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 #ifndef HANDY_H /* Guard against nested #inclusion */ 12 #define HANDY_H 13 14 #if !defined(__STDC__) 15 #ifdef NULL 16 #undef NULL 17 #endif 18 # define NULL 0 19 #endif 20 21 #ifndef PERL_CORE 22 # define Null(type) ((type)NULL) 23 24 /* 25 =head1 Handy Values 26 27 =for apidoc AmU||Nullch 28 Null character pointer. (No longer available when C<PERL_CORE> is defined.) 29 30 =for apidoc AmU||Nullsv 31 Null SV pointer. (No longer available when C<PERL_CORE> is defined.) 32 33 =cut 34 */ 35 36 # define Nullch Null(char*) 37 # define Nullfp Null(PerlIO*) 38 # define Nullsv Null(SV*) 39 #endif 40 41 #ifdef TRUE 42 #undef TRUE 43 #endif 44 #ifdef FALSE 45 #undef FALSE 46 #endif 47 #define TRUE (1) 48 #define FALSE (0) 49 50 /* The MUTABLE_*() macros cast pointers to the types shown, in such a way 51 * (compiler permitting) that casting away const-ness will give a warning; 52 * e.g.: 53 * 54 * const SV *sv = ...; 55 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away 56 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn 57 */ 58 59 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN) 60 # define MUTABLE_PTR(p) ({ void *_p = (p); _p; }) 61 #else 62 # define MUTABLE_PTR(p) ((void *) (p)) 63 #endif 64 65 #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p)) 66 #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p)) 67 #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p)) 68 #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p)) 69 #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p)) 70 #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p)) 71 72 #ifdef I_STDBOOL 73 # include <stdbool.h> 74 # ifndef HAS_BOOL 75 # define HAS_BOOL 1 76 # endif 77 #endif 78 79 /* bool is built-in for g++-2.6.3 and later, which might be used 80 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't 81 be sure _G_config.h will be included before this file. _G_config.h 82 also defines _G_HAVE_BOOL for both gcc and g++, but only g++ 83 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us. 84 g++ can be identified by __GNUG__. 85 Andy Dougherty February 2000 86 */ 87 #ifdef __GNUG__ /* GNU g++ has bool built-in */ 88 # ifndef HAS_BOOL 89 # define HAS_BOOL 1 90 # endif 91 #endif 92 93 /* The NeXT dynamic loader headers will not build with the bool macro 94 So declare them now to clear confusion. 95 */ 96 #if defined(NeXT) || defined(__NeXT__) 97 # undef FALSE 98 # undef TRUE 99 typedef enum bool { FALSE = 0, TRUE = 1 } bool; 100 # define ENUM_BOOL 1 101 # ifndef HAS_BOOL 102 # define HAS_BOOL 1 103 # endif /* !HAS_BOOL */ 104 #endif /* NeXT || __NeXT__ */ 105 106 #ifndef HAS_BOOL 107 # if defined(VMS) 108 # define bool int 109 # else 110 # define bool char 111 # endif 112 # define HAS_BOOL 1 113 #endif 114 115 /* a simple (bool) cast may not do the right thing: if bool is defined 116 * as char for example, then the cast from int is implementation-defined 117 * (bool)!!(cbool) in a ternary triggers a bug in xlc on AIX 118 */ 119 120 #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0) 121 122 /* Try to figure out __func__ or __FUNCTION__ equivalent, if any. 123 * XXX Should really be a Configure probe, with HAS__FUNCTION__ 124 * and FUNCTION__ as results. 125 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */ 126 #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */ 127 # define FUNCTION__ __func__ 128 #else 129 # if (defined(_MSC_VER) && _MSC_VER < 1300) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \ 130 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */ 131 # define FUNCTION__ "" 132 # else 133 # define FUNCTION__ __FUNCTION__ /* Common extension. */ 134 # endif 135 #endif 136 137 /* XXX A note on the perl source internal type system. The 138 original intent was that I32 be *exactly* 32 bits. 139 140 Currently, we only guarantee that I32 is *at least* 32 bits. 141 Specifically, if int is 64 bits, then so is I32. (This is the case 142 for the Cray.) This has the advantage of meshing nicely with 143 standard library calls (where we pass an I32 and the library is 144 expecting an int), but the disadvantage that an I32 is not 32 bits. 145 Andy Dougherty August 1996 146 147 There is no guarantee that there is *any* integral type with 148 exactly 32 bits. It is perfectly legal for a system to have 149 sizeof(short) == sizeof(int) == sizeof(long) == 8. 150 151 Similarly, there is no guarantee that I16 and U16 have exactly 16 152 bits. 153 154 For dealing with issues that may arise from various 32/64-bit 155 systems, we will ask Configure to check out 156 157 SHORTSIZE == sizeof(short) 158 INTSIZE == sizeof(int) 159 LONGSIZE == sizeof(long) 160 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG) 161 PTRSIZE == sizeof(void *) 162 DOUBLESIZE == sizeof(double) 163 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE). 164 165 */ 166 167 #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */ 168 # include <inttypes.h> 169 # ifdef INT32_MIN_BROKEN 170 # undef INT32_MIN 171 # define INT32_MIN (-2147483647-1) 172 # endif 173 # ifdef INT64_MIN_BROKEN 174 # undef INT64_MIN 175 # define INT64_MIN (-9223372036854775807LL-1) 176 # endif 177 #endif 178 179 typedef I8TYPE I8; 180 typedef U8TYPE U8; 181 typedef I16TYPE I16; 182 typedef U16TYPE U16; 183 typedef I32TYPE I32; 184 typedef U32TYPE U32; 185 #ifdef PERL_CORE 186 # ifdef HAS_QUAD 187 typedef I64TYPE I64; 188 typedef U64TYPE U64; 189 # endif 190 #endif /* PERL_CORE */ 191 192 #if defined(HAS_QUAD) && defined(USE_64_BIT_INT) 193 # if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG 194 # define PeRl_INT64_C(c) CAT2(c,LL) 195 # define PeRl_UINT64_C(c) CAT2(c,ULL) 196 # else 197 # if QUADKIND == QUAD_IS___INT64 198 # define PeRl_INT64_C(c) CAT2(c,I64) 199 # define PeRl_UINT64_C(c) CAT2(c,UI64) 200 # else 201 # if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG 202 # define PeRl_INT64_C(c) CAT2(c,L) 203 # define PeRl_UINT64_C(c) CAT2(c,UL) 204 # else 205 # define PeRl_INT64_C(c) ((I64TYPE)(c)) 206 # define PeRl_UINT64_C(c) ((U64TYPE)(c)) 207 # endif 208 # endif 209 # endif 210 # ifndef UINT64_C 211 # define UINT64_C(c) PeRl_UINT64_C(c) 212 # endif 213 # ifndef INT64_C 214 # define INT64_C(c) PeRl_INT64_C(c) 215 # endif 216 #endif 217 218 #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX) 219 220 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type. 221 Please search CHAR_MAX in perl.h for further details. */ 222 #define U8_MAX UINT8_MAX 223 #define U8_MIN UINT8_MIN 224 225 #define I16_MAX INT16_MAX 226 #define I16_MIN INT16_MIN 227 #define U16_MAX UINT16_MAX 228 #define U16_MIN UINT16_MIN 229 230 #define I32_MAX INT32_MAX 231 #define I32_MIN INT32_MIN 232 #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */ 233 # define U32_MAX UINT32_MAX 234 #else 235 # define U32_MAX 4294967295U 236 #endif 237 #define U32_MIN UINT32_MIN 238 239 #else 240 241 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type. 242 Please search CHAR_MAX in perl.h for further details. */ 243 #define U8_MAX PERL_UCHAR_MAX 244 #define U8_MIN PERL_UCHAR_MIN 245 246 #define I16_MAX PERL_SHORT_MAX 247 #define I16_MIN PERL_SHORT_MIN 248 #define U16_MAX PERL_USHORT_MAX 249 #define U16_MIN PERL_USHORT_MIN 250 251 #if LONGSIZE > 4 252 # define I32_MAX PERL_INT_MAX 253 # define I32_MIN PERL_INT_MIN 254 # define U32_MAX PERL_UINT_MAX 255 # define U32_MIN PERL_UINT_MIN 256 #else 257 # define I32_MAX PERL_LONG_MAX 258 # define I32_MIN PERL_LONG_MIN 259 # define U32_MAX PERL_ULONG_MAX 260 # define U32_MIN PERL_ULONG_MIN 261 #endif 262 263 #endif 264 265 /* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */ 266 #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */ 267 #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8) 268 #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */ 269 270 #define Ctl(ch) ((ch) & 037) 271 272 /* 273 =head1 SV-Body Allocation 274 275 =for apidoc Ama|SV*|newSVpvs|const char* s 276 Like C<newSVpvn>, but takes a literal string instead of a string/length pair. 277 278 =for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags 279 Like C<newSVpvn_flags>, but takes a literal string instead of a string/length 280 pair. 281 282 =for apidoc Ama|SV*|newSVpvs_share|const char* s 283 Like C<newSVpvn_share>, but takes a literal string instead of a string/length 284 pair and omits the hash parameter. 285 286 =for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags 287 Like C<sv_catpvn_flags>, but takes a literal string instead of a 288 string/length pair. 289 290 =for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s 291 Like C<sv_catpvn_nomg>, but takes a literal string instead of a 292 string/length pair. 293 294 =for apidoc Am|void|sv_catpvs|SV* sv|const char* s 295 Like C<sv_catpvn>, but takes a literal string instead of a string/length pair. 296 297 =for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s 298 Like C<sv_catpvn_mg>, but takes a literal string instead of a 299 string/length pair. 300 301 =for apidoc Am|void|sv_setpvs|SV* sv|const char* s 302 Like C<sv_setpvn>, but takes a literal string instead of a string/length pair. 303 304 =for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s 305 Like C<sv_setpvn_mg>, but takes a literal string instead of a 306 string/length pair. 307 308 =for apidoc Am|SV *|sv_setref_pvs|const char* s 309 Like C<sv_setref_pvn>, but takes a literal string instead of a 310 string/length pair. 311 312 =head1 Memory Management 313 314 =for apidoc Ama|char*|savepvs|const char* s 315 Like C<savepvn>, but takes a literal string instead of a string/length pair. 316 317 =for apidoc Ama|char*|savesharedpvs|const char* s 318 A version of C<savepvs()> which allocates the duplicate string in memory 319 which is shared between threads. 320 321 =head1 GV Functions 322 323 =for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create 324 Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair. 325 326 =head1 Hash Manipulation Functions 327 328 =for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval 329 Like C<hv_fetch>, but takes a literal string instead of a string/length pair. 330 331 =for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val 332 Like C<hv_store>, but takes a literal string instead of a string/length pair 333 and omits the hash parameter. 334 335 =head1 Lexer interface 336 337 =for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags 338 339 Like L</lex_stuff_pvn>, but takes a literal string instead of a 340 string/length pair. 341 342 =cut 343 */ 344 345 /* concatenating with "" ensures that only literal strings are accepted as 346 * argument */ 347 #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1) 348 349 /* note that STR_WITH_LEN() can't be used as argument to macros or functions 350 * that under some configurations might be macros, which means that it requires 351 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used. 352 */ 353 354 /* STR_WITH_LEN() shortcuts */ 355 #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str)) 356 #define newSVpvs_flags(str,flags) \ 357 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags) 358 #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0) 359 #define sv_catpvs_flags(sv, str, flags) \ 360 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags) 361 #define sv_catpvs_nomg(sv, str) \ 362 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0) 363 #define sv_catpvs(sv, str) \ 364 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC) 365 #define sv_catpvs_mg(sv, str) \ 366 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC) 367 #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str)) 368 #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str)) 369 #define sv_setref_pvs(rv, classname, str) \ 370 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str)) 371 #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str)) 372 #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str)) 373 #define gv_stashpvs(str, create) \ 374 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create) 375 #define gv_fetchpvs(namebeg, add, sv_type) \ 376 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type) 377 #define gv_fetchpvn(namebeg, len, add, sv_type) \ 378 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type) 379 #define sv_catxmlpvs(dsv, str, utf8) \ 380 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8) 381 #define hv_fetchs(hv,key,lval) \ 382 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \ 383 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \ 384 : HV_FETCH_JUST_SV, NULL, 0)) 385 386 #define hv_stores(hv,key,val) \ 387 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \ 388 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0)) 389 390 #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags) 391 392 #define get_cvs(str, flags) \ 393 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags)) 394 395 /* 396 =head1 Miscellaneous Functions 397 398 =for apidoc Am|bool|strNE|char* s1|char* s2 399 Test two strings to see if they are different. Returns true or 400 false. 401 402 =for apidoc Am|bool|strEQ|char* s1|char* s2 403 Test two strings to see if they are equal. Returns true or false. 404 405 =for apidoc Am|bool|strLT|char* s1|char* s2 406 Test two strings to see if the first, C<s1>, is less than the second, 407 C<s2>. Returns true or false. 408 409 =for apidoc Am|bool|strLE|char* s1|char* s2 410 Test two strings to see if the first, C<s1>, is less than or equal to the 411 second, C<s2>. Returns true or false. 412 413 =for apidoc Am|bool|strGT|char* s1|char* s2 414 Test two strings to see if the first, C<s1>, is greater than the second, 415 C<s2>. Returns true or false. 416 417 =for apidoc Am|bool|strGE|char* s1|char* s2 418 Test two strings to see if the first, C<s1>, is greater than or equal to 419 the second, C<s2>. Returns true or false. 420 421 =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len 422 Test two strings to see if they are different. The C<len> parameter 423 indicates the number of bytes to compare. Returns true or false. (A 424 wrapper for C<strncmp>). 425 426 =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len 427 Test two strings to see if they are equal. The C<len> parameter indicates 428 the number of bytes to compare. Returns true or false. (A wrapper for 429 C<strncmp>). 430 431 =cut 432 */ 433 434 #define strNE(s1,s2) (strcmp(s1,s2)) 435 #define strEQ(s1,s2) (!strcmp(s1,s2)) 436 #define strLT(s1,s2) (strcmp(s1,s2) < 0) 437 #define strLE(s1,s2) (strcmp(s1,s2) <= 0) 438 #define strGT(s1,s2) (strcmp(s1,s2) > 0) 439 #define strGE(s1,s2) (strcmp(s1,s2) >= 0) 440 #define strnNE(s1,s2,l) (strncmp(s1,s2,l)) 441 #define strnEQ(s1,s2,l) (!strncmp(s1,s2,l)) 442 443 #ifdef HAS_MEMCMP 444 # define memNE(s1,s2,l) (memcmp(s1,s2,l)) 445 # define memEQ(s1,s2,l) (!memcmp(s1,s2,l)) 446 #else 447 # define memNE(s1,s2,l) (bcmp(s1,s2,l)) 448 # define memEQ(s1,s2,l) (!bcmp(s1,s2,l)) 449 #endif 450 451 #define memEQs(s1, l, s2) \ 452 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1))) 453 #define memNEs(s1, l, s2) !memEQs(s1, l, s2) 454 455 /* 456 * Character classes. 457 * 458 * Unfortunately, the introduction of locales means that we 459 * can't trust isupper(), etc. to tell the truth. And when 460 * it comes to /\w+/ with tainting enabled, we *must* be able 461 * to trust our character classes. 462 * 463 * Therefore, the default tests in the text of Perl will be 464 * independent of locale. Any code that wants to depend on 465 * the current locale will use the tests that begin with "lc". 466 */ 467 468 #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */ 469 # ifndef CTYPE256 470 # define CTYPE256 471 # endif 472 #endif 473 474 /* 475 476 =head1 Character classes 477 This section is about functions (really macros) that classify characters 478 into types, such as punctuation versus alphabetic, etc. Most of these are 479 analogous to regular expression character classes. (See 480 L<perlrecharclass/POSIX Character Classes>.) There are several variants for 481 each class. (Not all macros have all variants; each item below lists the 482 ones valid for it.) None are affected by C<use bytes>, and only the ones 483 with C<LC> in the name are affected by the current locale. 484 485 The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a 486 C<U8>) as input and returns a boolean as to whether or not the character 487 represented by that octet is (or on non-ASCII platforms, corresponds to) an 488 ASCII character in the named class based on platform, Unicode, and Perl rules. 489 If the input is a number that doesn't fit in an octet, FALSE is returned. 490 491 Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function 492 with no suffix C<"_A">. 493 494 Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set 495 onto the platform. That is, the code points that are ASCII are unaffected, 496 since ASCII is a subset of Latin-1. But the non-ASCII code points are treated 497 as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return 498 true when called with the code point 0xDF, which is a word character in both 499 ASCII and EBCDIC (though it represent different characters in each). 500 501 Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code 502 point as input. If the code point is larger than 255, Unicode rules are used 503 to determine if it is in the character class. For example, 504 C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A 505 WITH MACRON in Unicode, and is a word character. 506 507 Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a 508 (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The 509 classification of just the first (possibly multi-byte) character in the string 510 is tested. 511 512 Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but uses 513 the C library function that gives the named classification instead of 514 hard-coded rules. For example, C<isDIGIT_LC()> returns the result of calling 515 C<isdigit()>. This means that the result is based on the current locale, which 516 is what C<LC> in the name stands for. FALSE is always returned if the input 517 won't fit into an octet. 518 519 Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It 520 returns the same as C<isFOO_LC> for input code points less than 256, and 521 returns the hard-coded, not-affected-by-locale, Unicode results for larger ones. 522 523 Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, 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 =for apidoc Am|bool|isALPHA|char ch 529 Returns a boolean indicating whether the specified character is an 530 alphabetic character, analogous to C<m/[[:alpha:]]/>. 531 See the L<top of this section|/Character classes> for an explanation of variants 532 C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>, 533 C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>. 534 535 =for apidoc Am|bool|isALPHANUMERIC|char ch 536 Returns a boolean indicating whether the specified character is a either an 537 alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>. 538 See the L<top of this section|/Character classes> for an explanation of variants 539 C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>, 540 C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and 541 C<isALPHANUMERIC_LC_utf8>. 542 543 =for apidoc Am|bool|isASCII|char ch 544 Returns a boolean indicating whether the specified character is one of the 128 545 characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>. 546 On non-ASCII platforms, it returns TRUE iff this 547 character corresponds to an ASCII character. Variants C<isASCII_A()> and 548 C<isASCII_L1()> are identical to C<isASCII()>. 549 See the L<top of this section|/Character classes> for an explanation of variants 550 C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and 551 C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C 552 library routine C<isascii()>. In these cases, the variants whose names contain 553 C<LC> are the same as the corresponding ones without. 554 555 =for apidoc Am|bool|isBLANK|char ch 556 Returns a boolean indicating whether the specified character is a 557 character considered to be a blank, analogous to C<m/[[:blank:]]/>. 558 See the L<top of this section|/Character classes> for an explanation of variants 559 C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>, 560 C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some 561 platforms do not have the C library routine C<isblank()>. In these cases, the 562 variants whose names contain C<LC> are the same as the corresponding ones 563 without. 564 565 =for apidoc Am|bool|isCNTRL|char ch 566 Returns a boolean indicating whether the specified character is a 567 control character, analogous to C<m/[[:cntrl:]]/>. 568 See the L<top of this section|/Character classes> for an explanation of variants 569 C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>, 570 C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8> 571 On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant. 572 573 =for apidoc Am|bool|isDIGIT|char ch 574 Returns a boolean indicating whether the specified character is a 575 digit, analogous to C<m/[[:digit:]]/>. 576 Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>. 577 See the L<top of this section|/Character classes> for an explanation of variants 578 C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and 579 C<isDIGIT_LC_utf8>. 580 581 =for apidoc Am|bool|isGRAPH|char ch 582 Returns a boolean indicating whether the specified character is a 583 graphic character, analogous to C<m/[[:graph:]]/>. 584 See the L<top of this section|/Character classes> for an explanation of variants 585 C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>, 586 C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>. 587 588 =for apidoc Am|bool|isLOWER|char ch 589 Returns a boolean indicating whether the specified character is a 590 lowercase character, analogous to C<m/[[:lower:]]/>. 591 See the L<top of this section|/Character classes> for an explanation of variants 592 C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>, 593 C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>. 594 595 =for apidoc Am|bool|isOCTAL|char ch 596 Returns a boolean indicating whether the specified character is an 597 octal digit, [0-7]. 598 The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to 599 C<isOCTAL>. 600 601 =for apidoc Am|bool|isPUNCT|char ch 602 Returns a boolean indicating whether the specified character is a 603 punctuation character, analogous to C<m/[[:punct:]]/>. 604 Note that the definition of what is punctuation isn't as 605 straightforward as one might desire. See L<perlrecharclass/POSIX Character 606 Classes> for details. 607 See the L<top of this section|/Character classes> for an explanation of variants 608 C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>, 609 C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>. 610 611 =for apidoc Am|bool|isSPACE|char ch 612 Returns a boolean indicating whether the specified character is a 613 whitespace character. This is analogous 614 to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18 615 (experimentally), this also matches what C<m/[[:space:]]/> does. 616 ("Experimentally" means that this change may be backed out in 5.20 or 5.22 if 617 field experience indicates that it was unwise.) Prior to 5.18, only the 618 locale forms of this macro (the ones with C<LC> in their names) matched 619 precisely what C<m/[[:space:]]/> does. In those releases, the only difference, 620 in the non-locale variants, was that C<isSPACE()> did not match a vertical tab. 621 (See L</isPSXSPC> for a macro that matches a vertical tab in all releases.) 622 See the L<top of this section|/Character classes> for an explanation of variants 623 C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>, 624 C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>. 625 626 =for apidoc Am|bool|isPSXSPC|char ch 627 (short for Posix Space) 628 Starting in 5.18, this is identical (experimentally) in all its forms to the 629 corresponding C<isSPACE()> macros. ("Experimentally" means that this change 630 may be backed out in 5.20 or 5.22 if field experience indicates that it 631 was unwise.) 632 The locale forms of this macro are identical to their corresponding 633 C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the 634 non-locale forms differ from their C<isSPACE()> forms only in that the 635 C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do. 636 Otherwise they are identical. Thus this macro is analogous to what 637 C<m/[[:space:]]/> matches in a regular expression. 638 See the L<top of this section|/Character classes> for an explanation of variants 639 C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>, 640 C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>. 641 642 =for apidoc Am|bool|isUPPER|char ch 643 Returns a boolean indicating whether the specified character is an 644 uppercase character, analogous to C<m/[[:upper:]]/>. 645 See the L<top of this section|/Character classes> for an explanation of variants 646 C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>, 647 C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>. 648 649 =for apidoc Am|bool|isPRINT|char ch 650 Returns a boolean indicating whether the specified character is a 651 printable character, analogous to C<m/[[:print:]]/>. 652 See the L<top of this section|/Character classes> for an explanation of variants 653 C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>, 654 C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>. 655 656 =for apidoc Am|bool|isWORDCHAR|char ch 657 Returns a boolean indicating whether the specified character is a character 658 that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match 659 in a regular expression. A word character is an alphabetic character, a 660 decimal digit, a connecting punctuation character (such as an underscore), or 661 a "mark" character that attaches to one of those (like some sort of accent). 662 C<isALNUM()> is a synonym provided for backward compatibility, even though a 663 word character includes more than the standard C language meaning of 664 alphanumeric. 665 See the L<top of this section|/Character classes> for an explanation of variants 666 C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>, 667 C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>. 668 669 =for apidoc Am|bool|isXDIGIT|char ch 670 Returns a boolean indicating whether the specified character is a hexadecimal 671 digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()> 672 and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>. 673 See the L<top of this section|/Character classes> for an explanation of variants 674 C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and 675 C<isXDIGIT_LC_utf8>. 676 677 =for apidoc Am|bool|isIDFIRST|char ch 678 Returns a boolean indicating whether the specified character can be the first 679 character of an identifier. This is very close to, but not quite the same as 680 the official Unicode property C<XID_Start>. The difference is that this 681 returns true only if the input character also matches L</isWORDCHAR>. 682 See the L<top of this section|/Character classes> for an explanation of variants 683 C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>, 684 C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>. 685 686 =for apidoc Am|bool|isIDCONT|char ch 687 Returns a boolean indicating whether the specified character can be the 688 second or succeeding character of an identifier. This is very close to, but 689 not quite the same as the official Unicode property C<XID_Continue>. The 690 difference is that this returns true only if the input character also matches 691 L</isWORDCHAR>. See the L<top of this section|/Character classes> for an 692 explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>, 693 C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and 694 C<isIDCONT_LC_utf8>. 695 696 =head1 Miscellaneous Functions 697 698 =for apidoc Am|U8|READ_XDIGIT|char str* 699 Returns the value of an ASCII-range hex digit and advances the string pointer. 700 Behaviour is only well defined when isXDIGIT(*str) is true. 701 702 =head1 Character case changing 703 704 =for apidoc Am|char|toUPPER|char ch 705 Converts the specified character to uppercase, if possible; otherwise returns 706 the input character itself. 707 708 =for apidoc Am|char|toLOWER|char ch 709 Converts the specified character to lowercase, if possible; otherwise returns 710 the input character itself. 711 712 =cut 713 714 XXX Still undocumented isVERTWS_uni and _utf8, and the other toUPPER etc functions 715 716 Note that these macros are repeated in Devel::PPPort, so should also be 717 patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc 718 719 */ 720 721 /* Specify the widest unsigned type on the platform. Use U64TYPE because U64 722 * is known only in the perl core, and this macro can be called from outside 723 * that */ 724 #ifdef HAS_QUAD 725 # define WIDEST_UTYPE U64TYPE 726 #else 727 # define WIDEST_UTYPE U32 728 #endif 729 730 /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in 731 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no 732 * bits of information are lost even on a 64-bit machine, but to get the 733 * compiler to optimize it out if possible. This is because Configure makes 734 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the 735 * sizeof() guarantees that this evaluates to a constant true at compile time. 736 */ 737 #define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF)) 738 739 #ifdef EBCDIC 740 # define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128)) 741 #else 742 # define isASCII(c) ((WIDEST_UTYPE)(c) < 128) 743 #endif 744 745 #define isASCII_A(c) isASCII(c) 746 #define isASCII_L1(c) isASCII(c) 747 748 /* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0, 749 * and the 8 possible permutations of those bits exactly comprise the 8 octal 750 * digits */ 751 #define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0') 752 753 /* ASCII range only */ 754 #ifdef H_PERL /* If have access to perl.h, lookup in its table */ 755 756 /* Character class numbers. For internal core Perl use only. The ones less 757 * than 32 are used in PL_charclass[] and the ones up through the one that 758 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and 759 * related files. PL_charclass ones use names used in l1_char_class_tab.h but 760 * their actual definitions are here. If that file has a name not used here, 761 * it won't compile. 762 * 763 * The first group of these is ordered in what I (khw) estimate to be the 764 * frequency of their use. This gives a slight edge to exiting a loop earlier 765 * (in reginclass() in regexec.c) */ 766 # define _CC_WORDCHAR 0 /* \w and [:word:] */ 767 # define _CC_DIGIT 1 /* \d and [:digit:] */ 768 # define _CC_ALPHA 2 /* [:alpha:] */ 769 # define _CC_LOWER 3 /* [:lower:] */ 770 # define _CC_UPPER 4 /* [:upper:] */ 771 # define _CC_PUNCT 5 /* [:punct:] */ 772 # define _CC_PRINT 6 /* [:print:] */ 773 # define _CC_ALPHANUMERIC 7 /* [:alnum:] */ 774 # define _CC_GRAPH 8 /* [:graph:] */ 775 # define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */ 776 777 #define _FIRST_NON_SWASH_CC 10 778 /* The character classes above are implemented with swashes. The second group 779 * (just below) contains the ones implemented without. These are also sorted 780 * in rough order of the frequency of their use, except that \v should be last, 781 * as it isn't a real Posix character class, and some (small) inefficiencies in 782 * regular expression handling would be introduced by putting it in the middle 783 * of those that are. Also, cntrl and ascii come after the others as it may be 784 * useful to group these which have no members that match above Latin1, (or 785 * above ASCII in the latter case) */ 786 787 # define _CC_SPACE 10 /* \s */ 788 # define _CC_BLANK 11 /* [:blank:] */ 789 # define _CC_XDIGIT 12 /* [:xdigit:] */ 790 # define _CC_PSXSPC 13 /* [:space:] */ 791 # define _CC_CNTRL 14 /* [:cntrl:] */ 792 # define _CC_ASCII 15 /* [:ascii:] */ 793 # define _CC_VERTSPACE 16 /* \v */ 794 795 # define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE 796 797 /* The members of the third group below do not need to be coordinated with data 798 * structures in regcomp.[ch] and regexec.c */ 799 # define _CC_IDFIRST 17 800 # define _CC_CHARNAME_CONT 18 801 # define _CC_NONLATIN1_FOLD 19 802 # define _CC_QUOTEMETA 20 803 # define _CC_NON_FINAL_FOLD 21 804 # define _CC_IS_IN_SOME_FOLD 22 805 # define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */ 806 /* Unused: 23-30 807 * If more bits are needed, one could add a second word for non-64bit 808 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd 809 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it 810 * is used only for optimization (as of this writing), and differs in the 811 * Latin1 range from the ALPHA bit only in two relatively unimportant 812 * characters: the masculine and feminine ordinal indicators, so removing it 813 * would just cause /i regexes which match them to run less efficiently */ 814 815 #if defined(PERL_CORE) || defined(PERL_EXT) 816 /* An enum version of the character class numbers, to help compilers 817 * optimize */ 818 typedef enum { 819 _CC_ENUM_ALPHA = _CC_ALPHA, 820 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC, 821 _CC_ENUM_ASCII = _CC_ASCII, 822 _CC_ENUM_BLANK = _CC_BLANK, 823 _CC_ENUM_CASED = _CC_CASED, 824 _CC_ENUM_CNTRL = _CC_CNTRL, 825 _CC_ENUM_DIGIT = _CC_DIGIT, 826 _CC_ENUM_GRAPH = _CC_GRAPH, 827 _CC_ENUM_LOWER = _CC_LOWER, 828 _CC_ENUM_PRINT = _CC_PRINT, 829 _CC_ENUM_PSXSPC = _CC_PSXSPC, 830 _CC_ENUM_PUNCT = _CC_PUNCT, 831 _CC_ENUM_SPACE = _CC_SPACE, 832 _CC_ENUM_UPPER = _CC_UPPER, 833 _CC_ENUM_VERTSPACE = _CC_VERTSPACE, 834 _CC_ENUM_WORDCHAR = _CC_WORDCHAR, 835 _CC_ENUM_XDIGIT = _CC_XDIGIT 836 } _char_class_number; 837 #endif 838 839 #define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC 840 #define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1) 841 842 #if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C) 843 # if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \ 844 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \ 845 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9 846 #error Need to adjust order of swash_property_names[] 847 # endif 848 849 /* This is declared static in each of the few files that this is #defined for 850 * to keep them from being publicly accessible. Hence there is a small amount 851 * of wasted space */ 852 853 static const char* const swash_property_names[] = { 854 "XPosixWord", 855 "XPosixDigit", 856 "XPosixAlpha", 857 "XPosixLower", 858 "XPosixUpper", 859 "XPosixPunct", 860 "XPosixPrint", 861 "XPosixAlnum", 862 "XPosixGraph", 863 "Cased" 864 }; 865 #endif 866 867 # ifdef DOINIT 868 EXTCONST U32 PL_charclass[] = { 869 # include "l1_char_class_tab.h" 870 }; 871 872 # else /* ! DOINIT */ 873 EXTCONST U32 PL_charclass[]; 874 # endif 875 876 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */ 877 # define _CC_mask(classnum) (1U << (classnum)) 878 # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \ 879 && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask(classnum))) 880 881 /* The mask for the _A versions of the macros; it just adds in the bit for 882 * ASCII. */ 883 # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII)) 884 885 /* The _A version makes sure that both the desired bit and the ASCII bit 886 * are present */ 887 # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \ 888 && ((PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask_A(classnum)) \ 889 == _CC_mask_A(classnum))) 890 891 # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA) 892 # define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC) 893 # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK) 894 # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL) 895 # define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT) 896 # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH) 897 # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER) 898 # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT) 899 # define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC) 900 # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT) 901 # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE) 902 # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER) 903 # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR) 904 # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT) 905 # define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST)) 906 907 /* Either participates in a fold with a character above 255, or is a 908 * multi-char fold */ 909 # 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) NATIVE_TO_UNI(c)] & _CC_mask(_CC_NONLATIN1_FOLD))) 910 911 # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA) 912 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \ 913 _generic_isCC(c, _CC_NON_FINAL_FOLD) 914 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \ 915 _generic_isCC(c, _CC_IS_IN_SOME_FOLD) 916 #else /* No perl.h. */ 917 # ifdef EBCDIC 918 # define isALPHA_A(c) (isASCII(c) && isALPHA(c)) 919 # define isALPHANUMERIC_A(c) (isASCII(c) && isALPHANUMERIC(c)) 920 # define isBLANK_A(c) (isASCII(c) && isBLANK(c)) 921 # define isCNTRL_A(c) (isASCII(c) && isCNTRL(c)) 922 # define isDIGIT_A(c) (isASCII(c) && isDIGIT(c)) 923 # define isGRAPH_A(c) (isASCII(c) && isGRAPH(c)) 924 # define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c)) 925 # define isLOWER_A(c) (isASCII(c) && isLOWER(c)) 926 # define isPRINT_A(c) (isASCII(c) && isPRINT(c)) 927 # define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c)) 928 # define isPUNCT_A(c) (isASCII(c) && isPUNCT(c)) 929 # define isSPACE_A(c) (isASCII(c) && isSPACE(c)) 930 # define isUPPER_A(c) (isASCII(c) && isUPPER(c)) 931 # define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c)) 932 # define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c)) 933 # else /* ASCII platform, no perl.h */ 934 # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c)) 935 # define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c)) 936 # define isBLANK_A(c) ((c) == ' ' || (c) == '\t') 937 # define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127)) 938 # define isDIGIT_A(c) ((c) <= '9' && (c) >= '0') 939 # define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c)) 940 # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_') 941 # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z') 942 # define isPRINT_A(c) (((c) >= 32 && (c) < 127)) 943 # define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v') 944 # define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) \ 945 || ((c) >= 58 && (c) <= 64) \ 946 || ((c) >= 91 && (c) <= 96) \ 947 || ((c) >= 123 && (c) <= 126)) 948 # define isSPACE_A(c) ((c) == ' ' \ 949 || (c) == '\t' \ 950 || (c) == '\n' \ 951 || (c) =='\r' \ 952 || (c) == '\f') 953 # define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A') 954 # define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_') 955 # define isXDIGIT_A(c) (isDIGIT_A(c) \ 956 || ((c) >= 'a' && (c) <= 'f') \ 957 || ((c) <= 'F' && (c) >= 'A')) 958 # endif 959 #endif /* ASCII range definitions */ 960 961 /* Latin1 definitions */ 962 #ifdef H_PERL 963 # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA) 964 # define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC) 965 # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK) 966 967 /* continuation character for legal NAME in \N{NAME} */ 968 # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT) 969 970 # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL) 971 # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH) 972 # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER) 973 # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT) 974 # define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC) 975 # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT) 976 # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE) 977 # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER) 978 # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR) 979 # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST) 980 #else /* No access to perl.h. Only a few provided here, just in case needed 981 * for backwards compatibility */ 982 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra 983 * >= AA test to speed up ASCII-only tests at the expense of the others */ 984 # define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \ 985 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \ 986 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \ 987 || NATIVE_TO_UNI((U8) c) == 0xAA \ 988 || NATIVE_TO_UNI((U8) c) == 0xB5 \ 989 || NATIVE_TO_UNI((U8) c) == 0xBA))) 990 # define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \ 991 || (c) == ' ' \ 992 || (c) == '-' \ 993 || (c) == '(' \ 994 || (c) == ')' \ 995 || (c) == ':' \ 996 || NATIVE_TO_UNI((U8) c) == 0xA0) 997 #endif 998 999 /* Macros that differ between EBCDIC and ASCII. Where C89 defines a function, 1000 * that is used in the EBCDIC form, because in EBCDIC we do not do locales: 1001 * therefore can use native functions. For those where C89 doesn't define a 1002 * function, use our function, assuming that the EBCDIC code page is isomorphic 1003 * with Latin1, which the three currently recognized by Perl are. Some libc's 1004 * have an isblank(), but it's not guaranteed. */ 1005 #ifdef EBCDIC 1006 # define isALPHA(c) isalpha(c) 1007 # define isALPHANUMERIC(c) isalnum(c) 1008 # define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0) 1009 # define isCNTRL(c) iscntrl(c) 1010 # define isDIGIT(c) isdigit(c) 1011 # define isGRAPH(c) isgraph(c) 1012 # define isIDFIRST(c) (isALPHA(c) || (c) == '_') 1013 # define isLOWER(c) islower(c) 1014 # define isPRINT(c) isprint(c) 1015 # define isPSXSPC(c) isspace(c) 1016 # define isPUNCT(c) ispunct(c) 1017 # define isSPACE(c) (isPSXSPC(c) /* && (c) != '\v' (Experimentally making 1018 these macros identical) */) 1019 # define isUPPER(c) isupper(c) 1020 # define isXDIGIT(c) isxdigit(c) 1021 # define isWORDCHAR(c) (isalnum(c) || (c) == '_') 1022 # define toLOWER(c) tolower(c) 1023 # define toUPPER(c) toupper(c) 1024 #else /* Not EBCDIC: ASCII-only matching */ 1025 # define isALPHANUMERIC(c) isALPHANUMERIC_A(c) 1026 # define isALPHA(c) isALPHA_A(c) 1027 # define isBLANK(c) isBLANK_A(c) 1028 # define isCNTRL(c) isCNTRL_A(c) 1029 # define isDIGIT(c) isDIGIT_A(c) 1030 # define isGRAPH(c) isGRAPH_A(c) 1031 # define isIDFIRST(c) isIDFIRST_A(c) 1032 # define isLOWER(c) isLOWER_A(c) 1033 # define isPRINT(c) isPRINT_A(c) 1034 # define isPSXSPC(c) isPSXSPC_A(c) 1035 # define isPUNCT(c) isPUNCT_A(c) 1036 # define isSPACE(c) isSPACE_A(c) 1037 # define isUPPER(c) isUPPER_A(c) 1038 # define isWORDCHAR(c) isWORDCHAR_A(c) 1039 # define isXDIGIT(c) isXDIGIT_A(c) 1040 1041 /* ASCII casing. These could also be written as 1042 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c)) 1043 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c)) 1044 which uses table lookup and mask instead of subtraction. (This would 1045 work because the _MOD does not apply in the ASCII range) */ 1046 # define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c)) 1047 # define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c)) 1048 #endif 1049 1050 1051 /* Use table lookup for speed; return error character for input 1052 * out-of-range */ 1053 #define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \ 1054 ? UNI_TO_NATIVE(PL_latin1_lc[ \ 1055 NATIVE_TO_UNI( (U8) (c)) ]) \ 1056 : UNICODE_REPLACEMENT) 1057 /* Modified uc. Is correct uc except for three non-ascii chars which are 1058 * all mapped to one of them, and these need special handling; error 1059 * character for input out-of-range */ 1060 #define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \ 1061 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \ 1062 NATIVE_TO_UNI( (U8) (c)) ]) \ 1063 : UNICODE_REPLACEMENT) 1064 1065 #ifdef USE_NEXT_CTYPE 1066 1067 # define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c)) 1068 # define isALPHA_LC(c) NXIsAlpha((unsigned int)(c)) 1069 # define isASCII_LC(c) isASCII((unsigned int)(c)) 1070 # define isBLANK_LC(c) isBLANK((unsigned int)(c)) 1071 # define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c)) 1072 # define isDIGIT_LC(c) NXIsDigit((unsigned int)(c)) 1073 # define isGRAPH_LC(c) NXIsGraph((unsigned int)(c)) 1074 # define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_') 1075 # define isLOWER_LC(c) NXIsLower((unsigned int)(c)) 1076 # define isPRINT_LC(c) NXIsPrint((unsigned int)(c)) 1077 # define isPUNCT_LC(c) NXIsPunct((unsigned int)(c)) 1078 # define isSPACE_LC(c) NXIsSpace((unsigned int)(c)) 1079 # define isUPPER_LC(c) NXIsUpper((unsigned int)(c)) 1080 # define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_') 1081 # define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c)) 1082 # define toLOWER_LC(c) NXToLower((unsigned int)(c)) 1083 # define toUPPER_LC(c) NXToUpper((unsigned int)(c)) 1084 1085 #else /* !USE_NEXT_CTYPE */ 1086 1087 # if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII)) 1088 1089 /* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */ 1090 1091 # define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c))) 1092 # define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \ 1093 && isalnum((unsigned char)(c))) 1094 # ifdef HAS_ISASCII 1095 # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c))) 1096 # else 1097 # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c))) 1098 # endif 1099 # ifdef HAS_ISBLANK 1100 # define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c))) 1101 # else 1102 # define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c))) 1103 # endif 1104 # define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c))) 1105 # define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c))) 1106 # define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c))) 1107 # define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \ 1108 && (isalpha((unsigned char)(c)) || (char)(c) == '_')) 1109 # define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c))) 1110 # define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c))) 1111 # define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c))) 1112 # define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c))) 1113 # define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c))) 1114 # define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \ 1115 && (isalnum((unsigned char)(c)) || (char)(c) == '_')) 1116 # define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c))) 1117 # define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? tolower((unsigned char)(c)) : (c)) 1118 # define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? toupper((unsigned char)(c)) : (c)) 1119 1120 # else 1121 1122 # define isALPHA_LC(c) (isascii(c) && isalpha(c)) 1123 # define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c)) 1124 # define isASCII_LC(c) isascii(c) 1125 # ifdef HAS_ISBLANK 1126 # define isBLANK_LC(c) (isascii(c) && isblank(c)) 1127 # else 1128 # define isBLANK_LC(c) isBLANK_A(c) 1129 # endif 1130 # define isCNTRL_LC(c) (isascii(c) && iscntrl(c)) 1131 # define isDIGIT_LC(c) (isascii(c) && isdigit(c)) 1132 # define isGRAPH_LC(c) (isascii(c) && isgraph(c)) 1133 # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_')) 1134 # define isLOWER_LC(c) (isascii(c) && islower(c)) 1135 # define isPRINT_LC(c) (isascii(c) && isprint(c)) 1136 # define isPUNCT_LC(c) (isascii(c) && ispunct(c)) 1137 # define isSPACE_LC(c) (isascii(c) && isspace(c)) 1138 # define isUPPER_LC(c) (isascii(c) && isupper(c)) 1139 # define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_')) 1140 # define isXDIGIT_LC(c) (isascii(c) && isxdigit(c)) 1141 # define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c)) 1142 # define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c)) 1143 1144 # endif 1145 #endif /* USE_NEXT_CTYPE */ 1146 1147 #define isIDCONT(c) isWORDCHAR(c) 1148 #define isIDCONT_A(c) isWORDCHAR_A(c) 1149 #define isIDCONT_L1(c) isWORDCHAR_L1(c) 1150 #define isIDCONT_LC(c) isWORDCHAR_LC(c) 1151 #define isPSXSPC_LC(c) isSPACE_LC(c) 1152 1153 /* For internal core Perl use only. If the input is Latin1, use the Latin1 1154 * macro; otherwise use the function 'above_latin1'. Won't compile if 'c' isn't unsigned, as 1155 * won't match above_latin1 prototype. The macros do bounds checking, so have 1156 * duplicate checks here, so could create versions of the macros that don't, 1157 * but experiments show that gcc optimizes them out anyway. */ 1158 1159 /* Note that all ignore 'use bytes' */ 1160 #define _generic_uni(classnum, above_latin1, c) ((c) < 256 \ 1161 ? _generic_isCC(c, classnum) \ 1162 : above_latin1(c)) 1163 #define _generic_swash_uni(classnum, c) ((c) < 256 \ 1164 ? _generic_isCC(c, classnum) \ 1165 : _is_uni_FOO(classnum, c)) 1166 #define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c) 1167 #define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c) 1168 #define isASCII_uni(c) isASCII(c) 1169 #define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c) 1170 #define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */ 1171 #define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c) 1172 #define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c) 1173 #define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c) 1174 #define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c) 1175 #define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c) 1176 #define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c) 1177 1178 /* Posix and regular space are identical above Latin1 */ 1179 #define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c) 1180 1181 #define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c) 1182 #define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c) 1183 #define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c) 1184 #define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c) 1185 #define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c) 1186 #define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c) 1187 1188 #define toFOLD_uni(c,s,l) to_uni_fold(c,s,l) 1189 #define toLOWER_uni(c,s,l) to_uni_lower(c,s,l) 1190 #define toTITLE_uni(c,s,l) to_uni_title(c,s,l) 1191 #define toUPPER_uni(c,s,l) to_uni_upper(c,s,l) 1192 1193 #define _generic_LC_uvchr(latin1, above_latin1, c) \ 1194 (c < 256 ? latin1(c) : above_latin1(c)) 1195 #define _generic_LC_swash_uvchr(latin1, classnum, c) \ 1196 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c)) 1197 1198 #define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c) 1199 #define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \ 1200 _CC_ALPHANUMERIC, c) 1201 #define isASCII_LC_uvchr(c) isASCII_LC(c) 1202 #define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c) 1203 #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0) 1204 #define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c) 1205 #define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c) 1206 #define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \ 1207 _is_uni_perl_idcont, c) 1208 #define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \ 1209 _is_uni_perl_idstart, c) 1210 #define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c) 1211 #define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c) 1212 #define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix 1213 space under locale */ 1214 #define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c) 1215 #define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \ 1216 is_XPERLSPACE_cp_high, c) 1217 #define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c) 1218 #define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \ 1219 _CC_WORDCHAR, c) 1220 #define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c) 1221 1222 1223 #define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c)) 1224 1225 /* Everything whose name begins with an underscore is for internal core Perl 1226 * use only. */ 1227 1228 /* If the input is in the Latin1 range, use 1229 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string. 1230 * Otherwise use the value given by the 'utf8' parameter. This relies on the 1231 * fact that ASCII characters have the same representation whether utf8 or not. 1232 * Note that it assumes that the utf8 has been validated, and ignores 'use 1233 * bytes' */ 1234 #define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \ 1235 ? _generic_isCC(*(p), classnum) \ 1236 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \ 1237 ? _generic_isCC( \ 1238 TWO_BYTE_UTF8_TO_UNI(*(p), \ 1239 *((p)+1 )), \ 1240 classnum) \ 1241 : utf8) 1242 /* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1' 1243 * can be a macro */ 1244 #define _generic_func_utf8(classnum, above_latin1, p) \ 1245 _generic_utf8(classnum, p, above_latin1(p)) 1246 /* Like the above, but passes classnum to _isFOO_utf8(), instead of having a 1247 * 'above_latin1' parameter */ 1248 #define _generic_swash_utf8(classnum, p) \ 1249 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p)) 1250 1251 /* Like the above, but should be used only when it is known that there are no 1252 * characters in the range 128-255 which the class is TRUE for. Hence it can 1253 * skip the tests for this range. 'above_latin1' should include its arguments */ 1254 #define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \ 1255 (UTF8_IS_INVARIANT(*(p)) \ 1256 ? _generic_isCC(*(p), classnum) \ 1257 : (UTF8_IS_ABOVE_LATIN1(*(p))) \ 1258 ? above_latin1 \ 1259 : 0) 1260 1261 /* NOTE that some of these macros have very similar ones in regcharclass.h. 1262 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()' 1263 * there, differing in name only by an underscore from the one here 1264 * 'isSPACE_utf8(). The difference is that the ones here are probably more 1265 * efficient and smaller, using an O(1) array lookup for Latin1-range code 1266 * points; the regcharclass.h ones are implemented as a series of 1267 * "if-else-if-else ..." */ 1268 1269 #define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p) 1270 #define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p) 1271 #define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under 1272 utf8, the non-utf8 macro works 1273 */ 1274 #define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p) 1275 #define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0) 1276 #define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \ 1277 _is_utf8_FOO(_CC_DIGIT, p)) 1278 #define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p) 1279 #define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \ 1280 _is_utf8_perl_idcont, p) 1281 1282 /* To prevent S_scan_word in toke.c from hanging, we have to make sure that 1283 * IDFIRST is an alnum. See 1284 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you 1285 * ever wanted to know about. (In the ASCII range, there isn't a difference.) 1286 * This used to be not the XID version, but we decided to go with the more 1287 * modern Unicode definition */ 1288 #define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \ 1289 _is_utf8_perl_idstart, p) 1290 1291 #define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p) 1292 #define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p) 1293 1294 /* Posix and regular space are identical above Latin1 */ 1295 #define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p) 1296 1297 #define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p) 1298 #define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p) 1299 #define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p) 1300 #define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p) 1301 #define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p) 1302 #define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \ 1303 is_XDIGIT_high(p)) 1304 1305 #define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l) 1306 #define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l) 1307 #define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l) 1308 1309 /* For internal core Perl use only. If the input is in the Latin1 range, use 1310 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise 1311 * use the value given by the 'utf8' parameter. This relies on the fact that 1312 * ASCII characters have the same representation whether utf8 or not. Note 1313 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */ 1314 #define _generic_LC_utf8(macro, p, utf8) \ 1315 (UTF8_IS_INVARIANT(*(p)) \ 1316 ? macro(*(p)) \ 1317 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \ 1318 ? macro(TWO_BYTE_UTF8_TO_UNI(*(p), *((p)+1))) \ 1319 : utf8) 1320 1321 #define _generic_LC_swash_utf8(macro, classnum, p) \ 1322 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p)) 1323 #define _generic_LC_func_utf8(macro, above_latin1, p) \ 1324 _generic_LC_utf8(macro, p, above_latin1(p)) 1325 1326 #define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \ 1327 _CC_ALPHANUMERIC, p) 1328 #define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p) 1329 #define isASCII_LC_utf8(p) isASCII_LC(*p) 1330 #define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p) 1331 #define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0) 1332 #define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p) 1333 #define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p) 1334 #define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p) 1335 #define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p) 1336 #define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p) 1337 #define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p) 1338 #define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix 1339 space under locale */ 1340 #define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p) 1341 #define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p) 1342 #define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p) 1343 #define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \ 1344 _CC_WORDCHAR, p) 1345 #define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p) 1346 1347 /* Macros for backwards compatibility and for completeness when the ASCII and 1348 * Latin1 values are identical */ 1349 #define isALPHAU(c) isALPHA_L1(c) 1350 #define isDIGIT_L1(c) isDIGIT_A(c) 1351 #define isOCTAL(c) isOCTAL_A(c) 1352 #define isOCTAL_L1(c) isOCTAL_A(c) 1353 #define isXDIGIT_L1(c) isXDIGIT_A(c) 1354 #define isALNUM(c) isWORDCHAR(c) 1355 #define isALNUMU(c) isWORDCHAR_L1(c) 1356 #define isALNUM_LC(c) isWORDCHAR_LC(c) 1357 #define isALNUM_uni(c) isWORDCHAR_uni(c) 1358 #define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c) 1359 #define isALNUM_utf8(p) isWORDCHAR_utf8(p) 1360 #define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p) 1361 #define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */ 1362 #define isALNUMC_L1(c) isALPHANUMERIC_L1(c) 1363 #define isALNUMC(c) isALPHANUMERIC(c) 1364 #define isALNUMC_LC(c) isALPHANUMERIC_LC(c) 1365 #define isALNUMC_uni(c) isALPHANUMERIC_uni(c) 1366 #define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c) 1367 #define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p) 1368 #define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p) 1369 1370 /* This conversion works both ways, strangely enough. On EBCDIC platforms, 1371 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't 1372 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but 1373 * that is EOT on ASCII; ST on EBCDIC */ 1374 # define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64) 1375 1376 /* Line numbers are unsigned, 32 bits. */ 1377 typedef U32 line_t; 1378 #define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */ 1379 1380 /* Helpful alias for version prescan */ 1381 #define is_LAX_VERSION(a,b) \ 1382 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL)) 1383 1384 #define is_STRICT_VERSION(a,b) \ 1385 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL)) 1386 1387 #define BADVERSION(a,b,c) \ 1388 if (b) { \ 1389 *b = c; \ 1390 } \ 1391 return a; 1392 1393 #define READ_XDIGIT(s) (isALPHA(*(s)) ? ((*(s)++ + 9) & 0xf) : (*(s)++ & 0xf)) 1394 1395 /* 1396 =head1 Memory Management 1397 1398 =for apidoc Am|void|Newx|void* ptr|int nitems|type 1399 The XSUB-writer's interface to the C C<malloc> function. 1400 1401 In 5.9.3, Newx() and friends replace the older New() API, and drops 1402 the first parameter, I<x>, a debug aid which allowed callers to identify 1403 themselves. This aid has been superseded by a new build option, 1404 PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still 1405 there for use in XS modules supporting older perls. 1406 1407 =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast 1408 The XSUB-writer's interface to the C C<malloc> function, with 1409 cast. See also C<Newx>. 1410 1411 =for apidoc Am|void|Newxz|void* ptr|int nitems|type 1412 The XSUB-writer's interface to the C C<malloc> function. The allocated 1413 memory is zeroed with C<memzero>. See also C<Newx>. 1414 1415 =for apidoc Am|void|Renew|void* ptr|int nitems|type 1416 The XSUB-writer's interface to the C C<realloc> function. 1417 1418 =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast 1419 The XSUB-writer's interface to the C C<realloc> function, with 1420 cast. 1421 1422 =for apidoc Am|void|Safefree|void* ptr 1423 The XSUB-writer's interface to the C C<free> function. 1424 1425 =for apidoc Am|void|Move|void* src|void* dest|int nitems|type 1426 The XSUB-writer's interface to the C C<memmove> function. The C<src> is the 1427 source, C<dest> is the destination, C<nitems> is the number of items, and 1428 C<type> is the type. Can do overlapping moves. See also C<Copy>. 1429 1430 =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type 1431 Like C<Move> but returns dest. Useful for encouraging compilers to tail-call 1432 optimise. 1433 1434 =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type 1435 The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the 1436 source, C<dest> is the destination, C<nitems> is the number of items, and 1437 C<type> is the type. May fail on overlapping copies. See also C<Move>. 1438 1439 =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type 1440 1441 Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call 1442 optimise. 1443 1444 =for apidoc Am|void|Zero|void* dest|int nitems|type 1445 1446 The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the 1447 destination, C<nitems> is the number of items, and C<type> is the type. 1448 1449 =for apidoc Am|void *|ZeroD|void* dest|int nitems|type 1450 1451 Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call 1452 optimise. 1453 1454 =for apidoc Am|void|StructCopy|type *src|type *dest|type 1455 This is an architecture-independent macro to copy one structure to another. 1456 1457 =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte 1458 1459 Fill up memory with a byte pattern (a byte repeated over and over 1460 again) that hopefully catches attempts to access uninitialized memory. 1461 1462 =for apidoc Am|void|PoisonNew|void* dest|int nitems|type 1463 1464 PoisonWith(0xAB) for catching access to allocated but uninitialized memory. 1465 1466 =for apidoc Am|void|PoisonFree|void* dest|int nitems|type 1467 1468 PoisonWith(0xEF) for catching access to freed memory. 1469 1470 =for apidoc Am|void|Poison|void* dest|int nitems|type 1471 1472 PoisonWith(0xEF) for catching access to freed memory. 1473 1474 =cut */ 1475 1476 /* Maintained for backwards-compatibility only. Use newSV() instead. */ 1477 #ifndef PERL_CORE 1478 #define NEWSV(x,len) newSV(len) 1479 #endif 1480 1481 #define MEM_SIZE_MAX ((MEM_SIZE)~0) 1482 1483 /* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil 1484 * overly eager compilers that will bleat about e.g. 1485 * (U16)n > (size_t)~0/sizeof(U16) always being false. */ 1486 #ifdef PERL_MALLOC_WRAP 1487 #define MEM_WRAP_CHECK(n,t) \ 1488 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_memory_wrap(),0)) 1489 #define MEM_WRAP_CHECK_1(n,t,a) \ 1490 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0)) 1491 #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t), 1492 1493 #define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (Perl_croak_memory_wrap(),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1))) 1494 #else 1495 1496 #define MEM_WRAP_CHECK(n,t) 1497 #define MEM_WRAP_CHECK_1(n,t,a) 1498 #define MEM_WRAP_CHECK_2(n,t,a,b) 1499 #define MEM_WRAP_CHECK_(n,t) 1500 1501 #define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1))) 1502 1503 #endif 1504 1505 #ifdef PERL_MEM_LOG 1506 /* 1507 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s 1508 * go through functions, which are handy for debugging breakpoints, but 1509 * which more importantly get the immediate calling environment (file and 1510 * line number, and C function name if available) passed in. This info can 1511 * then be used for logging the calls, for which one gets a sample 1512 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined. 1513 * 1514 * Known problems: 1515 * - not all memory allocs get logged, only those 1516 * that go through Newx() and derivatives (while all 1517 * Safefrees do get logged) 1518 * - __FILE__ and __LINE__ do not work everywhere 1519 * - __func__ or __FUNCTION__ even less so 1520 * - I think more goes on after the perlio frees but 1521 * the thing is that STDERR gets closed (as do all 1522 * the file descriptors) 1523 * - no deeper calling stack than the caller of the Newx() 1524 * or the kind, but do I look like a C reflection/introspection 1525 * utility to you? 1526 * - the function prototypes for the logging functions 1527 * probably should maybe be somewhere else than handy.h 1528 * - one could consider inlining (macrofying) the logging 1529 * for speed, but I am too lazy 1530 * - one could imagine recording the allocations in a hash, 1531 * (keyed by the allocation address?), and maintain that 1532 * through reallocs and frees, but how to do that without 1533 * any News() happening...? 1534 * - lots of -Ddefines to get useful/controllable output 1535 * - lots of ENV reads 1536 */ 1537 1538 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); 1539 1540 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); 1541 1542 PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname); 1543 1544 # ifdef PERL_CORE 1545 # ifndef PERL_MEM_LOG_NOIMPL 1546 enum mem_log_type { 1547 MLT_ALLOC, 1548 MLT_REALLOC, 1549 MLT_FREE, 1550 MLT_NEW_SV, 1551 MLT_DEL_SV 1552 }; 1553 # endif 1554 # if defined(PERL_IN_SV_C) /* those are only used in sv.c */ 1555 void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); 1556 void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); 1557 # endif 1558 # endif 1559 1560 #endif 1561 1562 #ifdef PERL_MEM_LOG 1563 #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__) 1564 #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__) 1565 #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__) 1566 #endif 1567 1568 #ifndef MEM_LOG_ALLOC 1569 #define MEM_LOG_ALLOC(n,t,a) (a) 1570 #endif 1571 #ifndef MEM_LOG_REALLOC 1572 #define MEM_LOG_REALLOC(n,t,v,a) (a) 1573 #endif 1574 #ifndef MEM_LOG_FREE 1575 #define MEM_LOG_FREE(a) (a) 1576 #endif 1577 1578 #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) 1579 #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) 1580 #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t))))) 1581 1582 #ifndef PERL_CORE 1583 /* pre 5.9.x compatibility */ 1584 #define New(x,v,n,t) Newx(v,n,t) 1585 #define Newc(x,v,n,t,c) Newxc(v,n,t,c) 1586 #define Newz(x,v,n,t) Newxz(v,n,t) 1587 #endif 1588 1589 #define Renew(v,n,t) \ 1590 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) 1591 #define Renewc(v,n,t,c) \ 1592 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) 1593 1594 #ifdef PERL_POISON 1595 #define Safefree(d) \ 1596 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0) 1597 #else 1598 #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d))) 1599 #endif 1600 1601 #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) 1602 #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) 1603 #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t))) 1604 1605 #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) 1606 #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) 1607 #ifdef HAS_MEMSET 1608 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t))) 1609 #else 1610 /* Using bzero(), which returns void. */ 1611 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d) 1612 #endif 1613 1614 #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t))) 1615 #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB) 1616 #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF) 1617 #define Poison(d,n,t) PoisonFree(d,n,t) 1618 1619 #ifdef USE_STRUCT_COPY 1620 #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s))) 1621 #else 1622 #define StructCopy(s,d,t) Copy(s,d,1,t) 1623 #endif 1624 1625 #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0])) 1626 #define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0])) 1627 1628 #ifdef NEED_VA_COPY 1629 # ifdef va_copy 1630 # define Perl_va_copy(s, d) va_copy(d, s) 1631 # else 1632 # if defined(__va_copy) 1633 # define Perl_va_copy(s, d) __va_copy(d, s) 1634 # else 1635 # define Perl_va_copy(s, d) Copy(s, d, 1, va_list) 1636 # endif 1637 # endif 1638 #endif 1639 1640 /* convenience debug macros */ 1641 #ifdef USE_ITHREADS 1642 #define pTHX_FORMAT "Perl interpreter: 0x%p" 1643 #define pTHX__FORMAT ", Perl interpreter: 0x%p" 1644 #define pTHX_VALUE_ (void *)my_perl, 1645 #define pTHX_VALUE (void *)my_perl 1646 #define pTHX__VALUE_ ,(void *)my_perl, 1647 #define pTHX__VALUE ,(void *)my_perl 1648 #else 1649 #define pTHX_FORMAT 1650 #define pTHX__FORMAT 1651 #define pTHX_VALUE_ 1652 #define pTHX_VALUE 1653 #define pTHX__VALUE_ 1654 #define pTHX__VALUE 1655 #endif /* USE_ITHREADS */ 1656 1657 /* Perl_deprecate was not part of the public API, and did not have a deprecate() 1658 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor 1659 CPAN::Unpack show any users outside the core. */ 1660 #ifdef PERL_CORE 1661 # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated") 1662 #endif 1663 1664 #endif /* HANDY_H */ 1665 1666 /* 1667 * Local variables: 1668 * c-indentation-style: bsd 1669 * c-basic-offset: 4 1670 * indent-tabs-mode: nil 1671 * End: 1672 * 1673 * ex: set ts=8 sts=4 sw=4 et: 1674 */ 1675