This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
https://perl5.git.perl.org / perl5.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
embed.fnc: Change formal param name to match docs
[perl5.git] / utf8.h
CommitLineData
a0ed51b3
LW		 1/* utf8.h
2 *
2eee27d7
SS		 3 * Copyright (C) 2000, 2001, 2002, 2005, 2006, 2007, 2009,
4 * 2010, 2011 by Larry Wall and others
a0ed51b3
LW		 5 *
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
8 *
9 */
10
39e02b42 11/* Use UTF-8 as the default script encoding?
1e54db1a 12 * Turning this on will break scripts having non-UTF-8 binary
39e02b42
JH		 13 * data (such as Latin-1) in string literals. */
14#ifdef USE_UTF8_SCRIPTS
15# define USE_UTF8_IN_NAMES (!IN_BYTES)
16#else
17# define USE_UTF8_IN_NAMES (PL_hints & HINT_UTF8)
18#endif
19
051a06d4
KW		 20/* For to_utf8_fold_flags, q.v. */
21#define FOLD_FLAGS_LOCALE 0x1
22#define FOLD_FLAGS_FULL 0x2
23
36bb2ab6 24#define to_uni_fold(c, p, lenp) _to_uni_fold_flags(c, p, lenp, 1)
051a06d4
KW		 25#define to_utf8_fold(c, p, lenp) _to_utf8_fold_flags(c, p, lenp, \
26 FOLD_FLAGS_FULL, NULL)
27#define to_utf8_lower(a,b,c) _to_utf8_lower_flags(a,b,c,0, NULL)
28#define to_utf8_upper(a,b,c) _to_utf8_upper_flags(a,b,c,0, NULL)
29#define to_utf8_title(a,b,c) _to_utf8_title_flags(a,b,c,0, NULL)
36bb2ab6 30
fd7cb289
RGS		 31/* Source backward compatibility. */
32#define uvuni_to_utf8(d, uv) uvuni_to_utf8_flags(d, uv, 0)
33#define is_utf8_string_loc(s, len, ep) is_utf8_string_loclen(s, len, ep, 0)
34
eda9cac1
KW		 35#define foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \
36 foldEQ_utf8_flags(s1, pe1, l1, u1, s2, pe2, l2, u2, 0)
a33c29bc 37#define FOLDEQ_UTF8_NOMIX_ASCII (1 << 0)
5e64d0fa 38#define FOLDEQ_UTF8_LOCALE (1 << 1)
18f762c3
KW		 39#define FOLDEQ_S1_ALREADY_FOLDED (1 << 2)
40#define FOLDEQ_S2_ALREADY_FOLDED (1 << 3)
a33c29bc 41
e6226b18
KW		 42/*
43=for apidoc ibcmp_utf8
44
45This is a synonym for (! foldEQ_utf8())
46
47=cut
48*/
49#define ibcmp_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \
50 cBOOL(! foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2))
51
1d72bdf6
NIS		 52#ifdef EBCDIC
53/* The equivalent of these macros but implementing UTF-EBCDIC
54 are in the following header file:
55 */
56
57#include "utfebcdic.h"
fd7cb289 58
d06134e5 59#else /* ! EBCDIC */
73c4f7a1
GS		 60START_EXTERN_C
61
a0ed51b3 62#ifdef DOINIT
6f06b55f 63EXTCONST unsigned char PL_utf8skip[] = {
a0ed51b3
LW		 641,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
651,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
661,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
671,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
681,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus */
691,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus */
702,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, /* scripts */
3c77ea2b 713,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4,5,5,5,5,6,6, /* cjk etc. */
7eb168e2
KW		 727,13, /* Perl extended (not official UTF-8). Up to 72bit allowed (64-bit +
73 reserved). */
a0ed51b3
LW		 74};
75#else
6f06b55f 76EXTCONST unsigned char PL_utf8skip[];
a0ed51b3
LW		 77#endif
78
73c4f7a1 79END_EXTERN_C
7e2040f0 80
1d72bdf6
NIS		 81/* Native character to iso-8859-1 */
82#define NATIVE_TO_ASCII(ch) (ch)
83#define ASCII_TO_NATIVE(ch) (ch)
84/* Transform after encoding */
85#define NATIVE_TO_UTF(ch) (ch)
86#define UTF_TO_NATIVE(ch) (ch)
87/* Transforms in wide UV chars */
88#define UNI_TO_NATIVE(ch) (ch)
89#define NATIVE_TO_UNI(ch) (ch)
90/* Transforms in invariant space */
91#define NATIVE_TO_NEED(enc,ch) (ch)
92#define ASCII_TO_NEED(enc,ch) (ch)
d7578b48 93
d06134e5 94/* As there are no translations, avoid the function wrapper */
1754c1a1
NC		 95#define utf8n_to_uvchr utf8n_to_uvuni
96#define uvchr_to_utf8 uvuni_to_utf8
2b9d42f0 97
877d9f0d 98/*
9041c2e3 99
8c007b5a 100 The following table is from Unicode 3.2.
877d9f0d
JH		 101
102 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
103
375122d7 104 U+0000..U+007F 00..7F
e1b711da 105 U+0080..U+07FF * C2..DF 80..BF
37e2e78e 106 U+0800..U+0FFF E0 * A0..BF 80..BF
375122d7 107 U+1000..U+CFFF E1..EC 80..BF 80..BF
e1b711da 108 U+D000..U+D7FF ED 80..9F 80..BF
37e2e78e 109 U+D800..U+DFFF +++++++ utf16 surrogates, not legal utf8 +++++++
375122d7 110 U+E000..U+FFFF EE..EF 80..BF 80..BF
37e2e78e 111 U+10000..U+3FFFF F0 * 90..BF 80..BF 80..BF
877d9f0d
JH		 112 U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF
113 U+100000..U+10FFFF F4 80..8F 80..BF 80..BF
114
e1b711da 115Note the gaps before several of the byte entries above marked by '*'. These are
37e2e78e
KW		 116caused by legal UTF-8 avoiding non-shortest encodings: it is technically
117possible to UTF-8-encode a single code point in different ways, but that is
118explicitly forbidden, and the shortest possible encoding should always be used
119(and that is what Perl does).
8c007b5a 120
877d9f0d
JH		 121 */
122
8c007b5a
JH		 123/*
124 Another way to look at it, as bits:
125
126 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
127
128 0aaaaaaa 0aaaaaaa
129 00000bbbbbaaaaaa 110bbbbb 10aaaaaa
130 ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa
131 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa
132
133As you can see, the continuation bytes all begin with C<10>, and the
e1b711da 134leading bits of the start byte tell how many bytes there are in the
8c007b5a
JH		 135encoded character.
136
65ab9279
TC		 137Perl's extended UTF-8 means we can have start bytes up to FF.
138
8c007b5a
JH		 139*/
140
c4d5f83a 141#define UNI_IS_INVARIANT(c) (((UV)c) < 0x80)
0447e8df 142#define UTF8_IS_START(c) (((U8)c) >= 0xc2)
c512ce4f
JH		 143#define UTF8_IS_CONTINUATION(c) (((U8)c) >= 0x80 && (((U8)c) <= 0xbf))
144#define UTF8_IS_CONTINUED(c) (((U8)c) & 0x80)
0ae1fa71
CH		 145
146/* Masking with 0xfe allows low bit to be 0 or 1; thus this matches 0xc[23] */
147#define UTF8_IS_DOWNGRADEABLE_START(c) (((U8)c & 0xfe) == 0xc2)
8850bf83 148
22901f30
RGS		 149#define UTF_START_MARK(len) (((len) > 7) ? 0xFF : (0xFE << (7-(len))))
150#define UTF_START_MASK(len) (((len) >= 7) ? 0x00 : (0x1F >> ((len)-2)))
1d72bdf6
NIS		 151
152#define UTF_CONTINUATION_MARK 0x80
153#define UTF_ACCUMULATION_SHIFT 6
154#define UTF_CONTINUATION_MASK ((U8)0x3f)
c512ce4f 155
eb83ed87
KW		 156/* This sets the UTF_CONTINUATION_MASK in the upper bits of a word. If a value
157 * is anded with it, and the result is non-zero, then using the original value
158 * in UTF8_ACCUMULATE will overflow, shifting bits off the left */
159#define UTF_ACCUMULATION_OVERFLOW_MASK \
160 (((UV) UTF_CONTINUATION_MASK) << ((sizeof(UV) * CHARBITS) - UTF_ACCUMULATION_SHIFT))
161
1d68d6cd 162#ifdef HAS_QUAD
5bbb0b5a 163#define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \
1d68d6cd
SC		 164 (uv) < 0x800 ? 2 : \
165 (uv) < 0x10000 ? 3 : \
166 (uv) < 0x200000 ? 4 : \
167 (uv) < 0x4000000 ? 5 : \
168 (uv) < 0x80000000 ? 6 : \
9041c2e3 169 (uv) < UTF8_QUAD_MAX ? 7 : 13 )
1d68d6cd
SC		 170#else
171/* No, I'm not even going to *TRY* putting #ifdef inside a #define */
5bbb0b5a 172#define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \
1d68d6cd
SC		 173 (uv) < 0x800 ? 2 : \
174 (uv) < 0x10000 ? 3 : \
175 (uv) < 0x200000 ? 4 : \
176 (uv) < 0x4000000 ? 5 : \
177 (uv) < 0x80000000 ? 6 : 7 )
178#endif
179
d06134e5
KW		 180#endif /* EBCDIC vs ASCII */
181
182/* Rest of these are attributes of Unicode and perl's internals rather than the
183 * encoding, or happen to be the same in both ASCII and EBCDIC (at least at
184 * this level; the macros that some of these call may have different
185 * definitions in the two encodings */
186
187#define NATIVE8_TO_UNI(ch) NATIVE_TO_ASCII(ch) /* a clearer synonym */
188
189#define UTF8_ACCUMULATE(old, new) (((old) << UTF_ACCUMULATION_SHIFT) | (((U8)new) & UTF_CONTINUATION_MASK))
190
2950f2a7
KW		 191/* Convert a two (not one) byte utf8 character to a unicode code point value.
192 * Needs just one iteration of accumulate. Should not be used unless it is
193 * known that the two bytes are legal: 1) two-byte start, and 2) continuation.
194 * Note that the result can be larger than 255 if the input character is not
195 * downgradable */
196#define TWO_BYTE_UTF8_TO_UNI(HI, LO) \
197 UTF8_ACCUMULATE((NATIVE_TO_UTF(HI) & UTF_START_MASK(2)), \
198 NATIVE_TO_UTF(LO))
199
d06134e5
KW		 200#define UTF8SKIP(s) PL_utf8skip[*(const U8*)(s)]
201
202#define UTF8_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE_TO_UTF(c))
203#define NATIVE_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE8_TO_UNI(c))
204
205#define MAX_PORTABLE_UTF8_TWO_BYTE 0x3FF /* constrained by EBCDIC */
206
207/* The macros in the next sets are used to generate the two utf8 or utfebcdic
208 * bytes from an ordinal that is known to fit into two bytes; it must be less
209 * than 0x3FF to work across both encodings. */
210/* Nocast allows these to be used in the case label of a switch statement */
0072151e 211#define UTF8_TWO_BYTE_HI_nocast(c) UTF_TO_NATIVE(((c) >> UTF_ACCUMULATION_SHIFT) | (0xFF & UTF_START_MARK(2)))
14c28784 212#define UTF8_TWO_BYTE_LO_nocast(c) UTF_TO_NATIVE(((c) & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK)
d06134e5
KW		 213
214#define UTF8_TWO_BYTE_HI(c) ((U8) (UTF8_TWO_BYTE_HI_nocast(c)))
215#define UTF8_TWO_BYTE_LO(c) ((U8) (UTF8_TWO_BYTE_LO_nocast(c)))
216
217/* This name is used when the source is a single byte */
218#define UTF8_EIGHT_BIT_HI(c) UTF8_TWO_BYTE_HI((U8)(c))
219#define UTF8_EIGHT_BIT_LO(c) UTF8_TWO_BYTE_LO((U8)(c))
220
7e2040f0 221/*
e3036cf4 222 * 'UTF' is whether or not p is encoded in UTF8. The names 'foo_lazy_if' stem
20df05f4
KW		 223 * from an earlier version of these macros in which they didn't call the
224 * foo_utf8() macros (i.e. were 'lazy') unless they decided that *p is the
225 * beginning of a utf8 character. Now that foo_utf8() determines that itself,
226 * no need to do it again here
7e2040f0 227 */
e3036cf4
KW		 228#define isIDFIRST_lazy_if(p,UTF) ((IN_BYTES || !UTF ) \
229 ? isIDFIRST(*(p)) \
230 : isIDFIRST_utf8((const U8*)p))
231#define isALNUM_lazy_if(p,UTF) ((IN_BYTES || (!UTF )) \
232 ? isALNUM(*(p)) \
233 : isALNUM_utf8((const U8*)p))
1d72bdf6 234
7e2040f0
GS		 235#define isIDFIRST_lazy(p) isIDFIRST_lazy_if(p,1)
236#define isALNUM_lazy(p) isALNUM_lazy_if(p,1)
3bd709b1 237
89ebb4a3
JH		 238#define UTF8_MAXBYTES 13
239/* How wide can a single UTF-8 encoded character become in bytes.
240 * NOTE: Strictly speaking Perl's UTF-8 should not be called UTF-8
241 * since UTF-8 is an encoding of Unicode and given Unicode's current
242 * upper limit only four bytes is possible. Perl thinks of UTF-8
243 * as a way to encode non-negative integers in a binary format. */
244#define UTF8_MAXLEN UTF8_MAXBYTES
245
89ebb4a3
JH		 246/* The maximum number of UTF-8 bytes a single Unicode character can
247 * uppercase/lowercase/fold into; this number depends on the Unicode
248 * version. An example of maximal expansion is the U+03B0 which
249 * uppercases to U+03C5 U+0308 U+0301. The Unicode databases that
d06134e5 250 * tell these things are UnicodeData.txt, CaseFolding.txt, and
88d45ddf
KW		 251 * SpecialCasing.txt. The value is 6 for strict Unicode characters, but it has
252 * to be as big as Perl allows for a single character */
253#define UTF8_MAXBYTES_CASE UTF8_MAXBYTES
3bd709b1 254
8cb75cc8
KW		 255/* A Unicode character can fold to up to 3 characters */
256#define UTF8_MAX_FOLD_CHAR_EXPAND 3
257
a98fe34d 258#define IN_BYTES (CopHINTS_get(PL_curcop) & HINT_BYTES)
0064a8a9 259#define DO_UTF8(sv) (SvUTF8(sv) && !IN_BYTES)
66cbab2c 260#define IN_UNI_8_BIT \
b36bf33f
KW		 261 (CopHINTS_get(PL_curcop) & (HINT_UNI_8_BIT|HINT_LOCALE_NOT_CHARS) \
262 && ! IN_LOCALE_RUNTIME && ! IN_BYTES)
263
1d72bdf6 264
c76687c5
KW		 265#define UTF8_ALLOW_EMPTY 0x0001 /* Allow a zero length string */
266
267/* Allow first byte to be a continuation byte */
1d72bdf6 268#define UTF8_ALLOW_CONTINUATION 0x0002
c76687c5
KW		 269
270/* Allow second... bytes to be non-continuation bytes */
1d72bdf6 271#define UTF8_ALLOW_NON_CONTINUATION 0x0004
949cf498
KW		 272
273/* expecting more bytes than were available in the string */
274#define UTF8_ALLOW_SHORT 0x0008
275
276/* Overlong sequence; i.e., the code point can be specified in fewer bytes. */
277#define UTF8_ALLOW_LONG 0x0010
278
279#define UTF8_DISALLOW_SURROGATE 0x0020 /* Unicode surrogates */
280#define UTF8_WARN_SURROGATE 0x0040
281
282#define UTF8_DISALLOW_NONCHAR 0x0080 /* Unicode non-character */
283#define UTF8_WARN_NONCHAR 0x0100 /* code points */
284
285#define UTF8_DISALLOW_SUPER 0x0200 /* Super-set of Unicode: code */
286#define UTF8_WARN_SUPER 0x0400 /* points above the legal max */
287
288/* Code points which never were part of the original UTF-8 standard, the first
289 * byte of which is a FE or FF on ASCII platforms. */
290#define UTF8_DISALLOW_FE_FF 0x0800
291#define UTF8_WARN_FE_FF 0x1000
292
293#define UTF8_CHECK_ONLY 0x2000
294
295/* For backwards source compatibility. They do nothing, as the default now
296 * includes what they used to mean. The first one's meaning was to allow the
297 * just the single non-character 0xFFFF */
298#define UTF8_ALLOW_FFFF 0
299#define UTF8_ALLOW_SURROGATE 0
300
33d9abfb 301#define UTF8_DISALLOW_ILLEGAL_INTERCHANGE (UTF8_DISALLOW_SUPER|UTF8_DISALLOW_NONCHAR|UTF8_DISALLOW_SURROGATE|UTF8_DISALLOW_FE_FF)
949cf498 302#define UTF8_WARN_ILLEGAL_INTERCHANGE \
33d9abfb 303 (UTF8_WARN_SUPER|UTF8_WARN_NONCHAR|UTF8_WARN_SURROGATE|UTF8_WARN_FE_FF)
949cf498
KW		 304#define UTF8_ALLOW_ANY \
305 (~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE|UTF8_WARN_ILLEGAL_INTERCHANGE))
306#define UTF8_ALLOW_ANYUV \
307 (UTF8_ALLOW_EMPTY \
308 & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE|UTF8_WARN_ILLEGAL_INTERCHANGE))
9f7f3913
TS		 309#define UTF8_ALLOW_DEFAULT (ckWARN(WARN_UTF8) ? 0 : \
310 UTF8_ALLOW_ANYUV)
1d72bdf6 311
7131f24d
KW		 312/* Surrogates, non-character code points and above-Unicode code points are
313 * problematic in some contexts. This allows code that needs to check for
314 * those to to quickly exclude the vast majority of code points it will
315 * encounter */
316#ifdef EBCDIC
317# define UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE UTF_TO_NATIVE(0xF1)
318#else
319# define UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE 0xED
320#endif
321
322/* ASCII EBCDIC I8
323 * U+D7FF: \xED\x9F\xBF \xF1\xB5\xBF\xBF last before surrogates
324 * U+D800: \xED\xA0\x80 \xF1\xB6\xA0\xA0 1st surrogate
325 * U+DFFF: \xED\xBF\xBF \xF1\xB7\xBF\xBF final surrogate
326 * U+E000: \xEE\x80\x80 \xF1\xB8\xA0\xA0 next after surrogates
327 */
328#ifdef EBCDIC /* Both versions assume well-formed UTF8 */
329# define UTF8_IS_SURROGATE(s) (*(s) == UTF_TO_NATIVE(0xF1) \
880d3d38 330 && ((*((s) +1) == UTF_TO_NATIVE(0xB6)) || *((s) + 1) == UTF_TO_NATIVE(0xB7)))
7131f24d
KW		 331#else
332# define UTF8_IS_SURROGATE(s) (*(s) == 0xED && *((s) + 1) >= 0xA0)
333#endif
334
335/* ASCII EBCDIC I8
336 * U+10FFFF: \xF4\x8F\xBF\xBF \xF9\xA1\xBF\xBF\xBF max legal Unicode
337 * U+110000: \xF4\x90\x80\x80 \xF9\xA2\xA0\xA0\xA0
338 * U+110001: \xF4\x90\x80\x81 \xF9\xA2\xA0\xA0\xA1
339 */
340#ifdef EBCDIC /* Both versions assume well-formed UTF8 */
341# define UTF8_IS_SUPER(s) (*(s) >= UTF_TO_NATIVE(0xF9) \
342 && (*(s) > UTF_TO_NATIVE(0xF9) || (*((s) + 1) >= UTF_TO_NATIVE(0xA2))))
343#else
344# define UTF8_IS_SUPER(s) (*(s) >= 0xF4 \
345 && (*(s) > 0xF4 || (*((s) + 1) >= 0x90)))
346#endif
347
348/* ASCII EBCDIC I8
349 * U+FDCF: \xEF\xB7\x8F \xF1\xBF\xAE\xAF last before non-char block
350 * U+FDD0: \xEF\xB7\x90 \xF1\xBF\xAE\xB0 first non-char in block
351 * U+FDEF: \xEF\xB7\xAF \xF1\xBF\xAF\xAF last non-char in block
352 * U+FDF0: \xEF\xB7\xB0 \xF1\xBF\xAF\xB0 first after non-char block
353 * U+FFFF: \xEF\xBF\xBF \xF1\xBF\xBF\xBF
354 * U+1FFFF: \xF0\x9F\xBF\xBF \xF3\xBF\xBF\xBF
355 * U+2FFFF: \xF0\xAF\xBF\xBF \xF5\xBF\xBF\xBF
356 * U+3FFFF: \xF0\xBF\xBF\xBF \xF7\xBF\xBF\xBF
357 * U+4FFFF: \xF1\x8F\xBF\xBF \xF8\xA9\xBF\xBF\xBF
358 * U+5FFFF: \xF1\x9F\xBF\xBF \xF8\xAB\xBF\xBF\xBF
359 * U+6FFFF: \xF1\xAF\xBF\xBF \xF8\xAD\xBF\xBF\xBF
360 * U+7FFFF: \xF1\xBF\xBF\xBF \xF8\xAF\xBF\xBF\xBF
361 * U+8FFFF: \xF2\x8F\xBF\xBF \xF8\xB1\xBF\xBF\xBF
362 * U+9FFFF: \xF2\x9F\xBF\xBF \xF8\xB3\xBF\xBF\xBF
363 * U+AFFFF: \xF2\xAF\xBF\xBF \xF8\xB5\xBF\xBF\xBF
364 * U+BFFFF: \xF2\xBF\xBF\xBF \xF8\xB7\xBF\xBF\xBF
365 * U+CFFFF: \xF3\x8F\xBF\xBF \xF8\xB9\xBF\xBF\xBF
366 * U+DFFFF: \xF3\x9F\xBF\xBF \xF8\xBB\xBF\xBF\xBF
367 * U+EFFFF: \xF3\xAF\xBF\xBF \xF8\xBD\xBF\xBF\xBF
368 * U+FFFFF: \xF3\xBF\xBF\xBF \xF8\xBF\xBF\xBF\xBF
369 * U+10FFFF: \xF4\x8F\xBF\xBF \xF9\xA1\xBF\xBF\xBF
370 */
371#define UTF8_IS_NONCHAR_(s) ( \
372 *(s) >= UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE \
373 && ! UTF8_IS_SUPER(s) \
374 && UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_FIRST_PROBLEMATIC(s) \
375
376#ifdef EBCDIC /* Both versions assume well-formed UTF8 */
377# define UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s) \
378 ((*(s) == UTF_TO_NATIVE(0xF1) \
379 && (*((s) + 1) == UTF_TO_NATIVE(0xBF) \
380 && ((*((s) + 2) == UTF_TO_NATIVE(0xAE) \
381 && *((s) + 3) >= UTF_TO_NATIVE(0xB0)) \
382 || (*((s) + 2) == UTF_TO_NATIVE(0xAF) \
383 && *((s) + 3) <= UTF_TO_NATIVE(0xAF))))) \
384 || (UTF8SKIP(*(s)) > 3 \
385 /* (These were all derived by inspection and experimentation with an */ \
386 /* editor) The next line checks the next to final byte in the char */ \
387 && *((s) + UTF8SKIP(*(s)) - 2) == UTF_TO_NATIVE(0xBF) \
388 && *((s) + UTF8SKIP(*(s)) - 3) == UTF_TO_NATIVE(0xBF) \
389 && (NATIVE_TO_UTF(*((s) + UTF8SKIP(*(s)) - 4)) & 0x81) == 0x81 \
390 && (NATIVE_TO_UTF(*((s) + UTF8SKIP(*(s)) - 1)) & 0xBE) == 0XBE))
391#else
392# define UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s) \
393 ((*(s) == 0xEF \
394 && ((*((s) + 1) == 0xB7 && (*((s) + 2) >= 0x90 && (*((s) + 2) <= 0xAF)))\
395 /* Gets U+FFF[EF] */ \
396 || (*((s) + 1) == 0xBF && ((*((s) + 2) & 0xBE) == 0xBE)))) \
397 || ((*((s) + 2) == 0xBF \
398 && (*((s) + 3) & 0xBE) == 0xBE \
399 /* Excludes things like U+10FFE = \xF0\x90\xBF\xBE */ \
400 && (*((s) + 1) & 0x8F) == 0x8F)))
401#endif
402
c867b360
JH		 403#define UNICODE_SURROGATE_FIRST 0xD800
404#define UNICODE_SURROGATE_LAST 0xDFFF
405#define UNICODE_REPLACEMENT 0xFFFD
406#define UNICODE_BYTE_ORDER_MARK 0xFEFF
1d72bdf6 407
b851fbc1 408/* Though our UTF-8 encoding can go beyond this,
c76687c5 409 * let's be conservative and do as Unicode says. */
b851fbc1
JH		 410#define PERL_UNICODE_MAX 0x10FFFF
411
949cf498
KW		 412#define UNICODE_WARN_SURROGATE 0x0001 /* UTF-16 surrogates */
413#define UNICODE_WARN_NONCHAR 0x0002 /* Non-char code points */
414#define UNICODE_WARN_SUPER 0x0004 /* Above 0x10FFFF */
415#define UNICODE_WARN_FE_FF 0x0008 /* Above 0x10FFFF */
416#define UNICODE_DISALLOW_SURROGATE 0x0010
417#define UNICODE_DISALLOW_NONCHAR 0x0020
418#define UNICODE_DISALLOW_SUPER 0x0040
419#define UNICODE_DISALLOW_FE_FF 0x0080
420#define UNICODE_WARN_ILLEGAL_INTERCHANGE \
421 (UNICODE_WARN_SURROGATE|UNICODE_WARN_NONCHAR|UNICODE_WARN_SUPER)
422#define UNICODE_DISALLOW_ILLEGAL_INTERCHANGE \
423 (UNICODE_DISALLOW_SURROGATE|UNICODE_DISALLOW_NONCHAR|UNICODE_DISALLOW_SUPER)
424
425/* For backward source compatibility, as are now the default */
426#define UNICODE_ALLOW_SURROGATE 0
427#define UNICODE_ALLOW_SUPER 0
428#define UNICODE_ALLOW_ANY 0
b851fbc1 429
1d72bdf6
NIS		 430#define UNICODE_IS_SURROGATE(c) ((c) >= UNICODE_SURROGATE_FIRST && \
431 (c) <= UNICODE_SURROGATE_LAST)
a10ec373 432#define UNICODE_IS_REPLACEMENT(c) ((c) == UNICODE_REPLACEMENT)
872c91ae 433#define UNICODE_IS_BYTE_ORDER_MARK(c) ((c) == UNICODE_BYTE_ORDER_MARK)
7131f24d
KW		 434#define UNICODE_IS_NONCHAR(c) ((c >= 0xFDD0 && c <= 0xFDEF) \
435 /* The other noncharacters end in FFFE or FFFF, which \
436 * the mask below catches both of, but beyond the last \
437 * official unicode code point, they aren't \
438 * noncharacters, since those aren't Unicode \
439 * characters at all */ \
440 || ((((c & 0xFFFE) == 0xFFFE)) && ! UNICODE_IS_SUPER(c)))
441#define UNICODE_IS_SUPER(c) ((c) > PERL_UNICODE_MAX)
442#define UNICODE_IS_FE_FF(c) ((c) > 0x7FFFFFFF)
1d72bdf6 443
f067b878
NA		 444#ifdef HAS_QUAD
445# define UTF8_QUAD_MAX UINT64_C(0x1000000000)
446#endif
3bd709b1 447
09091399
JH		 448#define UNICODE_GREEK_CAPITAL_LETTER_SIGMA 0x03A3
449#define UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA 0x03C2
450#define UNICODE_GREEK_SMALL_LETTER_SIGMA 0x03C3
9dcbe121 451#define GREEK_SMALL_LETTER_MU 0x03BC
78a0d3cc
KW		 452#define GREEK_CAPITAL_LETTER_MU 0x039C /* Upper and title case of MICRON */
453#define LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS 0x0178 /* Also is title case */
97298f37 454#define LATIN_CAPITAL_LETTER_SHARP_S 0x1E9E
09091399 455
9e55ce06 456#define UNI_DISPLAY_ISPRINT 0x0001
c728cb41
JH		 457#define UNI_DISPLAY_BACKSLASH 0x0002
458#define UNI_DISPLAY_QQ (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)
459#define UNI_DISPLAY_REGEX (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)
9e55ce06 460
5cd46e1f
KW		 461#ifndef EBCDIC
462# define LATIN_SMALL_LETTER_SHARP_S 0x00DF
463# define LATIN_SMALL_LETTER_Y_WITH_DIAERESIS 0x00FF
464# define MICRO_SIGN 0x00B5
f508a607
KW		 465# define LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE 0x00C5
466# define LATIN_SMALL_LETTER_A_WITH_RING_ABOVE 0x00E5
5cd46e1f
KW		 467#endif
468
469#define ANYOF_FOLD_SHARP_S(node, input, end) \
470 (ANYOF_BITMAP_TEST(node, LATIN_SMALL_LETTER_SHARP_S) && \
137165a6 471 (ANYOF_NONBITMAP(node)) && \
39065660 472 (ANYOF_FLAGS(node) & ANYOF_LOC_NONBITMAP_FOLD) && \
07b6858f
JH		 473 ((end) > (input) + 1) && \
474 toLOWER((input)[0]) == 's' && \
475 toLOWER((input)[1]) == 's')
ebc501f0 476#define SHARP_S_SKIP 2
3b0fc154 477
2f454f11 478#ifndef EBCDIC
c78f6c49 479# define IS_UTF8_CHAR_1(p) \
3b0fc154 480 ((p)[0] <= 0x7F)
c78f6c49 481# define IS_UTF8_CHAR_2(p) \
3b0fc154
JH		 482 ((p)[0] >= 0xC2 && (p)[0] <= 0xDF && \
483 (p)[1] >= 0x80 && (p)[1] <= 0xBF)
c78f6c49 484# define IS_UTF8_CHAR_3a(p) \
3b0fc154
JH		 485 ((p)[0] == 0xE0 && \
486 (p)[1] >= 0xA0 && (p)[1] <= 0xBF && \
487 (p)[2] >= 0x80 && (p)[2] <= 0xBF)
c78f6c49 488# define IS_UTF8_CHAR_3b(p) \
3b0fc154
JH		 489 ((p)[0] >= 0xE1 && (p)[0] <= 0xEC && \
490 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
491 (p)[2] >= 0x80 && (p)[2] <= 0xBF)
c78f6c49 492# define IS_UTF8_CHAR_3c(p) \
3b0fc154
JH		 493 ((p)[0] == 0xED && \
494 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
495 (p)[2] >= 0x80 && (p)[2] <= 0xBF)
c78f6c49
KW		 496 /* In IS_UTF8_CHAR_3c(p) one could use
497 * (p)[1] >= 0x80 && (p)[1] <= 0x9F
498 * if one wanted to exclude surrogates. */
499# define IS_UTF8_CHAR_3d(p) \
3b0fc154
JH		 500 ((p)[0] >= 0xEE && (p)[0] <= 0xEF && \
501 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
502 (p)[2] >= 0x80 && (p)[2] <= 0xBF)
c78f6c49 503# define IS_UTF8_CHAR_4a(p) \
3b0fc154
JH		 504 ((p)[0] == 0xF0 && \
505 (p)[1] >= 0x90 && (p)[1] <= 0xBF && \
506 (p)[2] >= 0x80 && (p)[2] <= 0xBF && \
507 (p)[3] >= 0x80 && (p)[3] <= 0xBF)
c78f6c49 508# define IS_UTF8_CHAR_4b(p) \
3b0fc154
JH		 509 ((p)[0] >= 0xF1 && (p)[0] <= 0xF3 && \
510 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
511 (p)[2] >= 0x80 && (p)[2] <= 0xBF && \
512 (p)[3] >= 0x80 && (p)[3] <= 0xBF)
513/* In IS_UTF8_CHAR_4c(p) one could use
514 * (p)[0] == 0xF4
515 * if one wanted to stop at the Unicode limit U+10FFFF.
516 * The 0xF7 allows us to go to 0x1fffff (0x200000 would
517 * require five bytes). Not doing any further code points
518 * since that is not needed (and that would not be strict
519 * UTF-8, anyway). The "slow path" in Perl_is_utf8_char()
520 * will take care of the "extended UTF-8". */
c78f6c49 521# define IS_UTF8_CHAR_4c(p) \
a18d6e6e 522 ((p)[0] >= 0xF4 && (p)[0] <= 0xF7 && \
3b0fc154
JH		 523 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
524 (p)[2] >= 0x80 && (p)[2] <= 0xBF && \
525 (p)[3] >= 0x80 && (p)[3] <= 0xBF)
526
c78f6c49 527# define IS_UTF8_CHAR_3(p) \
3b0fc154
JH		 528 (IS_UTF8_CHAR_3a(p) || \
529 IS_UTF8_CHAR_3b(p) || \
530 IS_UTF8_CHAR_3c(p) || \
531 IS_UTF8_CHAR_3d(p))
c78f6c49 532# define IS_UTF8_CHAR_4(p) \
3b0fc154
JH		 533 (IS_UTF8_CHAR_4a(p) || \
534 IS_UTF8_CHAR_4b(p) || \
535 IS_UTF8_CHAR_4c(p))
536
537/* IS_UTF8_CHAR(p) is strictly speaking wrong (not UTF-8) because it
538 * (1) allows UTF-8 encoded UTF-16 surrogates
539 * (2) it allows code points past U+10FFFF.
540 * The Perl_is_utf8_char() full "slow" code will handle the Perl
541 * "extended UTF-8". */
c78f6c49 542# define IS_UTF8_CHAR(p, n) \
3b0fc154
JH		 543 ((n) == 1 ? IS_UTF8_CHAR_1(p) : \
544 (n) == 2 ? IS_UTF8_CHAR_2(p) : \
545 (n) == 3 ? IS_UTF8_CHAR_3(p) : \
546 (n) == 4 ? IS_UTF8_CHAR_4(p) : 0)
547
c78f6c49 548# define IS_UTF8_CHAR_FAST(n) ((n) <= 4)
768c67ee 549
2f454f11
KW		 550#else /* EBCDIC */
551
552/* This is an attempt to port IS_UTF8_CHAR to EBCDIC based on eyeballing.
553 * untested. If want to exclude surrogates and above-Unicode, see the
554 * definitions for UTF8_IS_SURROGATE and UTF8_IS_SUPER */
555# define IS_UTF8_CHAR_1(p) \
556 (NATIVE_TO_ASCII((p)[0]) <= 0x9F)
557# define IS_UTF8_CHAR_2(p) \
558 (NATIVE_TO_I8((p)[0]) >= 0xC5 && NATIVE_TO_I8((p)[0]) <= 0xDF && \
559 NATIVE_TO_I8((p)[1]) >= 0xA0 && NATIVE_TO_I8((p)[1]) <= 0xBF)
560# define IS_UTF8_CHAR_3(p) \
561 (NATIVE_TO_I8((p)[0]) == 0xE1 && NATIVE_TO_I8((p)[1]) <= 0xEF && \
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# define IS_UTF8_CHAR_4a(p) \
565 (NATIVE_TO_I8((p)[0]) == 0xF0 && \
566 NATIVE_TO_I8((p)[1]) >= 0xB0 && NATIVE_TO_I8((p)[1]) <= 0xBF && \
567 NATIVE_TO_I8((p)[2]) >= 0xA0 && NATIVE_TO_I8((p)[2]) <= 0xBF && \
568 NATIVE_TO_I8((p)[3]) >= 0xA0 && NATIVE_TO_I8((p)[3]) <= 0xBF)
569# define IS_UTF8_CHAR_4b(p) \
570 (NATIVE_TO_I8((p)[0]) >= 0xF1 && NATIVE_TO_I8((p)[0]) <= 0xF7 && \
571 NATIVE_TO_I8((p)[1]) >= 0xA0 && NATIVE_TO_I8((p)[1]) <= 0xBF && \
572 NATIVE_TO_I8((p)[2]) >= 0xA0 && NATIVE_TO_I8((p)[2]) <= 0xBF && \
573 NATIVE_TO_I8((p)[3]) >= 0xA0 && NATIVE_TO_I8((p)[3]) <= 0xBF)
574# define IS_UTF8_CHAR_5a(p) \
575 (NATIVE_TO_I8((p)[0]) == 0xF8 && \
576 NATIVE_TO_I8((p)[1]) >= 0xA8 && NATIVE_TO_I8((p)[1]) <= 0xBF && \
577 NATIVE_TO_I8((p)[1]) >= 0xA0 && NATIVE_TO_I8((p)[1]) <= 0xBF && \
578 NATIVE_TO_I8((p)[2]) >= 0xA0 && NATIVE_TO_I8((p)[2]) <= 0xBF && \
579 NATIVE_TO_I8((p)[3]) >= 0xA0 && NATIVE_TO_I8((p)[3]) <= 0xBF)
580# define IS_UTF8_CHAR_5b(p) \
581 (NATIVE_TO_I8((p)[0]) >= 0xF9 && NATIVE_TO_I8((p)[1]) <= 0xFB && \
582 NATIVE_TO_I8((p)[1]) >= 0xA0 && NATIVE_TO_I8((p)[1]) <= 0xBF && \
583 NATIVE_TO_I8((p)[1]) >= 0xA0 && NATIVE_TO_I8((p)[1]) <= 0xBF && \
584 NATIVE_TO_I8((p)[2]) >= 0xA0 && NATIVE_TO_I8((p)[2]) <= 0xBF && \
585 NATIVE_TO_I8((p)[3]) >= 0xA0 && NATIVE_TO_I8((p)[3]) <= 0xBF)
586
587# define IS_UTF8_CHAR_4(p) \
588 (IS_UTF8_CHAR_4a(p) || \
589 IS_UTF8_CHAR_4b(p))
590# define IS_UTF8_CHAR_5(p) \
591 (IS_UTF8_CHAR_5a(p) || \
592 IS_UTF8_CHAR_5b(p))
593# define IS_UTF8_CHAR(p, n) \
594 ((n) == 1 ? IS_UTF8_CHAR_1(p) : \
595 (n) == 2 ? IS_UTF8_CHAR_2(p) : \
596 (n) == 3 ? IS_UTF8_CHAR_3(p) : \
597 (n) == 4 ? IS_UTF8_CHAR_4(p) : \
598 (n) == 5 ? IS_UTF8_CHAR_5(p) : 0)
599
600# define IS_UTF8_CHAR_FAST(n) ((n) <= 5)
601
77263263 602#endif /* IS_UTF8_CHAR() for UTF-8 */
e9a8c099
MHM		 603
604/*
605 * Local variables:
606 * c-indentation-style: bsd
607 * c-basic-offset: 4
608 * indent-tabs-mode: t
609 * End:
610 *
611 * ex: set ts=8 sts=4 sw=4 noet:
612 */
Unnamed repository; edit this file 'description' to name the repository.