xref: /openbsd/gnu/usr.bin/perl/utf8.h (revision 5af055cd)
1 /*    utf8.h
2  *
3  * This file contains definitions for use with the UTF-8 encoding.  It
4  * actually also works with the variant UTF-8 encoding called UTF-EBCDIC, and
5  * hides almost all of the differences between these from the caller.  In other
6  * words, someone should #include this file, and if the code is being compiled
7  * on an EBCDIC platform, things should mostly just work.
8  *
9  *    Copyright (C) 2000, 2001, 2002, 2005, 2006, 2007, 2009,
10  *    2010, 2011 by Larry Wall and others
11  *
12  *    You may distribute under the terms of either the GNU General Public
13  *    License or the Artistic License, as specified in the README file.
14  *
15  */
16 
17 #ifndef H_UTF8      /* Guard against recursive inclusion */
18 #define H_UTF8 1
19 
20 /* Use UTF-8 as the default script encoding?
21  * Turning this on will break scripts having non-UTF-8 binary
22  * data (such as Latin-1) in string literals. */
23 #ifdef USE_UTF8_SCRIPTS
24 #    define USE_UTF8_IN_NAMES (!IN_BYTES)
25 #else
26 #    define USE_UTF8_IN_NAMES (PL_hints & HINT_UTF8)
27 #endif
28 
29 #include "regcharclass.h"
30 #include "unicode_constants.h"
31 
32 /* For to_utf8_fold_flags, q.v. */
33 #define FOLD_FLAGS_LOCALE 0x1
34 #define FOLD_FLAGS_FULL   0x2
35 #define FOLD_FLAGS_NOMIX_ASCII 0x4
36 
37 /* For _core_swash_init(), internal core use only */
38 #define _CORE_SWASH_INIT_USER_DEFINED_PROPERTY 0x1
39 #define _CORE_SWASH_INIT_RETURN_IF_UNDEF       0x2
40 #define _CORE_SWASH_INIT_ACCEPT_INVLIST        0x4
41 
42 #define uvchr_to_utf8(a,b)          uvchr_to_utf8_flags(a,b,0)
43 #define uvchr_to_utf8_flags(d,uv,flags)                                        \
44                             uvoffuni_to_utf8_flags(d,NATIVE_TO_UNI(uv),flags)
45 #define utf8_to_uvchr_buf(s, e, lenp)                                          \
46                      utf8n_to_uvchr(s, (U8*)(e) - (U8*)(s), lenp,              \
47                                     ckWARN_d(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY)
48 
49 #define to_uni_fold(c, p, lenp) _to_uni_fold_flags(c, p, lenp, FOLD_FLAGS_FULL)
50 #define to_utf8_fold(c, p, lenp) _to_utf8_fold_flags(c, p, lenp, FOLD_FLAGS_FULL)
51 #define to_utf8_lower(a,b,c) _to_utf8_lower_flags(a,b,c,0)
52 #define to_utf8_upper(a,b,c) _to_utf8_upper_flags(a,b,c,0)
53 #define to_utf8_title(a,b,c) _to_utf8_title_flags(a,b,c,0)
54 
55 /* Source backward compatibility. */
56 #define is_utf8_string_loc(s, len, ep)	is_utf8_string_loclen(s, len, ep, 0)
57 
58 #define foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \
59 		    foldEQ_utf8_flags(s1, pe1, l1, u1, s2, pe2, l2, u2, 0)
60 #define FOLDEQ_UTF8_NOMIX_ASCII   (1 << 0)
61 #define FOLDEQ_LOCALE             (1 << 1)
62 #define FOLDEQ_S1_ALREADY_FOLDED  (1 << 2)
63 #define FOLDEQ_S2_ALREADY_FOLDED  (1 << 3)
64 
65 /*
66 =for apidoc ibcmp_utf8
67 
68 This is a synonym for (! foldEQ_utf8())
69 
70 =cut
71 */
72 #define ibcmp_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \
73 		    cBOOL(! foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2))
74 
75 #ifdef EBCDIC
76 /* The equivalent of these macros but implementing UTF-EBCDIC
77    are in the following header file:
78  */
79 
80 #include "utfebcdic.h"
81 
82 #else	/* ! EBCDIC */
83 START_EXTERN_C
84 
85 #ifdef DOINIT
86 EXTCONST unsigned char PL_utf8skip[] = {
87 /* 0x00 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
88 /* 0x10 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
89 /* 0x20 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
90 /* 0x30 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
91 /* 0x40 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
92 /* 0x50 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
93 /* 0x60 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
94 /* 0x70 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
95 /* 0x80 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
96 /* 0x90 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
97 /* 0xA0 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
98 /* 0xB0 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
99 /* 0xC0 */ 2,2,				    /* overlong */
100 /* 0xC2 */ 2,2,2,2,2,2,2,2,2,2,2,2,2,2,     /* U+0080 to U+03FF */
101 /* 0xD0 */ 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, /* U+0400 to U+07FF */
102 /* 0xE0 */ 3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3, /* U+0800 to U+FFFF */
103 /* 0xF0 */ 4,4,4,4,4,4,4,4,5,5,5,5,6,6,	    /* above BMP to 2**31 - 1 */
104 /* 0xFE */ 7,13, /* Perl extended (never was official UTF-8).  Up to 72bit
105 		    allowed (64-bit + reserved). */
106 };
107 #else
108 EXTCONST unsigned char PL_utf8skip[];
109 #endif
110 
111 END_EXTERN_C
112 
113 /* Native character to/from iso-8859-1.  Are the identity functions on ASCII
114  * platforms */
115 #define NATIVE_TO_LATIN1(ch)     (ch)
116 #define LATIN1_TO_NATIVE(ch)     (ch)
117 
118 /* I8 is an intermediate version of UTF-8 used only in UTF-EBCDIC.  We thus
119  * consider it to be identical to UTF-8 on ASCII platforms.  Strictly speaking
120  * UTF-8 and UTF-EBCDIC are two different things, but we often conflate them
121  * because they are 8-bit encodings that serve the same purpose in Perl, and
122  * rarely do we need to distinguish them.  The term "NATIVE_UTF8" applies to
123  * whichever one is applicable on the current platform */
124 #define NATIVE_UTF8_TO_I8(ch) (ch)
125 #define I8_TO_NATIVE_UTF8(ch) (ch)
126 
127 /* Transforms in wide UV chars */
128 #define UNI_TO_NATIVE(ch)        (ch)
129 #define NATIVE_TO_UNI(ch)        (ch)
130 
131 /*
132 
133  The following table is from Unicode 3.2.
134 
135  Code Points		1st Byte  2nd Byte  3rd Byte  4th Byte
136 
137    U+0000..U+007F	00..7F
138    U+0080..U+07FF     * C2..DF    80..BF
139    U+0800..U+0FFF	E0      * A0..BF    80..BF
140    U+1000..U+CFFF       E1..EC    80..BF    80..BF
141    U+D000..U+D7FF       ED        80..9F    80..BF
142    U+D800..U+DFFF       ED        A0..BF    80..BF  (surrogates)
143    U+E000..U+FFFF       EE..EF    80..BF    80..BF
144   U+10000..U+3FFFF	F0      * 90..BF    80..BF    80..BF
145   U+40000..U+FFFFF	F1..F3    80..BF    80..BF    80..BF
146  U+100000..U+10FFFF	F4        80..8F    80..BF    80..BF
147     Below are non-Unicode code points
148  U+110000..U+13FFFF	F4        90..BF    80..BF    80..BF
149  U+110000..U+1FFFFF	F5..F7    80..BF    80..BF    80..BF
150  U+200000..:            F8..    * 88..BF    80..BF    80..BF    80..BF
151 
152 Note the gaps before several of the byte entries above marked by '*'.  These are
153 caused by legal UTF-8 avoiding non-shortest encodings: it is technically
154 possible to UTF-8-encode a single code point in different ways, but that is
155 explicitly forbidden, and the shortest possible encoding should always be used
156 (and that is what Perl does).  The non-shortest ones are called 'overlongs'.
157 
158  */
159 
160 /*
161  Another way to look at it, as bits:
162 
163                   Code Points      1st Byte   2nd Byte   3rd Byte   4th Byte
164 
165                         0aaa aaaa  0aaa aaaa
166               0000 0bbb bbaa aaaa  110b bbbb  10aa aaaa
167               cccc bbbb bbaa aaaa  1110 cccc  10bb bbbb  10aa aaaa
168  00 000d ddcc cccc bbbb bbaa aaaa  1111 0ddd  10cc cccc  10bb bbbb  10aa aaaa
169 
170 As you can see, the continuation bytes all begin with C<10>, and the
171 leading bits of the start byte tell how many bytes there are in the
172 encoded character.
173 
174 Perl's extended UTF-8 means we can have start bytes up to FF.
175 
176 */
177 
178 /* Is the representation of the Unicode code point 'c' the same regardless of
179  * being encoded in UTF-8 or not? */
180 #define UNI_IS_INVARIANT(c)		(((UV)c) <  0x80)
181 
182 /* Is the UTF8-encoded byte 'c' part of a variant sequence in UTF-8?  This is
183  * the inverse of UTF8_IS_INVARIANT */
184 #define UTF8_IS_CONTINUED(c) 		(((U8)c) &  0x80)
185 
186 /* Is the byte 'c' the first byte of a multi-byte UTF8-8 encoded sequence?
187  * This doesn't catch invariants (they are single-byte).  It also excludes the
188  * illegal overlong sequences that begin with C0 and C1. */
189 #define UTF8_IS_START(c)		(((U8)c) >= 0xc2)
190 
191 /* Is the byte 'c' part of a multi-byte UTF8-8 encoded sequence, and not the
192  * first byte thereof?  */
193 #define UTF8_IS_CONTINUATION(c)		((((U8)c) & 0xC0) == 0x80)
194 
195 /* Is the UTF8-encoded byte 'c' the first byte of a two byte sequence?  Use
196  * UTF8_IS_NEXT_CHAR_DOWNGRADEABLE() instead if the input isn't known to
197  * be well-formed.  Masking with 0xfe allows the low bit to be 0 or 1; thus
198  * this matches 0xc[23]. */
199 #define UTF8_IS_DOWNGRADEABLE_START(c)	(((U8)(c) & 0xfe) == 0xc2)
200 
201 /* Is the UTF8-encoded byte 'c' the first byte of a sequence of bytes that
202  * represent a code point > 255? */
203 #define UTF8_IS_ABOVE_LATIN1(c)	((U8)(c) >= 0xc4)
204 
205 /* This defines the 1-bits that are to be in the first byte of a multi-byte
206  * UTF-8 encoded character that give the number of bytes that comprise the
207  * character. 'len' is the number of bytes in the multi-byte sequence. */
208 #define UTF_START_MARK(len) (((len) >  7) ? 0xFF : (0xFF & (0xFE << (7-(len)))))
209 
210 /* Masks out the initial one bits in a start byte, leaving the real data ones.
211  * Doesn't work on an invariant byte.  'len' is the number of bytes in the
212  * multi-byte sequence that comprises the character. */
213 #define UTF_START_MASK(len) (((len) >= 7) ? 0x00 : (0x1F >> ((len)-2)))
214 
215 /* This defines the bits that are to be in the continuation bytes of a multi-byte
216  * UTF-8 encoded character that indicate it is a continuation byte. */
217 #define UTF_CONTINUATION_MARK		0x80
218 
219 /* This is the number of low-order bits a continuation byte in a UTF-8 encoded
220  * sequence contributes to the specification of the code point.  In the bit
221  * maps above, you see that the first 2 bits are a constant '10', leaving 6 of
222  * real information */
223 #define UTF_ACCUMULATION_SHIFT		6
224 
225 /* 2**UTF_ACCUMULATION_SHIFT - 1 */
226 #define UTF_CONTINUATION_MASK		((U8)0x3f)
227 
228 /* If a value is anded with this, and the result is non-zero, then using the
229  * original value in UTF8_ACCUMULATE will overflow, shifting bits off the left
230  * */
231 #define UTF_ACCUMULATION_OVERFLOW_MASK					\
232     (((UV) UTF_CONTINUATION_MASK) << ((sizeof(UV) * CHARBITS)           \
233            - UTF_ACCUMULATION_SHIFT))
234 
235 #if UVSIZE >= 8
236 #  define UTF8_QUAD_MAX UINT64_C(0x1000000000)
237 
238 /* Input is a true Unicode (not-native) code point */
239 #define OFFUNISKIP(uv) ( (uv) < 0x80        ? 1 : \
240 		      (uv) < 0x800          ? 2 : \
241 		      (uv) < 0x10000        ? 3 : \
242 		      (uv) < 0x200000       ? 4 : \
243 		      (uv) < 0x4000000      ? 5 : \
244 		      (uv) < 0x80000000     ? 6 : \
245                       (uv) < UTF8_QUAD_MAX ? 7 : 13 )
246 #else
247 /* No, I'm not even going to *TRY* putting #ifdef inside a #define */
248 #define OFFUNISKIP(uv) ( (uv) < 0x80        ? 1 : \
249 		      (uv) < 0x800          ? 2 : \
250 		      (uv) < 0x10000        ? 3 : \
251 		      (uv) < 0x200000       ? 4 : \
252 		      (uv) < 0x4000000      ? 5 : \
253 		      (uv) < 0x80000000     ? 6 : 7 )
254 #endif
255 
256 /* How wide can a single UTF-8 encoded character become in bytes. */
257 /* NOTE: Strictly speaking Perl's UTF-8 should not be called UTF-8 since UTF-8
258  * is an encoding of Unicode, and Unicode's upper limit, 0x10FFFF, can be
259  * expressed with 4 bytes.  However, Perl thinks of UTF-8 as a way to encode
260  * non-negative integers in a binary format, even those above Unicode */
261 #define UTF8_MAXBYTES 13
262 
263 /* The maximum number of UTF-8 bytes a single Unicode character can
264  * uppercase/lowercase/fold into.  Unicode guarantees that the maximum
265  * expansion is 3 characters.  On ASCIIish platforms, the highest Unicode
266  * character occupies 4 bytes, therefore this number would be 12, but this is
267  * smaller than the maximum width a single above-Unicode character can occupy,
268  * so use that instead */
269 #if UTF8_MAXBYTES < 12
270 #error UTF8_MAXBYTES must be at least 12
271 #endif
272 
273 #define MAX_UTF8_TWO_BYTE 0x7FF
274 
275 #define UTF8_MAXBYTES_CASE	UTF8_MAXBYTES
276 
277 #endif /* EBCDIC vs ASCII */
278 
279 /* Rest of these are attributes of Unicode and perl's internals rather than the
280  * encoding, or happen to be the same in both ASCII and EBCDIC (at least at
281  * this level; the macros that some of these call may have different
282  * definitions in the two encodings */
283 
284 /* In domain restricted to ASCII, these may make more sense to the reader than
285  * the ones with Latin1 in the name */
286 #define NATIVE_TO_ASCII(ch)      NATIVE_TO_LATIN1(ch)
287 #define ASCII_TO_NATIVE(ch)      LATIN1_TO_NATIVE(ch)
288 
289 /* More or less misleadingly-named defines, retained for back compat */
290 #define NATIVE_TO_UTF(ch)        NATIVE_UTF8_TO_I8(ch)
291 #define NATIVE_TO_I8(ch)         NATIVE_UTF8_TO_I8(ch)
292 #define UTF_TO_NATIVE(ch)        I8_TO_NATIVE_UTF8(ch)
293 #define I8_TO_NATIVE(ch)         I8_TO_NATIVE_UTF8(ch)
294 #define NATIVE8_TO_UNI(ch)       NATIVE_TO_LATIN1(ch)
295 
296 /* Adds a UTF8 continuation byte 'new' of information to a running total code
297  * point 'old' of all the continuation bytes so far.  This is designed to be
298  * used in a loop to convert from UTF-8 to the code point represented.  Note
299  * that this is asymmetric on EBCDIC platforms, in that the 'new' parameter is
300  * the UTF-EBCDIC byte, whereas the 'old' parameter is a Unicode (not EBCDIC)
301  * code point in process of being generated */
302 #define UTF8_ACCUMULATE(old, new) (((old) << UTF_ACCUMULATION_SHIFT)           \
303                                    | ((NATIVE_UTF8_TO_I8((U8)new))             \
304                                        & UTF_CONTINUATION_MASK))
305 
306 /* This works in the face of malformed UTF-8. */
307 #define UTF8_IS_NEXT_CHAR_DOWNGRADEABLE(s, e) (UTF8_IS_DOWNGRADEABLE_START(*s) \
308                                                && ( (e) - (s) > 1)             \
309                                                && UTF8_IS_CONTINUATION(*((s)+1)))
310 
311 /* Number of bytes a code point occupies in UTF-8. */
312 #define NATIVE_SKIP(uv) OFFUNISKIP(NATIVE_TO_UNI(uv))
313 
314 /* Most code which says UNISKIP is really thinking in terms of native code
315  * points (0-255) plus all those beyond.  This is an imprecise term, but having
316  * it means existing code continues to work.  For precision, use NATIVE_SKIP
317  * and OFFUNISKIP */
318 #define UNISKIP(uv)   NATIVE_SKIP(uv)
319 
320 /* Convert a two (not one) byte utf8 character to a native code point value.
321  * Needs just one iteration of accumulate.  Should not be used unless it is
322  * known that the two bytes are legal: 1) two-byte start, and 2) continuation.
323  * Note that the result can be larger than 255 if the input character is not
324  * downgradable */
325 #define TWO_BYTE_UTF8_TO_NATIVE(HI, LO) \
326      UNI_TO_NATIVE(UTF8_ACCUMULATE((NATIVE_UTF8_TO_I8(HI) & UTF_START_MASK(2)), \
327                                    (LO)))
328 
329 /* Should never be used, and be deprecated */
330 #define TWO_BYTE_UTF8_TO_UNI(HI, LO) NATIVE_TO_UNI(TWO_BYTE_UTF8_TO_NATIVE(HI, LO))
331 
332 /* How many bytes in the UTF-8 encoded character whose first (perhaps only)
333  * byte is pointed to by 's' */
334 #define UTF8SKIP(s) PL_utf8skip[*(const U8*)(s)]
335 
336 /* Is the byte 'c' the same character when encoded in UTF-8 as when not.  This
337  * works on both UTF-8 encoded strings and non-encoded, as it returns TRUE in
338  * each for the exact same set of bit patterns.  (And it works on any byte in a
339  * UTF-8 encoded string) */
340 #define UTF8_IS_INVARIANT(c)		UNI_IS_INVARIANT(NATIVE_UTF8_TO_I8(c))
341 
342 /* Like the above, but its name implies a non-UTF8 input, and is implemented
343  * differently (for no particular reason) */
344 #define NATIVE_BYTE_IS_INVARIANT(c)	UNI_IS_INVARIANT(NATIVE_TO_LATIN1(c))
345 
346 /* Like the above, but accepts any UV as input */
347 #define UVCHR_IS_INVARIANT(uv)          UNI_IS_INVARIANT(NATIVE_TO_UNI(uv))
348 
349 #define MAX_PORTABLE_UTF8_TWO_BYTE 0x3FF    /* constrained by EBCDIC */
350 
351 /* The macros in the next 4 sets are used to generate the two utf8 or utfebcdic
352  * bytes from an ordinal that is known to fit into exactly two (not one) bytes;
353  * it must be less than 0x3FF to work across both encodings. */
354 
355 /* These two are helper macros for the other three sets, and should not be used
356  * directly anywhere else.  'translate_function' is either NATIVE_TO_LATIN1
357  * (which works for code points up to 0xFF) or NATIVE_TO_UNI which works for any
358  * code point */
359 #define __BASE_TWO_BYTE_HI(c, translate_function)                               \
360             I8_TO_NATIVE_UTF8((translate_function(c) >> UTF_ACCUMULATION_SHIFT) \
361                               | UTF_START_MARK(2))
362 #define __BASE_TWO_BYTE_LO(c, translate_function)                               \
363               I8_TO_NATIVE_UTF8((translate_function(c) & UTF_CONTINUATION_MASK) \
364                                  | UTF_CONTINUATION_MARK)
365 
366 /* The next two macros should not be used.  They were designed to be usable as
367  * the case label of a switch statement, but this doesn't work for EBCDIC.  Use
368  * regen/unicode_constants.pl instead */
369 #define UTF8_TWO_BYTE_HI_nocast(c)  __BASE_TWO_BYTE_HI(c, NATIVE_TO_UNI)
370 #define UTF8_TWO_BYTE_LO_nocast(c)  __BASE_TWO_BYTE_LO(c, NATIVE_TO_UNI)
371 
372 /* The next two macros are used when the source should be a single byte
373  * character; checked for under DEBUGGING */
374 #define UTF8_EIGHT_BIT_HI(c) (__ASSERT_(FITS_IN_8_BITS(c))                    \
375                              ((U8) __BASE_TWO_BYTE_HI(c, NATIVE_TO_LATIN1)))
376 #define UTF8_EIGHT_BIT_LO(c) (__ASSERT_(FITS_IN_8_BITS(c))                    \
377                              ((U8) __BASE_TWO_BYTE_LO(c, NATIVE_TO_LATIN1)))
378 
379 /* These final two macros in the series are used when the source can be any
380  * code point whose UTF-8 is known to occupy 2 bytes; they are less efficient
381  * than the EIGHT_BIT versions on EBCDIC platforms.  We use the logical '~'
382  * operator instead of "<=" to avoid getting compiler warnings.
383  * MAX_PORTABLE_UTF8_TWO_BYTE should be exactly all one bits in the lower few
384  * places, so the ~ works */
385 #define UTF8_TWO_BYTE_HI(c)                                                    \
386        (__ASSERT_((sizeof(c) ==  1)                                            \
387                   || !(((WIDEST_UTYPE)(c)) & ~MAX_PORTABLE_UTF8_TWO_BYTE))     \
388         ((U8) __BASE_TWO_BYTE_HI(c, NATIVE_TO_LATIN1)))
389 #define UTF8_TWO_BYTE_LO(c)                                                    \
390        (__ASSERT_((sizeof(c) ==  1)                                            \
391                   || !(((WIDEST_UTYPE)(c)) & ~MAX_PORTABLE_UTF8_TWO_BYTE))     \
392         ((U8) __BASE_TWO_BYTE_LO(c, NATIVE_TO_LATIN1)))
393 
394 /* This is illegal in any well-formed UTF-8 in both EBCDIC and ASCII
395  * as it is only in overlongs. */
396 #define ILLEGAL_UTF8_BYTE   I8_TO_NATIVE_UTF8(0xC1)
397 
398 /*
399  * 'UTF' is whether or not p is encoded in UTF8.  The names 'foo_lazy_if' stem
400  * from an earlier version of these macros in which they didn't call the
401  * foo_utf8() macros (i.e. were 'lazy') unless they decided that *p is the
402  * beginning of a utf8 character.  Now that foo_utf8() determines that itself,
403  * no need to do it again here
404  */
405 #define isIDFIRST_lazy_if(p,UTF) ((IN_BYTES || !UTF ) \
406 				 ? isIDFIRST(*(p)) \
407 				 : isIDFIRST_utf8((const U8*)p))
408 #define isWORDCHAR_lazy_if(p,UTF)   ((IN_BYTES || (!UTF )) \
409 				 ? isWORDCHAR(*(p)) \
410 				 : isWORDCHAR_utf8((const U8*)p))
411 #define isALNUM_lazy_if(p,UTF)   isWORDCHAR_lazy_if(p,UTF)
412 
413 #define UTF8_MAXLEN UTF8_MAXBYTES
414 
415 /* A Unicode character can fold to up to 3 characters */
416 #define UTF8_MAX_FOLD_CHAR_EXPAND 3
417 
418 #define IN_BYTES (CopHINTS_get(PL_curcop) & HINT_BYTES)
419 #define DO_UTF8(sv) (SvUTF8(sv) && !IN_BYTES)
420 #define IN_UNI_8_BIT \
421 	    (CopHINTS_get(PL_curcop) & (HINT_UNI_8_BIT|HINT_LOCALE_NOT_CHARS) \
422 	     && ! IN_LOCALE_RUNTIME && ! IN_BYTES)
423 
424 
425 #define UTF8_ALLOW_EMPTY		0x0001	/* Allow a zero length string */
426 
427 /* Allow first byte to be a continuation byte */
428 #define UTF8_ALLOW_CONTINUATION		0x0002
429 
430 /* Allow second... bytes to be non-continuation bytes */
431 #define UTF8_ALLOW_NON_CONTINUATION	0x0004
432 
433 /* expecting more bytes than were available in the string */
434 #define UTF8_ALLOW_SHORT		0x0008
435 
436 /* Overlong sequence; i.e., the code point can be specified in fewer bytes. */
437 #define UTF8_ALLOW_LONG                 0x0010
438 
439 #define UTF8_DISALLOW_SURROGATE		0x0020	/* Unicode surrogates */
440 #define UTF8_WARN_SURROGATE		0x0040
441 
442 #define UTF8_DISALLOW_NONCHAR           0x0080	/* Unicode non-character */
443 #define UTF8_WARN_NONCHAR               0x0100	/*  code points */
444 
445 #define UTF8_DISALLOW_SUPER		0x0200	/* Super-set of Unicode: code */
446 #define UTF8_WARN_SUPER		        0x0400	/* points above the legal max */
447 
448 /* Code points which never were part of the original UTF-8 standard, the first
449  * byte of which is a FE or FF on ASCII platforms. If the first byte is FF, it
450  * will overflow a 32-bit word.  If the first byte is FE, it will overflow a
451  * signed 32-bit word. */
452 #define UTF8_DISALLOW_FE_FF		0x0800
453 #define UTF8_WARN_FE_FF		        0x1000
454 
455 #define UTF8_CHECK_ONLY			0x2000
456 
457 /* For backwards source compatibility.  They do nothing, as the default now
458  * includes what they used to mean.  The first one's meaning was to allow the
459  * just the single non-character 0xFFFF */
460 #define UTF8_ALLOW_FFFF 0
461 #define UTF8_ALLOW_SURROGATE 0
462 
463 #define UTF8_DISALLOW_ILLEGAL_INTERCHANGE (UTF8_DISALLOW_SUPER|UTF8_DISALLOW_NONCHAR|UTF8_DISALLOW_SURROGATE|UTF8_DISALLOW_FE_FF)
464 #define UTF8_WARN_ILLEGAL_INTERCHANGE \
465 	(UTF8_WARN_SUPER|UTF8_WARN_NONCHAR|UTF8_WARN_SURROGATE|UTF8_WARN_FE_FF)
466 #define UTF8_ALLOW_ANY \
467 	    (~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE|UTF8_WARN_ILLEGAL_INTERCHANGE))
468 #define UTF8_ALLOW_ANYUV                                                        \
469          (UTF8_ALLOW_EMPTY                                                      \
470 	  & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE|UTF8_WARN_ILLEGAL_INTERCHANGE))
471 #define UTF8_ALLOW_DEFAULT		(ckWARN(WARN_UTF8) ? 0 : \
472 					 UTF8_ALLOW_ANYUV)
473 
474 /* Surrogates, non-character code points and above-Unicode code points are
475  * problematic in some contexts.  This allows code that needs to check for
476  * those to to quickly exclude the vast majority of code points it will
477  * encounter */
478 #define UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE \
479                                     FIRST_SURROGATE_UTF8_FIRST_BYTE
480 
481 #define UTF8_IS_SURROGATE(s) cBOOL(is_SURROGATE_utf8(s))
482 #define UTF8_IS_REPLACEMENT(s, send) cBOOL(is_REPLACEMENT_utf8_safe(s,send))
483 
484 /*		  ASCII		     EBCDIC I8
485  * U+10FFFF: \xF4\x8F\xBF\xBF	\xF9\xA1\xBF\xBF\xBF	max legal Unicode
486  * U+110000: \xF4\x90\x80\x80	\xF9\xA2\xA0\xA0\xA0
487  * U+110001: \xF4\x90\x80\x81	\xF9\xA2\xA0\xA0\xA1
488  */
489 #ifdef EBCDIC /* Both versions assume well-formed UTF8 */
490 #   define UTF8_IS_SUPER(s) (NATIVE_UTF8_TO_I8(* (U8*) (s)) >= 0xF9             \
491                          && (NATIVE_UTF8_TO_I8(* (U8*) (s)) > 0xF9              \
492                              || (NATIVE_UTF8_TO_I8(* ((U8*) (s) + 1)) >= 0xA2)))
493 #else
494 #   define UTF8_IS_SUPER(s) (*(U8*) (s) >= 0xF4                                 \
495                             && (*(U8*) (s) > 0xF4 || (*((U8*) (s) + 1) >= 0x90)))
496 #endif
497 
498 /* These are now machine generated, and the 'given' clause is no longer
499  * applicable */
500 #define UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s)             \
501                                                     cBOOL(is_NONCHAR_utf8(s))
502 #define UTF8_IS_NONCHAR_(s)                                                    \
503                     UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s)
504 
505 #define UNICODE_SURROGATE_FIRST		0xD800
506 #define UNICODE_SURROGATE_LAST		0xDFFF
507 #define UNICODE_REPLACEMENT		0xFFFD
508 #define UNICODE_BYTE_ORDER_MARK		0xFEFF
509 
510 /* Though our UTF-8 encoding can go beyond this,
511  * let's be conservative and do as Unicode says. */
512 #define PERL_UNICODE_MAX	0x10FFFF
513 
514 #define UNICODE_WARN_SURROGATE     0x0001	/* UTF-16 surrogates */
515 #define UNICODE_WARN_NONCHAR       0x0002	/* Non-char code points */
516 #define UNICODE_WARN_SUPER         0x0004	/* Above 0x10FFFF */
517 #define UNICODE_WARN_FE_FF         0x0008	/* Above 0x10FFFF */
518 #define UNICODE_DISALLOW_SURROGATE 0x0010
519 #define UNICODE_DISALLOW_NONCHAR   0x0020
520 #define UNICODE_DISALLOW_SUPER     0x0040
521 #define UNICODE_DISALLOW_FE_FF     0x0080
522 #define UNICODE_WARN_ILLEGAL_INTERCHANGE                                      \
523             (UNICODE_WARN_SURROGATE|UNICODE_WARN_NONCHAR|UNICODE_WARN_SUPER)
524 #define UNICODE_DISALLOW_ILLEGAL_INTERCHANGE                                  \
525  (UNICODE_DISALLOW_SURROGATE|UNICODE_DISALLOW_NONCHAR|UNICODE_DISALLOW_SUPER)
526 
527 /* For backward source compatibility, as are now the default */
528 #define UNICODE_ALLOW_SURROGATE 0
529 #define UNICODE_ALLOW_SUPER	0
530 #define UNICODE_ALLOW_ANY	0
531 
532 #define UNICODE_IS_SURROGATE(c)		((c) >= UNICODE_SURROGATE_FIRST && \
533 					 (c) <= UNICODE_SURROGATE_LAST)
534 #define UNICODE_IS_REPLACEMENT(c)	((c) == UNICODE_REPLACEMENT)
535 #define UNICODE_IS_BYTE_ORDER_MARK(c)	((c) == UNICODE_BYTE_ORDER_MARK)
536 #define UNICODE_IS_NONCHAR(c)		((c >= 0xFDD0 && c <= 0xFDEF) \
537 			/* The other noncharacters end in FFFE or FFFF, which  \
538 			 * the mask below catches both of, but beyond the last \
539 			 * official unicode code point, they aren't            \
540 			 * noncharacters, since those aren't Unicode           \
541 			 * characters at all */                                \
542 			|| ((((c & 0xFFFE) == 0xFFFE)) && ! UNICODE_IS_SUPER(c)))
543 #define UNICODE_IS_SUPER(c)		((c) > PERL_UNICODE_MAX)
544 #define UNICODE_IS_FE_FF(c)		((c) > 0x7FFFFFFF)
545 
546 #define LATIN_SMALL_LETTER_SHARP_S      LATIN_SMALL_LETTER_SHARP_S_NATIVE
547 #define LATIN_SMALL_LETTER_Y_WITH_DIAERESIS                                  \
548                                 LATIN_SMALL_LETTER_Y_WITH_DIAERESIS_NATIVE
549 #define MICRO_SIGN      MICRO_SIGN_NATIVE
550 #define LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE                               \
551                             LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE_NATIVE
552 #define LATIN_SMALL_LETTER_A_WITH_RING_ABOVE                                 \
553                                 LATIN_SMALL_LETTER_A_WITH_RING_ABOVE_NATIVE
554 #define UNICODE_GREEK_CAPITAL_LETTER_SIGMA	0x03A3
555 #define UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA	0x03C2
556 #define UNICODE_GREEK_SMALL_LETTER_SIGMA	0x03C3
557 #define GREEK_SMALL_LETTER_MU                   0x03BC
558 #define GREEK_CAPITAL_LETTER_MU                 0x039C	/* Upper and title case
559                                                            of MICRON */
560 #define LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS   0x0178	/* Also is title case */
561 #define LATIN_CAPITAL_LETTER_SHARP_S	        0x1E9E
562 #define LATIN_SMALL_LETTER_LONG_S               0x017F
563 #define LATIN_SMALL_LIGATURE_LONG_S_T           0xFB05
564 #define LATIN_SMALL_LIGATURE_ST                 0xFB06
565 #define KELVIN_SIGN                             0x212A
566 #define ANGSTROM_SIGN                           0x212B
567 
568 #define UNI_DISPLAY_ISPRINT	0x0001
569 #define UNI_DISPLAY_BACKSLASH	0x0002
570 #define UNI_DISPLAY_QQ		(UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)
571 #define UNI_DISPLAY_REGEX	(UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)
572 
573 #define ANYOF_FOLD_SHARP_S(node, input, end)	\
574 	(ANYOF_BITMAP_TEST(node, LATIN_SMALL_LETTER_SHARP_S) && \
575 	 (ANYOF_NONBITMAP(node)) && \
576 	 (ANYOF_FLAGS(node) & ANYOF_LOC_NONBITMAP_FOLD) && \
577 	 ((end) > (input) + 1) && \
578 	 toFOLD((input)[0]) == 's' && \
579 	 toFOLD((input)[1]) == 's')
580 #define SHARP_S_SKIP 2
581 
582 /* If you want to exclude surrogates, and beyond legal Unicode, see the blame
583  * log for earlier versions which gave details for these */
584 
585 #ifndef EBCDIC
586 /* This was generated by regen/regcharclass.pl, and then moved here.  The lines
587  * that generated it were then commented out.  This was done solely because it
588  * takes on the order of 10 minutes to generate, and is never going to change.
589  * The EBCDIC equivalent hasn't been commented out in regcharclass.pl, so it
590  * should generate and run the correct stuff */
591 /*
592 	UTF8_CHAR: Matches utf8 from 1 to 4 bytes
593 
594 	0x0 - 0x1FFFFF
595 */
596 /*** GENERATED CODE ***/
597 #define is_UTF8_CHAR_utf8_safe(s,e)                                         \
598 ( ((e)-(s) > 3) ?                                                           \
599     ( ( ( ((U8*)s)[0] & 0x80 ) == 0x00 ) ? 1                                \
600     : ( 0xC2 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xDF ) ?                      \
601 	( ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) ? 2 : 0 )                      \
602     : ( 0xE0 == ((U8*)s)[0] ) ?                                             \
603 	( ( ( ( ((U8*)s)[1] & 0xE0 ) == 0xA0 ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
604     : ( 0xE1 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xEF ) ?                      \
605 	( ( ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
606     : ( 0xF0 == ((U8*)s)[0] ) ?                                             \
607 	( ( ( ( 0x90 <= ((U8*)s)[1] && ((U8*)s)[1] <= 0xBF ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\
608     : ( ( ( ( 0xF1 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xF7 ) && ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\
609 : ((e)-(s) > 2) ?                                                           \
610     ( ( ( ((U8*)s)[0] & 0x80 ) == 0x00 ) ? 1                                \
611     : ( 0xC2 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xDF ) ?                      \
612 	( ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) ? 2 : 0 )                      \
613     : ( 0xE0 == ((U8*)s)[0] ) ?                                             \
614 	( ( ( ( ((U8*)s)[1] & 0xE0 ) == 0xA0 ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
615     : ( ( ( 0xE1 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xEF ) && ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) ) && ( ( ((U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
616 : ((e)-(s) > 1) ?                                                           \
617     ( ( ( ((U8*)s)[0] & 0x80 ) == 0x00 ) ? 1                                \
618     : ( ( 0xC2 <= ((U8*)s)[0] && ((U8*)s)[0] <= 0xDF ) && ( ( ((U8*)s)[1] & 0xC0 ) == 0x80 ) ) ? 2 : 0 )\
619 : ((e)-(s) > 0) ?                                                           \
620     ( ( ((U8*)s)[0] & 0x80 ) == 0x00 )                                      \
621 : 0 )
622 #endif
623 
624 /* IS_UTF8_CHAR(p) is strictly speaking wrong (not UTF-8) because it
625  * (1) allows UTF-8 encoded UTF-16 surrogates
626  * (2) it allows code points past U+10FFFF.
627  * The Perl_is_utf8_char() full "slow" code will handle the Perl
628  * "extended UTF-8". */
629 #define IS_UTF8_CHAR(p, n)      (is_UTF8_CHAR_utf8_safe(p, (p) + (n)) == n)
630 
631 /* regen/regcharclass.pl generates is_UTF8_CHAR_utf8_safe() macros for up to
632  * these number of bytes.  So this has to be coordinated with it */
633 #ifdef EBCDIC
634 #   define IS_UTF8_CHAR_FAST(n) ((n) <= 5)
635 #else
636 #   define IS_UTF8_CHAR_FAST(n) ((n) <= 4)
637 #endif
638 
639 #endif /* H_UTF8 */
640 
641 /*
642  * Local variables:
643  * c-indentation-style: bsd
644  * c-basic-offset: 4
645  * indent-tabs-mode: nil
646  * End:
647  *
648  * ex: set ts=8 sts=4 sw=4 et:
649  */
650