x)) |
394 | goto out; |
81cd54e3 |
395 | else { |
768c67ee |
396 | /* ... and call is_utf8_char() only if really needed. */ |
397 | #ifdef IS_UTF8_CHAR |
398 | c = UTF8SKIP(x); |
399 | if (IS_UTF8_CHAR_FAST(c)) { |
400 | if (!IS_UTF8_CHAR(x, c)) |
401 | c = 0; |
402 | } else |
403 | c = is_utf8_char_slow(x, c); |
404 | #else |
405 | c = is_utf8_char(x); |
406 | #endif /* #ifdef IS_UTF8_CHAR */ |
407 | if (!c) |
408 | goto out; |
81cd54e3 |
409 | } |
768c67ee |
410 | x += c; |
3ebfea28 |
411 | outlen++; |
81cd54e3 |
412 | } |
768c67ee |
413 | |
414 | out: |
3ebfea28 |
415 | if (el) |
416 | *el = outlen; |
417 | |
768c67ee |
418 | if (ep) |
419 | *ep = x; |
3ebfea28 |
420 | return (x == send); |
81cd54e3 |
421 | } |
422 | |
423 | /* |
768c67ee |
424 | |
87cea99e |
425 | =for apidoc utf8n_to_uvuni |
67e989fb |
426 | |
9041c2e3 |
427 | Bottom level UTF-8 decode routine. |
38a44b82 |
428 | Returns the Unicode code point value of the first character in the string C<s> |
1e54db1a |
429 | which is assumed to be in UTF-8 encoding and no longer than C<curlen>; |
7df053ec |
430 | C<retlen> will be set to the length, in bytes, of that character. |
67e989fb |
431 | |
1e54db1a |
432 | If C<s> does not point to a well-formed UTF-8 character, the behaviour |
dcad2880 |
433 | is dependent on the value of C<flags>: if it contains UTF8_CHECK_ONLY, |
434 | it is assumed that the caller will raise a warning, and this function |
28d3d195 |
435 | will silently just set C<retlen> to C<-1> and return zero. If the |
436 | C<flags> does not contain UTF8_CHECK_ONLY, warnings about |
437 | malformations will be given, C<retlen> will be set to the expected |
438 | length of the UTF-8 character in bytes, and zero will be returned. |
439 | |
440 | The C<flags> can also contain various flags to allow deviations from |
441 | the strict UTF-8 encoding (see F<utf8.h>). |
67e989fb |
442 | |
9041c2e3 |
443 | Most code should use utf8_to_uvchr() rather than call this directly. |
444 | |
37607a96 |
445 | =cut |
446 | */ |
67e989fb |
447 | |
a0ed51b3 |
448 | UV |
7fc63493 |
449 | Perl_utf8n_to_uvuni(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, U32 flags) |
a0ed51b3 |
450 | { |
97aff369 |
451 | dVAR; |
d4c19fe8 |
452 | const U8 * const s0 = s; |
9c5ffd7c |
453 | UV uv = *s, ouv = 0; |
ba210ebe |
454 | STRLEN len = 1; |
7fc63493 |
455 | const bool dowarn = ckWARN_d(WARN_UTF8); |
456 | const UV startbyte = *s; |
ba210ebe |
457 | STRLEN expectlen = 0; |
a0dbb045 |
458 | U32 warning = 0; |
5b311467 |
459 | SV* sv; |
a0dbb045 |
460 | |
7918f24d |
461 | PERL_ARGS_ASSERT_UTF8N_TO_UVUNI; |
462 | |
5b311467 |
463 | /* This list is a superset of the UTF8_ALLOW_XXX. BUT it isn't, eg SUPER missing XXX */ |
a0dbb045 |
464 | |
465 | #define UTF8_WARN_EMPTY 1 |
466 | #define UTF8_WARN_CONTINUATION 2 |
467 | #define UTF8_WARN_NON_CONTINUATION 3 |
468 | #define UTF8_WARN_FE_FF 4 |
469 | #define UTF8_WARN_SHORT 5 |
470 | #define UTF8_WARN_OVERFLOW 6 |
471 | #define UTF8_WARN_SURROGATE 7 |
c867b360 |
472 | #define UTF8_WARN_LONG 8 |
473 | #define UTF8_WARN_FFFF 9 /* Also FFFE. */ |
a0dbb045 |
474 | |
475 | if (curlen == 0 && |
476 | !(flags & UTF8_ALLOW_EMPTY)) { |
477 | warning = UTF8_WARN_EMPTY; |
0c443dc2 |
478 | goto malformed; |
479 | } |
480 | |
1d72bdf6 |
481 | if (UTF8_IS_INVARIANT(uv)) { |
a0ed51b3 |
482 | if (retlen) |
483 | *retlen = 1; |
c4d5f83a |
484 | return (UV) (NATIVE_TO_UTF(*s)); |
a0ed51b3 |
485 | } |
67e989fb |
486 | |
421a8bf2 |
487 | if (UTF8_IS_CONTINUATION(uv) && |
fcc8fcf6 |
488 | !(flags & UTF8_ALLOW_CONTINUATION)) { |
a0dbb045 |
489 | warning = UTF8_WARN_CONTINUATION; |
ba210ebe |
490 | goto malformed; |
491 | } |
492 | |
421a8bf2 |
493 | if (UTF8_IS_START(uv) && curlen > 1 && !UTF8_IS_CONTINUATION(s[1]) && |
fcc8fcf6 |
494 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { |
a0dbb045 |
495 | warning = UTF8_WARN_NON_CONTINUATION; |
ba210ebe |
496 | goto malformed; |
497 | } |
9041c2e3 |
498 | |
1d72bdf6 |
499 | #ifdef EBCDIC |
75383841 |
500 | uv = NATIVE_TO_UTF(uv); |
1d72bdf6 |
501 | #else |
fcc8fcf6 |
502 | if ((uv == 0xfe || uv == 0xff) && |
503 | !(flags & UTF8_ALLOW_FE_FF)) { |
a0dbb045 |
504 | warning = UTF8_WARN_FE_FF; |
ba210ebe |
505 | goto malformed; |
a0ed51b3 |
506 | } |
1d72bdf6 |
507 | #endif |
508 | |
ba210ebe |
509 | if (!(uv & 0x20)) { len = 2; uv &= 0x1f; } |
510 | else if (!(uv & 0x10)) { len = 3; uv &= 0x0f; } |
511 | else if (!(uv & 0x08)) { len = 4; uv &= 0x07; } |
512 | else if (!(uv & 0x04)) { len = 5; uv &= 0x03; } |
1d72bdf6 |
513 | #ifdef EBCDIC |
514 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } |
515 | else { len = 7; uv &= 0x01; } |
516 | #else |
ba210ebe |
517 | else if (!(uv & 0x02)) { len = 6; uv &= 0x01; } |
518 | else if (!(uv & 0x01)) { len = 7; uv = 0; } |
1d72bdf6 |
519 | else { len = 13; uv = 0; } /* whoa! */ |
520 | #endif |
521 | |
a0ed51b3 |
522 | if (retlen) |
523 | *retlen = len; |
9041c2e3 |
524 | |
ba210ebe |
525 | expectlen = len; |
526 | |
fcc8fcf6 |
527 | if ((curlen < expectlen) && |
528 | !(flags & UTF8_ALLOW_SHORT)) { |
a0dbb045 |
529 | warning = UTF8_WARN_SHORT; |
ba210ebe |
530 | goto malformed; |
531 | } |
532 | |
533 | len--; |
a0ed51b3 |
534 | s++; |
ba210ebe |
535 | ouv = uv; |
536 | |
a0ed51b3 |
537 | while (len--) { |
421a8bf2 |
538 | if (!UTF8_IS_CONTINUATION(*s) && |
539 | !(flags & UTF8_ALLOW_NON_CONTINUATION)) { |
a0dbb045 |
540 | s--; |
541 | warning = UTF8_WARN_NON_CONTINUATION; |
ba210ebe |
542 | goto malformed; |
a0ed51b3 |
543 | } |
544 | else |
8850bf83 |
545 | uv = UTF8_ACCUMULATE(uv, *s); |
a0dbb045 |
546 | if (!(uv > ouv)) { |
547 | /* These cannot be allowed. */ |
548 | if (uv == ouv) { |
75dbc644 |
549 | if (expectlen != 13 && !(flags & UTF8_ALLOW_LONG)) { |
a0dbb045 |
550 | warning = UTF8_WARN_LONG; |
551 | goto malformed; |
552 | } |
553 | } |
554 | else { /* uv < ouv */ |
555 | /* This cannot be allowed. */ |
556 | warning = UTF8_WARN_OVERFLOW; |
557 | goto malformed; |
558 | } |
ba210ebe |
559 | } |
560 | s++; |
561 | ouv = uv; |
562 | } |
563 | |
421a8bf2 |
564 | if (UNICODE_IS_SURROGATE(uv) && |
fcc8fcf6 |
565 | !(flags & UTF8_ALLOW_SURROGATE)) { |
a0dbb045 |
566 | warning = UTF8_WARN_SURROGATE; |
ba210ebe |
567 | goto malformed; |
eb160463 |
568 | } else if ((expectlen > (STRLEN)UNISKIP(uv)) && |
fcc8fcf6 |
569 | !(flags & UTF8_ALLOW_LONG)) { |
a0dbb045 |
570 | warning = UTF8_WARN_LONG; |
ba210ebe |
571 | goto malformed; |
421a8bf2 |
572 | } else if (UNICODE_IS_ILLEGAL(uv) && |
a9917092 |
573 | !(flags & UTF8_ALLOW_FFFF)) { |
a0dbb045 |
574 | warning = UTF8_WARN_FFFF; |
a9917092 |
575 | goto malformed; |
a0ed51b3 |
576 | } |
ba210ebe |
577 | |
a0ed51b3 |
578 | return uv; |
ba210ebe |
579 | |
580 | malformed: |
581 | |
fcc8fcf6 |
582 | if (flags & UTF8_CHECK_ONLY) { |
ba210ebe |
583 | if (retlen) |
10edeb5d |
584 | *retlen = ((STRLEN) -1); |
ba210ebe |
585 | return 0; |
586 | } |
587 | |
a0dbb045 |
588 | if (dowarn) { |
5b311467 |
589 | if (warning == UTF8_WARN_FFFF) { |
590 | sv = newSVpvs_flags("Unicode non-character ", SVs_TEMP); |
591 | Perl_sv_catpvf(aTHX_ sv, "0x%04"UVxf" is illegal for interchange", uv); |
592 | } |
593 | else { |
594 | sv = newSVpvs_flags("Malformed UTF-8 character ", SVs_TEMP); |
595 | |
596 | switch (warning) { |
597 | case 0: /* Intentionally empty. */ break; |
598 | case UTF8_WARN_EMPTY: |
599 | sv_catpvs(sv, "(empty string)"); |
600 | break; |
601 | case UTF8_WARN_CONTINUATION: |
602 | Perl_sv_catpvf(aTHX_ sv, "(unexpected continuation byte 0x%02"UVxf", with no preceding start byte)", uv); |
603 | break; |
604 | case UTF8_WARN_NON_CONTINUATION: |
605 | if (s == s0) |
606 | Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", immediately after start byte 0x%02"UVxf")", |
607 | (UV)s[1], startbyte); |
608 | else { |
609 | const int len = (int)(s-s0); |
610 | Perl_sv_catpvf(aTHX_ sv, "(unexpected non-continuation byte 0x%02"UVxf", %d byte%s after start byte 0x%02"UVxf", expected %d bytes)", |
611 | (UV)s[1], len, len > 1 ? "s" : "", startbyte, (int)expectlen); |
612 | } |
613 | |
614 | break; |
615 | case UTF8_WARN_FE_FF: |
616 | Perl_sv_catpvf(aTHX_ sv, "(byte 0x%02"UVxf")", uv); |
617 | break; |
618 | case UTF8_WARN_SHORT: |
619 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", |
620 | (int)curlen, curlen == 1 ? "" : "s", (int)expectlen, startbyte); |
621 | expectlen = curlen; /* distance for caller to skip */ |
622 | break; |
623 | case UTF8_WARN_OVERFLOW: |
624 | Perl_sv_catpvf(aTHX_ sv, "(overflow at 0x%"UVxf", byte 0x%02x, after start byte 0x%02"UVxf")", |
625 | ouv, *s, startbyte); |
626 | break; |
627 | case UTF8_WARN_SURROGATE: |
628 | Perl_sv_catpvf(aTHX_ sv, "(UTF-16 surrogate 0x%04"UVxf")", uv); |
629 | break; |
630 | case UTF8_WARN_LONG: |
631 | Perl_sv_catpvf(aTHX_ sv, "(%d byte%s, need %d, after start byte 0x%02"UVxf")", |
632 | (int)expectlen, expectlen == 1 ? "": "s", UNISKIP(uv), startbyte); |
633 | break; |
634 | default: |
635 | sv_catpvs(sv, "(unknown reason)"); |
636 | break; |
551405c4 |
637 | } |
a0dbb045 |
638 | } |
639 | |
640 | if (warning) { |
44f8325f |
641 | const char * const s = SvPVX_const(sv); |
a0dbb045 |
642 | |
643 | if (PL_op) |
9014280d |
644 | Perl_warner(aTHX_ packWARN(WARN_UTF8), |
53e06cf0 |
645 | "%s in %s", s, OP_DESC(PL_op)); |
a0dbb045 |
646 | else |
9014280d |
647 | Perl_warner(aTHX_ packWARN(WARN_UTF8), "%s", s); |
a0dbb045 |
648 | } |
649 | } |
650 | |
ba210ebe |
651 | if (retlen) |
28d3d195 |
652 | *retlen = expectlen ? expectlen : len; |
ba210ebe |
653 | |
28d3d195 |
654 | return 0; |
a0ed51b3 |
655 | } |
656 | |
8e84507e |
657 | /* |
87cea99e |
658 | =for apidoc utf8_to_uvchr |
9041c2e3 |
659 | |
660 | Returns the native character value of the first character in the string C<s> |
1e54db1a |
661 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
9041c2e3 |
662 | length, in bytes, of that character. |
663 | |
1e54db1a |
664 | If C<s> does not point to a well-formed UTF-8 character, zero is |
9041c2e3 |
665 | returned and retlen is set, if possible, to -1. |
666 | |
667 | =cut |
668 | */ |
669 | |
670 | UV |
7fc63493 |
671 | Perl_utf8_to_uvchr(pTHX_ const U8 *s, STRLEN *retlen) |
9041c2e3 |
672 | { |
7918f24d |
673 | PERL_ARGS_ASSERT_UTF8_TO_UVCHR; |
674 | |
1754c1a1 |
675 | return utf8n_to_uvchr(s, UTF8_MAXBYTES, retlen, |
676 | ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); |
9041c2e3 |
677 | } |
678 | |
679 | /* |
87cea99e |
680 | =for apidoc utf8_to_uvuni |
9041c2e3 |
681 | |
682 | Returns the Unicode code point of the first character in the string C<s> |
1e54db1a |
683 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
9041c2e3 |
684 | length, in bytes, of that character. |
685 | |
2bbc8d55 |
686 | This function should only be used when the returned UV is considered |
9041c2e3 |
687 | an index into the Unicode semantic tables (e.g. swashes). |
688 | |
1e54db1a |
689 | If C<s> does not point to a well-formed UTF-8 character, zero is |
ba210ebe |
690 | returned and retlen is set, if possible, to -1. |
8e84507e |
691 | |
692 | =cut |
693 | */ |
694 | |
695 | UV |
7fc63493 |
696 | Perl_utf8_to_uvuni(pTHX_ const U8 *s, STRLEN *retlen) |
8e84507e |
697 | { |
7918f24d |
698 | PERL_ARGS_ASSERT_UTF8_TO_UVUNI; |
699 | |
9041c2e3 |
700 | /* Call the low level routine asking for checks */ |
89ebb4a3 |
701 | return Perl_utf8n_to_uvuni(aTHX_ s, UTF8_MAXBYTES, retlen, |
872c91ae |
702 | ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); |
8e84507e |
703 | } |
704 | |
b76347f2 |
705 | /* |
87cea99e |
706 | =for apidoc utf8_length |
b76347f2 |
707 | |
708 | Return the length of the UTF-8 char encoded string C<s> in characters. |
02eb7b47 |
709 | Stops at C<e> (inclusive). If C<e E<lt> s> or if the scan would end |
710 | up past C<e>, croaks. |
b76347f2 |
711 | |
712 | =cut |
713 | */ |
714 | |
715 | STRLEN |
35a4481c |
716 | Perl_utf8_length(pTHX_ const U8 *s, const U8 *e) |
b76347f2 |
717 | { |
97aff369 |
718 | dVAR; |
b76347f2 |
719 | STRLEN len = 0; |
720 | |
7918f24d |
721 | PERL_ARGS_ASSERT_UTF8_LENGTH; |
722 | |
8850bf83 |
723 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g. |
724 | * the bitops (especially ~) can create illegal UTF-8. |
725 | * In other words: in Perl UTF-8 is not just for Unicode. */ |
726 | |
a3b680e6 |
727 | if (e < s) |
728 | goto warn_and_return; |
b76347f2 |
729 | while (s < e) { |
8e91ec7f |
730 | if (!UTF8_IS_INVARIANT(*s)) |
731 | s += UTF8SKIP(s); |
732 | else |
733 | s++; |
734 | len++; |
735 | } |
736 | |
737 | if (e != s) { |
738 | len--; |
739 | warn_and_return: |
9b387841 |
740 | if (PL_op) |
741 | Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), |
742 | "%s in %s", unees, OP_DESC(PL_op)); |
743 | else |
744 | Perl_ck_warner_d(aTHX_ packWARN(WARN_UTF8), unees); |
b76347f2 |
745 | } |
746 | |
747 | return len; |
748 | } |
749 | |
b06226ff |
750 | /* |
87cea99e |
751 | =for apidoc utf8_distance |
b06226ff |
752 | |
1e54db1a |
753 | Returns the number of UTF-8 characters between the UTF-8 pointers C<a> |
b06226ff |
754 | and C<b>. |
755 | |
756 | WARNING: use only if you *know* that the pointers point inside the |
757 | same UTF-8 buffer. |
758 | |
37607a96 |
759 | =cut |
760 | */ |
a0ed51b3 |
761 | |
02eb7b47 |
762 | IV |
35a4481c |
763 | Perl_utf8_distance(pTHX_ const U8 *a, const U8 *b) |
a0ed51b3 |
764 | { |
7918f24d |
765 | PERL_ARGS_ASSERT_UTF8_DISTANCE; |
766 | |
bf1665bc |
767 | return (a < b) ? -1 * (IV) utf8_length(a, b) : (IV) utf8_length(b, a); |
a0ed51b3 |
768 | } |
769 | |
b06226ff |
770 | /* |
87cea99e |
771 | =for apidoc utf8_hop |
b06226ff |
772 | |
8850bf83 |
773 | Return the UTF-8 pointer C<s> displaced by C<off> characters, either |
774 | forward or backward. |
b06226ff |
775 | |
776 | WARNING: do not use the following unless you *know* C<off> is within |
8850bf83 |
777 | the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned |
778 | on the first byte of character or just after the last byte of a character. |
b06226ff |
779 | |
37607a96 |
780 | =cut |
781 | */ |
a0ed51b3 |
782 | |
783 | U8 * |
4373e329 |
784 | Perl_utf8_hop(pTHX_ const U8 *s, I32 off) |
a0ed51b3 |
785 | { |
7918f24d |
786 | PERL_ARGS_ASSERT_UTF8_HOP; |
787 | |
96a5add6 |
788 | PERL_UNUSED_CONTEXT; |
8850bf83 |
789 | /* Note: cannot use UTF8_IS_...() too eagerly here since e.g |
790 | * the bitops (especially ~) can create illegal UTF-8. |
791 | * In other words: in Perl UTF-8 is not just for Unicode. */ |
792 | |
a0ed51b3 |
793 | if (off >= 0) { |
794 | while (off--) |
795 | s += UTF8SKIP(s); |
796 | } |
797 | else { |
798 | while (off++) { |
799 | s--; |
8850bf83 |
800 | while (UTF8_IS_CONTINUATION(*s)) |
801 | s--; |
a0ed51b3 |
802 | } |
803 | } |
4373e329 |
804 | return (U8 *)s; |
a0ed51b3 |
805 | } |
806 | |
6940069f |
807 | /* |
87cea99e |
808 | =for apidoc utf8_to_bytes |
6940069f |
809 | |
2bbc8d55 |
810 | Converts a string C<s> of length C<len> from UTF-8 into native byte encoding. |
246fae53 |
811 | Unlike C<bytes_to_utf8>, this over-writes the original string, and |
812 | updates len to contain the new length. |
67e989fb |
813 | Returns zero on failure, setting C<len> to -1. |
6940069f |
814 | |
95be277c |
815 | If you need a copy of the string, see C<bytes_from_utf8>. |
816 | |
6940069f |
817 | =cut |
818 | */ |
819 | |
820 | U8 * |
37607a96 |
821 | Perl_utf8_to_bytes(pTHX_ U8 *s, STRLEN *len) |
6940069f |
822 | { |
d4c19fe8 |
823 | U8 * const save = s; |
824 | U8 * const send = s + *len; |
6940069f |
825 | U8 *d; |
246fae53 |
826 | |
7918f24d |
827 | PERL_ARGS_ASSERT_UTF8_TO_BYTES; |
828 | |
1e54db1a |
829 | /* ensure valid UTF-8 and chars < 256 before updating string */ |
d4c19fe8 |
830 | while (s < send) { |
dcad2880 |
831 | U8 c = *s++; |
832 | |
1d72bdf6 |
833 | if (!UTF8_IS_INVARIANT(c) && |
834 | (!UTF8_IS_DOWNGRADEABLE_START(c) || (s >= send) |
835 | || !(c = *s++) || !UTF8_IS_CONTINUATION(c))) { |
10edeb5d |
836 | *len = ((STRLEN) -1); |
dcad2880 |
837 | return 0; |
838 | } |
246fae53 |
839 | } |
dcad2880 |
840 | |
841 | d = s = save; |
6940069f |
842 | while (s < send) { |
ed646e6e |
843 | STRLEN ulen; |
9041c2e3 |
844 | *d++ = (U8)utf8_to_uvchr(s, &ulen); |
ed646e6e |
845 | s += ulen; |
6940069f |
846 | } |
847 | *d = '\0'; |
246fae53 |
848 | *len = d - save; |
6940069f |
849 | return save; |
850 | } |
851 | |
852 | /* |
87cea99e |
853 | =for apidoc bytes_from_utf8 |
f9a63242 |
854 | |
2bbc8d55 |
855 | Converts a string C<s> of length C<len> from UTF-8 into native byte encoding. |
35a4481c |
856 | Unlike C<utf8_to_bytes> but like C<bytes_to_utf8>, returns a pointer to |
ef9edfd0 |
857 | the newly-created string, and updates C<len> to contain the new |
858 | length. Returns the original string if no conversion occurs, C<len> |
859 | is unchanged. Do nothing if C<is_utf8> points to 0. Sets C<is_utf8> to |
2bbc8d55 |
860 | 0 if C<s> is converted or consisted entirely of characters that are invariant |
861 | in utf8 (i.e., US-ASCII on non-EBCDIC machines). |
f9a63242 |
862 | |
37607a96 |
863 | =cut |
864 | */ |
f9a63242 |
865 | |
866 | U8 * |
e1ec3a88 |
867 | Perl_bytes_from_utf8(pTHX_ const U8 *s, STRLEN *len, bool *is_utf8) |
f9a63242 |
868 | { |
f9a63242 |
869 | U8 *d; |
e1ec3a88 |
870 | const U8 *start = s; |
871 | const U8 *send; |
f9a63242 |
872 | I32 count = 0; |
873 | |
7918f24d |
874 | PERL_ARGS_ASSERT_BYTES_FROM_UTF8; |
875 | |
96a5add6 |
876 | PERL_UNUSED_CONTEXT; |
f9a63242 |
877 | if (!*is_utf8) |
73d840c0 |
878 | return (U8 *)start; |
f9a63242 |
879 | |
1e54db1a |
880 | /* ensure valid UTF-8 and chars < 256 before converting string */ |
f9a63242 |
881 | for (send = s + *len; s < send;) { |
e1ec3a88 |
882 | U8 c = *s++; |
1d72bdf6 |
883 | if (!UTF8_IS_INVARIANT(c)) { |
db42d148 |
884 | if (UTF8_IS_DOWNGRADEABLE_START(c) && s < send && |
885 | (c = *s++) && UTF8_IS_CONTINUATION(c)) |
886 | count++; |
887 | else |
73d840c0 |
888 | return (U8 *)start; |
db42d148 |
889 | } |
f9a63242 |
890 | } |
891 | |
35da51f7 |
892 | *is_utf8 = FALSE; |
f9a63242 |
893 | |
212542aa |
894 | Newx(d, (*len) - count + 1, U8); |
ef9edfd0 |
895 | s = start; start = d; |
f9a63242 |
896 | while (s < send) { |
897 | U8 c = *s++; |
c4d5f83a |
898 | if (!UTF8_IS_INVARIANT(c)) { |
899 | /* Then it is two-byte encoded */ |
900 | c = UTF8_ACCUMULATE(NATIVE_TO_UTF(c), *s++); |
901 | c = ASCII_TO_NATIVE(c); |
902 | } |
903 | *d++ = c; |
f9a63242 |
904 | } |
905 | *d = '\0'; |
906 | *len = d - start; |
73d840c0 |
907 | return (U8 *)start; |
f9a63242 |
908 | } |
909 | |
910 | /* |
87cea99e |
911 | =for apidoc bytes_to_utf8 |
6940069f |
912 | |
2bbc8d55 |
913 | Converts a string C<s> of length C<len> from the native encoding into UTF-8. |
6662521e |
914 | Returns a pointer to the newly-created string, and sets C<len> to |
915 | reflect the new length. |
6940069f |
916 | |
2bbc8d55 |
917 | A NUL character will be written after the end of the string. |
918 | |
919 | If you want to convert to UTF-8 from encodings other than |
920 | the native (Latin1 or EBCDIC), |
c9ada85f |
921 | see sv_recode_to_utf8(). |
922 | |
497711e7 |
923 | =cut |
6940069f |
924 | */ |
925 | |
926 | U8* |
35a4481c |
927 | Perl_bytes_to_utf8(pTHX_ const U8 *s, STRLEN *len) |
6940069f |
928 | { |
35a4481c |
929 | const U8 * const send = s + (*len); |
6940069f |
930 | U8 *d; |
931 | U8 *dst; |
7918f24d |
932 | |
933 | PERL_ARGS_ASSERT_BYTES_TO_UTF8; |
96a5add6 |
934 | PERL_UNUSED_CONTEXT; |
6940069f |
935 | |
212542aa |
936 | Newx(d, (*len) * 2 + 1, U8); |
6940069f |
937 | dst = d; |
938 | |
939 | while (s < send) { |
35a4481c |
940 | const UV uv = NATIVE_TO_ASCII(*s++); |
c4d5f83a |
941 | if (UNI_IS_INVARIANT(uv)) |
eb160463 |
942 | *d++ = (U8)UTF_TO_NATIVE(uv); |
6940069f |
943 | else { |
eb160463 |
944 | *d++ = (U8)UTF8_EIGHT_BIT_HI(uv); |
945 | *d++ = (U8)UTF8_EIGHT_BIT_LO(uv); |
6940069f |
946 | } |
947 | } |
948 | *d = '\0'; |
6662521e |
949 | *len = d-dst; |
6940069f |
950 | return dst; |
951 | } |
952 | |
a0ed51b3 |
953 | /* |
dea0fc0b |
954 | * Convert native (big-endian) or reversed (little-endian) UTF-16 to UTF-8. |
a0ed51b3 |
955 | * |
956 | * Destination must be pre-extended to 3/2 source. Do not use in-place. |
957 | * We optimize for native, for obvious reasons. */ |
958 | |
959 | U8* |
dea0fc0b |
960 | Perl_utf16_to_utf8(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 |
961 | { |
dea0fc0b |
962 | U8* pend; |
963 | U8* dstart = d; |
964 | |
7918f24d |
965 | PERL_ARGS_ASSERT_UTF16_TO_UTF8; |
966 | |
dea0fc0b |
967 | if (bytelen & 1) |
f5992bc4 |
968 | Perl_croak(aTHX_ "panic: utf16_to_utf8: odd bytelen %"UVuf, (UV)bytelen); |
dea0fc0b |
969 | |
970 | pend = p + bytelen; |
971 | |
a0ed51b3 |
972 | while (p < pend) { |
dea0fc0b |
973 | UV uv = (p[0] << 8) + p[1]; /* UTF-16BE */ |
974 | p += 2; |
a0ed51b3 |
975 | if (uv < 0x80) { |
e294cc5d |
976 | #ifdef EBCDIC |
977 | *d++ = UNI_TO_NATIVE(uv); |
978 | #else |
eb160463 |
979 | *d++ = (U8)uv; |
e294cc5d |
980 | #endif |
a0ed51b3 |
981 | continue; |
982 | } |
983 | if (uv < 0x800) { |
eb160463 |
984 | *d++ = (U8)(( uv >> 6) | 0xc0); |
985 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
986 | continue; |
987 | } |
52b9aa85 |
988 | if (uv >= 0xd800 && uv <= 0xdbff) { /* surrogates */ |
01ea242b |
989 | if (p >= pend) { |
dea0fc0b |
990 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
01ea242b |
991 | } else { |
992 | UV low = (p[0] << 8) + p[1]; |
993 | p += 2; |
52b9aa85 |
994 | if (low < 0xdc00 || low > 0xdfff) |
01ea242b |
995 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
996 | uv = ((uv - 0xd800) << 10) + (low - 0xdc00) + 0x10000; |
997 | } |
dbde1951 |
998 | } else if (uv >= 0xdc00 && uv <= 0xdfff) { |
999 | Perl_croak(aTHX_ "Malformed UTF-16 surrogate"); |
a0ed51b3 |
1000 | } |
1001 | if (uv < 0x10000) { |
eb160463 |
1002 | *d++ = (U8)(( uv >> 12) | 0xe0); |
1003 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
1004 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
1005 | continue; |
1006 | } |
1007 | else { |
eb160463 |
1008 | *d++ = (U8)(( uv >> 18) | 0xf0); |
1009 | *d++ = (U8)(((uv >> 12) & 0x3f) | 0x80); |
1010 | *d++ = (U8)(((uv >> 6) & 0x3f) | 0x80); |
1011 | *d++ = (U8)(( uv & 0x3f) | 0x80); |
a0ed51b3 |
1012 | continue; |
1013 | } |
1014 | } |
dea0fc0b |
1015 | *newlen = d - dstart; |
a0ed51b3 |
1016 | return d; |
1017 | } |
1018 | |
1019 | /* Note: this one is slightly destructive of the source. */ |
1020 | |
1021 | U8* |
dea0fc0b |
1022 | Perl_utf16_to_utf8_reversed(pTHX_ U8* p, U8* d, I32 bytelen, I32 *newlen) |
a0ed51b3 |
1023 | { |
1024 | U8* s = (U8*)p; |
d4c19fe8 |
1025 | U8* const send = s + bytelen; |
7918f24d |
1026 | |
1027 | PERL_ARGS_ASSERT_UTF16_TO_UTF8_REVERSED; |
1028 | |
e0ea5e2d |
1029 | if (bytelen & 1) |
1030 | Perl_croak(aTHX_ "panic: utf16_to_utf8_reversed: odd bytelen %"UVuf, |
1031 | (UV)bytelen); |
1032 | |
a0ed51b3 |
1033 | while (s < send) { |
d4c19fe8 |
1034 | const U8 tmp = s[0]; |
a0ed51b3 |
1035 | s[0] = s[1]; |
1036 | s[1] = tmp; |
1037 | s += 2; |
1038 | } |
dea0fc0b |
1039 | return utf16_to_utf8(p, d, bytelen, newlen); |
a0ed51b3 |
1040 | } |
1041 | |
1042 | /* for now these are all defined (inefficiently) in terms of the utf8 versions */ |
1043 | |
1044 | bool |
84afefe6 |
1045 | Perl_is_uni_alnum(pTHX_ UV c) |
a0ed51b3 |
1046 | { |
89ebb4a3 |
1047 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1048 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1049 | return is_utf8_alnum(tmpbuf); |
1050 | } |
1051 | |
1052 | bool |
84afefe6 |
1053 | Perl_is_uni_idfirst(pTHX_ UV c) |
a0ed51b3 |
1054 | { |
89ebb4a3 |
1055 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1056 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1057 | return is_utf8_idfirst(tmpbuf); |
1058 | } |
1059 | |
1060 | bool |
84afefe6 |
1061 | Perl_is_uni_alpha(pTHX_ UV c) |
a0ed51b3 |
1062 | { |
89ebb4a3 |
1063 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1064 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1065 | return is_utf8_alpha(tmpbuf); |
1066 | } |
1067 | |
1068 | bool |
84afefe6 |
1069 | Perl_is_uni_ascii(pTHX_ UV c) |
4d61ec05 |
1070 | { |
89ebb4a3 |
1071 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1072 | uvchr_to_utf8(tmpbuf, c); |
4d61ec05 |
1073 | return is_utf8_ascii(tmpbuf); |
1074 | } |
1075 | |
1076 | bool |
84afefe6 |
1077 | Perl_is_uni_space(pTHX_ UV c) |
a0ed51b3 |
1078 | { |
89ebb4a3 |
1079 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1080 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1081 | return is_utf8_space(tmpbuf); |
1082 | } |
1083 | |
1084 | bool |
84afefe6 |
1085 | Perl_is_uni_digit(pTHX_ UV c) |
a0ed51b3 |
1086 | { |
89ebb4a3 |
1087 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1088 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1089 | return is_utf8_digit(tmpbuf); |
1090 | } |
1091 | |
1092 | bool |
84afefe6 |
1093 | Perl_is_uni_upper(pTHX_ UV c) |
a0ed51b3 |
1094 | { |
89ebb4a3 |
1095 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1096 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1097 | return is_utf8_upper(tmpbuf); |
1098 | } |
1099 | |
1100 | bool |
84afefe6 |
1101 | Perl_is_uni_lower(pTHX_ UV c) |
a0ed51b3 |
1102 | { |
89ebb4a3 |
1103 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1104 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1105 | return is_utf8_lower(tmpbuf); |
1106 | } |
1107 | |
1108 | bool |
84afefe6 |
1109 | Perl_is_uni_cntrl(pTHX_ UV c) |
b8c5462f |
1110 | { |
89ebb4a3 |
1111 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1112 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f |
1113 | return is_utf8_cntrl(tmpbuf); |
1114 | } |
1115 | |
1116 | bool |
84afefe6 |
1117 | Perl_is_uni_graph(pTHX_ UV c) |
b8c5462f |
1118 | { |
89ebb4a3 |
1119 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1120 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f |
1121 | return is_utf8_graph(tmpbuf); |
1122 | } |
1123 | |
1124 | bool |
84afefe6 |
1125 | Perl_is_uni_print(pTHX_ UV c) |
a0ed51b3 |
1126 | { |
89ebb4a3 |
1127 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1128 | uvchr_to_utf8(tmpbuf, c); |
a0ed51b3 |
1129 | return is_utf8_print(tmpbuf); |
1130 | } |
1131 | |
b8c5462f |
1132 | bool |
84afefe6 |
1133 | Perl_is_uni_punct(pTHX_ UV c) |
b8c5462f |
1134 | { |
89ebb4a3 |
1135 | U8 tmpbuf[UTF8_MAXBYTES+1]; |
230880c1 |
1136 | uvchr_to_utf8(tmpbuf, c); |
b8c5462f |
1137 | return is_utf8_punct(tmpbuf); |
1138 | } |
1139 | |
4d61ec05 |
1140 | bool |
84afefe6 |
1141 | Perl_is_uni_xdigit(pTHX_ UV c) |
4d61ec05 |
1142 | { |
89ebb4a3 |
1143 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
230880c1 |
1144 | uvchr_to_utf8(tmpbuf, c); |
4d61ec05 |
1145 | return is_utf8_xdigit(tmpbuf); |
1146 | } |
1147 | |
84afefe6 |
1148 | UV |
1149 | Perl_to_uni_upper(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
1150 | { |
7918f24d |
1151 | PERL_ARGS_ASSERT_TO_UNI_UPPER; |
1152 | |
0ebc6274 |
1153 | uvchr_to_utf8(p, c); |
1154 | return to_utf8_upper(p, p, lenp); |
a0ed51b3 |
1155 | } |
1156 | |
84afefe6 |
1157 | UV |
1158 | Perl_to_uni_title(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
1159 | { |
7918f24d |
1160 | PERL_ARGS_ASSERT_TO_UNI_TITLE; |
1161 | |
0ebc6274 |
1162 | uvchr_to_utf8(p, c); |
1163 | return to_utf8_title(p, p, lenp); |
a0ed51b3 |
1164 | } |
1165 | |
84afefe6 |
1166 | UV |
1167 | Perl_to_uni_lower(pTHX_ UV c, U8* p, STRLEN *lenp) |
a0ed51b3 |
1168 | { |
7918f24d |
1169 | PERL_ARGS_ASSERT_TO_UNI_LOWER; |
1170 | |
0ebc6274 |
1171 | uvchr_to_utf8(p, c); |
1172 | return to_utf8_lower(p, p, lenp); |
a0ed51b3 |
1173 | } |
1174 | |
84afefe6 |
1175 | UV |
1176 | Perl_to_uni_fold(pTHX_ UV c, U8* p, STRLEN *lenp) |
1177 | { |
7918f24d |
1178 | PERL_ARGS_ASSERT_TO_UNI_FOLD; |
1179 | |
0ebc6274 |
1180 | uvchr_to_utf8(p, c); |
1181 | return to_utf8_fold(p, p, lenp); |
84afefe6 |
1182 | } |
1183 | |
a0ed51b3 |
1184 | /* for now these all assume no locale info available for Unicode > 255 */ |
1185 | |
1186 | bool |
84afefe6 |
1187 | Perl_is_uni_alnum_lc(pTHX_ UV c) |
a0ed51b3 |
1188 | { |
1189 | return is_uni_alnum(c); /* XXX no locale support yet */ |
1190 | } |
1191 | |
1192 | bool |
84afefe6 |
1193 | Perl_is_uni_idfirst_lc(pTHX_ UV c) |
a0ed51b3 |
1194 | { |
1195 | return is_uni_idfirst(c); /* XXX no locale support yet */ |
1196 | } |
1197 | |
1198 | bool |
84afefe6 |
1199 | Perl_is_uni_alpha_lc(pTHX_ UV c) |
a0ed51b3 |
1200 | { |
1201 | return is_uni_alpha(c); /* XXX no locale support yet */ |
1202 | } |
1203 | |
1204 | bool |
84afefe6 |
1205 | Perl_is_uni_ascii_lc(pTHX_ UV c) |
4d61ec05 |
1206 | { |
1207 | return is_uni_ascii(c); /* XXX no locale support yet */ |
1208 | } |
1209 | |
1210 | bool |
84afefe6 |
1211 | Perl_is_uni_space_lc(pTHX_ UV c) |
a0ed51b3 |
1212 | { |
1213 | return is_uni_space(c); /* XXX no locale support yet */ |
1214 | } |
1215 | |
1216 | bool |
84afefe6 |
1217 | Perl_is_uni_digit_lc(pTHX_ UV c) |
a0ed51b3 |
1218 | { |
1219 | return is_uni_digit(c); /* XXX no locale support yet */ |
1220 | } |
1221 | |
1222 | bool |
84afefe6 |
1223 | Perl_is_uni_upper_lc(pTHX_ UV c) |
a0ed51b3 |
1224 | { |
1225 | return is_uni_upper(c); /* XXX no locale support yet */ |
1226 | } |
1227 | |
1228 | bool |
84afefe6 |
1229 | Perl_is_uni_lower_lc(pTHX_ UV c) |
a0ed51b3 |
1230 | { |
1231 | return is_uni_lower(c); /* XXX no locale support yet */ |
1232 | } |
1233 | |
1234 | bool |
84afefe6 |
1235 | Perl_is_uni_cntrl_lc(pTHX_ UV c) |
b8c5462f |
1236 | { |
1237 | return is_uni_cntrl(c); /* XXX no locale support yet */ |
1238 | } |
1239 | |
1240 | bool |
84afefe6 |
1241 | Perl_is_uni_graph_lc(pTHX_ UV c) |
b8c5462f |
1242 | { |
1243 | return is_uni_graph(c); /* XXX no locale support yet */ |
1244 | } |
1245 | |
1246 | bool |
84afefe6 |
1247 | Perl_is_uni_print_lc(pTHX_ UV c) |
a0ed51b3 |
1248 | { |
1249 | return is_uni_print(c); /* XXX no locale support yet */ |
1250 | } |
1251 | |
b8c5462f |
1252 | bool |
84afefe6 |
1253 | Perl_is_uni_punct_lc(pTHX_ UV c) |
b8c5462f |
1254 | { |
1255 | return is_uni_punct(c); /* XXX no locale support yet */ |
1256 | } |
1257 | |
4d61ec05 |
1258 | bool |
84afefe6 |
1259 | Perl_is_uni_xdigit_lc(pTHX_ UV c) |
4d61ec05 |
1260 | { |
1261 | return is_uni_xdigit(c); /* XXX no locale support yet */ |
1262 | } |
1263 | |
b7ac61fa |
1264 | U32 |
1265 | Perl_to_uni_upper_lc(pTHX_ U32 c) |
1266 | { |
ee099d14 |
1267 | /* XXX returns only the first character -- do not use XXX */ |
1268 | /* XXX no locale support yet */ |
1269 | STRLEN len; |
89ebb4a3 |
1270 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 |
1271 | return (U32)to_uni_upper(c, tmpbuf, &len); |
b7ac61fa |
1272 | } |
1273 | |
1274 | U32 |
1275 | Perl_to_uni_title_lc(pTHX_ U32 c) |
1276 | { |
ee099d14 |
1277 | /* XXX returns only the first character XXX -- do not use XXX */ |
1278 | /* XXX no locale support yet */ |
1279 | STRLEN len; |
89ebb4a3 |
1280 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 |
1281 | return (U32)to_uni_title(c, tmpbuf, &len); |
b7ac61fa |
1282 | } |
1283 | |
1284 | U32 |
1285 | Perl_to_uni_lower_lc(pTHX_ U32 c) |
1286 | { |
ee099d14 |
1287 | /* XXX returns only the first character -- do not use XXX */ |
1288 | /* XXX no locale support yet */ |
1289 | STRLEN len; |
89ebb4a3 |
1290 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
ee099d14 |
1291 | return (U32)to_uni_lower(c, tmpbuf, &len); |
b7ac61fa |
1292 | } |
1293 | |
7452cf6a |
1294 | static bool |
5141f98e |
1295 | S_is_utf8_common(pTHX_ const U8 *const p, SV **swash, |
bde6a22d |
1296 | const char *const swashname) |
1297 | { |
97aff369 |
1298 | dVAR; |
7918f24d |
1299 | |
1300 | PERL_ARGS_ASSERT_IS_UTF8_COMMON; |
1301 | |
bde6a22d |
1302 | if (!is_utf8_char(p)) |
1303 | return FALSE; |
1304 | if (!*swash) |
711a919c |
1305 | *swash = swash_init("utf8", swashname, &PL_sv_undef, 1, 0); |
bde6a22d |
1306 | return swash_fetch(*swash, p, TRUE) != 0; |
1307 | } |
1308 | |
1309 | bool |
7fc63493 |
1310 | Perl_is_utf8_alnum(pTHX_ const U8 *p) |
a0ed51b3 |
1311 | { |
97aff369 |
1312 | dVAR; |
7918f24d |
1313 | |
1314 | PERL_ARGS_ASSERT_IS_UTF8_ALNUM; |
1315 | |
671c33bf |
1316 | /* NOTE: "IsWord", not "IsAlnum", since Alnum is a true |
1317 | * descendant of isalnum(3), in other words, it doesn't |
1318 | * contain the '_'. --jhi */ |
d4c19fe8 |
1319 | return is_utf8_common(p, &PL_utf8_alnum, "IsWord"); |
a0ed51b3 |
1320 | } |
1321 | |
1322 | bool |
7fc63493 |
1323 | Perl_is_utf8_idfirst(pTHX_ const U8 *p) /* The naming is historical. */ |
a0ed51b3 |
1324 | { |
97aff369 |
1325 | dVAR; |
7918f24d |
1326 | |
1327 | PERL_ARGS_ASSERT_IS_UTF8_IDFIRST; |
1328 | |
82686b01 |
1329 | if (*p == '_') |
1330 | return TRUE; |
bde6a22d |
1331 | /* is_utf8_idstart would be more logical. */ |
d4c19fe8 |
1332 | return is_utf8_common(p, &PL_utf8_idstart, "IdStart"); |
82686b01 |
1333 | } |
1334 | |
1335 | bool |
7fc63493 |
1336 | Perl_is_utf8_idcont(pTHX_ const U8 *p) |
82686b01 |
1337 | { |
97aff369 |
1338 | dVAR; |
7918f24d |
1339 | |
1340 | PERL_ARGS_ASSERT_IS_UTF8_IDCONT; |
1341 | |
82686b01 |
1342 | if (*p == '_') |
1343 | return TRUE; |
d4c19fe8 |
1344 | return is_utf8_common(p, &PL_utf8_idcont, "IdContinue"); |
a0ed51b3 |
1345 | } |
1346 | |
1347 | bool |
7fc63493 |
1348 | Perl_is_utf8_alpha(pTHX_ const U8 *p) |
a0ed51b3 |
1349 | { |
97aff369 |
1350 | dVAR; |
7918f24d |
1351 | |
1352 | PERL_ARGS_ASSERT_IS_UTF8_ALPHA; |
1353 | |
d4c19fe8 |
1354 | return is_utf8_common(p, &PL_utf8_alpha, "IsAlpha"); |
a0ed51b3 |
1355 | } |
1356 | |
1357 | bool |
7fc63493 |
1358 | Perl_is_utf8_ascii(pTHX_ const U8 *p) |
b8c5462f |
1359 | { |
97aff369 |
1360 | dVAR; |
7918f24d |
1361 | |
1362 | PERL_ARGS_ASSERT_IS_UTF8_ASCII; |
1363 | |
d4c19fe8 |
1364 | return is_utf8_common(p, &PL_utf8_ascii, "IsAscii"); |
b8c5462f |
1365 | } |
1366 | |
1367 | bool |
7fc63493 |
1368 | Perl_is_utf8_space(pTHX_ const U8 *p) |
a0ed51b3 |
1369 | { |
97aff369 |
1370 | dVAR; |
7918f24d |
1371 | |
1372 | PERL_ARGS_ASSERT_IS_UTF8_SPACE; |
1373 | |
d4c19fe8 |
1374 | return is_utf8_common(p, &PL_utf8_space, "IsSpacePerl"); |
a0ed51b3 |
1375 | } |
1376 | |
1377 | bool |
d1eb3177 |
1378 | Perl_is_utf8_perl_space(pTHX_ const U8 *p) |
1379 | { |
1380 | dVAR; |
1381 | |
1382 | PERL_ARGS_ASSERT_IS_UTF8_PERL_SPACE; |
1383 | |
1384 | return is_utf8_common(p, &PL_utf8_perl_space, "IsPerlSpace"); |
1385 | } |
1386 | |
1387 | bool |
1388 | Perl_is_utf8_perl_word(pTHX_ const U8 *p) |
1389 | { |
1390 | dVAR; |
1391 | |
1392 | PERL_ARGS_ASSERT_IS_UTF8_PERL_WORD; |
1393 | |
1394 | return is_utf8_common(p, &PL_utf8_perl_word, "IsPerlWord"); |
1395 | } |
1396 | |
1397 | bool |
7fc63493 |
1398 | Perl_is_utf8_digit(pTHX_ const U8 *p) |
a0ed51b3 |
1399 | { |
97aff369 |
1400 | dVAR; |
7918f24d |
1401 | |
1402 | PERL_ARGS_ASSERT_IS_UTF8_DIGIT; |
1403 | |
d4c19fe8 |
1404 | return is_utf8_common(p, &PL_utf8_digit, "IsDigit"); |
a0ed51b3 |
1405 | } |
1406 | |
1407 | bool |
d1eb3177 |
1408 | Perl_is_utf8_posix_digit(pTHX_ const U8 *p) |
1409 | { |
1410 | dVAR; |
1411 | |
1412 | PERL_ARGS_ASSERT_IS_UTF8_POSIX_DIGIT; |
1413 | |
1414 | return is_utf8_common(p, &PL_utf8_posix_digit, "IsPosixDigit"); |
1415 | } |
1416 | |
1417 | bool |
7fc63493 |
1418 | Perl_is_utf8_upper(pTHX_ const U8 *p) |
a0ed51b3 |
1419 | { |
97aff369 |
1420 | dVAR; |
7918f24d |
1421 | |
1422 | PERL_ARGS_ASSERT_IS_UTF8_UPPER; |
1423 | |
d4c19fe8 |
1424 | return is_utf8_common(p, &PL_utf8_upper, "IsUppercase"); |
a0ed51b3 |
1425 | } |
1426 | |
1427 | bool |
7fc63493 |
1428 | Perl_is_utf8_lower(pTHX_ const U8 *p) |
a0ed51b3 |
1429 | { |
97aff369 |
1430 | dVAR; |
7918f24d |
1431 | |
1432 | PERL_ARGS_ASSERT_IS_UTF8_LOWER; |
1433 | |
d4c19fe8 |
1434 | return is_utf8_common(p, &PL_utf8_lower, "IsLowercase"); |
a0ed51b3 |
1435 | } |
1436 | |
1437 | bool |
7fc63493 |
1438 | Perl_is_utf8_cntrl(pTHX_ const U8 *p) |
b8c5462f |
1439 | { |
97aff369 |
1440 | dVAR; |
7918f24d |
1441 | |
1442 | PERL_ARGS_ASSERT_IS_UTF8_CNTRL; |
1443 | |
d4c19fe8 |
1444 | return is_utf8_common(p, &PL_utf8_cntrl, "IsCntrl"); |
b8c5462f |
1445 | } |
1446 | |
1447 | bool |
7fc63493 |
1448 | Perl_is_utf8_graph(pTHX_ const U8 *p) |
b8c5462f |
1449 | { |
97aff369 |
1450 | dVAR; |
7918f24d |
1451 | |
1452 | PERL_ARGS_ASSERT_IS_UTF8_GRAPH; |
1453 | |
d4c19fe8 |
1454 | return is_utf8_common(p, &PL_utf8_graph, "IsGraph"); |
b8c5462f |
1455 | } |
1456 | |
1457 | bool |
7fc63493 |
1458 | Perl_is_utf8_print(pTHX_ const U8 *p) |
a0ed51b3 |
1459 | { |
97aff369 |
1460 | dVAR; |
7918f24d |
1461 | |
1462 | PERL_ARGS_ASSERT_IS_UTF8_PRINT; |
1463 | |
d4c19fe8 |
1464 | return is_utf8_common(p, &PL_utf8_print, "IsPrint"); |
a0ed51b3 |
1465 | } |
1466 | |
1467 | bool |
7fc63493 |
1468 | Perl_is_utf8_punct(pTHX_ const U8 *p) |
b8c5462f |
1469 | { |
97aff369 |
1470 | dVAR; |
7918f24d |
1471 | |
1472 | PERL_ARGS_ASSERT_IS_UTF8_PUNCT; |
1473 | |
d4c19fe8 |
1474 | return is_utf8_common(p, &PL_utf8_punct, "IsPunct"); |
b8c5462f |
1475 | } |
1476 | |
1477 | bool |
7fc63493 |
1478 | Perl_is_utf8_xdigit(pTHX_ const U8 *p) |
b8c5462f |
1479 | { |
97aff369 |
1480 | dVAR; |
7918f24d |
1481 | |
1482 | PERL_ARGS_ASSERT_IS_UTF8_XDIGIT; |
1483 | |
d1eb3177 |
1484 | return is_utf8_common(p, &PL_utf8_xdigit, "IsXDigit"); |
b8c5462f |
1485 | } |
1486 | |
1487 | bool |
7fc63493 |
1488 | Perl_is_utf8_mark(pTHX_ const U8 *p) |
a0ed51b3 |
1489 | { |
97aff369 |
1490 | dVAR; |
7918f24d |
1491 | |
1492 | PERL_ARGS_ASSERT_IS_UTF8_MARK; |
1493 | |
d4c19fe8 |
1494 | return is_utf8_common(p, &PL_utf8_mark, "IsM"); |
a0ed51b3 |
1495 | } |
1496 | |
37e2e78e |
1497 | bool |
1498 | Perl_is_utf8_X_begin(pTHX_ const U8 *p) |
1499 | { |
1500 | dVAR; |
1501 | |
1502 | PERL_ARGS_ASSERT_IS_UTF8_X_BEGIN; |
1503 | |
1504 | return is_utf8_common(p, &PL_utf8_X_begin, "_X_Begin"); |
1505 | } |
1506 | |
1507 | bool |
1508 | Perl_is_utf8_X_extend(pTHX_ const U8 *p) |
1509 | { |
1510 | dVAR; |
1511 | |
1512 | PERL_ARGS_ASSERT_IS_UTF8_X_EXTEND; |
1513 | |
1514 | return is_utf8_common(p, &PL_utf8_X_extend, "_X_Extend"); |
1515 | } |
1516 | |
1517 | bool |
1518 | Perl_is_utf8_X_prepend(pTHX_ const U8 *p) |
1519 | { |
1520 | dVAR; |
1521 | |
1522 | PERL_ARGS_ASSERT_IS_UTF8_X_PREPEND; |
1523 | |
1524 | return is_utf8_common(p, &PL_utf8_X_prepend, "GCB=Prepend"); |
1525 | } |
1526 | |
1527 | bool |
1528 | Perl_is_utf8_X_non_hangul(pTHX_ const U8 *p) |
1529 | { |
1530 | dVAR; |
1531 | |
1532 | PERL_ARGS_ASSERT_IS_UTF8_X_NON_HANGUL; |
1533 | |
1534 | return is_utf8_common(p, &PL_utf8_X_non_hangul, "HST=Not_Applicable"); |
1535 | } |
1536 | |
1537 | bool |
1538 | Perl_is_utf8_X_L(pTHX_ const U8 *p) |
1539 | { |
1540 | dVAR; |
1541 | |
1542 | PERL_ARGS_ASSERT_IS_UTF8_X_L; |
1543 | |
1544 | return is_utf8_common(p, &PL_utf8_X_L, "GCB=L"); |
1545 | } |
1546 | |
1547 | bool |
1548 | Perl_is_utf8_X_LV(pTHX_ const U8 *p) |
1549 | { |
1550 | dVAR; |
1551 | |
1552 | PERL_ARGS_ASSERT_IS_UTF8_X_LV; |
1553 | |
1554 | return is_utf8_common(p, &PL_utf8_X_LV, "GCB=LV"); |
1555 | } |
1556 | |
1557 | bool |
1558 | Perl_is_utf8_X_LVT(pTHX_ const U8 *p) |
1559 | { |
1560 | dVAR; |
1561 | |
1562 | PERL_ARGS_ASSERT_IS_UTF8_X_LVT; |
1563 | |
1564 | return is_utf8_common(p, &PL_utf8_X_LVT, "GCB=LVT"); |
1565 | } |
1566 | |
1567 | bool |
1568 | Perl_is_utf8_X_T(pTHX_ const U8 *p) |
1569 | { |
1570 | dVAR; |
1571 | |
1572 | PERL_ARGS_ASSERT_IS_UTF8_X_T; |
1573 | |
1574 | return is_utf8_common(p, &PL_utf8_X_T, "GCB=T"); |
1575 | } |
1576 | |
1577 | bool |
1578 | Perl_is_utf8_X_V(pTHX_ const U8 *p) |
1579 | { |
1580 | dVAR; |
1581 | |
1582 | PERL_ARGS_ASSERT_IS_UTF8_X_V; |
1583 | |
1584 | return is_utf8_common(p, &PL_utf8_X_V, "GCB=V"); |
1585 | } |
1586 | |
1587 | bool |
1588 | Perl_is_utf8_X_LV_LVT_V(pTHX_ const U8 *p) |
1589 | { |
1590 | dVAR; |
1591 | |
1592 | PERL_ARGS_ASSERT_IS_UTF8_X_LV_LVT_V; |
1593 | |
1594 | return is_utf8_common(p, &PL_utf8_X_LV_LVT_V, "_X_LV_LVT_V"); |
1595 | } |
1596 | |
6b5c0936 |
1597 | /* |
87cea99e |
1598 | =for apidoc to_utf8_case |
6b5c0936 |
1599 | |
1600 | The "p" contains the pointer to the UTF-8 string encoding |
1601 | the character that is being converted. |
1602 | |
1603 | The "ustrp" is a pointer to the character buffer to put the |
1604 | conversion result to. The "lenp" is a pointer to the length |
1605 | of the result. |
1606 | |
0134edef |
1607 | The "swashp" is a pointer to the swash to use. |
6b5c0936 |
1608 | |
0134edef |
1609 | Both the special and normal mappings are stored lib/unicore/To/Foo.pl, |
8fe4d5b2 |
1610 | and loaded by SWASHNEW, using lib/utf8_heavy.pl. The special (usually, |
0134edef |
1611 | but not always, a multicharacter mapping), is tried first. |
6b5c0936 |
1612 | |
0134edef |
1613 | The "special" is a string like "utf8::ToSpecLower", which means the |
1614 | hash %utf8::ToSpecLower. The access to the hash is through |
1615 | Perl_to_utf8_case(). |
6b5c0936 |
1616 | |
0134edef |
1617 | The "normal" is a string like "ToLower" which means the swash |
1618 | %utf8::ToLower. |
1619 | |
1620 | =cut */ |
6b5c0936 |
1621 | |
2104c8d9 |
1622 | UV |
9a957fbc |
1623 | Perl_to_utf8_case(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp, |
1624 | SV **swashp, const char *normal, const char *special) |
a0ed51b3 |
1625 | { |
97aff369 |
1626 | dVAR; |
89ebb4a3 |
1627 | U8 tmpbuf[UTF8_MAXBYTES_CASE+1]; |
0134edef |
1628 | STRLEN len = 0; |
aec46f14 |
1629 | const UV uv0 = utf8_to_uvchr(p, NULL); |
1feea2c7 |
1630 | /* The NATIVE_TO_UNI() and UNI_TO_NATIVE() mappings |
1631 | * are necessary in EBCDIC, they are redundant no-ops |
1632 | * in ASCII-ish platforms, and hopefully optimized away. */ |
f54cb97a |
1633 | const UV uv1 = NATIVE_TO_UNI(uv0); |
7918f24d |
1634 | |
1635 | PERL_ARGS_ASSERT_TO_UTF8_CASE; |
1636 | |
1feea2c7 |
1637 | uvuni_to_utf8(tmpbuf, uv1); |
0134edef |
1638 | |
1639 | if (!*swashp) /* load on-demand */ |
1640 | *swashp = swash_init("utf8", normal, &PL_sv_undef, 4, 0); |
37e2e78e |
1641 | /* This is the beginnings of a skeleton of code to read the info section |
1642 | * that is in all the swashes in case we ever want to do that, so one can |
1643 | * read things whose maps aren't code points, and whose default if missing |
1644 | * is not to the code point itself. This was just to see if it actually |
1645 | * worked. Details on what the possibilities are are in perluniprops.pod |
1646 | HV * const hv = get_hv("utf8::SwashInfo", 0); |
1647 | if (hv) { |
1648 | SV **svp; |
1649 | svp = hv_fetch(hv, (const char*)normal, strlen(normal), FALSE); |
1650 | const char *s; |
1651 | |
1652 | HV * const this_hash = SvRV(*svp); |
1653 | svp = hv_fetch(this_hash, "type", strlen("type"), FALSE); |
1654 | s = SvPV_const(*svp, len); |
1655 | } |
1656 | }*/ |
0134edef |
1657 | |
a6f87d8c |
1658 | if (special) { |
0134edef |
1659 | /* It might be "special" (sometimes, but not always, |
2a37f04d |
1660 | * a multicharacter mapping) */ |
6673a63c |
1661 | HV * const hv = get_hv(special, 0); |
b08cf34e |
1662 | SV **svp; |
1663 | |
35da51f7 |
1664 | if (hv && |
b08cf34e |
1665 | (svp = hv_fetch(hv, (const char*)tmpbuf, UNISKIP(uv1), FALSE)) && |
1666 | (*svp)) { |
cfd0369c |
1667 | const char *s; |
47654450 |
1668 | |
cfd0369c |
1669 | s = SvPV_const(*svp, len); |
47654450 |
1670 | if (len == 1) |
1671 | len = uvuni_to_utf8(ustrp, NATIVE_TO_UNI(*(U8*)s)) - ustrp; |
2a37f04d |
1672 | else { |
2f9475ad |
1673 | #ifdef EBCDIC |
1674 | /* If we have EBCDIC we need to remap the characters |
1675 | * since any characters in the low 256 are Unicode |
1676 | * code points, not EBCDIC. */ |
7cda7a3d |
1677 | U8 *t = (U8*)s, *tend = t + len, *d; |
2f9475ad |
1678 | |
1679 | d = tmpbuf; |
b08cf34e |
1680 | if (SvUTF8(*svp)) { |
2f9475ad |
1681 | STRLEN tlen = 0; |
1682 | |
1683 | while (t < tend) { |
d4c19fe8 |
1684 | const UV c = utf8_to_uvchr(t, &tlen); |
2f9475ad |
1685 | if (tlen > 0) { |
1686 | d = uvchr_to_utf8(d, UNI_TO_NATIVE(c)); |
1687 | t += tlen; |
1688 | } |
1689 | else |
1690 | break; |
1691 | } |
1692 | } |
1693 | else { |
36fec512 |
1694 | while (t < tend) { |
1695 | d = uvchr_to_utf8(d, UNI_TO_NATIVE(*t)); |
1696 | t++; |
1697 | } |
2f9475ad |
1698 | } |
1699 | len = d - tmpbuf; |
1700 | Copy(tmpbuf, ustrp, len, U8); |
1701 | #else |
d2dcd0fb |
1702 | Copy(s, ustrp, len, U8); |
2f9475ad |
1703 | #endif |
29e98929 |
1704 | } |
983ffd37 |
1705 | } |
0134edef |
1706 | } |
1707 | |
1708 | if (!len && *swashp) { |
d4c19fe8 |
1709 | const UV uv2 = swash_fetch(*swashp, tmpbuf, TRUE); |
1710 | |
0134edef |
1711 | if (uv2) { |
1712 | /* It was "normal" (a single character mapping). */ |
d4c19fe8 |
1713 | const UV uv3 = UNI_TO_NATIVE(uv2); |
e9101d72 |
1714 | len = uvchr_to_utf8(ustrp, uv3) - ustrp; |
2a37f04d |
1715 | } |
1716 | } |
1feea2c7 |
1717 | |
37e2e78e |
1718 | if (!len) /* Neither: just copy. In other words, there was no mapping |
1719 | defined, which means that the code point maps to itself */ |
0134edef |
1720 | len = uvchr_to_utf8(ustrp, uv0) - ustrp; |
1721 | |
2a37f04d |
1722 | if (lenp) |
1723 | *lenp = len; |
1724 | |
0134edef |
1725 | return len ? utf8_to_uvchr(ustrp, 0) : 0; |
a0ed51b3 |
1726 | } |
1727 | |
d3e79532 |
1728 | /* |
87cea99e |
1729 | =for apidoc to_utf8_upper |
d3e79532 |
1730 | |
1731 | Convert the UTF-8 encoded character at p to its uppercase version and |
1732 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1733 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since |
1734 | the uppercase version may be longer than the original character. |
d3e79532 |
1735 | |
1736 | The first character of the uppercased version is returned |
1737 | (but note, as explained above, that there may be more.) |
1738 | |
1739 | =cut */ |
1740 | |
2104c8d9 |
1741 | UV |
7fc63493 |
1742 | Perl_to_utf8_upper(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 |
1743 | { |
97aff369 |
1744 | dVAR; |
7918f24d |
1745 | |
1746 | PERL_ARGS_ASSERT_TO_UTF8_UPPER; |
1747 | |
983ffd37 |
1748 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1749 | &PL_utf8_toupper, "ToUpper", "utf8::ToSpecUpper"); |
983ffd37 |
1750 | } |
a0ed51b3 |
1751 | |
d3e79532 |
1752 | /* |
87cea99e |
1753 | =for apidoc to_utf8_title |
d3e79532 |
1754 | |
1755 | Convert the UTF-8 encoded character at p to its titlecase version and |
1756 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1757 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
1758 | titlecase version may be longer than the original character. |
d3e79532 |
1759 | |
1760 | The first character of the titlecased version is returned |
1761 | (but note, as explained above, that there may be more.) |
1762 | |
1763 | =cut */ |
1764 | |
983ffd37 |
1765 | UV |
7fc63493 |
1766 | Perl_to_utf8_title(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
983ffd37 |
1767 | { |
97aff369 |
1768 | dVAR; |
7918f24d |
1769 | |
1770 | PERL_ARGS_ASSERT_TO_UTF8_TITLE; |
1771 | |
983ffd37 |
1772 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1773 | &PL_utf8_totitle, "ToTitle", "utf8::ToSpecTitle"); |
a0ed51b3 |
1774 | } |
1775 | |
d3e79532 |
1776 | /* |
87cea99e |
1777 | =for apidoc to_utf8_lower |
d3e79532 |
1778 | |
1779 | Convert the UTF-8 encoded character at p to its lowercase version and |
1780 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1781 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
1782 | lowercase version may be longer than the original character. |
d3e79532 |
1783 | |
1784 | The first character of the lowercased version is returned |
1785 | (but note, as explained above, that there may be more.) |
1786 | |
1787 | =cut */ |
1788 | |
2104c8d9 |
1789 | UV |
7fc63493 |
1790 | Perl_to_utf8_lower(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
a0ed51b3 |
1791 | { |
97aff369 |
1792 | dVAR; |
7918f24d |
1793 | |
1794 | PERL_ARGS_ASSERT_TO_UTF8_LOWER; |
1795 | |
983ffd37 |
1796 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
b4e400f9 |
1797 | &PL_utf8_tolower, "ToLower", "utf8::ToSpecLower"); |
1798 | } |
1799 | |
d3e79532 |
1800 | /* |
87cea99e |
1801 | =for apidoc to_utf8_fold |
d3e79532 |
1802 | |
1803 | Convert the UTF-8 encoded character at p to its foldcase version and |
1804 | store that in UTF-8 in ustrp and its length in bytes in lenp. Note |
89ebb4a3 |
1805 | that the ustrp needs to be at least UTF8_MAXBYTES_CASE+1 bytes since the |
d3e79532 |
1806 | foldcase version may be longer than the original character (up to |
1807 | three characters). |
1808 | |
1809 | The first character of the foldcased version is returned |
1810 | (but note, as explained above, that there may be more.) |
1811 | |
1812 | =cut */ |
1813 | |
b4e400f9 |
1814 | UV |
7fc63493 |
1815 | Perl_to_utf8_fold(pTHX_ const U8 *p, U8* ustrp, STRLEN *lenp) |
b4e400f9 |
1816 | { |
97aff369 |
1817 | dVAR; |
7918f24d |
1818 | |
1819 | PERL_ARGS_ASSERT_TO_UTF8_FOLD; |
1820 | |
b4e400f9 |
1821 | return Perl_to_utf8_case(aTHX_ p, ustrp, lenp, |
1822 | &PL_utf8_tofold, "ToFold", "utf8::ToSpecFold"); |
a0ed51b3 |
1823 | } |
1824 | |
711a919c |
1825 | /* Note: |
1826 | * A "swash" is a swatch hash. |
1827 | * A "swatch" is a bit vector generated by utf8.c:S_swash_get(). |
1828 | * C<pkg> is a pointer to a package name for SWASHNEW, should be "utf8". |
1829 | * For other parameters, see utf8::SWASHNEW in lib/utf8_heavy.pl. |
1830 | */ |
a0ed51b3 |
1831 | SV* |
7fc63493 |
1832 | Perl_swash_init(pTHX_ const char* pkg, const char* name, SV *listsv, I32 minbits, I32 none) |
a0ed51b3 |
1833 | { |
27da23d5 |
1834 | dVAR; |
a0ed51b3 |
1835 | SV* retval; |
8e84507e |
1836 | dSP; |
7fc63493 |
1837 | const size_t pkg_len = strlen(pkg); |
1838 | const size_t name_len = strlen(name); |
da51bb9b |
1839 | HV * const stash = gv_stashpvn(pkg, pkg_len, 0); |
f8be5cf0 |
1840 | SV* errsv_save; |
ce3b816e |
1841 | |
7918f24d |
1842 | PERL_ARGS_ASSERT_SWASH_INIT; |
1843 | |
96ca9f55 |
1844 | PUSHSTACKi(PERLSI_MAGIC); |
1845 | ENTER; |
ec34a119 |
1846 | SAVEHINTS(); |
96ca9f55 |
1847 | save_re_context(); |
1b026014 |
1848 | if (!gv_fetchmeth(stash, "SWASHNEW", 8, -1)) { /* demand load utf8 */ |
ce3b816e |
1849 | ENTER; |
f8be5cf0 |
1850 | errsv_save = newSVsv(ERRSV); |
dc0c6abb |
1851 | /* It is assumed that callers of this routine are not passing in any |
1852 | user derived data. */ |
1853 | /* Need to do this after save_re_context() as it will set PL_tainted to |
1854 | 1 while saving $1 etc (see the code after getrx: in Perl_magic_get). |
1855 | Even line to create errsv_save can turn on PL_tainted. */ |
1856 | SAVEBOOL(PL_tainted); |
1857 | PL_tainted = 0; |
71bed85a |
1858 | Perl_load_module(aTHX_ PERL_LOADMOD_NOIMPORT, newSVpvn(pkg,pkg_len), |
a0714e2c |
1859 | NULL); |
f8be5cf0 |
1860 | if (!SvTRUE(ERRSV)) |
1861 | sv_setsv(ERRSV, errsv_save); |
1862 | SvREFCNT_dec(errsv_save); |
ce3b816e |
1863 | LEAVE; |
1864 | } |
1865 | SPAGAIN; |
a0ed51b3 |
1866 | PUSHMARK(SP); |
1867 | EXTEND(SP,5); |
6e449a3a |
1868 | mPUSHp(pkg, pkg_len); |
1869 | mPUSHp(name, name_len); |
a0ed51b3 |
1870 | PUSHs(listsv); |
6e449a3a |
1871 | mPUSHi(minbits); |
1872 | mPUSHi(none); |
a0ed51b3 |
1873 | PUTBACK; |
f8be5cf0 |
1874 | errsv_save = newSVsv(ERRSV); |
864dbfa3 |
1875 | if (call_method("SWASHNEW", G_SCALAR)) |
8e84507e |
1876 | retval = newSVsv(*PL_stack_sp--); |
a0ed51b3 |
1877 | else |
e24b16f9 |
1878 | retval = &PL_sv_undef; |
f8be5cf0 |
1879 | if (!SvTRUE(ERRSV)) |
1880 | sv_setsv(ERRSV, errsv_save); |
1881 | SvREFCNT_dec(errsv_save); |
a0ed51b3 |
1882 | LEAVE; |
1883 | POPSTACK; |
923e4eb5 |
1884 | if (IN_PERL_COMPILETIME) { |
623e6609 |
1885 | CopHINTS_set(PL_curcop, PL_hints); |
a0ed51b3 |
1886 | } |
bc45ce41 |
1887 | if (!SvROK(retval) || SvTYPE(SvRV(retval)) != SVt_PVHV) { |
1888 | if (SvPOK(retval)) |
35c1215d |
1889 | Perl_croak(aTHX_ "Can't find Unicode property definition \"%"SVf"\"", |
be2597df |
1890 | SVfARG(retval)); |
cea2e8a9 |
1891 | Perl_croak(aTHX_ "SWASHNEW didn't return an HV ref"); |
bc45ce41 |
1892 | } |
a0ed51b3 |
1893 | return retval; |
1894 | } |
1895 | |
035d37be |
1896 | |
1897 | /* This API is wrong for special case conversions since we may need to |
1898 | * return several Unicode characters for a single Unicode character |
1899 | * (see lib/unicore/SpecCase.txt) The SWASHGET in lib/utf8_heavy.pl is |
1900 | * the lower-level routine, and it is similarly broken for returning |
1901 | * multiple values. --jhi */ |
979f2922 |
1902 | /* Now SWASHGET is recasted into S_swash_get in this file. */ |
680c470c |
1903 | |
1904 | /* Note: |
1905 | * Returns the value of property/mapping C<swash> for the first character |
1906 | * of the string C<ptr>. If C<do_utf8> is true, the string C<ptr> is |
1907 | * assumed to be in utf8. If C<do_utf8> is false, the string C<ptr> is |
1908 | * assumed to be in native 8-bit encoding. Caches the swatch in C<swash>. |
1909 | */ |
a0ed51b3 |
1910 | UV |
680c470c |
1911 | Perl_swash_fetch(pTHX_ SV *swash, const U8 *ptr, bool do_utf8) |
a0ed51b3 |
1912 | { |
27da23d5 |
1913 | dVAR; |
ef8f7699 |
1914 | HV *const hv = MUTABLE_HV(SvRV(swash)); |
3568d838 |
1915 | U32 klen; |
1916 | U32 off; |
a0ed51b3 |
1917 | STRLEN slen; |
7d85a32c |
1918 | STRLEN needents; |
cfd0369c |
1919 | const U8 *tmps = NULL; |
a0ed51b3 |
1920 | U32 bit; |
979f2922 |
1921 | SV *swatch; |
3568d838 |
1922 | U8 tmputf8[2]; |
35da51f7 |
1923 | const UV c = NATIVE_TO_ASCII(*ptr); |
3568d838 |
1924 | |
7918f24d |
1925 | PERL_ARGS_ASSERT_SWASH_FETCH; |
1926 | |
3568d838 |
1927 | if (!do_utf8 && !UNI_IS_INVARIANT(c)) { |
979f2922 |
1928 | tmputf8[0] = (U8)UTF8_EIGHT_BIT_HI(c); |
1929 | tmputf8[1] = (U8)UTF8_EIGHT_BIT_LO(c); |
1930 | ptr = tmputf8; |
3568d838 |
1931 | } |
1932 | /* Given a UTF-X encoded char 0xAA..0xYY,0xZZ |
37e2e78e |
1933 | * then the "swatch" is a vec() for all the chars which start |
3568d838 |
1934 | * with 0xAA..0xYY |
1935 | * So the key in the hash (klen) is length of encoded char -1 |
1936 | */ |
1937 | klen = UTF8SKIP(ptr) - 1; |
1938 | off = ptr[klen]; |
a0ed51b3 |
1939 | |
979f2922 |
1940 | if (klen == 0) { |
37e2e78e |
1941 | /* If char is invariant then swatch is for all the invariant chars |
1e54db1a |
1942 | * In both UTF-8 and UTF-8-MOD that happens to be UTF_CONTINUATION_MARK |
7d85a32c |
1943 | */ |
979f2922 |
1944 | needents = UTF_CONTINUATION_MARK; |
1945 | off = NATIVE_TO_UTF(ptr[klen]); |
1946 | } |
1947 | else { |
7d85a32c |
1948 | /* If char is encoded then swatch is for the prefix */ |
979f2922 |
1949 | needents = (1 << UTF_ACCUMULATION_SHIFT); |
1950 | off = NATIVE_TO_UTF(ptr[klen]) & UTF_CONTINUATION_MASK; |
1951 | } |
7d85a32c |
1952 | |
a0ed51b3 |
1953 | /* |
1954 | * This single-entry cache saves about 1/3 of the utf8 overhead in test |
1955 | * suite. (That is, only 7-8% overall over just a hash cache. Still, |
1956 | * it's nothing to sniff at.) Pity we usually come through at least |
1957 | * two function calls to get here... |
1958 | * |
1959 | * NB: this code assumes that swatches are never modified, once generated! |
1960 | */ |
1961 | |
3568d838 |
1962 | if (hv == PL_last_swash_hv && |
a0ed51b3 |
1963 | klen == PL_last_swash_klen && |
27da23d5 |
1964 | (!klen || memEQ((char *)ptr, (char *)PL_last_swash_key, klen)) ) |
a0ed51b3 |
1965 | { |
1966 | tmps = PL_last_swash_tmps; |
1967 | slen = PL_last_swash_slen; |
1968 | } |
1969 | else { |
1970 | /* Try our second-level swatch cache, kept in a hash. */ |
e1ec3a88 |
1971 | SV** svp = hv_fetch(hv, (const char*)ptr, klen, FALSE); |
a0ed51b3 |
1972 | |
979f2922 |
1973 | /* If not cached, generate it via swash_get */ |
1974 | if (!svp || !SvPOK(*svp) |
1975 | || !(tmps = (const U8*)SvPV_const(*svp, slen))) { |
2b9d42f0 |
1976 | /* We use utf8n_to_uvuni() as we want an index into |
1977 | Unicode tables, not a native character number. |
1978 | */ |
aec46f14 |
1979 | const UV code_point = utf8n_to_uvuni(ptr, UTF8_MAXBYTES, 0, |
872c91ae |
1980 | ckWARN(WARN_UTF8) ? |
1981 | 0 : UTF8_ALLOW_ANY); |
680c470c |
1982 | swatch = swash_get(swash, |
979f2922 |
1983 | /* On EBCDIC & ~(0xA0-1) isn't a useful thing to do */ |
1984 | (klen) ? (code_point & ~(needents - 1)) : 0, |
1985 | needents); |
1986 | |
923e4eb5 |
1987 | if (IN_PERL_COMPILETIME) |
623e6609 |
1988 | CopHINTS_set(PL_curcop, PL_hints); |
a0ed51b3 |
1989 | |
979f2922 |
1990 | svp = hv_store(hv, (const char *)ptr, klen, swatch, 0); |
a0ed51b3 |
1991 | |
979f2922 |
1992 | if (!svp || !(tmps = (U8*)SvPV(*svp, slen)) |
1993 | || (slen << 3) < needents) |
660a4616 |
1994 | Perl_croak(aTHX_ "panic: swash_fetch got improper swatch"); |
a0ed51b3 |
1995 | } |
1996 | |
1997 | PL_last_swash_hv = hv; |
16d8f38a |
1998 | assert(klen <= sizeof(PL_last_swash_key)); |
eac04b2e |
1999 | PL_last_swash_klen = (U8)klen; |
cfd0369c |
2000 | /* FIXME change interpvar.h? */ |
2001 | PL_last_swash_tmps = (U8 *) tmps; |
a0ed51b3 |
2002 | PL_last_swash_slen = slen; |
2003 | if (klen) |
2004 | Copy(ptr, PL_last_swash_key, klen, U8); |
2005 | } |
2006 | |
9faf8d75 |
2007 | switch ((int)((slen << 3) / needents)) { |
a0ed51b3 |
2008 | case 1: |
2009 | bit = 1 << (off & 7); |
2010 | off >>= 3; |
2011 | return (tmps[off] & bit) != 0; |
2012 | case 8: |
2013 | return tmps[off]; |
2014 | case 16: |
2015 | off <<= 1; |
2016 | return (tmps[off] << 8) + tmps[off + 1] ; |
2017 | case 32: |
2018 | off <<= 2; |
2019 | return (tmps[off] << 24) + (tmps[off+1] << 16) + (tmps[off+2] << 8) + tmps[off + 3] ; |
2020 | } |
660a4616 |
2021 | Perl_croak(aTHX_ "panic: swash_fetch got swatch of unexpected bit width"); |
670f1322 |
2022 | NORETURN_FUNCTION_END; |
a0ed51b3 |
2023 | } |
2b9d42f0 |
2024 | |
979f2922 |
2025 | /* Note: |
2026 | * Returns a swatch (a bit vector string) for a code point sequence |
2027 | * that starts from the value C<start> and comprises the number C<span>. |
2028 | * A C<swash> must be an object created by SWASHNEW (see lib/utf8_heavy.pl). |
2029 | * Should be used via swash_fetch, which will cache the swatch in C<swash>. |
2030 | */ |
2031 | STATIC SV* |
2032 | S_swash_get(pTHX_ SV* swash, UV start, UV span) |
2033 | { |
2034 | SV *swatch; |
711a919c |
2035 | U8 *l, *lend, *x, *xend, *s; |
979f2922 |
2036 | STRLEN lcur, xcur, scur; |
ef8f7699 |
2037 | HV *const hv = MUTABLE_HV(SvRV(swash)); |
017a3ce5 |
2038 | SV** const listsvp = hv_fetchs(hv, "LIST", FALSE); |
2039 | SV** const typesvp = hv_fetchs(hv, "TYPE", FALSE); |
2040 | SV** const bitssvp = hv_fetchs(hv, "BITS", FALSE); |
2041 | SV** const nonesvp = hv_fetchs(hv, "NONE", FALSE); |
2042 | SV** const extssvp = hv_fetchs(hv, "EXTRAS", FALSE); |
0bd48802 |
2043 | const U8* const typestr = (U8*)SvPV_nolen(*typesvp); |
2044 | const int typeto = typestr[0] == 'T' && typestr[1] == 'o'; |
2045 | const STRLEN bits = SvUV(*bitssvp); |
2046 | const STRLEN octets = bits >> 3; /* if bits == 1, then octets == 0 */ |
2047 | const UV none = SvUV(*nonesvp); |
2048 | const UV end = start + span; |
979f2922 |
2049 | |
7918f24d |
2050 | PERL_ARGS_ASSERT_SWASH_GET; |
2051 | |
979f2922 |
2052 | if (bits != 1 && bits != 8 && bits != 16 && bits != 32) { |
660a4616 |
2053 | Perl_croak(aTHX_ "panic: swash_get doesn't expect bits %"UVuf, |
2054 | (UV)bits); |
979f2922 |
2055 | } |
2056 | |
2057 | /* create and initialize $swatch */ |
979f2922 |
2058 | scur = octets ? (span * octets) : (span + 7) / 8; |
e524fe40 |
2059 | swatch = newSV(scur); |
2060 | SvPOK_on(swatch); |
979f2922 |
2061 | s = (U8*)SvPVX(swatch); |
2062 | if (octets && none) { |
0bd48802 |
2063 | const U8* const e = s + scur; |
979f2922 |
2064 | while (s < e) { |
2065 | if (bits == 8) |
2066 | *s++ = (U8)(none & 0xff); |
2067 | else if (bits == 16) { |
2068 | *s++ = (U8)((none >> 8) & 0xff); |
2069 | *s++ = (U8)( none & 0xff); |
2070 | } |
2071 | else if (bits == 32) { |
2072 | *s++ = (U8)((none >> 24) & 0xff); |
2073 | *s++ = (U8)((none >> 16) & 0xff); |
2074 | *s++ = (U8)((none >> 8) & 0xff); |
2075 | *s++ = (U8)( none & 0xff); |
2076 | } |
2077 | } |
2078 | *s = '\0'; |
2079 | } |
2080 | else { |
2081 | (void)memzero((U8*)s, scur + 1); |
2082 | } |
2083 | SvCUR_set(swatch, scur); |
2084 | s = (U8*)SvPVX(swatch); |
2085 | |
2086 | /* read $swash->{LIST} */ |
2087 | l = (U8*)SvPV(*listsvp, lcur); |
2088 | lend = l + lcur; |
2089 | while (l < lend) { |
35da51f7 |
2090 | UV min, max, val; |
979f2922 |
2091 | STRLEN numlen; |
2092 | I32 flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX; |
2093 | |
0bd48802 |
2094 | U8* const nl = (U8*)memchr(l, '\n', lend - l); |
979f2922 |
2095 | |
2096 | numlen = lend - l; |
2097 | min = grok_hex((char *)l, &numlen, &flags, NULL); |
2098 | if (numlen) |
2099 | l += numlen; |
2100 | else if (nl) { |
2101 | l = nl + 1; /* 1 is length of "\n" */ |
2102 | continue; |
2103 | } |
2104 | else { |
2105 | l = lend; /* to LIST's end at which \n is not found */ |
2106 | break; |
2107 | } |
2108 | |
2109 | if (isBLANK(*l)) { |
2110 | ++l; |
2111 | flags = PERL_SCAN_SILENT_ILLDIGIT | PERL_SCAN_DISALLOW_PREFIX; |
2112 | numlen = lend - l; |
2113 | max = grok_hex((char *)l, &numlen, &flags, NULL); |
2114 | if (numlen) |
2115 | l += numlen; |
2116 | else |
2117 | max = min; |
2118 | |
2119 | if (octets) { |
2120 | if (isBLANK(*l)) { |
2121 | ++l; |
2122 | flags = PERL_SCAN_SILENT_ILLDIGIT | |
2123 | PERL_SCAN_DISALLOW_PREFIX; |
2124 | numlen = lend - l; |
2125 | val = grok_hex((char *)l, &numlen, &flags, NULL); |
2126 | if (numlen) |
2127 | l += numlen; |
2128 | else |
2129 | val = 0; |
2130 | } |
2131 | else { |
2132 | val = 0; |
2133 | if (typeto) { |
2134 | Perl_croak(aTHX_ "%s: illegal mapping '%s'", |
2135 | typestr, l); |
2136 | } |
2137 | } |
2138 | } |
711a919c |
2139 | else |
2140 | val = 0; /* bits == 1, then val should be ignored */ |
979f2922 |
2141 | } |
2142 | else { |
2143 | max = min; |
2144 | if (octets) { |
2145 | val = 0; |
2146 | if (typeto) { |
2147 | Perl_croak(aTHX_ "%s: illegal mapping '%s'", typestr, l); |
2148 | } |
2149 | } |
711a919c |
2150 | else |
2151 | val = 0; /* bits == 1, then val should be ignored */ |
979f2922 |
2152 | } |
2153 | |
2154 | if (nl) |
2155 | l = nl + 1; |
2156 | else |
2157 | l = lend; |
2158 | |
2159 | if (max < start) |
2160 | continue; |
2161 | |
2162 | if (octets) { |
35da51f7 |
2163 | UV key; |
979f2922 |
2164 | if (min < start) { |
2165 | if (!none || val < none) { |
2166 | val += start - min; |
2167 | } |
2168 | min = start; |
2169 | } |
2170 | for (key = min; key <= max; key++) { |
2171 | STRLEN offset; |
2172 | if (key >= end) |
2173 | goto go_out_list; |
2174 | /* offset must be non-negative (start <= min <= key < end) */ |
2175 | offset = octets * (key - start); |
2176 | if (bits == 8) |
2177 | s[offset] = (U8)(val & 0xff); |
2178 | else if (bits == 16) { |
2179 | s[offset ] = (U8)((val >> 8) & 0xff); |
2180 | s[offset + 1] = (U8)( val & 0xff); |
2181 | } |
2182 | else if (bits == 32) { |
2183 | s[offset ] = (U8)((val >> 24) & 0xff); |
2184 | s[offset + 1] = (U8)((val >> 16) & 0xff); |
2185 | s[offset + 2] = (U8)((val >> 8) & 0xff); |
2186 | s[offset + 3] = (U8)( val & 0xff); |
2187 | } |
2188 | |
2189 | if (!none || val < none) |
2190 | ++val; |
2191 | } |
2192 | } |
711a919c |
2193 | else { /* bits == 1, then val should be ignored */ |
35da51f7 |
2194 | UV key; |
979f2922 |
2195 | if (min < start) |
2196 | min = start; |
2197 | for (key = min; key <= max; key++) { |
0bd48802 |
2198 | const STRLEN offset = (STRLEN)(key - start); |
979f2922 |
2199 | if (key >= end) |
2200 | goto go_out_list; |
2201 | s[offset >> 3] |= 1 << (offset & 7); |
2202 | } |
2203 | } |
2204 | } /* while */ |
2205 | go_out_list: |
2206 | |
2207 | /* read $swash->{EXTRAS} */ |
2208 | x = (U8*)SvPV(*extssvp, xcur); |
2209 | xend = x + xcur; |
2210 | while (x < xend) { |
2211 | STRLEN namelen; |
2212 | U8 *namestr; |
2213 | SV** othersvp; |
2214 | HV* otherhv; |
2215 | STRLEN otherbits; |
2216 | SV **otherbitssvp, *other; |
711a919c |
2217 | U8 *s, *o, *nl; |
979f2922 |
2218 | STRLEN slen, olen; |
2219 | |
35da51f7 |
2220 | const U8 opc = *x++; |
979f2922 |
2221 | if (opc == '\n') |
2222 | continue; |
2223 | |
2224 | nl = (U8*)memchr(x, '\n', xend - x); |
2225 | |
2226 | if (opc != '-' && opc != '+' && opc != '!' && opc != '&') { |
2227 | if (nl) { |
2228 | x = nl + 1; /* 1 is length of "\n" */ |
2229 | continue; |
2230 | } |
2231 | else { |
2232 | x = xend; /* to EXTRAS' end at which \n is not found */ |
2233 | break; |
2234 | } |
2235 | } |
2236 | |
2237 | namestr = x; |
2238 | if (nl) { |
2239 | namelen = nl - namestr; |
2240 | x = nl + 1; |
2241 | } |
2242 | else { |
2243 | namelen = xend - namestr; |
2244 | x = xend; |
2245 | } |
2246 | |
2247 | othersvp = hv_fetch(hv, (char *)namestr, namelen, FALSE); |
ef8f7699 |
2248 | otherhv = MUTABLE_HV(SvRV(*othersvp)); |
017a3ce5 |
2249 | otherbitssvp = hv_fetchs(otherhv, "BITS", FALSE); |
979f2922 |
2250 | otherbits = (STRLEN)SvUV(*otherbitssvp); |
2251 | if (bits < otherbits) |
660a4616 |
2252 | Perl_croak(aTHX_ "panic: swash_get found swatch size mismatch"); |
979f2922 |
2253 | |
2254 | /* The "other" swatch must be destroyed after. */ |
2255 | other = swash_get(*othersvp, start, span); |
2256 | o = (U8*)SvPV(other, olen); |
2257 | |
2258 | if (!olen) |
660a4616 |
2259 | Perl_croak(aTHX_ "panic: swash_get got improper swatch"); |
979f2922 |
2260 | |
2261 | s = (U8*)SvPV(swatch, slen); |
2262 | if (bits == 1 && otherbits == 1) { |
2263 | if (slen != olen) |
660a4616 |
2264 | Perl_croak(aTHX_ "panic: swash_get found swatch length mismatch"); |
979f2922 |
2265 | |
2266 | switch (opc) { |
2267 | case '+': |
2268 | while (slen--) |
2269 | *s++ |= *o++; |
2270 | break; |
2271 | case '!': |
2272 | while (slen--) |
2273 | *s++ |= ~*o++; |
2274 | break; |
2275 | case '-': |
2276 | while (slen--) |
2277 | *s++ &= ~*o++; |
2278 | break; |
2279 | case '&': |
2280 | while (slen--) |
2281 | *s++ &= *o++; |
2282 | break; |
2283 | default: |
2284 | break; |
2285 | } |
2286 | } |
711a919c |
2287 | else { |
979f2922 |
2288 | STRLEN otheroctets = otherbits >> 3; |
2289 | STRLEN offset = 0; |
35da51f7 |
2290 | U8* const send = s + slen; |
979f2922 |
2291 | |
2292 | while (s < send) { |
2293 | UV otherval = 0; |
2294 | |
2295 | if (otherbits == 1) { |
2296 | otherval = (o[offset >> 3] >> (offset & 7)) & 1; |
2297 | ++offset; |
2298 | } |
2299 | else { |
2300 | STRLEN vlen = otheroctets; |
2301 | otherval = *o++; |
2302 | while (--vlen) { |
2303 | otherval <<= 8; |
2304 | otherval |= *o++; |
2305 | } |
2306 | } |
2307 | |
711a919c |
2308 | if (opc == '+' && otherval) |
6f207bd3 |
2309 | NOOP; /* replace with otherval */ |
979f2922 |
2310 | else if (opc == '!' && !otherval) |
2311 | otherval = 1; |
2312 | else if (opc == '-' && otherval) |
2313 | otherval = 0; |
2314 | else if (opc == '&' && !otherval) |
2315 | otherval = 0; |
2316 | else { |
711a919c |
2317 | s += octets; /* no replacement */ |
979f2922 |
2318 | continue; |
2319 | } |
2320 | |
2321 | if (bits == 8) |
2322 | *s++ = (U8)( otherval & 0xff); |
2323 | else if (bits == 16) { |
2324 | *s++ = (U8)((otherval >> 8) & 0xff); |
2325 | *s++ = (U8)( otherval & 0xff); |
2326 | } |
2327 | else if (bits == 32) { |
2328 | *s++ = (U8)((otherval >> 24) & 0xff); |
2329 | *s++ = (U8)((otherval >> 16) & 0xff); |
2330 | *s++ = (U8)((otherval >> 8) & 0xff); |
2331 | *s++ = (U8)( otherval & 0xff); |
2332 | } |
2333 | } |
2334 | } |
2335 | sv_free(other); /* through with it! */ |
2336 | } /* while */ |
2337 | return swatch; |
2338 | } |
2339 | |
0f830e0b |
2340 | /* |
87cea99e |
2341 | =for apidoc uvchr_to_utf8 |
0f830e0b |
2342 | |
2343 | Adds the UTF-8 representation of the Native codepoint C<uv> to the end |
2344 | of the string C<d>; C<d> should be have at least C<UTF8_MAXBYTES+1> free |
2345 | bytes available. The return value is the pointer to the byte after the |
2346 | end of the new character. In other words, |
2347 | |
2348 | d = uvchr_to_utf8(d, uv); |
2349 | |
2350 | is the recommended wide native character-aware way of saying |
2351 | |
2352 | *(d++) = uv; |
2353 | |
2354 | =cut |
2355 | */ |
2356 | |
2357 | /* On ASCII machines this is normally a macro but we want a |
2358 | real function in case XS code wants it |
2359 | */ |
0f830e0b |
2360 | U8 * |
2361 | Perl_uvchr_to_utf8(pTHX_ U8 *d, UV uv) |
2362 | { |
7918f24d |
2363 | PERL_ARGS_ASSERT_UVCHR_TO_UTF8; |
2364 | |
0f830e0b |
2365 | return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), 0); |
2366 | } |
2367 | |
b851fbc1 |
2368 | U8 * |
2369 | Perl_uvchr_to_utf8_flags(pTHX_ U8 *d, UV uv, UV flags) |
2370 | { |
7918f24d |
2371 | PERL_ARGS_ASSERT_UVCHR_TO_UTF8_FLAGS; |
2372 | |
b851fbc1 |
2373 | return Perl_uvuni_to_utf8_flags(aTHX_ d, NATIVE_TO_UNI(uv), flags); |
2374 | } |
2b9d42f0 |
2375 | |
2376 | /* |
87cea99e |
2377 | =for apidoc utf8n_to_uvchr |
0f830e0b |
2378 | flags |
2379 | |
48ef279e |
2380 | Returns the native character value of the first character in the string |
0f830e0b |
2381 | C<s> |
2382 | which is assumed to be in UTF-8 encoding; C<retlen> will be set to the |
2383 | length, in bytes, of that character. |
2384 | |
2385 | Allows length and flags to be passed to low level routine. |
2386 | |
2387 | =cut |
2388 | */ |
2389 | /* On ASCII machines this is normally a macro but we want |
2390 | a real function in case XS code wants it |
2391 | */ |
0f830e0b |
2392 | UV |
48ef279e |
2393 | Perl_utf8n_to_uvchr(pTHX_ const U8 *s, STRLEN curlen, STRLEN *retlen, |
0f830e0b |
2394 | U32 flags) |
2395 | { |
2396 | const UV uv = Perl_utf8n_to_uvuni(aTHX_ s, curlen, retlen, flags); |
7918f24d |
2397 | |
2398 | PERL_ARGS_ASSERT_UTF8N_TO_UVCHR; |
2399 | |
0f830e0b |
2400 | return UNI_TO_NATIVE(uv); |
2401 | } |
2402 | |
2403 | /* |
87cea99e |
2404 | =for apidoc pv_uni_display |
d2cc3551 |
2405 | |
2406 | Build to the scalar dsv a displayable version of the string spv, |
2407 | length len, the displayable version being at most pvlim bytes long |
2408 | (if longer, the rest is truncated and "..." will be appended). |
0a2ef054 |
2409 | |
9e55ce06 |
2410 | The flags argument can have UNI_DISPLAY_ISPRINT set to display |
00e86452 |
2411 | isPRINT()able characters as themselves, UNI_DISPLAY_BACKSLASH |
0a2ef054 |
2412 | to display the \\[nrfta\\] as the backslashed versions (like '\n') |
2413 | (UNI_DISPLAY_BACKSLASH is preferred over UNI_DISPLAY_ISPRINT for \\). |
2414 | UNI_DISPLAY_QQ (and its alias UNI_DISPLAY_REGEX) have both |
2415 | UNI_DISPLAY_BACKSLASH and UNI_DISPLAY_ISPRINT turned on. |
2416 | |
d2cc3551 |
2417 | The pointer to the PV of the dsv is returned. |
2418 | |
2419 | =cut */ |
e6b2e755 |
2420 | char * |
e1ec3a88 |
2421 | Perl_pv_uni_display(pTHX_ SV *dsv, const U8 *spv, STRLEN len, STRLEN pvlim, UV flags) |
e6b2e755 |
2422 | { |
2423 | int truncated = 0; |
e1ec3a88 |
2424 | const char *s, *e; |
e6b2e755 |
2425 | |
7918f24d |
2426 | PERL_ARGS_ASSERT_PV_UNI_DISPLAY; |
2427 | |
76f68e9b |
2428 | sv_setpvs(dsv, ""); |
7fddd944 |
2429 | SvUTF8_off(dsv); |
e1ec3a88 |
2430 | for (s = (const char *)spv, e = s + len; s < e; s += UTF8SKIP(s)) { |
e6b2e755 |
2431 | UV u; |
a49f32c6 |
2432 | /* This serves double duty as a flag and a character to print after |
2433 | a \ when flags & UNI_DISPLAY_BACKSLASH is true. |
2434 | */ |
2435 | char ok = 0; |
c728cb41 |
2436 | |
e6b2e755 |
2437 | if (pvlim && SvCUR(dsv) >= pvlim) { |
2438 | truncated++; |
2439 | break; |
2440 | } |
2441 | u = utf8_to_uvchr((U8*)s, 0); |
c728cb41 |
2442 | if (u < 256) { |
a3b680e6 |
2443 | const unsigned char c = (unsigned char)u & 0xFF; |
0bd48802 |
2444 | if (flags & UNI_DISPLAY_BACKSLASH) { |
a49f32c6 |
2445 | switch (c) { |
c728cb41 |
2446 | case '\n': |
a49f32c6 |
2447 | ok = 'n'; break; |
c728cb41 |
2448 | case '\r': |
a49f32c6 |
2449 | ok = 'r'; break; |
c728cb41 |
2450 | case '\t': |
a49f32c6 |
2451 | ok = 't'; break; |
c728cb41 |
2452 | case '\f': |
a49f32c6 |
2453 | ok = 'f'; break; |
c728cb41 |
2454 | case '\a': |
a49f32c6 |
2455 | ok = 'a'; break; |
c728cb41 |
2456 | case '\\': |
a49f32c6 |
2457 | ok = '\\'; break; |
c728cb41 |
2458 | default: break; |
2459 | } |
a49f32c6 |
2460 | if (ok) { |
88c9ea1e |
2461 | const char string = ok; |
76f68e9b |
2462 | sv_catpvs(dsv, "\\"); |
5e7aa789 |
2463 | sv_catpvn(dsv, &string, 1); |
a49f32c6 |
2464 | } |
c728cb41 |
2465 | } |
00e86452 |
2466 | /* isPRINT() is the locale-blind version. */ |
a49f32c6 |
2467 | if (!ok && (flags & UNI_DISPLAY_ISPRINT) && isPRINT(c)) { |
88c9ea1e |
2468 | const char string = c; |
5e7aa789 |
2469 | sv_catpvn(dsv, &string, 1); |
a49f32c6 |
2470 | ok = 1; |
0a2ef054 |
2471 | } |
c728cb41 |
2472 | } |
2473 | if (!ok) |
9e55ce06 |
2474 | Perl_sv_catpvf(aTHX_ dsv, "\\x{%"UVxf"}", u); |
e6b2e755 |
2475 | } |
2476 | if (truncated) |
396482e1 |
2477 | sv_catpvs(dsv, "..."); |
48ef279e |
2478 | |
e6b2e755 |
2479 | return SvPVX(dsv); |
2480 | } |
2b9d42f0 |
2481 | |
d2cc3551 |
2482 | /* |
87cea99e |
2483 | =for apidoc sv_uni_display |
d2cc3551 |
2484 | |
2485 | Build to the scalar dsv a displayable version of the scalar sv, |
0a2ef054 |
2486 | the displayable version being at most pvlim bytes long |
d2cc3551 |
2487 | (if longer, the rest is truncated and "..." will be appended). |
0a2ef054 |
2488 | |
2489 | The flags argument is as in pv_uni_display(). |
2490 | |
d2cc3551 |
2491 | The pointer to the PV of the dsv is returned. |
2492 | |
d4c19fe8 |
2493 | =cut |
2494 | */ |
e6b2e755 |
2495 | char * |
2496 | Perl_sv_uni_display(pTHX_ SV *dsv, SV *ssv, STRLEN pvlim, UV flags) |
2497 | { |
7918f24d |
2498 | PERL_ARGS_ASSERT_SV_UNI_DISPLAY; |
2499 | |
cfd0369c |
2500 | return Perl_pv_uni_display(aTHX_ dsv, (const U8*)SvPVX_const(ssv), |
2501 | SvCUR(ssv), pvlim, flags); |
701a277b |
2502 | } |
2503 | |
d2cc3551 |
2504 | /* |
e6226b18 |
2505 | =for apidoc foldEQ_utf8 |
d2cc3551 |
2506 | |
d51c1b21 |
2507 | Returns true if the leading portions of the strings s1 and s2 (either or both |
e6226b18 |
2508 | of which may be in UTF-8) are the same case-insensitively; false otherwise. |
d51c1b21 |
2509 | How far into the strings to compare is determined by other input parameters. |
8b35872c |
2510 | |
2511 | If u1 is true, the string s1 is assumed to be in UTF-8-encoded Unicode; |
2512 | otherwise it is assumed to be in native 8-bit encoding. Correspondingly for u2 |
2513 | with respect to s2. |
2514 | |
d51c1b21 |
2515 | If the byte length l1 is non-zero, it says how far into s1 to check for fold |
2516 | equality. In other words, s1+l1 will be used as a goal to reach. The |
8b35872c |
2517 | scan will not be considered to be a match unless the goal is reached, and |
2518 | scanning won't continue past that goal. Correspondingly for l2 with respect to |
2519 | s2. |
2520 | |
2521 | If pe1 is non-NULL and the pointer it points to is not NULL, that pointer is |
2522 | considered an end pointer beyond which scanning of s1 will not continue under |
2523 | any circumstances. This means that if both l1 and pe1 are specified, and pe1 |
2524 | is less than s1+l1, the match will never be successful because it can never |
d51c1b21 |
2525 | get as far as its goal (and in fact is asserted against). Correspondingly for |
2526 | pe2 with respect to s2. |
8b35872c |
2527 | |
d51c1b21 |
2528 | At least one of s1 and s2 must have a goal (at least one of l1 and l2 must be |
2529 | non-zero), and if both do, both have to be |
8b35872c |
2530 | reached for a successful match. Also, if the fold of a character is multiple |
2531 | characters, all of them must be matched (see tr21 reference below for |
2532 | 'folding'). |
2533 | |
e6226b18 |
2534 | Upon a successful match, if pe1 is non-NULL, |
8b35872c |
2535 | it will be set to point to the beginning of the I<next> character of s1 beyond |
2536 | what was matched. Correspondingly for pe2 and s2. |
d2cc3551 |
2537 | |
2538 | For case-insensitiveness, the "casefolding" of Unicode is used |
2539 | instead of upper/lowercasing both the characters, see |
2540 | http://www.unicode.org/unicode/reports/tr21/ (Case Mappings). |
2541 | |
2542 | =cut */ |
701a277b |
2543 | I32 |
e6226b18 |
2544 | Perl_foldEQ_utf8(pTHX_ const char *s1, char **pe1, register UV l1, bool u1, const char *s2, char **pe2, register UV l2, bool u2) |
332ddc25 |
2545 | { |
8b35872c |
2546 | dVAR; |
2547 | register const U8 *p1 = (const U8*)s1; /* Point to current char */ |
2548 | register const U8 *p2 = (const U8*)s2; |
48ef279e |
2549 | register const U8 *g1 = NULL; /* goal for s1 */ |
8b35872c |
2550 | register const U8 *g2 = NULL; |
48ef279e |
2551 | register const U8 *e1 = NULL; /* Don't scan s1 past this */ |
2552 | register U8 *f1 = NULL; /* Point to current folded */ |
8b35872c |
2553 | register const U8 *e2 = NULL; |
2554 | register U8 *f2 = NULL; |
48ef279e |
2555 | STRLEN n1 = 0, n2 = 0; /* Number of bytes in current char */ |
8b35872c |
2556 | U8 foldbuf1[UTF8_MAXBYTES_CASE+1]; |
2557 | U8 foldbuf2[UTF8_MAXBYTES_CASE+1]; |
48ef279e |
2558 | U8 natbuf[2]; /* Holds native 8-bit char converted to utf8; |
2559 | these always fit in 2 bytes */ |
8b35872c |
2560 | |
e6226b18 |
2561 | PERL_ARGS_ASSERT_FOLDEQ_UTF8; |
8b35872c |
2562 | |
2563 | if (pe1) { |
48ef279e |
2564 | e1 = *(U8**)pe1; |
8b35872c |
2565 | } |
2566 | |
2567 | if (l1) { |
48ef279e |
2568 | g1 = (const U8*)s1 + l1; |
8b35872c |
2569 | } |
2570 | |
2571 | if (pe2) { |
48ef279e |
2572 | e2 = *(U8**)pe2; |
8b35872c |
2573 | } |
2574 | |
2575 | if (l2) { |
48ef279e |
2576 | g2 = (const U8*)s2 + l2; |
8b35872c |
2577 | } |
2578 | |
2579 | /* Must have at least one goal */ |
2580 | assert(g1 || g2); |
2581 | |
2582 | if (g1) { |
2583 | |
48ef279e |
2584 | /* Will never match if goal is out-of-bounds */ |
2585 | assert(! e1 || e1 >= g1); |
8b35872c |
2586 | |
48ef279e |
2587 | /* Here, there isn't an end pointer, or it is beyond the goal. We |
2588 | * only go as far as the goal */ |
2589 | e1 = g1; |
8b35872c |
2590 | } |
313b38e5 |
2591 | else { |
2592 | assert(e1); /* Must have an end for looking at s1 */ |
2593 | } |
8b35872c |
2594 | |
2595 | /* Same for goal for s2 */ |
2596 | if (g2) { |
48ef279e |
2597 | assert(! e2 || e2 >= g2); |
2598 | e2 = g2; |
8b35872c |
2599 | } |
313b38e5 |
2600 | else { |
2601 | assert(e2); |
2602 | } |
8b35872c |
2603 | |
2604 | /* Look through both strings, a character at a time */ |
2605 | while (p1 < e1 && p2 < e2) { |
2606 | |
d51c1b21 |
2607 | /* If at the beginning of a new character in s1, get its fold to use |
2608 | * and the length of the fold */ |
48ef279e |
2609 | if (n1 == 0) { |
2610 | if (u1) { |
2611 | to_utf8_fold(p1, foldbuf1, &n1); |
2612 | } |
2613 | else { /* Not utf8, convert to it first and then get fold */ |
2614 | uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p1))); |
2615 | to_utf8_fold(natbuf, foldbuf1, &n1); |
2616 | } |
2617 | f1 = foldbuf1; |
2618 | } |
8b35872c |
2619 | |
48ef279e |
2620 | if (n2 == 0) { /* Same for s2 */ |
2621 | if (u2) { |
2622 | to_utf8_fold(p2, foldbuf2, &n2); |
2623 | } |
2624 | else { |
2625 | uvuni_to_utf8(natbuf, (UV) NATIVE_TO_UNI(((UV)*p2))); |
2626 | to_utf8_fold(natbuf, foldbuf2, &n2); |
2627 | } |
2628 | f2 = foldbuf2; |
2629 | } |
8b35872c |
2630 | |
48ef279e |
2631 | /* While there is more to look for in both folds, see if they |
2632 | * continue to match */ |
2633 | while (n1 && n2) { |
2634 | U8 fold_length = UTF8SKIP(f1); |
2635 | if (fold_length != UTF8SKIP(f2) |
2636 | || (fold_length == 1 && *f1 != *f2) /* Short circuit memNE |
2637 | function call for single |
2638 | character */ |
2639 | || memNE((char*)f1, (char*)f2, fold_length)) |
2640 | { |
e6226b18 |
2641 | return 0; /* mismatch */ |
48ef279e |
2642 | } |
2643 | |
2644 | /* Here, they matched, advance past them */ |
2645 | n1 -= fold_length; |
2646 | f1 += fold_length; |
2647 | n2 -= fold_length; |
2648 | f2 += fold_length; |
2649 | } |
8b35872c |
2650 | |
48ef279e |
2651 | /* When reach the end of any fold, advance the input past it */ |
2652 | if (n1 == 0) { |
2653 | p1 += u1 ? UTF8SKIP(p1) : 1; |
2654 | } |
2655 | if (n2 == 0) { |
2656 | p2 += u2 ? UTF8SKIP(p2) : 1; |
2657 | } |
8b35872c |
2658 | } /* End of loop through both strings */ |
2659 | |
2660 | /* A match is defined by each scan that specified an explicit length |
2661 | * reaching its final goal, and the other not having matched a partial |
2662 | * character (which can happen when the fold of a character is more than one |
2663 | * character). */ |
2664 | if (! ((g1 == 0 || p1 == g1) && (g2 == 0 || p2 == g2)) || n1 || n2) { |
e6226b18 |
2665 | return 0; |
8b35872c |
2666 | } |
2667 | |
2668 | /* Successful match. Set output pointers */ |
2669 | if (pe1) { |
48ef279e |
2670 | *pe1 = (char*)p1; |
8b35872c |
2671 | } |
2672 | if (pe2) { |
48ef279e |
2673 | *pe2 = (char*)p2; |
8b35872c |
2674 | } |
e6226b18 |
2675 | return 1; |
e6b2e755 |
2676 | } |
701a277b |
2677 | |
a49f32c6 |
2678 | /* |
2679 | * Local variables: |
2680 | * c-indentation-style: bsd |
2681 | * c-basic-offset: 4 |
2682 | * indent-tabs-mode: t |
2683 | * End: |
2684 | * |
37442d52 |
2685 | * ex: set ts=8 sts=4 sw=4 noet: |
2686 | */ |