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__)
23 # define Null(type) ((type)NULL)
28 =for apidoc AmU||Nullch
29 Null character pointer. (No longer available when C<PERL_CORE> is defined.)
31 =for apidoc AmU||Nullsv
32 Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
37 # define Nullch Null(char*)
38 # define Nullfp Null(PerlIO*)
39 # define Nullsv Null(SV*)
51 /* The MUTABLE_*() macros cast pointers to the types shown, in such a way
52 * (compiler permitting) that casting away const-ness will give a warning;
56 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
57 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
60 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
61 # define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
63 # define MUTABLE_PTR(p) ((void *) (p))
66 #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
67 #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
68 #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
69 #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
70 #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
71 #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
80 /* bool is built-in for g++-2.6.3 and later, which might be used
81 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
82 be sure _G_config.h will be included before this file. _G_config.h
83 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
84 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
85 g++ can be identified by __GNUG__.
86 Andy Dougherty February 2000
88 #ifdef __GNUG__ /* GNU g++ has bool built-in */
94 /* The NeXT dynamic loader headers will not build with the bool macro
95 So declare them now to clear confusion.
97 #if defined(NeXT) || defined(__NeXT__)
100 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
104 # endif /* !HAS_BOOL */
105 #endif /* NeXT || __NeXT__ */
108 # if defined(UTS) || defined(VMS)
116 /* a simple (bool) cast may not do the right thing: if bool is defined
117 * as char for example, then the cast from int is implementation-defined
118 * (bool)!!(cbool) in a ternary triggers a bug in xlc on AIX
121 #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
123 /* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
124 * XXX Should really be a Configure probe, with HAS__FUNCTION__
125 * and FUNCTION__ as results.
126 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
127 #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
128 # define FUNCTION__ __func__
130 # if (defined(_MSC_VER) && _MSC_VER < 1300) || /* Pre-MSVC 7.0 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
131 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
132 # define FUNCTION__ ""
134 # define FUNCTION__ __FUNCTION__ /* Common extension. */
138 /* XXX A note on the perl source internal type system. The
139 original intent was that I32 be *exactly* 32 bits.
141 Currently, we only guarantee that I32 is *at least* 32 bits.
142 Specifically, if int is 64 bits, then so is I32. (This is the case
143 for the Cray.) This has the advantage of meshing nicely with
144 standard library calls (where we pass an I32 and the library is
145 expecting an int), but the disadvantage that an I32 is not 32 bits.
146 Andy Dougherty August 1996
148 There is no guarantee that there is *any* integral type with
149 exactly 32 bits. It is perfectly legal for a system to have
150 sizeof(short) == sizeof(int) == sizeof(long) == 8.
152 Similarly, there is no guarantee that I16 and U16 have exactly 16
155 For dealing with issues that may arise from various 32/64-bit
156 systems, we will ask Configure to check out
158 SHORTSIZE == sizeof(short)
159 INTSIZE == sizeof(int)
160 LONGSIZE == sizeof(long)
161 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
162 PTRSIZE == sizeof(void *)
163 DOUBLESIZE == sizeof(double)
164 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
168 #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
169 # include <inttypes.h>
170 # ifdef INT32_MIN_BROKEN
172 # define INT32_MIN (-2147483647-1)
174 # ifdef INT64_MIN_BROKEN
176 # define INT64_MIN (-9223372036854775807LL-1)
191 #endif /* PERL_CORE */
193 #if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
194 # if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
195 # define PeRl_INT64_C(c) CAT2(c,LL)
196 # define PeRl_UINT64_C(c) CAT2(c,ULL)
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 # if defined(_WIN64) && defined(_MSC_VER)
203 # define PeRl_INT64_C(c) CAT2(c,I64)
204 # define PeRl_UINT64_C(c) CAT2(c,UI64)
206 # define PeRl_INT64_C(c) ((I64TYPE)(c))
207 # define PeRl_UINT64_C(c) ((U64TYPE)(c))
212 # define UINT64_C(c) PeRl_UINT64_C(c)
215 # define INT64_C(c) PeRl_INT64_C(c)
219 #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
221 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
222 Please search CHAR_MAX in perl.h for further details. */
223 #define U8_MAX UINT8_MAX
224 #define U8_MIN UINT8_MIN
226 #define I16_MAX INT16_MAX
227 #define I16_MIN INT16_MIN
228 #define U16_MAX UINT16_MAX
229 #define U16_MIN UINT16_MIN
231 #define I32_MAX INT32_MAX
232 #define I32_MIN INT32_MIN
233 #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
234 # define U32_MAX UINT32_MAX
236 # define U32_MAX 4294967295U
238 #define U32_MIN UINT32_MIN
242 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
243 Please search CHAR_MAX in perl.h for further details. */
244 #define U8_MAX PERL_UCHAR_MAX
245 #define U8_MIN PERL_UCHAR_MIN
247 #define I16_MAX PERL_SHORT_MAX
248 #define I16_MIN PERL_SHORT_MIN
249 #define U16_MAX PERL_USHORT_MAX
250 #define U16_MIN PERL_USHORT_MIN
253 # define I32_MAX PERL_INT_MAX
254 # define I32_MIN PERL_INT_MIN
255 # define U32_MAX PERL_UINT_MAX
256 # define U32_MIN PERL_UINT_MIN
258 # define I32_MAX PERL_LONG_MAX
259 # define I32_MIN PERL_LONG_MIN
260 # define U32_MAX PERL_ULONG_MAX
261 # define U32_MIN PERL_ULONG_MIN
266 /* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
267 #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
268 #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
269 #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
271 #define Ctl(ch) ((ch) & 037)
274 =head1 SV-Body Allocation
276 =for apidoc Ama|SV*|newSVpvs|const char* s
277 Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
279 =for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
280 Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
283 =for apidoc Ama|SV*|newSVpvs_share|const char* s
284 Like C<newSVpvn_share>, but takes a literal string instead of a string/length
285 pair and omits the hash parameter.
287 =for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
288 Like C<sv_catpvn_flags>, but takes a literal string instead of a
291 =for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
292 Like C<sv_catpvn_nomg>, but takes a literal string instead of a
295 =for apidoc Am|void|sv_catpvs|SV* sv|const char* s
296 Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
298 =for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
299 Like C<sv_catpvn_mg>, but takes a literal string instead of a
302 =for apidoc Am|void|sv_setpvs|SV* sv|const char* s
303 Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
305 =for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
306 Like C<sv_setpvn_mg>, but takes a literal string instead of a
309 =for apidoc Am|SV *|sv_setref_pvs|const char* s
310 Like C<sv_setref_pvn>, but takes a literal string instead of a
313 =head1 Memory Management
315 =for apidoc Ama|char*|savepvs|const char* s
316 Like C<savepvn>, but takes a literal string instead of a string/length pair.
318 =for apidoc Ama|char*|savesharedpvs|const char* s
319 A version of C<savepvs()> which allocates the duplicate string in memory
320 which is shared between threads.
324 =for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
325 Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
327 =head1 Hash Manipulation Functions
329 =for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
330 Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
332 =for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
333 Like C<hv_store>, but takes a literal string instead of a string/length pair
334 and omits the hash parameter.
336 =head1 Lexer interface
338 =for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
340 Like L</lex_stuff_pvn>, but takes a literal string instead of a
346 /* concatenating with "" ensures that only literal strings are accepted as argument */
347 #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
349 /* note that STR_WITH_LEN() can't be used as argument to macros or functions that
350 * under some configurations might be macros, which means that it requires the full
351 * Perl_xxx(aTHX_ ...) form for any API calls where it's used.
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))
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))
390 #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
392 #define get_cvs(str, flags) \
393 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
396 =head1 Miscellaneous Functions
398 =for apidoc Am|bool|strNE|char* s1|char* s2
399 Test two strings to see if they are different. Returns true or
402 =for apidoc Am|bool|strEQ|char* s1|char* s2
403 Test two strings to see if they are equal. Returns true or false.
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.
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.
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.
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.
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>).
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
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))
444 # define memNE(s1,s2,l) (memcmp(s1,s2,l))
445 # define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
447 # define memNE(s1,s2,l) (bcmp(s1,s2,l))
448 # define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
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)
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.
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".
468 #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
476 =head1 Character classes
477 There are three variants for all the functions in this section. The base ones
478 operate using the character set of the platform Perl is running on. The ones
479 with an C<_A> suffix operate on the ASCII character set, and the ones with an
480 C<_L1> suffix operate on the full Latin1 character set. All are unaffected by
481 locale and by C<use bytes>.
483 For ASCII platforms, the base function with no suffix and the one with the
484 C<_A> suffix are identical. The function with the C<_L1> suffix imposes the
485 Latin-1 character set onto the platform. That is, the code points that are
486 ASCII are unaffected, since ASCII is a subset of Latin-1. But the non-ASCII
487 code points are treated as if they are Latin-1 characters. For example,
488 C<isSPACE_L1()> will return true when called with the code point 0xA0, which is
489 the Latin-1 NO-BREAK SPACE.
491 For EBCDIC platforms, the base function with no suffix and the one with the
492 C<_L1> suffix should be identical, since, as of this writing, the EBCDIC code
493 pages that Perl knows about all are equivalent to Latin-1. The function that
494 ends in an C<_A> suffix will not return true unless the specified character also
495 has an ASCII equivalent.
497 =for apidoc Am|bool|isALPHA|char ch
498 Returns a boolean indicating whether the specified character is an
499 alphabetic character in the platform's native character set.
500 See the L<top of this section|/Character classes> for an explanation of variants
501 C<isALPHA_A> and C<isALPHA_L1>.
503 =for apidoc Am|bool|isASCII|char ch
504 Returns a boolean indicating whether the specified character is one of the 128
505 characters in the ASCII character set. On non-ASCII platforms, it is if this
506 character corresponds to an ASCII character. Variants C<isASCII_A()> and
507 C<isASCII_L1()> are identical to C<isASCII()>.
509 =for apidoc Am|bool|isDIGIT|char ch
510 Returns a boolean indicating whether the specified character is a
511 digit in the platform's native character set.
512 Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
514 =for apidoc Am|bool|isLOWER|char ch
515 Returns a boolean indicating whether the specified character is a
516 lowercase character in the platform's native character set.
517 See the L<top of this section|/Character classes> for an explanation of variants
518 C<isLOWER_A> and C<isLOWER_L1>.
520 =for apidoc Am|bool|isOCTAL|char ch
521 Returns a boolean indicating whether the specified character is an
522 octal digit, [0-7] in the platform's native character set.
523 Variants C<isOCTAL_A> and C<isOCTAL_L1> are identical to C<isOCTAL>.
525 =for apidoc Am|bool|isSPACE|char ch
526 Returns a boolean indicating whether the specified character is a
527 whitespace character in the platform's native character set. This is the same
528 as what C<\s> matches in a regular expression.
529 See the L<top of this section|/Character classes> for an explanation of variants
530 C<isSPACE_A> and C<isSPACE_L1>.
532 =for apidoc Am|bool|isUPPER|char ch
533 Returns a boolean indicating whether the specified character is an
534 uppercase character in the platform's native character set.
535 See the L<top of this section|/Character classes> for an explanation of variants
536 C<isUPPER_A> and C<isUPPER_L1>.
538 =for apidoc Am|bool|isWORDCHAR|char ch
539 Returns a boolean indicating whether the specified character is a
540 character that is any of: alphabetic, numeric, or an underscore. This is the
541 same as what C<\w> matches in a regular expression.
542 C<isALNUM()> is a synonym provided for backward compatibility. Note that it
543 does not have the standard C language meaning of alphanumeric, since it matches
544 an underscore and the standard meaning does not.
545 See the L<top of this section|/Character classes> for an explanation of variants
546 C<isWORDCHAR_A> and C<isWORDCHAR_L1>.
548 =for apidoc Am|bool|isXDIGIT|char ch
549 Returns a boolean indicating whether the specified character is a hexadecimal
550 digit, [0-9A-Fa-f]. Variants C<isXDIGIT_A()> and C<isXDIGIT_L1()> are
551 identical to C<isXDIGIT()>.
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. These are used in PL_charclass[] and the ones
606 * up through the one that corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are
607 * used by regcomp.h. These use names used in l1_char_class_tab.h but their
608 * actual definitions are here. If that has a name not used here, it won't
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 _HIGHEST_REGCOMP_DOT_H_SYNC _CC_BLANK
627 # define _CC_IDFIRST 15
628 # define _CC_CHARNAME_CONT 16
629 # define _CC_NONLATIN1_FOLD 17
630 # define _CC_QUOTEMETA 18
632 * If more bits are needed, one could add a second word for non-64bit
633 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
637 EXTCONST U32 PL_charclass[] = {
638 # include "l1_char_class_tab.h"
641 # else /* ! DOINIT */
642 EXTCONST U32 PL_charclass[];
645 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
646 # define _CC_mask(classnum) (1U << (classnum))
647 # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
648 && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask(classnum)))
650 /* The mask for the _A versions of the macros; it just adds in the bit for
652 # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
654 /* The _A version makes sure that both the desired bit and the ASCII bit
656 # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
657 && ((PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask_A(classnum)) \
658 == _CC_mask_A(classnum)))
660 # define isALNUMC_A(c) _generic_isCC_A(c, _CC_ALNUMC)
661 # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
662 # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
663 # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
664 # define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
665 # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
666 # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
667 # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
668 # define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
669 # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
670 # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
671 # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
672 # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
673 # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
674 # define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST))
676 /* Either participates in a fold with a character above 255, or is a
678 # 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)))
680 # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
681 #else /* No perl.h. */
683 # define isALNUMC_A(c) (isASCII(c) && isALNUMC(c))
684 # define isALPHA_A(c) (isASCII(c) && isALPHA(c))
685 # define isBLANK_A(c) (isASCII(c) && isBLANK(c))
686 # define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
687 # define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
688 # define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
689 # define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
690 # define isLOWER_A(c) (isASCII(c) && isLOWER(c))
691 # define isPRINT_A(c) (isASCII(c) && isPRINT(c))
692 # define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
693 # define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
694 # define isSPACE_A(c) (isASCII(c) && isSPACE(c))
695 # define isUPPER_A(c) (isASCII(c) && isUPPER(c))
696 # define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
697 # define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
698 # else /* ASCII platform, no perl.h */
699 # define isALNUMC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
700 # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
701 # define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
702 # define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
703 # define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
704 # define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
705 # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
706 # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
707 # define isPRINT_A(c) (((c) >= 32 && (c) < 127))
708 # define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
709 # define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) || ((c) >= 58 && (c) <= 64) || ((c) >= 91 && (c) <= 96) || ((c) >= 123 && (c) <= 126))
710 # define isSPACE_A(c) ((c) == ' ' || (c) == '\t' || (c) == '\n' || (c) =='\r' || (c) == '\f')
711 # define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
712 # define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
713 # define isXDIGIT_A(c) (isDIGIT_A(c) || ((c) >= 'a' && (c) <= 'f') || ((c) <= 'F' && (c) >= 'A'))
715 #endif /* ASCII range definitions */
717 /* Latin1 definitions */
719 # define isALNUMC_L1(c) _generic_isCC(c, _CC_ALNUMC)
720 # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
721 # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
722 /* continuation character for legal NAME in \N{NAME} */
723 # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
724 # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
725 # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
726 # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
727 # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
728 # define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
729 # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
730 # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
731 # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
732 # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
733 # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
734 #else /* No access to perl.h. Only a few provided here, just in case needed
735 * for backwards compatibility */
736 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
737 * >= AA test to speed up ASCII-only tests at the expense of the others */
738 # define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
739 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
740 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
741 || NATIVE_TO_UNI((U8) c) == 0xAA \
742 || NATIVE_TO_UNI((U8) c) == 0xB5 \
743 || NATIVE_TO_UNI((U8) c) == 0xBA)))
744 # define isCHARNAME_CONT(c) (isALNUM_L1(c) || (c) == ' ' || (c) == '-' || (c) == '(' || (c) == ')' || (c) == ':' || NATIVE_TO_UNI((U8) c) == 0xA0)
747 /* Macros for backwards compatibility and for completeness when the ASCII and
748 * Latin1 values are identical */
749 #define isALNUM(c) isWORDCHAR(c)
750 #define isALNUMU(c) isWORDCHAR_L1(c)
751 #define isALPHAU(c) isALPHA_L1(c)
752 #define isDIGIT_L1(c) isDIGIT_A(c)
753 #define isOCTAL(c) isOCTAL_A(c)
754 #define isOCTAL_L1(c) isOCTAL_A(c)
755 #define isXDIGIT_L1(c) isXDIGIT_A(c)
757 /* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
758 * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
759 * therefore can use native functions. For those where C89 doesn't define a
760 * function, use our function, assuming that the EBCDIC code page is isomorphic
761 * with Latin1, which the three currently recognized by Perl are. Some libc's
762 * have an isblank(), but it's not guaranteed. */
764 # define isALNUMC(c) isalnum(c)
765 # define isALPHA(c) isalpha(c)
766 # define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
767 # define isCNTRL(c) iscntrl(c)
768 # define isDIGIT(c) isdigit(c)
769 # define isGRAPH(c) isgraph(c)
770 # define isIDFIRST(c) (isALPHA(c) || (c) == '_')
771 # define isLOWER(c) islower(c)
772 # define isPRINT(c) isprint(c)
773 # define isPSXSPC(c) isspace(c)
774 # define isPUNCT(c) ispunct(c)
775 # define isSPACE(c) (isPSXSPC(c) && (c) != '\v')
776 # define isUPPER(c) isupper(c)
777 # define isXDIGIT(c) isxdigit(c)
778 # define isWORDCHAR(c) (isalnum(c) || (c) == '_')
779 # define toLOWER(c) tolower(c)
780 # define toUPPER(c) toupper(c)
781 #else /* Not EBCDIC: ASCII-only matching */
782 # define isALNUMC(c) isALNUMC_A(c) /* Mnemonic: "C's alnum" = alpha + digit */
783 # define isALPHA(c) isALPHA_A(c)
784 # define isBLANK(c) isBLANK_A(c)
785 # define isCNTRL(c) isCNTRL_A(c)
786 # define isDIGIT(c) isDIGIT_A(c)
787 # define isGRAPH(c) isGRAPH_A(c)
788 # define isIDFIRST(c) isIDFIRST_A(c)
789 # define isLOWER(c) isLOWER_A(c)
790 # define isPRINT(c) isPRINT_A(c)
791 # define isPSXSPC(c) isPSXSPC_A(c)
792 # define isPUNCT(c) isPUNCT_A(c)
793 # define isSPACE(c) isSPACE_A(c)
794 # define isUPPER(c) isUPPER_A(c)
795 # define isWORDCHAR(c) isWORDCHAR_A(c)
796 # define isXDIGIT(c) isXDIGIT_A(c)
798 /* ASCII casing. These could also be written as
799 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
800 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
801 which uses table lookup and mask instead of subtraction. (This would
802 work because the _MOD does not apply in the ASCII range) */
803 # define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
804 # define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
808 /* Use table lookup for speed; return error character for input
810 #define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \
811 ? UNI_TO_NATIVE(PL_latin1_lc[ \
812 NATIVE_TO_UNI( (U8) (c)) ]) \
813 : UNICODE_REPLACEMENT)
814 /* Modified uc. Is correct uc except for three non-ascii chars which are
815 * all mapped to one of them, and these need special handling; error
816 * character for input out-of-range */
817 #define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \
818 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \
819 NATIVE_TO_UNI( (U8) (c)) ]) \
820 : UNICODE_REPLACEMENT)
822 #ifdef USE_NEXT_CTYPE
824 # define isALNUM_LC(c) \
825 (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
826 # define isIDFIRST_LC(c) \
827 (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
828 # define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
829 # define isASCII_LC(c) isASCII((unsigned int)(c))
830 # define isBLANK_LC(c) isBLANK((unsigned int)(c))
831 # define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
832 # define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
833 # define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
834 # define isLOWER_LC(c) NXIsLower((unsigned int)(c))
835 # define isALNUMC_LC(c) NXIsAlNum((unsigned int)(c))
836 # define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
837 # define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
838 # define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
839 # define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
840 # define toUPPER_LC(c) NXToUpper((unsigned int)(c))
841 # define toLOWER_LC(c) NXToLower((unsigned int)(c))
843 #else /* !USE_NEXT_CTYPE */
845 # if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
847 /* Note that the foo_LC() macros in this case generally are defined only on
848 * code points 0-256, and give undefined, unwarned results if called with
849 * values outside that range */
851 # define isALNUM_LC(c) (isalnum((unsigned char)(c)) || (char)(c) == '_')
852 # define isIDFIRST_LC(c) (isalpha((unsigned char)(c)) || (char)(c) == '_')
853 # define isALPHA_LC(c) isalpha((unsigned char)(c))
855 # define isASCII_LC(c) isascii((unsigned char)(c))
857 # define isASCII_LC(c) isASCII((unsigned char)(c))
860 # define isBLANK_LC(c) isblank((unsigned char)(c))
862 # define isBLANK_LC(c) isBLANK((unsigned char)(c))
864 # define isSPACE_LC(c) isspace((unsigned char)(c))
865 # define isDIGIT_LC(c) isdigit((unsigned char)(c))
866 # define isUPPER_LC(c) isupper((unsigned char)(c))
867 # define isLOWER_LC(c) islower((unsigned char)(c))
868 # define isALNUMC_LC(c) isalnum((unsigned char)(c))
869 # define isCNTRL_LC(c) iscntrl((unsigned char)(c))
870 # define isGRAPH_LC(c) isgraph((unsigned char)(c))
871 # define isPRINT_LC(c) isprint((unsigned char)(c))
872 # define isPUNCT_LC(c) ispunct((unsigned char)(c))
873 # define toUPPER_LC(c) toupper((unsigned char)(c))
874 # define toLOWER_LC(c) tolower((unsigned char)(c))
878 # define isALNUM_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
879 # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
880 # define isALPHA_LC(c) (isascii(c) && isalpha(c))
881 # define isASCII_LC(c) isascii(c)
883 # define isBLANK_LC(c) (isascii(c) && isblank(c))
885 # define isBLANK_LC(c) isBLANK(c)
887 # define isSPACE_LC(c) (isascii(c) && isspace(c))
888 # define isDIGIT_LC(c) (isascii(c) && isdigit(c))
889 # define isUPPER_LC(c) (isascii(c) && isupper(c))
890 # define isLOWER_LC(c) (isascii(c) && islower(c))
891 # define isALNUMC_LC(c) (isascii(c) && isalnum(c))
892 # define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
893 # define isGRAPH_LC(c) (isascii(c) && isgraph(c))
894 # define isPRINT_LC(c) (isascii(c) && isprint(c))
895 # define isPUNCT_LC(c) (isascii(c) && ispunct(c))
896 # define toUPPER_LC(c) toupper(c)
897 # define toLOWER_LC(c) tolower(c)
900 #endif /* USE_NEXT_CTYPE */
902 #define isPSXSPC_LC(c) (isSPACE_LC(c) || (c) == '\v')
904 /* For use in the macros just below. If the input is Latin1, use the Latin1
905 * (_L1) version of the macro; otherwise use the function. Won't compile if
906 * 'c' isn't unsigned, as won't match function prototype. The macros do bounds
907 * checking, so have duplicate checks here, so could create versions of the
908 * macros that don't, but experiments show that gcc optimizes them out anyway.
910 #define generic_uni(macro, function, c) ((c) < 256 \
911 ? CAT2(macro, _L1)(c) \
913 /* Note that all ignore 'use bytes' */
915 #define isALNUM_uni(c) generic_uni(isWORDCHAR, is_uni_alnum, c)
916 #define isBLANK_uni(c) generic_uni(isBLANK, is_uni_blank, c)
917 #define isIDFIRST_uni(c) generic_uni(isIDFIRST, is_uni_idfirst, c)
918 #define isALPHA_uni(c) generic_uni(isALPHA, is_uni_alpha, c)
919 #define isSPACE_uni(c) generic_uni(isSPACE, is_uni_space, c)
920 #define isDIGIT_uni(c) generic_uni(isDIGIT, is_uni_digit, c)
921 #define isUPPER_uni(c) generic_uni(isUPPER, is_uni_upper, c)
922 #define isLOWER_uni(c) generic_uni(isLOWER, is_uni_lower, c)
923 #define isASCII_uni(c) isASCII(c)
924 /* All controls are in Latin1 */
925 #define isCNTRL_uni(c) isCNTRL_L1(c)
926 #define isGRAPH_uni(c) generic_uni(isGRAPH, is_uni_graph, c)
927 #define isPRINT_uni(c) generic_uni(isPRINT, is_uni_print, c)
928 #define isPUNCT_uni(c) generic_uni(isPUNCT, is_uni_punct, c)
929 #define isXDIGIT_uni(c) generic_uni(isXDIGIT, is_uni_xdigit, c)
931 /* Posix and regular space differ only in U+000B, which is in Latin1 */
932 #define isPSXSPC_uni(c) ((c) < 256 ? isPSXSPC_L1(c) : isSPACE_uni(c))
934 #define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
935 #define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
936 #define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
937 #define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
939 #define isALNUM_LC_uvchr(c) (c < 256 ? isALNUM_LC(c) : is_uni_alnum_lc(c))
940 #define isIDFIRST_LC_uvchr(c) (c < 256 ? isIDFIRST_LC(c) : is_uni_idfirst_lc(c))
941 #define isALPHA_LC_uvchr(c) (c < 256 ? isALPHA_LC(c) : is_uni_alpha_lc(c))
942 #define isSPACE_LC_uvchr(c) (c < 256 ? isSPACE_LC(c) : is_uni_space_lc(c))
943 #define isDIGIT_LC_uvchr(c) (c < 256 ? isDIGIT_LC(c) : is_uni_digit_lc(c))
944 #define isUPPER_LC_uvchr(c) (c < 256 ? isUPPER_LC(c) : is_uni_upper_lc(c))
945 #define isLOWER_LC_uvchr(c) (c < 256 ? isLOWER_LC(c) : is_uni_lower_lc(c))
946 #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : is_uni_cntrl_lc(c))
947 #define isGRAPH_LC_uvchr(c) (c < 256 ? isGRAPH_LC(c) : is_uni_graph_lc(c))
948 #define isPRINT_LC_uvchr(c) (c < 256 ? isPRINT_LC(c) : is_uni_print_lc(c))
949 #define isPUNCT_LC_uvchr(c) (c < 256 ? isPUNCT_LC(c) : is_uni_punct_lc(c))
951 #define isPSXSPC_LC_uni(c) (isSPACE_LC_uni(c) ||(c) == '\f')
952 #define isBLANK_LC_uni(c) isBLANK(c) /* could be wrong */
954 /* For use in the macros just below. If the input is ASCII, use the ASCII (_A)
955 * version of the macro; if the input is in the upper Latin1 range, use the
956 * Latin1 (_L1) version of the macro, after converting from utf8; otherwise use
957 * the function. This relies on the fact that ASCII characters have the same
958 * representation whether utf8 or not */
959 #define generic_utf8(macro, function, p) (isASCII(*(p)) \
960 ? CAT2(CAT2(macro,_),A)(*(p)) \
961 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
963 (TWO_BYTE_UTF8_TO_UNI(*(p), \
967 /* Note that all assume that the utf8 has been validated, and ignore 'use
970 #define isALNUM_utf8(p) generic_utf8(isWORDCHAR, is_utf8_alnum, p)
971 /* To prevent S_scan_word in toke.c from hanging, we have to make sure that
972 * IDFIRST is an alnum. See
973 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
974 * ever wanted to know about. XXX It is unclear if this should extend to
975 * isIDFIRST_uni() which it hasn't so far. (In the ASCII range, there isn't a
976 * difference.) This used to be not the XID version, but we decided to go with
977 * the more modern Unicode definition */
978 #define isIDFIRST_utf8(p) (isASCII(*(p)) \
979 ? isIDFIRST_A(*(p)) \
980 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
981 ? isIDFIRST_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \
983 : Perl__is_utf8__perl_idstart(aTHX_ p))
984 #define isIDCONT_utf8(p) generic_utf8(isWORDCHAR, is_utf8_xidcont, p)
985 #define isALPHA_utf8(p) generic_utf8(isALPHA, is_utf8_alpha, p)
986 #define isBLANK_utf8(p) generic_utf8(isBLANK, is_utf8_blank, p)
987 #define isSPACE_utf8(p) generic_utf8(isSPACE, is_utf8_space, p)
988 #define isDIGIT_utf8(p) generic_utf8(isDIGIT, is_utf8_digit, p)
989 #define isUPPER_utf8(p) generic_utf8(isUPPER, is_utf8_upper, p)
990 #define isLOWER_utf8(p) generic_utf8(isLOWER, is_utf8_lower, p)
991 /* Because ASCII is invariant under utf8, the non-utf8 macro works */
992 #define isASCII_utf8(p) isASCII(*p)
993 #define isCNTRL_utf8(p) generic_utf8(isCNTRL, is_utf8_cntrl, p)
994 #define isGRAPH_utf8(p) generic_utf8(isGRAPH, is_utf8_graph, p)
995 #define isPRINT_utf8(p) generic_utf8(isPRINT, is_utf8_print, p)
996 #define isPUNCT_utf8(p) generic_utf8(isPUNCT, is_utf8_punct, p)
997 #define isXDIGIT_utf8(p) generic_utf8(isXDIGIT, is_utf8_xdigit, p)
998 #define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
999 #define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1000 #define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1002 /* Posix and regular space differ only in U+000B, which is in ASCII (and hence
1004 #define isPSXSPC_utf8(p) ((isASCII(*(p))) \
1005 ? isPSXSPC_A(*(p)) \
1006 : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \
1007 ? isPSXSPC_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \
1010 #define isALNUM_LC_utf8(p) isALNUM_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1011 #define isIDFIRST_LC_utf8(p) isIDFIRST_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1012 #define isALPHA_LC_utf8(p) isALPHA_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1013 #define isBLANK_LC_utf8(p) isBLANK_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1014 #define isSPACE_LC_utf8(p) isSPACE_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1015 #define isDIGIT_LC_utf8(p) isDIGIT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1016 #define isUPPER_LC_utf8(p) isUPPER_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1017 #define isLOWER_LC_utf8(p) isLOWER_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1018 #define isALNUMC_LC_utf8(p) isALNUMC_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1019 #define isCNTRL_LC_utf8(p) isCNTRL_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1020 #define isGRAPH_LC_utf8(p) isGRAPH_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1021 #define isPRINT_LC_utf8(p) isPRINT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1022 #define isPUNCT_LC_utf8(p) isPUNCT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1024 #define isPSXSPC_LC_utf8(c) (isSPACE_LC_utf8(c) ||(c) == '\f')
1026 /* This conversion works both ways, strangely enough. On EBCDIC platforms,
1027 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII */
1028 # define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64)
1030 /* Line numbers are unsigned, 32 bits. */
1032 #define NOLINE ((line_t) 4294967295UL)
1034 /* Helpful alias for version prescan */
1035 #define is_LAX_VERSION(a,b) \
1036 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1038 #define is_STRICT_VERSION(a,b) \
1039 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1041 #define BADVERSION(a,b,c) \
1048 =head1 Memory Management
1050 =for apidoc Am|void|Newx|void* ptr|int nitems|type
1051 The XSUB-writer's interface to the C C<malloc> function.
1053 In 5.9.3, Newx() and friends replace the older New() API, and drops
1054 the first parameter, I<x>, a debug aid which allowed callers to identify
1055 themselves. This aid has been superseded by a new build option,
1056 PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
1057 there for use in XS modules supporting older perls.
1059 =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
1060 The XSUB-writer's interface to the C C<malloc> function, with
1061 cast. See also C<Newx>.
1063 =for apidoc Am|void|Newxz|void* ptr|int nitems|type
1064 The XSUB-writer's interface to the C C<malloc> function. The allocated
1065 memory is zeroed with C<memzero>. See also C<Newx>.
1067 =for apidoc Am|void|Renew|void* ptr|int nitems|type
1068 The XSUB-writer's interface to the C C<realloc> function.
1070 =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1071 The XSUB-writer's interface to the C C<realloc> function, with
1074 =for apidoc Am|void|Safefree|void* ptr
1075 The XSUB-writer's interface to the C C<free> function.
1077 =for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1078 The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
1079 source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
1080 the type. Can do overlapping moves. See also C<Copy>.
1082 =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1083 Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1086 =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1087 The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
1088 source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
1089 the type. May fail on overlapping copies. See also C<Move>.
1091 =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1093 Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1096 =for apidoc Am|void|Zero|void* dest|int nitems|type
1098 The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1099 destination, C<nitems> is the number of items, and C<type> is the type.
1101 =for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1103 Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1106 =for apidoc Am|void|StructCopy|type *src|type *dest|type
1107 This is an architecture-independent macro to copy one structure to another.
1109 =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1111 Fill up memory with a byte pattern (a byte repeated over and over
1112 again) that hopefully catches attempts to access uninitialized memory.
1114 =for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1116 PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1118 =for apidoc Am|void|PoisonFree|void* dest|int nitems|type
1120 PoisonWith(0xEF) for catching access to freed memory.
1122 =for apidoc Am|void|Poison|void* dest|int nitems|type
1124 PoisonWith(0xEF) for catching access to freed memory.
1128 /* Maintained for backwards-compatibility only. Use newSV() instead. */
1130 #define NEWSV(x,len) newSV(len)
1133 #define MEM_SIZE_MAX ((MEM_SIZE)~0)
1135 /* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1136 * overly eager compilers that will bleat about e.g.
1137 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
1138 #ifdef PERL_MALLOC_WRAP
1139 #define MEM_WRAP_CHECK(n,t) MEM_WRAP_CHECK_1(n,t,PL_memory_wrap)
1140 #define MEM_WRAP_CHECK_1(n,t,a) \
1141 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
1142 #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
1144 #define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (Perl_croak_nocontext("%s",PL_memory_wrap),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1148 #define MEM_WRAP_CHECK(n,t)
1149 #define MEM_WRAP_CHECK_1(n,t,a)
1150 #define MEM_WRAP_CHECK_2(n,t,a,b)
1151 #define MEM_WRAP_CHECK_(n,t)
1153 #define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1159 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
1160 * go through functions, which are handy for debugging breakpoints, but
1161 * which more importantly get the immediate calling environment (file and
1162 * line number, and C function name if available) passed in. This info can
1163 * then be used for logging the calls, for which one gets a sample
1164 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
1167 * - not all memory allocs get logged, only those
1168 * that go through Newx() and derivatives (while all
1169 * Safefrees do get logged)
1170 * - __FILE__ and __LINE__ do not work everywhere
1171 * - __func__ or __FUNCTION__ even less so
1172 * - I think more goes on after the perlio frees but
1173 * the thing is that STDERR gets closed (as do all
1174 * the file descriptors)
1175 * - no deeper calling stack than the caller of the Newx()
1176 * or the kind, but do I look like a C reflection/introspection
1178 * - the function prototypes for the logging functions
1179 * probably should maybe be somewhere else than handy.h
1180 * - one could consider inlining (macrofying) the logging
1181 * for speed, but I am too lazy
1182 * - one could imagine recording the allocations in a hash,
1183 * (keyed by the allocation address?), and maintain that
1184 * through reallocs and frees, but how to do that without
1185 * any News() happening...?
1186 * - lots of -Ddefines to get useful/controllable output
1187 * - lots of ENV reads
1190 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);
1192 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);
1194 PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
1197 # ifndef PERL_MEM_LOG_NOIMPL
1206 # if defined(PERL_IN_SV_C) /* those are only used in sv.c */
1207 void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1208 void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1215 #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1216 #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
1217 #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
1220 #ifndef MEM_LOG_ALLOC
1221 #define MEM_LOG_ALLOC(n,t,a) (a)
1223 #ifndef MEM_LOG_REALLOC
1224 #define MEM_LOG_REALLOC(n,t,v,a) (a)
1226 #ifndef MEM_LOG_FREE
1227 #define MEM_LOG_FREE(a) (a)
1230 #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1231 #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1232 #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
1235 /* pre 5.9.x compatibility */
1236 #define New(x,v,n,t) Newx(v,n,t)
1237 #define Newc(x,v,n,t,c) Newxc(v,n,t,c)
1238 #define Newz(x,v,n,t) Newxz(v,n,t)
1241 #define Renew(v,n,t) \
1242 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1243 #define Renewc(v,n,t,c) \
1244 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1247 #define Safefree(d) \
1248 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
1250 #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
1253 #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1254 #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1255 #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
1257 #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1258 #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1260 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
1262 /* Using bzero(), which returns void. */
1263 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
1266 #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1267 #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1268 #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1269 #define Poison(d,n,t) PoisonFree(d,n,t)
1271 #ifdef USE_STRUCT_COPY
1272 #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
1274 #define StructCopy(s,d,t) Copy(s,d,1,t)
1277 #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
1278 #define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
1282 # define Perl_va_copy(s, d) va_copy(d, s)
1284 # if defined(__va_copy)
1285 # define Perl_va_copy(s, d) __va_copy(d, s)
1287 # define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1292 /* convenience debug macros */
1294 #define pTHX_FORMAT "Perl interpreter: 0x%p"
1295 #define pTHX__FORMAT ", Perl interpreter: 0x%p"
1296 #define pTHX_VALUE_ (void *)my_perl,
1297 #define pTHX_VALUE (void *)my_perl
1298 #define pTHX__VALUE_ ,(void *)my_perl,
1299 #define pTHX__VALUE ,(void *)my_perl
1302 #define pTHX__FORMAT
1305 #define pTHX__VALUE_
1307 #endif /* USE_ITHREADS */
1309 /* Perl_deprecate was not part of the public API, and did not have a deprecate()
1310 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1311 CPAN::Unpack show any users outside the core. */
1313 # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
1318 * c-indentation-style: bsd
1320 * indent-tabs-mode: nil
1323 * ex: set ts=8 sts=4 sw=4 et: