3 * Copyright (C) 2000, 2001, 2002, 2005, 2006, 2007, 2009,
4 * 2010, 2011 by Larry Wall and others
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.
11 #ifndef H_UTF8 /* Guard against recursive inclusion */
14 /* Use UTF-8 as the default script encoding?
15 * Turning this on will break scripts having non-UTF-8 binary
16 * data (such as Latin-1) in string literals. */
17 #ifdef USE_UTF8_SCRIPTS
18 # define USE_UTF8_IN_NAMES (!IN_BYTES)
20 # define USE_UTF8_IN_NAMES (PL_hints & HINT_UTF8)
23 /* For to_utf8_fold_flags, q.v. */
24 #define FOLD_FLAGS_LOCALE 0x1
25 #define FOLD_FLAGS_FULL 0x2
26 #define FOLD_FLAGS_NOMIX_ASCII 0x4
28 /* For _core_swash_init(), internal core use only */
29 #define _CORE_SWASH_INIT_USER_DEFINED_PROPERTY 0x1
30 #define _CORE_SWASH_INIT_RETURN_IF_UNDEF 0x2
31 #define _CORE_SWASH_INIT_ACCEPT_INVLIST 0x4
33 #define to_uni_fold(c, p, lenp) _to_uni_fold_flags(c, p, lenp, FOLD_FLAGS_FULL)
34 #define to_utf8_fold(c, p, lenp) _to_utf8_fold_flags(c, p, lenp, \
35 FOLD_FLAGS_FULL, NULL)
36 #define to_utf8_lower(a,b,c) _to_utf8_lower_flags(a,b,c,0, NULL)
37 #define to_utf8_upper(a,b,c) _to_utf8_upper_flags(a,b,c,0, NULL)
38 #define to_utf8_title(a,b,c) _to_utf8_title_flags(a,b,c,0, NULL)
40 /* Source backward compatibility. */
41 #define uvuni_to_utf8(d, uv) uvuni_to_utf8_flags(d, uv, 0)
42 #define is_utf8_string_loc(s, len, ep) is_utf8_string_loclen(s, len, ep, 0)
44 #define foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \
45 foldEQ_utf8_flags(s1, pe1, l1, u1, s2, pe2, l2, u2, 0)
46 #define FOLDEQ_UTF8_NOMIX_ASCII (1 << 0)
47 #define FOLDEQ_UTF8_LOCALE (1 << 1)
48 #define FOLDEQ_S1_ALREADY_FOLDED (1 << 2)
49 #define FOLDEQ_S2_ALREADY_FOLDED (1 << 3)
52 =for apidoc ibcmp_utf8
54 This is a synonym for (! foldEQ_utf8())
58 #define ibcmp_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \
59 cBOOL(! foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2))
62 /* The equivalent of these macros but implementing UTF-EBCDIC
63 are in the following header file:
66 #include "utfebcdic.h"
72 EXTCONST unsigned char PL_utf8skip[] = {
73 /* 0x00 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
74 /* 0x10 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
75 /* 0x20 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
76 /* 0x30 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
77 /* 0x40 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
78 /* 0x50 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
79 /* 0x60 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
80 /* 0x70 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
81 /* 0x80 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
82 /* 0x90 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
83 /* 0xA0 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
84 /* 0xB0 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
85 /* 0xC0 */ 2,2, /* overlong */
86 /* 0xC2 */ 2,2,2,2,2,2,2,2,2,2,2,2,2,2, /* U+0080 to U+03FF */
87 /* 0xD0 */ 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, /* U+0400 to U+07FF */
88 /* 0xE0 */ 3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3, /* U+0800 to U+FFFF */
89 /* 0xF0 */ 4,4,4,4,4,4,4,4,5,5,5,5,6,6, /* above BMP to 2**31 - 1 */
90 /* 0xFE */ 7,13, /* Perl extended (never was official UTF-8). Up to 72bit
91 allowed (64-bit + reserved). */
94 EXTCONST unsigned char PL_utf8skip[];
99 #include "regcharclass.h"
100 #include "unicode_constants.h"
102 /* Native character to iso-8859-1 */
103 #define NATIVE_TO_ASCII(ch) (ch)
104 #define ASCII_TO_NATIVE(ch) (ch)
105 /* Transform after encoding */
106 #define NATIVE_TO_UTF(ch) (ch)
107 #define NATIVE_TO_I8(ch) NATIVE_TO_UTF(ch) /* a clearer synonym */
108 #define UTF_TO_NATIVE(ch) (ch)
109 #define I8_TO_NATIVE(ch) UTF_TO_NATIVE(ch)
110 /* Transforms in wide UV chars */
111 #define UNI_TO_NATIVE(ch) (ch)
112 #define NATIVE_TO_UNI(ch) (ch)
113 /* Transforms in invariant space */
114 #define NATIVE_TO_NEED(enc,ch) (ch)
115 #define ASCII_TO_NEED(enc,ch) (ch)
117 /* As there are no translations, avoid the function wrapper */
118 #define utf8n_to_uvchr utf8n_to_uvuni
119 #define valid_utf8_to_uvchr valid_utf8_to_uvuni
120 #define uvchr_to_utf8 uvuni_to_utf8
124 The following table is from Unicode 3.2.
126 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
128 U+0000..U+007F 00..7F
129 U+0080..U+07FF * C2..DF 80..BF
130 U+0800..U+0FFF E0 * A0..BF 80..BF
131 U+1000..U+CFFF E1..EC 80..BF 80..BF
132 U+D000..U+D7FF ED 80..9F 80..BF
133 U+D800..U+DFFF +++++++ utf16 surrogates, not legal utf8 +++++++
134 U+E000..U+FFFF EE..EF 80..BF 80..BF
135 U+10000..U+3FFFF F0 * 90..BF 80..BF 80..BF
136 U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF
137 U+100000..U+10FFFF F4 80..8F 80..BF 80..BF
138 Below are non-Unicode code points
139 U+110000..U+13FFFF F4 90..BF 80..BF 80..BF
140 U+110000..U+1FFFFF F5..F7 80..BF 80..BF 80..BF
141 U+200000: F8.. * 88..BF 80..BF 80..BF 80..BF
143 Note the gaps before several of the byte entries above marked by '*'. These are
144 caused by legal UTF-8 avoiding non-shortest encodings: it is technically
145 possible to UTF-8-encode a single code point in different ways, but that is
146 explicitly forbidden, and the shortest possible encoding should always be used
147 (and that is what Perl does).
152 Another way to look at it, as bits:
154 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
157 0000 0bbb bbaa aaaa 110b bbbb 10aa aaaa
158 cccc bbbb bbaa aaaa 1110 cccc 10bb bbbb 10aa aaaa
159 00 000d ddcc cccc bbbb bbaa aaaa 1111 0ddd 10cc cccc 10bb bbbb 10aa aaaa
161 As you can see, the continuation bytes all begin with C<10>, and the
162 leading bits of the start byte tell how many bytes there are in the
165 Perl's extended UTF-8 means we can have start bytes up to FF.
169 #define UNI_IS_INVARIANT(c) (((UV)c) < 0x80)
170 #define UTF8_IS_START(c) (((U8)c) >= 0xc2)
171 #define UTF8_IS_CONTINUATION(c) ((((U8)c) & 0xC0) == 0x80)
172 #define UTF8_IS_CONTINUED(c) (((U8)c) & 0x80)
174 /* Masking with 0xfe allows low bit to be 0 or 1; thus this matches 0xc[23] */
175 #define UTF8_IS_DOWNGRADEABLE_START(c) (((U8)c & 0xfe) == 0xc2)
176 #define UTF8_IS_ABOVE_LATIN1(c) ((U8)(c) >= 0xc4)
178 #define UTF_START_MARK(len) (((len) > 7) ? 0xFF : (0xFE << (7-(len))))
180 /* Masks out the initial one bits in a start byte, leaving the real data ones.
181 * Doesn't work on an invariant byte */
182 #define UTF_START_MASK(len) (((len) >= 7) ? 0x00 : (0x1F >> ((len)-2)))
184 #define UTF_CONTINUATION_MARK 0x80
185 #define UTF_ACCUMULATION_SHIFT 6
187 /* 2**UTF_ACCUMULATION_SHIFT - 1 */
188 #define UTF_CONTINUATION_MASK ((U8)0x3f)
190 /* This sets the UTF_CONTINUATION_MASK in the upper bits of a word. If a value
191 * is anded with it, and the result is non-zero, then using the original value
192 * in UTF8_ACCUMULATE will overflow, shifting bits off the left */
193 #define UTF_ACCUMULATION_OVERFLOW_MASK \
194 (((UV) UTF_CONTINUATION_MASK) << ((sizeof(UV) * CHARBITS) \
195 - UTF_ACCUMULATION_SHIFT))
198 #define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \
200 (uv) < 0x10000 ? 3 : \
201 (uv) < 0x200000 ? 4 : \
202 (uv) < 0x4000000 ? 5 : \
203 (uv) < 0x80000000 ? 6 : \
204 (uv) < UTF8_QUAD_MAX ? 7 : 13 )
206 /* No, I'm not even going to *TRY* putting #ifdef inside a #define */
207 #define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \
209 (uv) < 0x10000 ? 3 : \
210 (uv) < 0x200000 ? 4 : \
211 (uv) < 0x4000000 ? 5 : \
212 (uv) < 0x80000000 ? 6 : 7 )
215 /* How wide can a single UTF-8 encoded character become in bytes. */
216 /* NOTE: Strictly speaking Perl's UTF-8 should not be called UTF-8 since UTF-8
217 * is an encoding of Unicode, and Unicode's upper limit, 0x10FFFF, can be
218 * expressed with 4 bytes. However, Perl thinks of UTF-8 as a way to encode
219 * non-negative integers in a binary format, even those above Unicode */
220 #define UTF8_MAXBYTES 13
222 /* The maximum number of UTF-8 bytes a single Unicode character can
223 * uppercase/lowercase/fold into. Unicode guarantees that the maximum
224 * expansion is 3 characters. On ASCIIish platforms, the highest Unicode
225 * character occupies 4 bytes, therefore this number would be 12, but this is
226 * smaller than the maximum width a single above-Unicode character can occupy,
227 * so use that instead */
228 #if UTF8_MAXBYTES < 12
229 #error UTF8_MAXBYTES must be at least 12
232 #define UTF8_MAXBYTES_CASE UTF8_MAXBYTES
234 #endif /* EBCDIC vs ASCII */
236 /* Rest of these are attributes of Unicode and perl's internals rather than the
237 * encoding, or happen to be the same in both ASCII and EBCDIC (at least at
238 * this level; the macros that some of these call may have different
239 * definitions in the two encodings */
241 #define NATIVE8_TO_UNI(ch) NATIVE_TO_ASCII(ch) /* a clearer synonym */
243 #define UTF8_ACCUMULATE(old, new) (((old) << UTF_ACCUMULATION_SHIFT) \
244 | (((U8)new) & UTF_CONTINUATION_MASK))
246 /* Convert a two (not one) byte utf8 character to a unicode code point value.
247 * Needs just one iteration of accumulate. Should not be used unless it is
248 * known that the two bytes are legal: 1) two-byte start, and 2) continuation.
249 * Note that the result can be larger than 255 if the input character is not
251 #define TWO_BYTE_UTF8_TO_UNI(HI, LO) \
252 UTF8_ACCUMULATE((NATIVE_TO_UTF(HI) & UTF_START_MASK(2)), \
255 #define UTF8SKIP(s) PL_utf8skip[*(const U8*)(s)]
257 #define UTF8_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE_TO_UTF(c))
258 #define NATIVE_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE8_TO_UNI(c))
260 #define MAX_PORTABLE_UTF8_TWO_BYTE 0x3FF /* constrained by EBCDIC */
262 /* The macros in the next sets are used to generate the two utf8 or utfebcdic
263 * bytes from an ordinal that is known to fit into two bytes; it must be less
264 * than 0x3FF to work across both encodings. */
265 /* Nocast allows these to be used in the case label of a switch statement */
266 #define UTF8_TWO_BYTE_HI_nocast(c) NATIVE_TO_I8(((c) \
267 >> UTF_ACCUMULATION_SHIFT) | (0xFF & UTF_START_MARK(2)))
268 #define UTF8_TWO_BYTE_LO_nocast(c) NATIVE_TO_I8(((c) & UTF_CONTINUATION_MASK) \
269 | UTF_CONTINUATION_MARK)
271 #define UTF8_TWO_BYTE_HI(c) ((U8) (UTF8_TWO_BYTE_HI_nocast(c)))
272 #define UTF8_TWO_BYTE_LO(c) ((U8) (UTF8_TWO_BYTE_LO_nocast(c)))
274 /* This name is used when the source is a single byte */
275 #define UTF8_EIGHT_BIT_HI(c) UTF8_TWO_BYTE_HI((U8)(c))
276 #define UTF8_EIGHT_BIT_LO(c) UTF8_TWO_BYTE_LO((U8)(c))
279 * 'UTF' is whether or not p is encoded in UTF8. The names 'foo_lazy_if' stem
280 * from an earlier version of these macros in which they didn't call the
281 * foo_utf8() macros (i.e. were 'lazy') unless they decided that *p is the
282 * beginning of a utf8 character. Now that foo_utf8() determines that itself,
283 * no need to do it again here
285 #define isIDFIRST_lazy_if(p,UTF) ((IN_BYTES || !UTF ) \
287 : isIDFIRST_utf8((const U8*)p))
288 #define isALNUM_lazy_if(p,UTF) ((IN_BYTES || (!UTF )) \
290 : isALNUM_utf8((const U8*)p))
292 #define isIDFIRST_lazy(p) isIDFIRST_lazy_if(p,1)
293 #define isALNUM_lazy(p) isALNUM_lazy_if(p,1)
295 #define UTF8_MAXLEN UTF8_MAXBYTES
297 /* A Unicode character can fold to up to 3 characters */
298 #define UTF8_MAX_FOLD_CHAR_EXPAND 3
300 #define IN_BYTES (CopHINTS_get(PL_curcop) & HINT_BYTES)
301 #define DO_UTF8(sv) (SvUTF8(sv) && !IN_BYTES)
302 #define IN_UNI_8_BIT \
303 (CopHINTS_get(PL_curcop) & (HINT_UNI_8_BIT|HINT_LOCALE_NOT_CHARS) \
304 && ! IN_LOCALE_RUNTIME && ! IN_BYTES)
307 #define UTF8_ALLOW_EMPTY 0x0001 /* Allow a zero length string */
309 /* Allow first byte to be a continuation byte */
310 #define UTF8_ALLOW_CONTINUATION 0x0002
312 /* Allow second... bytes to be non-continuation bytes */
313 #define UTF8_ALLOW_NON_CONTINUATION 0x0004
315 /* expecting more bytes than were available in the string */
316 #define UTF8_ALLOW_SHORT 0x0008
318 /* Overlong sequence; i.e., the code point can be specified in fewer bytes. */
319 #define UTF8_ALLOW_LONG 0x0010
321 #define UTF8_DISALLOW_SURROGATE 0x0020 /* Unicode surrogates */
322 #define UTF8_WARN_SURROGATE 0x0040
324 #define UTF8_DISALLOW_NONCHAR 0x0080 /* Unicode non-character */
325 #define UTF8_WARN_NONCHAR 0x0100 /* code points */
327 #define UTF8_DISALLOW_SUPER 0x0200 /* Super-set of Unicode: code */
328 #define UTF8_WARN_SUPER 0x0400 /* points above the legal max */
330 /* Code points which never were part of the original UTF-8 standard, the first
331 * byte of which is a FE or FF on ASCII platforms. */
332 #define UTF8_DISALLOW_FE_FF 0x0800
333 #define UTF8_WARN_FE_FF 0x1000
335 #define UTF8_CHECK_ONLY 0x2000
337 /* For backwards source compatibility. They do nothing, as the default now
338 * includes what they used to mean. The first one's meaning was to allow the
339 * just the single non-character 0xFFFF */
340 #define UTF8_ALLOW_FFFF 0
341 #define UTF8_ALLOW_SURROGATE 0
343 #define UTF8_DISALLOW_ILLEGAL_INTERCHANGE (UTF8_DISALLOW_SUPER|UTF8_DISALLOW_NONCHAR|UTF8_DISALLOW_SURROGATE|UTF8_DISALLOW_FE_FF)
344 #define UTF8_WARN_ILLEGAL_INTERCHANGE \
345 (UTF8_WARN_SUPER|UTF8_WARN_NONCHAR|UTF8_WARN_SURROGATE|UTF8_WARN_FE_FF)
346 #define UTF8_ALLOW_ANY \
347 (~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE|UTF8_WARN_ILLEGAL_INTERCHANGE))
348 #define UTF8_ALLOW_ANYUV \
350 & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE|UTF8_WARN_ILLEGAL_INTERCHANGE))
351 #define UTF8_ALLOW_DEFAULT (ckWARN(WARN_UTF8) ? 0 : \
354 /* Surrogates, non-character code points and above-Unicode code points are
355 * problematic in some contexts. This allows code that needs to check for
356 * those to to quickly exclude the vast majority of code points it will
358 #define UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE \
359 FIRST_SURROGATE_UTF8_FIRST_BYTE
361 #define UTF8_IS_SURROGATE(s) cBOOL(is_SURROGATE_utf8(s))
362 #define UTF8_IS_REPLACEMENT(s, send) cBOOL(is_REPLACEMENT_utf8_safe(s,send))
365 * U+10FFFF: \xF4\x8F\xBF\xBF \xF9\xA1\xBF\xBF\xBF max legal Unicode
366 * U+110000: \xF4\x90\x80\x80 \xF9\xA2\xA0\xA0\xA0
367 * U+110001: \xF4\x90\x80\x81 \xF9\xA2\xA0\xA0\xA1
369 #ifdef EBCDIC /* Both versions assume well-formed UTF8 */
370 # define UTF8_IS_SUPER(s) (NATIVE_TO_I8(*(s)) >= 0xF9 \
371 && (NATIVE_TO_I8(*(s)) > 0xF9) || (NATIVE_TO_I8(*((s)) + 1 >= 0xA2)))
373 # define UTF8_IS_SUPER(s) (*(s) >= 0xF4 \
374 && (*(s) > 0xF4 || (*((s) + 1) >= 0x90)))
377 /* These are now machine generated, and the 'given' clause is no longer
379 #define UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s) \
380 cBOOL(is_NONCHAR_utf8(s))
381 #define UTF8_IS_NONCHAR_(s) \
382 UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s)
384 #define UNICODE_SURROGATE_FIRST 0xD800
385 #define UNICODE_SURROGATE_LAST 0xDFFF
386 #define UNICODE_REPLACEMENT 0xFFFD
387 #define UNICODE_BYTE_ORDER_MARK 0xFEFF
389 /* Though our UTF-8 encoding can go beyond this,
390 * let's be conservative and do as Unicode says. */
391 #define PERL_UNICODE_MAX 0x10FFFF
393 #define UNICODE_WARN_SURROGATE 0x0001 /* UTF-16 surrogates */
394 #define UNICODE_WARN_NONCHAR 0x0002 /* Non-char code points */
395 #define UNICODE_WARN_SUPER 0x0004 /* Above 0x10FFFF */
396 #define UNICODE_WARN_FE_FF 0x0008 /* Above 0x10FFFF */
397 #define UNICODE_DISALLOW_SURROGATE 0x0010
398 #define UNICODE_DISALLOW_NONCHAR 0x0020
399 #define UNICODE_DISALLOW_SUPER 0x0040
400 #define UNICODE_DISALLOW_FE_FF 0x0080
401 #define UNICODE_WARN_ILLEGAL_INTERCHANGE \
402 (UNICODE_WARN_SURROGATE|UNICODE_WARN_NONCHAR|UNICODE_WARN_SUPER)
403 #define UNICODE_DISALLOW_ILLEGAL_INTERCHANGE \
404 (UNICODE_DISALLOW_SURROGATE|UNICODE_DISALLOW_NONCHAR|UNICODE_DISALLOW_SUPER)
406 /* For backward source compatibility, as are now the default */
407 #define UNICODE_ALLOW_SURROGATE 0
408 #define UNICODE_ALLOW_SUPER 0
409 #define UNICODE_ALLOW_ANY 0
411 #define UNICODE_IS_SURROGATE(c) ((c) >= UNICODE_SURROGATE_FIRST && \
412 (c) <= UNICODE_SURROGATE_LAST)
413 #define UNICODE_IS_REPLACEMENT(c) ((c) == UNICODE_REPLACEMENT)
414 #define UNICODE_IS_BYTE_ORDER_MARK(c) ((c) == UNICODE_BYTE_ORDER_MARK)
415 #define UNICODE_IS_NONCHAR(c) ((c >= 0xFDD0 && c <= 0xFDEF) \
416 /* The other noncharacters end in FFFE or FFFF, which \
417 * the mask below catches both of, but beyond the last \
418 * official unicode code point, they aren't \
419 * noncharacters, since those aren't Unicode \
420 * characters at all */ \
421 || ((((c & 0xFFFE) == 0xFFFE)) && ! UNICODE_IS_SUPER(c)))
422 #define UNICODE_IS_SUPER(c) ((c) > PERL_UNICODE_MAX)
423 #define UNICODE_IS_FE_FF(c) ((c) > 0x7FFFFFFF)
426 # define UTF8_QUAD_MAX UINT64_C(0x1000000000)
429 #define LATIN_SMALL_LETTER_SHARP_S LATIN_SMALL_LETTER_SHARP_S_NATIVE
430 #define LATIN_SMALL_LETTER_Y_WITH_DIAERESIS \
431 LATIN_SMALL_LETTER_Y_WITH_DIAERESIS_NATIVE
432 #define MICRO_SIGN MICRO_SIGN_NATIVE
433 #define LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE \
434 LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE_NATIVE
435 #define LATIN_SMALL_LETTER_A_WITH_RING_ABOVE \
436 LATIN_SMALL_LETTER_A_WITH_RING_ABOVE_NATIVE
437 #define UNICODE_GREEK_CAPITAL_LETTER_SIGMA 0x03A3
438 #define UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA 0x03C2
439 #define UNICODE_GREEK_SMALL_LETTER_SIGMA 0x03C3
440 #define GREEK_SMALL_LETTER_MU 0x03BC
441 #define GREEK_CAPITAL_LETTER_MU 0x039C /* Upper and title case of MICRON */
442 #define LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS 0x0178 /* Also is title case */
443 #define LATIN_CAPITAL_LETTER_SHARP_S 0x1E9E
444 #define LATIN_SMALL_LETTER_LONG_S 0x017F
445 #define KELVIN_SIGN 0x212A
446 #define ANGSTROM_SIGN 0x212B
448 #define UNI_DISPLAY_ISPRINT 0x0001
449 #define UNI_DISPLAY_BACKSLASH 0x0002
450 #define UNI_DISPLAY_QQ (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)
451 #define UNI_DISPLAY_REGEX (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)
453 #define ANYOF_FOLD_SHARP_S(node, input, end) \
454 (ANYOF_BITMAP_TEST(node, LATIN_SMALL_LETTER_SHARP_S) && \
455 (ANYOF_NONBITMAP(node)) && \
456 (ANYOF_FLAGS(node) & ANYOF_LOC_NONBITMAP_FOLD) && \
457 ((end) > (input) + 1) && \
458 toLOWER((input)[0]) == 's' && \
459 toLOWER((input)[1]) == 's')
460 #define SHARP_S_SKIP 2
462 /* If you want to exclude surrogates, and beyond legal Unicode, see the blame
463 * log for earlier versions which gave details for these */
466 /* This was generated by regen/regcharclass.pl, and then moved here. The lines
467 * that generated it were then commented out. This was done solely because it
468 * takes on the order of 10 minutes to generate, and is never going to change.
469 * The EBCDIC equivalent hasn't been commented out in regcharclass.pl, so it
470 * should generate and run the correct stuff */
471 /*** GENERATED CODE ***/
472 #define is_UTF8_CHAR_utf8_safe(s,e) \
474 ( ( ( ((U8*)s)[0] & 0x80 ) == 0x00 ) ? 1 \
475 : ( 0xC2 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xDF ) ? \
476 ( ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) ? 2 : 0 ) \
477 : ( 0xE0 == ((U8*)s)[0] ) ? \
478 ( ( ( ( ((U8*)s)[1] & 0xE0 ) == 0xA0 ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
479 : ( 0xE1 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xEF ) ? \
480 ( ( ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
481 : ( 0xF0 == ((U8*)s)[0] ) ? \
482 ( ( ( ( 0x90 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0xBF ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\
483 : ( 0xF1 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xF7 ) ? \
484 ( ( ( ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\
487 ( ( ( ((U8*)s)[0] & 0x80 ) == 0x00 ) ? 1 \
488 : ( 0xC2 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xDF ) ? \
489 ( ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) ? 2 : 0 ) \
490 : ( 0xE0 == ((U8*)s)[0] ) ? \
491 ( ( ( ( ((U8*)s)[1] & 0xE0 ) == 0xA0 ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
492 : ( 0xE1 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xEF ) ? \
493 ( ( ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
496 ( ( ( ((U8*)s)[0] & 0x80 ) == 0x00 ) ? 1 \
497 : ( 0xC2 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xDF ) ? \
498 ( ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) ? 2 : 0 ) \
501 ( ( ((U8*)s)[0] & 0x80 ) == 0x00 ) \
505 /* IS_UTF8_CHAR(p) is strictly speaking wrong (not UTF-8) because it
506 * (1) allows UTF-8 encoded UTF-16 surrogates
507 * (2) it allows code points past U+10FFFF.
508 * The Perl_is_utf8_char() full "slow" code will handle the Perl
509 * "extended UTF-8". */
510 #define IS_UTF8_CHAR(p, n) (is_UTF8_CHAR_utf8_safe(p, (p) + (n)) == n)
512 /* regen/regcharclass.pl generates is_UTF8_CHAR_utf8_safe() macros for up to
513 * these number of bytes. So this has to be coordinated with it */
515 # define IS_UTF8_CHAR_FAST(n) ((n) <= 5)
517 # define IS_UTF8_CHAR_FAST(n) ((n) <= 4)
524 * c-indentation-style: bsd
526 * indent-tabs-mode: nil
529 * ex: set ts=8 sts=4 sw=4 et: