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 HANDY_H /* Guard against nested #inclusion */ 15 #define HANDY_H 16 17 #if !defined(__STDC__) 18 #ifdef NULL 19 #undef NULL 20 #endif 21 # define NULL 0 22 #endif 23 24 #ifndef PERL_CORE 25 # define Null(type) ((type)NULL) 26 27 /* 28 =head1 Handy Values 29 30 =for apidoc AmU||Nullch 31 Null character pointer. (No longer available when C<PERL_CORE> is 32 defined.) 33 34 =for apidoc AmU||Nullsv 35 Null SV pointer. (No longer available when C<PERL_CORE> is defined.) 36 37 =cut 38 */ 39 40 # define Nullch Null(char*) 41 # define Nullfp Null(PerlIO*) 42 # define Nullsv Null(SV*) 43 #endif 44 45 #ifdef TRUE 46 #undef TRUE 47 #endif 48 #ifdef FALSE 49 #undef FALSE 50 #endif 51 #define TRUE (1) 52 #define FALSE (0) 53 54 /* The MUTABLE_*() macros cast pointers to the types shown, in such a way 55 * (compiler permitting) that casting away const-ness will give a warning; 56 * e.g.: 57 * 58 * const SV *sv = ...; 59 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away 60 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn 61 */ 62 63 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN) 64 # define MUTABLE_PTR(p) ({ void *_p = (p); _p; }) 65 #else 66 # define MUTABLE_PTR(p) ((void *) (p)) 67 #endif 68 69 #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p)) 70 #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p)) 71 #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p)) 72 #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p)) 73 #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p)) 74 #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p)) 75 76 #if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR) 77 # include <stdbool.h> 78 # ifndef HAS_BOOL 79 # define HAS_BOOL 1 80 # endif 81 #endif 82 83 /* bool is built-in for g++-2.6.3 and later, which might be used 84 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't 85 be sure _G_config.h will be included before this file. _G_config.h 86 also defines _G_HAVE_BOOL for both gcc and g++, but only g++ 87 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us. 88 g++ can be identified by __GNUG__. 89 Andy Dougherty February 2000 90 */ 91 #ifdef __GNUG__ /* GNU g++ has bool built-in */ 92 # ifndef PERL_BOOL_AS_CHAR 93 # ifndef HAS_BOOL 94 # define HAS_BOOL 1 95 # endif 96 # endif 97 #endif 98 99 /* The NeXT dynamic loader headers will not build with the bool macro 100 So declare them now to clear confusion. 101 */ 102 #if defined(NeXT) || defined(__NeXT__) 103 # undef FALSE 104 # undef TRUE 105 typedef enum bool { FALSE = 0, TRUE = 1 } bool; 106 # define ENUM_BOOL 1 107 # ifndef HAS_BOOL 108 # define HAS_BOOL 1 109 # endif /* !HAS_BOOL */ 110 #endif /* NeXT || __NeXT__ */ 111 112 #ifndef HAS_BOOL 113 # ifdef bool 114 # undef bool 115 # endif 116 # define bool char 117 # define HAS_BOOL 1 118 #endif 119 120 /* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is 121 * defined as char for example, then the cast from int is 122 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on 123 * AIX */ 124 #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0) 125 126 /* Try to figure out __func__ or __FUNCTION__ equivalent, if any. 127 * XXX Should really be a Configure probe, with HAS__FUNCTION__ 128 * and FUNCTION__ as results. 129 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */ 130 #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */ 131 # define FUNCTION__ __func__ 132 #else 133 # if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \ 134 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */ 135 # define FUNCTION__ "" 136 # else 137 # define FUNCTION__ __FUNCTION__ /* Common extension. */ 138 # endif 139 #endif 140 141 /* XXX A note on the perl source internal type system. The 142 original intent was that I32 be *exactly* 32 bits. 143 144 Currently, we only guarantee that I32 is *at least* 32 bits. 145 Specifically, if int is 64 bits, then so is I32. (This is the case 146 for the Cray.) This has the advantage of meshing nicely with 147 standard library calls (where we pass an I32 and the library is 148 expecting an int), but the disadvantage that an I32 is not 32 bits. 149 Andy Dougherty August 1996 150 151 There is no guarantee that there is *any* integral type with 152 exactly 32 bits. It is perfectly legal for a system to have 153 sizeof(short) == sizeof(int) == sizeof(long) == 8. 154 155 Similarly, there is no guarantee that I16 and U16 have exactly 16 156 bits. 157 158 For dealing with issues that may arise from various 32/64-bit 159 systems, we will ask Configure to check out 160 161 SHORTSIZE == sizeof(short) 162 INTSIZE == sizeof(int) 163 LONGSIZE == sizeof(long) 164 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG) 165 PTRSIZE == sizeof(void *) 166 DOUBLESIZE == sizeof(double) 167 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE). 168 169 */ 170 171 #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */ 172 # include <inttypes.h> 173 # ifdef INT32_MIN_BROKEN 174 # undef INT32_MIN 175 # define INT32_MIN (-2147483647-1) 176 # endif 177 # ifdef INT64_MIN_BROKEN 178 # undef INT64_MIN 179 # define INT64_MIN (-9223372036854775807LL-1) 180 # endif 181 #endif 182 183 typedef I8TYPE I8; 184 typedef U8TYPE U8; 185 typedef I16TYPE I16; 186 typedef U16TYPE U16; 187 typedef I32TYPE I32; 188 typedef U32TYPE U32; 189 #ifdef PERL_CORE 190 # ifdef HAS_QUAD 191 typedef I64TYPE I64; 192 typedef U64TYPE U64; 193 # endif 194 #endif /* PERL_CORE */ 195 196 #if defined(HAS_QUAD) && defined(USE_64_BIT_INT) 197 # if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG 198 # define PeRl_INT64_C(c) CAT2(c,LL) 199 # define PeRl_UINT64_C(c) CAT2(c,ULL) 200 # else 201 # if QUADKIND == QUAD_IS___INT64 202 # define PeRl_INT64_C(c) CAT2(c,I64) 203 # define PeRl_UINT64_C(c) CAT2(c,UI64) 204 # else 205 # if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG 206 # define PeRl_INT64_C(c) CAT2(c,L) 207 # define PeRl_UINT64_C(c) CAT2(c,UL) 208 # else 209 # define PeRl_INT64_C(c) ((I64TYPE)(c)) 210 # define PeRl_UINT64_C(c) ((U64TYPE)(c)) 211 # endif 212 # endif 213 # endif 214 # ifndef UINT64_C 215 # define UINT64_C(c) PeRl_UINT64_C(c) 216 # endif 217 # ifndef INT64_C 218 # define INT64_C(c) PeRl_INT64_C(c) 219 # endif 220 #endif 221 222 #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX) 223 224 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type. 225 Please search CHAR_MAX in perl.h for further details. */ 226 #define U8_MAX UINT8_MAX 227 #define U8_MIN UINT8_MIN 228 229 #define I16_MAX INT16_MAX 230 #define I16_MIN INT16_MIN 231 #define U16_MAX UINT16_MAX 232 #define U16_MIN UINT16_MIN 233 234 #define I32_MAX INT32_MAX 235 #define I32_MIN INT32_MIN 236 #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */ 237 # define U32_MAX UINT32_MAX 238 #else 239 # define U32_MAX 4294967295U 240 #endif 241 #define U32_MIN UINT32_MIN 242 243 #else 244 245 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type. 246 Please search CHAR_MAX in perl.h for further details. */ 247 #define U8_MAX PERL_UCHAR_MAX 248 #define U8_MIN PERL_UCHAR_MIN 249 250 #define I16_MAX PERL_SHORT_MAX 251 #define I16_MIN PERL_SHORT_MIN 252 #define U16_MAX PERL_USHORT_MAX 253 #define U16_MIN PERL_USHORT_MIN 254 255 #if LONGSIZE > 4 256 # define I32_MAX PERL_INT_MAX 257 # define I32_MIN PERL_INT_MIN 258 # define U32_MAX PERL_UINT_MAX 259 # define U32_MIN PERL_UINT_MIN 260 #else 261 # define I32_MAX PERL_LONG_MAX 262 # define I32_MIN PERL_LONG_MIN 263 # define U32_MAX PERL_ULONG_MAX 264 # define U32_MIN PERL_ULONG_MIN 265 #endif 266 267 #endif 268 269 /* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */ 270 #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */ 271 #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8) 272 #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */ 273 274 #define Ctl(ch) ((ch) & 037) 275 276 /* This is a helper macro to avoid preprocessor issues, replaced by nothing 277 * unless under DEBUGGING, where it expands to an assert of its argument, 278 * followed by a comma (hence the comma operator). If we just used a straight 279 * assert(), we would get a comma with nothing before it when not DEBUGGING */ 280 #ifdef DEBUGGING 281 # define __ASSERT_(statement) assert(statement), 282 #else 283 # define __ASSERT_(statement) 284 #endif 285 286 /* 287 =head1 SV-Body Allocation 288 289 =for apidoc Ama|SV*|newSVpvs|const char* s 290 Like C<newSVpvn>, but takes a literal C<NUL>-terminated string instead of a 291 string/length pair. 292 293 =for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags 294 Like C<newSVpvn_flags>, but takes a literal C<NUL>-terminated string instead of 295 a string/length pair. 296 297 =for apidoc Ama|SV*|newSVpvs_share|const char* s 298 Like C<newSVpvn_share>, but takes a literal C<NUL>-terminated string instead of 299 a string/length pair and omits the hash parameter. 300 301 =for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags 302 Like C<sv_catpvn_flags>, but takes a literal C<NUL>-terminated string instead 303 of a string/length pair. 304 305 =for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s 306 Like C<sv_catpvn_nomg>, but takes a literal string instead of a 307 string/length pair. 308 309 =for apidoc Am|void|sv_catpvs|SV* sv|const char* s 310 Like C<sv_catpvn>, but takes a literal string instead of a string/length pair. 311 312 =for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s 313 Like C<sv_catpvn_mg>, but takes a literal string instead of a 314 string/length pair. 315 316 =for apidoc Am|void|sv_setpvs|SV* sv|const char* s 317 Like C<sv_setpvn>, but takes a literal string instead of a string/length pair. 318 319 =for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s 320 Like C<sv_setpvn_mg>, but takes a literal string instead of a 321 string/length pair. 322 323 =for apidoc Am|SV *|sv_setref_pvs|const char* s 324 Like C<sv_setref_pvn>, but takes a literal string instead of a 325 string/length pair. 326 327 =head1 Memory Management 328 329 =for apidoc Ama|char*|savepvs|const char* s 330 Like C<savepvn>, but takes a literal C<NUL>-terminated string instead of a 331 string/length pair. 332 333 =for apidoc Ama|char*|savesharedpvs|const char* s 334 A version of C<savepvs()> which allocates the duplicate string in memory 335 which is shared between threads. 336 337 =head1 GV Functions 338 339 =for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create 340 Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair. 341 342 =head1 Hash Manipulation Functions 343 344 =for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval 345 Like C<hv_fetch>, but takes a literal string instead of a string/length pair. 346 347 =for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val 348 Like C<hv_store>, but takes a literal string instead of a string/length pair 349 and omits the hash parameter. 350 351 =head1 Lexer interface 352 353 =for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags 354 355 Like L</lex_stuff_pvn>, but takes a literal string instead of a 356 string/length pair. 357 358 =cut 359 */ 360 361 /* concatenating with "" ensures that only literal strings are accepted as 362 * argument */ 363 #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1) 364 365 /* note that STR_WITH_LEN() can't be used as argument to macros or functions 366 * that under some configurations might be macros, which means that it requires 367 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used. 368 */ 369 370 /* STR_WITH_LEN() shortcuts */ 371 #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str)) 372 #define newSVpvs_flags(str,flags) \ 373 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags) 374 #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0) 375 #define sv_catpvs_flags(sv, str, flags) \ 376 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags) 377 #define sv_catpvs_nomg(sv, str) \ 378 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0) 379 #define sv_catpvs(sv, str) \ 380 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC) 381 #define sv_catpvs_mg(sv, str) \ 382 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC) 383 #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str)) 384 #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str)) 385 #define sv_setref_pvs(rv, classname, str) \ 386 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str)) 387 #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str)) 388 #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str)) 389 #define gv_stashpvs(str, create) \ 390 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create) 391 #define gv_fetchpvs(namebeg, add, sv_type) \ 392 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type) 393 #define gv_fetchpvn(namebeg, len, add, sv_type) \ 394 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type) 395 #define sv_catxmlpvs(dsv, str, utf8) \ 396 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8) 397 #define hv_fetchs(hv,key,lval) \ 398 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \ 399 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \ 400 : HV_FETCH_JUST_SV, NULL, 0)) 401 402 #define hv_stores(hv,key,val) \ 403 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \ 404 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0)) 405 406 #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags) 407 408 #define get_cvs(str, flags) \ 409 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags)) 410 411 /* 412 =head1 Miscellaneous Functions 413 414 =for apidoc Am|bool|strNE|char* s1|char* s2 415 Test two strings to see if they are different. Returns true or 416 false. 417 418 =for apidoc Am|bool|strEQ|char* s1|char* s2 419 Test two strings to see if they are equal. Returns true or false. 420 421 =for apidoc Am|bool|strLT|char* s1|char* s2 422 Test two strings to see if the first, C<s1>, is less than the second, 423 C<s2>. Returns true or false. 424 425 =for apidoc Am|bool|strLE|char* s1|char* s2 426 Test two strings to see if the first, C<s1>, is less than or equal to the 427 second, C<s2>. Returns true or false. 428 429 =for apidoc Am|bool|strGT|char* s1|char* s2 430 Test two strings to see if the first, C<s1>, is greater than the second, 431 C<s2>. Returns true or false. 432 433 =for apidoc Am|bool|strGE|char* s1|char* s2 434 Test two strings to see if the first, C<s1>, is greater than or equal to 435 the second, C<s2>. Returns true or false. 436 437 =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len 438 Test two strings to see if they are different. The C<len> parameter 439 indicates the number of bytes to compare. Returns true or false. (A 440 wrapper for C<strncmp>). 441 442 =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len 443 Test two strings to see if they are equal. The C<len> parameter indicates 444 the number of bytes to compare. Returns true or false. (A wrapper for 445 C<strncmp>). 446 447 =cut 448 */ 449 450 #define strNE(s1,s2) (strcmp(s1,s2)) 451 #define strEQ(s1,s2) (!strcmp(s1,s2)) 452 #define strLT(s1,s2) (strcmp(s1,s2) < 0) 453 #define strLE(s1,s2) (strcmp(s1,s2) <= 0) 454 #define strGT(s1,s2) (strcmp(s1,s2) > 0) 455 #define strGE(s1,s2) (strcmp(s1,s2) >= 0) 456 #define strnNE(s1,s2,l) (strncmp(s1,s2,l)) 457 #define strnEQ(s1,s2,l) (!strncmp(s1,s2,l)) 458 459 #ifdef HAS_MEMCMP 460 # define memNE(s1,s2,l) (memcmp(s1,s2,l)) 461 # define memEQ(s1,s2,l) (!memcmp(s1,s2,l)) 462 #else 463 # define memNE(s1,s2,l) (bcmp(s1,s2,l)) 464 # define memEQ(s1,s2,l) (!bcmp(s1,s2,l)) 465 #endif 466 467 #define memEQs(s1, l, s2) \ 468 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1))) 469 #define memNEs(s1, l, s2) !memEQs(s1, l, s2) 470 471 /* 472 * Character classes. 473 * 474 * Unfortunately, the introduction of locales means that we 475 * can't trust isupper(), etc. to tell the truth. And when 476 * it comes to /\w+/ with tainting enabled, we *must* be able 477 * to trust our character classes. 478 * 479 * Therefore, the default tests in the text of Perl will be 480 * independent of locale. Any code that wants to depend on 481 * the current locale will use the tests that begin with "lc". 482 */ 483 484 #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */ 485 # ifndef CTYPE256 486 # define CTYPE256 487 # endif 488 #endif 489 490 /* 491 492 =head1 Character classes 493 This section is about functions (really macros) that classify characters 494 into types, such as punctuation versus alphabetic, etc. Most of these are 495 analogous to regular expression character classes. (See 496 L<perlrecharclass/POSIX Character Classes>.) There are several variants for 497 each class. (Not all macros have all variants; each item below lists the 498 ones valid for it.) None are affected by C<use bytes>, and only the ones 499 with C<LC> in the name are affected by the current locale. 500 501 The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a 502 C<U8>) as input and returns a boolean as to whether or not the character 503 represented by that octet is (or on non-ASCII platforms, corresponds to) an 504 ASCII character in the named class based on platform, Unicode, and Perl rules. 505 If the input is a number that doesn't fit in an octet, FALSE is returned. 506 507 Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function 508 with no suffix C<"_A">. 509 510 Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set 511 onto the platform. That is, the code points that are ASCII are unaffected, 512 since ASCII is a subset of Latin-1. But the non-ASCII code points are treated 513 as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return 514 true when called with the code point 0xDF, which is a word character in both 515 ASCII and EBCDIC (though it represents different characters in each). 516 517 Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code 518 point as input. If the code point is larger than 255, Unicode rules are used 519 to determine if it is in the character class. For example, 520 C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A 521 WITH MACRON in Unicode, and is a word character. 522 523 Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a 524 (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The 525 classification of just the first (possibly multi-byte) character in the string 526 is tested. 527 528 Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but the 529 result is based on the current locale, which is what C<LC> in the name stands 530 for. If Perl can determine that the current locale is a UTF-8 locale, it uses 531 the published Unicode rules; otherwise, it uses the C library function that 532 gives the named classification. For example, C<isDIGIT_LC()> when not in a 533 UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always 534 returned if the input won't fit into an octet. 535 536 Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It 537 returns the same as C<isFOO_LC> for input code points less than 256, and 538 returns the hard-coded, not-affected-by-locale, Unicode results for larger ones. 539 540 Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a 541 (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The 542 classification of just the first (possibly multi-byte) character in the string 543 is tested. 544 545 =for apidoc Am|bool|isALPHA|char ch 546 Returns a boolean indicating whether the specified character is an 547 alphabetic character, analogous to C<m/[[:alpha:]]/>. 548 See the L<top of this section|/Character classes> for an explanation of variants 549 C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>, 550 C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>. 551 552 =for apidoc Am|bool|isALPHANUMERIC|char ch 553 Returns a boolean indicating whether the specified character is a either an 554 alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>. 555 See the L<top of this section|/Character classes> for an explanation of variants 556 C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>, 557 C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and 558 C<isALPHANUMERIC_LC_utf8>. 559 560 =for apidoc Am|bool|isASCII|char ch 561 Returns a boolean indicating whether the specified character is one of the 128 562 characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>. 563 On non-ASCII platforms, it returns TRUE iff this 564 character corresponds to an ASCII character. Variants C<isASCII_A()> and 565 C<isASCII_L1()> are identical to C<isASCII()>. 566 See the L<top of this section|/Character classes> for an explanation of variants 567 C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and 568 C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C 569 library routine C<isascii()>. In these cases, the variants whose names contain 570 C<LC> are the same as the corresponding ones without. 571 572 Also note, that because all ASCII characters are UTF-8 invariant (meaning they 573 have the exact same representation (always a single byte) whether encoded in 574 UTF-8 or not), C<isASCII> will give the correct results when called with any 575 byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will 576 work properly on any string encoded or not in UTF-8. 577 578 =for apidoc Am|bool|isBLANK|char ch 579 Returns a boolean indicating whether the specified character is a 580 character considered to be a blank, analogous to C<m/[[:blank:]]/>. 581 See the L<top of this section|/Character classes> for an explanation of variants 582 C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>, 583 C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some 584 platforms do not have the C library routine C<isblank()>. In these cases, the 585 variants whose names contain C<LC> are the same as the corresponding ones 586 without. 587 588 =for apidoc Am|bool|isCNTRL|char ch 589 Returns a boolean indicating whether the specified character is a 590 control character, analogous to C<m/[[:cntrl:]]/>. 591 See the L<top of this section|/Character classes> for an explanation of variants 592 C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>, 593 C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8> 594 On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant. 595 596 =for apidoc Am|bool|isDIGIT|char ch 597 Returns a boolean indicating whether the specified character is a 598 digit, analogous to C<m/[[:digit:]]/>. 599 Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>. 600 See the L<top of this section|/Character classes> for an explanation of variants 601 C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and 602 C<isDIGIT_LC_utf8>. 603 604 =for apidoc Am|bool|isGRAPH|char ch 605 Returns a boolean indicating whether the specified character is a 606 graphic character, analogous to C<m/[[:graph:]]/>. 607 See the L<top of this section|/Character classes> for an explanation of variants 608 C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>, 609 C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>. 610 611 =for apidoc Am|bool|isLOWER|char ch 612 Returns a boolean indicating whether the specified character is a 613 lowercase character, analogous to C<m/[[:lower:]]/>. 614 See the L<top of this section|/Character classes> for an explanation of variants 615 C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>, 616 C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>. 617 618 =for apidoc Am|bool|isOCTAL|char ch 619 Returns a boolean indicating whether the specified character is an 620 octal digit, [0-7]. 621 The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to 622 C<isOCTAL>. 623 624 =for apidoc Am|bool|isPUNCT|char ch 625 Returns a boolean indicating whether the specified character is a 626 punctuation character, analogous to C<m/[[:punct:]]/>. 627 Note that the definition of what is punctuation isn't as 628 straightforward as one might desire. See L<perlrecharclass/POSIX Character 629 Classes> for details. 630 See the L<top of this section|/Character classes> for an explanation of variants 631 C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>, 632 C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>. 633 634 =for apidoc Am|bool|isSPACE|char ch 635 Returns a boolean indicating whether the specified character is a 636 whitespace character. This is analogous 637 to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18 638 (experimentally), this also matches what C<m/[[:space:]]/> does. 639 ("Experimentally" means that this change may be backed out in 5.22 if 640 field experience indicates that it was unwise.) Prior to 5.18, only the 641 locale forms of this macro (the ones with C<LC> in their names) matched 642 precisely what C<m/[[:space:]]/> does. In those releases, the only difference, 643 in the non-locale variants, was that C<isSPACE()> did not match a vertical tab. 644 (See L</isPSXSPC> for a macro that matches a vertical tab in all releases.) 645 See the L<top of this section|/Character classes> for an explanation of variants 646 C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>, 647 C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>. 648 649 =for apidoc Am|bool|isPSXSPC|char ch 650 (short for Posix Space) 651 Starting in 5.18, this is identical (experimentally) in all its forms to the 652 corresponding C<isSPACE()> macros. ("Experimentally" means that this change 653 may be backed out in 5.22 if field experience indicates that it 654 was unwise.) 655 The locale forms of this macro are identical to their corresponding 656 C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the 657 non-locale forms differ from their C<isSPACE()> forms only in that the 658 C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do. 659 Otherwise they are identical. Thus this macro is analogous to what 660 C<m/[[:space:]]/> matches in a regular expression. 661 See the L<top of this section|/Character classes> for an explanation of variants 662 C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>, 663 C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>. 664 665 =for apidoc Am|bool|isUPPER|char ch 666 Returns a boolean indicating whether the specified character is an 667 uppercase character, analogous to C<m/[[:upper:]]/>. 668 See the L<top of this section|/Character classes> for an explanation of variants 669 C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>, 670 C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>. 671 672 =for apidoc Am|bool|isPRINT|char ch 673 Returns a boolean indicating whether the specified character is a 674 printable character, analogous to C<m/[[:print:]]/>. 675 See the L<top of this section|/Character classes> for an explanation of variants 676 C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>, 677 C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>. 678 679 =for apidoc Am|bool|isWORDCHAR|char ch 680 Returns a boolean indicating whether the specified character is a character 681 that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match 682 in a regular expression. A word character is an alphabetic character, a 683 decimal digit, a connecting punctuation character (such as an underscore), or 684 a "mark" character that attaches to one of those (like some sort of accent). 685 C<isALNUM()> is a synonym provided for backward compatibility, even though a 686 word character includes more than the standard C language meaning of 687 alphanumeric. 688 See the L<top of this section|/Character classes> for an explanation of variants 689 C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>, 690 C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>. 691 692 =for apidoc Am|bool|isXDIGIT|char ch 693 Returns a boolean indicating whether the specified character is a hexadecimal 694 digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()> 695 and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>. 696 See the L<top of this section|/Character classes> for an explanation of variants 697 C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and 698 C<isXDIGIT_LC_utf8>. 699 700 =for apidoc Am|bool|isIDFIRST|char ch 701 Returns a boolean indicating whether the specified character can be the first 702 character of an identifier. This is very close to, but not quite the same as 703 the official Unicode property C<XID_Start>. The difference is that this 704 returns true only if the input character also matches L</isWORDCHAR>. 705 See the L<top of this section|/Character classes> for an explanation of variants 706 C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>, 707 C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>. 708 709 =for apidoc Am|bool|isIDCONT|char ch 710 Returns a boolean indicating whether the specified character can be the 711 second or succeeding character of an identifier. This is very close to, but 712 not quite the same as the official Unicode property C<XID_Continue>. The 713 difference is that this returns true only if the input character also matches 714 L</isWORDCHAR>. See the L<top of this section|/Character classes> for an 715 explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>, 716 C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and 717 C<isIDCONT_LC_utf8>. 718 719 =head1 Miscellaneous Functions 720 721 =for apidoc Am|U8|READ_XDIGIT|char str* 722 Returns the value of an ASCII-range hex digit and advances the string pointer. 723 Behaviour is only well defined when isXDIGIT(*str) is true. 724 725 =head1 Character case changing 726 727 =for apidoc Am|U8|toUPPER|U8 ch 728 Converts the specified character to uppercase. If the input is anything but an 729 ASCII lowercase character, that input character itself is returned. Variant 730 C<toUPPER_A> is equivalent. 731 732 =for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp 733 Converts the Unicode code point C<cp> to its uppercase version, and 734 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 735 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 736 bytes since the uppercase version may be longer than the original character. 737 738 The first code point of the uppercased version is returned 739 (but note, as explained just above, that there may be more.) 740 741 =for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp 742 Converts the UTF-8 encoded character at C<p> to its uppercase version, and 743 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 744 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 745 bytes since the uppercase version may be longer than the original character. 746 747 The first code point of the uppercased version is returned 748 (but note, as explained just above, that there may be more.) 749 750 The input character at C<p> is assumed to be well-formed. 751 752 =for apidoc Am|U8|toFOLD|U8 ch 753 Converts the specified character to foldcase. If the input is anything but an 754 ASCII uppercase character, that input character itself is returned. Variant 755 C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full 756 Latin1 range, as the full generality of L</toFOLD_uni> is needed there.) 757 758 =for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp 759 Converts the Unicode code point C<cp> to its foldcase version, and 760 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 761 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 762 bytes since the foldcase version may be longer than the original character. 763 764 The first code point of the foldcased version is returned 765 (but note, as explained just above, that there may be more.) 766 767 =for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp 768 Converts the UTF-8 encoded character at C<p> to its foldcase version, and 769 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 770 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 771 bytes since the foldcase version may be longer than the original character. 772 773 The first code point of the foldcased version is returned 774 (but note, as explained just above, that there may be more.) 775 776 The input character at C<p> is assumed to be well-formed. 777 778 =for apidoc Am|U8|toLOWER|U8 ch 779 Converts the specified character to lowercase. If the input is anything but an 780 ASCII uppercase character, that input character itself is returned. Variant 781 C<toLOWER_A> is equivalent. 782 783 =for apidoc Am|U8|toLOWER_L1|U8 ch 784 Converts the specified Latin1 character to lowercase. The results are undefined if 785 the input doesn't fit in a byte. 786 787 =for apidoc Am|U8|toLOWER_LC|U8 ch 788 Converts the specified character to lowercase using the current locale's rules, 789 if possible; otherwise returns the input character itself. 790 791 =for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp 792 Converts the Unicode code point C<cp> to its lowercase version, and 793 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 794 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 795 bytes since the lowercase version may be longer than the original character. 796 797 The first code point of the lowercased version is returned 798 (but note, as explained just above, that there may be more.) 799 800 =for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp 801 Converts the UTF-8 encoded character at C<p> to its lowercase version, and 802 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 803 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 804 bytes since the lowercase version may be longer than the original character. 805 806 The first code point of the lowercased version is returned 807 (but note, as explained just above, that there may be more.) 808 809 The input character at C<p> is assumed to be well-formed. 810 811 =for apidoc Am|U8|toLOWER_LC|U8 ch 812 Converts the specified character to lowercase using the current locale's rules, 813 if possible; otherwise returns the input character itself. 814 815 =for apidoc Am|U8|toTITLE|U8 ch 816 Converts the specified character to titlecase. If the input is anything but an 817 ASCII lowercase character, that input character itself is returned. Variant 818 C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range, 819 as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a 820 concept used in locale handling, so there is no functionality for that.) 821 822 =for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp 823 Converts the Unicode code point C<cp> to its titlecase version, and 824 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 825 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 826 bytes since the titlecase version may be longer than the original character. 827 828 The first code point of the titlecased version is returned 829 (but note, as explained just above, that there may be more.) 830 831 =for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp 832 Converts the UTF-8 encoded character at C<p> to its titlecase version, and 833 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note 834 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> 835 bytes since the titlecase version may be longer than the original character. 836 837 The first code point of the titlecased version is returned 838 (but note, as explained just above, that there may be more.) 839 840 The input character at C<p> is assumed to be well-formed. 841 842 =cut 843 844 XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names 845 really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change. 846 847 Note that these macros are repeated in Devel::PPPort, so should also be 848 patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc 849 850 */ 851 852 /* Specify the widest unsigned type on the platform. Use U64TYPE because U64 853 * is known only in the perl core, and this macro can be called from outside 854 * that */ 855 #ifdef HAS_QUAD 856 # define WIDEST_UTYPE U64TYPE 857 #else 858 # define WIDEST_UTYPE U32 859 #endif 860 861 /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in 862 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no 863 * bits of information are lost even on a 64-bit machine, but to get the 864 * compiler to optimize it out if possible. This is because Configure makes 865 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the 866 * sizeof() guarantees that this evaluates to a constant true at compile time. 867 */ 868 #define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF)) 869 870 #ifdef EBCDIC 871 # ifndef _ALL_SOURCE 872 /* This returns the wrong results on at least z/OS unless this is 873 * defined. */ 874 # error _ALL_SOURCE should probably be defined 875 # endif 876 877 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is 878 * likely not defined, and so we use the native function */ 879 # define isASCII(c) cBOOL(isascii(c)) 880 #else 881 # define isASCII(c) ((WIDEST_UTYPE)(c) < 128) 882 #endif 883 884 #define isASCII_A(c) isASCII(c) 885 #define isASCII_L1(c) isASCII(c) 886 887 /* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0, 888 * and the 8 possible permutations of those bits exactly comprise the 8 octal 889 * digits */ 890 #define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0') 891 892 /* ASCII range only */ 893 #ifdef H_PERL /* If have access to perl.h, lookup in its table */ 894 895 /* Character class numbers. For internal core Perl use only. The ones less 896 * than 32 are used in PL_charclass[] and the ones up through the one that 897 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and 898 * related files. PL_charclass ones use names used in l1_char_class_tab.h but 899 * their actual definitions are here. If that file has a name not used here, 900 * it won't compile. 901 * 902 * The first group of these is ordered in what I (khw) estimate to be the 903 * frequency of their use. This gives a slight edge to exiting a loop earlier 904 * (in reginclass() in regexec.c) */ 905 # define _CC_WORDCHAR 0 /* \w and [:word:] */ 906 # define _CC_DIGIT 1 /* \d and [:digit:] */ 907 # define _CC_ALPHA 2 /* [:alpha:] */ 908 # define _CC_LOWER 3 /* [:lower:] */ 909 # define _CC_UPPER 4 /* [:upper:] */ 910 # define _CC_PUNCT 5 /* [:punct:] */ 911 # define _CC_PRINT 6 /* [:print:] */ 912 # define _CC_ALPHANUMERIC 7 /* [:alnum:] */ 913 # define _CC_GRAPH 8 /* [:graph:] */ 914 # define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */ 915 916 #define _FIRST_NON_SWASH_CC 10 917 /* The character classes above are implemented with swashes. The second group 918 * (just below) contains the ones implemented without. These are also sorted 919 * in rough order of the frequency of their use, except that \v should be last, 920 * as it isn't a real Posix character class, and some (small) inefficiencies in 921 * regular expression handling would be introduced by putting it in the middle 922 * of those that are. Also, cntrl and ascii come after the others as it may be 923 * useful to group these which have no members that match above Latin1, (or 924 * above ASCII in the latter case) */ 925 926 # define _CC_SPACE 10 /* \s */ 927 # define _CC_BLANK 11 /* [:blank:] */ 928 # define _CC_XDIGIT 12 /* [:xdigit:] */ 929 # define _CC_PSXSPC 13 /* [:space:] */ 930 # define _CC_CNTRL 14 /* [:cntrl:] */ 931 # define _CC_ASCII 15 /* [:ascii:] */ 932 # define _CC_VERTSPACE 16 /* \v */ 933 934 # define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE 935 936 /* The members of the third group below do not need to be coordinated with data 937 * structures in regcomp.[ch] and regexec.c. But they should be added to 938 * bootstrap_ctype() */ 939 # define _CC_IDFIRST 17 940 # define _CC_CHARNAME_CONT 18 941 # define _CC_NONLATIN1_FOLD 19 942 # define _CC_QUOTEMETA 20 943 # define _CC_NON_FINAL_FOLD 21 944 # define _CC_IS_IN_SOME_FOLD 22 945 # define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */ 946 /* Unused: 23-30 947 * If more bits are needed, one could add a second word for non-64bit 948 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd 949 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it 950 * is used only for optimization (as of this writing), and differs in the 951 * Latin1 range from the ALPHA bit only in two relatively unimportant 952 * characters: the masculine and feminine ordinal indicators, so removing it 953 * would just cause /i regexes which match them to run less efficiently */ 954 955 #if defined(PERL_CORE) || defined(PERL_EXT) 956 /* An enum version of the character class numbers, to help compilers 957 * optimize */ 958 typedef enum { 959 _CC_ENUM_ALPHA = _CC_ALPHA, 960 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC, 961 _CC_ENUM_ASCII = _CC_ASCII, 962 _CC_ENUM_BLANK = _CC_BLANK, 963 _CC_ENUM_CASED = _CC_CASED, 964 _CC_ENUM_CNTRL = _CC_CNTRL, 965 _CC_ENUM_DIGIT = _CC_DIGIT, 966 _CC_ENUM_GRAPH = _CC_GRAPH, 967 _CC_ENUM_LOWER = _CC_LOWER, 968 _CC_ENUM_PRINT = _CC_PRINT, 969 _CC_ENUM_PSXSPC = _CC_PSXSPC, 970 _CC_ENUM_PUNCT = _CC_PUNCT, 971 _CC_ENUM_SPACE = _CC_SPACE, 972 _CC_ENUM_UPPER = _CC_UPPER, 973 _CC_ENUM_VERTSPACE = _CC_VERTSPACE, 974 _CC_ENUM_WORDCHAR = _CC_WORDCHAR, 975 _CC_ENUM_XDIGIT = _CC_XDIGIT 976 } _char_class_number; 977 #endif 978 979 #define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC 980 #define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1) 981 982 #if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C) 983 # if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \ 984 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \ 985 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9 986 #error Need to adjust order of swash_property_names[] 987 # endif 988 989 /* This is declared static in each of the few files that this is #defined for 990 * to keep them from being publicly accessible. Hence there is a small amount 991 * of wasted space */ 992 993 static const char* const swash_property_names[] = { 994 "XPosixWord", 995 "XPosixDigit", 996 "XPosixAlpha", 997 "XPosixLower", 998 "XPosixUpper", 999 "XPosixPunct", 1000 "XPosixPrint", 1001 "XPosixAlnum", 1002 "XPosixGraph", 1003 "Cased" 1004 }; 1005 #endif 1006 1007 # ifdef DOINIT 1008 EXTCONST U32 PL_charclass[] = { 1009 # include "l1_char_class_tab.h" 1010 }; 1011 1012 # else /* ! DOINIT */ 1013 EXTCONST U32 PL_charclass[]; 1014 # endif 1015 #endif /* Has perl.h */ 1016 1017 #if defined(H_PERL) && ! defined(BOOTSTRAP_CHARSET) 1018 1019 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */ 1020 # define _CC_mask(classnum) (1U << (classnum)) 1021 1022 /* For internal core Perl use only: the base macro for defining macros like 1023 * isALPHA */ 1024 # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \ 1025 && (PL_charclass[(U8) (c)] & _CC_mask(classnum))) 1026 1027 /* The mask for the _A versions of the macros; it just adds in the bit for 1028 * ASCII. */ 1029 # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII)) 1030 1031 /* For internal core Perl use only: the base macro for defining macros like 1032 * isALPHA_A. The foo_A version makes sure that both the desired bit and 1033 * the ASCII bit are present */ 1034 # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \ 1035 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \ 1036 == _CC_mask_A(classnum))) 1037 1038 # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA) 1039 # define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC) 1040 # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK) 1041 # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL) 1042 # define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT) 1043 # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH) 1044 # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER) 1045 # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT) 1046 # define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC) 1047 # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT) 1048 # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE) 1049 # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER) 1050 # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR) 1051 # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT) 1052 # define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST) 1053 # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA) 1054 # define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC) 1055 # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK) 1056 1057 /* continuation character for legal NAME in \N{NAME} */ 1058 # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT) 1059 1060 # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL) 1061 # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH) 1062 # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER) 1063 # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT) 1064 # define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC) 1065 # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT) 1066 # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE) 1067 # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER) 1068 # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR) 1069 # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST) 1070 1071 /* Either participates in a fold with a character above 255, or is a 1072 * multi-char fold */ 1073 # 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))) 1074 1075 # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA) 1076 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \ 1077 _generic_isCC(c, _CC_NON_FINAL_FOLD) 1078 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \ 1079 _generic_isCC(c, _CC_IS_IN_SOME_FOLD) 1080 #else /* Either don't have perl.h or don't want to use char_class_tab.h */ 1081 1082 /* If we don't have perl.h, we are compiling a utility program. Below we 1083 * hard-code various macro definitions that wouldn't otherwise be available 1084 * to it. We can also get here if we are configured to bootstrap up Perl 1085 * on a non-ASCII platform that doesn't have a working Perl (currently only 1086 * EBCDIC). For these we currently use the native definitions to get 1087 * things going. (It should also be possible to use the translation 1088 * function NATIVE_TO_LATIN1(), but that is an extra layer of dependence on 1089 * Perl, so it is currently avoided for the macros where it's possible to 1090 * do so.) */ 1091 # ifdef EBCDIC 1092 /* Use the native functions. They likely will return false for all 1093 * non-ASCII values, but this makes sure */ 1094 # define isALPHA_A(c) (isASCII(c) && isalpha(c)) 1095 # define isALPHANUMERIC_A(c) (isASCII(c) && isalnum(c)) 1096 # define isCNTRL_A(c) (isASCII(c) && iscntrl(c)) 1097 # define isDIGIT_A(c) (isASCII(c) && isdigit(c)) 1098 # define isGRAPH_A(c) (isASCII(c) && isgraph(c)) 1099 # define isLOWER_A(c) (isASCII(c) && islower(c)) 1100 # define isPRINT_A(c) (isASCII(c) && isprint(c)) 1101 # define isPUNCT_A(c) (isASCII(c) && ispunct(c)) 1102 # define isSPACE_A(c) (isASCII(c) && isspace(c)) 1103 # define isUPPER_A(c) (isASCII(c) && isupper(c)) 1104 # define isXDIGIT_A(c) (isASCII(c) && isxdigit(c)) 1105 # else /* ASCII platform. These are coded based on first principals */ 1106 # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c)) 1107 # define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c)) 1108 # define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c))) 1109 # define isDIGIT_A(c) ((c) <= '9' && (c) >= '0') 1110 # define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ') 1111 # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z') 1112 # define isPRINT_A(c) (((c) >= 32 && (c) < 127)) 1113 # define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c))) 1114 # define isSPACE_A(c) ((c) == ' ' \ 1115 || (c) == '\t' \ 1116 || (c) == '\n' \ 1117 || (c) == '\r' \ 1118 || (c) =='\v' \ 1119 || (c) == '\f') 1120 # define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A') 1121 # define isXDIGIT_A(c) (isDIGIT_A(c) \ 1122 || ((c) >= 'a' && (c) <= 'f') \ 1123 || ((c) <= 'F' && (c) >= 'A')) 1124 # endif /* Below are common definitions for ASCII and non-ASCII */ 1125 # define isBLANK_A(c) ((c) == ' ' || (c) == '\t') 1126 # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_') 1127 # define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_') 1128 1129 /* The _L1 macros may be unnecessary for both the utilities and for 1130 * bootstrapping; I (khw) added them during debugging of bootstrapping, and 1131 * it seems best to keep them. */ 1132 # define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */ 1133 # define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c)) 1134 # define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c)) 1135 # define isBLANK_L1(c) (isBLANK_A(c) \ 1136 || (FITS_IN_8_BITS(c) \ 1137 && NATIVE_TO_LATIN1((U8) c) == 0xA0)) 1138 # define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c))) 1139 # define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c))) 1140 # define isLOWER_L1(c) (isLOWER_A(c) \ 1141 || (FITS_IN_8_BITS(c) \ 1142 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \ 1143 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \ 1144 || NATIVE_TO_LATIN1((U8) c) == 0xAA \ 1145 || NATIVE_TO_LATIN1((U8) c) == 0xBA \ 1146 || NATIVE_TO_LATIN1((U8) c) == 0xB5))) 1147 # define isPRINT_L1(c) (isPRINT_A(c) \ 1148 || (FITS_IN_8_BITS(c) \ 1149 && NATIVE_TO_LATIN1((U8) c) >= 0xA0)) 1150 # define isPSXSPC_L1(c) isSPACE_L1(c) 1151 # define isPUNCT_L1(c) (isPUNCT_A(c) \ 1152 || (FITS_IN_8_BITS(c) \ 1153 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \ 1154 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \ 1155 || NATIVE_TO_LATIN1((U8) c) == 0xAB \ 1156 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \ 1157 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \ 1158 || NATIVE_TO_LATIN1((U8) c) == 0xBB \ 1159 || NATIVE_TO_LATIN1((U8) c) == 0xBF))) 1160 # define isSPACE_L1(c) (isSPACE_A(c) \ 1161 || (FITS_IN_8_BITS(c) \ 1162 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \ 1163 || NATIVE_TO_LATIN1((U8) c) == 0xA0))) 1164 # define isUPPER_L1(c) (isUPPER_A(c) \ 1165 || (FITS_IN_8_BITS(c) \ 1166 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \ 1167 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \ 1168 && NATIVE_TO_LATIN1((U8) c) != 0xD7))) 1169 # define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c)) 1170 # define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_') 1171 # define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \ 1172 || isBLANK_L1(c) \ 1173 || (c) == '-' \ 1174 || (c) == '(' \ 1175 || (c) == ')') 1176 /* The following are not fully accurate in the above-ASCII range. I (khw) 1177 * don't think it's necessary to be so for the purposes where this gets 1178 * compiled */ 1179 # define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c)) 1180 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c) 1181 1182 /* And these aren't accurate at all. They are useful only for above 1183 * Latin1, which utilities and bootstrapping don't deal with */ 1184 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0 1185 # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0 1186 1187 /* Many of the macros later in this file are defined in terms of these. By 1188 * implementing them with a function, which converts the class number into 1189 * a call to the desired macro, all of the later ones work. However, that 1190 * function won't be actually defined when building a utility program (no 1191 * perl.h), and so a compiler error will be generated if one is attempted 1192 * to be used. And the above-Latin1 code points require Unicode tables to 1193 * be present, something unlikely to be the case when bootstrapping */ 1194 # define _generic_isCC(c, classnum) \ 1195 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE)) 1196 # define _generic_isCC_A(c, classnum) \ 1197 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE)) 1198 #endif /* End of no perl.h or have BOOTSTRAP_CHARSET */ 1199 1200 #define isALPHANUMERIC(c) isALPHANUMERIC_A(c) 1201 #define isALPHA(c) isALPHA_A(c) 1202 #define isBLANK(c) isBLANK_A(c) 1203 #define isCNTRL(c) isCNTRL_A(c) 1204 #define isDIGIT(c) isDIGIT_A(c) 1205 #define isGRAPH(c) isGRAPH_A(c) 1206 #define isIDFIRST(c) isIDFIRST_A(c) 1207 #define isLOWER(c) isLOWER_A(c) 1208 #define isPRINT(c) isPRINT_A(c) 1209 #define isPSXSPC(c) isPSXSPC_A(c) 1210 #define isPUNCT(c) isPUNCT_A(c) 1211 #define isSPACE(c) isSPACE_A(c) 1212 #define isUPPER(c) isUPPER_A(c) 1213 #define isWORDCHAR(c) isWORDCHAR_A(c) 1214 #define isXDIGIT(c) isXDIGIT_A(c) 1215 1216 /* ASCII casing. These could also be written as 1217 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c)) 1218 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c)) 1219 which uses table lookup and mask instead of subtraction. (This would 1220 work because the _MOD does not apply in the ASCII range) */ 1221 #define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c)) 1222 #define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c)) 1223 1224 /* In the ASCII range, these are equivalent to what they're here defined to be. 1225 * But by creating these definitions, other code doesn't have to be aware of 1226 * this detail */ 1227 #define toFOLD(c) toLOWER(c) 1228 #define toTITLE(c) toUPPER(c) 1229 1230 #define toLOWER_A(c) toLOWER(c) 1231 #define toUPPER_A(c) toUPPER(c) 1232 #define toFOLD_A(c) toFOLD(c) 1233 #define toTITLE_A(c) toTITLE(c) 1234 1235 /* Use table lookup for speed; returns the input itself if is out-of-range */ 1236 #define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \ 1237 ? (c) \ 1238 : PL_latin1_lc[ (U8) (c) ]) 1239 #define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */ 1240 1241 /* Modified uc. Is correct uc except for three non-ascii chars which are 1242 * all mapped to one of them, and these need special handling; returns the 1243 * input itself if is out-of-range */ 1244 #define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \ 1245 ? (c) \ 1246 : PL_mod_latin1_uc[ (U8) (c) ]) 1247 #define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale 1248 1249 /* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */ 1250 1251 /* For internal core Perl use only: the base macro for defining macros like 1252 * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point 1253 * (0-255) to check. In a UTF-8 locale, the result is the same as calling 1254 * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like 1255 * _CC_UPPER, which gives the class number for doing this. For non-UTF-8 1256 * locales, the code to actually do the test this is passed in 'non_utf8'. If 1257 * 'c' is above 255, 0 is returned. For accessing the full range of possible 1258 * code points under locale rules, use the macros based on _generic_LC_uvchr 1259 * instead of this. */ 1260 #define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \ 1261 (! FITS_IN_8_BITS(c) \ 1262 ? 0 \ 1263 : IN_UTF8_CTYPE_LOCALE \ 1264 ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \ 1265 : cBOOL(non_utf8)) 1266 1267 /* For internal core Perl use only: a helper macro for defining macros like 1268 * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to 1269 * actually do this test is passed in 'non_utf8_func', which is called on 'c', 1270 * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See 1271 * _generic_LC_base for more info */ 1272 #define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \ 1273 _generic_LC_base(c,utf8_locale_classnum, \ 1274 non_utf8_func( (_LC_CAST) (c))) 1275 1276 /* For internal core Perl use only: like _generic_LC, but also returns TRUE if 1277 * 'c' is the platform's native underscore character */ 1278 #define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \ 1279 _generic_LC_base(c, utf8_locale_classnum, \ 1280 (non_utf8_func( (_LC_CAST) (c)) \ 1281 || (char)(c) == '_')) 1282 1283 /* These next three are also for internal core Perl use only: case-change 1284 * helper macros */ 1285 #define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \ 1286 ? (c) \ 1287 : (IN_UTF8_CTYPE_LOCALE) \ 1288 ? PL_latin1_lc[ (U8) (c) ] \ 1289 : function((cast)(c))) 1290 1291 /* Note that the result can be larger than a byte in a UTF-8 locale. It 1292 * returns a single value, so can't adequately return the upper case of LATIN 1293 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two 1294 * values "SS"); instead it asserts against that under DEBUGGING, and 1295 * otherwise returns its input */ 1296 #define _generic_toUPPER_LC(c, function, cast) \ 1297 (! FITS_IN_8_BITS(c) \ 1298 ? (c) \ 1299 : ((! IN_UTF8_CTYPE_LOCALE) \ 1300 ? function((cast)(c)) \ 1301 : ((((U8)(c)) == MICRO_SIGN) \ 1302 ? GREEK_CAPITAL_LETTER_MU \ 1303 : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \ 1304 ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \ 1305 : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \ 1306 ? (__ASSERT_(0) (c)) \ 1307 : PL_mod_latin1_uc[ (U8) (c) ]))))) 1308 1309 /* Note that the result can be larger than a byte in a UTF-8 locale. It 1310 * returns a single value, so can't adequately return the fold case of LATIN 1311 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two 1312 * values "ss"); instead it asserts against that under DEBUGGING, and 1313 * otherwise returns its input */ 1314 #define _generic_toFOLD_LC(c, function, cast) \ 1315 ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \ 1316 ? GREEK_SMALL_LETTER_MU \ 1317 : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \ 1318 || (c) != LATIN_SMALL_LETTER_SHARP_S) \ 1319 _generic_toLOWER_LC(c, function, cast))) 1320 1321 /* Use the libc versions for these if available. */ 1322 #if defined(HAS_ISASCII) && ! defined(USE_NEXT_CTYPE) 1323 # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c))) 1324 #else 1325 # define isASCII_LC(c) isASCII(c) 1326 #endif 1327 1328 #if defined(HAS_ISBLANK) && ! defined(USE_NEXT_CTYPE) 1329 # define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank) 1330 #else /* Unlike isASCII, varies if in a UTF-8 locale */ 1331 # define isBLANK_LC(c) (IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c) 1332 #endif 1333 1334 #ifdef USE_NEXT_CTYPE /* NeXT computers */ 1335 1336 # define _LC_CAST unsigned int /* Needed by _generic_LC. NeXT functions 1337 use this as their input type */ 1338 1339 # define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, NXIsAlpha) 1340 # define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, NXIsAlNum) 1341 # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, NXIsCntrl) 1342 # define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, NXIsDigit) 1343 # define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, NXIsGraph) 1344 # define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, NXIsAlpha) 1345 # define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, NXIsLower) 1346 # define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, NXIsPrint) 1347 # define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, NXIsPunct) 1348 # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, NXIsSpace) 1349 # define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, NXIsUpper) 1350 # define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, NXIsAlNum) 1351 # define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, NXIsXdigit) 1352 1353 # define toLOWER_LC(c) _generic_toLOWER_LC((c), NXToLower, unsigned int) 1354 # define toUPPER_LC(c) _generic_toUPPER_LC((c), NXToUpper, unsigned int) 1355 # define toFOLD_LC(c) _generic_toFOLD_LC((c), NXToLower, unsigned int) 1356 1357 #else /* !USE_NEXT_CTYPE */ 1358 1359 # define _LC_CAST U8 1360 1361 # if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII)) 1362 /* For most other platforms */ 1363 1364 # define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha) 1365 # define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum) 1366 # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl) 1367 # define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit) 1368 # define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph) 1369 # define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha) 1370 # define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower) 1371 # define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint) 1372 # define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct) 1373 # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace) 1374 # define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper) 1375 # define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum) 1376 # define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit) 1377 1378 1379 # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8) 1380 # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8) 1381 # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8) 1382 1383 # else /* The final fallback position */ 1384 1385 # define isALPHA_LC(c) (isascii(c) && isalpha(c)) 1386 # define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c)) 1387 # define isCNTRL_LC(c) (isascii(c) && iscntrl(c)) 1388 # define isDIGIT_LC(c) (isascii(c) && isdigit(c)) 1389 # define isGRAPH_LC(c) (isascii(c) && isgraph(c)) 1390 # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_')) 1391 # define isLOWER_LC(c) (isascii(c) && islower(c)) 1392 # define isPRINT_LC(c) (isascii(c) && isprint(c)) 1393 # define isPUNCT_LC(c) (isascii(c) && ispunct(c)) 1394 # define isSPACE_LC(c) (isascii(c) && isspace(c)) 1395 # define isUPPER_LC(c) (isascii(c) && isupper(c)) 1396 # define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_')) 1397 # define isXDIGIT_LC(c) (isascii(c) && isxdigit(c)) 1398 1399 # define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c)) 1400 # define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c)) 1401 # define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c)) 1402 1403 # endif 1404 #endif /* USE_NEXT_CTYPE */ 1405 1406 #define isIDCONT(c) isWORDCHAR(c) 1407 #define isIDCONT_A(c) isWORDCHAR_A(c) 1408 #define isIDCONT_L1(c) isWORDCHAR_L1(c) 1409 #define isIDCONT_LC(c) isWORDCHAR_LC(c) 1410 #define isPSXSPC_LC(c) isSPACE_LC(c) 1411 1412 /* For internal core Perl use only: the base macros for defining macros like 1413 * isALPHA_uni. 'c' is the code point to check. 'classnum' is the POSIX class 1414 * number defined earlier in this file. _generic_uni() is used for POSIX 1415 * classes where there is a macro or function 'above_latin1' that takes the 1416 * single argument 'c' and returns the desired value. These exist for those 1417 * classes which have simple definitions, avoiding the overhead of a hash 1418 * lookup or inversion list binary search. _generic_swash_uni() can be used 1419 * for classes where that overhead is faster than a direct lookup. 1420 * _generic_uni() won't compile if 'c' isn't unsigned, as it won't match the 1421 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so 1422 * have duplicate checks here, so could create versions of the macros that 1423 * don't, but experiments show that gcc optimizes them out anyway. */ 1424 1425 /* Note that all ignore 'use bytes' */ 1426 #define _generic_uni(classnum, above_latin1, c) ((c) < 256 \ 1427 ? _generic_isCC(c, classnum) \ 1428 : above_latin1(c)) 1429 #define _generic_swash_uni(classnum, c) ((c) < 256 \ 1430 ? _generic_isCC(c, classnum) \ 1431 : _is_uni_FOO(classnum, c)) 1432 #define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c) 1433 #define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c) 1434 #define isASCII_uni(c) isASCII(c) 1435 #define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c) 1436 #define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */ 1437 #define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c) 1438 #define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c) 1439 #define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c) 1440 #define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c) 1441 #define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c) 1442 #define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c) 1443 1444 /* Posix and regular space are identical above Latin1 */ 1445 #define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c) 1446 1447 #define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c) 1448 #define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c) 1449 #define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c) 1450 #define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c) 1451 #define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c) 1452 #define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c) 1453 1454 #define toFOLD_uni(c,s,l) to_uni_fold(c,s,l) 1455 #define toLOWER_uni(c,s,l) to_uni_lower(c,s,l) 1456 #define toTITLE_uni(c,s,l) to_uni_title(c,s,l) 1457 #define toUPPER_uni(c,s,l) to_uni_upper(c,s,l) 1458 1459 /* For internal core Perl use only: the base macros for defining macros like 1460 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code 1461 * point, not just 0-255. Like _generic_uni, there are two versions, one for 1462 * simple class definitions; the other for more complex. These are like 1463 * _generic_uni, so see it for more info. */ 1464 #define _generic_LC_uvchr(latin1, above_latin1, c) \ 1465 (c < 256 ? latin1(c) : above_latin1(c)) 1466 #define _generic_LC_swash_uvchr(latin1, classnum, c) \ 1467 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c)) 1468 1469 #define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c) 1470 #define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \ 1471 _CC_ALPHANUMERIC, c) 1472 #define isASCII_LC_uvchr(c) isASCII_LC(c) 1473 #define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c) 1474 #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0) 1475 #define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c) 1476 #define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c) 1477 #define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \ 1478 _is_uni_perl_idcont, c) 1479 #define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \ 1480 _is_uni_perl_idstart, c) 1481 #define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c) 1482 #define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c) 1483 #define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix 1484 space under locale */ 1485 #define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c) 1486 #define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \ 1487 is_XPERLSPACE_cp_high, c) 1488 #define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c) 1489 #define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \ 1490 _CC_WORDCHAR, c) 1491 #define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c) 1492 1493 #define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c)) 1494 1495 /* For internal core Perl use only: the base macros for defining macros like 1496 * isALPHA_utf8. These are like the earlier defined macros, but take an input 1497 * UTF-8 encoded string 'p'. If the input is in the Latin1 range, use 1498 * the Latin1 macro 'classnum' on 'p'. Otherwise use the value given by the 1499 * 'utf8' parameter. This relies on the fact that ASCII characters have the 1500 * same representation whether utf8 or not. Note that it assumes that the utf8 1501 * has been validated, and ignores 'use bytes' */ 1502 #define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \ 1503 ? _generic_isCC(*(p), classnum) \ 1504 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \ 1505 ? _generic_isCC( \ 1506 TWO_BYTE_UTF8_TO_NATIVE(*(p), \ 1507 *((p)+1 )), \ 1508 classnum) \ 1509 : utf8) 1510 /* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1' 1511 * can be a macro */ 1512 #define _generic_func_utf8(classnum, above_latin1, p) \ 1513 _generic_utf8(classnum, p, above_latin1(p)) 1514 /* Like the above, but passes classnum to _isFOO_utf8(), instead of having an 1515 * 'above_latin1' parameter */ 1516 #define _generic_swash_utf8(classnum, p) \ 1517 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p)) 1518 1519 /* Like the above, but should be used only when it is known that there are no 1520 * characters in the range 128-255 which the class is TRUE for. Hence it can 1521 * skip the tests for this range. 'above_latin1' should include its arguments */ 1522 #define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \ 1523 (UTF8_IS_INVARIANT(*(p)) \ 1524 ? _generic_isCC(*(p), classnum) \ 1525 : (UTF8_IS_ABOVE_LATIN1(*(p))) \ 1526 ? above_latin1 \ 1527 : 0) 1528 1529 /* NOTE that some of these macros have very similar ones in regcharclass.h. 1530 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()' 1531 * there, differing in name only by an underscore from the one here 1532 * 'isSPACE_utf8(). The difference is that the ones here are probably more 1533 * efficient and smaller, using an O(1) array lookup for Latin1-range code 1534 * points; the regcharclass.h ones are implemented as a series of 1535 * "if-else-if-else ..." */ 1536 1537 #define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p) 1538 #define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p) 1539 #define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under 1540 utf8, the non-utf8 macro works 1541 */ 1542 #define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p) 1543 #define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0) 1544 #define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \ 1545 _is_utf8_FOO(_CC_DIGIT, p)) 1546 #define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p) 1547 #define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \ 1548 _is_utf8_perl_idcont, p) 1549 1550 /* To prevent S_scan_word in toke.c from hanging, we have to make sure that 1551 * IDFIRST is an alnum. See 1552 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you 1553 * ever wanted to know about. (In the ASCII range, there isn't a difference.) 1554 * This used to be not the XID version, but we decided to go with the more 1555 * modern Unicode definition */ 1556 #define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \ 1557 _is_utf8_perl_idstart, p) 1558 1559 #define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p) 1560 #define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p) 1561 1562 /* Posix and regular space are identical above Latin1 */ 1563 #define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p) 1564 1565 #define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p) 1566 #define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p) 1567 #define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p) 1568 #define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p) 1569 #define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p) 1570 #define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \ 1571 is_XDIGIT_high(p)) 1572 1573 #define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l) 1574 #define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l) 1575 #define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l) 1576 #define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l) 1577 1578 /* For internal core Perl use only: the base macros for defining macros like 1579 * isALPHA_LC_utf8. These are like _generic_utf8, but if the first code point 1580 * in 'p' is within the 0-255 range, it uses locale rules from the passed-in 1581 * 'macro' parameter */ 1582 #define _generic_LC_utf8(macro, p, utf8) \ 1583 (UTF8_IS_INVARIANT(*(p)) \ 1584 ? macro(*(p)) \ 1585 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \ 1586 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \ 1587 : utf8) 1588 1589 #define _generic_LC_swash_utf8(macro, classnum, p) \ 1590 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p)) 1591 #define _generic_LC_func_utf8(macro, above_latin1, p) \ 1592 _generic_LC_utf8(macro, p, above_latin1(p)) 1593 1594 #define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \ 1595 _CC_ALPHANUMERIC, p) 1596 #define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p) 1597 #define isASCII_LC_utf8(p) isASCII_LC(*p) 1598 #define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p) 1599 #define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0) 1600 #define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p) 1601 #define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p) 1602 #define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p) 1603 #define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p) 1604 #define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p) 1605 #define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p) 1606 #define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix 1607 space under locale */ 1608 #define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p) 1609 #define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p) 1610 #define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p) 1611 #define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \ 1612 _CC_WORDCHAR, p) 1613 #define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p) 1614 1615 /* Macros for backwards compatibility and for completeness when the ASCII and 1616 * Latin1 values are identical */ 1617 #define isALPHAU(c) isALPHA_L1(c) 1618 #define isDIGIT_L1(c) isDIGIT_A(c) 1619 #define isOCTAL(c) isOCTAL_A(c) 1620 #define isOCTAL_L1(c) isOCTAL_A(c) 1621 #define isXDIGIT_L1(c) isXDIGIT_A(c) 1622 #define isALNUM(c) isWORDCHAR(c) 1623 #define isALNUMU(c) isWORDCHAR_L1(c) 1624 #define isALNUM_LC(c) isWORDCHAR_LC(c) 1625 #define isALNUM_uni(c) isWORDCHAR_uni(c) 1626 #define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c) 1627 #define isALNUM_utf8(p) isWORDCHAR_utf8(p) 1628 #define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p) 1629 #define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */ 1630 #define isALNUMC_L1(c) isALPHANUMERIC_L1(c) 1631 #define isALNUMC(c) isALPHANUMERIC(c) 1632 #define isALNUMC_LC(c) isALPHANUMERIC_LC(c) 1633 #define isALNUMC_uni(c) isALPHANUMERIC_uni(c) 1634 #define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c) 1635 #define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p) 1636 #define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p) 1637 1638 /* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, 1639 * except that they don't necessarily mean the same characters, e.g. CTRL-D is 1640 * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC. 1641 * '?' is special-cased on EBCDIC to APC, which is the control there that is 1642 * the outlier from the block that contains the other controls, just like 1643 * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0 1644 * block. If it weren't special cased, it would yield a non-control. 1645 * The conversion works both ways, so CTRL('D') is 4, and CTRL(4) is D, etc. */ 1646 #ifndef EBCDIC 1647 # define toCTRL(c) (toUPPER(c) ^ 64) 1648 #else 1649 # define toCTRL(c) ((c) == '?' \ 1650 ? LATIN1_TO_NATIVE(0x9F) \ 1651 : (c) == LATIN1_TO_NATIVE(0x9F) \ 1652 ? '?' \ 1653 : (NATIVE_TO_LATIN1(toUPPER(c)) ^ 64)) 1654 #endif 1655 1656 /* Line numbers are unsigned, 32 bits. */ 1657 typedef U32 line_t; 1658 #define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */ 1659 1660 /* Helpful alias for version prescan */ 1661 #define is_LAX_VERSION(a,b) \ 1662 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL)) 1663 1664 #define is_STRICT_VERSION(a,b) \ 1665 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL)) 1666 1667 #define BADVERSION(a,b,c) \ 1668 if (b) { \ 1669 *b = c; \ 1670 } \ 1671 return a; 1672 1673 /* Converts a character known to represent a hexadecimal digit (0-9, A-F, or 1674 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer, 1675 * which is advanced. The input is validated only by an assert() in DEBUGGING 1676 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and 1677 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */ 1678 #define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \ 1679 ? (c) \ 1680 : ((c) + 9)))) 1681 #define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \ 1682 ? (*(s)++) \ 1683 : (*(s)++ + 9)))) 1684 1685 /* Converts a character known to represent an octal digit (0-7) to its numeric 1686 * value. The input is validated only by an assert() in DEBUGGING builds. In 1687 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */ 1688 #define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c))) 1689 1690 /* 1691 =head1 Memory Management 1692 1693 =for apidoc Am|void|Newx|void* ptr|int nitems|type 1694 The XSUB-writer's interface to the C C<malloc> function. 1695 1696 Memory obtained by this should B<ONLY> be freed with L<"Safefree">. 1697 1698 In 5.9.3, Newx() and friends replace the older New() API, and drops 1699 the first parameter, I<x>, a debug aid which allowed callers to identify 1700 themselves. This aid has been superseded by a new build option, 1701 PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still 1702 there for use in XS modules supporting older perls. 1703 1704 =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast 1705 The XSUB-writer's interface to the C C<malloc> function, with 1706 cast. See also C<Newx>. 1707 1708 Memory obtained by this should B<ONLY> be freed with L<"Safefree">. 1709 1710 =for apidoc Am|void|Newxz|void* ptr|int nitems|type 1711 The XSUB-writer's interface to the C C<malloc> function. The allocated 1712 memory is zeroed with C<memzero>. See also C<Newx>. 1713 1714 Memory obtained by this should B<ONLY> be freed with L<"Safefree">. 1715 1716 =for apidoc Am|void|Renew|void* ptr|int nitems|type 1717 The XSUB-writer's interface to the C C<realloc> function. 1718 1719 Memory obtained by this should B<ONLY> be freed with L<"Safefree">. 1720 1721 =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast 1722 The XSUB-writer's interface to the C C<realloc> function, with 1723 cast. 1724 1725 Memory obtained by this should B<ONLY> be freed with L<"Safefree">. 1726 1727 =for apidoc Am|void|Safefree|void* ptr 1728 The XSUB-writer's interface to the C C<free> function. 1729 1730 This should B<ONLY> be used on memory obtained using L<"Newx"> and friends. 1731 1732 =for apidoc Am|void|Move|void* src|void* dest|int nitems|type 1733 The XSUB-writer's interface to the C C<memmove> function. The C<src> is the 1734 source, C<dest> is the destination, C<nitems> is the number of items, and 1735 C<type> is the type. Can do overlapping moves. See also C<Copy>. 1736 1737 =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type 1738 Like C<Move> but returns dest. Useful 1739 for encouraging compilers to tail-call 1740 optimise. 1741 1742 =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type 1743 The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the 1744 source, C<dest> is the destination, C<nitems> is the number of items, and 1745 C<type> is the type. May fail on overlapping copies. See also C<Move>. 1746 1747 =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type 1748 1749 Like C<Copy> but returns dest. Useful 1750 for encouraging compilers to tail-call 1751 optimise. 1752 1753 =for apidoc Am|void|Zero|void* dest|int nitems|type 1754 1755 The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the 1756 destination, C<nitems> is the number of items, and C<type> is the type. 1757 1758 =for apidoc Am|void *|ZeroD|void* dest|int nitems|type 1759 1760 Like C<Zero> but returns dest. Useful 1761 for encouraging compilers to tail-call 1762 optimise. 1763 1764 =for apidoc Am|void|StructCopy|type *src|type *dest|type 1765 This is an architecture-independent macro to copy one structure to another. 1766 1767 =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte 1768 1769 Fill up memory with a byte pattern (a byte repeated over and over 1770 again) that hopefully catches attempts to access uninitialized memory. 1771 1772 =for apidoc Am|void|PoisonNew|void* dest|int nitems|type 1773 1774 PoisonWith(0xAB) for catching access to allocated but uninitialized memory. 1775 1776 =for apidoc Am|void|PoisonFree|void* dest|int nitems|type 1777 1778 PoisonWith(0xEF) for catching access to freed memory. 1779 1780 =for apidoc Am|void|Poison|void* dest|int nitems|type 1781 1782 PoisonWith(0xEF) for catching access to freed memory. 1783 1784 =cut */ 1785 1786 /* Maintained for backwards-compatibility only. Use newSV() instead. */ 1787 #ifndef PERL_CORE 1788 #define NEWSV(x,len) newSV(len) 1789 #endif 1790 1791 #define MEM_SIZE_MAX ((MEM_SIZE)~0) 1792 1793 /* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil 1794 * overly eager compilers that will bleat about e.g. 1795 * (U16)n > (size_t)~0/sizeof(U16) always being false. */ 1796 #ifdef PERL_MALLOC_WRAP 1797 #define MEM_WRAP_CHECK(n,t) \ 1798 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (croak_memory_wrap(),0)) 1799 #define MEM_WRAP_CHECK_1(n,t,a) \ 1800 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (Perl_croak_nocontext("%s",(a)),0)) 1801 #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t), 1802 1803 #define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (croak_memory_wrap(),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1))) 1804 #else 1805 1806 #define MEM_WRAP_CHECK(n,t) 1807 #define MEM_WRAP_CHECK_1(n,t,a) 1808 #define MEM_WRAP_CHECK_2(n,t,a,b) 1809 #define MEM_WRAP_CHECK_(n,t) 1810 1811 #define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1))) 1812 1813 #endif 1814 1815 #ifdef PERL_MEM_LOG 1816 /* 1817 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s 1818 * go through functions, which are handy for debugging breakpoints, but 1819 * which more importantly get the immediate calling environment (file and 1820 * line number, and C function name if available) passed in. This info can 1821 * then be used for logging the calls, for which one gets a sample 1822 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined. 1823 * 1824 * Known problems: 1825 * - not all memory allocs get logged, only those 1826 * that go through Newx() and derivatives (while all 1827 * Safefrees do get logged) 1828 * - __FILE__ and __LINE__ do not work everywhere 1829 * - __func__ or __FUNCTION__ even less so 1830 * - I think more goes on after the perlio frees but 1831 * the thing is that STDERR gets closed (as do all 1832 * the file descriptors) 1833 * - no deeper calling stack than the caller of the Newx() 1834 * or the kind, but do I look like a C reflection/introspection 1835 * utility to you? 1836 * - the function prototypes for the logging functions 1837 * probably should maybe be somewhere else than handy.h 1838 * - one could consider inlining (macrofying) the logging 1839 * for speed, but I am too lazy 1840 * - one could imagine recording the allocations in a hash, 1841 * (keyed by the allocation address?), and maintain that 1842 * through reallocs and frees, but how to do that without 1843 * any News() happening...? 1844 * - lots of -Ddefines to get useful/controllable output 1845 * - lots of ENV reads 1846 */ 1847 1848 PERL_EXPORT_C Malloc_t Perl_mem_log_alloc(const UV n, const UV typesize, const char *type_name, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname); 1849 1850 PERL_EXPORT_C Malloc_t Perl_mem_log_realloc(const UV n, const UV typesize, const char *type_name, Malloc_t oldalloc, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname); 1851 1852 PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname); 1853 1854 # ifdef PERL_CORE 1855 # ifndef PERL_MEM_LOG_NOIMPL 1856 enum mem_log_type { 1857 MLT_ALLOC, 1858 MLT_REALLOC, 1859 MLT_FREE, 1860 MLT_NEW_SV, 1861 MLT_DEL_SV 1862 }; 1863 # endif 1864 # if defined(PERL_IN_SV_C) /* those are only used in sv.c */ 1865 void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); 1866 void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); 1867 # endif 1868 # endif 1869 1870 #endif 1871 1872 #ifdef PERL_MEM_LOG 1873 #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__) 1874 #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__) 1875 #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__) 1876 #endif 1877 1878 #ifndef MEM_LOG_ALLOC 1879 #define MEM_LOG_ALLOC(n,t,a) (a) 1880 #endif 1881 #ifndef MEM_LOG_REALLOC 1882 #define MEM_LOG_REALLOC(n,t,v,a) (a) 1883 #endif 1884 #ifndef MEM_LOG_FREE 1885 #define MEM_LOG_FREE(a) (a) 1886 #endif 1887 1888 #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) 1889 #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) 1890 #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t))))) 1891 1892 #ifndef PERL_CORE 1893 /* pre 5.9.x compatibility */ 1894 #define New(x,v,n,t) Newx(v,n,t) 1895 #define Newc(x,v,n,t,c) Newxc(v,n,t,c) 1896 #define Newz(x,v,n,t) Newxz(v,n,t) 1897 #endif 1898 1899 #define Renew(v,n,t) \ 1900 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) 1901 #define Renewc(v,n,t,c) \ 1902 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) 1903 1904 #ifdef PERL_POISON 1905 #define Safefree(d) \ 1906 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0) 1907 #else 1908 #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d))) 1909 #endif 1910 1911 #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) 1912 #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) 1913 #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t))) 1914 1915 #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) 1916 #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) 1917 #ifdef HAS_MEMSET 1918 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t))) 1919 #else 1920 /* Using bzero(), which returns void. */ 1921 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d) 1922 #endif 1923 1924 #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t))) 1925 #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB) 1926 #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF) 1927 #define Poison(d,n,t) PoisonFree(d,n,t) 1928 1929 #ifdef PERL_POISON 1930 # define PERL_POISON_EXPR(x) x 1931 #else 1932 # define PERL_POISON_EXPR(x) 1933 #endif 1934 1935 #ifdef USE_STRUCT_COPY 1936 #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s))) 1937 #else 1938 #define StructCopy(s,d,t) Copy(s,d,1,t) 1939 #endif 1940 1941 /* C_ARRAY_LENGTH is the number of elements in the C array (so you 1942 * want your zero-based indices to be less than but not equal to). 1943 * 1944 * C_ARRAY_END is one past the last: half-open/half-closed range, 1945 * not last-inclusive range. */ 1946 #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0])) 1947 #define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a)) 1948 1949 #ifdef NEED_VA_COPY 1950 # ifdef va_copy 1951 # define Perl_va_copy(s, d) va_copy(d, s) 1952 # else 1953 # if defined(__va_copy) 1954 # define Perl_va_copy(s, d) __va_copy(d, s) 1955 # else 1956 # define Perl_va_copy(s, d) Copy(s, d, 1, va_list) 1957 # endif 1958 # endif 1959 #endif 1960 1961 /* convenience debug macros */ 1962 #ifdef USE_ITHREADS 1963 #define pTHX_FORMAT "Perl interpreter: 0x%p" 1964 #define pTHX__FORMAT ", Perl interpreter: 0x%p" 1965 #define pTHX_VALUE_ (void *)my_perl, 1966 #define pTHX_VALUE (void *)my_perl 1967 #define pTHX__VALUE_ ,(void *)my_perl, 1968 #define pTHX__VALUE ,(void *)my_perl 1969 #else 1970 #define pTHX_FORMAT 1971 #define pTHX__FORMAT 1972 #define pTHX_VALUE_ 1973 #define pTHX_VALUE 1974 #define pTHX__VALUE_ 1975 #define pTHX__VALUE 1976 #endif /* USE_ITHREADS */ 1977 1978 /* Perl_deprecate was not part of the public API, and did not have a deprecate() 1979 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor 1980 CPAN::Unpack show any users outside the core. */ 1981 #ifdef PERL_CORE 1982 # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated") 1983 #endif 1984 1985 /* Internal macros to deal with gids and uids */ 1986 #ifdef PERL_CORE 1987 1988 # if Uid_t_size > IVSIZE 1989 # define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid)) 1990 # define SvUID(sv) SvNV(sv) 1991 # else 1992 # if Uid_t_sign <= 0 1993 # define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid)) 1994 # define SvUID(sv) SvIV(sv) 1995 # else 1996 # define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid)) 1997 # define SvUID(sv) SvUV(sv) 1998 # endif 1999 # endif /* Uid_t_size */ 2000 2001 # if Gid_t_size > IVSIZE 2002 # define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid)) 2003 # define SvGID(sv) SvNV(sv) 2004 # else 2005 # if Gid_t_sign <= 0 2006 # define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid)) 2007 # define SvGID(sv) SvIV(sv) 2008 # else 2009 # define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid)) 2010 # define SvGID(sv) SvUV(sv) 2011 # endif 2012 # endif /* Gid_t_size */ 2013 2014 #endif 2015 2016 #endif /* HANDY_H */ 2017 2018 /* 2019 * Local variables: 2020 * c-indentation-style: bsd 2021 * c-basic-offset: 4 2022 * indent-tabs-mode: nil 2023 * End: 2024 * 2025 * ex: set ts=8 sts=4 sw=4 et: 2026 */ 2027