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
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.
11 #if !defined(__STDC__)
19 # define Null(type) ((type)NULL)
24 =for apidoc AmU||Nullch
25 Null character pointer. (No longer available when C<PERL_CORE> is defined.)
27 =for apidoc AmU||Nullsv
28 Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
33 # define Nullch Null(char*)
34 # define Nullfp Null(PerlIO*)
35 # define Nullsv Null(SV*)
47 /* The MUTABLE_*() macros cast pointers to the types shown, in such a way
48 * (compiler permitting) that casting away const-ness will give a warning;
52 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
53 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
56 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
57 # define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
59 # define MUTABLE_PTR(p) ((void *) (p))
62 #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
63 #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
64 #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
65 #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
66 #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
67 #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
76 /* bool is built-in for g++-2.6.3 and later, which might be used
77 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
78 be sure _G_config.h will be included before this file. _G_config.h
79 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
80 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
81 g++ can be identified by __GNUG__.
82 Andy Dougherty February 2000
84 #ifdef __GNUG__ /* GNU g++ has bool built-in */
90 /* The NeXT dynamic loader headers will not build with the bool macro
91 So declare them now to clear confusion.
93 #if defined(NeXT) || defined(__NeXT__)
96 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
100 # endif /* !HAS_BOOL */
101 #endif /* NeXT || __NeXT__ */
112 /* a simple (bool) cast may not do the right thing: if bool is defined
113 * as char for example, then the cast from int is implementation-defined
114 * (bool)!!(cbool) in a ternary triggers a bug in xlc on AIX
117 #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
119 /* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
120 * XXX Should really be a Configure probe, with HAS__FUNCTION__
121 * and FUNCTION__ as results.
122 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
123 #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
124 # define FUNCTION__ __func__
126 # if (defined(_MSC_VER) && _MSC_VER < 1300) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
127 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
128 # define FUNCTION__ ""
130 # define FUNCTION__ __FUNCTION__ /* Common extension. */
134 /* XXX A note on the perl source internal type system. The
135 original intent was that I32 be *exactly* 32 bits.
137 Currently, we only guarantee that I32 is *at least* 32 bits.
138 Specifically, if int is 64 bits, then so is I32. (This is the case
139 for the Cray.) This has the advantage of meshing nicely with
140 standard library calls (where we pass an I32 and the library is
141 expecting an int), but the disadvantage that an I32 is not 32 bits.
142 Andy Dougherty August 1996
144 There is no guarantee that there is *any* integral type with
145 exactly 32 bits. It is perfectly legal for a system to have
146 sizeof(short) == sizeof(int) == sizeof(long) == 8.
148 Similarly, there is no guarantee that I16 and U16 have exactly 16
151 For dealing with issues that may arise from various 32/64-bit
152 systems, we will ask Configure to check out
154 SHORTSIZE == sizeof(short)
155 INTSIZE == sizeof(int)
156 LONGSIZE == sizeof(long)
157 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
158 PTRSIZE == sizeof(void *)
159 DOUBLESIZE == sizeof(double)
160 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
164 #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
165 # include <inttypes.h>
166 # ifdef INT32_MIN_BROKEN
168 # define INT32_MIN (-2147483647-1)
170 # ifdef INT64_MIN_BROKEN
172 # define INT64_MIN (-9223372036854775807LL-1)
187 #endif /* PERL_CORE */
189 #if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
190 # if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
191 # define PeRl_INT64_C(c) CAT2(c,LL)
192 # define PeRl_UINT64_C(c) CAT2(c,ULL)
194 # if QUADKIND == QUAD_IS___INT64
195 # define PeRl_INT64_C(c) CAT2(c,I64)
196 # define PeRl_UINT64_C(c) CAT2(c,UI64)
198 # if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
199 # define PeRl_INT64_C(c) CAT2(c,L)
200 # define PeRl_UINT64_C(c) CAT2(c,UL)
202 # define PeRl_INT64_C(c) ((I64TYPE)(c))
203 # define PeRl_UINT64_C(c) ((U64TYPE)(c))
208 # define UINT64_C(c) PeRl_UINT64_C(c)
211 # define INT64_C(c) PeRl_INT64_C(c)
215 #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
217 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
218 Please search CHAR_MAX in perl.h for further details. */
219 #define U8_MAX UINT8_MAX
220 #define U8_MIN UINT8_MIN
222 #define I16_MAX INT16_MAX
223 #define I16_MIN INT16_MIN
224 #define U16_MAX UINT16_MAX
225 #define U16_MIN UINT16_MIN
227 #define I32_MAX INT32_MAX
228 #define I32_MIN INT32_MIN
229 #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
230 # define U32_MAX UINT32_MAX
232 # define U32_MAX 4294967295U
234 #define U32_MIN UINT32_MIN
238 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
239 Please search CHAR_MAX in perl.h for further details. */
240 #define U8_MAX PERL_UCHAR_MAX
241 #define U8_MIN PERL_UCHAR_MIN
243 #define I16_MAX PERL_SHORT_MAX
244 #define I16_MIN PERL_SHORT_MIN
245 #define U16_MAX PERL_USHORT_MAX
246 #define U16_MIN PERL_USHORT_MIN
249 # define I32_MAX PERL_INT_MAX
250 # define I32_MIN PERL_INT_MIN
251 # define U32_MAX PERL_UINT_MAX
252 # define U32_MIN PERL_UINT_MIN
254 # define I32_MAX PERL_LONG_MAX
255 # define I32_MIN PERL_LONG_MIN
256 # define U32_MAX PERL_ULONG_MAX
257 # define U32_MIN PERL_ULONG_MIN
262 /* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
263 #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
264 #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
265 #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
267 #define Ctl(ch) ((ch) & 037)
270 =head1 SV-Body Allocation
272 =for apidoc Ama|SV*|newSVpvs|const char* s
273 Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
275 =for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
276 Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
279 =for apidoc Ama|SV*|newSVpvs_share|const char* s
280 Like C<newSVpvn_share>, but takes a literal string instead of a string/length
281 pair and omits the hash parameter.
283 =for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
284 Like C<sv_catpvn_flags>, but takes a literal string instead of a
287 =for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
288 Like C<sv_catpvn_nomg>, but takes a literal string instead of a
291 =for apidoc Am|void|sv_catpvs|SV* sv|const char* s
292 Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
294 =for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
295 Like C<sv_catpvn_mg>, but takes a literal string instead of a
298 =for apidoc Am|void|sv_setpvs|SV* sv|const char* s
299 Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
301 =for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
302 Like C<sv_setpvn_mg>, but takes a literal string instead of a
305 =for apidoc Am|SV *|sv_setref_pvs|const char* s
306 Like C<sv_setref_pvn>, but takes a literal string instead of a
309 =head1 Memory Management
311 =for apidoc Ama|char*|savepvs|const char* s
312 Like C<savepvn>, but takes a literal string instead of a string/length pair.
314 =for apidoc Ama|char*|savesharedpvs|const char* s
315 A version of C<savepvs()> which allocates the duplicate string in memory
316 which is shared between threads.
320 =for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
321 Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
323 =head1 Hash Manipulation Functions
325 =for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
326 Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
328 =for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
329 Like C<hv_store>, but takes a literal string instead of a string/length pair
330 and omits the hash parameter.
332 =head1 Lexer interface
334 =for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
336 Like L</lex_stuff_pvn>, but takes a literal string instead of a
342 /* concatenating with "" ensures that only literal strings are accepted as argument */
343 #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
345 /* note that STR_WITH_LEN() can't be used as argument to macros or functions that
346 * under some configurations might be macros, which means that it requires the full
347 * Perl_xxx(aTHX_ ...) form for any API calls where it's used.
350 /* STR_WITH_LEN() shortcuts */
351 #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
352 #define newSVpvs_flags(str,flags) \
353 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
354 #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
355 #define sv_catpvs_flags(sv, str, flags) \
356 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
357 #define sv_catpvs_nomg(sv, str) \
358 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
359 #define sv_catpvs(sv, str) \
360 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
361 #define sv_catpvs_mg(sv, str) \
362 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
363 #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
364 #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
365 #define sv_setref_pvs(rv, classname, str) \
366 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
367 #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
368 #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
369 #define gv_stashpvs(str, create) \
370 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
371 #define gv_fetchpvs(namebeg, add, sv_type) \
372 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
373 #define gv_fetchpvn(namebeg, len, add, sv_type) \
374 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
375 #define sv_catxmlpvs(dsv, str, utf8) \
376 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
377 #define hv_fetchs(hv,key,lval) \
378 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
379 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
380 : HV_FETCH_JUST_SV, NULL, 0))
382 #define hv_stores(hv,key,val) \
383 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
384 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
386 #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
388 #define get_cvs(str, flags) \
389 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
392 =head1 Miscellaneous Functions
394 =for apidoc Am|bool|strNE|char* s1|char* s2
395 Test two strings to see if they are different. Returns true or
398 =for apidoc Am|bool|strEQ|char* s1|char* s2
399 Test two strings to see if they are equal. Returns true or false.
401 =for apidoc Am|bool|strLT|char* s1|char* s2
402 Test two strings to see if the first, C<s1>, is less than the second,
403 C<s2>. Returns true or false.
405 =for apidoc Am|bool|strLE|char* s1|char* s2
406 Test two strings to see if the first, C<s1>, is less than or equal to the
407 second, C<s2>. Returns true or false.
409 =for apidoc Am|bool|strGT|char* s1|char* s2
410 Test two strings to see if the first, C<s1>, is greater than the second,
411 C<s2>. Returns true or false.
413 =for apidoc Am|bool|strGE|char* s1|char* s2
414 Test two strings to see if the first, C<s1>, is greater than or equal to
415 the second, C<s2>. Returns true or false.
417 =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
418 Test two strings to see if they are different. The C<len> parameter
419 indicates the number of bytes to compare. Returns true or false. (A
420 wrapper for C<strncmp>).
422 =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
423 Test two strings to see if they are equal. The C<len> parameter indicates
424 the number of bytes to compare. Returns true or false. (A wrapper for
430 #define strNE(s1,s2) (strcmp(s1,s2))
431 #define strEQ(s1,s2) (!strcmp(s1,s2))
432 #define strLT(s1,s2) (strcmp(s1,s2) < 0)
433 #define strLE(s1,s2) (strcmp(s1,s2) <= 0)
434 #define strGT(s1,s2) (strcmp(s1,s2) > 0)
435 #define strGE(s1,s2) (strcmp(s1,s2) >= 0)
436 #define strnNE(s1,s2,l) (strncmp(s1,s2,l))
437 #define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
440 # define memNE(s1,s2,l) (memcmp(s1,s2,l))
441 # define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
443 # define memNE(s1,s2,l) (bcmp(s1,s2,l))
444 # define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
447 #define memEQs(s1, l, s2) \
448 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
449 #define memNEs(s1, l, s2) !memEQs(s1, l, s2)
454 * Unfortunately, the introduction of locales means that we
455 * can't trust isupper(), etc. to tell the truth. And when
456 * it comes to /\w+/ with tainting enabled, we *must* be able
457 * to trust our character classes.
459 * Therefore, the default tests in the text of Perl will be
460 * independent of locale. Any code that wants to depend on
461 * the current locale will use the tests that begin with "lc".
464 #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
472 =head1 Character classes
473 There are three variants for all the functions in this section. The base ones
474 operate using the character set of the platform Perl is running on. The ones
475 with an C<_A> suffix operate on the ASCII character set, and the ones with an
476 C<_L1> suffix operate on the full Latin1 character set. All are unaffected by
477 locale and by C<use bytes>.
479 For ASCII platforms, the base function with no suffix and the one with the
480 C<_A> suffix are identical. The function with the C<_L1> suffix imposes the
481 Latin-1 character set onto the platform. That is, the code points that are
482 ASCII are unaffected, since ASCII is a subset of Latin-1. But the non-ASCII
483 code points are treated as if they are Latin-1 characters. For example,
484 C<isSPACE_L1()> will return true when called with the code point 0xA0, which is
485 the Latin-1 NO-BREAK SPACE.
487 For EBCDIC platforms, the base function with no suffix and the one with the
488 C<_L1> suffix should be identical, since, as of this writing, the EBCDIC code
489 pages that Perl knows about all are equivalent to Latin-1. The function that
490 ends in an C<_A> suffix will not return true unless the specified character also
491 has an ASCII equivalent.
493 =for apidoc Am|bool|isALPHA|char ch
494 Returns a boolean indicating whether the specified character is an
495 alphabetic character in the platform's native character set.
496 See the L<top of this section|/Character classes> for an explanation of variants
497 C<isALPHA_A> and C<isALPHA_L1>.
499 =for apidoc Am|bool|isASCII|char ch
500 Returns a boolean indicating whether the specified character is one of the 128
501 characters in the ASCII character set. On non-ASCII platforms, it is if this
502 character corresponds to an ASCII character. Variants C<isASCII_A()> and
503 C<isASCII_L1()> are identical to C<isASCII()>.
505 =for apidoc Am|bool|isDIGIT|char ch
506 Returns a boolean indicating whether the specified character is a
507 digit in the platform's native character set.
508 Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
510 =for apidoc Am|bool|isLOWER|char ch
511 Returns a boolean indicating whether the specified character is a
512 lowercase character in the platform's native character set.
513 See the L<top of this section|/Character classes> for an explanation of variants
514 C<isLOWER_A> and C<isLOWER_L1>.
516 =for apidoc Am|bool|isOCTAL|char ch
517 Returns a boolean indicating whether the specified character is an
518 octal digit, [0-7] in the platform's native character set.
519 Variants C<isOCTAL_A> and C<isOCTAL_L1> are identical to C<isOCTAL>.
521 =for apidoc Am|bool|isSPACE|char ch
522 Returns a boolean indicating whether the specified character is a
523 whitespace character in the platform's native character set. This is the same
524 as what C<\s> matches in a regular expression.
525 See the L<top of this section|/Character classes> for an explanation of variants
526 C<isSPACE_A> and C<isSPACE_L1>.
528 =for apidoc Am|bool|isUPPER|char ch
529 Returns a boolean indicating whether the specified character is an
530 uppercase character in the platform's native character set.
531 See the L<top of this section|/Character classes> for an explanation of variants
532 C<isUPPER_A> and C<isUPPER_L1>.
534 =for apidoc Am|bool|isWORDCHAR|char ch
535 Returns a boolean indicating whether the specified character is a
536 character that is any of: alphabetic, numeric, or an underscore. This is the
537 same as what C<\w> matches in a regular expression.
538 C<isALNUM()> is a synonym provided for backward compatibility. Note that it
539 does not have the standard C language meaning of alphanumeric, since it matches
540 an underscore and the standard meaning does not.
541 See the L<top of this section|/Character classes> for an explanation of variants
542 C<isWORDCHAR_A> and C<isWORDCHAR_L1>.
544 =for apidoc Am|bool|isXDIGIT|char ch
545 Returns a boolean indicating whether the specified character is a hexadecimal
546 digit, [0-9A-Fa-f]. Variants C<isXDIGIT_A()> and C<isXDIGIT_L1()> are
547 identical to C<isXDIGIT()>.
549 =for apidoc Am|U8|READ_XDIGIT|char str*
550 Returns the value of a hex digit and advances the string pointer.
551 Behaviour is only well defined when isXDIGIT(*str) is true.
553 =head1 Character case changing
555 =for apidoc Am|char|toUPPER|char ch
556 Converts the specified character to uppercase in the platform's native
557 character set, if possible; otherwise returns the input character itself.
559 =for apidoc Am|char|toLOWER|char ch
560 Converts the specified character to lowercase in the platform's native
561 character set, if possible; otherwise returns the input character itself.
565 Note that these macros are repeated in Devel::PPPort, so should also be
566 patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
570 /* Specify the widest unsigned type on the platform. Use U64TYPE because U64
571 * is known only in the perl core, and this macro can be called from outside
574 # define WIDEST_UTYPE U64TYPE
576 # define WIDEST_UTYPE U32
579 /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
580 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
581 * bits of information are lost even on a 64-bit machine, but to get the
582 * compiler to optimize it out if possible. This is because Configure makes
583 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
584 * sizeof() guarantees that this evaluates to a constant true at compile time.
586 #define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
589 # define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128))
591 # define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
594 #define isASCII_A(c) isASCII(c)
595 #define isASCII_L1(c) isASCII(c)
597 /* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
598 * and the 8 possible permutations of those bits exactly comprise the 8 octal
600 #define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
602 /* ASCII range only */
603 #ifdef H_PERL /* If have access to perl.h, lookup in its table */
605 /* Character class numbers. For internal core Perl use only. These are used
606 * in PL_charclass[] and the ones up through the one that corresponds to
607 * <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h. These use names used
608 * in l1_char_class_tab.h but their actual definitions are here. If that has a
609 * name not used here, it won't compile. */
610 # define _CC_WORDCHAR 0
613 # define _CC_ALNUMC 3
620 # define _CC_PUNCT 10
621 # define _CC_UPPER 11
622 # define _CC_XDIGIT 12
623 # define _CC_PSXSPC 13
624 # define _CC_BLANK 14
625 # define _CC_VERTSPACE 15
626 # define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
628 # define _CC_IDFIRST 16
629 # define _CC_CHARNAME_CONT 17
630 # define _CC_NONLATIN1_FOLD 18
631 # define _CC_QUOTEMETA 19
632 # define _CC_NON_FINAL_FOLD 20
633 # define _CC_IS_IN_SOME_FOLD 21
635 * If more bits are needed, one could add a second word for non-64bit
636 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
637 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
638 * is used only for optimization (as of this writing), and differs in the
639 * Latin1 range from the ALPHA bit only in two relatively unimportant
640 * characters: the masculine and feminine ordinal indicators */
643 EXTCONST U32 PL_charclass[] = {
644 # include "l1_char_class_tab.h"
647 # else /* ! DOINIT */
648 EXTCONST U32 PL_charclass[];
651 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
652 # define _CC_mask(classnum) (1U << (classnum))
653 # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
654 && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask(classnum)))
656 /* The mask for the _A versions of the macros; it just adds in the bit for
658 # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
660 /* The _A version makes sure that both the desired bit and the ASCII bit
662 # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
663 && ((PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask_A(classnum)) \
664 == _CC_mask_A(classnum)))
666 # define isALNUMC_A(c) _generic_isCC_A(c, _CC_ALNUMC)
667 # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
668 # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
669 # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
670 # define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
671 # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
672 # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
673 # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
674 # define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
675 # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
676 # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
677 # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
678 # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
679 # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
680 # define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST))
682 /* Either participates in a fold with a character above 255, or is a
684 # 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)))
686 # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
687 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
688 _generic_isCC(c, _CC_NON_FINAL_FOLD)
689 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
690 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
691 #else /* No perl.h. */
693 # define isALNUMC_A(c) (isASCII(c) && isALNUMC(c))
694 # define isALPHA_A(c) (isASCII(c) && isALPHA(c))
695 # define isBLANK_A(c) (isASCII(c) && isBLANK(c))
696 # define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
697 # define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
698 # define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
699 # define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
700 # define isLOWER_A(c) (isASCII(c) && isLOWER(c))
701 # define isPRINT_A(c) (isASCII(c) && isPRINT(c))
702 # define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
703 # define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
704 # define isSPACE_A(c) (isASCII(c) && isSPACE(c))
705 # define isUPPER_A(c) (isASCII(c) && isUPPER(c))
706 # define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
707 # define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
708 # else /* ASCII platform, no perl.h */
709 # define isALNUMC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
710 # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
711 # define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
712 # define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
713 # define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
714 # define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
715 # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
716 # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
717 # define isPRINT_A(c) (((c) >= 32 && (c) < 127))
718 # define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
719 # define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) || ((c) >= 58 && (c) <= 64) || ((c) >= 91 && (c) <= 96) || ((c) >= 123 && (c) <= 126))
720 # define isSPACE_A(c) ((c) == ' ' || (c) == '\t' || (c) == '\n' || (c) =='\r' || (c) == '\f')
721 # define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
722 # define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
723 # define isXDIGIT_A(c) (isDIGIT_A(c) || ((c) >= 'a' && (c) <= 'f') || ((c) <= 'F' && (c) >= 'A'))
725 #endif /* ASCII range definitions */
727 /* Latin1 definitions */
729 # define isALNUMC_L1(c) _generic_isCC(c, _CC_ALNUMC)
730 # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
731 # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
733 /* continuation character for legal NAME in \N{NAME} */
734 # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
736 # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
737 # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
738 # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
739 # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
740 # define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
741 # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
742 # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
743 # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
744 # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
745 # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
746 #else /* No access to perl.h. Only a few provided here, just in case needed
747 * for backwards compatibility */
748 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
749 * >= AA test to speed up ASCII-only tests at the expense of the others */
750 # define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
751 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
752 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
753 || NATIVE_TO_UNI((U8) c) == 0xAA \
754 || NATIVE_TO_UNI((U8) c) == 0xB5 \
755 || NATIVE_TO_UNI((U8) c) == 0xBA)))
756 # define isCHARNAME_CONT(c) (isALNUM_L1(c) || (c) == ' ' || (c) == '-' || (c) == '(' || (c) == ')' || (c) == ':' || NATIVE_TO_UNI((U8) c) == 0xA0)
759 /* Macros for backwards compatibility and for completeness when the ASCII and
760 * Latin1 values are identical */
761 #define isALNUM(c) isWORDCHAR(c)
762 #define isALNUMU(c) isWORDCHAR_L1(c)
763 #define isALPHAU(c) isALPHA_L1(c)
764 #define isDIGIT_L1(c) isDIGIT_A(c)
765 #define isOCTAL(c) isOCTAL_A(c)
766 #define isOCTAL_L1(c) isOCTAL_A(c)
767 #define isXDIGIT_L1(c) isXDIGIT_A(c)
769 /* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
770 * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
771 * therefore can use native functions. For those where C89 doesn't define a
772 * function, use our function, assuming that the EBCDIC code page is isomorphic
773 * with Latin1, which the three currently recognized by Perl are. Some libc's
774 * have an isblank(), but it's not guaranteed. */
776 # define isALNUMC(c) isalnum(c)
777 # define isALPHA(c) isalpha(c)
778 # define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
779 # define isCNTRL(c) iscntrl(c)
780 # define isDIGIT(c) isdigit(c)
781 # define isGRAPH(c) isgraph(c)
782 # define isIDFIRST(c) (isALPHA(c) || (c) == '_')
783 # define isLOWER(c) islower(c)
784 # define isPRINT(c) isprint(c)
785 # define isPSXSPC(c) isspace(c)
786 # define isPUNCT(c) ispunct(c)
787 # define isSPACE(c) (isPSXSPC(c) && (c) != '\v')
788 # define isUPPER(c) isupper(c)
789 # define isXDIGIT(c) isxdigit(c)
790 # define isWORDCHAR(c) (isalnum(c) || (c) == '_')
791 # define toLOWER(c) tolower(c)
792 # define toUPPER(c) toupper(c)
793 #else /* Not EBCDIC: ASCII-only matching */
794 # define isALNUMC(c) isALNUMC_A(c) /* Mnemonic: "C's alnum" = alpha + digit */
795 # define isALPHA(c) isALPHA_A(c)
796 # define isBLANK(c) isBLANK_A(c)
797 # define isCNTRL(c) isCNTRL_A(c)
798 # define isDIGIT(c) isDIGIT_A(c)
799 # define isGRAPH(c) isGRAPH_A(c)
800 # define isIDFIRST(c) isIDFIRST_A(c)
801 # define isLOWER(c) isLOWER_A(c)
802 # define isPRINT(c) isPRINT_A(c)
803 # define isPSXSPC(c) isPSXSPC_A(c)
804 # define isPUNCT(c) isPUNCT_A(c)
805 # define isSPACE(c) isSPACE_A(c)
806 # define isUPPER(c) isUPPER_A(c)
807 # define isWORDCHAR(c) isWORDCHAR_A(c)
808 # define isXDIGIT(c) isXDIGIT_A(c)
810 /* ASCII casing. These could also be written as
811 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
812 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
813 which uses table lookup and mask instead of subtraction. (This would
814 work because the _MOD does not apply in the ASCII range) */
815 # define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
816 # define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
820 /* Use table lookup for speed; return error character for input
822 #define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \
823 ? UNI_TO_NATIVE(PL_latin1_lc[ \
824 NATIVE_TO_UNI( (U8) (c)) ]) \
825 : UNICODE_REPLACEMENT)
826 /* Modified uc. Is correct uc except for three non-ascii chars which are
827 * all mapped to one of them, and these need special handling; error
828 * character for input out-of-range */
829 #define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \
830 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \
831 NATIVE_TO_UNI( (U8) (c)) ]) \
832 : UNICODE_REPLACEMENT)
834 #ifdef USE_NEXT_CTYPE
836 # define isALNUM_LC(c) \
837 (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
838 # define isIDFIRST_LC(c) \
839 (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
840 # define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
841 # define isASCII_LC(c) isASCII((unsigned int)(c))
842 # define isBLANK_LC(c) isBLANK((unsigned int)(c))
843 # define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
844 # define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
845 # define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
846 # define isLOWER_LC(c) NXIsLower((unsigned int)(c))
847 # define isALNUMC_LC(c) NXIsAlNum((unsigned int)(c))
848 # define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
849 # define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
850 # define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
851 # define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
852 # define toUPPER_LC(c) NXToUpper((unsigned int)(c))
853 # define toLOWER_LC(c) NXToLower((unsigned int)(c))
855 #else /* !USE_NEXT_CTYPE */
857 # if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
859 /* Note that the foo_LC() macros in this case generally are defined only on
860 * code points 0-256, and give undefined, unwarned results if called with
861 * values outside that range */
863 # define isALNUM_LC(c) (isalnum((unsigned char)(c)) || (char)(c) == '_')
864 # define isIDFIRST_LC(c) (isalpha((unsigned char)(c)) || (char)(c) == '_')
865 # define isALPHA_LC(c) isalpha((unsigned char)(c))
867 # define isASCII_LC(c) isascii((unsigned char)(c))
869 # define isASCII_LC(c) isASCII((unsigned char)(c))
872 # define isBLANK_LC(c) isblank((unsigned char)(c))
874 # define isBLANK_LC(c) isBLANK((unsigned char)(c))
876 # define isSPACE_LC(c) isspace((unsigned char)(c))
877 # define isDIGIT_LC(c) isdigit((unsigned char)(c))
878 # define isUPPER_LC(c) isupper((unsigned char)(c))
879 # define isLOWER_LC(c) islower((unsigned char)(c))
880 # define isALNUMC_LC(c) isalnum((unsigned char)(c))
881 # define isCNTRL_LC(c) iscntrl((unsigned char)(c))
882 # define isGRAPH_LC(c) isgraph((unsigned char)(c))
883 # define isPRINT_LC(c) isprint((unsigned char)(c))
884 # define isPUNCT_LC(c) ispunct((unsigned char)(c))
885 # define toUPPER_LC(c) toupper((unsigned char)(c))
886 # define toLOWER_LC(c) tolower((unsigned char)(c))
890 # define isALNUM_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
891 # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
892 # define isALPHA_LC(c) (isascii(c) && isalpha(c))
893 # define isASCII_LC(c) isascii(c)
895 # define isBLANK_LC(c) (isascii(c) && isblank(c))
897 # define isBLANK_LC(c) isBLANK(c)
899 # define isSPACE_LC(c) (isascii(c) && isspace(c))
900 # define isDIGIT_LC(c) (isascii(c) && isdigit(c))
901 # define isUPPER_LC(c) (isascii(c) && isupper(c))
902 # define isLOWER_LC(c) (isascii(c) && islower(c))
903 # define isALNUMC_LC(c) (isascii(c) && isalnum(c))
904 # define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
905 # define isGRAPH_LC(c) (isascii(c) && isgraph(c))
906 # define isPRINT_LC(c) (isascii(c) && isprint(c))
907 # define isPUNCT_LC(c) (isascii(c) && ispunct(c))
908 # define toUPPER_LC(c) toupper(c)
909 # define toLOWER_LC(c) tolower(c)
912 #endif /* USE_NEXT_CTYPE */
914 #define isPSXSPC_LC(c) (isSPACE_LC(c) || (c) == '\v')
916 /* For internal core Perl use only. If the input is Latin1, use the Latin1
917 * macro; otherwise use the function. Won't compile if 'c' isn't unsigned, as
918 * won't match function prototype. The macros do bounds checking, so have
919 * duplicate checks here, so could create versions of the macros that don't,
920 * but experiments show that gcc optimizes them out anyway. */
922 /* Note that all ignore 'use bytes' */
923 #define _generic_uni(classnum, function, c) ((c) < 256 \
924 ? _generic_isCC(c, classnum) \
927 #define isWORDCHAR_uni(c) _generic_uni(_CC_WORDCHAR, is_uni_alnum, c)
928 #define isALNUM_uni(c) isWORDCHAR_uni(c)
929 #define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
930 #define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, is_uni_idfirst, c)
931 #define isALPHA_uni(c) _generic_uni(_CC_ALPHA, is_uni_alpha, c)
932 #define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
933 #define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
934 #define isDIGIT_uni(c) _generic_uni(_CC_DIGIT, is_uni_digit, c)
935 #define isUPPER_uni(c) _generic_uni(_CC_UPPER, is_uni_upper, c)
936 #define isLOWER_uni(c) _generic_uni(_CC_LOWER, is_uni_lower, c)
937 #define isASCII_uni(c) isASCII(c)
939 /* All controls are in Latin1 */
940 #define isCNTRL_uni(c) isCNTRL_L1(c)
942 #define isGRAPH_uni(c) _generic_uni(_CC_GRAPH, is_uni_graph, c)
943 #define isPRINT_uni(c) _generic_uni(_CC_PRINT, is_uni_print, c)
944 #define isPUNCT_uni(c) _generic_uni(_CC_PUNCT, is_uni_punct, c)
945 #define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
947 /* Posix and regular space differ only in U+000B, which is in Latin1 */
948 #define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, \
949 is_XPERLSPACE_cp_high, c)
951 #define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
952 #define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
953 #define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
954 #define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
956 #define isALNUM_LC_uvchr(c) (c < 256 ? isALNUM_LC(c) : is_uni_alnum_lc(c))
957 #define isIDFIRST_LC_uvchr(c) (c < 256 ? isIDFIRST_LC(c) : is_uni_idfirst_lc(c))
958 #define isALPHA_LC_uvchr(c) (c < 256 ? isALPHA_LC(c) : is_uni_alpha_lc(c))
959 #define isSPACE_LC_uvchr(c) (c < 256 ? isSPACE_LC(c) : is_uni_space_lc(c))
960 #define isDIGIT_LC_uvchr(c) (c < 256 ? isDIGIT_LC(c) : is_uni_digit_lc(c))
961 #define isUPPER_LC_uvchr(c) (c < 256 ? isUPPER_LC(c) : is_uni_upper_lc(c))
962 #define isLOWER_LC_uvchr(c) (c < 256 ? isLOWER_LC(c) : is_uni_lower_lc(c))
963 #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : is_uni_cntrl_lc(c))
964 #define isGRAPH_LC_uvchr(c) (c < 256 ? isGRAPH_LC(c) : is_uni_graph_lc(c))
965 #define isPRINT_LC_uvchr(c) (c < 256 ? isPRINT_LC(c) : is_uni_print_lc(c))
966 #define isPUNCT_LC_uvchr(c) (c < 256 ? isPUNCT_LC(c) : is_uni_punct_lc(c))
968 #define isPSXSPC_LC_uni(c) (isSPACE_LC_uni(c) ||(c) == '\f')
969 #define isBLANK_LC_uni(c) isBLANK(c) /* could be wrong */
971 /* For internal core Perl use only. If the input is in the Latin1 range, use
972 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
973 * Otherwise use the value given by the 'utf8' parameter. This relies on the
974 * fact that ASCII characters have the same representation whether utf8 or not.
975 * Note that it assumes that the utf8 has been validated, and ignores 'use
977 #define _generic_utf8_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
978 ? _generic_isCC(*(p), classnum) \
979 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
981 TWO_BYTE_UTF8_TO_UNI(*(p), \
985 /* Like the above, but calls 'function(p)' to get the utf8 value */
986 #define _generic_utf8(classnum, function, p) \
987 _generic_utf8_utf8(classnum, p, function(p))
989 #define isWORDCHAR_utf8(p) _generic_utf8(_CC_WORDCHAR, is_utf8_alnum, p)
990 #define isALNUM_utf8(p) isWORDCHAR_utf8(p) /* back compat */
992 /* To prevent S_scan_word in toke.c from hanging, we have to make sure that
993 * IDFIRST is an alnum. See
994 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
995 * ever wanted to know about. XXX It is unclear if this should extend to
996 * isIDFIRST_uni() which it hasn't so far. (In the ASCII range, there isn't a
997 * difference.) This used to be not the XID version, but we decided to go with
998 * the more modern Unicode definition */
999 #define isIDFIRST_utf8(p) _generic_utf8(_CC_IDFIRST, \
1000 _is_utf8__perl_idstart, p)
1002 #define isIDCONT_utf8(p) _generic_utf8(_CC_WORDCHAR, is_utf8_xidcont, p)
1003 #define isALPHA_utf8(p) _generic_utf8(_CC_ALPHA, is_utf8_alpha, p)
1004 #define isBLANK_utf8(p) _generic_utf8(_CC_BLANK, is_HORIZWS_high, p)
1005 #define isSPACE_utf8(p) _generic_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
1006 #define isVERTWS_utf8(p) _generic_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
1007 #define isDIGIT_utf8(p) _generic_utf8(_CC_DIGIT, is_utf8_digit, p)
1008 #define isUPPER_utf8(p) _generic_utf8(_CC_UPPER, is_utf8_upper, p)
1009 #define isLOWER_utf8(p) _generic_utf8(_CC_LOWER, is_utf8_lower, p)
1011 /* Because ASCII is invariant under utf8, the non-utf8 macro works */
1012 #define isASCII_utf8(p) isASCII(*p)
1014 #define isCNTRL_utf8(p) _generic_utf8_utf8(_CC_CNTRL, p, 0)
1015 #define isGRAPH_utf8(p) _generic_utf8(_CC_GRAPH, is_utf8_graph, p)
1016 #define isPRINT_utf8(p) _generic_utf8(_CC_PRINT, is_utf8_print, p)
1017 #define isPUNCT_utf8(p) _generic_utf8(_CC_PUNCT, is_utf8_punct, p)
1018 #define isXDIGIT_utf8(p) _generic_utf8(_CC_XDIGIT, is_XDIGIT_high, p)
1019 #define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
1020 #define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1021 #define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1023 /* Posix and regular space differ only in U+000B, which is in ASCII (and hence
1025 #define isPSXSPC_utf8(p) _generic_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
1027 #define isALNUM_LC_utf8(p) isALNUM_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1028 #define isIDFIRST_LC_utf8(p) isIDFIRST_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1029 #define isALPHA_LC_utf8(p) isALPHA_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1030 #define isBLANK_LC_utf8(p) isBLANK_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1031 #define isSPACE_LC_utf8(p) isSPACE_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1032 #define isDIGIT_LC_utf8(p) isDIGIT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1033 #define isUPPER_LC_utf8(p) isUPPER_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1034 #define isLOWER_LC_utf8(p) isLOWER_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1035 #define isALNUMC_LC_utf8(p) isALNUMC_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1036 #define isCNTRL_LC_utf8(p) isCNTRL_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1037 #define isGRAPH_LC_utf8(p) isGRAPH_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1038 #define isPRINT_LC_utf8(p) isPRINT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1039 #define isPUNCT_LC_utf8(p) isPUNCT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1041 #define isPSXSPC_LC_utf8(c) (isSPACE_LC_utf8(c) ||(c) == '\f')
1043 /* This conversion works both ways, strangely enough. On EBCDIC platforms,
1044 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1045 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1046 * that is EOT on ASCII; ST on EBCDIC */
1047 # define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64)
1049 /* Line numbers are unsigned, 32 bits. */
1051 #define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
1053 /* Helpful alias for version prescan */
1054 #define is_LAX_VERSION(a,b) \
1055 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1057 #define is_STRICT_VERSION(a,b) \
1058 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1060 #define BADVERSION(a,b,c) \
1066 #define READ_XDIGIT(s) (isALPHA(*(s)) ? ((*(s)++ + 9) & 0xf) : (*(s)++ & 0xf))
1069 =head1 Memory Management
1071 =for apidoc Am|void|Newx|void* ptr|int nitems|type
1072 The XSUB-writer's interface to the C C<malloc> function.
1074 In 5.9.3, Newx() and friends replace the older New() API, and drops
1075 the first parameter, I<x>, a debug aid which allowed callers to identify
1076 themselves. This aid has been superseded by a new build option,
1077 PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
1078 there for use in XS modules supporting older perls.
1080 =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
1081 The XSUB-writer's interface to the C C<malloc> function, with
1082 cast. See also C<Newx>.
1084 =for apidoc Am|void|Newxz|void* ptr|int nitems|type
1085 The XSUB-writer's interface to the C C<malloc> function. The allocated
1086 memory is zeroed with C<memzero>. See also C<Newx>.
1088 =for apidoc Am|void|Renew|void* ptr|int nitems|type
1089 The XSUB-writer's interface to the C C<realloc> function.
1091 =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1092 The XSUB-writer's interface to the C C<realloc> function, with
1095 =for apidoc Am|void|Safefree|void* ptr
1096 The XSUB-writer's interface to the C C<free> function.
1098 =for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1099 The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
1100 source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
1101 the type. Can do overlapping moves. See also C<Copy>.
1103 =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1104 Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1107 =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1108 The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
1109 source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
1110 the type. May fail on overlapping copies. See also C<Move>.
1112 =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1114 Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1117 =for apidoc Am|void|Zero|void* dest|int nitems|type
1119 The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1120 destination, C<nitems> is the number of items, and C<type> is the type.
1122 =for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1124 Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1127 =for apidoc Am|void|StructCopy|type *src|type *dest|type
1128 This is an architecture-independent macro to copy one structure to another.
1130 =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1132 Fill up memory with a byte pattern (a byte repeated over and over
1133 again) that hopefully catches attempts to access uninitialized memory.
1135 =for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1137 PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1139 =for apidoc Am|void|PoisonFree|void* dest|int nitems|type
1141 PoisonWith(0xEF) for catching access to freed memory.
1143 =for apidoc Am|void|Poison|void* dest|int nitems|type
1145 PoisonWith(0xEF) for catching access to freed memory.
1149 /* Maintained for backwards-compatibility only. Use newSV() instead. */
1151 #define NEWSV(x,len) newSV(len)
1154 #define MEM_SIZE_MAX ((MEM_SIZE)~0)
1156 /* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1157 * overly eager compilers that will bleat about e.g.
1158 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
1159 #ifdef PERL_MALLOC_WRAP
1160 #define MEM_WRAP_CHECK(n,t) \
1161 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
1162 #define MEM_WRAP_CHECK_1(n,t,a) \
1163 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
1164 #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
1166 #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)))
1169 #define MEM_WRAP_CHECK(n,t)
1170 #define MEM_WRAP_CHECK_1(n,t,a)
1171 #define MEM_WRAP_CHECK_2(n,t,a,b)
1172 #define MEM_WRAP_CHECK_(n,t)
1174 #define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1180 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
1181 * go through functions, which are handy for debugging breakpoints, but
1182 * which more importantly get the immediate calling environment (file and
1183 * line number, and C function name if available) passed in. This info can
1184 * then be used for logging the calls, for which one gets a sample
1185 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
1188 * - not all memory allocs get logged, only those
1189 * that go through Newx() and derivatives (while all
1190 * Safefrees do get logged)
1191 * - __FILE__ and __LINE__ do not work everywhere
1192 * - __func__ or __FUNCTION__ even less so
1193 * - I think more goes on after the perlio frees but
1194 * the thing is that STDERR gets closed (as do all
1195 * the file descriptors)
1196 * - no deeper calling stack than the caller of the Newx()
1197 * or the kind, but do I look like a C reflection/introspection
1199 * - the function prototypes for the logging functions
1200 * probably should maybe be somewhere else than handy.h
1201 * - one could consider inlining (macrofying) the logging
1202 * for speed, but I am too lazy
1203 * - one could imagine recording the allocations in a hash,
1204 * (keyed by the allocation address?), and maintain that
1205 * through reallocs and frees, but how to do that without
1206 * any News() happening...?
1207 * - lots of -Ddefines to get useful/controllable output
1208 * - lots of ENV reads
1211 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);
1213 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);
1215 PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
1218 # ifndef PERL_MEM_LOG_NOIMPL
1227 # if defined(PERL_IN_SV_C) /* those are only used in sv.c */
1228 void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1229 void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1236 #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1237 #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
1238 #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
1241 #ifndef MEM_LOG_ALLOC
1242 #define MEM_LOG_ALLOC(n,t,a) (a)
1244 #ifndef MEM_LOG_REALLOC
1245 #define MEM_LOG_REALLOC(n,t,v,a) (a)
1247 #ifndef MEM_LOG_FREE
1248 #define MEM_LOG_FREE(a) (a)
1251 #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1252 #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1253 #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
1256 /* pre 5.9.x compatibility */
1257 #define New(x,v,n,t) Newx(v,n,t)
1258 #define Newc(x,v,n,t,c) Newxc(v,n,t,c)
1259 #define Newz(x,v,n,t) Newxz(v,n,t)
1262 #define Renew(v,n,t) \
1263 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1264 #define Renewc(v,n,t,c) \
1265 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1268 #define Safefree(d) \
1269 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
1271 #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
1274 #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1275 #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1276 #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
1278 #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1279 #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1281 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
1283 /* Using bzero(), which returns void. */
1284 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
1287 #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1288 #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1289 #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1290 #define Poison(d,n,t) PoisonFree(d,n,t)
1292 #ifdef USE_STRUCT_COPY
1293 #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
1295 #define StructCopy(s,d,t) Copy(s,d,1,t)
1298 #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
1299 #define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
1303 # define Perl_va_copy(s, d) va_copy(d, s)
1305 # if defined(__va_copy)
1306 # define Perl_va_copy(s, d) __va_copy(d, s)
1308 # define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1313 /* convenience debug macros */
1315 #define pTHX_FORMAT "Perl interpreter: 0x%p"
1316 #define pTHX__FORMAT ", Perl interpreter: 0x%p"
1317 #define pTHX_VALUE_ (void *)my_perl,
1318 #define pTHX_VALUE (void *)my_perl
1319 #define pTHX__VALUE_ ,(void *)my_perl,
1320 #define pTHX__VALUE ,(void *)my_perl
1323 #define pTHX__FORMAT
1326 #define pTHX__VALUE_
1328 #endif /* USE_ITHREADS */
1330 /* Perl_deprecate was not part of the public API, and did not have a deprecate()
1331 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1332 CPAN::Unpack show any users outside the core. */
1334 # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
1339 * c-indentation-style: bsd
1341 * indent-tabs-mode: nil
1344 * ex: set ts=8 sts=4 sw=4 et: