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 | ||
051a06d4 KW |
19 | /* For to_utf8_fold_flags, q.v. */ |
20 | #define FOLD_FLAGS_LOCALE 0x1 | |
21 | #define FOLD_FLAGS_FULL 0x2 | |
22 | ||
36bb2ab6 | 23 | #define to_uni_fold(c, p, lenp) _to_uni_fold_flags(c, p, lenp, 1) |
051a06d4 KW |
24 | #define to_utf8_fold(c, p, lenp) _to_utf8_fold_flags(c, p, lenp, \ |
25 | FOLD_FLAGS_FULL, NULL) | |
26 | #define to_utf8_lower(a,b,c) _to_utf8_lower_flags(a,b,c,0, NULL) | |
27 | #define to_utf8_upper(a,b,c) _to_utf8_upper_flags(a,b,c,0, NULL) | |
28 | #define to_utf8_title(a,b,c) _to_utf8_title_flags(a,b,c,0, NULL) | |
36bb2ab6 | 29 | |
fd7cb289 RGS |
30 | /* Source backward compatibility. */ |
31 | #define uvuni_to_utf8(d, uv) uvuni_to_utf8_flags(d, uv, 0) | |
32 | #define is_utf8_string_loc(s, len, ep) is_utf8_string_loclen(s, len, ep, 0) | |
33 | ||
eda9cac1 KW |
34 | #define foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \ |
35 | foldEQ_utf8_flags(s1, pe1, l1, u1, s2, pe2, l2, u2, 0) | |
a33c29bc | 36 | #define FOLDEQ_UTF8_NOMIX_ASCII (1 << 0) |
5e64d0fa | 37 | #define FOLDEQ_UTF8_LOCALE (1 << 1) |
18f762c3 KW |
38 | #define FOLDEQ_S1_ALREADY_FOLDED (1 << 2) |
39 | #define FOLDEQ_S2_ALREADY_FOLDED (1 << 3) | |
a33c29bc | 40 | |
e6226b18 KW |
41 | /* |
42 | =for apidoc ibcmp_utf8 | |
43 | ||
44 | This is a synonym for (! foldEQ_utf8()) | |
45 | ||
46 | =cut | |
47 | */ | |
48 | #define ibcmp_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \ | |
49 | cBOOL(! foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2)) | |
50 | ||
1d72bdf6 NIS |
51 | #ifdef EBCDIC |
52 | /* The equivalent of these macros but implementing UTF-EBCDIC | |
53 | are in the following header file: | |
54 | */ | |
55 | ||
56 | #include "utfebcdic.h" | |
fd7cb289 | 57 | |
d06134e5 | 58 | #else /* ! EBCDIC */ |
73c4f7a1 GS |
59 | START_EXTERN_C |
60 | ||
a0ed51b3 | 61 | #ifdef DOINIT |
6f06b55f | 62 | EXTCONST unsigned char PL_utf8skip[] = { |
a0ed51b3 LW |
63 | 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 */ |
64 | 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 */ | |
65 | 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 */ | |
66 | 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 */ | |
67 | 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 */ | |
68 | 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 */ | |
69 | 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 | 70 | 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. */ |
7eb168e2 KW |
71 | 7,13, /* Perl extended (not official UTF-8). Up to 72bit allowed (64-bit + |
72 | reserved). */ | |
a0ed51b3 LW |
73 | }; |
74 | #else | |
6f06b55f | 75 | EXTCONST unsigned char PL_utf8skip[]; |
a0ed51b3 LW |
76 | #endif |
77 | ||
73c4f7a1 | 78 | END_EXTERN_C |
7e2040f0 | 79 | |
1d72bdf6 NIS |
80 | /* Native character to iso-8859-1 */ |
81 | #define NATIVE_TO_ASCII(ch) (ch) | |
82 | #define ASCII_TO_NATIVE(ch) (ch) | |
83 | /* Transform after encoding */ | |
84 | #define NATIVE_TO_UTF(ch) (ch) | |
85 | #define UTF_TO_NATIVE(ch) (ch) | |
86 | /* Transforms in wide UV chars */ | |
87 | #define UNI_TO_NATIVE(ch) (ch) | |
88 | #define NATIVE_TO_UNI(ch) (ch) | |
89 | /* Transforms in invariant space */ | |
90 | #define NATIVE_TO_NEED(enc,ch) (ch) | |
91 | #define ASCII_TO_NEED(enc,ch) (ch) | |
d7578b48 | 92 | |
d06134e5 | 93 | /* As there are no translations, avoid the function wrapper */ |
1754c1a1 NC |
94 | #define utf8n_to_uvchr utf8n_to_uvuni |
95 | #define uvchr_to_utf8 uvuni_to_utf8 | |
2b9d42f0 | 96 | |
877d9f0d | 97 | /* |
9041c2e3 | 98 | |
8c007b5a | 99 | The following table is from Unicode 3.2. |
877d9f0d JH |
100 | |
101 | Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte | |
102 | ||
375122d7 | 103 | U+0000..U+007F 00..7F |
e1b711da | 104 | U+0080..U+07FF * C2..DF 80..BF |
37e2e78e | 105 | U+0800..U+0FFF E0 * A0..BF 80..BF |
375122d7 | 106 | U+1000..U+CFFF E1..EC 80..BF 80..BF |
e1b711da | 107 | U+D000..U+D7FF ED 80..9F 80..BF |
37e2e78e | 108 | U+D800..U+DFFF +++++++ utf16 surrogates, not legal utf8 +++++++ |
375122d7 | 109 | U+E000..U+FFFF EE..EF 80..BF 80..BF |
37e2e78e | 110 | U+10000..U+3FFFF F0 * 90..BF 80..BF 80..BF |
877d9f0d JH |
111 | U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF |
112 | U+100000..U+10FFFF F4 80..8F 80..BF 80..BF | |
113 | ||
e1b711da | 114 | Note the gaps before several of the byte entries above marked by '*'. These are |
37e2e78e KW |
115 | caused by legal UTF-8 avoiding non-shortest encodings: it is technically |
116 | possible to UTF-8-encode a single code point in different ways, but that is | |
117 | explicitly forbidden, and the shortest possible encoding should always be used | |
118 | (and that is what Perl does). | |
8c007b5a | 119 | |
877d9f0d JH |
120 | */ |
121 | ||
8c007b5a JH |
122 | /* |
123 | Another way to look at it, as bits: | |
124 | ||
125 | Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte | |
126 | ||
127 | 0aaaaaaa 0aaaaaaa | |
128 | 00000bbbbbaaaaaa 110bbbbb 10aaaaaa | |
129 | ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa | |
130 | 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa | |
131 | ||
132 | As you can see, the continuation bytes all begin with C<10>, and the | |
e1b711da | 133 | leading bits of the start byte tell how many bytes there are in the |
8c007b5a JH |
134 | encoded character. |
135 | ||
65ab9279 TC |
136 | Perl's extended UTF-8 means we can have start bytes up to FF. |
137 | ||
8c007b5a JH |
138 | */ |
139 | ||
c4d5f83a | 140 | #define UNI_IS_INVARIANT(c) (((UV)c) < 0x80) |
d06134e5 KW |
141 | /* Note that C0 and C1 are invalid in legal UTF8, so the lower bound of the |
142 | * below might ought to be C2 */ | |
65ab9279 | 143 | #define UTF8_IS_START(c) (((U8)c) >= 0xc0) |
c512ce4f JH |
144 | #define UTF8_IS_CONTINUATION(c) (((U8)c) >= 0x80 && (((U8)c) <= 0xbf)) |
145 | #define UTF8_IS_CONTINUED(c) (((U8)c) & 0x80) | |
db42d148 | 146 | #define UTF8_IS_DOWNGRADEABLE_START(c) (((U8)c & 0xfc) == 0xc0) |
8850bf83 | 147 | |
22901f30 RGS |
148 | #define UTF_START_MARK(len) (((len) > 7) ? 0xFF : (0xFE << (7-(len)))) |
149 | #define UTF_START_MASK(len) (((len) >= 7) ? 0x00 : (0x1F >> ((len)-2))) | |
1d72bdf6 NIS |
150 | |
151 | #define UTF_CONTINUATION_MARK 0x80 | |
152 | #define UTF_ACCUMULATION_SHIFT 6 | |
153 | #define UTF_CONTINUATION_MASK ((U8)0x3f) | |
c512ce4f | 154 | |
1d68d6cd | 155 | #ifdef HAS_QUAD |
5bbb0b5a | 156 | #define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \ |
1d68d6cd SC |
157 | (uv) < 0x800 ? 2 : \ |
158 | (uv) < 0x10000 ? 3 : \ | |
159 | (uv) < 0x200000 ? 4 : \ | |
160 | (uv) < 0x4000000 ? 5 : \ | |
161 | (uv) < 0x80000000 ? 6 : \ | |
9041c2e3 | 162 | (uv) < UTF8_QUAD_MAX ? 7 : 13 ) |
1d68d6cd SC |
163 | #else |
164 | /* No, I'm not even going to *TRY* putting #ifdef inside a #define */ | |
5bbb0b5a | 165 | #define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \ |
1d68d6cd SC |
166 | (uv) < 0x800 ? 2 : \ |
167 | (uv) < 0x10000 ? 3 : \ | |
168 | (uv) < 0x200000 ? 4 : \ | |
169 | (uv) < 0x4000000 ? 5 : \ | |
170 | (uv) < 0x80000000 ? 6 : 7 ) | |
171 | #endif | |
172 | ||
d06134e5 KW |
173 | #endif /* EBCDIC vs ASCII */ |
174 | ||
175 | /* Rest of these are attributes of Unicode and perl's internals rather than the | |
176 | * encoding, or happen to be the same in both ASCII and EBCDIC (at least at | |
177 | * this level; the macros that some of these call may have different | |
178 | * definitions in the two encodings */ | |
179 | ||
180 | #define NATIVE8_TO_UNI(ch) NATIVE_TO_ASCII(ch) /* a clearer synonym */ | |
181 | ||
182 | #define UTF8_ACCUMULATE(old, new) (((old) << UTF_ACCUMULATION_SHIFT) | (((U8)new) & UTF_CONTINUATION_MASK)) | |
183 | ||
2950f2a7 KW |
184 | /* Convert a two (not one) byte utf8 character to a unicode code point value. |
185 | * Needs just one iteration of accumulate. Should not be used unless it is | |
186 | * known that the two bytes are legal: 1) two-byte start, and 2) continuation. | |
187 | * Note that the result can be larger than 255 if the input character is not | |
188 | * downgradable */ | |
189 | #define TWO_BYTE_UTF8_TO_UNI(HI, LO) \ | |
190 | UTF8_ACCUMULATE((NATIVE_TO_UTF(HI) & UTF_START_MASK(2)), \ | |
191 | NATIVE_TO_UTF(LO)) | |
192 | ||
d06134e5 KW |
193 | #define UTF8SKIP(s) PL_utf8skip[*(const U8*)(s)] |
194 | ||
195 | #define UTF8_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE_TO_UTF(c)) | |
196 | #define NATIVE_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE8_TO_UNI(c)) | |
197 | ||
198 | #define MAX_PORTABLE_UTF8_TWO_BYTE 0x3FF /* constrained by EBCDIC */ | |
199 | ||
200 | /* The macros in the next sets are used to generate the two utf8 or utfebcdic | |
201 | * bytes from an ordinal that is known to fit into two bytes; it must be less | |
202 | * than 0x3FF to work across both encodings. */ | |
203 | /* Nocast allows these to be used in the case label of a switch statement */ | |
0072151e | 204 | #define UTF8_TWO_BYTE_HI_nocast(c) UTF_TO_NATIVE(((c) >> UTF_ACCUMULATION_SHIFT) | (0xFF & UTF_START_MARK(2))) |
14c28784 | 205 | #define UTF8_TWO_BYTE_LO_nocast(c) UTF_TO_NATIVE(((c) & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK) |
d06134e5 KW |
206 | |
207 | #define UTF8_TWO_BYTE_HI(c) ((U8) (UTF8_TWO_BYTE_HI_nocast(c))) | |
208 | #define UTF8_TWO_BYTE_LO(c) ((U8) (UTF8_TWO_BYTE_LO_nocast(c))) | |
209 | ||
210 | /* This name is used when the source is a single byte */ | |
211 | #define UTF8_EIGHT_BIT_HI(c) UTF8_TWO_BYTE_HI((U8)(c)) | |
212 | #define UTF8_EIGHT_BIT_LO(c) UTF8_TWO_BYTE_LO((U8)(c)) | |
213 | ||
7e2040f0 | 214 | /* |
e3036cf4 | 215 | * 'UTF' is whether or not p is encoded in UTF8. The names 'foo_lazy_if' stem |
20df05f4 KW |
216 | * from an earlier version of these macros in which they didn't call the |
217 | * foo_utf8() macros (i.e. were 'lazy') unless they decided that *p is the | |
218 | * beginning of a utf8 character. Now that foo_utf8() determines that itself, | |
219 | * no need to do it again here | |
7e2040f0 | 220 | */ |
e3036cf4 KW |
221 | #define isIDFIRST_lazy_if(p,UTF) ((IN_BYTES || !UTF ) \ |
222 | ? isIDFIRST(*(p)) \ | |
223 | : isIDFIRST_utf8((const U8*)p)) | |
224 | #define isALNUM_lazy_if(p,UTF) ((IN_BYTES || (!UTF )) \ | |
225 | ? isALNUM(*(p)) \ | |
226 | : isALNUM_utf8((const U8*)p)) | |
1d72bdf6 | 227 | |
7e2040f0 GS |
228 | #define isIDFIRST_lazy(p) isIDFIRST_lazy_if(p,1) |
229 | #define isALNUM_lazy(p) isALNUM_lazy_if(p,1) | |
3bd709b1 | 230 | |
89ebb4a3 JH |
231 | #define UTF8_MAXBYTES 13 |
232 | /* How wide can a single UTF-8 encoded character become in bytes. | |
233 | * NOTE: Strictly speaking Perl's UTF-8 should not be called UTF-8 | |
234 | * since UTF-8 is an encoding of Unicode and given Unicode's current | |
235 | * upper limit only four bytes is possible. Perl thinks of UTF-8 | |
236 | * as a way to encode non-negative integers in a binary format. */ | |
237 | #define UTF8_MAXLEN UTF8_MAXBYTES | |
238 | ||
89ebb4a3 JH |
239 | /* The maximum number of UTF-8 bytes a single Unicode character can |
240 | * uppercase/lowercase/fold into; this number depends on the Unicode | |
241 | * version. An example of maximal expansion is the U+03B0 which | |
242 | * uppercases to U+03C5 U+0308 U+0301. The Unicode databases that | |
d06134e5 | 243 | * tell these things are UnicodeData.txt, CaseFolding.txt, and |
88d45ddf KW |
244 | * SpecialCasing.txt. The value is 6 for strict Unicode characters, but it has |
245 | * to be as big as Perl allows for a single character */ | |
246 | #define UTF8_MAXBYTES_CASE UTF8_MAXBYTES | |
3bd709b1 | 247 | |
8cb75cc8 KW |
248 | /* A Unicode character can fold to up to 3 characters */ |
249 | #define UTF8_MAX_FOLD_CHAR_EXPAND 3 | |
250 | ||
a98fe34d | 251 | #define IN_BYTES (CopHINTS_get(PL_curcop) & HINT_BYTES) |
0064a8a9 | 252 | #define DO_UTF8(sv) (SvUTF8(sv) && !IN_BYTES) |
1863b879 | 253 | #define IN_UNI_8_BIT ( (CopHINTS_get(PL_curcop) & HINT_UNI_8_BIT) \ |
61fc5122 | 254 | && ! IN_LOCALE_RUNTIME && ! IN_BYTES) |
1d72bdf6 | 255 | |
c76687c5 KW |
256 | #define UTF8_ALLOW_EMPTY 0x0001 /* Allow a zero length string */ |
257 | ||
258 | /* Allow first byte to be a continuation byte */ | |
1d72bdf6 | 259 | #define UTF8_ALLOW_CONTINUATION 0x0002 |
c76687c5 KW |
260 | |
261 | /* Allow second... bytes to be non-continuation bytes */ | |
1d72bdf6 | 262 | #define UTF8_ALLOW_NON_CONTINUATION 0x0004 |
949cf498 KW |
263 | |
264 | /* expecting more bytes than were available in the string */ | |
265 | #define UTF8_ALLOW_SHORT 0x0008 | |
266 | ||
267 | /* Overlong sequence; i.e., the code point can be specified in fewer bytes. */ | |
268 | #define UTF8_ALLOW_LONG 0x0010 | |
269 | ||
270 | #define UTF8_DISALLOW_SURROGATE 0x0020 /* Unicode surrogates */ | |
271 | #define UTF8_WARN_SURROGATE 0x0040 | |
272 | ||
273 | #define UTF8_DISALLOW_NONCHAR 0x0080 /* Unicode non-character */ | |
274 | #define UTF8_WARN_NONCHAR 0x0100 /* code points */ | |
275 | ||
276 | #define UTF8_DISALLOW_SUPER 0x0200 /* Super-set of Unicode: code */ | |
277 | #define UTF8_WARN_SUPER 0x0400 /* points above the legal max */ | |
278 | ||
279 | /* Code points which never were part of the original UTF-8 standard, the first | |
280 | * byte of which is a FE or FF on ASCII platforms. */ | |
281 | #define UTF8_DISALLOW_FE_FF 0x0800 | |
282 | #define UTF8_WARN_FE_FF 0x1000 | |
283 | ||
284 | #define UTF8_CHECK_ONLY 0x2000 | |
285 | ||
286 | /* For backwards source compatibility. They do nothing, as the default now | |
287 | * includes what they used to mean. The first one's meaning was to allow the | |
288 | * just the single non-character 0xFFFF */ | |
289 | #define UTF8_ALLOW_FFFF 0 | |
290 | #define UTF8_ALLOW_SURROGATE 0 | |
291 | ||
33d9abfb | 292 | #define UTF8_DISALLOW_ILLEGAL_INTERCHANGE (UTF8_DISALLOW_SUPER|UTF8_DISALLOW_NONCHAR|UTF8_DISALLOW_SURROGATE|UTF8_DISALLOW_FE_FF) |
949cf498 | 293 | #define UTF8_WARN_ILLEGAL_INTERCHANGE \ |
33d9abfb | 294 | (UTF8_WARN_SUPER|UTF8_WARN_NONCHAR|UTF8_WARN_SURROGATE|UTF8_WARN_FE_FF) |
949cf498 KW |
295 | #define UTF8_ALLOW_ANY \ |
296 | (~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE|UTF8_WARN_ILLEGAL_INTERCHANGE)) | |
297 | #define UTF8_ALLOW_ANYUV \ | |
298 | (UTF8_ALLOW_EMPTY \ | |
299 | & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE|UTF8_WARN_ILLEGAL_INTERCHANGE)) | |
9f7f3913 TS |
300 | #define UTF8_ALLOW_DEFAULT (ckWARN(WARN_UTF8) ? 0 : \ |
301 | UTF8_ALLOW_ANYUV) | |
1d72bdf6 | 302 | |
7131f24d KW |
303 | /* Surrogates, non-character code points and above-Unicode code points are |
304 | * problematic in some contexts. This allows code that needs to check for | |
305 | * those to to quickly exclude the vast majority of code points it will | |
306 | * encounter */ | |
307 | #ifdef EBCDIC | |
308 | # define UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE UTF_TO_NATIVE(0xF1) | |
309 | #else | |
310 | # define UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE 0xED | |
311 | #endif | |
312 | ||
313 | /* ASCII EBCDIC I8 | |
314 | * U+D7FF: \xED\x9F\xBF \xF1\xB5\xBF\xBF last before surrogates | |
315 | * U+D800: \xED\xA0\x80 \xF1\xB6\xA0\xA0 1st surrogate | |
316 | * U+DFFF: \xED\xBF\xBF \xF1\xB7\xBF\xBF final surrogate | |
317 | * U+E000: \xEE\x80\x80 \xF1\xB8\xA0\xA0 next after surrogates | |
318 | */ | |
319 | #ifdef EBCDIC /* Both versions assume well-formed UTF8 */ | |
320 | # define UTF8_IS_SURROGATE(s) (*(s) == UTF_TO_NATIVE(0xF1) \ | |
880d3d38 | 321 | && ((*((s) +1) == UTF_TO_NATIVE(0xB6)) || *((s) + 1) == UTF_TO_NATIVE(0xB7))) |
7131f24d KW |
322 | #else |
323 | # define UTF8_IS_SURROGATE(s) (*(s) == 0xED && *((s) + 1) >= 0xA0) | |
324 | #endif | |
325 | ||
326 | /* ASCII EBCDIC I8 | |
327 | * U+10FFFF: \xF4\x8F\xBF\xBF \xF9\xA1\xBF\xBF\xBF max legal Unicode | |
328 | * U+110000: \xF4\x90\x80\x80 \xF9\xA2\xA0\xA0\xA0 | |
329 | * U+110001: \xF4\x90\x80\x81 \xF9\xA2\xA0\xA0\xA1 | |
330 | */ | |
331 | #ifdef EBCDIC /* Both versions assume well-formed UTF8 */ | |
332 | # define UTF8_IS_SUPER(s) (*(s) >= UTF_TO_NATIVE(0xF9) \ | |
333 | && (*(s) > UTF_TO_NATIVE(0xF9) || (*((s) + 1) >= UTF_TO_NATIVE(0xA2)))) | |
334 | #else | |
335 | # define UTF8_IS_SUPER(s) (*(s) >= 0xF4 \ | |
336 | && (*(s) > 0xF4 || (*((s) + 1) >= 0x90))) | |
337 | #endif | |
338 | ||
339 | /* ASCII EBCDIC I8 | |
340 | * U+FDCF: \xEF\xB7\x8F \xF1\xBF\xAE\xAF last before non-char block | |
341 | * U+FDD0: \xEF\xB7\x90 \xF1\xBF\xAE\xB0 first non-char in block | |
342 | * U+FDEF: \xEF\xB7\xAF \xF1\xBF\xAF\xAF last non-char in block | |
343 | * U+FDF0: \xEF\xB7\xB0 \xF1\xBF\xAF\xB0 first after non-char block | |
344 | * U+FFFF: \xEF\xBF\xBF \xF1\xBF\xBF\xBF | |
345 | * U+1FFFF: \xF0\x9F\xBF\xBF \xF3\xBF\xBF\xBF | |
346 | * U+2FFFF: \xF0\xAF\xBF\xBF \xF5\xBF\xBF\xBF | |
347 | * U+3FFFF: \xF0\xBF\xBF\xBF \xF7\xBF\xBF\xBF | |
348 | * U+4FFFF: \xF1\x8F\xBF\xBF \xF8\xA9\xBF\xBF\xBF | |
349 | * U+5FFFF: \xF1\x9F\xBF\xBF \xF8\xAB\xBF\xBF\xBF | |
350 | * U+6FFFF: \xF1\xAF\xBF\xBF \xF8\xAD\xBF\xBF\xBF | |
351 | * U+7FFFF: \xF1\xBF\xBF\xBF \xF8\xAF\xBF\xBF\xBF | |
352 | * U+8FFFF: \xF2\x8F\xBF\xBF \xF8\xB1\xBF\xBF\xBF | |
353 | * U+9FFFF: \xF2\x9F\xBF\xBF \xF8\xB3\xBF\xBF\xBF | |
354 | * U+AFFFF: \xF2\xAF\xBF\xBF \xF8\xB5\xBF\xBF\xBF | |
355 | * U+BFFFF: \xF2\xBF\xBF\xBF \xF8\xB7\xBF\xBF\xBF | |
356 | * U+CFFFF: \xF3\x8F\xBF\xBF \xF8\xB9\xBF\xBF\xBF | |
357 | * U+DFFFF: \xF3\x9F\xBF\xBF \xF8\xBB\xBF\xBF\xBF | |
358 | * U+EFFFF: \xF3\xAF\xBF\xBF \xF8\xBD\xBF\xBF\xBF | |
359 | * U+FFFFF: \xF3\xBF\xBF\xBF \xF8\xBF\xBF\xBF\xBF | |
360 | * U+10FFFF: \xF4\x8F\xBF\xBF \xF9\xA1\xBF\xBF\xBF | |
361 | */ | |
362 | #define UTF8_IS_NONCHAR_(s) ( \ | |
363 | *(s) >= UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE \ | |
364 | && ! UTF8_IS_SUPER(s) \ | |
365 | && UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_FIRST_PROBLEMATIC(s) \ | |
366 | ||
367 | #ifdef EBCDIC /* Both versions assume well-formed UTF8 */ | |
368 | # define UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s) \ | |
369 | ((*(s) == UTF_TO_NATIVE(0xF1) \ | |
370 | && (*((s) + 1) == UTF_TO_NATIVE(0xBF) \ | |
371 | && ((*((s) + 2) == UTF_TO_NATIVE(0xAE) \ | |
372 | && *((s) + 3) >= UTF_TO_NATIVE(0xB0)) \ | |
373 | || (*((s) + 2) == UTF_TO_NATIVE(0xAF) \ | |
374 | && *((s) + 3) <= UTF_TO_NATIVE(0xAF))))) \ | |
375 | || (UTF8SKIP(*(s)) > 3 \ | |
376 | /* (These were all derived by inspection and experimentation with an */ \ | |
377 | /* editor) The next line checks the next to final byte in the char */ \ | |
378 | && *((s) + UTF8SKIP(*(s)) - 2) == UTF_TO_NATIVE(0xBF) \ | |
379 | && *((s) + UTF8SKIP(*(s)) - 3) == UTF_TO_NATIVE(0xBF) \ | |
380 | && (NATIVE_TO_UTF(*((s) + UTF8SKIP(*(s)) - 4)) & 0x81) == 0x81 \ | |
381 | && (NATIVE_TO_UTF(*((s) + UTF8SKIP(*(s)) - 1)) & 0xBE) == 0XBE)) | |
382 | #else | |
383 | # define UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s) \ | |
384 | ((*(s) == 0xEF \ | |
385 | && ((*((s) + 1) == 0xB7 && (*((s) + 2) >= 0x90 && (*((s) + 2) <= 0xAF)))\ | |
386 | /* Gets U+FFF[EF] */ \ | |
387 | || (*((s) + 1) == 0xBF && ((*((s) + 2) & 0xBE) == 0xBE)))) \ | |
388 | || ((*((s) + 2) == 0xBF \ | |
389 | && (*((s) + 3) & 0xBE) == 0xBE \ | |
390 | /* Excludes things like U+10FFE = \xF0\x90\xBF\xBE */ \ | |
391 | && (*((s) + 1) & 0x8F) == 0x8F))) | |
392 | #endif | |
393 | ||
c867b360 JH |
394 | #define UNICODE_SURROGATE_FIRST 0xD800 |
395 | #define UNICODE_SURROGATE_LAST 0xDFFF | |
396 | #define UNICODE_REPLACEMENT 0xFFFD | |
397 | #define UNICODE_BYTE_ORDER_MARK 0xFEFF | |
1d72bdf6 | 398 | |
b851fbc1 | 399 | /* Though our UTF-8 encoding can go beyond this, |
c76687c5 | 400 | * let's be conservative and do as Unicode says. */ |
b851fbc1 JH |
401 | #define PERL_UNICODE_MAX 0x10FFFF |
402 | ||
949cf498 KW |
403 | #define UNICODE_WARN_SURROGATE 0x0001 /* UTF-16 surrogates */ |
404 | #define UNICODE_WARN_NONCHAR 0x0002 /* Non-char code points */ | |
405 | #define UNICODE_WARN_SUPER 0x0004 /* Above 0x10FFFF */ | |
406 | #define UNICODE_WARN_FE_FF 0x0008 /* Above 0x10FFFF */ | |
407 | #define UNICODE_DISALLOW_SURROGATE 0x0010 | |
408 | #define UNICODE_DISALLOW_NONCHAR 0x0020 | |
409 | #define UNICODE_DISALLOW_SUPER 0x0040 | |
410 | #define UNICODE_DISALLOW_FE_FF 0x0080 | |
411 | #define UNICODE_WARN_ILLEGAL_INTERCHANGE \ | |
412 | (UNICODE_WARN_SURROGATE|UNICODE_WARN_NONCHAR|UNICODE_WARN_SUPER) | |
413 | #define UNICODE_DISALLOW_ILLEGAL_INTERCHANGE \ | |
414 | (UNICODE_DISALLOW_SURROGATE|UNICODE_DISALLOW_NONCHAR|UNICODE_DISALLOW_SUPER) | |
415 | ||
416 | /* For backward source compatibility, as are now the default */ | |
417 | #define UNICODE_ALLOW_SURROGATE 0 | |
418 | #define UNICODE_ALLOW_SUPER 0 | |
419 | #define UNICODE_ALLOW_ANY 0 | |
b851fbc1 | 420 | |
1d72bdf6 NIS |
421 | #define UNICODE_IS_SURROGATE(c) ((c) >= UNICODE_SURROGATE_FIRST && \ |
422 | (c) <= UNICODE_SURROGATE_LAST) | |
a10ec373 | 423 | #define UNICODE_IS_REPLACEMENT(c) ((c) == UNICODE_REPLACEMENT) |
872c91ae | 424 | #define UNICODE_IS_BYTE_ORDER_MARK(c) ((c) == UNICODE_BYTE_ORDER_MARK) |
7131f24d KW |
425 | #define UNICODE_IS_NONCHAR(c) ((c >= 0xFDD0 && c <= 0xFDEF) \ |
426 | /* The other noncharacters end in FFFE or FFFF, which \ | |
427 | * the mask below catches both of, but beyond the last \ | |
428 | * official unicode code point, they aren't \ | |
429 | * noncharacters, since those aren't Unicode \ | |
430 | * characters at all */ \ | |
431 | || ((((c & 0xFFFE) == 0xFFFE)) && ! UNICODE_IS_SUPER(c))) | |
432 | #define UNICODE_IS_SUPER(c) ((c) > PERL_UNICODE_MAX) | |
433 | #define UNICODE_IS_FE_FF(c) ((c) > 0x7FFFFFFF) | |
1d72bdf6 | 434 | |
f067b878 NA |
435 | #ifdef HAS_QUAD |
436 | # define UTF8_QUAD_MAX UINT64_C(0x1000000000) | |
437 | #endif | |
3bd709b1 | 438 | |
09091399 JH |
439 | #define UNICODE_GREEK_CAPITAL_LETTER_SIGMA 0x03A3 |
440 | #define UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA 0x03C2 | |
441 | #define UNICODE_GREEK_SMALL_LETTER_SIGMA 0x03C3 | |
9dcbe121 | 442 | #define GREEK_SMALL_LETTER_MU 0x03BC |
78a0d3cc KW |
443 | #define GREEK_CAPITAL_LETTER_MU 0x039C /* Upper and title case of MICRON */ |
444 | #define LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS 0x0178 /* Also is title case */ | |
97298f37 | 445 | #define LATIN_CAPITAL_LETTER_SHARP_S 0x1E9E |
09091399 | 446 | |
9e55ce06 | 447 | #define UNI_DISPLAY_ISPRINT 0x0001 |
c728cb41 JH |
448 | #define UNI_DISPLAY_BACKSLASH 0x0002 |
449 | #define UNI_DISPLAY_QQ (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH) | |
450 | #define UNI_DISPLAY_REGEX (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH) | |
9e55ce06 | 451 | |
5cd46e1f KW |
452 | #ifndef EBCDIC |
453 | # define LATIN_SMALL_LETTER_SHARP_S 0x00DF | |
454 | # define LATIN_SMALL_LETTER_Y_WITH_DIAERESIS 0x00FF | |
455 | # define MICRO_SIGN 0x00B5 | |
f508a607 KW |
456 | # define LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE 0x00C5 |
457 | # define LATIN_SMALL_LETTER_A_WITH_RING_ABOVE 0x00E5 | |
5cd46e1f KW |
458 | #endif |
459 | ||
460 | #define ANYOF_FOLD_SHARP_S(node, input, end) \ | |
461 | (ANYOF_BITMAP_TEST(node, LATIN_SMALL_LETTER_SHARP_S) && \ | |
137165a6 | 462 | (ANYOF_NONBITMAP(node)) && \ |
39065660 | 463 | (ANYOF_FLAGS(node) & ANYOF_LOC_NONBITMAP_FOLD) && \ |
07b6858f JH |
464 | ((end) > (input) + 1) && \ |
465 | toLOWER((input)[0]) == 's' && \ | |
466 | toLOWER((input)[1]) == 's') | |
ebc501f0 | 467 | #define SHARP_S_SKIP 2 |
3b0fc154 | 468 | |
2f454f11 | 469 | #ifndef EBCDIC |
c78f6c49 | 470 | # define IS_UTF8_CHAR_1(p) \ |
3b0fc154 | 471 | ((p)[0] <= 0x7F) |
c78f6c49 | 472 | # define IS_UTF8_CHAR_2(p) \ |
3b0fc154 JH |
473 | ((p)[0] >= 0xC2 && (p)[0] <= 0xDF && \ |
474 | (p)[1] >= 0x80 && (p)[1] <= 0xBF) | |
c78f6c49 | 475 | # define IS_UTF8_CHAR_3a(p) \ |
3b0fc154 JH |
476 | ((p)[0] == 0xE0 && \ |
477 | (p)[1] >= 0xA0 && (p)[1] <= 0xBF && \ | |
478 | (p)[2] >= 0x80 && (p)[2] <= 0xBF) | |
c78f6c49 | 479 | # define IS_UTF8_CHAR_3b(p) \ |
3b0fc154 JH |
480 | ((p)[0] >= 0xE1 && (p)[0] <= 0xEC && \ |
481 | (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ | |
482 | (p)[2] >= 0x80 && (p)[2] <= 0xBF) | |
c78f6c49 | 483 | # define IS_UTF8_CHAR_3c(p) \ |
3b0fc154 JH |
484 | ((p)[0] == 0xED && \ |
485 | (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ | |
486 | (p)[2] >= 0x80 && (p)[2] <= 0xBF) | |
c78f6c49 KW |
487 | /* In IS_UTF8_CHAR_3c(p) one could use |
488 | * (p)[1] >= 0x80 && (p)[1] <= 0x9F | |
489 | * if one wanted to exclude surrogates. */ | |
490 | # define IS_UTF8_CHAR_3d(p) \ | |
3b0fc154 JH |
491 | ((p)[0] >= 0xEE && (p)[0] <= 0xEF && \ |
492 | (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ | |
493 | (p)[2] >= 0x80 && (p)[2] <= 0xBF) | |
c78f6c49 | 494 | # define IS_UTF8_CHAR_4a(p) \ |
3b0fc154 JH |
495 | ((p)[0] == 0xF0 && \ |
496 | (p)[1] >= 0x90 && (p)[1] <= 0xBF && \ | |
497 | (p)[2] >= 0x80 && (p)[2] <= 0xBF && \ | |
498 | (p)[3] >= 0x80 && (p)[3] <= 0xBF) | |
c78f6c49 | 499 | # define IS_UTF8_CHAR_4b(p) \ |
3b0fc154 JH |
500 | ((p)[0] >= 0xF1 && (p)[0] <= 0xF3 && \ |
501 | (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ | |
502 | (p)[2] >= 0x80 && (p)[2] <= 0xBF && \ | |
503 | (p)[3] >= 0x80 && (p)[3] <= 0xBF) | |
504 | /* In IS_UTF8_CHAR_4c(p) one could use | |
505 | * (p)[0] == 0xF4 | |
506 | * if one wanted to stop at the Unicode limit U+10FFFF. | |
507 | * The 0xF7 allows us to go to 0x1fffff (0x200000 would | |
508 | * require five bytes). Not doing any further code points | |
509 | * since that is not needed (and that would not be strict | |
510 | * UTF-8, anyway). The "slow path" in Perl_is_utf8_char() | |
511 | * will take care of the "extended UTF-8". */ | |
c78f6c49 | 512 | # define IS_UTF8_CHAR_4c(p) \ |
a18d6e6e | 513 | ((p)[0] >= 0xF4 && (p)[0] <= 0xF7 && \ |
3b0fc154 JH |
514 | (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ |
515 | (p)[2] >= 0x80 && (p)[2] <= 0xBF && \ | |
516 | (p)[3] >= 0x80 && (p)[3] <= 0xBF) | |
517 | ||
c78f6c49 | 518 | # define IS_UTF8_CHAR_3(p) \ |
3b0fc154 JH |
519 | (IS_UTF8_CHAR_3a(p) || \ |
520 | IS_UTF8_CHAR_3b(p) || \ | |
521 | IS_UTF8_CHAR_3c(p) || \ | |
522 | IS_UTF8_CHAR_3d(p)) | |
c78f6c49 | 523 | # define IS_UTF8_CHAR_4(p) \ |
3b0fc154 JH |
524 | (IS_UTF8_CHAR_4a(p) || \ |
525 | IS_UTF8_CHAR_4b(p) || \ | |
526 | IS_UTF8_CHAR_4c(p)) | |
527 | ||
528 | /* IS_UTF8_CHAR(p) is strictly speaking wrong (not UTF-8) because it | |
529 | * (1) allows UTF-8 encoded UTF-16 surrogates | |
530 | * (2) it allows code points past U+10FFFF. | |
531 | * The Perl_is_utf8_char() full "slow" code will handle the Perl | |
532 | * "extended UTF-8". */ | |
c78f6c49 | 533 | # define IS_UTF8_CHAR(p, n) \ |
3b0fc154 JH |
534 | ((n) == 1 ? IS_UTF8_CHAR_1(p) : \ |
535 | (n) == 2 ? IS_UTF8_CHAR_2(p) : \ | |
536 | (n) == 3 ? IS_UTF8_CHAR_3(p) : \ | |
537 | (n) == 4 ? IS_UTF8_CHAR_4(p) : 0) | |
538 | ||
c78f6c49 | 539 | # define IS_UTF8_CHAR_FAST(n) ((n) <= 4) |
768c67ee | 540 | |
2f454f11 KW |
541 | #else /* EBCDIC */ |
542 | ||
543 | /* This is an attempt to port IS_UTF8_CHAR to EBCDIC based on eyeballing. | |
544 | * untested. If want to exclude surrogates and above-Unicode, see the | |
545 | * definitions for UTF8_IS_SURROGATE and UTF8_IS_SUPER */ | |
546 | # define IS_UTF8_CHAR_1(p) \ | |
547 | (NATIVE_TO_ASCII((p)[0]) <= 0x9F) | |
548 | # define IS_UTF8_CHAR_2(p) \ | |
549 | (NATIVE_TO_I8((p)[0]) >= 0xC5 && NATIVE_TO_I8((p)[0]) <= 0xDF && \ | |
550 | NATIVE_TO_I8((p)[1]) >= 0xA0 && NATIVE_TO_I8((p)[1]) <= 0xBF) | |
551 | # define IS_UTF8_CHAR_3(p) \ | |
552 | (NATIVE_TO_I8((p)[0]) == 0xE1 && NATIVE_TO_I8((p)[1]) <= 0xEF && \ | |
553 | NATIVE_TO_I8((p)[1]) >= 0xA0 && NATIVE_TO_I8((p)[1]) <= 0xBF && \ | |
554 | NATIVE_TO_I8((p)[2]) >= 0xA0 && NATIVE_TO_I8((p)[2]) <= 0xBF) | |
555 | # define IS_UTF8_CHAR_4a(p) \ | |
556 | (NATIVE_TO_I8((p)[0]) == 0xF0 && \ | |
557 | NATIVE_TO_I8((p)[1]) >= 0xB0 && NATIVE_TO_I8((p)[1]) <= 0xBF && \ | |
558 | NATIVE_TO_I8((p)[2]) >= 0xA0 && NATIVE_TO_I8((p)[2]) <= 0xBF && \ | |
559 | NATIVE_TO_I8((p)[3]) >= 0xA0 && NATIVE_TO_I8((p)[3]) <= 0xBF) | |
560 | # define IS_UTF8_CHAR_4b(p) \ | |
561 | (NATIVE_TO_I8((p)[0]) >= 0xF1 && NATIVE_TO_I8((p)[0]) <= 0xF7 && \ | |
562 | NATIVE_TO_I8((p)[1]) >= 0xA0 && NATIVE_TO_I8((p)[1]) <= 0xBF && \ | |
563 | NATIVE_TO_I8((p)[2]) >= 0xA0 && NATIVE_TO_I8((p)[2]) <= 0xBF && \ | |
564 | NATIVE_TO_I8((p)[3]) >= 0xA0 && NATIVE_TO_I8((p)[3]) <= 0xBF) | |
565 | # define IS_UTF8_CHAR_5a(p) \ | |
566 | (NATIVE_TO_I8((p)[0]) == 0xF8 && \ | |
567 | NATIVE_TO_I8((p)[1]) >= 0xA8 && NATIVE_TO_I8((p)[1]) <= 0xBF && \ | |
568 | NATIVE_TO_I8((p)[1]) >= 0xA0 && NATIVE_TO_I8((p)[1]) <= 0xBF && \ | |
569 | NATIVE_TO_I8((p)[2]) >= 0xA0 && NATIVE_TO_I8((p)[2]) <= 0xBF && \ | |
570 | NATIVE_TO_I8((p)[3]) >= 0xA0 && NATIVE_TO_I8((p)[3]) <= 0xBF) | |
571 | # define IS_UTF8_CHAR_5b(p) \ | |
572 | (NATIVE_TO_I8((p)[0]) >= 0xF9 && NATIVE_TO_I8((p)[1]) <= 0xFB && \ | |
573 | NATIVE_TO_I8((p)[1]) >= 0xA0 && NATIVE_TO_I8((p)[1]) <= 0xBF && \ | |
574 | NATIVE_TO_I8((p)[1]) >= 0xA0 && NATIVE_TO_I8((p)[1]) <= 0xBF && \ | |
575 | NATIVE_TO_I8((p)[2]) >= 0xA0 && NATIVE_TO_I8((p)[2]) <= 0xBF && \ | |
576 | NATIVE_TO_I8((p)[3]) >= 0xA0 && NATIVE_TO_I8((p)[3]) <= 0xBF) | |
577 | ||
578 | # define IS_UTF8_CHAR_4(p) \ | |
579 | (IS_UTF8_CHAR_4a(p) || \ | |
580 | IS_UTF8_CHAR_4b(p)) | |
581 | # define IS_UTF8_CHAR_5(p) \ | |
582 | (IS_UTF8_CHAR_5a(p) || \ | |
583 | IS_UTF8_CHAR_5b(p)) | |
584 | # define IS_UTF8_CHAR(p, n) \ | |
585 | ((n) == 1 ? IS_UTF8_CHAR_1(p) : \ | |
586 | (n) == 2 ? IS_UTF8_CHAR_2(p) : \ | |
587 | (n) == 3 ? IS_UTF8_CHAR_3(p) : \ | |
588 | (n) == 4 ? IS_UTF8_CHAR_4(p) : \ | |
589 | (n) == 5 ? IS_UTF8_CHAR_5(p) : 0) | |
590 | ||
591 | # define IS_UTF8_CHAR_FAST(n) ((n) <= 5) | |
592 | ||
77263263 | 593 | #endif /* IS_UTF8_CHAR() for UTF-8 */ |
e9a8c099 MHM |
594 | |
595 | /* | |
596 | * Local variables: | |
597 | * c-indentation-style: bsd | |
598 | * c-basic-offset: 4 | |
599 | * indent-tabs-mode: t | |
600 | * End: | |
601 | * | |
602 | * ex: set ts=8 sts=4 sw=4 noet: | |
603 | */ |