Commit | Line | Data |
---|---|---|
a0ed51b3 LW |
1 | /* utf8.h |
2 | * | |
83706693 | 3 | * Copyright (C) 2000, 2001, 2002, 2005, 2006, 2007, 2009 by Larry Wall and others |
a0ed51b3 LW |
4 | * |
5 | * You may distribute under the terms of either the GNU General Public | |
6 | * License or the Artistic License, as specified in the README file. | |
7 | * | |
8 | */ | |
9 | ||
39e02b42 | 10 | /* Use UTF-8 as the default script encoding? |
1e54db1a | 11 | * Turning this on will break scripts having non-UTF-8 binary |
39e02b42 JH |
12 | * data (such as Latin-1) in string literals. */ |
13 | #ifdef USE_UTF8_SCRIPTS | |
14 | # define USE_UTF8_IN_NAMES (!IN_BYTES) | |
15 | #else | |
16 | # define USE_UTF8_IN_NAMES (PL_hints & HINT_UTF8) | |
17 | #endif | |
18 | ||
36bb2ab6 KW |
19 | #define to_uni_fold(c, p, lenp) _to_uni_fold_flags(c, p, lenp, 1) |
20 | #define to_utf8_fold(c, p, lenp) _to_utf8_fold_flags(c, p, lenp, 1) | |
21 | ||
fd7cb289 RGS |
22 | /* Source backward compatibility. */ |
23 | #define uvuni_to_utf8(d, uv) uvuni_to_utf8_flags(d, uv, 0) | |
24 | #define is_utf8_string_loc(s, len, ep) is_utf8_string_loclen(s, len, ep, 0) | |
25 | ||
eda9cac1 KW |
26 | #define foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \ |
27 | foldEQ_utf8_flags(s1, pe1, l1, u1, s2, pe2, l2, u2, 0) | |
a33c29bc | 28 | #define FOLDEQ_UTF8_NOMIX_ASCII (1 << 0) |
5e64d0fa | 29 | #define FOLDEQ_UTF8_LOCALE (1 << 1) |
a33c29bc | 30 | |
e6226b18 KW |
31 | /* |
32 | =for apidoc ibcmp_utf8 | |
33 | ||
34 | This is a synonym for (! foldEQ_utf8()) | |
35 | ||
36 | =cut | |
37 | */ | |
38 | #define ibcmp_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \ | |
39 | cBOOL(! foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2)) | |
40 | ||
1d72bdf6 NIS |
41 | #ifdef EBCDIC |
42 | /* The equivalent of these macros but implementing UTF-EBCDIC | |
43 | are in the following header file: | |
44 | */ | |
45 | ||
46 | #include "utfebcdic.h" | |
fd7cb289 | 47 | |
d06134e5 | 48 | #else /* ! EBCDIC */ |
73c4f7a1 GS |
49 | START_EXTERN_C |
50 | ||
a0ed51b3 | 51 | #ifdef DOINIT |
6f06b55f | 52 | EXTCONST unsigned char PL_utf8skip[] = { |
a0ed51b3 LW |
53 | 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */ |
54 | 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */ | |
55 | 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */ | |
56 | 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */ | |
57 | 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus */ | |
58 | 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus */ | |
59 | 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, /* scripts */ | |
3c77ea2b GS |
60 | 3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4,5,5,5,5,6,6, /* cjk etc. */ |
61 | 7,13, /* Perl extended (not UTF-8). Up to 72bit allowed (64-bit + reserved). */ | |
a0ed51b3 LW |
62 | }; |
63 | #else | |
6f06b55f | 64 | EXTCONST unsigned char PL_utf8skip[]; |
a0ed51b3 LW |
65 | #endif |
66 | ||
73c4f7a1 | 67 | END_EXTERN_C |
7e2040f0 | 68 | |
1d72bdf6 NIS |
69 | /* Native character to iso-8859-1 */ |
70 | #define NATIVE_TO_ASCII(ch) (ch) | |
71 | #define ASCII_TO_NATIVE(ch) (ch) | |
72 | /* Transform after encoding */ | |
73 | #define NATIVE_TO_UTF(ch) (ch) | |
74 | #define UTF_TO_NATIVE(ch) (ch) | |
75 | /* Transforms in wide UV chars */ | |
76 | #define UNI_TO_NATIVE(ch) (ch) | |
77 | #define NATIVE_TO_UNI(ch) (ch) | |
78 | /* Transforms in invariant space */ | |
79 | #define NATIVE_TO_NEED(enc,ch) (ch) | |
80 | #define ASCII_TO_NEED(enc,ch) (ch) | |
d7578b48 | 81 | |
d06134e5 | 82 | /* As there are no translations, avoid the function wrapper */ |
1754c1a1 NC |
83 | #define utf8n_to_uvchr utf8n_to_uvuni |
84 | #define uvchr_to_utf8 uvuni_to_utf8 | |
2b9d42f0 | 85 | |
877d9f0d | 86 | /* |
9041c2e3 | 87 | |
8c007b5a | 88 | The following table is from Unicode 3.2. |
877d9f0d JH |
89 | |
90 | Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte | |
91 | ||
375122d7 | 92 | U+0000..U+007F 00..7F |
e1b711da | 93 | U+0080..U+07FF * C2..DF 80..BF |
37e2e78e | 94 | U+0800..U+0FFF E0 * A0..BF 80..BF |
375122d7 | 95 | U+1000..U+CFFF E1..EC 80..BF 80..BF |
e1b711da | 96 | U+D000..U+D7FF ED 80..9F 80..BF |
37e2e78e | 97 | U+D800..U+DFFF +++++++ utf16 surrogates, not legal utf8 +++++++ |
375122d7 | 98 | U+E000..U+FFFF EE..EF 80..BF 80..BF |
37e2e78e | 99 | U+10000..U+3FFFF F0 * 90..BF 80..BF 80..BF |
877d9f0d JH |
100 | U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF |
101 | U+100000..U+10FFFF F4 80..8F 80..BF 80..BF | |
102 | ||
e1b711da | 103 | Note the gaps before several of the byte entries above marked by '*'. These are |
37e2e78e KW |
104 | caused by legal UTF-8 avoiding non-shortest encodings: it is technically |
105 | possible to UTF-8-encode a single code point in different ways, but that is | |
106 | explicitly forbidden, and the shortest possible encoding should always be used | |
107 | (and that is what Perl does). | |
8c007b5a | 108 | |
877d9f0d JH |
109 | */ |
110 | ||
8c007b5a JH |
111 | /* |
112 | Another way to look at it, as bits: | |
113 | ||
114 | Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte | |
115 | ||
116 | 0aaaaaaa 0aaaaaaa | |
117 | 00000bbbbbaaaaaa 110bbbbb 10aaaaaa | |
118 | ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa | |
119 | 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa | |
120 | ||
121 | As you can see, the continuation bytes all begin with C<10>, and the | |
e1b711da | 122 | leading bits of the start byte tell how many bytes there are in the |
8c007b5a JH |
123 | encoded character. |
124 | ||
65ab9279 TC |
125 | Perl's extended UTF-8 means we can have start bytes up to FF. |
126 | ||
8c007b5a JH |
127 | */ |
128 | ||
c4d5f83a | 129 | #define UNI_IS_INVARIANT(c) (((UV)c) < 0x80) |
d06134e5 KW |
130 | /* Note that C0 and C1 are invalid in legal UTF8, so the lower bound of the |
131 | * below might ought to be C2 */ | |
65ab9279 | 132 | #define UTF8_IS_START(c) (((U8)c) >= 0xc0) |
c512ce4f JH |
133 | #define UTF8_IS_CONTINUATION(c) (((U8)c) >= 0x80 && (((U8)c) <= 0xbf)) |
134 | #define UTF8_IS_CONTINUED(c) (((U8)c) & 0x80) | |
db42d148 | 135 | #define UTF8_IS_DOWNGRADEABLE_START(c) (((U8)c & 0xfc) == 0xc0) |
8850bf83 | 136 | |
22901f30 RGS |
137 | #define UTF_START_MARK(len) (((len) > 7) ? 0xFF : (0xFE << (7-(len)))) |
138 | #define UTF_START_MASK(len) (((len) >= 7) ? 0x00 : (0x1F >> ((len)-2))) | |
1d72bdf6 NIS |
139 | |
140 | #define UTF_CONTINUATION_MARK 0x80 | |
141 | #define UTF_ACCUMULATION_SHIFT 6 | |
142 | #define UTF_CONTINUATION_MASK ((U8)0x3f) | |
c512ce4f | 143 | |
1d68d6cd | 144 | #ifdef HAS_QUAD |
5bbb0b5a | 145 | #define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \ |
1d68d6cd SC |
146 | (uv) < 0x800 ? 2 : \ |
147 | (uv) < 0x10000 ? 3 : \ | |
148 | (uv) < 0x200000 ? 4 : \ | |
149 | (uv) < 0x4000000 ? 5 : \ | |
150 | (uv) < 0x80000000 ? 6 : \ | |
9041c2e3 | 151 | (uv) < UTF8_QUAD_MAX ? 7 : 13 ) |
1d68d6cd SC |
152 | #else |
153 | /* No, I'm not even going to *TRY* putting #ifdef inside a #define */ | |
5bbb0b5a | 154 | #define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \ |
1d68d6cd SC |
155 | (uv) < 0x800 ? 2 : \ |
156 | (uv) < 0x10000 ? 3 : \ | |
157 | (uv) < 0x200000 ? 4 : \ | |
158 | (uv) < 0x4000000 ? 5 : \ | |
159 | (uv) < 0x80000000 ? 6 : 7 ) | |
160 | #endif | |
161 | ||
d06134e5 KW |
162 | #endif /* EBCDIC vs ASCII */ |
163 | ||
164 | /* Rest of these are attributes of Unicode and perl's internals rather than the | |
165 | * encoding, or happen to be the same in both ASCII and EBCDIC (at least at | |
166 | * this level; the macros that some of these call may have different | |
167 | * definitions in the two encodings */ | |
168 | ||
169 | #define NATIVE8_TO_UNI(ch) NATIVE_TO_ASCII(ch) /* a clearer synonym */ | |
170 | ||
171 | #define UTF8_ACCUMULATE(old, new) (((old) << UTF_ACCUMULATION_SHIFT) | (((U8)new) & UTF_CONTINUATION_MASK)) | |
172 | ||
2950f2a7 KW |
173 | /* Convert a two (not one) byte utf8 character to a unicode code point value. |
174 | * Needs just one iteration of accumulate. Should not be used unless it is | |
175 | * known that the two bytes are legal: 1) two-byte start, and 2) continuation. | |
176 | * Note that the result can be larger than 255 if the input character is not | |
177 | * downgradable */ | |
178 | #define TWO_BYTE_UTF8_TO_UNI(HI, LO) \ | |
179 | UTF8_ACCUMULATE((NATIVE_TO_UTF(HI) & UTF_START_MASK(2)), \ | |
180 | NATIVE_TO_UTF(LO)) | |
181 | ||
d06134e5 KW |
182 | #define UTF8SKIP(s) PL_utf8skip[*(const U8*)(s)] |
183 | ||
184 | #define UTF8_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE_TO_UTF(c)) | |
185 | #define NATIVE_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE8_TO_UNI(c)) | |
186 | ||
187 | #define MAX_PORTABLE_UTF8_TWO_BYTE 0x3FF /* constrained by EBCDIC */ | |
188 | ||
189 | /* The macros in the next sets are used to generate the two utf8 or utfebcdic | |
190 | * bytes from an ordinal that is known to fit into two bytes; it must be less | |
191 | * than 0x3FF to work across both encodings. */ | |
192 | /* Nocast allows these to be used in the case label of a switch statement */ | |
0072151e | 193 | #define UTF8_TWO_BYTE_HI_nocast(c) UTF_TO_NATIVE(((c) >> UTF_ACCUMULATION_SHIFT) | (0xFF & UTF_START_MARK(2))) |
14c28784 | 194 | #define UTF8_TWO_BYTE_LO_nocast(c) UTF_TO_NATIVE(((c) & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK) |
d06134e5 KW |
195 | |
196 | #define UTF8_TWO_BYTE_HI(c) ((U8) (UTF8_TWO_BYTE_HI_nocast(c))) | |
197 | #define UTF8_TWO_BYTE_LO(c) ((U8) (UTF8_TWO_BYTE_LO_nocast(c))) | |
198 | ||
199 | /* This name is used when the source is a single byte */ | |
200 | #define UTF8_EIGHT_BIT_HI(c) UTF8_TWO_BYTE_HI((U8)(c)) | |
201 | #define UTF8_EIGHT_BIT_LO(c) UTF8_TWO_BYTE_LO((U8)(c)) | |
202 | ||
7e2040f0 GS |
203 | /* |
204 | * Note: we try to be careful never to call the isXXX_utf8() functions | |
d06134e5 KW |
205 | * unless we're pretty sure we've seen the beginning of a UTF-8 or UTFEBCDIC |
206 | * character. Otherwise we risk loading in the heavy-duty swash_init and | |
207 | * swash_fetch routines unnecessarily. | |
7e2040f0 | 208 | */ |
d06134e5 | 209 | #define isIDFIRST_lazy_if(p,c) ((IN_BYTES || (!c || ! UTF8_IS_START(*((const U8*)p)))) \ |
7e2040f0 | 210 | ? isIDFIRST(*(p)) \ |
e1ec3a88 | 211 | : isIDFIRST_utf8((const U8*)p)) |
d06134e5 | 212 | #define isALNUM_lazy_if(p,c) ((IN_BYTES || (!c || ! UTF8_IS_START(*((const U8*)p)))) \ |
7e2040f0 | 213 | ? isALNUM(*(p)) \ |
e1ec3a88 | 214 | : isALNUM_utf8((const U8*)p)) |
1d72bdf6 | 215 | |
7e2040f0 GS |
216 | #define isIDFIRST_lazy(p) isIDFIRST_lazy_if(p,1) |
217 | #define isALNUM_lazy(p) isALNUM_lazy_if(p,1) | |
3bd709b1 | 218 | |
89ebb4a3 JH |
219 | #define UTF8_MAXBYTES 13 |
220 | /* How wide can a single UTF-8 encoded character become in bytes. | |
221 | * NOTE: Strictly speaking Perl's UTF-8 should not be called UTF-8 | |
222 | * since UTF-8 is an encoding of Unicode and given Unicode's current | |
223 | * upper limit only four bytes is possible. Perl thinks of UTF-8 | |
224 | * as a way to encode non-negative integers in a binary format. */ | |
225 | #define UTF8_MAXLEN UTF8_MAXBYTES | |
226 | ||
89ebb4a3 JH |
227 | /* The maximum number of UTF-8 bytes a single Unicode character can |
228 | * uppercase/lowercase/fold into; this number depends on the Unicode | |
229 | * version. An example of maximal expansion is the U+03B0 which | |
230 | * uppercases to U+03C5 U+0308 U+0301. The Unicode databases that | |
d06134e5 | 231 | * tell these things are UnicodeData.txt, CaseFolding.txt, and |
88d45ddf KW |
232 | * SpecialCasing.txt. The value is 6 for strict Unicode characters, but it has |
233 | * to be as big as Perl allows for a single character */ | |
234 | #define UTF8_MAXBYTES_CASE UTF8_MAXBYTES | |
3bd709b1 | 235 | |
8cb75cc8 KW |
236 | /* A Unicode character can fold to up to 3 characters */ |
237 | #define UTF8_MAX_FOLD_CHAR_EXPAND 3 | |
238 | ||
a98fe34d | 239 | #define IN_BYTES (CopHINTS_get(PL_curcop) & HINT_BYTES) |
0064a8a9 | 240 | #define DO_UTF8(sv) (SvUTF8(sv) && !IN_BYTES) |
1863b879 | 241 | #define IN_UNI_8_BIT ( (CopHINTS_get(PL_curcop) & HINT_UNI_8_BIT) \ |
61fc5122 | 242 | && ! IN_LOCALE_RUNTIME && ! IN_BYTES) |
1d72bdf6 | 243 | |
c76687c5 KW |
244 | #define UTF8_ALLOW_EMPTY 0x0001 /* Allow a zero length string */ |
245 | ||
246 | /* Allow first byte to be a continuation byte */ | |
1d72bdf6 | 247 | #define UTF8_ALLOW_CONTINUATION 0x0002 |
c76687c5 KW |
248 | |
249 | /* Allow second... bytes to be non-continuation bytes */ | |
1d72bdf6 | 250 | #define UTF8_ALLOW_NON_CONTINUATION 0x0004 |
949cf498 KW |
251 | |
252 | /* expecting more bytes than were available in the string */ | |
253 | #define UTF8_ALLOW_SHORT 0x0008 | |
254 | ||
255 | /* Overlong sequence; i.e., the code point can be specified in fewer bytes. */ | |
256 | #define UTF8_ALLOW_LONG 0x0010 | |
257 | ||
258 | #define UTF8_DISALLOW_SURROGATE 0x0020 /* Unicode surrogates */ | |
259 | #define UTF8_WARN_SURROGATE 0x0040 | |
260 | ||
261 | #define UTF8_DISALLOW_NONCHAR 0x0080 /* Unicode non-character */ | |
262 | #define UTF8_WARN_NONCHAR 0x0100 /* code points */ | |
263 | ||
264 | #define UTF8_DISALLOW_SUPER 0x0200 /* Super-set of Unicode: code */ | |
265 | #define UTF8_WARN_SUPER 0x0400 /* points above the legal max */ | |
266 | ||
267 | /* Code points which never were part of the original UTF-8 standard, the first | |
268 | * byte of which is a FE or FF on ASCII platforms. */ | |
269 | #define UTF8_DISALLOW_FE_FF 0x0800 | |
270 | #define UTF8_WARN_FE_FF 0x1000 | |
271 | ||
272 | #define UTF8_CHECK_ONLY 0x2000 | |
273 | ||
274 | /* For backwards source compatibility. They do nothing, as the default now | |
275 | * includes what they used to mean. The first one's meaning was to allow the | |
276 | * just the single non-character 0xFFFF */ | |
277 | #define UTF8_ALLOW_FFFF 0 | |
278 | #define UTF8_ALLOW_SURROGATE 0 | |
279 | ||
33d9abfb | 280 | #define UTF8_DISALLOW_ILLEGAL_INTERCHANGE (UTF8_DISALLOW_SUPER|UTF8_DISALLOW_NONCHAR|UTF8_DISALLOW_SURROGATE|UTF8_DISALLOW_FE_FF) |
949cf498 | 281 | #define UTF8_WARN_ILLEGAL_INTERCHANGE \ |
33d9abfb | 282 | (UTF8_WARN_SUPER|UTF8_WARN_NONCHAR|UTF8_WARN_SURROGATE|UTF8_WARN_FE_FF) |
949cf498 KW |
283 | #define UTF8_ALLOW_ANY \ |
284 | (~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE|UTF8_WARN_ILLEGAL_INTERCHANGE)) | |
285 | #define UTF8_ALLOW_ANYUV \ | |
286 | (UTF8_ALLOW_EMPTY \ | |
287 | & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE|UTF8_WARN_ILLEGAL_INTERCHANGE)) | |
9f7f3913 TS |
288 | #define UTF8_ALLOW_DEFAULT (ckWARN(WARN_UTF8) ? 0 : \ |
289 | UTF8_ALLOW_ANYUV) | |
1d72bdf6 | 290 | |
7131f24d KW |
291 | /* Surrogates, non-character code points and above-Unicode code points are |
292 | * problematic in some contexts. This allows code that needs to check for | |
293 | * those to to quickly exclude the vast majority of code points it will | |
294 | * encounter */ | |
295 | #ifdef EBCDIC | |
296 | # define UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE UTF_TO_NATIVE(0xF1) | |
297 | #else | |
298 | # define UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE 0xED | |
299 | #endif | |
300 | ||
301 | /* ASCII EBCDIC I8 | |
302 | * U+D7FF: \xED\x9F\xBF \xF1\xB5\xBF\xBF last before surrogates | |
303 | * U+D800: \xED\xA0\x80 \xF1\xB6\xA0\xA0 1st surrogate | |
304 | * U+DFFF: \xED\xBF\xBF \xF1\xB7\xBF\xBF final surrogate | |
305 | * U+E000: \xEE\x80\x80 \xF1\xB8\xA0\xA0 next after surrogates | |
306 | */ | |
307 | #ifdef EBCDIC /* Both versions assume well-formed UTF8 */ | |
308 | # define UTF8_IS_SURROGATE(s) (*(s) == UTF_TO_NATIVE(0xF1) \ | |
309 | && (*((s) +1) == UTF_TO_NATIVE(0xB6)) || *((s) + 1) == UTF_TO_NATIVE(0xB7)) | |
310 | #else | |
311 | # define UTF8_IS_SURROGATE(s) (*(s) == 0xED && *((s) + 1) >= 0xA0) | |
312 | #endif | |
313 | ||
314 | /* ASCII EBCDIC I8 | |
315 | * U+10FFFF: \xF4\x8F\xBF\xBF \xF9\xA1\xBF\xBF\xBF max legal Unicode | |
316 | * U+110000: \xF4\x90\x80\x80 \xF9\xA2\xA0\xA0\xA0 | |
317 | * U+110001: \xF4\x90\x80\x81 \xF9\xA2\xA0\xA0\xA1 | |
318 | */ | |
319 | #ifdef EBCDIC /* Both versions assume well-formed UTF8 */ | |
320 | # define UTF8_IS_SUPER(s) (*(s) >= UTF_TO_NATIVE(0xF9) \ | |
321 | && (*(s) > UTF_TO_NATIVE(0xF9) || (*((s) + 1) >= UTF_TO_NATIVE(0xA2)))) | |
322 | #else | |
323 | # define UTF8_IS_SUPER(s) (*(s) >= 0xF4 \ | |
324 | && (*(s) > 0xF4 || (*((s) + 1) >= 0x90))) | |
325 | #endif | |
326 | ||
327 | /* ASCII EBCDIC I8 | |
328 | * U+FDCF: \xEF\xB7\x8F \xF1\xBF\xAE\xAF last before non-char block | |
329 | * U+FDD0: \xEF\xB7\x90 \xF1\xBF\xAE\xB0 first non-char in block | |
330 | * U+FDEF: \xEF\xB7\xAF \xF1\xBF\xAF\xAF last non-char in block | |
331 | * U+FDF0: \xEF\xB7\xB0 \xF1\xBF\xAF\xB0 first after non-char block | |
332 | * U+FFFF: \xEF\xBF\xBF \xF1\xBF\xBF\xBF | |
333 | * U+1FFFF: \xF0\x9F\xBF\xBF \xF3\xBF\xBF\xBF | |
334 | * U+2FFFF: \xF0\xAF\xBF\xBF \xF5\xBF\xBF\xBF | |
335 | * U+3FFFF: \xF0\xBF\xBF\xBF \xF7\xBF\xBF\xBF | |
336 | * U+4FFFF: \xF1\x8F\xBF\xBF \xF8\xA9\xBF\xBF\xBF | |
337 | * U+5FFFF: \xF1\x9F\xBF\xBF \xF8\xAB\xBF\xBF\xBF | |
338 | * U+6FFFF: \xF1\xAF\xBF\xBF \xF8\xAD\xBF\xBF\xBF | |
339 | * U+7FFFF: \xF1\xBF\xBF\xBF \xF8\xAF\xBF\xBF\xBF | |
340 | * U+8FFFF: \xF2\x8F\xBF\xBF \xF8\xB1\xBF\xBF\xBF | |
341 | * U+9FFFF: \xF2\x9F\xBF\xBF \xF8\xB3\xBF\xBF\xBF | |
342 | * U+AFFFF: \xF2\xAF\xBF\xBF \xF8\xB5\xBF\xBF\xBF | |
343 | * U+BFFFF: \xF2\xBF\xBF\xBF \xF8\xB7\xBF\xBF\xBF | |
344 | * U+CFFFF: \xF3\x8F\xBF\xBF \xF8\xB9\xBF\xBF\xBF | |
345 | * U+DFFFF: \xF3\x9F\xBF\xBF \xF8\xBB\xBF\xBF\xBF | |
346 | * U+EFFFF: \xF3\xAF\xBF\xBF \xF8\xBD\xBF\xBF\xBF | |
347 | * U+FFFFF: \xF3\xBF\xBF\xBF \xF8\xBF\xBF\xBF\xBF | |
348 | * U+10FFFF: \xF4\x8F\xBF\xBF \xF9\xA1\xBF\xBF\xBF | |
349 | */ | |
350 | #define UTF8_IS_NONCHAR_(s) ( \ | |
351 | *(s) >= UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE \ | |
352 | && ! UTF8_IS_SUPER(s) \ | |
353 | && UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_FIRST_PROBLEMATIC(s) \ | |
354 | ||
355 | #ifdef EBCDIC /* Both versions assume well-formed UTF8 */ | |
356 | # define UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s) \ | |
357 | ((*(s) == UTF_TO_NATIVE(0xF1) \ | |
358 | && (*((s) + 1) == UTF_TO_NATIVE(0xBF) \ | |
359 | && ((*((s) + 2) == UTF_TO_NATIVE(0xAE) \ | |
360 | && *((s) + 3) >= UTF_TO_NATIVE(0xB0)) \ | |
361 | || (*((s) + 2) == UTF_TO_NATIVE(0xAF) \ | |
362 | && *((s) + 3) <= UTF_TO_NATIVE(0xAF))))) \ | |
363 | || (UTF8SKIP(*(s)) > 3 \ | |
364 | /* (These were all derived by inspection and experimentation with an */ \ | |
365 | /* editor) The next line checks the next to final byte in the char */ \ | |
366 | && *((s) + UTF8SKIP(*(s)) - 2) == UTF_TO_NATIVE(0xBF) \ | |
367 | && *((s) + UTF8SKIP(*(s)) - 3) == UTF_TO_NATIVE(0xBF) \ | |
368 | && (NATIVE_TO_UTF(*((s) + UTF8SKIP(*(s)) - 4)) & 0x81) == 0x81 \ | |
369 | && (NATIVE_TO_UTF(*((s) + UTF8SKIP(*(s)) - 1)) & 0xBE) == 0XBE)) | |
370 | #else | |
371 | # define UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s) \ | |
372 | ((*(s) == 0xEF \ | |
373 | && ((*((s) + 1) == 0xB7 && (*((s) + 2) >= 0x90 && (*((s) + 2) <= 0xAF)))\ | |
374 | /* Gets U+FFF[EF] */ \ | |
375 | || (*((s) + 1) == 0xBF && ((*((s) + 2) & 0xBE) == 0xBE)))) \ | |
376 | || ((*((s) + 2) == 0xBF \ | |
377 | && (*((s) + 3) & 0xBE) == 0xBE \ | |
378 | /* Excludes things like U+10FFE = \xF0\x90\xBF\xBE */ \ | |
379 | && (*((s) + 1) & 0x8F) == 0x8F))) | |
380 | #endif | |
381 | ||
c867b360 JH |
382 | #define UNICODE_SURROGATE_FIRST 0xD800 |
383 | #define UNICODE_SURROGATE_LAST 0xDFFF | |
384 | #define UNICODE_REPLACEMENT 0xFFFD | |
385 | #define UNICODE_BYTE_ORDER_MARK 0xFEFF | |
1d72bdf6 | 386 | |
b851fbc1 | 387 | /* Though our UTF-8 encoding can go beyond this, |
c76687c5 | 388 | * let's be conservative and do as Unicode says. */ |
b851fbc1 JH |
389 | #define PERL_UNICODE_MAX 0x10FFFF |
390 | ||
949cf498 KW |
391 | #define UNICODE_WARN_SURROGATE 0x0001 /* UTF-16 surrogates */ |
392 | #define UNICODE_WARN_NONCHAR 0x0002 /* Non-char code points */ | |
393 | #define UNICODE_WARN_SUPER 0x0004 /* Above 0x10FFFF */ | |
394 | #define UNICODE_WARN_FE_FF 0x0008 /* Above 0x10FFFF */ | |
395 | #define UNICODE_DISALLOW_SURROGATE 0x0010 | |
396 | #define UNICODE_DISALLOW_NONCHAR 0x0020 | |
397 | #define UNICODE_DISALLOW_SUPER 0x0040 | |
398 | #define UNICODE_DISALLOW_FE_FF 0x0080 | |
399 | #define UNICODE_WARN_ILLEGAL_INTERCHANGE \ | |
400 | (UNICODE_WARN_SURROGATE|UNICODE_WARN_NONCHAR|UNICODE_WARN_SUPER) | |
401 | #define UNICODE_DISALLOW_ILLEGAL_INTERCHANGE \ | |
402 | (UNICODE_DISALLOW_SURROGATE|UNICODE_DISALLOW_NONCHAR|UNICODE_DISALLOW_SUPER) | |
403 | ||
404 | /* For backward source compatibility, as are now the default */ | |
405 | #define UNICODE_ALLOW_SURROGATE 0 | |
406 | #define UNICODE_ALLOW_SUPER 0 | |
407 | #define UNICODE_ALLOW_ANY 0 | |
b851fbc1 | 408 | |
1d72bdf6 NIS |
409 | #define UNICODE_IS_SURROGATE(c) ((c) >= UNICODE_SURROGATE_FIRST && \ |
410 | (c) <= UNICODE_SURROGATE_LAST) | |
a10ec373 | 411 | #define UNICODE_IS_REPLACEMENT(c) ((c) == UNICODE_REPLACEMENT) |
872c91ae | 412 | #define UNICODE_IS_BYTE_ORDER_MARK(c) ((c) == UNICODE_BYTE_ORDER_MARK) |
7131f24d KW |
413 | #define UNICODE_IS_NONCHAR(c) ((c >= 0xFDD0 && c <= 0xFDEF) \ |
414 | /* The other noncharacters end in FFFE or FFFF, which \ | |
415 | * the mask below catches both of, but beyond the last \ | |
416 | * official unicode code point, they aren't \ | |
417 | * noncharacters, since those aren't Unicode \ | |
418 | * characters at all */ \ | |
419 | || ((((c & 0xFFFE) == 0xFFFE)) && ! UNICODE_IS_SUPER(c))) | |
420 | #define UNICODE_IS_SUPER(c) ((c) > PERL_UNICODE_MAX) | |
421 | #define UNICODE_IS_FE_FF(c) ((c) > 0x7FFFFFFF) | |
1d72bdf6 | 422 | |
f067b878 NA |
423 | #ifdef HAS_QUAD |
424 | # define UTF8_QUAD_MAX UINT64_C(0x1000000000) | |
425 | #endif | |
3bd709b1 | 426 | |
09091399 JH |
427 | #define UNICODE_GREEK_CAPITAL_LETTER_SIGMA 0x03A3 |
428 | #define UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA 0x03C2 | |
429 | #define UNICODE_GREEK_SMALL_LETTER_SIGMA 0x03C3 | |
9dcbe121 | 430 | #define GREEK_SMALL_LETTER_MU 0x03BC |
78a0d3cc KW |
431 | #define GREEK_CAPITAL_LETTER_MU 0x039C /* Upper and title case of MICRON */ |
432 | #define LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS 0x0178 /* Also is title case */ | |
97298f37 | 433 | #define LATIN_CAPITAL_LETTER_SHARP_S 0x1E9E |
09091399 | 434 | |
9e55ce06 | 435 | #define UNI_DISPLAY_ISPRINT 0x0001 |
c728cb41 JH |
436 | #define UNI_DISPLAY_BACKSLASH 0x0002 |
437 | #define UNI_DISPLAY_QQ (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH) | |
438 | #define UNI_DISPLAY_REGEX (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH) | |
9e55ce06 | 439 | |
5cd46e1f KW |
440 | #ifndef EBCDIC |
441 | # define LATIN_SMALL_LETTER_SHARP_S 0x00DF | |
442 | # define LATIN_SMALL_LETTER_Y_WITH_DIAERESIS 0x00FF | |
443 | # define MICRO_SIGN 0x00B5 | |
f508a607 KW |
444 | # define LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE 0x00C5 |
445 | # define LATIN_SMALL_LETTER_A_WITH_RING_ABOVE 0x00E5 | |
5cd46e1f KW |
446 | #endif |
447 | ||
448 | #define ANYOF_FOLD_SHARP_S(node, input, end) \ | |
449 | (ANYOF_BITMAP_TEST(node, LATIN_SMALL_LETTER_SHARP_S) && \ | |
137165a6 | 450 | (ANYOF_NONBITMAP(node)) && \ |
39065660 | 451 | (ANYOF_FLAGS(node) & ANYOF_LOC_NONBITMAP_FOLD) && \ |
07b6858f JH |
452 | ((end) > (input) + 1) && \ |
453 | toLOWER((input)[0]) == 's' && \ | |
454 | toLOWER((input)[1]) == 's') | |
ebc501f0 | 455 | #define SHARP_S_SKIP 2 |
3b0fc154 | 456 | |
77263263 TS |
457 | #ifdef EBCDIC |
458 | /* IS_UTF8_CHAR() is not ported to EBCDIC */ | |
459 | #else | |
3b0fc154 JH |
460 | #define IS_UTF8_CHAR_1(p) \ |
461 | ((p)[0] <= 0x7F) | |
462 | #define IS_UTF8_CHAR_2(p) \ | |
463 | ((p)[0] >= 0xC2 && (p)[0] <= 0xDF && \ | |
464 | (p)[1] >= 0x80 && (p)[1] <= 0xBF) | |
465 | #define IS_UTF8_CHAR_3a(p) \ | |
466 | ((p)[0] == 0xE0 && \ | |
467 | (p)[1] >= 0xA0 && (p)[1] <= 0xBF && \ | |
468 | (p)[2] >= 0x80 && (p)[2] <= 0xBF) | |
469 | #define IS_UTF8_CHAR_3b(p) \ | |
470 | ((p)[0] >= 0xE1 && (p)[0] <= 0xEC && \ | |
471 | (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ | |
472 | (p)[2] >= 0x80 && (p)[2] <= 0xBF) | |
473 | #define IS_UTF8_CHAR_3c(p) \ | |
474 | ((p)[0] == 0xED && \ | |
475 | (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ | |
476 | (p)[2] >= 0x80 && (p)[2] <= 0xBF) | |
477 | /* In IS_UTF8_CHAR_3c(p) one could use | |
478 | * (p)[1] >= 0x80 && (p)[1] <= 0x9F | |
479 | * if one wanted to exclude surrogates. */ | |
480 | #define IS_UTF8_CHAR_3d(p) \ | |
481 | ((p)[0] >= 0xEE && (p)[0] <= 0xEF && \ | |
482 | (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ | |
483 | (p)[2] >= 0x80 && (p)[2] <= 0xBF) | |
484 | #define IS_UTF8_CHAR_4a(p) \ | |
485 | ((p)[0] == 0xF0 && \ | |
486 | (p)[1] >= 0x90 && (p)[1] <= 0xBF && \ | |
487 | (p)[2] >= 0x80 && (p)[2] <= 0xBF && \ | |
488 | (p)[3] >= 0x80 && (p)[3] <= 0xBF) | |
489 | #define IS_UTF8_CHAR_4b(p) \ | |
490 | ((p)[0] >= 0xF1 && (p)[0] <= 0xF3 && \ | |
491 | (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ | |
492 | (p)[2] >= 0x80 && (p)[2] <= 0xBF && \ | |
493 | (p)[3] >= 0x80 && (p)[3] <= 0xBF) | |
494 | /* In IS_UTF8_CHAR_4c(p) one could use | |
495 | * (p)[0] == 0xF4 | |
496 | * if one wanted to stop at the Unicode limit U+10FFFF. | |
497 | * The 0xF7 allows us to go to 0x1fffff (0x200000 would | |
498 | * require five bytes). Not doing any further code points | |
499 | * since that is not needed (and that would not be strict | |
500 | * UTF-8, anyway). The "slow path" in Perl_is_utf8_char() | |
501 | * will take care of the "extended UTF-8". */ | |
502 | #define IS_UTF8_CHAR_4c(p) \ | |
a18d6e6e | 503 | ((p)[0] >= 0xF4 && (p)[0] <= 0xF7 && \ |
3b0fc154 JH |
504 | (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ |
505 | (p)[2] >= 0x80 && (p)[2] <= 0xBF && \ | |
506 | (p)[3] >= 0x80 && (p)[3] <= 0xBF) | |
507 | ||
508 | #define IS_UTF8_CHAR_3(p) \ | |
509 | (IS_UTF8_CHAR_3a(p) || \ | |
510 | IS_UTF8_CHAR_3b(p) || \ | |
511 | IS_UTF8_CHAR_3c(p) || \ | |
512 | IS_UTF8_CHAR_3d(p)) | |
513 | #define IS_UTF8_CHAR_4(p) \ | |
514 | (IS_UTF8_CHAR_4a(p) || \ | |
515 | IS_UTF8_CHAR_4b(p) || \ | |
516 | IS_UTF8_CHAR_4c(p)) | |
517 | ||
518 | /* IS_UTF8_CHAR(p) is strictly speaking wrong (not UTF-8) because it | |
519 | * (1) allows UTF-8 encoded UTF-16 surrogates | |
520 | * (2) it allows code points past U+10FFFF. | |
521 | * The Perl_is_utf8_char() full "slow" code will handle the Perl | |
522 | * "extended UTF-8". */ | |
523 | #define IS_UTF8_CHAR(p, n) \ | |
524 | ((n) == 1 ? IS_UTF8_CHAR_1(p) : \ | |
525 | (n) == 2 ? IS_UTF8_CHAR_2(p) : \ | |
526 | (n) == 3 ? IS_UTF8_CHAR_3(p) : \ | |
527 | (n) == 4 ? IS_UTF8_CHAR_4(p) : 0) | |
528 | ||
768c67ee JH |
529 | #define IS_UTF8_CHAR_FAST(n) ((n) <= 4) |
530 | ||
77263263 | 531 | #endif /* IS_UTF8_CHAR() for UTF-8 */ |
e9a8c099 MHM |
532 | |
533 | /* | |
534 | * Local variables: | |
535 | * c-indentation-style: bsd | |
536 | * c-basic-offset: 4 | |
537 | * indent-tabs-mode: t | |
538 | * End: | |
539 | * | |
540 | * ex: set ts=8 sts=4 sw=4 noet: | |
541 | */ |