| 1 | /* handy.h |
| 2 | * |
| 3 | * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000, |
| 4 | * 2001, 2002, 2004, 2005, 2006, 2007, 2008, 2012 by Larry Wall and others |
| 5 | * |
| 6 | * You may distribute under the terms of either the GNU General Public |
| 7 | * License or the Artistic License, as specified in the README file. |
| 8 | * |
| 9 | */ |
| 10 | |
| 11 | /* IMPORTANT NOTE: Everything whose name begins with an underscore is for |
| 12 | * internal core Perl use only. */ |
| 13 | |
| 14 | #ifndef PERL_HANDY_H_ /* Guard against nested #inclusion */ |
| 15 | #define PERL_HANDY_H_ |
| 16 | |
| 17 | #ifndef PERL_CORE |
| 18 | # define Null(type) ((type)NULL) |
| 19 | |
| 20 | /* |
| 21 | =for apidoc_section String Handling |
| 22 | =for apidoc AmnU||Nullch |
| 23 | Null character pointer. (No longer available when C<PERL_CORE> is |
| 24 | defined.) |
| 25 | |
| 26 | =for apidoc_section SV Handling |
| 27 | =for apidoc AmnU||Nullsv |
| 28 | Null SV pointer. (No longer available when C<PERL_CORE> is defined.) |
| 29 | |
| 30 | =cut |
| 31 | |
| 32 | Below are signatures of functions from config.h which can't easily be gleaned |
| 33 | from it, and are very unlikely to change |
| 34 | |
| 35 | =for apidoc_section Signals |
| 36 | =for apidoc Am|int|Sigsetjmp|jmp_buf env|int savesigs |
| 37 | =for apidoc Am|void|Siglongjmp|jmp_buf env|int val |
| 38 | |
| 39 | =for apidoc_section Filesystem configuration values |
| 40 | =for apidoc Am|void *|FILE_ptr|FILE * f |
| 41 | =for apidoc Am|Size_t|FILE_cnt|FILE * f |
| 42 | =for apidoc Am|void *|FILE_base|FILE * f |
| 43 | =for apidoc Am|Size_t|FILE_bufsiz |
| 44 | |
| 45 | =for apidoc_section String Handling |
| 46 | =for apidoc Am|token|CAT2|token x|token y |
| 47 | =for apidoc Am|string|STRINGIFY|token x |
| 48 | |
| 49 | =for apidoc_section Numeric Functions |
| 50 | =for apidoc Am|double|Drand01|double x |
| 51 | =for apidoc Am|void|seedDrand01|Rand_seed_t x |
| 52 | =for apidoc Am|char *|Gconvert|double x|Size_t n|bool t|char * b |
| 53 | |
| 54 | =cut |
| 55 | */ |
| 56 | |
| 57 | # define Nullch Null(char*) |
| 58 | # define Nullfp Null(PerlIO*) |
| 59 | # define Nullsv Null(SV*) |
| 60 | #endif |
| 61 | |
| 62 | #ifdef TRUE |
| 63 | #undef TRUE |
| 64 | #endif |
| 65 | #ifdef FALSE |
| 66 | #undef FALSE |
| 67 | #endif |
| 68 | #define TRUE (1) |
| 69 | #define FALSE (0) |
| 70 | |
| 71 | /* The MUTABLE_*() macros cast pointers to the types shown, in such a way |
| 72 | * (compiler permitting) that casting away const-ness will give a warning; |
| 73 | * e.g.: |
| 74 | * |
| 75 | * const SV *sv = ...; |
| 76 | * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away |
| 77 | * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn |
| 78 | */ |
| 79 | |
| 80 | #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN) |
| 81 | # define MUTABLE_PTR(p) ({ void *p_ = (p); p_; }) |
| 82 | #else |
| 83 | # define MUTABLE_PTR(p) ((void *) (p)) |
| 84 | #endif |
| 85 | |
| 86 | #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p)) |
| 87 | #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p)) |
| 88 | #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p)) |
| 89 | #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p)) |
| 90 | #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p)) |
| 91 | #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p)) |
| 92 | |
| 93 | #if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR) |
| 94 | # include <stdbool.h> |
| 95 | # ifndef HAS_BOOL |
| 96 | # define HAS_BOOL 1 |
| 97 | # endif |
| 98 | #endif |
| 99 | |
| 100 | /* bool is built-in for g++-2.6.3 and later, which might be used |
| 101 | for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't |
| 102 | be sure _G_config.h will be included before this file. _G_config.h |
| 103 | also defines _G_HAVE_BOOL for both gcc and g++, but only g++ |
| 104 | actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us. |
| 105 | g++ can be identified by __GNUG__. |
| 106 | Andy Dougherty February 2000 |
| 107 | */ |
| 108 | #ifdef __GNUG__ /* GNU g++ has bool built-in */ |
| 109 | # ifndef PERL_BOOL_AS_CHAR |
| 110 | # ifndef HAS_BOOL |
| 111 | # define HAS_BOOL 1 |
| 112 | # endif |
| 113 | # endif |
| 114 | #endif |
| 115 | |
| 116 | #ifndef HAS_BOOL |
| 117 | # ifdef bool |
| 118 | # undef bool |
| 119 | # endif |
| 120 | # define bool char |
| 121 | # define HAS_BOOL 1 |
| 122 | #endif |
| 123 | |
| 124 | /* |
| 125 | =for apidoc_section Casting |
| 126 | =for apidoc Am|bool|cBOOL|bool expr |
| 127 | |
| 128 | Cast-to-bool. A simple S<C<(bool) I<expr>>> cast may not do the right thing: |
| 129 | if C<bool> is defined as C<char>, for example, then the cast from C<int> is |
| 130 | implementation-defined. |
| 131 | |
| 132 | C<(bool)!!(cbool)> in a ternary triggers a bug in xlc on AIX |
| 133 | |
| 134 | =cut |
| 135 | */ |
| 136 | #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0) |
| 137 | |
| 138 | /* Try to figure out __func__ or __FUNCTION__ equivalent, if any. |
| 139 | * XXX Should really be a Configure probe, with HAS__FUNCTION__ |
| 140 | * and FUNCTION__ as results. |
| 141 | * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */ |
| 142 | #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */ |
| 143 | # define FUNCTION__ __func__ |
| 144 | #elif (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */ |
| 145 | # define FUNCTION__ "" |
| 146 | #else |
| 147 | # define FUNCTION__ __FUNCTION__ /* Common extension. */ |
| 148 | #endif |
| 149 | |
| 150 | /* XXX A note on the perl source internal type system. The |
| 151 | original intent was that I32 be *exactly* 32 bits. |
| 152 | |
| 153 | Currently, we only guarantee that I32 is *at least* 32 bits. |
| 154 | Specifically, if int is 64 bits, then so is I32. (This is the case |
| 155 | for the Cray.) This has the advantage of meshing nicely with |
| 156 | standard library calls (where we pass an I32 and the library is |
| 157 | expecting an int), but the disadvantage that an I32 is not 32 bits. |
| 158 | Andy Dougherty August 1996 |
| 159 | |
| 160 | There is no guarantee that there is *any* integral type with |
| 161 | exactly 32 bits. It is perfectly legal for a system to have |
| 162 | sizeof(short) == sizeof(int) == sizeof(long) == 8. |
| 163 | |
| 164 | Similarly, there is no guarantee that I16 and U16 have exactly 16 |
| 165 | bits. |
| 166 | |
| 167 | For dealing with issues that may arise from various 32/64-bit |
| 168 | systems, we will ask Configure to check out |
| 169 | |
| 170 | SHORTSIZE == sizeof(short) |
| 171 | INTSIZE == sizeof(int) |
| 172 | LONGSIZE == sizeof(long) |
| 173 | LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG) |
| 174 | PTRSIZE == sizeof(void *) |
| 175 | DOUBLESIZE == sizeof(double) |
| 176 | LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE). |
| 177 | |
| 178 | */ |
| 179 | |
| 180 | #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */ |
| 181 | # include <inttypes.h> |
| 182 | # ifdef INT32_MIN_BROKEN |
| 183 | # undef INT32_MIN |
| 184 | # define INT32_MIN (-2147483647-1) |
| 185 | # endif |
| 186 | # ifdef INT64_MIN_BROKEN |
| 187 | # undef INT64_MIN |
| 188 | # define INT64_MIN (-9223372036854775807LL-1) |
| 189 | # endif |
| 190 | #endif |
| 191 | |
| 192 | typedef I8TYPE I8; |
| 193 | typedef U8TYPE U8; |
| 194 | typedef I16TYPE I16; |
| 195 | typedef U16TYPE U16; |
| 196 | typedef I32TYPE I32; |
| 197 | typedef U32TYPE U32; |
| 198 | |
| 199 | #ifdef QUADKIND |
| 200 | typedef I64TYPE I64; |
| 201 | typedef U64TYPE U64; |
| 202 | #endif |
| 203 | |
| 204 | #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX) |
| 205 | |
| 206 | /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type. |
| 207 | Please search CHAR_MAX in perl.h for further details. */ |
| 208 | #define U8_MAX UINT8_MAX |
| 209 | #define U8_MIN UINT8_MIN |
| 210 | |
| 211 | #define I16_MAX INT16_MAX |
| 212 | #define I16_MIN INT16_MIN |
| 213 | #define U16_MAX UINT16_MAX |
| 214 | #define U16_MIN UINT16_MIN |
| 215 | |
| 216 | #define I32_MAX INT32_MAX |
| 217 | #define I32_MIN INT32_MIN |
| 218 | #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */ |
| 219 | # define U32_MAX UINT32_MAX |
| 220 | #else |
| 221 | # define U32_MAX 4294967295U |
| 222 | #endif |
| 223 | #define U32_MIN UINT32_MIN |
| 224 | |
| 225 | #else |
| 226 | |
| 227 | /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type. |
| 228 | Please search CHAR_MAX in perl.h for further details. */ |
| 229 | #define U8_MAX PERL_UCHAR_MAX |
| 230 | #define U8_MIN PERL_UCHAR_MIN |
| 231 | |
| 232 | #define I16_MAX PERL_SHORT_MAX |
| 233 | #define I16_MIN PERL_SHORT_MIN |
| 234 | #define U16_MAX PERL_USHORT_MAX |
| 235 | #define U16_MIN PERL_USHORT_MIN |
| 236 | |
| 237 | #if LONGSIZE > 4 |
| 238 | # define I32_MAX PERL_INT_MAX |
| 239 | # define I32_MIN PERL_INT_MIN |
| 240 | # define U32_MAX PERL_UINT_MAX |
| 241 | # define U32_MIN PERL_UINT_MIN |
| 242 | #else |
| 243 | # define I32_MAX PERL_LONG_MAX |
| 244 | # define I32_MIN PERL_LONG_MIN |
| 245 | # define U32_MAX PERL_ULONG_MAX |
| 246 | # define U32_MIN PERL_ULONG_MIN |
| 247 | #endif |
| 248 | |
| 249 | #endif |
| 250 | |
| 251 | /* These C99 typedefs are useful sometimes for, say, loop variables whose |
| 252 | * maximum values are small, but for which speed trumps size. If we have a C99 |
| 253 | * compiler, use that. Otherwise, a plain 'int' should be good enough. |
| 254 | * |
| 255 | * Restrict these to core for now until we are more certain this is a good |
| 256 | * idea. */ |
| 257 | #if defined(PERL_CORE) || defined(PERL_EXT) |
| 258 | # ifdef I_STDINT |
| 259 | typedef int_fast8_t PERL_INT_FAST8_T; |
| 260 | typedef uint_fast8_t PERL_UINT_FAST8_T; |
| 261 | typedef int_fast16_t PERL_INT_FAST16_T; |
| 262 | typedef uint_fast16_t PERL_UINT_FAST16_T; |
| 263 | # else |
| 264 | typedef int PERL_INT_FAST8_T; |
| 265 | typedef unsigned int PERL_UINT_FAST8_T; |
| 266 | typedef int PERL_INT_FAST16_T; |
| 267 | typedef unsigned int PERL_UINT_FAST16_T; |
| 268 | # endif |
| 269 | #endif |
| 270 | |
| 271 | /* log(2) (i.e., log base 10 of 2) is pretty close to 0.30103, just in case |
| 272 | * anyone is grepping for it. So BIT_DIGITS gives the number of decimal digits |
| 273 | * required to represent any possible unsigned number containing N bits. |
| 274 | * TYPE_DIGITS gives the number of decimal digits required to represent any |
| 275 | * possible unsigned number of type T. */ |
| 276 | #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log10(2) =~ 146/485 */ |
| 277 | #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8) |
| 278 | #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */ |
| 279 | |
| 280 | /* Unused by core; should be deprecated */ |
| 281 | #define Ctl(ch) ((ch) & 037) |
| 282 | |
| 283 | #if defined(PERL_CORE) || defined(PERL_EXT) |
| 284 | # ifndef MIN |
| 285 | # define MIN(a,b) ((a) < (b) ? (a) : (b)) |
| 286 | # endif |
| 287 | # ifndef MAX |
| 288 | # define MAX(a,b) ((a) > (b) ? (a) : (b)) |
| 289 | # endif |
| 290 | #endif |
| 291 | |
| 292 | /* Returns a boolean as to whether the input unsigned number is a power of 2 |
| 293 | * (2**0, 2**1, etc). In other words if it has just a single bit set. |
| 294 | * If not, subtracting 1 would leave the uppermost bit set, so the & would |
| 295 | * yield non-zero */ |
| 296 | #if defined(PERL_CORE) || defined(PERL_EXT) |
| 297 | # define isPOWER_OF_2(n) ((n) && ((n) & ((n)-1)) == 0) |
| 298 | #endif |
| 299 | |
| 300 | /* Returns a mask with the lowest n bits set */ |
| 301 | #define nBIT_MASK(n) ((UINTMAX_C(1) << (n)) - 1) |
| 302 | |
| 303 | /* The largest unsigned number that will fit into n bits */ |
| 304 | #define nBIT_UMAX(n) nBIT_MASK(n) |
| 305 | |
| 306 | /* |
| 307 | =for apidoc_section Compiler directives |
| 308 | =for apidoc Am|void|__ASSERT_|bool expr |
| 309 | |
| 310 | This is a helper macro to avoid preprocessor issues, replaced by nothing |
| 311 | unless under DEBUGGING, where it expands to an assert of its argument, |
| 312 | followed by a comma (hence the comma operator). If we just used a straight |
| 313 | assert(), we would get a comma with nothing before it when not DEBUGGING. |
| 314 | |
| 315 | =cut |
| 316 | |
| 317 | We also use empty definition under Coverity since the __ASSERT__ |
| 318 | checks often check for things that Really Cannot Happen, and Coverity |
| 319 | detects that and gets all excited. */ |
| 320 | |
| 321 | #if defined(DEBUGGING) && !defined(__COVERITY__) \ |
| 322 | && ! defined(PERL_SMALL_MACRO_BUFFER) |
| 323 | # define __ASSERT_(statement) assert(statement), |
| 324 | #else |
| 325 | # define __ASSERT_(statement) |
| 326 | #endif |
| 327 | |
| 328 | /* |
| 329 | =for apidoc_section SV Handling |
| 330 | |
| 331 | =for apidoc Ama|SV*|newSVpvs|"literal string" |
| 332 | Like C<newSVpvn>, but takes a literal string instead of a |
| 333 | string/length pair. |
| 334 | |
| 335 | =for apidoc Ama|SV*|newSVpvs_flags|"literal string"|U32 flags |
| 336 | Like C<newSVpvn_flags>, but takes a literal string instead of |
| 337 | a string/length pair. |
| 338 | |
| 339 | =for apidoc Ama|SV*|newSVpvs_share|"literal string" |
| 340 | Like C<newSVpvn_share>, but takes a literal string instead of |
| 341 | a string/length pair and omits the hash parameter. |
| 342 | |
| 343 | =for apidoc Am|void|sv_catpvs_flags|SV* sv|"literal string"|I32 flags |
| 344 | Like C<sv_catpvn_flags>, but takes a literal string instead |
| 345 | of a string/length pair. |
| 346 | |
| 347 | =for apidoc Am|void|sv_catpvs_nomg|SV* sv|"literal string" |
| 348 | Like C<sv_catpvn_nomg>, but takes a literal string instead of |
| 349 | a string/length pair. |
| 350 | |
| 351 | =for apidoc Am|void|sv_catpvs|SV* sv|"literal string" |
| 352 | Like C<sv_catpvn>, but takes a literal string instead of a |
| 353 | string/length pair. |
| 354 | |
| 355 | =for apidoc Am|void|sv_catpvs_mg|SV* sv|"literal string" |
| 356 | Like C<sv_catpvn_mg>, but takes a literal string instead of a |
| 357 | string/length pair. |
| 358 | |
| 359 | =for apidoc Am|void|sv_setpvs|SV* sv|"literal string" |
| 360 | Like C<sv_setpvn>, but takes a literal string instead of a |
| 361 | string/length pair. |
| 362 | |
| 363 | =for apidoc Am|void|sv_setpvs_mg|SV* sv|"literal string" |
| 364 | Like C<sv_setpvn_mg>, but takes a literal string instead of a |
| 365 | string/length pair. |
| 366 | |
| 367 | =for apidoc Am|SV *|sv_setref_pvs|SV *const rv|const char *const classname|"literal string" |
| 368 | Like C<sv_setref_pvn>, but takes a literal string instead of |
| 369 | a string/length pair. |
| 370 | |
| 371 | =for apidoc_section String Handling |
| 372 | |
| 373 | =for apidoc Ama|char*|savepvs|"literal string" |
| 374 | Like C<savepvn>, but takes a literal string instead of a |
| 375 | string/length pair. |
| 376 | |
| 377 | =for apidoc Ama|char*|savesharedpvs|"literal string" |
| 378 | A version of C<savepvs()> which allocates the duplicate string in memory |
| 379 | which is shared between threads. |
| 380 | |
| 381 | =for apidoc_section GV Handling |
| 382 | |
| 383 | =for apidoc Am|HV*|gv_stashpvs|"name"|I32 create |
| 384 | Like C<gv_stashpvn>, but takes a literal string instead of a |
| 385 | string/length pair. |
| 386 | |
| 387 | =for apidoc_section HV Handling |
| 388 | |
| 389 | =for apidoc Am|SV**|hv_fetchs|HV* tb|"key"|I32 lval |
| 390 | Like C<hv_fetch>, but takes a literal string instead of a |
| 391 | string/length pair. |
| 392 | |
| 393 | =for apidoc Am|SV**|hv_stores|HV* tb|"key"|SV* val |
| 394 | Like C<hv_store>, but takes a literal string instead of a |
| 395 | string/length pair |
| 396 | and omits the hash parameter. |
| 397 | |
| 398 | =for apidoc_section Lexer interface |
| 399 | |
| 400 | =for apidoc Amx|void|lex_stuff_pvs|"pv"|U32 flags |
| 401 | |
| 402 | Like L</lex_stuff_pvn>, but takes a literal string instead of |
| 403 | a string/length pair. |
| 404 | |
| 405 | =cut |
| 406 | */ |
| 407 | |
| 408 | /* |
| 409 | =for apidoc_section String Handling |
| 410 | |
| 411 | =for apidoc Amu|pair|STR_WITH_LEN|"literal string" |
| 412 | |
| 413 | Returns two comma separated tokens of the input literal string, and its length. |
| 414 | This is convenience macro which helps out in some API calls. |
| 415 | Note that it can't be used as an argument to macros or functions that under |
| 416 | some configurations might be macros, which means that it requires the full |
| 417 | Perl_xxx(aTHX_ ...) form for any API calls where it's used. |
| 418 | |
| 419 | =cut |
| 420 | */ |
| 421 | |
| 422 | #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1) |
| 423 | |
| 424 | /* STR_WITH_LEN() shortcuts */ |
| 425 | #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str)) |
| 426 | #define newSVpvs_flags(str,flags) \ |
| 427 | Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags) |
| 428 | #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0) |
| 429 | #define sv_catpvs_flags(sv, str, flags) \ |
| 430 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags) |
| 431 | #define sv_catpvs_nomg(sv, str) \ |
| 432 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0) |
| 433 | #define sv_catpvs(sv, str) \ |
| 434 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC) |
| 435 | #define sv_catpvs_mg(sv, str) \ |
| 436 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC) |
| 437 | #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str)) |
| 438 | #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str)) |
| 439 | #define sv_setref_pvs(rv, classname, str) \ |
| 440 | Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str)) |
| 441 | #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str)) |
| 442 | #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str)) |
| 443 | #define gv_stashpvs(str, create) \ |
| 444 | Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create) |
| 445 | |
| 446 | #define gv_fetchpvs(namebeg, flags, sv_type) \ |
| 447 | Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), flags, sv_type) |
| 448 | #define gv_fetchpvn(namebeg, len, flags, sv_type) \ |
| 449 | Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, flags, sv_type) |
| 450 | #define sv_catxmlpvs(dsv, str, utf8) \ |
| 451 | Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8) |
| 452 | |
| 453 | |
| 454 | #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags) |
| 455 | |
| 456 | #define get_cvs(str, flags) \ |
| 457 | Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags)) |
| 458 | |
| 459 | /* internal helpers */ |
| 460 | /* Transitional */ |
| 461 | #ifndef PERL_VERSION_MAJOR |
| 462 | # define PERL_VERSION_MAJOR PERL_REVISION |
| 463 | #else |
| 464 | # undef PERL_REVISION /* We don't want code to be using these */ |
| 465 | #endif |
| 466 | #ifndef PERL_VERSION_MINOR |
| 467 | # define PERL_VERSION_MINOR PERL_VERSION |
| 468 | #else |
| 469 | # undef PERL_VERSION |
| 470 | #endif |
| 471 | #ifndef PERL_VERSION_PATCH |
| 472 | # define PERL_VERSION_PATCH PERL_SUBVERSION |
| 473 | #else |
| 474 | # undef PERL_SUBVERSION |
| 475 | #endif |
| 476 | |
| 477 | #define PERL_JNP_TO_DECIMAL_(maJor,miNor,Patch) \ |
| 478 | /* '10*' leaves room for things like alpha, beta, releases */ \ |
| 479 | (10 * ((maJor) * 1000000) + ((miNor) * 1000) + (Patch)) |
| 480 | #define PERL_DECIMAL_VERSION_ \ |
| 481 | PERL_JNP_TO_DECIMAL_(PERL_VERSION_MAJOR, PERL_VERSION_MINOR, \ |
| 482 | PERL_VERSION_PATCH) |
| 483 | |
| 484 | /* |
| 485 | =for apidoc_section Versioning |
| 486 | =for apidoc AmR|bool|PERL_VERSION_EQ|const U8 major|const U8 minor|const U8 patch |
| 487 | |
| 488 | Returns whether or not the perl currently being compiled has the specified |
| 489 | relationship to the perl given by the parameters. For example, |
| 490 | |
| 491 | #if PERL_VERSION_GT(5,24,2) |
| 492 | code that will only be compiled on perls after v5.24.2 |
| 493 | #else |
| 494 | fallback code |
| 495 | #endif |
| 496 | |
| 497 | Note that this is usable in making compile-time decisions |
| 498 | |
| 499 | The possible comparisons are C<PERL_VERSION_EQ>, C<PERL_VERSION_NE>, |
| 500 | C<PERL_VERSION_GE>, C<PERL_VERSION_GT>, C<PERL_VERSION_LE>, and |
| 501 | C<PERL_VERSION_LT>. |
| 502 | |
| 503 | You may use the special value '*' for the final number to mean ALL possible |
| 504 | values for it. Thus, |
| 505 | |
| 506 | #if PERL_VERSION_EQ(5,31,'*') |
| 507 | |
| 508 | means all perls in the 5.31 series. And |
| 509 | |
| 510 | #if PERL_VERSION_NE(5,24,'*') |
| 511 | |
| 512 | means all perls EXCEPT 5.24 ones. And |
| 513 | |
| 514 | #if PERL_VERSION_LE(5,9,'*') |
| 515 | |
| 516 | is effectively |
| 517 | |
| 518 | #if PERL_VERSION_LT(5,10,0) |
| 519 | |
| 520 | This means you don't have to think so much when converting from the existing |
| 521 | deprecated C<PERL_VERSION> to using this macro: |
| 522 | |
| 523 | #if PERL_VERSION <= 9 |
| 524 | |
| 525 | becomes |
| 526 | |
| 527 | #if PERL_VERSION_LE(5,9,'*') |
| 528 | |
| 529 | =for apidoc AmRh|bool|PERL_VERSION_NE|const U8 major|const U8 minor|const U8 patch |
| 530 | =for apidoc AmRh|bool|PERL_VERSION_GE|const U8 major|const U8 minor|const U8 patch |
| 531 | =for apidoc AmRh|bool|PERL_VERSION_GT|const U8 major|const U8 minor|const U8 patch |
| 532 | =for apidoc AmRh|bool|PERL_VERSION_LE|const U8 major|const U8 minor|const U8 patch |
| 533 | =for apidoc AmRh|bool|PERL_VERSION_LT|const U8 major|const U8 minor|const U8 patch |
| 534 | |
| 535 | =cut |
| 536 | */ |
| 537 | |
| 538 | /* N.B. These don't work if the patch version is 42 or 92, as those are what |
| 539 | * '*' is in ASCII and EBCDIC respectively */ |
| 540 | # define PERL_VERSION_EQ(j,n,p) \ |
| 541 | (((p) == '*') \ |
| 542 | ? ( (j) == PERL_VERSION_MAJOR \ |
| 543 | && (n) == PERL_VERSION_MINOR) \ |
| 544 | : (PERL_DECIMAL_VERSION_ == PERL_JNP_TO_DECIMAL_(j,n,p))) |
| 545 | # define PERL_VERSION_NE(j,n,p) (! PERL_VERSION_EQ(j,n,p)) |
| 546 | |
| 547 | # define PERL_VERSION_LT(j,n,p) /* < '*' effectively means < 0 */ \ |
| 548 | (PERL_DECIMAL_VERSION_ < PERL_JNP_TO_DECIMAL_( (j), \ |
| 549 | (n), \ |
| 550 | (((p) == '*') ? 0 : p))) |
| 551 | # define PERL_VERSION_GE(j,n,p) (! PERL_VERSION_LT(j,n,p)) |
| 552 | |
| 553 | # define PERL_VERSION_LE(j,n,p) /* <= '*' effectively means < n+1 */ \ |
| 554 | (PERL_DECIMAL_VERSION_ < PERL_JNP_TO_DECIMAL_( (j), \ |
| 555 | (((p) == '*') ? ((n)+1) : (n)), \ |
| 556 | (((p) == '*') ? 0 : p))) |
| 557 | # define PERL_VERSION_GT(j,n,p) (! PERL_VERSION_LE(j,n,p)) |
| 558 | |
| 559 | /* |
| 560 | =for apidoc_section String Handling |
| 561 | |
| 562 | =for apidoc Am|bool|strNE|char* s1|char* s2 |
| 563 | Test two C<NUL>-terminated strings to see if they are different. Returns true |
| 564 | or false. |
| 565 | |
| 566 | =for apidoc Am|bool|strEQ|char* s1|char* s2 |
| 567 | Test two C<NUL>-terminated strings to see if they are equal. Returns true or |
| 568 | false. |
| 569 | |
| 570 | =for apidoc Am|bool|strLT|char* s1|char* s2 |
| 571 | Test two C<NUL>-terminated strings to see if the first, C<s1>, is less than the |
| 572 | second, C<s2>. Returns true or false. |
| 573 | |
| 574 | =for apidoc Am|bool|strLE|char* s1|char* s2 |
| 575 | Test two C<NUL>-terminated strings to see if the first, C<s1>, is less than or |
| 576 | equal to the second, C<s2>. Returns true or false. |
| 577 | |
| 578 | =for apidoc Am|bool|strGT|char* s1|char* s2 |
| 579 | Test two C<NUL>-terminated strings to see if the first, C<s1>, is greater than |
| 580 | the second, C<s2>. Returns true or false. |
| 581 | |
| 582 | =for apidoc Am|bool|strGE|char* s1|char* s2 |
| 583 | Test two C<NUL>-terminated strings to see if the first, C<s1>, is greater than |
| 584 | or equal to the second, C<s2>. Returns true or false. |
| 585 | |
| 586 | =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len |
| 587 | Test two C<NUL>-terminated strings to see if they are different. The C<len> |
| 588 | parameter indicates the number of bytes to compare. Returns true or false. (A |
| 589 | wrapper for C<strncmp>). |
| 590 | |
| 591 | =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len |
| 592 | Test two C<NUL>-terminated strings to see if they are equal. The C<len> |
| 593 | parameter indicates the number of bytes to compare. Returns true or false. (A |
| 594 | wrapper for C<strncmp>). |
| 595 | |
| 596 | =for apidoc Am|bool|memEQ|char* s1|char* s2|STRLEN len |
| 597 | Test two buffers (which may contain embedded C<NUL> characters, to see if they |
| 598 | are equal. The C<len> parameter indicates the number of bytes to compare. |
| 599 | Returns zero if equal, or non-zero if non-equal. |
| 600 | |
| 601 | =for apidoc Am|bool|memEQs|char* s1|STRLEN l1|"s2" |
| 602 | Like L</memEQ>, but the second string is a literal enclosed in double quotes, |
| 603 | C<l1> gives the number of bytes in C<s1>. |
| 604 | Returns zero if equal, or non-zero if non-equal. |
| 605 | |
| 606 | =for apidoc Am|bool|memNE|char* s1|char* s2|STRLEN len |
| 607 | Test two buffers (which may contain embedded C<NUL> characters, to see if they |
| 608 | are not equal. The C<len> parameter indicates the number of bytes to compare. |
| 609 | Returns zero if non-equal, or non-zero if equal. |
| 610 | |
| 611 | =for apidoc Am|bool|memNEs|char* s1|STRLEN l1|"s2" |
| 612 | Like L</memNE>, but the second string is a literal enclosed in double quotes, |
| 613 | C<l1> gives the number of bytes in C<s1>. |
| 614 | Returns zero if non-equal, or zero if non-equal. |
| 615 | |
| 616 | =for apidoc Am|bool|memCHRs|"list"|char c |
| 617 | Returns the position of the first occurence of the byte C<c> in the literal |
| 618 | string C<"list">, or NULL if C<c> doesn't appear in C<"list">. All bytes are |
| 619 | treated as unsigned char. Thus this macro can be used to determine if C<c> is |
| 620 | in a set of particular characters. Unlike L<strchr(3)>, it works even if C<c> |
| 621 | is C<NUL> (and the set doesn't include C<NUL>). |
| 622 | |
| 623 | =cut |
| 624 | |
| 625 | New macros should use the following conventions for their names (which are |
| 626 | based on the underlying C library functions): |
| 627 | |
| 628 | (mem | str n? ) (EQ | NE | LT | GT | GE | (( BEGIN | END ) P? )) l? s? |
| 629 | |
| 630 | Each has two main parameters, string-like operands that are compared |
| 631 | against each other, as specified by the macro name. Some macros may |
| 632 | additionally have one or potentially even two length parameters. If a length |
| 633 | parameter applies to both string parameters, it will be positioned third; |
| 634 | otherwise any length parameter immediately follows the string parameter it |
| 635 | applies to. |
| 636 | |
| 637 | If the prefix to the name is 'str', the string parameter is a pointer to a C |
| 638 | language string. Such a string does not contain embedded NUL bytes; its |
| 639 | length may be unknown, but can be calculated by C<strlen()>, since it is |
| 640 | terminated by a NUL, which isn't included in its length. |
| 641 | |
| 642 | The optional 'n' following 'str' means that there is a third parameter, |
| 643 | giving the maximum number of bytes to look at in each string. Even if both |
| 644 | strings are longer than the length parameter, those extra bytes will be |
| 645 | unexamined. |
| 646 | |
| 647 | The 's' suffix means that the 2nd byte string parameter is a literal C |
| 648 | double-quoted string. Its length will automatically be calculated by the |
| 649 | macro, so no length parameter will ever be needed for it. |
| 650 | |
| 651 | If the prefix is 'mem', the string parameters don't have to be C strings; |
| 652 | they may contain embedded NUL bytes, do not necessarily have a terminating |
| 653 | NUL, and their lengths can be known only through other means, which in |
| 654 | practice are additional parameter(s) passed to the function. All 'mem' |
| 655 | functions have at least one length parameter. Barring any 'l' or 's' suffix, |
| 656 | there is a single length parameter, in position 3, which applies to both |
| 657 | string parameters. The 's' suffix means, as described above, that the 2nd |
| 658 | string is a literal double-quoted C string (hence its length is calculated by |
| 659 | the macro, and the length parameter to the function applies just to the first |
| 660 | string parameter, and hence is positioned just after it). An 'l' suffix |
| 661 | means that the 2nd string parameter has its own length parameter, and the |
| 662 | signature will look like memFOOl(s1, l1, s2, l2). |
| 663 | |
| 664 | BEGIN (and END) are for testing if the 2nd string is an initial (or final) |
| 665 | substring of the 1st string. 'P' if present indicates that the substring |
| 666 | must be a "proper" one in tha mathematical sense that the first one must be |
| 667 | strictly larger than the 2nd. |
| 668 | |
| 669 | */ |
| 670 | |
| 671 | |
| 672 | #define strNE(s1,s2) (strcmp(s1,s2) != 0) |
| 673 | #define strEQ(s1,s2) (strcmp(s1,s2) == 0) |
| 674 | #define strLT(s1,s2) (strcmp(s1,s2) < 0) |
| 675 | #define strLE(s1,s2) (strcmp(s1,s2) <= 0) |
| 676 | #define strGT(s1,s2) (strcmp(s1,s2) > 0) |
| 677 | #define strGE(s1,s2) (strcmp(s1,s2) >= 0) |
| 678 | |
| 679 | #define strnNE(s1,s2,l) (strncmp(s1,s2,l) != 0) |
| 680 | #define strnEQ(s1,s2,l) (strncmp(s1,s2,l) == 0) |
| 681 | |
| 682 | #define memEQ(s1,s2,l) (memcmp(((const void *) (s1)), ((const void *) (s2)), l) == 0) |
| 683 | #define memNE(s1,s2,l) (! memEQ(s1,s2,l)) |
| 684 | |
| 685 | /* memEQ and memNE where second comparand is a string constant */ |
| 686 | #define memEQs(s1, l, s2) \ |
| 687 | (((sizeof(s2)-1) == (l)) && memEQ((s1), ("" s2 ""), (sizeof(s2)-1))) |
| 688 | #define memNEs(s1, l, s2) (! memEQs(s1, l, s2)) |
| 689 | |
| 690 | /* Keep these private until we decide it was a good idea */ |
| 691 | #if defined(PERL_CORE) || defined(PERL_EXT) || defined(PERL_EXT_POSIX) |
| 692 | |
| 693 | #define strBEGINs(s1,s2) (strncmp(s1,"" s2 "", sizeof(s2)-1) == 0) |
| 694 | |
| 695 | #define memBEGINs(s1, l, s2) \ |
| 696 | ( (Ptrdiff_t) (l) >= (Ptrdiff_t) sizeof(s2) - 1 \ |
| 697 | && memEQ(s1, "" s2 "", sizeof(s2)-1)) |
| 698 | #define memBEGINPs(s1, l, s2) \ |
| 699 | ( (Ptrdiff_t) (l) > (Ptrdiff_t) sizeof(s2) - 1 \ |
| 700 | && memEQ(s1, "" s2 "", sizeof(s2)-1)) |
| 701 | #define memENDs(s1, l, s2) \ |
| 702 | ( (Ptrdiff_t) (l) >= (Ptrdiff_t) sizeof(s2) - 1 \ |
| 703 | && memEQ(s1 + (l) - (sizeof(s2) - 1), "" s2 "", sizeof(s2)-1)) |
| 704 | #define memENDPs(s1, l, s2) \ |
| 705 | ( (Ptrdiff_t) (l) > (Ptrdiff_t) sizeof(s2) \ |
| 706 | && memEQ(s1 + (l) - (sizeof(s2) - 1), "" s2 "", sizeof(s2)-1)) |
| 707 | #endif /* End of making macros private */ |
| 708 | |
| 709 | #define memLT(s1,s2,l) (memcmp(s1,s2,l) < 0) |
| 710 | #define memLE(s1,s2,l) (memcmp(s1,s2,l) <= 0) |
| 711 | #define memGT(s1,s2,l) (memcmp(s1,s2,l) > 0) |
| 712 | #define memGE(s1,s2,l) (memcmp(s1,s2,l) >= 0) |
| 713 | |
| 714 | #define memCHRs(s1,c) ((const char *) memchr("" s1 "" , c, sizeof(s1)-1)) |
| 715 | |
| 716 | /* |
| 717 | * Character classes. |
| 718 | * |
| 719 | * Unfortunately, the introduction of locales means that we |
| 720 | * can't trust isupper(), etc. to tell the truth. And when |
| 721 | * it comes to /\w+/ with tainting enabled, we *must* be able |
| 722 | * to trust our character classes. |
| 723 | * |
| 724 | * Therefore, the default tests in the text of Perl will be |
| 725 | * independent of locale. Any code that wants to depend on |
| 726 | * the current locale will use the tests that begin with "lc". |
| 727 | */ |
| 728 | |
| 729 | #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */ |
| 730 | # ifndef CTYPE256 |
| 731 | # define CTYPE256 |
| 732 | # endif |
| 733 | #endif |
| 734 | |
| 735 | /* |
| 736 | |
| 737 | =head1 Character classification |
| 738 | This section is about functions (really macros) that classify characters |
| 739 | into types, such as punctuation versus alphabetic, etc. Most of these are |
| 740 | analogous to regular expression character classes. (See |
| 741 | L<perlrecharclass/POSIX Character Classes>.) There are several variants for |
| 742 | each class. (Not all macros have all variants; each item below lists the |
| 743 | ones valid for it.) None are affected by C<use bytes>, and only the ones |
| 744 | with C<LC> in the name are affected by the current locale. |
| 745 | |
| 746 | The base function, e.g., C<isALPHA()>, takes any signed or unsigned value, |
| 747 | treating it as a code point, and returns a boolean as to whether or not the |
| 748 | character represented by it is (or on non-ASCII platforms, corresponds to) an |
| 749 | ASCII character in the named class based on platform, Unicode, and Perl rules. |
| 750 | If the input is a number that doesn't fit in an octet, FALSE is returned. |
| 751 | |
| 752 | Variant C<isI<FOO>_A> (e.g., C<isALPHA_A()>) is identical to the base function |
| 753 | with no suffix C<"_A">. This variant is used to emphasize by its name that |
| 754 | only ASCII-range characters can return TRUE. |
| 755 | |
| 756 | Variant C<isI<FOO>_L1> imposes the Latin-1 (or EBCDIC equivalent) character set |
| 757 | onto the platform. That is, the code points that are ASCII are unaffected, |
| 758 | since ASCII is a subset of Latin-1. But the non-ASCII code points are treated |
| 759 | as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return |
| 760 | true when called with the code point 0xDF, which is a word character in both |
| 761 | ASCII and EBCDIC (though it represents different characters in each). |
| 762 | If the input is a number that doesn't fit in an octet, FALSE is returned. |
| 763 | (Perl's documentation uses a colloquial definition of Latin-1, to include all |
| 764 | code points below 256.) |
| 765 | |
| 766 | Variant C<isI<FOO>_uvchr> is exactly like the C<isI<FOO>_L1> variant, for |
| 767 | inputs below 256, but if the code point is larger than 255, Unicode rules are |
| 768 | used to determine if it is in the character class. For example, |
| 769 | C<isWORDCHAR_uvchr(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A |
| 770 | WITH MACRON in Unicode, and is a word character. |
| 771 | |
| 772 | Variants C<isI<FOO>_utf8> and C<isI<FOO>_utf8_safe> are like C<isI<FOO>_uvchr>, |
| 773 | but are used for UTF-8 encoded strings. The two forms are different names for |
| 774 | the same thing. Each call to one of these classifies the first character of |
| 775 | the string starting at C<p>. The second parameter, C<e>, points to anywhere in |
| 776 | the string beyond the first character, up to one byte past the end of the |
| 777 | entire string. Although both variants are identical, the suffix C<_safe> in |
| 778 | one name emphasizes that it will not attempt to read beyond S<C<e - 1>>, |
| 779 | provided that the constraint S<C<s E<lt> e>> is true (this is asserted for in |
| 780 | C<-DDEBUGGING> builds). If the UTF-8 for the input character is malformed in |
| 781 | some way, the program may croak, or the function may return FALSE, at the |
| 782 | discretion of the implementation, and subject to change in future releases. |
| 783 | |
| 784 | Variant C<isI<FOO>_LC> is like the C<isI<FOO>_A> and C<isI<FOO>_L1> variants, |
| 785 | but the result is based on the current locale, which is what C<LC> in the name |
| 786 | stands for. If Perl can determine that the current locale is a UTF-8 locale, |
| 787 | it uses the published Unicode rules; otherwise, it uses the C library function |
| 788 | that gives the named classification. For example, C<isDIGIT_LC()> when not in |
| 789 | a UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always |
| 790 | returned if the input won't fit into an octet. On some platforms where the C |
| 791 | library function is known to be defective, Perl changes its result to follow |
| 792 | the POSIX standard's rules. |
| 793 | |
| 794 | Variant C<isI<FOO>_LC_uvchr> acts exactly like C<isI<FOO>_LC> for inputs less |
| 795 | than 256, but for larger ones it returns the Unicode classification of the code |
| 796 | point. |
| 797 | |
| 798 | Variants C<isI<FOO>_LC_utf8> and C<isI<FOO>_LC_utf8_safe> are like |
| 799 | C<isI<FOO>_LC_uvchr>, but are used for UTF-8 encoded strings. The two forms |
| 800 | are different names for the same thing. Each call to one of these classifies |
| 801 | the first character of the string starting at C<p>. The second parameter, |
| 802 | C<e>, points to anywhere in the string beyond the first character, up to one |
| 803 | byte past the end of the entire string. Although both variants are identical, |
| 804 | the suffix C<_safe> in one name emphasizes that it will not attempt to read |
| 805 | beyond S<C<e - 1>>, provided that the constraint S<C<s E<lt> e>> is true (this |
| 806 | is asserted for in C<-DDEBUGGING> builds). If the UTF-8 for the input |
| 807 | character is malformed in some way, the program may croak, or the function may |
| 808 | return FALSE, at the discretion of the implementation, and subject to change in |
| 809 | future releases. |
| 810 | |
| 811 | =for apidoc Am|bool|isALPHA|int ch |
| 812 | =for apidoc_item ||isALPHA_A|int ch |
| 813 | =for apidoc_item ||isALPHA_L1|int ch |
| 814 | =for apidoc_item ||isALPHA_uvchr|int ch |
| 815 | =for apidoc_item ||isALPHA_utf8_safe|U8 * s|U8 * end |
| 816 | =for apidoc_item ||isALPHA_utf8|U8 * s|U8 * end |
| 817 | =for apidoc_item ||isALPHA_LC|int ch |
| 818 | =for apidoc_item ||isALPHA_LC_uvchr|int ch |
| 819 | =for apidoc_item ||isALPHA_LC_utf8_safe|U8 * s| U8 *end |
| 820 | Returns a boolean indicating whether the specified input is one of C<[A-Za-z]>, |
| 821 | analogous to C<m/[[:alpha:]]/>. |
| 822 | See the L<top of this section|/Character classification> for an explanation of |
| 823 | the variants. |
| 824 | |
| 825 | =cut |
| 826 | |
| 827 | Here and below, we add the protoypes of these macros for downstream programs |
| 828 | that would be interested in them, such as Devel::PPPort |
| 829 | |
| 830 | =for apidoc Am|bool|isALPHANUMERIC|int ch |
| 831 | =for apidoc_item ||isALPHANUMERIC_A|int ch |
| 832 | =for apidoc_item ||isALPHANUMERIC_L1|int ch |
| 833 | =for apidoc_item ||isALPHANUMERIC_uvchr|int ch |
| 834 | =for apidoc_item ||isALPHANUMERIC_utf8_safe|U8 * s|U8 * end |
| 835 | =for apidoc_item ||isALPHANUMERIC_utf8|U8 * s|U8 * end |
| 836 | =for apidoc_item ||isALPHANUMERIC_LC|int ch |
| 837 | =for apidoc_item ||isALPHANUMERIC_LC_uvchr|int ch |
| 838 | =for apidoc_item ||isALPHANUMERIC_LC_utf8_safe|U8 * s| U8 *end |
| 839 | =for apidoc_item ||isALNUMC|int ch |
| 840 | =for apidoc_item ||isALNUMC_A|int ch |
| 841 | =for apidoc_item ||isALNUMC_L1|int ch |
| 842 | =for apidoc_item ||isALNUMC_LC|int ch |
| 843 | =for apidoc_item ||isALNUMC_LC_uvchr|int ch |
| 844 | Returns a boolean indicating whether the specified character is one of |
| 845 | C<[A-Za-z0-9]>, analogous to C<m/[[:alnum:]]/>. |
| 846 | See the L<top of this section|/Character classification> for an explanation of |
| 847 | the variants. |
| 848 | |
| 849 | A (discouraged from use) synonym is C<isALNUMC> (where the C<C> suffix means |
| 850 | this corresponds to the C language alphanumeric definition). Also |
| 851 | there are the variants |
| 852 | C<isALNUMC_A>, C<isALNUMC_L1> |
| 853 | C<isALNUMC_LC>, and C<isALNUMC_LC_uvchr>. |
| 854 | |
| 855 | =for apidoc Am|bool|isASCII|int ch |
| 856 | =for apidoc_item ||isASCII_A|int ch |
| 857 | =for apidoc_item ||isASCII_L1|int ch |
| 858 | =for apidoc_item ||isASCII_uvchr|int ch |
| 859 | =for apidoc_item ||isASCII_utf8_safe|U8 * s|U8 * end |
| 860 | =for apidoc_item ||isASCII_utf8|U8 * s|U8 * end |
| 861 | =for apidoc_item ||isASCII_LC|int ch |
| 862 | =for apidoc_item ||isASCII_LC_uvchr|int ch |
| 863 | =for apidoc_item ||isASCII_LC_utf8_safe|U8 * s| U8 *end |
| 864 | Returns a boolean indicating whether the specified character is one of the 128 |
| 865 | characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>. |
| 866 | On non-ASCII platforms, it returns TRUE iff this |
| 867 | character corresponds to an ASCII character. Variants C<isASCII_A()> and |
| 868 | C<isASCII_L1()> are identical to C<isASCII()>. |
| 869 | See the L<top of this section|/Character classification> for an explanation of |
| 870 | the variants. |
| 871 | Note, however, that some platforms do not have the C library routine |
| 872 | C<isascii()>. In these cases, the variants whose names contain C<LC> are the |
| 873 | same as the corresponding ones without. |
| 874 | |
| 875 | Also note, that because all ASCII characters are UTF-8 invariant (meaning they |
| 876 | have the exact same representation (always a single byte) whether encoded in |
| 877 | UTF-8 or not), C<isASCII> will give the correct results when called with any |
| 878 | byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> and |
| 879 | C<isASCII_utf8_safe> will work properly on any string encoded or not in UTF-8. |
| 880 | |
| 881 | =for apidoc Am|bool|isBLANK|int ch |
| 882 | =for apidoc_item ||isBLANK_A|int ch |
| 883 | =for apidoc_item ||isBLANK_L1|int ch |
| 884 | =for apidoc_item ||isBLANK_uvchr|int ch |
| 885 | =for apidoc_item ||isBLANK_utf8_safe|U8 * s|U8 * end |
| 886 | =for apidoc_item ||isBLANK_utf8|U8 * s|U8 * end |
| 887 | =for apidoc_item ||isBLANK_LC|int ch |
| 888 | =for apidoc_item ||isBLANK_LC_uvchr|int ch |
| 889 | =for apidoc_item ||isBLANK_LC_utf8_safe|U8 * s| U8 *end |
| 890 | Returns a boolean indicating whether the specified character is a |
| 891 | character considered to be a blank, analogous to C<m/[[:blank:]]/>. |
| 892 | See the L<top of this section|/Character classification> for an explanation of |
| 893 | the variants. |
| 894 | Note, |
| 895 | however, that some platforms do not have the C library routine |
| 896 | C<isblank()>. In these cases, the variants whose names contain C<LC> are |
| 897 | the same as the corresponding ones without. |
| 898 | |
| 899 | =for apidoc Am|bool|isCNTRL|int ch |
| 900 | =for apidoc_item ||isCNTRL_A|int ch |
| 901 | =for apidoc_item ||isCNTRL_L1|int ch |
| 902 | =for apidoc_item ||isCNTRL_uvchr|int ch |
| 903 | =for apidoc_item ||isCNTRL_utf8_safe|U8 * s|U8 * end |
| 904 | =for apidoc_item ||isCNTRL_utf8|U8 * s|U8 * end |
| 905 | =for apidoc_item ||isCNTRL_LC|int ch |
| 906 | =for apidoc_item ||isCNTRL_LC_uvchr|int ch |
| 907 | =for apidoc_item ||isCNTRL_LC_utf8_safe|U8 * s| U8 *end |
| 908 | |
| 909 | Returns a boolean indicating whether the specified character is a |
| 910 | control character, analogous to C<m/[[:cntrl:]]/>. |
| 911 | See the L<top of this section|/Character classification> for an explanation of |
| 912 | the variants. |
| 913 | On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant. |
| 914 | |
| 915 | =for apidoc Am|bool|isDIGIT|int ch |
| 916 | =for apidoc_item ||isDIGIT_A|int ch |
| 917 | =for apidoc_item ||isDIGIT_L1|int ch |
| 918 | =for apidoc_item ||isDIGIT_uvchr|int ch |
| 919 | =for apidoc_item ||isDIGIT_utf8_safe|U8 * s|U8 * end |
| 920 | =for apidoc_item ||isDIGIT_utf8|U8 * s|U8 * end |
| 921 | =for apidoc_item ||isDIGIT_LC|int ch |
| 922 | =for apidoc_item ||isDIGIT_LC_uvchr|int ch |
| 923 | =for apidoc_item ||isDIGIT_LC_utf8_safe|U8 * s| U8 *end |
| 924 | |
| 925 | Returns a boolean indicating whether the specified character is a |
| 926 | digit, analogous to C<m/[[:digit:]]/>. |
| 927 | Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>. |
| 928 | See the L<top of this section|/Character classification> for an explanation of |
| 929 | the variants. |
| 930 | |
| 931 | =for apidoc Am|bool|isGRAPH|int ch |
| 932 | =for apidoc_item ||isGRAPH_A|int ch |
| 933 | =for apidoc_item ||isGRAPH_L1|int ch |
| 934 | =for apidoc_item ||isGRAPH_uvchr|int ch |
| 935 | =for apidoc_item ||isGRAPH_utf8_safe|U8 * s|U8 * end |
| 936 | =for apidoc_item ||isGRAPH_utf8|U8 * s|U8 * end |
| 937 | =for apidoc_item ||isGRAPH_LC|int ch |
| 938 | =for apidoc_item ||isGRAPH_LC_uvchr|int ch |
| 939 | =for apidoc_item ||isGRAPH_LC_utf8_safe|U8 * s| U8 *end |
| 940 | Returns a boolean indicating whether the specified character is a |
| 941 | graphic character, analogous to C<m/[[:graph:]]/>. |
| 942 | See the L<top of this section|/Character classification> for an explanation of |
| 943 | the variants. |
| 944 | |
| 945 | =for apidoc Am|bool|isLOWER|int ch |
| 946 | =for apidoc_item ||isLOWER_A|int ch |
| 947 | =for apidoc_item ||isLOWER_L1|int ch |
| 948 | =for apidoc_item ||isLOWER_uvchr|int ch |
| 949 | =for apidoc_item ||isLOWER_utf8_safe|U8 * s|U8 * end |
| 950 | =for apidoc_item ||isLOWER_utf8|U8 * s|U8 * end |
| 951 | =for apidoc_item ||isLOWER_LC|int ch |
| 952 | =for apidoc_item ||isLOWER_LC_uvchr|int ch |
| 953 | =for apidoc_item ||isLOWER_LC_utf8_safe|U8 * s| U8 *end |
| 954 | Returns a boolean indicating whether the specified character is a |
| 955 | lowercase character, analogous to C<m/[[:lower:]]/>. |
| 956 | See the L<top of this section|/Character classification> for an explanation of |
| 957 | the variants |
| 958 | |
| 959 | =for apidoc Am|bool|isOCTAL|int ch |
| 960 | =for apidoc_item ||isOCTAL_A|int ch |
| 961 | =for apidoc_item ||isOCTAL_L1|int ch |
| 962 | Returns a boolean indicating whether the specified character is an |
| 963 | octal digit, [0-7]. |
| 964 | The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to |
| 965 | C<isOCTAL>. |
| 966 | |
| 967 | =for apidoc Am|bool|isPUNCT|int ch |
| 968 | =for apidoc_item ||isPUNCT_A|int ch |
| 969 | =for apidoc_item ||isPUNCT_L1|int ch |
| 970 | =for apidoc_item ||isPUNCT_uvchr|int ch |
| 971 | =for apidoc_item ||isPUNCT_utf8_safe|U8 * s|U8 * end |
| 972 | =for apidoc_item ||isPUNCT_utf8|U8 * s|U8 * end |
| 973 | =for apidoc_item ||isPUNCT_LC|int ch |
| 974 | =for apidoc_item ||isPUNCT_LC_uvchr|int ch |
| 975 | =for apidoc_item ||isPUNCT_LC_utf8_safe|U8 * s| U8 *end |
| 976 | Returns a boolean indicating whether the specified character is a |
| 977 | punctuation character, analogous to C<m/[[:punct:]]/>. |
| 978 | Note that the definition of what is punctuation isn't as |
| 979 | straightforward as one might desire. See L<perlrecharclass/POSIX Character |
| 980 | Classes> for details. |
| 981 | See the L<top of this section|/Character classification> for an explanation of |
| 982 | the variants. |
| 983 | |
| 984 | =for apidoc Am|bool|isSPACE|int ch |
| 985 | =for apidoc_item ||isSPACE_A|int ch |
| 986 | =for apidoc_item ||isSPACE_L1|int ch |
| 987 | =for apidoc_item ||isSPACE_uvchr|int ch |
| 988 | =for apidoc_item ||isSPACE_utf8_safe|U8 * s|U8 * end |
| 989 | =for apidoc_item ||isSPACE_utf8|U8 * s|U8 * end |
| 990 | =for apidoc_item ||isSPACE_LC|int ch |
| 991 | =for apidoc_item ||isSPACE_LC_uvchr|int ch |
| 992 | =for apidoc_item ||isSPACE_LC_utf8_safe|U8 * s| U8 *end |
| 993 | Returns a boolean indicating whether the specified character is a |
| 994 | whitespace character. This is analogous |
| 995 | to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18 |
| 996 | this also matches what C<m/[[:space:]]/> does. Prior to 5.18, only the |
| 997 | locale forms of this macro (the ones with C<LC> in their names) matched |
| 998 | precisely what C<m/[[:space:]]/> does. In those releases, the only difference, |
| 999 | in the non-locale variants, was that C<isSPACE()> did not match a vertical tab. |
| 1000 | (See L</isPSXSPC> for a macro that matches a vertical tab in all releases.) |
| 1001 | See the L<top of this section|/Character classification> for an explanation of |
| 1002 | the variants. |
| 1003 | |
| 1004 | =for apidoc Am|bool|isPSXSPC|int ch |
| 1005 | =for apidoc_item ||isPSXSPC_A|int ch |
| 1006 | =for apidoc_item ||isPSXSPC_L1|int ch |
| 1007 | =for apidoc_item ||isPSXSPC_uvchr|int ch |
| 1008 | =for apidoc_item ||isPSXSPC_utf8_safe|U8 * s|U8 * end |
| 1009 | =for apidoc_item ||isPSXSPC_utf8|U8 * s|U8 * end |
| 1010 | =for apidoc_item ||isPSXSPC_LC|int ch |
| 1011 | =for apidoc_item ||isPSXSPC_LC_uvchr|int ch |
| 1012 | =for apidoc_item ||isPSXSPC_LC_utf8_safe|U8 * s| U8 *end |
| 1013 | (short for Posix Space) |
| 1014 | Starting in 5.18, this is identical in all its forms to the |
| 1015 | corresponding C<isSPACE()> macros. |
| 1016 | The locale forms of this macro are identical to their corresponding |
| 1017 | C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the |
| 1018 | non-locale forms differ from their C<isSPACE()> forms only in that the |
| 1019 | C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do. |
| 1020 | Otherwise they are identical. Thus this macro is analogous to what |
| 1021 | C<m/[[:space:]]/> matches in a regular expression. |
| 1022 | See the L<top of this section|/Character classification> for an explanation of |
| 1023 | the variants. |
| 1024 | |
| 1025 | =for apidoc Am|bool|isUPPER|int ch |
| 1026 | =for apidoc_item ||isUPPER_A|int ch |
| 1027 | =for apidoc_item ||isUPPER_L1|int ch |
| 1028 | =for apidoc_item ||isUPPER_uvchr|int ch |
| 1029 | =for apidoc_item ||isUPPER_utf8_safe|U8 * s|U8 * end |
| 1030 | =for apidoc_item ||isUPPER_utf8|U8 * s|U8 * end |
| 1031 | =for apidoc_item ||isUPPER_LC|int ch |
| 1032 | =for apidoc_item ||isUPPER_LC_uvchr|int ch |
| 1033 | =for apidoc_item ||isUPPER_LC_utf8_safe|U8 * s| U8 *end |
| 1034 | Returns a boolean indicating whether the specified character is an |
| 1035 | uppercase character, analogous to C<m/[[:upper:]]/>. |
| 1036 | See the L<top of this section|/Character classification> for an explanation of |
| 1037 | the variants. |
| 1038 | |
| 1039 | =for apidoc Am|bool|isPRINT|int ch |
| 1040 | =for apidoc_item ||isPRINT_A|int ch |
| 1041 | =for apidoc_item ||isPRINT_L1|int ch |
| 1042 | =for apidoc_item ||isPRINT_uvchr|int ch |
| 1043 | =for apidoc_item ||isPRINT_utf8_safe|U8 * s|U8 * end |
| 1044 | =for apidoc_item ||isPRINT_utf8|U8 * s|U8 * end |
| 1045 | =for apidoc_item ||isPRINT_LC|int ch |
| 1046 | =for apidoc_item ||isPRINT_LC_uvchr|int ch |
| 1047 | =for apidoc_item ||isPRINT_LC_utf8_safe|U8 * s| U8 *end |
| 1048 | Returns a boolean indicating whether the specified character is a |
| 1049 | printable character, analogous to C<m/[[:print:]]/>. |
| 1050 | See the L<top of this section|/Character classification> for an explanation of |
| 1051 | the variants. |
| 1052 | |
| 1053 | =for apidoc Am|bool|isWORDCHAR|int ch |
| 1054 | =for apidoc_item ||isWORDCHAR_A|int ch |
| 1055 | =for apidoc_item ||isWORDCHAR_L1|int ch |
| 1056 | =for apidoc_item ||isWORDCHAR_uvchr|int ch |
| 1057 | =for apidoc_item ||isWORDCHAR_utf8_safe|U8 * s|U8 * end |
| 1058 | =for apidoc_item ||isWORDCHAR_utf8|U8 * s|U8 * end |
| 1059 | =for apidoc_item ||isWORDCHAR_LC|int ch |
| 1060 | =for apidoc_item ||isWORDCHAR_LC_uvchr|int ch |
| 1061 | =for apidoc_item ||isWORDCHAR_LC_utf8_safe|U8 * s| U8 *end |
| 1062 | =for apidoc_item ||isALNUM|int ch |
| 1063 | =for apidoc_item ||isALNUM_A|int ch |
| 1064 | =for apidoc_item ||isALNUM_LC|int ch |
| 1065 | =for apidoc_item ||isALNUM_LC_uvchr|int ch |
| 1066 | Returns a boolean indicating whether the specified character is a character |
| 1067 | that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match |
| 1068 | in a regular expression. A word character is an alphabetic character, a |
| 1069 | decimal digit, a connecting punctuation character (such as an underscore), or |
| 1070 | a "mark" character that attaches to one of those (like some sort of accent). |
| 1071 | C<isALNUM()> is a synonym provided for backward compatibility, even though a |
| 1072 | word character includes more than the standard C language meaning of |
| 1073 | alphanumeric. |
| 1074 | See the L<top of this section|/Character classification> for an explanation of |
| 1075 | the variants. |
| 1076 | C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uvchr>, |
| 1077 | C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, C<isWORDCHAR_LC_utf8>, and |
| 1078 | C<isWORDCHAR_LC_utf8_safe> are also as described there, but additionally |
| 1079 | include the platform's native underscore. |
| 1080 | |
| 1081 | =for apidoc Am|bool|isXDIGIT|int ch |
| 1082 | =for apidoc_item ||isXDIGIT_A|int ch |
| 1083 | =for apidoc_item ||isXDIGIT_L1|int ch |
| 1084 | =for apidoc_item ||isXDIGIT_uvchr|int ch |
| 1085 | =for apidoc_item ||isXDIGIT_utf8_safe|U8 * s|U8 * end |
| 1086 | =for apidoc_item ||isXDIGIT_utf8|U8 * s|U8 * end |
| 1087 | =for apidoc_item ||isXDIGIT_LC|int ch |
| 1088 | =for apidoc_item ||isXDIGIT_LC_uvchr|int ch |
| 1089 | =for apidoc_item ||isXDIGIT_LC_utf8_safe|U8 * s| U8 *end |
| 1090 | Returns a boolean indicating whether the specified character is a hexadecimal |
| 1091 | digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()> |
| 1092 | and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>. |
| 1093 | See the L<top of this section|/Character classification> for an explanation of |
| 1094 | the variants. |
| 1095 | |
| 1096 | =for apidoc Am|bool|isIDFIRST|int ch |
| 1097 | =for apidoc_item ||isIDFIRST_A|int ch |
| 1098 | =for apidoc_item ||isIDFIRST_L1|int ch |
| 1099 | =for apidoc_item ||isIDFIRST_uvchr|int ch |
| 1100 | =for apidoc_item ||isIDFIRST_utf8_safe|U8 * s|U8 * end |
| 1101 | =for apidoc_item ||isIDFIRST_utf8|U8 * s|U8 * end |
| 1102 | =for apidoc_item ||isIDFIRST_LC|int ch |
| 1103 | =for apidoc_item ||isIDFIRST_LC_uvchr|int ch |
| 1104 | =for apidoc_item ||isIDFIRST_LC_utf8_safe|U8 * s| U8 *end |
| 1105 | Returns a boolean indicating whether the specified character can be the first |
| 1106 | character of an identifier. This is very close to, but not quite the same as |
| 1107 | the official Unicode property C<XID_Start>. The difference is that this |
| 1108 | returns true only if the input character also matches L</isWORDCHAR>. |
| 1109 | See the L<top of this section|/Character classification> for an explanation of |
| 1110 | the variants. |
| 1111 | |
| 1112 | =for apidoc Am|bool|isIDCONT|int ch |
| 1113 | =for apidoc_item ||isIDCONT_A|int ch |
| 1114 | =for apidoc_item ||isIDCONT_L1|int ch |
| 1115 | =for apidoc_item ||isIDCONT_uvchr|int ch |
| 1116 | =for apidoc_item ||isIDCONT_utf8_safe|U8 * s|U8 * end |
| 1117 | =for apidoc_item ||isIDCONT_utf8|U8 * s|U8 * end |
| 1118 | =for apidoc_item ||isIDCONT_LC|int ch |
| 1119 | =for apidoc_item ||isIDCONT_LC_uvchr|int ch |
| 1120 | =for apidoc_item ||isIDCONT_LC_utf8_safe|U8 * s| U8 *end |
| 1121 | Returns a boolean indicating whether the specified character can be the |
| 1122 | second or succeeding character of an identifier. This is very close to, but |
| 1123 | not quite the same as the official Unicode property C<XID_Continue>. The |
| 1124 | difference is that this returns true only if the input character also matches |
| 1125 | L</isWORDCHAR>. See the L<top of this section|/Character classification> for |
| 1126 | an explanation of the variants. |
| 1127 | |
| 1128 | =for apidoc_section Numeric Functions |
| 1129 | |
| 1130 | =for apidoc Am|U8|READ_XDIGIT|char str* |
| 1131 | Returns the value of an ASCII-range hex digit and advances the string pointer. |
| 1132 | Behaviour is only well defined when isXDIGIT(*str) is true. |
| 1133 | |
| 1134 | =head1 Character case changing |
| 1135 | Perl uses "full" Unicode case mappings. This means that converting a single |
| 1136 | character to another case may result in a sequence of more than one character. |
| 1137 | For example, the uppercase of C<E<223>> (LATIN SMALL LETTER SHARP S) is the two |
| 1138 | character sequence C<SS>. This presents some complications The lowercase of |
| 1139 | all characters in the range 0..255 is a single character, and thus |
| 1140 | C<L</toLOWER_L1>> is furnished. But, C<toUPPER_L1> can't exist, as it couldn't |
| 1141 | return a valid result for all legal inputs. Instead C<L</toUPPER_uvchr>> has |
| 1142 | an API that does allow every possible legal result to be returned.) Likewise |
| 1143 | no other function that is crippled by not being able to give the correct |
| 1144 | results for the full range of possible inputs has been implemented here. |
| 1145 | |
| 1146 | =for apidoc Am|U8|toUPPER|int ch |
| 1147 | Converts the specified character to uppercase. If the input is anything but an |
| 1148 | ASCII lowercase character, that input character itself is returned. Variant |
| 1149 | C<toUPPER_A> is equivalent. |
| 1150 | |
| 1151 | =for apidoc Am|UV|toUPPER_uvchr|UV cp|U8* s|STRLEN* lenp |
| 1152 | Converts the code point C<cp> to its uppercase version, and |
| 1153 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code |
| 1154 | point is interpreted as native if less than 256; otherwise as Unicode. Note |
| 1155 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> |
| 1156 | bytes since the uppercase version may be longer than the original character. |
| 1157 | |
| 1158 | The first code point of the uppercased version is returned |
| 1159 | (but note, as explained at L<the top of this section|/Character case |
| 1160 | changing>, that there may be more.) |
| 1161 | |
| 1162 | =for apidoc Am|UV|toUPPER_utf8|U8* p|U8* e|U8* s|STRLEN* lenp |
| 1163 | =for apidoc_item toUPPER_utf8_safe |
| 1164 | Converts the first UTF-8 encoded character in the sequence starting at C<p> and |
| 1165 | extending no further than S<C<e - 1>> to its uppercase version, and |
| 1166 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note |
| 1167 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> |
| 1168 | bytes since the uppercase version may be longer than the original character. |
| 1169 | |
| 1170 | The first code point of the uppercased version is returned |
| 1171 | (but note, as explained at L<the top of this section|/Character case |
| 1172 | changing>, that there may be more). |
| 1173 | |
| 1174 | It will not attempt to read beyond S<C<e - 1>>, provided that the constraint |
| 1175 | S<C<s E<lt> e>> is true (this is asserted for in C<-DDEBUGGING> builds). If |
| 1176 | the UTF-8 for the input character is malformed in some way, the program may |
| 1177 | croak, or the function may return the REPLACEMENT CHARACTER, at the discretion |
| 1178 | of the implementation, and subject to change in future releases. |
| 1179 | |
| 1180 | C<toUPPER_utf8_safe> is now just a different spelling of plain C<toUPPER_utf8> |
| 1181 | |
| 1182 | =for apidoc Am|U8|toFOLD|U8 ch |
| 1183 | Converts the specified character to foldcase. If the input is anything but an |
| 1184 | ASCII uppercase character, that input character itself is returned. Variant |
| 1185 | C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full |
| 1186 | Latin1 range, as the full generality of L</toFOLD_uvchr> is needed there.) |
| 1187 | |
| 1188 | =for apidoc Am|UV|toFOLD_uvchr|UV cp|U8* s|STRLEN* lenp |
| 1189 | Converts the code point C<cp> to its foldcase version, and |
| 1190 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code |
| 1191 | point is interpreted as native if less than 256; otherwise as Unicode. Note |
| 1192 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> |
| 1193 | bytes since the foldcase version may be longer than the original character. |
| 1194 | |
| 1195 | The first code point of the foldcased version is returned |
| 1196 | (but note, as explained at L<the top of this section|/Character case |
| 1197 | changing>, that there may be more). |
| 1198 | |
| 1199 | =for apidoc Am|UV|toFOLD_utf8|U8* p|U8* e|U8* s|STRLEN* lenp |
| 1200 | =for apidoc_item toFOLD_utf8_safe |
| 1201 | Converts the first UTF-8 encoded character in the sequence starting at C<p> and |
| 1202 | extending no further than S<C<e - 1>> to its foldcase version, and |
| 1203 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note |
| 1204 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> |
| 1205 | bytes since the foldcase version may be longer than the original character. |
| 1206 | |
| 1207 | The first code point of the foldcased version is returned |
| 1208 | (but note, as explained at L<the top of this section|/Character case |
| 1209 | changing>, that there may be more). |
| 1210 | |
| 1211 | It will not attempt |
| 1212 | to read beyond S<C<e - 1>>, provided that the constraint S<C<s E<lt> e>> is |
| 1213 | true (this is asserted for in C<-DDEBUGGING> builds). If the UTF-8 for the |
| 1214 | input character is malformed in some way, the program may croak, or the |
| 1215 | function may return the REPLACEMENT CHARACTER, at the discretion of the |
| 1216 | implementation, and subject to change in future releases. |
| 1217 | |
| 1218 | C<toFOLD_utf8_safe> is now just a different spelling of plain C<toFOLD_utf8> |
| 1219 | |
| 1220 | =for apidoc Am|U8|toLOWER|U8 ch |
| 1221 | Converts the specified character to lowercase. If the input is anything but an |
| 1222 | ASCII uppercase character, that input character itself is returned. Variant |
| 1223 | C<toLOWER_A> is equivalent. |
| 1224 | |
| 1225 | =for apidoc Am|U8|toLOWER_L1|U8 ch |
| 1226 | Converts the specified Latin1 character to lowercase. The results are |
| 1227 | undefined if the input doesn't fit in a byte. |
| 1228 | |
| 1229 | =for apidoc Am|U8|toLOWER_LC|U8 ch |
| 1230 | Converts the specified character to lowercase using the current locale's rules, |
| 1231 | if possible; otherwise returns the input character itself. |
| 1232 | |
| 1233 | =for apidoc Am|UV|toLOWER_uvchr|UV cp|U8* s|STRLEN* lenp |
| 1234 | Converts the code point C<cp> to its lowercase version, and |
| 1235 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code |
| 1236 | point is interpreted as native if less than 256; otherwise as Unicode. Note |
| 1237 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> |
| 1238 | bytes since the lowercase version may be longer than the original character. |
| 1239 | |
| 1240 | The first code point of the lowercased version is returned |
| 1241 | (but note, as explained at L<the top of this section|/Character case |
| 1242 | changing>, that there may be more). |
| 1243 | |
| 1244 | =for apidoc Am|UV|toLOWER_utf8|U8* p|U8* e|U8* s|STRLEN* lenp |
| 1245 | =for apidoc_item toLOWER_utf8_safe |
| 1246 | Converts the first UTF-8 encoded character in the sequence starting at C<p> and |
| 1247 | extending no further than S<C<e - 1>> to its lowercase version, and |
| 1248 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note |
| 1249 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> |
| 1250 | bytes since the lowercase version may be longer than the original character. |
| 1251 | |
| 1252 | The first code point of the lowercased version is returned |
| 1253 | (but note, as explained at L<the top of this section|/Character case |
| 1254 | changing>, that there may be more). |
| 1255 | It will not attempt to read beyond S<C<e - 1>>, provided that the constraint |
| 1256 | S<C<s E<lt> e>> is true (this is asserted for in C<-DDEBUGGING> builds). If |
| 1257 | the UTF-8 for the input character is malformed in some way, the program may |
| 1258 | croak, or the function may return the REPLACEMENT CHARACTER, at the discretion |
| 1259 | of the implementation, and subject to change in future releases. |
| 1260 | |
| 1261 | C<toLOWER_utf8_safe> is now just a different spelling of plain C<toLOWER_utf8> |
| 1262 | |
| 1263 | =for apidoc Am|U8|toTITLE|U8 ch |
| 1264 | Converts the specified character to titlecase. If the input is anything but an |
| 1265 | ASCII lowercase character, that input character itself is returned. Variant |
| 1266 | C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 |
| 1267 | range, as the full generality of L</toTITLE_uvchr> is needed there. Titlecase is |
| 1268 | not a concept used in locale handling, so there is no functionality for that.) |
| 1269 | |
| 1270 | =for apidoc Am|UV|toTITLE_uvchr|UV cp|U8* s|STRLEN* lenp |
| 1271 | Converts the code point C<cp> to its titlecase version, and |
| 1272 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code |
| 1273 | point is interpreted as native if less than 256; otherwise as Unicode. Note |
| 1274 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> |
| 1275 | bytes since the titlecase version may be longer than the original character. |
| 1276 | |
| 1277 | The first code point of the titlecased version is returned |
| 1278 | (but note, as explained at L<the top of this section|/Character case |
| 1279 | changing>, that there may be more). |
| 1280 | |
| 1281 | =for apidoc Am|UV|toTITLE_utf8|U8* p|U8* e|U8* s|STRLEN* lenp |
| 1282 | =for apidoc_item toTITLE_utf8_safe |
| 1283 | Convert the first UTF-8 encoded character in the sequence starting at C<p> and |
| 1284 | extending no further than S<C<e - 1>> to its titlecase version, and |
| 1285 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note |
| 1286 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> |
| 1287 | bytes since the titlecase version may be longer than the original character. |
| 1288 | |
| 1289 | The first code point of the titlecased version is returned |
| 1290 | (but note, as explained at L<the top of this section|/Character case |
| 1291 | changing>, that there may be more). |
| 1292 | |
| 1293 | It will not attempt |
| 1294 | to read beyond S<C<e - 1>>, provided that the constraint S<C<s E<lt> e>> is |
| 1295 | true (this is asserted for in C<-DDEBUGGING> builds). If the UTF-8 for the |
| 1296 | input character is malformed in some way, the program may croak, or the |
| 1297 | function may return the REPLACEMENT CHARACTER, at the discretion of the |
| 1298 | implementation, and subject to change in future releases. |
| 1299 | |
| 1300 | C<toTITLE_utf8_safe> is now just a different spelling of plain C<toTITLE_utf8> |
| 1301 | |
| 1302 | =cut |
| 1303 | |
| 1304 | XXX Still undocumented isVERTWS_uvchr and _utf8; it's unclear what their names |
| 1305 | really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change, |
| 1306 | and aren't general purpose as they don't work on U+DF, and assert against that. |
| 1307 | |
| 1308 | Note that these macros are repeated in Devel::PPPort, so should also be |
| 1309 | patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc |
| 1310 | |
| 1311 | */ |
| 1312 | |
| 1313 | /* |
| 1314 | void below because that's the best fit, and works for Devel::PPPort |
| 1315 | =for apidoc_section Integer configuration values |
| 1316 | =for apidoc AmnU|void|WIDEST_UTYPE |
| 1317 | |
| 1318 | Yields the widest unsigned integer type on the platform, currently either |
| 1319 | C<U32> or C<64>. This can be used in declarations such as |
| 1320 | |
| 1321 | WIDEST_UTYPE my_uv; |
| 1322 | |
| 1323 | or casts |
| 1324 | |
| 1325 | my_uv = (WIDEST_UTYPE) val; |
| 1326 | |
| 1327 | =cut |
| 1328 | |
| 1329 | */ |
| 1330 | #ifdef QUADKIND |
| 1331 | # define WIDEST_UTYPE U64 |
| 1332 | #else |
| 1333 | # define WIDEST_UTYPE U32 |
| 1334 | #endif |
| 1335 | |
| 1336 | /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in |
| 1337 | * the lower 8. It is designed to be hopefully bomb-proof, making sure that no |
| 1338 | * bits of information are lost even on a 64-bit machine, but to get the |
| 1339 | * compiler to optimize it out if possible. This is because Configure makes |
| 1340 | * sure that the machine has an 8-bit byte, so if c is stored in a byte, the |
| 1341 | * sizeof() guarantees that this evaluates to a constant true at compile time. |
| 1342 | * |
| 1343 | * For Coverity, be always true, because otherwise Coverity thinks |
| 1344 | * it finds several expressions that are always true, independent |
| 1345 | * of operands. Well, they are, but that is kind of the point. |
| 1346 | */ |
| 1347 | #ifndef __COVERITY__ |
| 1348 | /* The '| 0' part ensures a compiler error if c is not integer (like e.g., a |
| 1349 | * pointer) */ |
| 1350 | #define FITS_IN_8_BITS(c) ( (sizeof(c) == 1) \ |
| 1351 | || !(((WIDEST_UTYPE)((c) | 0)) & ~0xFF)) |
| 1352 | #else |
| 1353 | #define FITS_IN_8_BITS(c) (1) |
| 1354 | #endif |
| 1355 | |
| 1356 | /* Returns true if l <= c <= (l + n), where 'l' and 'n' are non-negative |
| 1357 | * Written this way so that after optimization, only one conditional test is |
| 1358 | * needed. (The NV casts stop any warnings about comparison always being true |
| 1359 | * if called with an unsigned. The cast preserves the sign, which is all we |
| 1360 | * care about.) */ |
| 1361 | #define withinCOUNT(c, l, n) (__ASSERT_((NV) (l) >= 0) \ |
| 1362 | __ASSERT_((NV) (n) >= 0) \ |
| 1363 | (((WIDEST_UTYPE) (((c)) - ((l) | 0))) <= (((WIDEST_UTYPE) ((n) | 0))))) |
| 1364 | |
| 1365 | /* Returns true if c is in the range l..u, where 'l' is non-negative |
| 1366 | * Written this way so that after optimization, only one conditional test is |
| 1367 | * needed. */ |
| 1368 | #define inRANGE(c, l, u) (__ASSERT_((u) >= (l)) \ |
| 1369 | ( (sizeof(c) == sizeof(U8)) ? withinCOUNT(((U8) (c)), (l), ((u) - (l))) \ |
| 1370 | : (sizeof(c) == sizeof(U32)) ? withinCOUNT(((U32) (c)), (l), ((u) - (l))) \ |
| 1371 | : (__ASSERT_(sizeof(c) == sizeof(WIDEST_UTYPE)) \ |
| 1372 | withinCOUNT(((WIDEST_UTYPE) (c)), (l), ((u) - (l)))))) |
| 1373 | |
| 1374 | #ifdef EBCDIC |
| 1375 | # ifndef _ALL_SOURCE |
| 1376 | /* The native libc isascii() et.al. functions return the wrong results |
| 1377 | * on at least z/OS unless this is defined. */ |
| 1378 | # error _ALL_SOURCE should probably be defined |
| 1379 | # endif |
| 1380 | #else |
| 1381 | /* There is a simple definition of ASCII for ASCII platforms. But the |
| 1382 | * EBCDIC one isn't so simple, so is defined using table look-up like the |
| 1383 | * other macros below. |
| 1384 | * |
| 1385 | * The cast here is used instead of '(c) >= 0', because some compilers emit |
| 1386 | * a warning that that test is always true when the parameter is an |
| 1387 | * unsigned type. khw supposes that it could be written as |
| 1388 | * && ((c) == '\0' || (c) > 0) |
| 1389 | * to avoid the message, but the cast will likely avoid extra branches even |
| 1390 | * with stupid compilers. |
| 1391 | * |
| 1392 | * The '| 0' part ensures a compiler error if c is not integer (like e.g., |
| 1393 | * a pointer) */ |
| 1394 | # define isASCII(c) ((WIDEST_UTYPE)((c) | 0) < 128) |
| 1395 | #endif |
| 1396 | |
| 1397 | /* Take the eight possible bit patterns of the lower 3 bits and you get the |
| 1398 | * lower 3 bits of the 8 octal digits, in both ASCII and EBCDIC, so those bits |
| 1399 | * can be ignored. If the rest match '0', we have an octal */ |
| 1400 | #define isOCTAL_A(c) (((WIDEST_UTYPE)((c) | 0) & ~7) == '0') |
| 1401 | |
| 1402 | #ifdef H_PERL /* If have access to perl.h, lookup in its table */ |
| 1403 | |
| 1404 | /* Character class numbers. For internal core Perl use only. The ones less |
| 1405 | * than 32 are used in PL_charclass[] and the ones up through the one that |
| 1406 | * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and |
| 1407 | * related files. PL_charclass ones use names used in l1_char_class_tab.h but |
| 1408 | * their actual definitions are here. If that file has a name not used here, |
| 1409 | * it won't compile. |
| 1410 | * |
| 1411 | * The first group of these is ordered in what I (khw) estimate to be the |
| 1412 | * frequency of their use. This gives a slight edge to exiting a loop earlier |
| 1413 | * (in reginclass() in regexec.c). Except \v should be last, as it isn't a |
| 1414 | * real Posix character class, and some (small) inefficiencies in regular |
| 1415 | * expression handling would be introduced by putting it in the middle of those |
| 1416 | * that are. Also, cntrl and ascii come after the others as it may be useful |
| 1417 | * to group these which have no members that match above Latin1, (or above |
| 1418 | * ASCII in the latter case) */ |
| 1419 | |
| 1420 | # define _CC_WORDCHAR 0 /* \w and [:word:] */ |
| 1421 | # define _CC_DIGIT 1 /* \d and [:digit:] */ |
| 1422 | # define _CC_ALPHA 2 /* [:alpha:] */ |
| 1423 | # define _CC_LOWER 3 /* [:lower:] */ |
| 1424 | # define _CC_UPPER 4 /* [:upper:] */ |
| 1425 | # define _CC_PUNCT 5 /* [:punct:] */ |
| 1426 | # define _CC_PRINT 6 /* [:print:] */ |
| 1427 | # define _CC_ALPHANUMERIC 7 /* [:alnum:] */ |
| 1428 | # define _CC_GRAPH 8 /* [:graph:] */ |
| 1429 | # define _CC_CASED 9 /* [:lower:] or [:upper:] under /i */ |
| 1430 | # define _CC_SPACE 10 /* \s, [:space:] */ |
| 1431 | # define _CC_BLANK 11 /* [:blank:] */ |
| 1432 | # define _CC_XDIGIT 12 /* [:xdigit:] */ |
| 1433 | # define _CC_CNTRL 13 /* [:cntrl:] */ |
| 1434 | # define _CC_ASCII 14 /* [:ascii:] */ |
| 1435 | # define _CC_VERTSPACE 15 /* \v */ |
| 1436 | |
| 1437 | # define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE |
| 1438 | |
| 1439 | /* The members of the third group below do not need to be coordinated with data |
| 1440 | * structures in regcomp.[ch] and regexec.c. */ |
| 1441 | # define _CC_IDFIRST 16 |
| 1442 | # define _CC_CHARNAME_CONT 17 |
| 1443 | # define _CC_NONLATIN1_FOLD 18 |
| 1444 | # define _CC_NONLATIN1_SIMPLE_FOLD 19 |
| 1445 | # define _CC_QUOTEMETA 20 |
| 1446 | # define _CC_NON_FINAL_FOLD 21 |
| 1447 | # define _CC_IS_IN_SOME_FOLD 22 |
| 1448 | # define _CC_BINDIGIT 23 |
| 1449 | # define _CC_OCTDIGIT 24 |
| 1450 | # define _CC_MNEMONIC_CNTRL 25 |
| 1451 | |
| 1452 | /* This next group is only used on EBCDIC platforms, so theoretically could be |
| 1453 | * shared with something entirely different that's only on ASCII platforms */ |
| 1454 | # define _CC_UTF8_START_BYTE_IS_FOR_AT_LEAST_SURROGATE 31 |
| 1455 | /* Unused: 26-30 |
| 1456 | * If more bits are needed, one could add a second word for non-64bit |
| 1457 | * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd |
| 1458 | * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it |
| 1459 | * is used only for optimization (as of this writing), and differs in the |
| 1460 | * Latin1 range from the ALPHA bit only in two relatively unimportant |
| 1461 | * characters: the masculine and feminine ordinal indicators, so removing it |
| 1462 | * would just cause /i regexes which match them to run less efficiently. |
| 1463 | * Similarly the EBCDIC-only bits are used just for speed, and could be |
| 1464 | * replaced by other means */ |
| 1465 | |
| 1466 | #if defined(PERL_CORE) || defined(PERL_EXT) |
| 1467 | /* An enum version of the character class numbers, to help compilers |
| 1468 | * optimize */ |
| 1469 | typedef enum { |
| 1470 | _CC_ENUM_ALPHA = _CC_ALPHA, |
| 1471 | _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC, |
| 1472 | _CC_ENUM_ASCII = _CC_ASCII, |
| 1473 | _CC_ENUM_BLANK = _CC_BLANK, |
| 1474 | _CC_ENUM_CASED = _CC_CASED, |
| 1475 | _CC_ENUM_CNTRL = _CC_CNTRL, |
| 1476 | _CC_ENUM_DIGIT = _CC_DIGIT, |
| 1477 | _CC_ENUM_GRAPH = _CC_GRAPH, |
| 1478 | _CC_ENUM_LOWER = _CC_LOWER, |
| 1479 | _CC_ENUM_PRINT = _CC_PRINT, |
| 1480 | _CC_ENUM_PUNCT = _CC_PUNCT, |
| 1481 | _CC_ENUM_SPACE = _CC_SPACE, |
| 1482 | _CC_ENUM_UPPER = _CC_UPPER, |
| 1483 | _CC_ENUM_VERTSPACE = _CC_VERTSPACE, |
| 1484 | _CC_ENUM_WORDCHAR = _CC_WORDCHAR, |
| 1485 | _CC_ENUM_XDIGIT = _CC_XDIGIT |
| 1486 | } _char_class_number; |
| 1487 | #endif |
| 1488 | |
| 1489 | #define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1) |
| 1490 | |
| 1491 | START_EXTERN_C |
| 1492 | # ifdef DOINIT |
| 1493 | EXTCONST U32 PL_charclass[] = { |
| 1494 | # include "l1_char_class_tab.h" |
| 1495 | }; |
| 1496 | |
| 1497 | # else /* ! DOINIT */ |
| 1498 | EXTCONST U32 PL_charclass[]; |
| 1499 | # endif |
| 1500 | END_EXTERN_C |
| 1501 | |
| 1502 | /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */ |
| 1503 | # define _CC_mask(classnum) (1U << (classnum)) |
| 1504 | |
| 1505 | /* For internal core Perl use only: the base macro for defining macros like |
| 1506 | * isALPHA */ |
| 1507 | # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \ |
| 1508 | && (PL_charclass[(U8) (c)] & _CC_mask(classnum))) |
| 1509 | |
| 1510 | /* The mask for the _A versions of the macros; it just adds in the bit for |
| 1511 | * ASCII. */ |
| 1512 | # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII)) |
| 1513 | |
| 1514 | /* For internal core Perl use only: the base macro for defining macros like |
| 1515 | * isALPHA_A. The foo_A version makes sure that both the desired bit and |
| 1516 | * the ASCII bit are present */ |
| 1517 | # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \ |
| 1518 | && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \ |
| 1519 | == _CC_mask_A(classnum))) |
| 1520 | |
| 1521 | /* On ASCII platforms certain classes form a single range. It's faster to |
| 1522 | * special case these. isDIGIT is a single range on all platforms */ |
| 1523 | # ifdef EBCDIC |
| 1524 | # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA) |
| 1525 | # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH) |
| 1526 | # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER) |
| 1527 | # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT) |
| 1528 | # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER) |
| 1529 | # else |
| 1530 | /* By folding the upper and lowercase, we can use a single range */ |
| 1531 | # define isALPHA_A(c) inRANGE((~('A' ^ 'a') & (c)), 'A', 'Z') |
| 1532 | # define isGRAPH_A(c) inRANGE(c, ' ' + 1, 0x7e) |
| 1533 | # define isLOWER_A(c) inRANGE(c, 'a', 'z') |
| 1534 | # define isPRINT_A(c) inRANGE(c, ' ', 0x7e) |
| 1535 | # define isUPPER_A(c) inRANGE(c, 'A', 'Z') |
| 1536 | # endif |
| 1537 | # define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC) |
| 1538 | # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK) |
| 1539 | # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL) |
| 1540 | # define isDIGIT_A(c) inRANGE(c, '0', '9') |
| 1541 | # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT) |
| 1542 | # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE) |
| 1543 | # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR) |
| 1544 | # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT) /* No non-ASCII xdigits |
| 1545 | */ |
| 1546 | # define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST) |
| 1547 | # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA) |
| 1548 | # define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC) |
| 1549 | # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK) |
| 1550 | |
| 1551 | /* continuation character for legal NAME in \N{NAME} */ |
| 1552 | # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT) |
| 1553 | |
| 1554 | # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL) |
| 1555 | # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH) |
| 1556 | # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER) |
| 1557 | # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT) |
| 1558 | # define isPSXSPC_L1(c) isSPACE_L1(c) |
| 1559 | # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT) |
| 1560 | # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE) |
| 1561 | # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER) |
| 1562 | # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR) |
| 1563 | # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST) |
| 1564 | |
| 1565 | # ifdef EBCDIC |
| 1566 | # define isASCII(c) _generic_isCC(c, _CC_ASCII) |
| 1567 | # endif |
| 1568 | |
| 1569 | /* Participates in a single-character fold with a character above 255 */ |
| 1570 | # define _HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_SIMPLE_FOLD))) |
| 1571 | |
| 1572 | /* Like the above, but also can be part of a multi-char fold */ |
| 1573 | # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_FOLD))) |
| 1574 | |
| 1575 | # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA) |
| 1576 | # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \ |
| 1577 | _generic_isCC(c, _CC_NON_FINAL_FOLD) |
| 1578 | # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \ |
| 1579 | _generic_isCC(c, _CC_IS_IN_SOME_FOLD) |
| 1580 | |
| 1581 | /* is c a control character for which we have a mnemonic? */ |
| 1582 | # if defined(PERL_CORE) || defined(PERL_EXT) |
| 1583 | # define isMNEMONIC_CNTRL(c) _generic_isCC(c, _CC_MNEMONIC_CNTRL) |
| 1584 | # endif |
| 1585 | #else /* else we don't have perl.h H_PERL */ |
| 1586 | |
| 1587 | /* If we don't have perl.h, we are compiling a utility program. Below we |
| 1588 | * hard-code various macro definitions that wouldn't otherwise be available |
| 1589 | * to it. Most are coded based on first principles. These are written to |
| 1590 | * avoid EBCDIC vs. ASCII #ifdef's as much as possible. */ |
| 1591 | # define isDIGIT_A(c) inRANGE(c, '0', '9') |
| 1592 | # define isBLANK_A(c) ((c) == ' ' || (c) == '\t') |
| 1593 | # define isSPACE_A(c) (isBLANK_A(c) \ |
| 1594 | || (c) == '\n' \ |
| 1595 | || (c) == '\r' \ |
| 1596 | || (c) == '\v' \ |
| 1597 | || (c) == '\f') |
| 1598 | /* On EBCDIC, there are gaps between 'i' and 'j'; 'r' and 's'. Same for |
| 1599 | * uppercase. The tests for those aren't necessary on ASCII, but hurt only |
| 1600 | * performance (if optimization isn't on), and allow the same code to be |
| 1601 | * used for both platform types */ |
| 1602 | # define isLOWER_A(c) inRANGE((c), 'a', 'i') \ |
| 1603 | || inRANGE((c), 'j', 'r') \ |
| 1604 | || inRANGE((c), 's', 'z') |
| 1605 | # define isUPPER_A(c) inRANGE((c), 'A', 'I') \ |
| 1606 | || inRANGE((c), 'J', 'R') \ |
| 1607 | || inRANGE((c), 'S', 'Z') |
| 1608 | # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c)) |
| 1609 | # define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c)) |
| 1610 | # define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_') |
| 1611 | # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_') |
| 1612 | # define isXDIGIT_A(c) ( isDIGIT_A(c) \ |
| 1613 | || inRANGE((c), 'a', 'f') \ |
| 1614 | || inRANGE((c), 'A', 'F') |
| 1615 | # define isPUNCT_A(c) ((c) == '-' || (c) == '!' || (c) == '"' \ |
| 1616 | || (c) == '#' || (c) == '$' || (c) == '%' \ |
| 1617 | || (c) == '&' || (c) == '\'' || (c) == '(' \ |
| 1618 | || (c) == ')' || (c) == '*' || (c) == '+' \ |
| 1619 | || (c) == ',' || (c) == '.' || (c) == '/' \ |
| 1620 | || (c) == ':' || (c) == ';' || (c) == '<' \ |
| 1621 | || (c) == '=' || (c) == '>' || (c) == '?' \ |
| 1622 | || (c) == '@' || (c) == '[' || (c) == '\\' \ |
| 1623 | || (c) == ']' || (c) == '^' || (c) == '_' \ |
| 1624 | || (c) == '`' || (c) == '{' || (c) == '|' \ |
| 1625 | || (c) == '}' || (c) == '~') |
| 1626 | # define isGRAPH_A(c) (isALPHANUMERIC_A(c) || isPUNCT_A(c)) |
| 1627 | # define isPRINT_A(c) (isGRAPH_A(c) || (c) == ' ') |
| 1628 | |
| 1629 | # ifdef EBCDIC |
| 1630 | /* The below is accurate for the 3 EBCDIC code pages traditionally |
| 1631 | * supported by perl. The only difference between them in the controls |
| 1632 | * is the position of \n, and that is represented symbolically below */ |
| 1633 | # define isCNTRL_A(c) ((c) == '\0' || (c) == '\a' || (c) == '\b' \ |
| 1634 | || (c) == '\f' || (c) == '\n' || (c) == '\r' \ |
| 1635 | || (c) == '\t' || (c) == '\v' \ |
| 1636 | || inRANGE((c), 1, 3) /* SOH, STX, ETX */ \ |
| 1637 | || (c) == 7F /* U+7F DEL */ \ |
| 1638 | || inRANGE((c), 0x0E, 0x13) /* SO SI DLE \ |
| 1639 | DC[1-3] */ \ |
| 1640 | || (c) == 0x18 /* U+18 CAN */ \ |
| 1641 | || (c) == 0x19 /* U+19 EOM */ \ |
| 1642 | || inRANGE((c), 0x1C, 0x1F) /* [FGRU]S */ \ |
| 1643 | || (c) == 0x26 /* U+17 ETB */ \ |
| 1644 | || (c) == 0x27 /* U+1B ESC */ \ |
| 1645 | || (c) == 0x2D /* U+05 ENQ */ \ |
| 1646 | || (c) == 0x2E /* U+06 ACK */ \ |
| 1647 | || (c) == 0x32 /* U+16 SYN */ \ |
| 1648 | || (c) == 0x37 /* U+04 EOT */ \ |
| 1649 | || (c) == 0x3C /* U+14 DC4 */ \ |
| 1650 | || (c) == 0x3D /* U+15 NAK */ \ |
| 1651 | || (c) == 0x3F)/* U+1A SUB */ |
| 1652 | # define isASCII(c) (isCNTRL_A(c) || isPRINT_A(c)) |
| 1653 | # else /* isASCII is already defined for ASCII platforms, so can use that to |
| 1654 | define isCNTRL */ |
| 1655 | # define isCNTRL_A(c) (isASCII(c) && ! isPRINT_A(c)) |
| 1656 | # endif |
| 1657 | |
| 1658 | /* The _L1 macros may be unnecessary for the utilities; I (khw) added them |
| 1659 | * during debugging, and it seems best to keep them. We may be called |
| 1660 | * without NATIVE_TO_LATIN1 being defined. On ASCII platforms, it doesn't |
| 1661 | * do anything anyway, so make it not a problem */ |
| 1662 | # if ! defined(EBCDIC) && ! defined(NATIVE_TO_LATIN1) |
| 1663 | # define NATIVE_TO_LATIN1(ch) (ch) |
| 1664 | # endif |
| 1665 | # define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c)) |
| 1666 | # define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c)) |
| 1667 | # define isBLANK_L1(c) (isBLANK_A(c) \ |
| 1668 | || (FITS_IN_8_BITS(c) \ |
| 1669 | && NATIVE_TO_LATIN1((U8) c) == 0xA0)) |
| 1670 | # define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c))) |
| 1671 | # define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c))) |
| 1672 | # define isLOWER_L1(c) (isLOWER_A(c) \ |
| 1673 | || (FITS_IN_8_BITS(c) \ |
| 1674 | && (( NATIVE_TO_LATIN1((U8) c) >= 0xDF \ |
| 1675 | && NATIVE_TO_LATIN1((U8) c) != 0xF7) \ |
| 1676 | || NATIVE_TO_LATIN1((U8) c) == 0xAA \ |
| 1677 | || NATIVE_TO_LATIN1((U8) c) == 0xBA \ |
| 1678 | || NATIVE_TO_LATIN1((U8) c) == 0xB5))) |
| 1679 | # define isPRINT_L1(c) (isPRINT_A(c) \ |
| 1680 | || (FITS_IN_8_BITS(c) \ |
| 1681 | && NATIVE_TO_LATIN1((U8) c) >= 0xA0)) |
| 1682 | # define isPUNCT_L1(c) (isPUNCT_A(c) \ |
| 1683 | || (FITS_IN_8_BITS(c) \ |
| 1684 | && ( NATIVE_TO_LATIN1((U8) c) == 0xA1 \ |
| 1685 | || NATIVE_TO_LATIN1((U8) c) == 0xA7 \ |
| 1686 | || NATIVE_TO_LATIN1((U8) c) == 0xAB \ |
| 1687 | || NATIVE_TO_LATIN1((U8) c) == 0xB6 \ |
| 1688 | || NATIVE_TO_LATIN1((U8) c) == 0xB7 \ |
| 1689 | || NATIVE_TO_LATIN1((U8) c) == 0xBB \ |
| 1690 | || NATIVE_TO_LATIN1((U8) c) == 0xBF))) |
| 1691 | # define isSPACE_L1(c) (isSPACE_A(c) \ |
| 1692 | || (FITS_IN_8_BITS(c) \ |
| 1693 | && ( NATIVE_TO_LATIN1((U8) c) == 0x85 \ |
| 1694 | || NATIVE_TO_LATIN1((U8) c) == 0xA0))) |
| 1695 | # define isUPPER_L1(c) (isUPPER_A(c) \ |
| 1696 | || (FITS_IN_8_BITS(c) \ |
| 1697 | && ( IN_RANGE(NATIVE_TO_LATIN1((U8) c), \ |
| 1698 | 0xC0, 0xDE) \ |
| 1699 | && NATIVE_TO_LATIN1((U8) c) != 0xD7))) |
| 1700 | # define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c)) |
| 1701 | # define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_') |
| 1702 | # define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \ |
| 1703 | || isBLANK_L1(c) \ |
| 1704 | || (c) == '-' \ |
| 1705 | || (c) == '(' \ |
| 1706 | || (c) == ')') |
| 1707 | /* The following are not fully accurate in the above-ASCII range. I (khw) |
| 1708 | * don't think it's necessary to be so for the purposes where this gets |
| 1709 | * compiled */ |
| 1710 | # define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c)) |
| 1711 | # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c) |
| 1712 | |
| 1713 | /* And these aren't accurate at all. They are useful only for above |
| 1714 | * Latin1, which utilities and bootstrapping don't deal with */ |
| 1715 | # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0 |
| 1716 | # define _HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0 |
| 1717 | # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0 |
| 1718 | |
| 1719 | /* Many of the macros later in this file are defined in terms of these. By |
| 1720 | * implementing them with a function, which converts the class number into |
| 1721 | * a call to the desired macro, all of the later ones work. However, that |
| 1722 | * function won't be actually defined when building a utility program (no |
| 1723 | * perl.h), and so a compiler error will be generated if one is attempted |
| 1724 | * to be used. And the above-Latin1 code points require Unicode tables to |
| 1725 | * be present, something unlikely to be the case when bootstrapping */ |
| 1726 | # define _generic_isCC(c, classnum) \ |
| 1727 | (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE)) |
| 1728 | # define _generic_isCC_A(c, classnum) \ |
| 1729 | (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE)) |
| 1730 | #endif /* End of no perl.h H_PERL */ |
| 1731 | |
| 1732 | #define isALPHANUMERIC(c) isALPHANUMERIC_A(c) |
| 1733 | #define isALPHA(c) isALPHA_A(c) |
| 1734 | #define isASCII_A(c) isASCII(c) |
| 1735 | #define isASCII_L1(c) isASCII(c) |
| 1736 | #define isBLANK(c) isBLANK_A(c) |
| 1737 | #define isCNTRL(c) isCNTRL_A(c) |
| 1738 | #define isDIGIT(c) isDIGIT_A(c) |
| 1739 | #define isGRAPH(c) isGRAPH_A(c) |
| 1740 | #define isIDFIRST(c) isIDFIRST_A(c) |
| 1741 | #define isLOWER(c) isLOWER_A(c) |
| 1742 | #define isPRINT(c) isPRINT_A(c) |
| 1743 | #define isPSXSPC_A(c) isSPACE_A(c) |
| 1744 | #define isPSXSPC(c) isPSXSPC_A(c) |
| 1745 | #define isPSXSPC_L1(c) isSPACE_L1(c) |
| 1746 | #define isPUNCT(c) isPUNCT_A(c) |
| 1747 | #define isSPACE(c) isSPACE_A(c) |
| 1748 | #define isUPPER(c) isUPPER_A(c) |
| 1749 | #define isWORDCHAR(c) isWORDCHAR_A(c) |
| 1750 | #define isXDIGIT(c) isXDIGIT_A(c) |
| 1751 | |
| 1752 | /* ASCII casing. These could also be written as |
| 1753 | #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c)) |
| 1754 | #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c)) |
| 1755 | which uses table lookup and mask instead of subtraction. (This would |
| 1756 | work because the _MOD does not apply in the ASCII range). |
| 1757 | |
| 1758 | These actually are UTF-8 invariant casing, not just ASCII, as any non-ASCII |
| 1759 | UTF-8 invariants are neither upper nor lower. (Only on EBCDIC platforms are |
| 1760 | there non-ASCII invariants, and all of them are controls.) */ |
| 1761 | #define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c)) |
| 1762 | #define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c)) |
| 1763 | |
| 1764 | /* In the ASCII range, these are equivalent to what they're here defined to be. |
| 1765 | * But by creating these definitions, other code doesn't have to be aware of |
| 1766 | * this detail. Actually this works for all UTF-8 invariants, not just the |
| 1767 | * ASCII range. (EBCDIC platforms can have non-ASCII invariants.) */ |
| 1768 | #define toFOLD(c) toLOWER(c) |
| 1769 | #define toTITLE(c) toUPPER(c) |
| 1770 | |
| 1771 | #define toLOWER_A(c) toLOWER(c) |
| 1772 | #define toUPPER_A(c) toUPPER(c) |
| 1773 | #define toFOLD_A(c) toFOLD(c) |
| 1774 | #define toTITLE_A(c) toTITLE(c) |
| 1775 | |
| 1776 | /* Use table lookup for speed; returns the input itself if is out-of-range */ |
| 1777 | #define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \ |
| 1778 | ? (c) \ |
| 1779 | : PL_latin1_lc[ (U8) (c) ]) |
| 1780 | #define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */ |
| 1781 | |
| 1782 | /* Modified uc. Is correct uc except for three non-ascii chars which are |
| 1783 | * all mapped to one of them, and these need special handling; returns the |
| 1784 | * input itself if is out-of-range */ |
| 1785 | #define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \ |
| 1786 | ? (c) \ |
| 1787 | : PL_mod_latin1_uc[ (U8) (c) ]) |
| 1788 | #define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale |
| 1789 | |
| 1790 | /* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */ |
| 1791 | |
| 1792 | /* For internal core Perl use only: the base macro for defining macros like |
| 1793 | * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point |
| 1794 | * (0-255) to check. In a UTF-8 locale, the result is the same as calling |
| 1795 | * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like |
| 1796 | * _CC_UPPER, which gives the class number for doing this. For non-UTF-8 |
| 1797 | * locales, the code to actually do the test this is passed in 'non_utf8'. If |
| 1798 | * 'c' is above 255, 0 is returned. For accessing the full range of possible |
| 1799 | * code points under locale rules, use the macros based on _generic_LC_uvchr |
| 1800 | * instead of this. */ |
| 1801 | #define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \ |
| 1802 | (! FITS_IN_8_BITS(c) \ |
| 1803 | ? 0 \ |
| 1804 | : IN_UTF8_CTYPE_LOCALE \ |
| 1805 | ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \ |
| 1806 | : cBOOL(non_utf8)) |
| 1807 | |
| 1808 | /* For internal core Perl use only: a helper macro for defining macros like |
| 1809 | * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to |
| 1810 | * actually do this test is passed in 'non_utf8_func', which is called on 'c', |
| 1811 | * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See |
| 1812 | * _generic_LC_base for more info */ |
| 1813 | #define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \ |
| 1814 | _generic_LC_base(c,utf8_locale_classnum, \ |
| 1815 | non_utf8_func( (_LC_CAST) (c))) |
| 1816 | |
| 1817 | /* For internal core Perl use only: like _generic_LC, but also returns TRUE if |
| 1818 | * 'c' is the platform's native underscore character */ |
| 1819 | #define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \ |
| 1820 | _generic_LC_base(c, utf8_locale_classnum, \ |
| 1821 | (non_utf8_func( (_LC_CAST) (c)) \ |
| 1822 | || (char)(c) == '_')) |
| 1823 | |
| 1824 | /* These next three are also for internal core Perl use only: case-change |
| 1825 | * helper macros. The reason for using the PL_latin arrays is in case the |
| 1826 | * system function is defective; it ensures uniform results that conform to the |
| 1827 | * Unicod standard. It does not handle the anomalies in UTF-8 Turkic locales */ |
| 1828 | #define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \ |
| 1829 | ? (c) \ |
| 1830 | : (IN_UTF8_CTYPE_LOCALE) \ |
| 1831 | ? PL_latin1_lc[ (U8) (c) ] \ |
| 1832 | : (cast)function((cast)(c))) |
| 1833 | |
| 1834 | /* Note that the result can be larger than a byte in a UTF-8 locale. It |
| 1835 | * returns a single value, so can't adequately return the upper case of LATIN |
| 1836 | * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two |
| 1837 | * values "SS"); instead it asserts against that under DEBUGGING, and |
| 1838 | * otherwise returns its input. It does not handle the anomalies in UTF-8 |
| 1839 | * Turkic locales. */ |
| 1840 | #define _generic_toUPPER_LC(c, function, cast) \ |
| 1841 | (! FITS_IN_8_BITS(c) \ |
| 1842 | ? (c) \ |
| 1843 | : ((! IN_UTF8_CTYPE_LOCALE) \ |
| 1844 | ? (cast)function((cast)(c)) \ |
| 1845 | : ((((U8)(c)) == MICRO_SIGN) \ |
| 1846 | ? GREEK_CAPITAL_LETTER_MU \ |
| 1847 | : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \ |
| 1848 | ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \ |
| 1849 | : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \ |
| 1850 | ? (__ASSERT_(0) (c)) \ |
| 1851 | : PL_mod_latin1_uc[ (U8) (c) ]))))) |
| 1852 | |
| 1853 | /* Note that the result can be larger than a byte in a UTF-8 locale. It |
| 1854 | * returns a single value, so can't adequately return the fold case of LATIN |
| 1855 | * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two |
| 1856 | * values "ss"); instead it asserts against that under DEBUGGING, and |
| 1857 | * otherwise returns its input. It does not handle the anomalies in UTF-8 |
| 1858 | * Turkic locales */ |
| 1859 | #define _generic_toFOLD_LC(c, function, cast) \ |
| 1860 | ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \ |
| 1861 | ? GREEK_SMALL_LETTER_MU \ |
| 1862 | : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \ |
| 1863 | || (c) != LATIN_SMALL_LETTER_SHARP_S) \ |
| 1864 | _generic_toLOWER_LC(c, function, cast))) |
| 1865 | |
| 1866 | /* Use the libc versions for these if available. */ |
| 1867 | #if defined(HAS_ISASCII) |
| 1868 | # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c))) |
| 1869 | #else |
| 1870 | # define isASCII_LC(c) isASCII(c) |
| 1871 | #endif |
| 1872 | |
| 1873 | #if defined(HAS_ISBLANK) |
| 1874 | # define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank) |
| 1875 | #else /* Unlike isASCII, varies if in a UTF-8 locale */ |
| 1876 | # define isBLANK_LC(c) ((IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c)) |
| 1877 | #endif |
| 1878 | |
| 1879 | #define _LC_CAST U8 |
| 1880 | |
| 1881 | #ifdef WIN32 |
| 1882 | /* The Windows functions don't bother to follow the POSIX standard, which |
| 1883 | * for example says that something can't both be a printable and a control. |
| 1884 | * But Windows treats the \t control as a printable, and does such things |
| 1885 | * as making superscripts into both digits and punctuation. This tames |
| 1886 | * these flaws by assuming that the definitions of both controls and space |
| 1887 | * are correct, and then making sure that other definitions don't have |
| 1888 | * weirdnesses, by making sure that isalnum() isn't also ispunct(), etc. |
| 1889 | * Not all possible weirdnesses are checked for, just the ones that were |
| 1890 | * detected on actual Microsoft code pages */ |
| 1891 | |
| 1892 | # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl) |
| 1893 | # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace) |
| 1894 | |
| 1895 | # define isALPHA_LC(c) (_generic_LC(c, _CC_ALPHA, isalpha) \ |
| 1896 | && isALPHANUMERIC_LC(c)) |
| 1897 | # define isALPHANUMERIC_LC(c) (_generic_LC(c, _CC_ALPHANUMERIC, isalnum) && \ |
| 1898 | ! isPUNCT_LC(c)) |
| 1899 | # define isDIGIT_LC(c) (_generic_LC(c, _CC_DIGIT, isdigit) && \ |
| 1900 | isALPHANUMERIC_LC(c)) |
| 1901 | # define isGRAPH_LC(c) (_generic_LC(c, _CC_GRAPH, isgraph) && isPRINT_LC(c)) |
| 1902 | # define isIDFIRST_LC(c) (((c) == '_') \ |
| 1903 | || (_generic_LC(c, _CC_IDFIRST, isalpha) && ! isPUNCT_LC(c))) |
| 1904 | # define isLOWER_LC(c) (_generic_LC(c, _CC_LOWER, islower) && isALPHA_LC(c)) |
| 1905 | # define isPRINT_LC(c) (_generic_LC(c, _CC_PRINT, isprint) && ! isCNTRL_LC(c)) |
| 1906 | # define isPUNCT_LC(c) (_generic_LC(c, _CC_PUNCT, ispunct) && ! isCNTRL_LC(c)) |
| 1907 | # define isUPPER_LC(c) (_generic_LC(c, _CC_UPPER, isupper) && isALPHA_LC(c)) |
| 1908 | # define isWORDCHAR_LC(c) (((c) == '_') || isALPHANUMERIC_LC(c)) |
| 1909 | # define isXDIGIT_LC(c) (_generic_LC(c, _CC_XDIGIT, isxdigit) \ |
| 1910 | && isALPHANUMERIC_LC(c)) |
| 1911 | |
| 1912 | # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8) |
| 1913 | # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8) |
| 1914 | # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8) |
| 1915 | |
| 1916 | #elif defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII)) |
| 1917 | /* For most other platforms */ |
| 1918 | |
| 1919 | # define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha) |
| 1920 | # define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum) |
| 1921 | # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl) |
| 1922 | # define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit) |
| 1923 | # define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph) |
| 1924 | # define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha) |
| 1925 | # define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower) |
| 1926 | # define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint) |
| 1927 | # define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct) |
| 1928 | # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace) |
| 1929 | # define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper) |
| 1930 | # define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum) |
| 1931 | # define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit) |
| 1932 | |
| 1933 | |
| 1934 | # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8) |
| 1935 | # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8) |
| 1936 | # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8) |
| 1937 | |
| 1938 | #else /* The final fallback position */ |
| 1939 | |
| 1940 | # define isALPHA_LC(c) (isascii(c) && isalpha(c)) |
| 1941 | # define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c)) |
| 1942 | # define isCNTRL_LC(c) (isascii(c) && iscntrl(c)) |
| 1943 | # define isDIGIT_LC(c) (isascii(c) && isdigit(c)) |
| 1944 | # define isGRAPH_LC(c) (isascii(c) && isgraph(c)) |
| 1945 | # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_')) |
| 1946 | # define isLOWER_LC(c) (isascii(c) && islower(c)) |
| 1947 | # define isPRINT_LC(c) (isascii(c) && isprint(c)) |
| 1948 | # define isPUNCT_LC(c) (isascii(c) && ispunct(c)) |
| 1949 | # define isSPACE_LC(c) (isascii(c) && isspace(c)) |
| 1950 | # define isUPPER_LC(c) (isascii(c) && isupper(c)) |
| 1951 | # define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_')) |
| 1952 | # define isXDIGIT_LC(c) (isascii(c) && isxdigit(c)) |
| 1953 | |
| 1954 | # define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c)) |
| 1955 | # define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c)) |
| 1956 | # define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c)) |
| 1957 | |
| 1958 | #endif |
| 1959 | |
| 1960 | #define isIDCONT(c) isWORDCHAR(c) |
| 1961 | #define isIDCONT_A(c) isWORDCHAR_A(c) |
| 1962 | #define isIDCONT_L1(c) isWORDCHAR_L1(c) |
| 1963 | #define isIDCONT_LC(c) isWORDCHAR_LC(c) |
| 1964 | #define isPSXSPC_LC(c) isSPACE_LC(c) |
| 1965 | |
| 1966 | /* For internal core Perl use only: the base macros for defining macros like |
| 1967 | * isALPHA_uvchr. 'c' is the code point to check. 'classnum' is the POSIX class |
| 1968 | * number defined earlier in this file. _generic_uvchr() is used for POSIX |
| 1969 | * classes where there is a macro or function 'above_latin1' that takes the |
| 1970 | * single argument 'c' and returns the desired value. These exist for those |
| 1971 | * classes which have simple definitions, avoiding the overhead of an inversion |
| 1972 | * list binary search. _generic_invlist_uvchr() can be used |
| 1973 | * for classes where that overhead is faster than a direct lookup. |
| 1974 | * _generic_uvchr() won't compile if 'c' isn't unsigned, as it won't match the |
| 1975 | * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so |
| 1976 | * have duplicate checks here, so could create versions of the macros that |
| 1977 | * don't, but experiments show that gcc optimizes them out anyway. */ |
| 1978 | |
| 1979 | /* Note that all ignore 'use bytes' */ |
| 1980 | #define _generic_uvchr(classnum, above_latin1, c) ((c) < 256 \ |
| 1981 | ? _generic_isCC(c, classnum) \ |
| 1982 | : above_latin1(c)) |
| 1983 | #define _generic_invlist_uvchr(classnum, c) ((c) < 256 \ |
| 1984 | ? _generic_isCC(c, classnum) \ |
| 1985 | : _is_uni_FOO(classnum, c)) |
| 1986 | #define isALPHA_uvchr(c) _generic_invlist_uvchr(_CC_ALPHA, c) |
| 1987 | #define isALPHANUMERIC_uvchr(c) _generic_invlist_uvchr(_CC_ALPHANUMERIC, c) |
| 1988 | #define isASCII_uvchr(c) isASCII(c) |
| 1989 | #define isBLANK_uvchr(c) _generic_uvchr(_CC_BLANK, is_HORIZWS_cp_high, c) |
| 1990 | #define isCNTRL_uvchr(c) isCNTRL_L1(c) /* All controls are in Latin1 */ |
| 1991 | #define isDIGIT_uvchr(c) _generic_invlist_uvchr(_CC_DIGIT, c) |
| 1992 | #define isGRAPH_uvchr(c) _generic_invlist_uvchr(_CC_GRAPH, c) |
| 1993 | #define isIDCONT_uvchr(c) \ |
| 1994 | _generic_uvchr(_CC_WORDCHAR, _is_uni_perl_idcont, c) |
| 1995 | #define isIDFIRST_uvchr(c) \ |
| 1996 | _generic_uvchr(_CC_IDFIRST, _is_uni_perl_idstart, c) |
| 1997 | #define isLOWER_uvchr(c) _generic_invlist_uvchr(_CC_LOWER, c) |
| 1998 | #define isPRINT_uvchr(c) _generic_invlist_uvchr(_CC_PRINT, c) |
| 1999 | |
| 2000 | #define isPUNCT_uvchr(c) _generic_invlist_uvchr(_CC_PUNCT, c) |
| 2001 | #define isSPACE_uvchr(c) _generic_uvchr(_CC_SPACE, is_XPERLSPACE_cp_high, c) |
| 2002 | #define isPSXSPC_uvchr(c) isSPACE_uvchr(c) |
| 2003 | |
| 2004 | #define isUPPER_uvchr(c) _generic_invlist_uvchr(_CC_UPPER, c) |
| 2005 | #define isVERTWS_uvchr(c) _generic_uvchr(_CC_VERTSPACE, is_VERTWS_cp_high, c) |
| 2006 | #define isWORDCHAR_uvchr(c) _generic_invlist_uvchr(_CC_WORDCHAR, c) |
| 2007 | #define isXDIGIT_uvchr(c) _generic_uvchr(_CC_XDIGIT, is_XDIGIT_cp_high, c) |
| 2008 | |
| 2009 | #define toFOLD_uvchr(c,s,l) to_uni_fold(c,s,l) |
| 2010 | #define toLOWER_uvchr(c,s,l) to_uni_lower(c,s,l) |
| 2011 | #define toTITLE_uvchr(c,s,l) to_uni_title(c,s,l) |
| 2012 | #define toUPPER_uvchr(c,s,l) to_uni_upper(c,s,l) |
| 2013 | |
| 2014 | /* For backwards compatibility, even though '_uni' should mean official Unicode |
| 2015 | * code points, in Perl it means native for those below 256 */ |
| 2016 | #define isALPHA_uni(c) isALPHA_uvchr(c) |
| 2017 | #define isALPHANUMERIC_uni(c) isALPHANUMERIC_uvchr(c) |
| 2018 | #define isASCII_uni(c) isASCII_uvchr(c) |
| 2019 | #define isBLANK_uni(c) isBLANK_uvchr(c) |
| 2020 | #define isCNTRL_uni(c) isCNTRL_uvchr(c) |
| 2021 | #define isDIGIT_uni(c) isDIGIT_uvchr(c) |
| 2022 | #define isGRAPH_uni(c) isGRAPH_uvchr(c) |
| 2023 | #define isIDCONT_uni(c) isIDCONT_uvchr(c) |
| 2024 | #define isIDFIRST_uni(c) isIDFIRST_uvchr(c) |
| 2025 | #define isLOWER_uni(c) isLOWER_uvchr(c) |
| 2026 | #define isPRINT_uni(c) isPRINT_uvchr(c) |
| 2027 | #define isPUNCT_uni(c) isPUNCT_uvchr(c) |
| 2028 | #define isSPACE_uni(c) isSPACE_uvchr(c) |
| 2029 | #define isPSXSPC_uni(c) isPSXSPC_uvchr(c) |
| 2030 | #define isUPPER_uni(c) isUPPER_uvchr(c) |
| 2031 | #define isVERTWS_uni(c) isVERTWS_uvchr(c) |
| 2032 | #define isWORDCHAR_uni(c) isWORDCHAR_uvchr(c) |
| 2033 | #define isXDIGIT_uni(c) isXDIGIT_uvchr(c) |
| 2034 | #define toFOLD_uni(c,s,l) toFOLD_uvchr(c,s,l) |
| 2035 | #define toLOWER_uni(c,s,l) toLOWER_uvchr(c,s,l) |
| 2036 | #define toTITLE_uni(c,s,l) toTITLE_uvchr(c,s,l) |
| 2037 | #define toUPPER_uni(c,s,l) toUPPER_uvchr(c,s,l) |
| 2038 | |
| 2039 | /* For internal core Perl use only: the base macros for defining macros like |
| 2040 | * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code |
| 2041 | * point, not just 0-255. Like _generic_uvchr, there are two versions, one for |
| 2042 | * simple class definitions; the other for more complex. These are like |
| 2043 | * _generic_uvchr, so see it for more info. */ |
| 2044 | #define _generic_LC_uvchr(latin1, above_latin1, c) \ |
| 2045 | (c < 256 ? latin1(c) : above_latin1(c)) |
| 2046 | #define _generic_LC_invlist_uvchr(latin1, classnum, c) \ |
| 2047 | (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c)) |
| 2048 | |
| 2049 | #define isALPHA_LC_uvchr(c) _generic_LC_invlist_uvchr(isALPHA_LC, _CC_ALPHA, c) |
| 2050 | #define isALPHANUMERIC_LC_uvchr(c) _generic_LC_invlist_uvchr(isALPHANUMERIC_LC, \ |
| 2051 | _CC_ALPHANUMERIC, c) |
| 2052 | #define isASCII_LC_uvchr(c) isASCII_LC(c) |
| 2053 | #define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, \ |
| 2054 | is_HORIZWS_cp_high, c) |
| 2055 | #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0) |
| 2056 | #define isDIGIT_LC_uvchr(c) _generic_LC_invlist_uvchr(isDIGIT_LC, _CC_DIGIT, c) |
| 2057 | #define isGRAPH_LC_uvchr(c) _generic_LC_invlist_uvchr(isGRAPH_LC, _CC_GRAPH, c) |
| 2058 | #define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \ |
| 2059 | _is_uni_perl_idcont, c) |
| 2060 | #define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \ |
| 2061 | _is_uni_perl_idstart, c) |
| 2062 | #define isLOWER_LC_uvchr(c) _generic_LC_invlist_uvchr(isLOWER_LC, _CC_LOWER, c) |
| 2063 | #define isPRINT_LC_uvchr(c) _generic_LC_invlist_uvchr(isPRINT_LC, _CC_PRINT, c) |
| 2064 | #define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) |
| 2065 | #define isPUNCT_LC_uvchr(c) _generic_LC_invlist_uvchr(isPUNCT_LC, _CC_PUNCT, c) |
| 2066 | #define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \ |
| 2067 | is_XPERLSPACE_cp_high, c) |
| 2068 | #define isUPPER_LC_uvchr(c) _generic_LC_invlist_uvchr(isUPPER_LC, _CC_UPPER, c) |
| 2069 | #define isWORDCHAR_LC_uvchr(c) _generic_LC_invlist_uvchr(isWORDCHAR_LC, \ |
| 2070 | _CC_WORDCHAR, c) |
| 2071 | #define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, \ |
| 2072 | is_XDIGIT_cp_high, c) |
| 2073 | |
| 2074 | #define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c)) |
| 2075 | |
| 2076 | /* The "_safe" macros make sure that we don't attempt to read beyond 'e', but |
| 2077 | * they don't otherwise go out of their way to look for malformed UTF-8. If |
| 2078 | * they can return accurate results without knowing if the input is otherwise |
| 2079 | * malformed, they do so. For example isASCII is accurate in spite of any |
| 2080 | * non-length malformations because it looks only at a single byte. Likewise |
| 2081 | * isDIGIT looks just at the first byte for code points 0-255, as all UTF-8 |
| 2082 | * variant ones return FALSE. But, if the input has to be well-formed in order |
| 2083 | * for the results to be accurate, the macros will test and if malformed will |
| 2084 | * call a routine to die |
| 2085 | * |
| 2086 | * Except for toke.c, the macros do assume that e > p, asserting that on |
| 2087 | * DEBUGGING builds. Much code that calls these depends on this being true, |
| 2088 | * for other reasons. toke.c is treated specially as using the regular |
| 2089 | * assertion breaks it in many ways. All strings that these operate on there |
| 2090 | * are supposed to have an extra NUL character at the end, so that *e = \0. A |
| 2091 | * bunch of code in toke.c assumes that this is true, so the assertion allows |
| 2092 | * for that */ |
| 2093 | #ifdef PERL_IN_TOKE_C |
| 2094 | # define _utf8_safe_assert(p,e) ((e) > (p) || ((e) == (p) && *(p) == '\0')) |
| 2095 | #else |
| 2096 | # define _utf8_safe_assert(p,e) ((e) > (p)) |
| 2097 | #endif |
| 2098 | |
| 2099 | #define _generic_utf8_safe(classnum, p, e, above_latin1) \ |
| 2100 | ((! _utf8_safe_assert(p, e)) \ |
| 2101 | ? (_force_out_malformed_utf8_message((U8 *) (p), (U8 *) (e), 0, 1), 0)\ |
| 2102 | : (UTF8_IS_INVARIANT(*(p))) \ |
| 2103 | ? _generic_isCC(*(p), classnum) \ |
| 2104 | : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \ |
| 2105 | ? ((LIKELY((e) - (p) > 1 && UTF8_IS_CONTINUATION(*((p)+1)))) \ |
| 2106 | ? _generic_isCC(EIGHT_BIT_UTF8_TO_NATIVE(*(p), *((p)+1 )), \ |
| 2107 | classnum) \ |
| 2108 | : (_force_out_malformed_utf8_message( \ |
| 2109 | (U8 *) (p), (U8 *) (e), 0, 1), 0)) \ |
| 2110 | : above_latin1)) |
| 2111 | /* Like the above, but calls 'above_latin1(p)' to get the utf8 value. |
| 2112 | * 'above_latin1' can be a macro */ |
| 2113 | #define _generic_func_utf8_safe(classnum, above_latin1, p, e) \ |
| 2114 | _generic_utf8_safe(classnum, p, e, above_latin1(p, e)) |
| 2115 | #define _generic_non_invlist_utf8_safe(classnum, above_latin1, p, e) \ |
| 2116 | _generic_utf8_safe(classnum, p, e, \ |
| 2117 | (UNLIKELY((e) - (p) < UTF8SKIP(p)) \ |
| 2118 | ? (_force_out_malformed_utf8_message( \ |
| 2119 | (U8 *) (p), (U8 *) (e), 0, 1), 0) \ |
| 2120 | : above_latin1(p))) |
| 2121 | /* Like the above, but passes classnum to _isFOO_utf8(), instead of having an |
| 2122 | * 'above_latin1' parameter */ |
| 2123 | #define _generic_invlist_utf8_safe(classnum, p, e) \ |
| 2124 | _generic_utf8_safe(classnum, p, e, _is_utf8_FOO(classnum, p, e)) |
| 2125 | |
| 2126 | /* Like the above, but should be used only when it is known that there are no |
| 2127 | * characters in the upper-Latin1 range (128-255 on ASCII platforms) which the |
| 2128 | * class is TRUE for. Hence it can skip the tests for this range. |
| 2129 | * 'above_latin1' should include its arguments */ |
| 2130 | #define _generic_utf8_safe_no_upper_latin1(classnum, p, e, above_latin1) \ |
| 2131 | (__ASSERT_(_utf8_safe_assert(p, e)) \ |
| 2132 | (UTF8_IS_INVARIANT(*(p))) \ |
| 2133 | ? _generic_isCC(*(p), classnum) \ |
| 2134 | : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \ |
| 2135 | ? 0 /* Note that doesn't check validity for latin1 */ \ |
| 2136 | : above_latin1) |
| 2137 | |
| 2138 | |
| 2139 | #define isALPHA_utf8(p, e) isALPHA_utf8_safe(p, e) |
| 2140 | #define isALPHANUMERIC_utf8(p, e) isALPHANUMERIC_utf8_safe(p, e) |
| 2141 | #define isASCII_utf8(p, e) isASCII_utf8_safe(p, e) |
| 2142 | #define isBLANK_utf8(p, e) isBLANK_utf8_safe(p, e) |
| 2143 | #define isCNTRL_utf8(p, e) isCNTRL_utf8_safe(p, e) |
| 2144 | #define isDIGIT_utf8(p, e) isDIGIT_utf8_safe(p, e) |
| 2145 | #define isGRAPH_utf8(p, e) isGRAPH_utf8_safe(p, e) |
| 2146 | #define isIDCONT_utf8(p, e) isIDCONT_utf8_safe(p, e) |
| 2147 | #define isIDFIRST_utf8(p, e) isIDFIRST_utf8_safe(p, e) |
| 2148 | #define isLOWER_utf8(p, e) isLOWER_utf8_safe(p, e) |
| 2149 | #define isPRINT_utf8(p, e) isPRINT_utf8_safe(p, e) |
| 2150 | #define isPSXSPC_utf8(p, e) isPSXSPC_utf8_safe(p, e) |
| 2151 | #define isPUNCT_utf8(p, e) isPUNCT_utf8_safe(p, e) |
| 2152 | #define isSPACE_utf8(p, e) isSPACE_utf8_safe(p, e) |
| 2153 | #define isUPPER_utf8(p, e) isUPPER_utf8_safe(p, e) |
| 2154 | #define isVERTWS_utf8(p, e) isVERTWS_utf8_safe(p, e) |
| 2155 | #define isWORDCHAR_utf8(p, e) isWORDCHAR_utf8_safe(p, e) |
| 2156 | #define isXDIGIT_utf8(p, e) isXDIGIT_utf8_safe(p, e) |
| 2157 | |
| 2158 | #define isALPHA_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_ALPHA, p, e) |
| 2159 | #define isALPHANUMERIC_utf8_safe(p, e) \ |
| 2160 | _generic_invlist_utf8_safe(_CC_ALPHANUMERIC, p, e) |
| 2161 | #define isASCII_utf8_safe(p, e) \ |
| 2162 | /* Because ASCII is invariant under utf8, the non-utf8 macro \ |
| 2163 | * works */ \ |
| 2164 | (__ASSERT_(_utf8_safe_assert(p, e)) isASCII(*(p))) |
| 2165 | #define isBLANK_utf8_safe(p, e) \ |
| 2166 | _generic_non_invlist_utf8_safe(_CC_BLANK, is_HORIZWS_high, p, e) |
| 2167 | |
| 2168 | #ifdef EBCDIC |
| 2169 | /* Because all controls are UTF-8 invariants in EBCDIC, we can use this |
| 2170 | * more efficient macro instead of the more general one */ |
| 2171 | # define isCNTRL_utf8_safe(p, e) \ |
| 2172 | (__ASSERT_(_utf8_safe_assert(p, e)) isCNTRL_L1(*(p))) |
| 2173 | #else |
| 2174 | # define isCNTRL_utf8_safe(p, e) _generic_utf8_safe(_CC_CNTRL, p, e, 0) |
| 2175 | #endif |
| 2176 | |
| 2177 | #define isDIGIT_utf8_safe(p, e) \ |
| 2178 | _generic_utf8_safe_no_upper_latin1(_CC_DIGIT, p, e, \ |
| 2179 | _is_utf8_FOO(_CC_DIGIT, p, e)) |
| 2180 | #define isGRAPH_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_GRAPH, p, e) |
| 2181 | #define isIDCONT_utf8_safe(p, e) _generic_func_utf8_safe(_CC_WORDCHAR, \ |
| 2182 | _is_utf8_perl_idcont, p, e) |
| 2183 | |
| 2184 | /* To prevent S_scan_word in toke.c from hanging, we have to make sure that |
| 2185 | * IDFIRST is an alnum. See |
| 2186 | * https://github.com/Perl/perl5/issues/10275 for more detail than you |
| 2187 | * ever wanted to know about. (In the ASCII range, there isn't a difference.) |
| 2188 | * This used to be not the XID version, but we decided to go with the more |
| 2189 | * modern Unicode definition */ |
| 2190 | #define isIDFIRST_utf8_safe(p, e) \ |
| 2191 | _generic_func_utf8_safe(_CC_IDFIRST, \ |
| 2192 | _is_utf8_perl_idstart, (U8 *) (p), (U8 *) (e)) |
| 2193 | |
| 2194 | #define isLOWER_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_LOWER, p, e) |
| 2195 | #define isPRINT_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_PRINT, p, e) |
| 2196 | #define isPSXSPC_utf8_safe(p, e) isSPACE_utf8_safe(p, e) |
| 2197 | #define isPUNCT_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_PUNCT, p, e) |
| 2198 | #define isSPACE_utf8_safe(p, e) \ |
| 2199 | _generic_non_invlist_utf8_safe(_CC_SPACE, is_XPERLSPACE_high, p, e) |
| 2200 | #define isUPPER_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_UPPER, p, e) |
| 2201 | #define isVERTWS_utf8_safe(p, e) \ |
| 2202 | _generic_non_invlist_utf8_safe(_CC_VERTSPACE, is_VERTWS_high, p, e) |
| 2203 | #define isWORDCHAR_utf8_safe(p, e) \ |
| 2204 | _generic_invlist_utf8_safe(_CC_WORDCHAR, p, e) |
| 2205 | #define isXDIGIT_utf8_safe(p, e) \ |
| 2206 | _generic_utf8_safe_no_upper_latin1(_CC_XDIGIT, p, e, \ |
| 2207 | (UNLIKELY((e) - (p) < UTF8SKIP(p)) \ |
| 2208 | ? (_force_out_malformed_utf8_message( \ |
| 2209 | (U8 *) (p), (U8 *) (e), 0, 1), 0) \ |
| 2210 | : is_XDIGIT_high(p))) |
| 2211 | |
| 2212 | #define toFOLD_utf8(p,e,s,l) toFOLD_utf8_safe(p,e,s,l) |
| 2213 | #define toLOWER_utf8(p,e,s,l) toLOWER_utf8_safe(p,e,s,l) |
| 2214 | #define toTITLE_utf8(p,e,s,l) toTITLE_utf8_safe(p,e,s,l) |
| 2215 | #define toUPPER_utf8(p,e,s,l) toUPPER_utf8_safe(p,e,s,l) |
| 2216 | |
| 2217 | /* For internal core use only, subject to change */ |
| 2218 | #define _toFOLD_utf8_flags(p,e,s,l,f) _to_utf8_fold_flags (p,e,s,l,f) |
| 2219 | #define _toLOWER_utf8_flags(p,e,s,l,f) _to_utf8_lower_flags(p,e,s,l,f) |
| 2220 | #define _toTITLE_utf8_flags(p,e,s,l,f) _to_utf8_title_flags(p,e,s,l,f) |
| 2221 | #define _toUPPER_utf8_flags(p,e,s,l,f) _to_utf8_upper_flags(p,e,s,l,f) |
| 2222 | |
| 2223 | #define toFOLD_utf8_safe(p,e,s,l) _toFOLD_utf8_flags(p,e,s,l, FOLD_FLAGS_FULL) |
| 2224 | #define toLOWER_utf8_safe(p,e,s,l) _toLOWER_utf8_flags(p,e,s,l, 0) |
| 2225 | #define toTITLE_utf8_safe(p,e,s,l) _toTITLE_utf8_flags(p,e,s,l, 0) |
| 2226 | #define toUPPER_utf8_safe(p,e,s,l) _toUPPER_utf8_flags(p,e,s,l, 0) |
| 2227 | |
| 2228 | #define isALPHA_LC_utf8(p, e) isALPHA_LC_utf8_safe(p, e) |
| 2229 | #define isALPHANUMERIC_LC_utf8(p, e) isALPHANUMERIC_LC_utf8_safe(p, e) |
| 2230 | #define isASCII_LC_utf8(p, e) isASCII_LC_utf8_safe(p, e) |
| 2231 | #define isBLANK_LC_utf8(p, e) isBLANK_LC_utf8_safe(p, e) |
| 2232 | #define isCNTRL_LC_utf8(p, e) isCNTRL_LC_utf8_safe(p, e) |
| 2233 | #define isDIGIT_LC_utf8(p, e) isDIGIT_LC_utf8_safe(p, e) |
| 2234 | #define isGRAPH_LC_utf8(p, e) isGRAPH_LC_utf8_safe(p, e) |
| 2235 | #define isIDCONT_LC_utf8(p, e) isIDCONT_LC_utf8_safe(p, e) |
| 2236 | #define isIDFIRST_LC_utf8(p, e) isIDFIRST_LC_utf8_safe(p, e) |
| 2237 | #define isLOWER_LC_utf8(p, e) isLOWER_LC_utf8_safe(p, e) |
| 2238 | #define isPRINT_LC_utf8(p, e) isPRINT_LC_utf8_safe(p, e) |
| 2239 | #define isPSXSPC_LC_utf8(p, e) isPSXSPC_LC_utf8_safe(p, e) |
| 2240 | #define isPUNCT_LC_utf8(p, e) isPUNCT_LC_utf8_safe(p, e) |
| 2241 | #define isSPACE_LC_utf8(p, e) isSPACE_LC_utf8_safe(p, e) |
| 2242 | #define isUPPER_LC_utf8(p, e) isUPPER_LC_utf8_safe(p, e) |
| 2243 | #define isWORDCHAR_LC_utf8(p, e) isWORDCHAR_LC_utf8_safe(p, e) |
| 2244 | #define isXDIGIT_LC_utf8(p, e) isXDIGIT_LC_utf8_safe(p, e) |
| 2245 | |
| 2246 | /* For internal core Perl use only: the base macros for defining macros like |
| 2247 | * isALPHA_LC_utf8_safe. These are like _generic_utf8, but if the first code |
| 2248 | * point in 'p' is within the 0-255 range, it uses locale rules from the |
| 2249 | * passed-in 'macro' parameter */ |
| 2250 | #define _generic_LC_utf8_safe(macro, p, e, above_latin1) \ |
| 2251 | (__ASSERT_(_utf8_safe_assert(p, e)) \ |
| 2252 | (UTF8_IS_INVARIANT(*(p))) \ |
| 2253 | ? macro(*(p)) \ |
| 2254 | : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \ |
| 2255 | ? ((LIKELY((e) - (p) > 1 && UTF8_IS_CONTINUATION(*((p)+1)))) \ |
| 2256 | ? macro(EIGHT_BIT_UTF8_TO_NATIVE(*(p), *((p)+1))) \ |
| 2257 | : (_force_out_malformed_utf8_message( \ |
| 2258 | (U8 *) (p), (U8 *) (e), 0, 1), 0)) \ |
| 2259 | : above_latin1)) |
| 2260 | |
| 2261 | #define _generic_LC_invlist_utf8_safe(macro, classnum, p, e) \ |
| 2262 | _generic_LC_utf8_safe(macro, p, e, \ |
| 2263 | _is_utf8_FOO(classnum, p, e)) |
| 2264 | |
| 2265 | #define _generic_LC_func_utf8_safe(macro, above_latin1, p, e) \ |
| 2266 | _generic_LC_utf8_safe(macro, p, e, above_latin1(p, e)) |
| 2267 | |
| 2268 | #define _generic_LC_non_invlist_utf8_safe(classnum, above_latin1, p, e) \ |
| 2269 | _generic_LC_utf8_safe(classnum, p, e, \ |
| 2270 | (UNLIKELY((e) - (p) < UTF8SKIP(p)) \ |
| 2271 | ? (_force_out_malformed_utf8_message( \ |
| 2272 | (U8 *) (p), (U8 *) (e), 0, 1), 0) \ |
| 2273 | : above_latin1(p))) |
| 2274 | |
| 2275 | #define isALPHANUMERIC_LC_utf8_safe(p, e) \ |
| 2276 | _generic_LC_invlist_utf8_safe(isALPHANUMERIC_LC, \ |
| 2277 | _CC_ALPHANUMERIC, p, e) |
| 2278 | #define isALPHA_LC_utf8_safe(p, e) \ |
| 2279 | _generic_LC_invlist_utf8_safe(isALPHA_LC, _CC_ALPHA, p, e) |
| 2280 | #define isASCII_LC_utf8_safe(p, e) \ |
| 2281 | (__ASSERT_(_utf8_safe_assert(p, e)) isASCII_LC(*(p))) |
| 2282 | #define isBLANK_LC_utf8_safe(p, e) \ |
| 2283 | _generic_LC_non_invlist_utf8_safe(isBLANK_LC, is_HORIZWS_high, p, e) |
| 2284 | #define isCNTRL_LC_utf8_safe(p, e) \ |
| 2285 | _generic_LC_utf8_safe(isCNTRL_LC, p, e, 0) |
| 2286 | #define isDIGIT_LC_utf8_safe(p, e) \ |
| 2287 | _generic_LC_invlist_utf8_safe(isDIGIT_LC, _CC_DIGIT, p, e) |
| 2288 | #define isGRAPH_LC_utf8_safe(p, e) \ |
| 2289 | _generic_LC_invlist_utf8_safe(isGRAPH_LC, _CC_GRAPH, p, e) |
| 2290 | #define isIDCONT_LC_utf8_safe(p, e) \ |
| 2291 | _generic_LC_func_utf8_safe(isIDCONT_LC, \ |
| 2292 | _is_utf8_perl_idcont, p, e) |
| 2293 | #define isIDFIRST_LC_utf8_safe(p, e) \ |
| 2294 | _generic_LC_func_utf8_safe(isIDFIRST_LC, \ |
| 2295 | _is_utf8_perl_idstart, p, e) |
| 2296 | #define isLOWER_LC_utf8_safe(p, e) \ |
| 2297 | _generic_LC_invlist_utf8_safe(isLOWER_LC, _CC_LOWER, p, e) |
| 2298 | #define isPRINT_LC_utf8_safe(p, e) \ |
| 2299 | _generic_LC_invlist_utf8_safe(isPRINT_LC, _CC_PRINT, p, e) |
| 2300 | #define isPSXSPC_LC_utf8_safe(p, e) isSPACE_LC_utf8_safe(p, e) |
| 2301 | #define isPUNCT_LC_utf8_safe(p, e) \ |
| 2302 | _generic_LC_invlist_utf8_safe(isPUNCT_LC, _CC_PUNCT, p, e) |
| 2303 | #define isSPACE_LC_utf8_safe(p, e) \ |
| 2304 | _generic_LC_non_invlist_utf8_safe(isSPACE_LC, is_XPERLSPACE_high, p, e) |
| 2305 | #define isUPPER_LC_utf8_safe(p, e) \ |
| 2306 | _generic_LC_invlist_utf8_safe(isUPPER_LC, _CC_UPPER, p, e) |
| 2307 | #define isWORDCHAR_LC_utf8_safe(p, e) \ |
| 2308 | _generic_LC_invlist_utf8_safe(isWORDCHAR_LC, _CC_WORDCHAR, p, e) |
| 2309 | #define isXDIGIT_LC_utf8_safe(p, e) \ |
| 2310 | _generic_LC_non_invlist_utf8_safe(isXDIGIT_LC, is_XDIGIT_high, p, e) |
| 2311 | |
| 2312 | /* Macros for backwards compatibility and for completeness when the ASCII and |
| 2313 | * Latin1 values are identical */ |
| 2314 | #define isALPHAU(c) isALPHA_L1(c) |
| 2315 | #define isDIGIT_L1(c) isDIGIT_A(c) |
| 2316 | #define isOCTAL(c) isOCTAL_A(c) |
| 2317 | #define isOCTAL_L1(c) isOCTAL_A(c) |
| 2318 | #define isXDIGIT_L1(c) isXDIGIT_A(c) |
| 2319 | #define isALNUM(c) isWORDCHAR(c) |
| 2320 | #define isALNUM_A(c) isALNUM(c) |
| 2321 | #define isALNUMU(c) isWORDCHAR_L1(c) |
| 2322 | #define isALNUM_LC(c) isWORDCHAR_LC(c) |
| 2323 | #define isALNUM_uni(c) isWORDCHAR_uni(c) |
| 2324 | #define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c) |
| 2325 | #define isALNUM_utf8(p,e) isWORDCHAR_utf8(p,e) |
| 2326 | #define isALNUM_utf8_safe(p,e) isWORDCHAR_utf8_safe(p,e) |
| 2327 | #define isALNUM_LC_utf8(p,e)isWORDCHAR_LC_utf8(p,e) |
| 2328 | #define isALNUM_LC_utf8_safe(p,e)isWORDCHAR_LC_utf8_safe(p,e) |
| 2329 | #define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */ |
| 2330 | #define isALNUMC_L1(c) isALPHANUMERIC_L1(c) |
| 2331 | #define isALNUMC(c) isALPHANUMERIC(c) |
| 2332 | #define isALNUMC_LC(c) isALPHANUMERIC_LC(c) |
| 2333 | #define isALNUMC_uni(c) isALPHANUMERIC_uni(c) |
| 2334 | #define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c) |
| 2335 | #define isALNUMC_utf8(p,e) isALPHANUMERIC_utf8(p,e) |
| 2336 | #define isALNUMC_utf8_safe(p,e) isALPHANUMERIC_utf8_safe(p,e) |
| 2337 | #define isALNUMC_LC_utf8_safe(p,e) isALPHANUMERIC_LC_utf8_safe(p,e) |
| 2338 | |
| 2339 | /* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, |
| 2340 | * except that they don't necessarily mean the same characters, e.g. CTRL-D is |
| 2341 | * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC. |
| 2342 | * '?' is special-cased on EBCDIC to APC, which is the control there that is |
| 2343 | * the outlier from the block that contains the other controls, just like |
| 2344 | * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0 |
| 2345 | * block. If it weren't special cased, it would yield a non-control. |
| 2346 | * The conversion works both ways, so toCTRL('D') is 4, and toCTRL(4) is D, |
| 2347 | * etc. */ |
| 2348 | #ifndef EBCDIC |
| 2349 | # define toCTRL(c) (__ASSERT_(FITS_IN_8_BITS(c)) toUPPER(((U8)(c))) ^ 64) |
| 2350 | #else |
| 2351 | # define toCTRL(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ |
| 2352 | ((isPRINT_A(c)) \ |
| 2353 | ? (UNLIKELY((c) == '?') \ |
| 2354 | ? QUESTION_MARK_CTRL \ |
| 2355 | : (NATIVE_TO_LATIN1(toUPPER((U8) (c))) ^ 64)) \ |
| 2356 | : (UNLIKELY((c) == QUESTION_MARK_CTRL) \ |
| 2357 | ? '?' \ |
| 2358 | : (LATIN1_TO_NATIVE(((U8) (c)) ^ 64))))) |
| 2359 | #endif |
| 2360 | |
| 2361 | /* Line numbers are unsigned, 32 bits. */ |
| 2362 | typedef U32 line_t; |
| 2363 | #define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */ |
| 2364 | |
| 2365 | /* Helpful alias for version prescan */ |
| 2366 | #define is_LAX_VERSION(a,b) \ |
| 2367 | (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL)) |
| 2368 | |
| 2369 | #define is_STRICT_VERSION(a,b) \ |
| 2370 | (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL)) |
| 2371 | |
| 2372 | #define BADVERSION(a,b,c) \ |
| 2373 | if (b) { \ |
| 2374 | *b = c; \ |
| 2375 | } \ |
| 2376 | return a; |
| 2377 | |
| 2378 | /* Converts a character KNOWN to represent a hexadecimal digit (0-9, A-F, or |
| 2379 | * a-f) to its numeric value without using any branches. The input is |
| 2380 | * validated only by an assert() in DEBUGGING builds. |
| 2381 | * |
| 2382 | * It works by right shifting and isolating the bit that is 0 for the digits, |
| 2383 | * and 1 for at least the alphas A-F, a-f. The bit is shifted to the ones |
| 2384 | * position, and then to the eights position. Both are added together to form |
| 2385 | * 0 if the input is '0'-'9' and to form 9 if alpha. This is added to the |
| 2386 | * final four bits of the input to form the correct value. */ |
| 2387 | #define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) \ |
| 2388 | ((NATIVE_TO_LATIN1(c) >> 6) & 1) /* 1 if alpha; 0 if not */ \ |
| 2389 | + ((NATIVE_TO_LATIN1(c) >> 3) & 8) /* 8 if alpha; 0 if not */ \ |
| 2390 | + ((c) & 0xF)) /* 0-9 if input valid hex digit */ |
| 2391 | |
| 2392 | /* The argument is a string pointer, which is advanced. */ |
| 2393 | #define READ_XDIGIT(s) ((s)++, XDIGIT_VALUE(*((s) - 1))) |
| 2394 | |
| 2395 | /* Converts a character known to represent an octal digit (0-7) to its numeric |
| 2396 | * value. The input is validated only by an assert() in DEBUGGING builds. In |
| 2397 | * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */ |
| 2398 | #define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c))) |
| 2399 | |
| 2400 | /* Efficiently returns a boolean as to if two native characters are equivalent |
| 2401 | * case-insenstively. At least one of the characters must be one of [A-Za-z]; |
| 2402 | * the ALPHA in the name is to remind you of that. This is asserted() in |
| 2403 | * DEBUGGING builds. Because [A-Za-z] are invariant under UTF-8, this macro |
| 2404 | * works (on valid input) for both non- and UTF-8-encoded bytes. |
| 2405 | * |
| 2406 | * When one of the inputs is a compile-time constant and gets folded by the |
| 2407 | * compiler, this reduces to an AND and a TEST. On both EBCDIC and ASCII |
| 2408 | * machines, 'A' and 'a' differ by a single bit; the same with the upper and |
| 2409 | * lower case of all other ASCII-range alphabetics. On ASCII platforms, they |
| 2410 | * are 32 apart; on EBCDIC, they are 64. At compile time, this uses an |
| 2411 | * exclusive 'or' to find that bit and then inverts it to form a mask, with |
| 2412 | * just a single 0, in the bit position where the upper- and lowercase differ. |
| 2413 | * */ |
| 2414 | #define isALPHA_FOLD_EQ(c1, c2) \ |
| 2415 | (__ASSERT_(isALPHA_A(c1) || isALPHA_A(c2)) \ |
| 2416 | ((c1) & ~('A' ^ 'a')) == ((c2) & ~('A' ^ 'a'))) |
| 2417 | #define isALPHA_FOLD_NE(c1, c2) (! isALPHA_FOLD_EQ((c1), (c2))) |
| 2418 | |
| 2419 | /* |
| 2420 | =for apidoc_section Memory Management |
| 2421 | |
| 2422 | =for apidoc Am|void|Newx|void* ptr|int nitems|type |
| 2423 | The XSUB-writer's interface to the C C<malloc> function. |
| 2424 | |
| 2425 | Memory obtained by this should B<ONLY> be freed with L</"Safefree">. |
| 2426 | |
| 2427 | In 5.9.3, Newx() and friends replace the older New() API, and drops |
| 2428 | the first parameter, I<x>, a debug aid which allowed callers to identify |
| 2429 | themselves. This aid has been superseded by a new build option, |
| 2430 | PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still |
| 2431 | there for use in XS modules supporting older perls. |
| 2432 | |
| 2433 | =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast |
| 2434 | The XSUB-writer's interface to the C C<malloc> function, with |
| 2435 | cast. See also C<L</Newx>>. |
| 2436 | |
| 2437 | Memory obtained by this should B<ONLY> be freed with L</"Safefree">. |
| 2438 | |
| 2439 | =for apidoc Am|void|Newxz|void* ptr|int nitems|type |
| 2440 | The XSUB-writer's interface to the C C<malloc> function. The allocated |
| 2441 | memory is zeroed with C<memzero>. See also C<L</Newx>>. |
| 2442 | |
| 2443 | Memory obtained by this should B<ONLY> be freed with L</"Safefree">. |
| 2444 | |
| 2445 | =for apidoc Am|void|Renew|void* ptr|int nitems|type |
| 2446 | The XSUB-writer's interface to the C C<realloc> function. |
| 2447 | |
| 2448 | Memory obtained by this should B<ONLY> be freed with L</"Safefree">. |
| 2449 | |
| 2450 | =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast |
| 2451 | The XSUB-writer's interface to the C C<realloc> function, with |
| 2452 | cast. |
| 2453 | |
| 2454 | Memory obtained by this should B<ONLY> be freed with L</"Safefree">. |
| 2455 | |
| 2456 | =for apidoc Am|void|Safefree|void* ptr |
| 2457 | The XSUB-writer's interface to the C C<free> function. |
| 2458 | |
| 2459 | This should B<ONLY> be used on memory obtained using L</"Newx"> and friends. |
| 2460 | |
| 2461 | =for apidoc_section String Handling |
| 2462 | =for apidoc Am|void|Move|void* src|void* dest|int nitems|type |
| 2463 | The XSUB-writer's interface to the C C<memmove> function. The C<src> is the |
| 2464 | source, C<dest> is the destination, C<nitems> is the number of items, and |
| 2465 | C<type> is the type. Can do overlapping moves. See also C<L</Copy>>. |
| 2466 | |
| 2467 | =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type |
| 2468 | Like C<Move> but returns C<dest>. Useful |
| 2469 | for encouraging compilers to tail-call |
| 2470 | optimise. |
| 2471 | |
| 2472 | =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type |
| 2473 | The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the |
| 2474 | source, C<dest> is the destination, C<nitems> is the number of items, and |
| 2475 | C<type> is the type. May fail on overlapping copies. See also C<L</Move>>. |
| 2476 | |
| 2477 | =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type |
| 2478 | |
| 2479 | Like C<Copy> but returns C<dest>. Useful |
| 2480 | for encouraging compilers to tail-call |
| 2481 | optimise. |
| 2482 | |
| 2483 | =for apidoc Am|void|Zero|void* dest|int nitems|type |
| 2484 | |
| 2485 | The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the |
| 2486 | destination, C<nitems> is the number of items, and C<type> is the type. |
| 2487 | |
| 2488 | =for apidoc Am|void *|ZeroD|void* dest|int nitems|type |
| 2489 | |
| 2490 | Like C<Zero> but returns dest. Useful |
| 2491 | for encouraging compilers to tail-call |
| 2492 | optimise. |
| 2493 | |
| 2494 | =for apidoc_section Utility Functions |
| 2495 | =for apidoc Am|void|StructCopy|type *src|type *dest|type |
| 2496 | This is an architecture-independent macro to copy one structure to another. |
| 2497 | |
| 2498 | =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte |
| 2499 | |
| 2500 | Fill up memory with a byte pattern (a byte repeated over and over |
| 2501 | again) that hopefully catches attempts to access uninitialized memory. |
| 2502 | |
| 2503 | =for apidoc Am|void|PoisonNew|void* dest|int nitems|type |
| 2504 | |
| 2505 | PoisonWith(0xAB) for catching access to allocated but uninitialized memory. |
| 2506 | |
| 2507 | =for apidoc Am|void|PoisonFree|void* dest|int nitems|type |
| 2508 | |
| 2509 | PoisonWith(0xEF) for catching access to freed memory. |
| 2510 | |
| 2511 | =for apidoc Am|void|Poison|void* dest|int nitems|type |
| 2512 | |
| 2513 | PoisonWith(0xEF) for catching access to freed memory. |
| 2514 | |
| 2515 | =cut */ |
| 2516 | |
| 2517 | /* Maintained for backwards-compatibility only. Use newSV() instead. */ |
| 2518 | #ifndef PERL_CORE |
| 2519 | #define NEWSV(x,len) newSV(len) |
| 2520 | #endif |
| 2521 | |
| 2522 | #define MEM_SIZE_MAX ((MEM_SIZE)-1) |
| 2523 | |
| 2524 | #define _PERL_STRLEN_ROUNDUP_UNCHECKED(n) (((n) - 1 + PERL_STRLEN_ROUNDUP_QUANTUM) & ~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM - 1)) |
| 2525 | |
| 2526 | #ifdef PERL_MALLOC_WRAP |
| 2527 | |
| 2528 | /* This expression will be constant-folded at compile time. It checks |
| 2529 | * whether or not the type of the count n is so small (e.g. U8 or U16, or |
| 2530 | * U32 on 64-bit systems) that there's no way a wrap-around could occur. |
| 2531 | * As well as avoiding the need for a run-time check in some cases, it's |
| 2532 | * designed to avoid compiler warnings like: |
| 2533 | * comparison is always false due to limited range of data type |
| 2534 | * It's mathematically equivalent to |
| 2535 | * max(n) * sizeof(t) > MEM_SIZE_MAX |
| 2536 | */ |
| 2537 | |
| 2538 | # define _MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) \ |
| 2539 | ( sizeof(MEM_SIZE) < sizeof(n) \ |
| 2540 | || sizeof(t) > ((MEM_SIZE)1 << 8*(sizeof(MEM_SIZE) - sizeof(n)))) |
| 2541 | |
| 2542 | /* This is written in a slightly odd way to avoid various spurious |
| 2543 | * compiler warnings. We *want* to write the expression as |
| 2544 | * _MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) && (n > C) |
| 2545 | * (for some compile-time constant C), but even when the LHS |
| 2546 | * constant-folds to false at compile-time, g++ insists on emitting |
| 2547 | * warnings about the RHS (e.g. "comparison is always false"), so instead |
| 2548 | * we write it as |
| 2549 | * |
| 2550 | * (cond ? n : X) > C |
| 2551 | * |
| 2552 | * where X is a constant with X > C always false. Choosing a value for X |
| 2553 | * is tricky. If 0, some compilers will complain about 0 > C always being |
| 2554 | * false; if 1, Coverity complains when n happens to be the constant value |
| 2555 | * '1', that cond ? 1 : 1 has the same value on both branches; so use C |
| 2556 | * for X and hope that nothing else whines. |
| 2557 | */ |
| 2558 | |
| 2559 | # define _MEM_WRAP_WILL_WRAP(n,t) \ |
| 2560 | ((_MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) ? (MEM_SIZE)(n) : \ |
| 2561 | MEM_SIZE_MAX/sizeof(t)) > MEM_SIZE_MAX/sizeof(t)) |
| 2562 | |
| 2563 | # define MEM_WRAP_CHECK(n,t) \ |
| 2564 | (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \ |
| 2565 | && (croak_memory_wrap(),0)) |
| 2566 | |
| 2567 | # define MEM_WRAP_CHECK_1(n,t,a) \ |
| 2568 | (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \ |
| 2569 | && (Perl_croak_nocontext("%s",(a)),0)) |
| 2570 | |
| 2571 | /* "a" arg must be a string literal */ |
| 2572 | # define MEM_WRAP_CHECK_s(n,t,a) \ |
| 2573 | (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \ |
| 2574 | && (Perl_croak_nocontext("" a ""),0)) |
| 2575 | |
| 2576 | #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t), |
| 2577 | |
| 2578 | #define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (croak_memory_wrap(),0) : 0), _PERL_STRLEN_ROUNDUP_UNCHECKED(n)) |
| 2579 | #else |
| 2580 | |
| 2581 | #define MEM_WRAP_CHECK(n,t) |
| 2582 | #define MEM_WRAP_CHECK_1(n,t,a) |
| 2583 | #define MEM_WRAP_CHECK_s(n,t,a) |
| 2584 | #define MEM_WRAP_CHECK_(n,t) |
| 2585 | |
| 2586 | #define PERL_STRLEN_ROUNDUP(n) _PERL_STRLEN_ROUNDUP_UNCHECKED(n) |
| 2587 | |
| 2588 | #endif |
| 2589 | |
| 2590 | #ifdef PERL_MEM_LOG |
| 2591 | /* |
| 2592 | * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s |
| 2593 | * go through functions, which are handy for debugging breakpoints, but |
| 2594 | * which more importantly get the immediate calling environment (file and |
| 2595 | * line number, and C function name if available) passed in. This info can |
| 2596 | * then be used for logging the calls, for which one gets a sample |
| 2597 | * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined. |
| 2598 | * |
| 2599 | * Known problems: |
| 2600 | * - not all memory allocs get logged, only those |
| 2601 | * that go through Newx() and derivatives (while all |
| 2602 | * Safefrees do get logged) |
| 2603 | * - __FILE__ and __LINE__ do not work everywhere |
| 2604 | * - __func__ or __FUNCTION__ even less so |
| 2605 | * - I think more goes on after the perlio frees but |
| 2606 | * the thing is that STDERR gets closed (as do all |
| 2607 | * the file descriptors) |
| 2608 | * - no deeper calling stack than the caller of the Newx() |
| 2609 | * or the kind, but do I look like a C reflection/introspection |
| 2610 | * utility to you? |
| 2611 | * - the function prototypes for the logging functions |
| 2612 | * probably should maybe be somewhere else than handy.h |
| 2613 | * - one could consider inlining (macrofying) the logging |
| 2614 | * for speed, but I am too lazy |
| 2615 | * - one could imagine recording the allocations in a hash, |
| 2616 | * (keyed by the allocation address?), and maintain that |
| 2617 | * through reallocs and frees, but how to do that without |
| 2618 | * any News() happening...? |
| 2619 | * - lots of -Ddefines to get useful/controllable output |
| 2620 | * - lots of ENV reads |
| 2621 | */ |
| 2622 | |
| 2623 | # ifdef PERL_CORE |
| 2624 | # ifndef PERL_MEM_LOG_NOIMPL |
| 2625 | enum mem_log_type { |
| 2626 | MLT_ALLOC, |
| 2627 | MLT_REALLOC, |
| 2628 | MLT_FREE, |
| 2629 | MLT_NEW_SV, |
| 2630 | MLT_DEL_SV |
| 2631 | }; |
| 2632 | # endif |
| 2633 | # if defined(PERL_IN_SV_C) /* those are only used in sv.c */ |
| 2634 | void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); |
| 2635 | void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); |
| 2636 | # endif |
| 2637 | # endif |
| 2638 | |
| 2639 | #endif |
| 2640 | |
| 2641 | #ifdef PERL_MEM_LOG |
| 2642 | #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__) |
| 2643 | #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__) |
| 2644 | #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__) |
| 2645 | #endif |
| 2646 | |
| 2647 | #ifndef MEM_LOG_ALLOC |
| 2648 | #define MEM_LOG_ALLOC(n,t,a) (a) |
| 2649 | #endif |
| 2650 | #ifndef MEM_LOG_REALLOC |
| 2651 | #define MEM_LOG_REALLOC(n,t,v,a) (a) |
| 2652 | #endif |
| 2653 | #ifndef MEM_LOG_FREE |
| 2654 | #define MEM_LOG_FREE(a) (a) |
| 2655 | #endif |
| 2656 | |
| 2657 | #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) |
| 2658 | #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) |
| 2659 | #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t))))) |
| 2660 | |
| 2661 | #ifndef PERL_CORE |
| 2662 | /* pre 5.9.x compatibility */ |
| 2663 | #define New(x,v,n,t) Newx(v,n,t) |
| 2664 | #define Newc(x,v,n,t,c) Newxc(v,n,t,c) |
| 2665 | #define Newz(x,v,n,t) Newxz(v,n,t) |
| 2666 | #endif |
| 2667 | |
| 2668 | #define Renew(v,n,t) \ |
| 2669 | (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) |
| 2670 | #define Renewc(v,n,t,c) \ |
| 2671 | (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) |
| 2672 | |
| 2673 | #ifdef PERL_POISON |
| 2674 | #define Safefree(d) \ |
| 2675 | ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0) |
| 2676 | #else |
| 2677 | #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d))) |
| 2678 | #endif |
| 2679 | |
| 2680 | /* assert that a valid ptr has been supplied - use this instead of assert(ptr) * |
| 2681 | * as it handles cases like constant string arguments without throwing warnings * |
| 2682 | * the cast is required, as is the inequality check, to avoid warnings */ |
| 2683 | #define perl_assert_ptr(p) assert( ((void*)(p)) != 0 ) |
| 2684 | |
| 2685 | |
| 2686 | #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) |
| 2687 | #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) |
| 2688 | #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), (void)memzero((char*)(d), (n) * sizeof(t))) |
| 2689 | |
| 2690 | /* Like above, but returns a pointer to 'd' */ |
| 2691 | #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) |
| 2692 | #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) |
| 2693 | #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), memzero((char*)(d), (n) * sizeof(t))) |
| 2694 | |
| 2695 | #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t))) |
| 2696 | #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB) |
| 2697 | #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF) |
| 2698 | #define Poison(d,n,t) PoisonFree(d,n,t) |
| 2699 | |
| 2700 | #ifdef PERL_POISON |
| 2701 | # define PERL_POISON_EXPR(x) x |
| 2702 | #else |
| 2703 | # define PERL_POISON_EXPR(x) |
| 2704 | #endif |
| 2705 | |
| 2706 | #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s))) |
| 2707 | |
| 2708 | /* |
| 2709 | =for apidoc_section Utility Functions |
| 2710 | |
| 2711 | =for apidoc Am|STRLEN|C_ARRAY_LENGTH|void *a |
| 2712 | |
| 2713 | Returns the number of elements in the input C array (so you want your |
| 2714 | zero-based indices to be less than but not equal to). |
| 2715 | |
| 2716 | =for apidoc Am|void *|C_ARRAY_END|void *a |
| 2717 | |
| 2718 | Returns a pointer to one element past the final element of the input C array. |
| 2719 | |
| 2720 | =cut |
| 2721 | |
| 2722 | C_ARRAY_END is one past the last: half-open/half-closed range, not |
| 2723 | last-inclusive range. |
| 2724 | */ |
| 2725 | #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0])) |
| 2726 | #define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a)) |
| 2727 | |
| 2728 | #ifdef NEED_VA_COPY |
| 2729 | # ifdef va_copy |
| 2730 | # define Perl_va_copy(s, d) va_copy(d, s) |
| 2731 | # elif defined(__va_copy) |
| 2732 | # define Perl_va_copy(s, d) __va_copy(d, s) |
| 2733 | # else |
| 2734 | # define Perl_va_copy(s, d) Copy(s, d, 1, va_list) |
| 2735 | # endif |
| 2736 | #endif |
| 2737 | |
| 2738 | /* convenience debug macros */ |
| 2739 | #ifdef USE_ITHREADS |
| 2740 | #define pTHX_FORMAT "Perl interpreter: 0x%p" |
| 2741 | #define pTHX__FORMAT ", Perl interpreter: 0x%p" |
| 2742 | #define pTHX_VALUE_ (void *)my_perl, |
| 2743 | #define pTHX_VALUE (void *)my_perl |
| 2744 | #define pTHX__VALUE_ ,(void *)my_perl, |
| 2745 | #define pTHX__VALUE ,(void *)my_perl |
| 2746 | #else |
| 2747 | #define pTHX_FORMAT |
| 2748 | #define pTHX__FORMAT |
| 2749 | #define pTHX_VALUE_ |
| 2750 | #define pTHX_VALUE |
| 2751 | #define pTHX__VALUE_ |
| 2752 | #define pTHX__VALUE |
| 2753 | #endif /* USE_ITHREADS */ |
| 2754 | |
| 2755 | /* Perl_deprecate was not part of the public API, and did not have a deprecate() |
| 2756 | shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor |
| 2757 | CPAN::Unpack show any users outside the core. */ |
| 2758 | #ifdef PERL_CORE |
| 2759 | # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \ |
| 2760 | "Use of " s " is deprecated") |
| 2761 | # define deprecate_disappears_in(when,message) \ |
| 2762 | Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \ |
| 2763 | message ", and will disappear in Perl " when) |
| 2764 | # define deprecate_fatal_in(when,message) \ |
| 2765 | Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \ |
| 2766 | message ". Its use will be fatal in Perl " when) |
| 2767 | #endif |
| 2768 | |
| 2769 | /* Internal macros to deal with gids and uids */ |
| 2770 | #ifdef PERL_CORE |
| 2771 | |
| 2772 | # if Uid_t_size > IVSIZE |
| 2773 | # define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid)) |
| 2774 | # define SvUID(sv) SvNV(sv) |
| 2775 | # elif Uid_t_sign <= 0 |
| 2776 | # define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid)) |
| 2777 | # define SvUID(sv) SvIV(sv) |
| 2778 | # else |
| 2779 | # define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid)) |
| 2780 | # define SvUID(sv) SvUV(sv) |
| 2781 | # endif /* Uid_t_size */ |
| 2782 | |
| 2783 | # if Gid_t_size > IVSIZE |
| 2784 | # define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid)) |
| 2785 | # define SvGID(sv) SvNV(sv) |
| 2786 | # elif Gid_t_sign <= 0 |
| 2787 | # define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid)) |
| 2788 | # define SvGID(sv) SvIV(sv) |
| 2789 | # else |
| 2790 | # define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid)) |
| 2791 | # define SvGID(sv) SvUV(sv) |
| 2792 | # endif /* Gid_t_size */ |
| 2793 | |
| 2794 | #endif |
| 2795 | |
| 2796 | #endif /* PERL_HANDY_H_ */ |
| 2797 | |
| 2798 | /* |
| 2799 | * ex: set ts=8 sts=4 sw=4 et: |
| 2800 | */ |