utf8.h 59.6 KB
Newer Older
1
/*    utf8.h
2 3 4 5 6 7
 *
 * This file contains definitions for use with the UTF-8 encoding.  It
 * actually also works with the variant UTF-8 encoding called UTF-EBCDIC, and
 * hides almost all of the differences between these from the caller.  In other
 * words, someone should #include this file, and if the code is being compiled
 * on an EBCDIC platform, things should mostly just work.
8
 *
9 10
 *    Copyright (C) 2000, 2001, 2002, 2005, 2006, 2007, 2009,
 *    2010, 2011 by Larry Wall and others
11 12 13 14 15 16
 *
 *    You may distribute under the terms of either the GNU General Public
 *    License or the Artistic License, as specified in the README file.
 *
 */

17 18
#ifndef PERL_UTF8_H_      /* Guard against recursive inclusion */
#define PERL_UTF8_H_ 1
19

20 21 22 23 24 25 26 27 28
/* Use UTF-8 as the default script encoding?
 * Turning this on will break scripts having non-UTF-8 binary
 * data (such as Latin-1) in string literals. */
#ifdef USE_UTF8_SCRIPTS
#    define USE_UTF8_IN_NAMES (!IN_BYTES)
#else
#    define USE_UTF8_IN_NAMES (PL_hints & HINT_UTF8)
#endif

29 30 31
#include "regcharclass.h"
#include "unicode_constants.h"

32
/* For to_utf8_fold_flags, q.v. */
33 34 35
#define FOLD_FLAGS_LOCALE       0x1
#define FOLD_FLAGS_FULL         0x2
#define FOLD_FLAGS_NOMIX_ASCII  0x4
36

37 38 39 40 41
/* For _core_swash_init(), internal core use only */
#define _CORE_SWASH_INIT_USER_DEFINED_PROPERTY 0x1
#define _CORE_SWASH_INIT_RETURN_IF_UNDEF       0x2
#define _CORE_SWASH_INIT_ACCEPT_INVLIST        0x4

42 43 44 45 46 47 48 49 50 51 52
/*
=head1 Unicode Support
L<perlguts/Unicode Support> has an introduction to this API.

See also L</Character classification>,
and L</Character case changing>.
Various functions outside this section also work specially with Unicode.
Search for the string "utf8" in this document.

=for apidoc is_ascii_string

53
This is a misleadingly-named synonym for L</is_utf8_invariant_string>.
54 55
On ASCII-ish platforms, the name isn't misleading: the ASCII-range characters
are exactly the UTF-8 invariants.  But EBCDIC machines have more invariants
56 57 58 59 60 61 62
than just the ASCII characters, so C<is_utf8_invariant_string> is preferred.

=for apidoc is_invariant_string

This is a somewhat misleadingly-named synonym for L</is_utf8_invariant_string>.
C<is_utf8_invariant_string> is preferred, as it indicates under what conditions
the string is invariant.
63 64 65

=cut
*/
66 67
#define is_ascii_string(s, len)     is_utf8_invariant_string(s, len)
#define is_invariant_string(s, len) is_utf8_invariant_string(s, len)
68

69 70
#define uvoffuni_to_utf8_flags(d,uv,flags)                                     \
                               uvoffuni_to_utf8_flags_msgs(d, uv, flags, 0)
71 72
#define uvchr_to_utf8(a,b)          uvchr_to_utf8_flags(a,b,0)
#define uvchr_to_utf8_flags(d,uv,flags)                                        \
73 74 75
                                    uvchr_to_utf8_flags_msgs(d,uv,flags, 0)
#define uvchr_to_utf8_flags_msgs(d,uv,flags,msgs)                              \
                uvoffuni_to_utf8_flags_msgs(d,NATIVE_TO_UNI(uv),flags, msgs)
76 77 78
#define utf8_to_uvchr_buf(s, e, lenp)                                          \
                     utf8n_to_uvchr(s, (U8*)(e) - (U8*)(s), lenp,              \
                                    ckWARN_d(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY)
79 80
#define utf8n_to_uvchr(s, len, lenp, flags)                                    \
                                utf8n_to_uvchr_error(s, len, lenp, flags, 0)
81 82
#define utf8n_to_uvchr_error(s, len, lenp, flags, errors)                      \
                        utf8n_to_uvchr_msgs(s, len, lenp, flags, errors, 0)
83

84
#define to_uni_fold(c, p, lenp) _to_uni_fold_flags(c, p, lenp, FOLD_FLAGS_FULL)
85

86 87 88 89 90 91 92 93
#define to_utf8_fold(s, r, lenr)                                                \
    _to_utf8_fold_flags (s, NULL, r, lenr, FOLD_FLAGS_FULL, __FILE__, __LINE__)
#define to_utf8_lower(s, r, lenr)                                               \
                  _to_utf8_lower_flags(s, NULL, r ,lenr, 0, __FILE__, __LINE__)
#define to_utf8_upper(s, r, lenr)                                               \
                  _to_utf8_upper_flags(s, NULL, r, lenr, 0, __FILE__, __LINE__)
#define to_utf8_title(s, r, lenr)                                               \
                  _to_utf8_title_flags(s, NULL, r, lenr ,0, __FILE__, __LINE__)
94

95 96
#define foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \
		    foldEQ_utf8_flags(s1, pe1, l1, u1, s2, pe2, l2, u2, 0)
97 98
#define FOLDEQ_UTF8_NOMIX_ASCII   (1 << 0)
#define FOLDEQ_LOCALE             (1 << 1)
99 100
#define FOLDEQ_S1_ALREADY_FOLDED  (1 << 2)
#define FOLDEQ_S2_ALREADY_FOLDED  (1 << 3)
101 102
#define FOLDEQ_S1_FOLDS_SANE      (1 << 4)
#define FOLDEQ_S2_FOLDS_SANE      (1 << 5)
103 104 105 106

#define ibcmp_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \
		    cBOOL(! foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2))

107 108 109 110 111 112
#ifdef EBCDIC
/* The equivalent of these macros but implementing UTF-EBCDIC
   are in the following header file:
 */

#include "utfebcdic.h"
113

114
#else	/* ! EBCDIC */
115 116
START_EXTERN_C

117 118 119 120 121 122 123
/* How wide can a single UTF-8 encoded character become in bytes. */
/* NOTE: Strictly speaking Perl's UTF-8 should not be called UTF-8 since UTF-8
 * is an encoding of Unicode, and Unicode's upper limit, 0x10FFFF, can be
 * expressed with 4 bytes.  However, Perl thinks of UTF-8 as a way to encode
 * non-negative integers in a binary format, even those above Unicode */
#define UTF8_MAXBYTES 13

124 125
#ifdef DOINIT
EXTCONST unsigned char PL_utf8skip[] = {
126 127 128 129 130 131 132 133 134 135 136 137 138
/* 0x00 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
/* 0x10 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
/* 0x20 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
/* 0x30 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
/* 0x40 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
/* 0x50 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
/* 0x60 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
/* 0x70 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
/* 0x80 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
/* 0x90 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
/* 0xA0 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
/* 0xB0 */ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus: continuation byte */
/* 0xC0 */ 2,2,				    /* overlong */
139
/* 0xC2 */     2,2,2,2,2,2,2,2,2,2,2,2,2,2, /* U+0080 to U+03FF */
140 141 142
/* 0xD0 */ 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, /* U+0400 to U+07FF */
/* 0xE0 */ 3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3, /* U+0800 to U+FFFF */
/* 0xF0 */ 4,4,4,4,4,4,4,4,5,5,5,5,6,6,	    /* above BMP to 2**31 - 1 */
143 144 145 146
           /* Perl extended (never was official UTF-8).  Up to 36 bit */
/* 0xFE */                             7,
           /* More extended, Up to 72 bits (64-bit + reserved) */
/* 0xFF */                               UTF8_MAXBYTES
147 148 149 150 151 152 153
};
#else
EXTCONST unsigned char PL_utf8skip[];
#endif

END_EXTERN_C

154 155 156 157 158
#if defined(_MSC_VER) && _MSC_VER < 1400
/* older MSVC versions have a smallish macro buffer */
#define PERL_SMALL_MACRO_BUFFER
#endif

159 160
/* Native character to/from iso-8859-1.  Are the identity functions on ASCII
 * platforms */
161 162 163 164 165 166 167
#ifdef PERL_SMALL_MACRO_BUFFER
#define NATIVE_TO_LATIN1(ch)     ((U8)(ch))
#define LATIN1_TO_NATIVE(ch)     ((U8)(ch))
#else
#define NATIVE_TO_LATIN1(ch)     (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) (ch)))
#define LATIN1_TO_NATIVE(ch)     (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) (ch)))
#endif
168 169 170 171 172 173 174

/* I8 is an intermediate version of UTF-8 used only in UTF-EBCDIC.  We thus
 * consider it to be identical to UTF-8 on ASCII platforms.  Strictly speaking
 * UTF-8 and UTF-EBCDIC are two different things, but we often conflate them
 * because they are 8-bit encodings that serve the same purpose in Perl, and
 * rarely do we need to distinguish them.  The term "NATIVE_UTF8" applies to
 * whichever one is applicable on the current platform */
175
#ifdef PERL_SMALL_MACRO_BUFFER
176 177
#define NATIVE_UTF8_TO_I8(ch) ((U8) (ch))
#define I8_TO_NATIVE_UTF8(ch) ((U8) (ch))
178 179 180 181
#else
#define NATIVE_UTF8_TO_I8(ch) (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) (ch)))
#define I8_TO_NATIVE_UTF8(ch) (__ASSERT_(FITS_IN_8_BITS(ch)) ((U8) (ch)))
#endif
182

183
/* Transforms in wide UV chars */
184 185
#define UNI_TO_NATIVE(ch)        ((UV) (ch))
#define NATIVE_TO_UNI(ch)        ((UV) (ch))
186 187

/*
188

189 190
 The following table is from Unicode 3.2, plus the Perl extensions for above
 U+10FFFF
191

192
 Code Points		1st Byte  2nd Byte  3rd    4th     5th     6th       7th   8th-13th
193

194
   U+0000..U+007F	00..7F
195
   U+0080..U+07FF     * C2..DF    80..BF
196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212
   U+0800..U+0FFF	E0      * A0..BF  80..BF
   U+1000..U+CFFF       E1..EC    80..BF  80..BF
   U+D000..U+D7FF       ED        80..9F  80..BF
   U+D800..U+DFFF       ED        A0..BF  80..BF  (surrogates)
   U+E000..U+FFFF       EE..EF    80..BF  80..BF
  U+10000..U+3FFFF	F0      * 90..BF  80..BF  80..BF
  U+40000..U+FFFFF	F1..F3    80..BF  80..BF  80..BF
 U+100000..U+10FFFF	F4        80..8F  80..BF  80..BF
    Below are above-Unicode code points
 U+110000..U+13FFFF	F4        90..BF  80..BF  80..BF
 U+110000..U+1FFFFF	F5..F7    80..BF  80..BF  80..BF
 U+200000..U+FFFFFF     F8      * 88..BF  80..BF  80..BF  80..BF
U+1000000..U+3FFFFFF    F9..FB    80..BF  80..BF  80..BF  80..BF
U+4000000..U+3FFFFFFF    FC     * 84..BF  80..BF  80..BF  80..BF  80..BF
U+40000000..U+7FFFFFFF   FD       80..BF  80..BF  80..BF  80..BF  80..BF
U+80000000..U+FFFFFFFFF  FE     * 82..BF  80..BF  80..BF  80..BF  80..BF    80..BF
U+1000000000..           FF       80..BF  80..BF  80..BF  80..BF  80..BF  * 81..BF  80..BF
213

214 215 216 217
Note the gaps before several of the byte entries above marked by '*'.  These are
caused by legal UTF-8 avoiding non-shortest encodings: it is technically
possible to UTF-8-encode a single code point in different ways, but that is
explicitly forbidden, and the shortest possible encoding should always be used
218
(and that is what Perl does).  The non-shortest ones are called 'overlongs'.
219

220 221
 */

222 223 224
/*
 Another way to look at it, as bits:

225
                  Code Points      1st Byte   2nd Byte   3rd Byte   4th Byte
226

227 228 229 230
                        0aaa aaaa  0aaa aaaa
              0000 0bbb bbaa aaaa  110b bbbb  10aa aaaa
              cccc bbbb bbaa aaaa  1110 cccc  10bb bbbb  10aa aaaa
 00 000d ddcc cccc bbbb bbaa aaaa  1111 0ddd  10cc cccc  10bb bbbb  10aa aaaa
231 232

As you can see, the continuation bytes all begin with C<10>, and the
233
leading bits of the start byte tell how many bytes there are in the
234 235
encoded character.

236 237 238 239 240
Perl's extended UTF-8 means we can have start bytes up through FF, though any
beginning with FF yields a code point that is too large for 32-bit ASCII
platforms.  FF signals to use 13 bytes for the encoded character.  This breaks
the paradigm that the number of leading bits gives how many total bytes there
are in the character.
241

242
*/
243

244
/* Is the representation of the Unicode code point 'cp' the same regardless of
245
 * being encoded in UTF-8 or not? */
246
#define OFFUNI_IS_INVARIANT(cp)     isASCII(cp)
247

248 249 250 251 252 253 254 255 256 257 258
/*
=for apidoc Am|bool|UVCHR_IS_INVARIANT|UV cp

Evaluates to 1 if the representation of code point C<cp> is the same whether or
not it is encoded in UTF-8; otherwise evaluates to 0.  UTF-8 invariant
characters can be copied as-is when converting to/from UTF-8, saving time.
C<cp> is Unicode if above 255; otherwise is platform-native.

=cut
 */

259 260 261 262 263 264 265 266 267
#define UVCHR_IS_INVARIANT(cp)      OFFUNI_IS_INVARIANT(cp)

/* This defines the bits that are to be in the continuation bytes of a multi-byte
 * UTF-8 encoded character that mark it is a continuation byte. */
#define UTF_CONTINUATION_MARK		0x80

/* Misleadingly named: is the UTF8-encoded byte 'c' part of a variant sequence
 * in UTF-8?  This is the inverse of UTF8_IS_INVARIANT.  The |0 makes sure this
 * isn't mistakenly called with a ptr argument */
268 269
#define UTF8_IS_CONTINUED(c)  (__ASSERT_(FITS_IN_8_BITS(c))                 \
                               ((U8)((c) | 0)) &  UTF_CONTINUATION_MARK)
270

271 272
/* Is the byte 'c' the first byte of a multi-byte UTF8-8 encoded sequence?
 * This doesn't catch invariants (they are single-byte).  It also excludes the
273 274
 * illegal overlong sequences that begin with C0 and C1.  The |0 makes sure
 * this isn't mistakenly called with a ptr argument */
275 276
#define UTF8_IS_START(c)      (__ASSERT_(FITS_IN_8_BITS(c))                 \
                               ((U8)((c) | 0)) >= 0xc2)
277 278 279

/* For use in UTF8_IS_CONTINUATION() below */
#define UTF_IS_CONTINUATION_MASK    0xC0
280 281

/* Is the byte 'c' part of a multi-byte UTF8-8 encoded sequence, and not the
282 283
 * first byte thereof?  The |0 makes sure this isn't mistakenly called with a
 * ptr argument */
284 285
#define UTF8_IS_CONTINUATION(c)     (__ASSERT_(FITS_IN_8_BITS(c))           \
     (((U8)((c) | 0)) & UTF_IS_CONTINUATION_MASK) == UTF_CONTINUATION_MARK)
286 287 288 289

/* Is the UTF8-encoded byte 'c' the first byte of a two byte sequence?  Use
 * UTF8_IS_NEXT_CHAR_DOWNGRADEABLE() instead if the input isn't known to
 * be well-formed.  Masking with 0xfe allows the low bit to be 0 or 1; thus
290 291
 * this matches 0xc[23].  The |0 makes sure this isn't mistakenly called with a
 * ptr argument */
292 293
#define UTF8_IS_DOWNGRADEABLE_START(c)	(__ASSERT_(FITS_IN_8_BITS(c))       \
                                         (((U8)((c) | 0)) & 0xfe) == 0xc2)
294

295
/* Is the UTF8-encoded byte 'c' the first byte of a sequence of bytes that
296 297
 * represent a code point > 255?  The |0 makes sure this isn't mistakenly
 * called with a ptr argument */
298 299
#define UTF8_IS_ABOVE_LATIN1(c)     (__ASSERT_(FITS_IN_8_BITS(c))           \
                                     ((U8)((c) | 0)) >= 0xc4)
300 301 302 303 304

/* This is the number of low-order bits a continuation byte in a UTF-8 encoded
 * sequence contributes to the specification of the code point.  In the bit
 * maps above, you see that the first 2 bits are a constant '10', leaving 6 of
 * real information */
305
#define UTF_ACCUMULATION_SHIFT		6
306

307 308 309
/* ^? is defined to be DEL on ASCII systems.  See the definition of toCTRL()
 * for more */
#define QUESTION_MARK_CTRL  DEL_NATIVE
310

311 312 313 314
/* Surrogates, non-character code points and above-Unicode code points are
 * problematic in some contexts.  This allows code that needs to check for
 * those to to quickly exclude the vast majority of code points it will
 * encounter */
315 316
#define isUTF8_POSSIBLY_PROBLEMATIC(c) (__ASSERT_(FITS_IN_8_BITS(c))        \
                                        (U8) c >= 0xED)
317

318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415
/* A helper macro for isUTF8_CHAR, so use that one instead of this.  This was
 * generated by regen/regcharclass.pl, and then moved here.  Then it was
 * hand-edited to add some LIKELY() calls, presuming that malformations are
 * unlikely.  The lines that generated it were then commented out.  This was
 * done because it takes on the order of 10 minutes to generate, and is never
 * going to change, unless the generated code is improved, and figuring out
 * the LIKELYs there would be hard.
 *
        UTF8_CHAR: Matches legal UTF-8 variant code points up through 0x1FFFFFF

	0x80 - 0x1FFFFF
*/
/*** GENERATED CODE ***/
#define is_UTF8_CHAR_utf8_no_length_checks(s)                               \
( ( 0xC2 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xDF ) ?                          \
    ( LIKELY( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) ? 2 : 0 )                    \
: ( 0xE0 == ((const U8*)s)[0] ) ?                                                 \
    ( LIKELY( ( ( ((const U8*)s)[1] & 0xE0 ) == 0xA0 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
: ( 0xE1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xEF ) ?                          \
    ( LIKELY( ( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
: ( 0xF0 == ((const U8*)s)[0] ) ?                                                 \
    ( LIKELY( ( ( 0x90 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xBF ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\
: ( ( ( ( 0xF1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xF7 ) && LIKELY( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) ) && LIKELY( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) && LIKELY( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )

/* The above macro handles UTF-8 that has this start byte as the maximum */
#define _IS_UTF8_CHAR_HIGHEST_START_BYTE 0xF7

/* A helper macro for isSTRICT_UTF8_CHAR, so use that one instead of this.
 * Like is_UTF8_CHAR_utf8_no_length_checks(), this was moved here and LIKELYs
 * added manually.
 *
	STRICT_UTF8_CHAR: Matches legal Unicode UTF-8 variant code points, no
                          surrrogates nor non-character code points
*/
/*** GENERATED CODE ***/
#define is_STRICT_UTF8_CHAR_utf8_no_length_checks(s)                        \
( ( 0xC2 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xDF ) ?                          \
    ( LIKELY( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) ? 2 : 0 )                          \
: ( 0xE0 == ((const U8*)s)[0] ) ?                                                 \
    ( LIKELY( ( ( ((const U8*)s)[1] & 0xE0 ) == 0xA0 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
: ( ( 0xE1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xEC ) || 0xEE == ((const U8*)s)[0] ) ?\
    ( ( ( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
: ( 0xED == ((const U8*)s)[0] ) ?                                                 \
    ( LIKELY( ( ( ((const U8*)s)[1] & 0xE0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
: ( 0xEF == ((const U8*)s)[0] ) ?                                                 \
    ( ( ( 0x80 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xB6 ) || ( 0xB8 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xBE ) ) ?\
	( LIKELY( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ? 3 : 0 )                      \
    : ( 0xB7 == ((const U8*)s)[1] ) ?                                             \
	( LIKELY( ( ((const U8*)s)[2] & 0xF0 ) == 0x80 || ( ((const U8*)s)[2] & 0xF0 ) == 0xB0 ) ? 3 : 0 )\
    : ( ( 0xBF == ((const U8*)s)[1] ) && ( 0x80 <= ((const U8*)s)[2] && ((const U8*)s)[2] <= 0xBD ) ) ? 3 : 0 )\
: ( 0xF0 == ((const U8*)s)[0] ) ?                                                 \
    ( ( ( 0x90 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0x9E ) || ( 0xA0 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xAE ) || ( 0xB0 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xBE ) ) ?\
	( LIKELY( ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\
    : ( ((const U8*)s)[1] == 0x9F || ( ( ((const U8*)s)[1] & 0xEF ) == 0xAF ) ) ?       \
	( ( 0x80 <= ((const U8*)s)[2] && ((const U8*)s)[2] <= 0xBE ) ?                  \
	    ( LIKELY( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ? 4 : 0 )                  \
	: LIKELY( ( 0xBF == ((const U8*)s)[2] ) && ( 0x80 <= ((const U8*)s)[3] && ((const U8*)s)[3] <= 0xBD ) ) ? 4 : 0 )\
    : 0 )                                                                   \
: ( 0xF1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xF3 ) ?                          \
    ( ( ( ( ((const U8*)s)[1] & 0xC8 ) == 0x80 ) || ( ( ((const U8*)s)[1] & 0xCC ) == 0x88 ) || ( ( ((const U8*)s)[1] & 0xCE ) == 0x8C ) || ( ( ((const U8*)s)[1] & 0xCF ) == 0x8E ) ) ?\
	( LIKELY( ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\
    : ( ( ((const U8*)s)[1] & 0xCF ) == 0x8F ) ?                                  \
	( ( 0x80 <= ((const U8*)s)[2] && ((const U8*)s)[2] <= 0xBE ) ?                  \
	    ( LIKELY( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ? 4 : 0 )                  \
	: LIKELY( ( 0xBF == ((const U8*)s)[2] ) && ( 0x80 <= ((const U8*)s)[3] && ((const U8*)s)[3] <= 0xBD ) ) ? 4 : 0 )\
    : 0 )                                                                   \
: ( 0xF4 == ((const U8*)s)[0] ) ?                                                 \
    ( ( 0x80 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0x8E ) ?                      \
	( LIKELY( ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\
    : ( 0x8F == ((const U8*)s)[1] ) ?                                             \
	( ( 0x80 <= ((const U8*)s)[2] && ((const U8*)s)[2] <= 0xBE ) ?                  \
	    ( LIKELY( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ? 4 : 0 )                  \
	: LIKELY( ( 0xBF == ((const U8*)s)[2] ) && ( 0x80 <= ((const U8*)s)[3] && ((const U8*)s)[3] <= 0xBD ) ) ? 4 : 0 )\
    : 0 )                                                                   \
: 0 )

/*  Similarly,
        C9_STRICT_UTF8_CHAR: Matches legal Unicode UTF-8 variant code
                                     points, no surrogates
	0x0080 - 0xD7FF
	0xE000 - 0x10FFFF
*/
/*** GENERATED CODE ***/
#define is_C9_STRICT_UTF8_CHAR_utf8_no_length_checks(s)                     \
( ( 0xC2 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xDF ) ?                          \
    ( LIKELY( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) ? 2 : 0 )                    \
: ( 0xE0 == ((const U8*)s)[0] ) ?                                                 \
    ( LIKELY( ( ( ((const U8*)s)[1] & 0xE0 ) == 0xA0 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
: ( ( 0xE1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xEC ) || ( ((const U8*)s)[0] & 0xFE ) == 0xEE ) ?\
    ( LIKELY( ( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
: ( 0xED == ((const U8*)s)[0] ) ?                                                 \
    ( LIKELY( ( ( ((const U8*)s)[1] & 0xE0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) ? 3 : 0 )\
: ( 0xF0 == ((const U8*)s)[0] ) ?                                                 \
    ( LIKELY( ( ( 0x90 <= ((const U8*)s)[1] && ((const U8*)s)[1] <= 0xBF ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\
: ( 0xF1 <= ((const U8*)s)[0] && ((const U8*)s)[0] <= 0xF3 ) ?                          \
    ( LIKELY( ( ( ( ((const U8*)s)[1] & 0xC0 ) == 0x80 ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )\
: LIKELY( ( ( ( 0xF4 == ((const U8*)s)[0] ) && ( ( ((const U8*)s)[1] & 0xF0 ) == 0x80 ) ) && ( ( ((const U8*)s)[2] & 0xC0 ) == 0x80 ) ) && ( ( ((const U8*)s)[3] & 0xC0 ) == 0x80 ) ) ? 4 : 0 )

416 417
#define UNICODE_IS_PERL_EXTENDED(uv)    UNLIKELY((UV) (uv) > 0x7FFFFFFF)

418
#endif /* EBCDIC vs ASCII */
419

420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458
/* 2**UTF_ACCUMULATION_SHIFT - 1 */
#define UTF_CONTINUATION_MASK  ((U8) ((1U << UTF_ACCUMULATION_SHIFT) - 1))

/* Internal macro to be used only in this file to aid in constructing other
 * publicly accessible macros.
 * The number of bytes required to express this uv in UTF-8, for just those
 * uv's requiring 2 through 6 bytes, as these are common to all platforms and
 * word sizes.  The number of bytes needed is given by the number of leading 1
 * bits in the start byte.  There are 32 start bytes that have 2 initial 1 bits
 * (C0-DF); there are 16 that have 3 initial 1 bits (E0-EF); 8 that have 4
 * initial 1 bits (F0-F8); 4 that have 5 initial 1 bits (F9-FB), and 2 that
 * have 6 initial 1 bits (FC-FD).  The largest number a string of n bytes can
 * represent is       (the number of possible start bytes for 'n')
 *                  * (the number of possiblities for each start byte
 * The latter in turn is
 *                  2  ** (  (how many continuation bytes there are)
 *                         * (the number of bits of information each
 *                            continuation byte holds))
 *
 * If we were on a platform where we could use a fast find first set bit
 * instruction (or count leading zeros instruction) this could be replaced by
 * using that to find the log2 of the uv, and divide that by the number of bits
 * of information in each continuation byte, adjusting for large cases and how
 * much information is in a start byte for that length */
#define __COMMON_UNI_SKIP(uv)                                               \
          (UV) (uv) < (32 * (1U << (    UTF_ACCUMULATION_SHIFT))) ? 2 :     \
          (UV) (uv) < (16 * (1U << (2 * UTF_ACCUMULATION_SHIFT))) ? 3 :     \
          (UV) (uv) < ( 8 * (1U << (3 * UTF_ACCUMULATION_SHIFT))) ? 4 :     \
          (UV) (uv) < ( 4 * (1U << (4 * UTF_ACCUMULATION_SHIFT))) ? 5 :     \
          (UV) (uv) < ( 2 * (1U << (5 * UTF_ACCUMULATION_SHIFT))) ? 6 :

/* Internal macro to be used only in this file.
 * This adds to __COMMON_UNI_SKIP the details at this platform's upper range.
 * For any-sized EBCDIC platforms, or 64-bit ASCII ones, we need one more test
 * to see if just 7 bytes is needed, or if the maximum is needed.  For 32-bit
 * ASCII platforms, everything is representable by 7 bytes */
#if defined(UV_IS_QUAD) || defined(EBCDIC)
#   define __BASE_UNI_SKIP(uv) (__COMMON_UNI_SKIP(uv)                       \
     (UV) (uv) < ((UV) 1U << (6 * UTF_ACCUMULATION_SHIFT)) ? 7 : UTF8_MAXBYTES)
459
#else
460
#   define __BASE_UNI_SKIP(uv) (__COMMON_UNI_SKIP(uv) 7)
461 462
#endif

463 464 465
/* The next two macros use the base macro defined above, and add in the tests
 * at the low-end of the range, for just 1 byte, yielding complete macros,
 * publicly accessible. */
466

467 468
/* Input is a true Unicode (not-native) code point */
#define OFFUNISKIP(uv) (OFFUNI_IS_INVARIANT(uv) ? 1 : __BASE_UNI_SKIP(uv))
469

470
/*
471

472 473 474 475
=for apidoc Am|STRLEN|UVCHR_SKIP|UV cp
returns the number of bytes required to represent the code point C<cp> when
encoded as UTF-8.  C<cp> is a native (ASCII or EBCDIC) code point if less than
255; a Unicode code point otherwise.
476

477 478 479
=cut
 */
#define UVCHR_SKIP(uv) ( UVCHR_IS_INVARIANT(uv) ? 1 : __BASE_UNI_SKIP(uv))
480

481 482
/* The largest code point representable by two UTF-8 bytes on this platform.
 * As explained in the comments for __COMMON_UNI_SKIP, 32 start bytes with
483 484 485
 * UTF_ACCUMULATION_SHIFT bits of information each */
#define MAX_UTF8_TWO_BYTE (32 * (1U << UTF_ACCUMULATION_SHIFT) - 1)

486 487 488
/* The largest code point representable by two UTF-8 bytes on any platform that
 * Perl runs on.  This value is constrained by EBCDIC which has 5 bits per
 * continuation byte */
489 490 491 492 493 494 495 496 497 498 499 500 501 502
#define MAX_PORTABLE_UTF8_TWO_BYTE (32 * (1U << 5) - 1)

/* The maximum number of UTF-8 bytes a single Unicode character can
 * uppercase/lowercase/fold into.  Unicode guarantees that the maximum
 * expansion is UTF8_MAX_FOLD_CHAR_EXPAND characters, but any above-Unicode
 * code point will fold to itself, so we only have to look at the expansion of
 * the maximum Unicode code point.  But this number may be less than the space
 * occupied by a very large code point under Perl's extended UTF-8.  We have to
 * make it large enough to fit any single character.  (It turns out that ASCII
 * and EBCDIC differ in which is larger) */
#define UTF8_MAXBYTES_CASE	                                                \
        (UTF8_MAXBYTES >= (UTF8_MAX_FOLD_CHAR_EXPAND * OFFUNISKIP(0x10FFFF))    \
                           ? UTF8_MAXBYTES                                      \
                           : (UTF8_MAX_FOLD_CHAR_EXPAND * OFFUNISKIP(0x10FFFF)))
503 504 505 506 507 508

/* Rest of these are attributes of Unicode and perl's internals rather than the
 * encoding, or happen to be the same in both ASCII and EBCDIC (at least at
 * this level; the macros that some of these call may have different
 * definitions in the two encodings */

509 510 511 512 513 514 515 516 517 518 519
/* In domain restricted to ASCII, these may make more sense to the reader than
 * the ones with Latin1 in the name */
#define NATIVE_TO_ASCII(ch)      NATIVE_TO_LATIN1(ch)
#define ASCII_TO_NATIVE(ch)      LATIN1_TO_NATIVE(ch)

/* More or less misleadingly-named defines, retained for back compat */
#define NATIVE_TO_UTF(ch)        NATIVE_UTF8_TO_I8(ch)
#define NATIVE_TO_I8(ch)         NATIVE_UTF8_TO_I8(ch)
#define UTF_TO_NATIVE(ch)        I8_TO_NATIVE_UTF8(ch)
#define I8_TO_NATIVE(ch)         I8_TO_NATIVE_UTF8(ch)
#define NATIVE8_TO_UNI(ch)       NATIVE_TO_LATIN1(ch)
520

521 522 523 524 525 526 527 528 529 530 531
/* This defines the 1-bits that are to be in the first byte of a multi-byte
 * UTF-8 encoded character that mark it as a start byte and give the number of
 * bytes that comprise the character. 'len' is the number of bytes in the
 * multi-byte sequence. */
#define UTF_START_MARK(len) (((len) >  7) ? 0xFF : (0xFF & (0xFE << (7-(len)))))

/* Masks out the initial one bits in a start byte, leaving the real data ones.
 * Doesn't work on an invariant byte.  'len' is the number of bytes in the
 * multi-byte sequence that comprises the character. */
#define UTF_START_MASK(len) (((len) >= 7) ? 0x00 : (0x1F >> ((len)-2)))

532 533
/* Adds a UTF8 continuation byte 'new' of information to a running total code
 * point 'old' of all the continuation bytes so far.  This is designed to be
534 535 536 537
 * used in a loop to convert from UTF-8 to the code point represented.  Note
 * that this is asymmetric on EBCDIC platforms, in that the 'new' parameter is
 * the UTF-EBCDIC byte, whereas the 'old' parameter is a Unicode (not EBCDIC)
 * code point in process of being generated */
538 539
#define UTF8_ACCUMULATE(old, new) (__ASSERT_(FITS_IN_8_BITS(new))              \
                                   ((old) << UTF_ACCUMULATION_SHIFT)           \
540 541
                                   | ((NATIVE_UTF8_TO_I8((U8)new))             \
                                       & UTF_CONTINUATION_MASK))
542 543

/* This works in the face of malformed UTF-8. */
544 545 546 547
#define UTF8_IS_NEXT_CHAR_DOWNGRADEABLE(s, e)                                 \
                                       (   UTF8_IS_DOWNGRADEABLE_START(*(s))  \
                                        && ( (e) - (s) > 1)                   \
                                        && UTF8_IS_CONTINUATION(*((s)+1)))
548

549
/* Number of bytes a code point occupies in UTF-8. */
550
#define NATIVE_SKIP(uv) UVCHR_SKIP(uv)
551

552 553
/* Most code which says UNISKIP is really thinking in terms of native code
 * points (0-255) plus all those beyond.  This is an imprecise term, but having
554
 * it means existing code continues to work.  For precision, use UVCHR_SKIP,
555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574
 * NATIVE_SKIP, or OFFUNISKIP */
#define UNISKIP(uv)   UVCHR_SKIP(uv)

/* Longer, but more accurate name */
#define UTF8_IS_ABOVE_LATIN1_START(c)     UTF8_IS_ABOVE_LATIN1(c)

/* Convert a UTF-8 variant Latin1 character to a native code point value.
 * Needs just one iteration of accumulate.  Should be used only if it is known
 * that the code point is < 256, and is not UTF-8 invariant.  Use the slower
 * but more general TWO_BYTE_UTF8_TO_NATIVE() which handles any code point
 * representable by two bytes (which turns out to be up through
 * MAX_PORTABLE_UTF8_TWO_BYTE).  The two parameters are:
 *  HI: a downgradable start byte;
 *  LO: continuation.
 * */
#define EIGHT_BIT_UTF8_TO_NATIVE(HI, LO)                                        \
    ( __ASSERT_(UTF8_IS_DOWNGRADEABLE_START(HI))                                \
      __ASSERT_(UTF8_IS_CONTINUATION(LO))                                       \
     LATIN1_TO_NATIVE(UTF8_ACCUMULATE((                                         \
                            NATIVE_UTF8_TO_I8(HI) & UTF_START_MASK(2)), (LO))))
575 576

/* Convert a two (not one) byte utf8 character to a native code point value.
577 578 579 580
 * Needs just one iteration of accumulate.  Should not be used unless it is
 * known that the two bytes are legal: 1) two-byte start, and 2) continuation.
 * Note that the result can be larger than 255 if the input character is not
 * downgradable */
581
#define TWO_BYTE_UTF8_TO_NATIVE(HI, LO) \
582 583 584 585
    (__ASSERT_(FITS_IN_8_BITS(HI))                                              \
     __ASSERT_(FITS_IN_8_BITS(LO))                                              \
     __ASSERT_(PL_utf8skip[HI] == 2)                                            \
     __ASSERT_(UTF8_IS_CONTINUATION(LO))                                        \
586
     UNI_TO_NATIVE(UTF8_ACCUMULATE((NATIVE_UTF8_TO_I8(HI) & UTF_START_MASK(2)), \
587
                                   (LO))))
588 589 590

/* Should never be used, and be deprecated */
#define TWO_BYTE_UTF8_TO_UNI(HI, LO) NATIVE_TO_UNI(TWO_BYTE_UTF8_TO_NATIVE(HI, LO))
591

592 593 594 595 596 597 598 599
/*

=for apidoc Am|STRLEN|UTF8SKIP|char* s
returns the number of bytes in the UTF-8 encoded character whose first (perhaps
only) byte is pointed to by C<s>.

=cut
 */
600 601 602 603 604 605
#define UTF8SKIP(s)  PL_utf8skip[*(const U8*)(s)]
#define UTF8_SKIP(s) UTF8SKIP(s)

/* Most code that says 'UNI_' really means the native value for code points up
 * through 255 */
#define UNI_IS_INVARIANT(cp)   UVCHR_IS_INVARIANT(cp)
606

607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627
/*
=for apidoc Am|bool|UTF8_IS_INVARIANT|char c

Evaluates to 1 if the byte C<c> represents the same character when encoded in
UTF-8 as when not; otherwise evaluates to 0.  UTF-8 invariant characters can be
copied as-is when converting to/from UTF-8, saving time.

In spite of the name, this macro gives the correct result if the input string
from which C<c> comes is not encoded in UTF-8.

See C<L</UVCHR_IS_INVARIANT>> for checking if a UV is invariant.

=cut

The reason it works on both UTF-8 encoded strings and non-UTF-8 encoded, is
that it returns TRUE in each for the exact same set of bit patterns.  It is
valid on a subset of what UVCHR_IS_INVARIANT is valid on, so can just use that;
and the compiler should optimize out anything extraneous given the
implementation of the latter.  The |0 makes sure this isn't mistakenly called
with a ptr argument.
*/
628
#define UTF8_IS_INVARIANT(c)	UVCHR_IS_INVARIANT((c) | 0)
629

630 631 632
/* Like the above, but its name implies a non-UTF8 input, which as the comments
 * above show, doesn't matter as to its implementation */
#define NATIVE_BYTE_IS_INVARIANT(c)	UVCHR_IS_INVARIANT(c)
633

634 635 636 637 638 639
/* The macros in the next 4 sets are used to generate the two utf8 or utfebcdic
 * bytes from an ordinal that is known to fit into exactly two (not one) bytes;
 * it must be less than 0x3FF to work across both encodings. */

/* These two are helper macros for the other three sets, and should not be used
 * directly anywhere else.  'translate_function' is either NATIVE_TO_LATIN1
640 641
 * (which works for code points up through 0xFF) or NATIVE_TO_UNI which works
 * for any code point */
642
#define __BASE_TWO_BYTE_HI(c, translate_function)                               \
643
           (__ASSERT_(! UVCHR_IS_INVARIANT(c))                                  \
644
            I8_TO_NATIVE_UTF8((translate_function(c) >> UTF_ACCUMULATION_SHIFT) \
645
                              | UTF_START_MARK(2)))
646
#define __BASE_TWO_BYTE_LO(c, translate_function)                               \
647
             (__ASSERT_(! UVCHR_IS_INVARIANT(c))                                \
648
              I8_TO_NATIVE_UTF8((translate_function(c) & UTF_CONTINUATION_MASK) \
649
                                 | UTF_CONTINUATION_MARK))
650 651 652 653 654 655 656 657 658 659

/* The next two macros should not be used.  They were designed to be usable as
 * the case label of a switch statement, but this doesn't work for EBCDIC.  Use
 * regen/unicode_constants.pl instead */
#define UTF8_TWO_BYTE_HI_nocast(c)  __BASE_TWO_BYTE_HI(c, NATIVE_TO_UNI)
#define UTF8_TWO_BYTE_LO_nocast(c)  __BASE_TWO_BYTE_LO(c, NATIVE_TO_UNI)

/* The next two macros are used when the source should be a single byte
 * character; checked for under DEBUGGING */
#define UTF8_EIGHT_BIT_HI(c) (__ASSERT_(FITS_IN_8_BITS(c))                    \
660
                             ( __BASE_TWO_BYTE_HI(c, NATIVE_TO_LATIN1)))
661
#define UTF8_EIGHT_BIT_LO(c) (__ASSERT_(FITS_IN_8_BITS(c))                    \
662
                             (__BASE_TWO_BYTE_LO(c, NATIVE_TO_LATIN1)))
663 664 665 666 667

/* These final two macros in the series are used when the source can be any
 * code point whose UTF-8 is known to occupy 2 bytes; they are less efficient
 * than the EIGHT_BIT versions on EBCDIC platforms.  We use the logical '~'
 * operator instead of "<=" to avoid getting compiler warnings.
668
 * MAX_UTF8_TWO_BYTE should be exactly all one bits in the lower few
669 670 671
 * places, so the ~ works */
#define UTF8_TWO_BYTE_HI(c)                                                    \
       (__ASSERT_((sizeof(c) ==  1)                                            \
672
                  || !(((WIDEST_UTYPE)(c)) & ~MAX_UTF8_TWO_BYTE))              \
673
        (__BASE_TWO_BYTE_HI(c, NATIVE_TO_UNI)))
674 675
#define UTF8_TWO_BYTE_LO(c)                                                    \
       (__ASSERT_((sizeof(c) ==  1)                                            \
676
                  || !(((WIDEST_UTYPE)(c)) & ~MAX_UTF8_TWO_BYTE))              \
677
        (__BASE_TWO_BYTE_LO(c, NATIVE_TO_UNI)))
678 679 680 681

/* This is illegal in any well-formed UTF-8 in both EBCDIC and ASCII
 * as it is only in overlongs. */
#define ILLEGAL_UTF8_BYTE   I8_TO_NATIVE_UTF8(0xC1)
682

683
/*
684 685 686 687 688
 * 'UTF' is whether or not p is encoded in UTF8.  The names 'foo_lazy_if' stem
 * from an earlier version of these macros in which they didn't call the
 * foo_utf8() macros (i.e. were 'lazy') unless they decided that *p is the
 * beginning of a utf8 character.  Now that foo_utf8() determines that itself,
 * no need to do it again here
689
 */
690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713
#define isIDFIRST_lazy_if(p,UTF)                                            \
            _is_utf8_FOO(_CC_IDFIRST, (const U8 *) p, "isIDFIRST_lazy_if",  \
                         "isIDFIRST_lazy_if_safe",                          \
                         cBOOL(UTF && ! IN_BYTES), 0, __FILE__,__LINE__)

#define isIDFIRST_lazy_if_safe(p, e, UTF)                                   \
                   ((IN_BYTES || !UTF)                                      \
                     ? isIDFIRST(*(p))                                      \
                     : isIDFIRST_utf8_safe(p, e))

#define isWORDCHAR_lazy_if(p,UTF)                                           \
            _is_utf8_FOO(_CC_IDFIRST, (const U8 *) p, "isWORDCHAR_lazy_if", \
                         "isWORDCHAR_lazy_if_safe",                         \
                         cBOOL(UTF && ! IN_BYTES), 0, __FILE__,__LINE__)

#define isWORDCHAR_lazy_if_safe(p, e, UTF)                                  \
                   ((IN_BYTES || !UTF)                                      \
                     ? isWORDCHAR(*(p))                                     \
                     : isWORDCHAR_utf8_safe((U8 *) p, (U8 *) e))

#define isALNUM_lazy_if(p,UTF)                                              \
            _is_utf8_FOO(_CC_IDFIRST, (const U8 *) p, "isALNUM_lazy_if",    \
                         "isWORDCHAR_lazy_if_safe",                         \
                         cBOOL(UTF && ! IN_BYTES), 0, __FILE__,__LINE__)
714

715 716
#define UTF8_MAXLEN UTF8_MAXBYTES

717 718
/* A Unicode character can fold to up to 3 characters */
#define UTF8_MAX_FOLD_CHAR_EXPAND 3
719

720
#define IN_BYTES UNLIKELY(CopHINTS_get(PL_curcop) & HINT_BYTES)
721 722 723 724 725 726 727 728 729 730 731 732

/*

=for apidoc Am|bool|DO_UTF8|SV* sv
Returns a bool giving whether or not the PV in C<sv> is to be treated as being
encoded in UTF-8.

You should use this I<after> a call to C<SvPV()> or one of its variants, in
case any call to string overloading updates the internal UTF-8 encoding flag.

=cut
*/
733
#define DO_UTF8(sv) (SvUTF8(sv) && !IN_BYTES)
734 735 736 737 738

/* Should all strings be treated as Unicode, and not just UTF-8 encoded ones?
 * Is so within 'feature unicode_strings' or 'locale :not_characters', and not
 * within 'use bytes'.  UTF-8 locales are not tested for here, but perhaps
 * could be */
739 740 741 742 743 744
#define IN_UNI_8_BIT                                                    \
	    ((    (      (CopHINTS_get(PL_curcop) & HINT_UNI_8_BIT))    \
                   || (   CopHINTS_get(PL_curcop) & HINT_LOCALE_PARTIAL \
                            /* -1 below is for :not_characters */       \
                       && _is_in_locale_category(FALSE, -1)))           \
              && (! IN_BYTES))
745

746

747
#define UTF8_ALLOW_EMPTY		0x0001	/* Allow a zero length string */
748
#define UTF8_GOT_EMPTY                  UTF8_ALLOW_EMPTY
749 750

/* Allow first byte to be a continuation byte */
751
#define UTF8_ALLOW_CONTINUATION		0x0002
752
#define UTF8_GOT_CONTINUATION		UTF8_ALLOW_CONTINUATION
753

754
/* Unexpected continuation byte */
755
#define UTF8_ALLOW_NON_CONTINUATION	0x0004
756
#define UTF8_GOT_NON_CONTINUATION	UTF8_ALLOW_NON_CONTINUATION
757 758 759

/* expecting more bytes than were available in the string */
#define UTF8_ALLOW_SHORT		0x0008
760
#define UTF8_GOT_SHORT		        UTF8_ALLOW_SHORT
761

762 763 764
/* Overlong sequence; i.e., the code point can be specified in fewer bytes.
 * First one will convert the overlong to the REPLACEMENT CHARACTER; second
 * will return what the overlong evaluates to */
765
#define UTF8_ALLOW_LONG                 0x0010
766 767 768 769 770
#define UTF8_ALLOW_LONG_AND_ITS_VALUE   (UTF8_ALLOW_LONG|0x0020)
#define UTF8_GOT_LONG                   UTF8_ALLOW_LONG

#define UTF8_ALLOW_OVERFLOW             0x0080
#define UTF8_GOT_OVERFLOW               UTF8_ALLOW_OVERFLOW
771

772 773 774
#define UTF8_DISALLOW_SURROGATE		0x0100	/* Unicode surrogates */
#define UTF8_GOT_SURROGATE		UTF8_DISALLOW_SURROGATE
#define UTF8_WARN_SURROGATE		0x0200
775

776 777
/* Unicode non-character  code points */
#define UTF8_DISALLOW_NONCHAR           0x0400
778
#define UTF8_GOT_NONCHAR                UTF8_DISALLOW_NONCHAR
779
#define UTF8_WARN_NONCHAR               0x0800
780

781 782
/* Super-set of Unicode: code points above the legal max */
#define UTF8_DISALLOW_SUPER		0x1000
783
#define UTF8_GOT_SUPER		        UTF8_DISALLOW_SUPER
784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805
#define UTF8_WARN_SUPER		        0x2000

/* The original UTF-8 standard did not define UTF-8 with start bytes of 0xFE or
 * 0xFF, though UTF-EBCDIC did.  This allowed both versions to represent code
 * points up to 2 ** 31 - 1.  Perl extends UTF-8 so that 0xFE and 0xFF are
 * usable on ASCII platforms, and 0xFF means something different than
 * UTF-EBCDIC defines.  These changes allow code points of 64 bits (actually
 * somewhat more) to be represented on both platforms.  But these are Perl
 * extensions, and not likely to be interchangeable with other languages.  Note
 * that on ASCII platforms, FE overflows a signed 32-bit word, and FF an
 * unsigned one. */
#define UTF8_DISALLOW_PERL_EXTENDED     0x4000
#define UTF8_GOT_PERL_EXTENDED          UTF8_DISALLOW_PERL_EXTENDED
#define UTF8_WARN_PERL_EXTENDED         0x8000

/* For back compat, these old names are misleading for overlongs and
 * UTF_EBCDIC. */
#define UTF8_DISALLOW_ABOVE_31_BIT      UTF8_DISALLOW_PERL_EXTENDED
#define UTF8_GOT_ABOVE_31_BIT           UTF8_GOT_PERL_EXTENDED
#define UTF8_WARN_ABOVE_31_BIT          UTF8_WARN_PERL_EXTENDED
#define UTF8_DISALLOW_FE_FF             UTF8_DISALLOW_PERL_EXTENDED
#define UTF8_WARN_FE_FF                 UTF8_WARN_PERL_EXTENDED
806

807 808
#define UTF8_CHECK_ONLY			0x10000
#define _UTF8_NO_CONFIDENCE_IN_CURLEN   0x20000  /* Internal core use only */
809 810 811 812 813

/* For backwards source compatibility.  They do nothing, as the default now
 * includes what they used to mean.  The first one's meaning was to allow the
 * just the single non-character 0xFFFF */
#define UTF8_ALLOW_FFFF 0
814
#define UTF8_ALLOW_FE_FF 0
815 816
#define UTF8_ALLOW_SURROGATE 0

817 818 819 820 821
/* C9 refers to Unicode Corrigendum #9: allows but discourages non-chars */
#define UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE                                    \
                                 (UTF8_DISALLOW_SUPER|UTF8_DISALLOW_SURROGATE)
#define UTF8_WARN_ILLEGAL_C9_INTERCHANGE (UTF8_WARN_SUPER|UTF8_WARN_SURROGATE)

822
#define UTF8_DISALLOW_ILLEGAL_INTERCHANGE                                       \
823
                  (UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE|UTF8_DISALLOW_NONCHAR)
824
#define UTF8_WARN_ILLEGAL_INTERCHANGE \
825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880
                          (UTF8_WARN_ILLEGAL_C9_INTERCHANGE|UTF8_WARN_NONCHAR)

/* This is typically used for code that processes UTF-8 input and doesn't want
 * to have to deal with any malformations that might be present.  All such will
 * be safely replaced by the REPLACEMENT CHARACTER, unless other flags
 * overriding this are also present. */
#define UTF8_ALLOW_ANY ( UTF8_ALLOW_CONTINUATION                                \
                        |UTF8_ALLOW_NON_CONTINUATION                            \
                        |UTF8_ALLOW_SHORT                                       \
                        |UTF8_ALLOW_LONG                                        \
                        |UTF8_ALLOW_OVERFLOW)

/* Accept any Perl-extended UTF-8 that evaluates to any UV on the platform, but
 * not any malformed.  This is the default.  (Note that UVs above IV_MAX are
 * deprecated. */
#define UTF8_ALLOW_ANYUV   0
#define UTF8_ALLOW_DEFAULT UTF8_ALLOW_ANYUV

/*
=for apidoc Am|bool|UTF8_IS_SURROGATE|const U8 *s|const U8 *e

Evaluates to non-zero if the first few bytes of the string starting at C<s> and
looking no further than S<C<e - 1>> are well-formed UTF-8 that represents one
of the Unicode surrogate code points; otherwise it evaluates to 0.  If
non-zero, the value gives how many bytes starting at C<s> comprise the code
point's representation.

=cut
 */
#define UTF8_IS_SURROGATE(s, e)      is_SURROGATE_utf8_safe(s, e)


#define UTF8_IS_REPLACEMENT(s, send) is_REPLACEMENT_utf8_safe(s,send)

/*
=for apidoc Am|bool|UTF8_IS_SUPER|const U8 *s|const U8 *e

Recall that Perl recognizes an extension to UTF-8 that can encode code
points larger than the ones defined by Unicode, which are 0..0x10FFFF.

This macro evaluates to non-zero if the first few bytes of the string starting
at C<s> and looking no further than S<C<e - 1>> are from this UTF-8 extension;
otherwise it evaluates to 0.  If non-zero, the value gives how many bytes
starting at C<s> comprise the code point's representation.

0 is returned if the bytes are not well-formed extended UTF-8, or if they
represent a code point that cannot fit in a UV on the current platform.  Hence
this macro can give different results when run on a 64-bit word machine than on
one with a 32-bit word size.

Note that it is deprecated to have code points that are larger than what can
fit in an IV on the current machine.

=cut

 *		  ASCII		     EBCDIC I8
881 882 883
 * U+10FFFF: \xF4\x8F\xBF\xBF	\xF9\xA1\xBF\xBF\xBF	max legal Unicode
 * U+110000: \xF4\x90\x80\x80	\xF9\xA2\xA0\xA0\xA0
 * U+110001: \xF4\x90\x80\x81	\xF9\xA2\xA0\xA0\xA1
884 885 886 887 888 889 890 891 892
 */
#ifdef EBCDIC
#   define UTF8_IS_SUPER(s, e)                                              \
                  ((    LIKELY((e) > (s) + 4)                               \
                    &&      NATIVE_UTF8_TO_I8(*(s)) >= 0xF9                 \
                    && (    NATIVE_UTF8_TO_I8(*(s)) >  0xF9                 \
                        || (NATIVE_UTF8_TO_I8(*((s) + 1)) >= 0xA2))         \
                    &&  LIKELY((s) + UTF8SKIP(s) <= (e)))                   \
                    ? _is_utf8_char_helper(s, s + UTF8SKIP(s), 0) : 0)
893
#else
894 895 896 897 898 899
#   define UTF8_IS_SUPER(s, e)                                              \
                   ((    LIKELY((e) > (s) + 3)                              \
                     &&  (*(U8*) (s)) >= 0xF4                               \
                     && ((*(U8*) (s)) >  0xF4 || (*((U8*) (s) + 1) >= 0x90))\
                     &&  LIKELY((s) + UTF8SKIP(s) <= (e)))                  \
                    ? _is_utf8_char_helper(s, s + UTF8SKIP(s), 0) : 0)
900 901
#endif

902 903
/* These are now machine generated, and the 'given' clause is no longer
 * applicable */
904
#define UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s, e)          \
905 906 907 908 909 910 911 912 913 914 915 916 917
                                            cBOOL(is_NONCHAR_utf8_safe(s,e))

/*
=for apidoc Am|bool|UTF8_IS_NONCHAR|const U8 *s|const U8 *e

Evaluates to non-zero if the first few bytes of the string starting at C<s> and
looking no further than S<C<e - 1>> are well-formed UTF-8 that represents one
of the Unicode non-character code points; otherwise it evaluates to 0.  If
non-zero, the value gives how many bytes starting at C<s> comprise the code
point's representation.

=cut
 */
918 919
#define UTF8_IS_NONCHAR(s, e)                                                  \
                UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s, e)
920

921 922 923 924 925 926
#define UNICODE_SURROGATE_FIRST		0xD800
#define UNICODE_SURROGATE_LAST		0xDFFF
#define UNICODE_REPLACEMENT		0xFFFD
#define UNICODE_BYTE_ORDER_MARK		0xFEFF

/* Though our UTF-8 encoding can go beyond this,
927
 * let's be conservative and do as Unicode says. */
928 929
#define PERL_UNICODE_MAX	0x10FFFF

930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945
#define UNICODE_WARN_SURROGATE         0x0001	/* UTF-16 surrogates */
#define UNICODE_WARN_NONCHAR           0x0002	/* Non-char code points */
#define UNICODE_WARN_SUPER             0x0004	/* Above 0x10FFFF */
#define UNICODE_WARN_PERL_EXTENDED     0x0008	/* Above 0x7FFF_FFFF */
#define UNICODE_WARN_ABOVE_31_BIT      UNICODE_WARN_PERL_EXTENDED
#define UNICODE_DISALLOW_SURROGATE     0x0010
#define UNICODE_DISALLOW_NONCHAR       0x0020
#define UNICODE_DISALLOW_SUPER         0x0040
#define UNICODE_DISALLOW_PERL_EXTENDED 0x0080
#define UNICODE_DISALLOW_ABOVE_31_BIT  UNICODE_DISALLOW_PERL_EXTENDED

#define UNICODE_GOT_SURROGATE       UNICODE_DISALLOW_SURROGATE
#define UNICODE_GOT_NONCHAR         UNICODE_DISALLOW_NONCHAR
#define UNICODE_GOT_SUPER           UNICODE_DISALLOW_SUPER
#define UNICODE_GOT_PERL_EXTENDED   UNICODE_DISALLOW_PERL_EXTENDED

946 947
#define UNICODE_WARN_ILLEGAL_C9_INTERCHANGE                                   \
                                  (UNICODE_WARN_SURROGATE|UNICODE_WARN_SUPER)
948
#define UNICODE_WARN_ILLEGAL_INTERCHANGE                                      \
949 950 951
                   (UNICODE_WARN_ILLEGAL_C9_INTERCHANGE|UNICODE_WARN_NONCHAR)
#define UNICODE_DISALLOW_ILLEGAL_C9_INTERCHANGE                               \
                          (UNICODE_DISALLOW_SURROGATE|UNICODE_DISALLOW_SUPER)
952
#define UNICODE_DISALLOW_ILLEGAL_INTERCHANGE                                  \
953
           (UNICODE_DISALLOW_ILLEGAL_C9_INTERCHANGE|UNICODE_DISALLOW_NONCHAR)
954 955 956 957 958

/* For backward source compatibility, as are now the default */
#define UNICODE_ALLOW_SURROGATE 0
#define UNICODE_ALLOW_SUPER	0
#define UNICODE_ALLOW_ANY	0
959

960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983
/* This matches the 2048 code points between UNICODE_SURROGATE_FIRST (0xD800) and
 * UNICODE_SURROGATE_LAST (0xDFFF) */
#define UNICODE_IS_SURROGATE(uv)        (((UV) (uv) & (~0xFFFF | 0xF800))       \
                                                                    == 0xD800)

#define UNICODE_IS_REPLACEMENT(uv)	((UV) (uv) == UNICODE_REPLACEMENT)
#define UNICODE_IS_BYTE_ORDER_MARK(uv)	((UV) (uv) == UNICODE_BYTE_ORDER_MARK)

/* Is 'uv' one of the 32 contiguous-range noncharacters? */
#define UNICODE_IS_32_CONTIGUOUS_NONCHARS(uv)      ((UV) (uv) >= 0xFDD0         \
                                                 && (UV) (uv) <= 0xFDEF)

/* Is 'uv' one of the 34 plane-ending noncharacters 0xFFFE, 0xFFFF, 0x1FFFE,
 * 0x1FFFF, ... 0x10FFFE, 0x10FFFF, given that we know that 'uv' is not above
 * the Unicode legal max */
#define UNICODE_IS_END_PLANE_NONCHAR_GIVEN_NOT_SUPER(uv)                        \
                                              (((UV) (uv) & 0xFFFE) == 0xFFFE)

#define UNICODE_IS_NONCHAR(uv)                                                  \
    (   UNICODE_IS_32_CONTIGUOUS_NONCHARS(uv)                                   \
     || (   LIKELY( ! UNICODE_IS_SUPER(uv))                                     \
         && UNICODE_IS_END_PLANE_NONCHAR_GIVEN_NOT_SUPER(uv)))

#define UNICODE_IS_SUPER(uv)    ((UV) (uv) > PERL_UNICODE_MAX)
984

985 986 987 988 989 990 991 992
#define LATIN_SMALL_LETTER_SHARP_S      LATIN_SMALL_LETTER_SHARP_S_NATIVE
#define LATIN_SMALL_LETTER_Y_WITH_DIAERESIS                                  \
                                LATIN_SMALL_LETTER_Y_WITH_DIAERESIS_NATIVE
#define MICRO_SIGN      MICRO_SIGN_NATIVE
#define LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE                               \
                            LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE_NATIVE
#define LATIN_SMALL_LETTER_A_WITH_RING_ABOVE                                 \
                                LATIN_SMALL_LETTER_A_WITH_RING_ABOVE_NATIVE
993 994 995
#define UNICODE_GREEK_CAPITAL_LETTER_SIGMA	0x03A3
#define UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA	0x03C2
#define UNICODE_GREEK_SMALL_LETTER_SIGMA	0x03C3
996
#define GREEK_SMALL_LETTER_MU                   0x03BC
997 998 999
#define GREEK_CAPITAL_LETTER_MU                 0x039C	/* Upper and title case
                                                           of MICRON */
#define LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS   0x0178	/* Also is title case */
1000 1001 1002 1003 1004
#ifdef LATIN_CAPITAL_LETTER_SHARP_S_UTF8
#   define LATIN_CAPITAL_LETTER_SHARP_S	        0x1E9E
#endif
#define LATIN_CAPITAL_LETTER_I_WITH_DOT_ABOVE   0x130
#define LATIN_SMALL_LETTER_DOTLESS_I            0x131
1005 1006 1007 1008 1009
#define LATIN_SMALL_LETTER_LONG_S               0x017F
#define LATIN_SMALL_LIGATURE_LONG_S_T           0xFB05
#define LATIN_SMALL_LIGATURE_ST                 0xFB06
#define KELVIN_SIGN                             0x212A
#define ANGSTROM_SIGN                           0x212B
1010 1011 1012 1013 1014 1015

#define UNI_DISPLAY_ISPRINT	0x0001
#define UNI_DISPLAY_BACKSLASH	0x0002
#define UNI_DISPLAY_QQ		(UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)
#define UNI_DISPLAY_REGEX	(UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)

1016 1017
#define ANYOF_FOLD_SHARP_S(node, input, end)	\
	(ANYOF_BITMAP_TEST(node, LATIN_SMALL_LETTER_SHARP_S) && \
1018 1019
	 (ANYOF_NONBITMAP(node)) && \
	 (ANYOF_FLAGS(node) & ANYOF_LOC_NONBITMAP_FOLD) && \
1020
	 ((end) > (input) + 1) && \
1021 1022
	 isALPHA_FOLD_EQ((input)[0], 's'))

1023
#define SHARP_S_SKIP 2
1024

1025
/*
1026

1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071
=for apidoc Am|STRLEN|isUTF8_CHAR|const U8 *s|const U8 *e

Evaluates to non-zero if the first few bytes of the string starting at C<s> and
looking no further than S<C<e - 1>> are well-formed UTF-8, as extended by Perl,
that represents some code point; otherwise it evaluates to 0.  If non-zero, the
value gives how many bytes starting at C<s> comprise the code point's
representation.  Any bytes remaining before C<e>, but beyond the ones needed to
form the first code point in C<s>, are not examined.

The code point can be any that will fit in a UV on this machine, using Perl's
extension to official UTF-8 to represent those higher than the Unicode maximum
of 0x10FFFF.  That means that this macro is used to efficiently decide if the
next few bytes in C<s> is legal UTF-8 for a single character.

Use C<L</isSTRICT_UTF8_CHAR>> to restrict the acceptable code points to those
defined by Unicode to be fully interchangeable across applications;
C<L</isC9_STRICT_UTF8_CHAR>> to use the L<Unicode Corrigendum
#9|http://www.unicode.org/versions/corrigendum9.html> definition of allowable
code points; and C<L</isUTF8_CHAR_flags>> for a more customized definition.

Use C<L</is_utf8_string>>, C<L</is_utf8_string_loc>>, and
C<L</is_utf8_string_loclen>> to check entire strings.

Note that it is deprecated to use code points higher than what will fit in an
IV.  This macro does not raise any warnings for such code points, treating them
as valid.

Note also that a UTF-8 INVARIANT character (i.e. ASCII on non-EBCDIC machines)
is a valid UTF-8 character.

=cut
*/

#define isUTF8_CHAR(s, e)                                                   \
    (UNLIKELY((e) <= (s))                                                   \
    ? 0                                                                     \
    : (UTF8_IS_INVARIANT(*s))                                               \
      ? 1                                                                   \
      : UNLIKELY(((e) - (s)) < UTF8SKIP(s))                                 \
        ? 0                                                                 \
        : LIKELY(NATIVE_UTF8_TO_I8(*s) <= _IS_UTF8_CHAR_HIGHEST_START_BYTE) \
          ? is_UTF8_CHAR_utf8_no_length_checks(s)                           \
          : _is_utf8_char_helper(s, e, 0))

#define is_utf8_char_buf(buf, buf_end) isUTF8_CHAR(buf, buf_end)
1072 1073
#define bytes_from_utf8(s, lenp, is_utf8p)                                  \
                            bytes_from_utf8_loc(s, lenp, is_utf8p, 0)
1074

1075 1076
/*

1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102
=for apidoc Am|STRLEN|isSTRICT_UTF8_CHAR|const U8 *s|const U8 *e

Evaluates to non-zero if the first few bytes of the string starting at C<s> and
looking no further than S<C<e - 1>> are well-formed UTF-8 that represents some
Unicode code point completely acceptable for open interchange between all
applications; otherwise it evaluates to 0.  If non-zero, the value gives how
many bytes starting at C<s> comprise the code point's representation.  Any
bytes remaining before C<e>, but beyond the ones needed to form the first code
point in C<s>, are not examined.

The largest acceptable code point is the Unicode maximum 0x10FFFF, and must not
be a surrogate nor a non-character code point.  Thus this excludes any code
point from Perl's extended UTF-8.

This is used to efficiently decide if the next few bytes in C<s> is
legal Unicode-acceptable UTF-8 for a single character.

Use C<L</isC9_STRICT_UTF8_CHAR>> to use the L<Unicode Corrigendum
#9|http://www.unicode.org/versions/corrigendum9.html> definition of allowable
code points; C<L</isUTF8_CHAR>> to check for Perl's extended UTF-8;
and C<L</isUTF8_CHAR_flags>> for a more customized definition.

Use C<L</is_strict_utf8_string>>, C<L</is_strict_utf8_string_loc>>, and
C<L</is_strict_utf8_string_loclen>> to check entire strings.

=cut
1103
*/
1104 1105 1106 1107 1108 1109 1110 1111 1112

#define isSTRICT_UTF8_CHAR(s, e)                                            \
    (UNLIKELY((e) <= (s))                                                   \
    ? 0                                                                     \
    : (UTF8_IS_INVARIANT(*s))                                               \
      ? 1                                                                   \
      : UNLIKELY(((e) - (s)) < UTF8SKIP(s))                                 \
        ? 0                                                                 \
        : is_STRICT_UTF8_CHAR_utf8_no_length_checks(s))
1113

1114
/*
1115

1116
=for apidoc Am|STRLEN|isC9_STRICT_UTF8_CHAR|const U8 *s|const U8 *e
1117

1118 1119 1120 1121 1122 1123
Evaluates to non-zero if the first few bytes of the string starting at C<s> and
looking no further than S<C<e - 1>> are well-formed UTF-8 that represents some
Unicode non-surrogate code point; otherwise it evaluates to 0.  If non-zero,
the value gives how many bytes starting at C<s> comprise the code point's
representation.  Any bytes remaining before C<e>, but beyond the ones needed to
form the first code point in C<s>, are not examined.
1124

1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137
The largest acceptable code point is the Unicode maximum 0x10FFFF.  This
differs from C<L</isSTRICT_UTF8_CHAR>> only in that it accepts non-character
code points.  This corresponds to
L<Unicode Corrigendum #9|http://www.unicode.org/versions/corrigendum9.html>.
which said that non-character code points are merely discouraged rather than
completely forbidden in open interchange.  See
L<perlunicode/Noncharacter code points>.

Use C<L</isUTF8_CHAR>> to check for Perl's extended UTF-8; and
C<L</isUTF8_CHAR_flags>> for a more customized definition.

Use C<L</is_c9strict_utf8_string>>, C<L</is_c9strict_utf8_string_loc>>, and
C<L</is_c9strict_utf8_string_loclen>> to check entire strings.
1138 1139 1140 1141

=cut
*/

1142 1143 1144 1145 1146 1147 1148 1149
#define isC9_STRICT_UTF8_CHAR(s, e)                                         \
    (UNLIKELY((e) <= (s))                                                   \
    ? 0                                                                     \
    : (UTF8_IS_INVARIANT(*s))                                               \
      ? 1                                                                   \
      : UNLIKELY(((e) - (s)) < UTF8SKIP(s))                                 \
        ? 0                                                                 \
        : is_C9_STRICT_UTF8_CHAR_utf8_no_length_checks(s))
1150

1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188
/*

=for apidoc Am|STRLEN|isUTF8_CHAR_flags|const U8 *s|const U8 *e| const U32 flags

Evaluates to non-zero if the first few bytes of the string starting at C<s> and
looking no further than S<C<e - 1>> are well-formed UTF-8, as extended by Perl,
that represents some code point, subject to the restrictions given by C<flags>;
otherwise it evaluates to 0.  If non-zero, the value gives how many bytes
starting at C<s> comprise the code point's representation.  Any bytes remaining
before C<e>, but beyond the ones needed to form the first code point in C<s>,
are not examined.

If C<flags> is 0, this gives the same results as C<L</isUTF8_CHAR>>;
if C<flags> is C<UTF8_DISALLOW_ILLEGAL_INTERCHANGE>, this gives the same results
as C<L</isSTRICT_UTF8_CHAR>>;
and if C<flags> is C<UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE>, this gives
the same results as C<L</isC9_STRICT_UTF8_CHAR>>.
Otherwise C<flags> may be any combination of the C<UTF8_DISALLOW_I<foo>> flags
understood by C<L</utf8n_to_uvchr>>, with the same meanings.

The three alternative macros are for the most commonly needed validations; they
are likely to run somewhat faster than this more general one, as they can be
inlined into your code.

Use L</is_utf8_string_flags>, L</is_utf8_string_loc_flags>, and
L</is_utf8_string_loclen_flags> to check entire strings.

=cut
*/

#define isUTF8_CHAR_flags(s, e, flags)                                      \
    (UNLIKELY((e) <= (s))                                                   \
    ? 0                                                                     \
    : (UTF8_IS_INVARIANT(*s))                                               \
      ? 1                                                                   \
      : UNLIKELY(((e) - (s)) < UTF8SKIP(s))                                 \
        ? 0                                                                 \
        : _is_utf8_char_helper(s, e, flags))
1189 1190 1191 1192

/* Do not use; should be deprecated.  Use isUTF8_CHAR() instead; this is
 * retained solely for backwards compatibility */
#define IS_UTF8_CHAR(p, n)      (isUTF8_CHAR(p, (p) + (n)) == n)
1193

1194
#endif /* PERL_UTF8_H_ */
1195 1196

/*
1197
 * ex: set ts=8 sts=4 sw=4 et:
1198
 */