3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000,
4 * 2001, 2002, 2004, 2005, 2006, 2007, 2008 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
120 #define cBOOL(cbool) ((bool)!!(cbool))
122 /* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
123 * XXX Should really be a Configure probe, with HAS__FUNCTION__
124 * and FUNCTION__ as results.
125 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
126 #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
127 # define FUNCTION__ __func__
129 # if (defined(_MSC_VER) && _MSC_VER < 1300) || /* Pre-MSVC 7.0 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
130 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
131 # define FUNCTION__ ""
133 # define FUNCTION__ __FUNCTION__ /* Common extension. */
137 /* XXX A note on the perl source internal type system. The
138 original intent was that I32 be *exactly* 32 bits.
140 Currently, we only guarantee that I32 is *at least* 32 bits.
141 Specifically, if int is 64 bits, then so is I32. (This is the case
142 for the Cray.) This has the advantage of meshing nicely with
143 standard library calls (where we pass an I32 and the library is
144 expecting an int), but the disadvantage that an I32 is not 32 bits.
145 Andy Dougherty August 1996
147 There is no guarantee that there is *any* integral type with
148 exactly 32 bits. It is perfectly legal for a system to have
149 sizeof(short) == sizeof(int) == sizeof(long) == 8.
151 Similarly, there is no guarantee that I16 and U16 have exactly 16
154 For dealing with issues that may arise from various 32/64-bit
155 systems, we will ask Configure to check out
157 SHORTSIZE == sizeof(short)
158 INTSIZE == sizeof(int)
159 LONGSIZE == sizeof(long)
160 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
161 PTRSIZE == sizeof(void *)
162 DOUBLESIZE == sizeof(double)
163 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
167 #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
168 # include <inttypes.h>
169 # ifdef INT32_MIN_BROKEN
171 # define INT32_MIN (-2147483647-1)
173 # ifdef INT64_MIN_BROKEN
175 # define INT64_MIN (-9223372036854775807LL-1)
190 #endif /* PERL_CORE */
192 #if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
193 # ifndef UINT64_C /* usually from <inttypes.h> */
194 # if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
195 # define INT64_C(c) CAT2(c,LL)
196 # define UINT64_C(c) CAT2(c,ULL)
198 # if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
199 # define INT64_C(c) CAT2(c,L)
200 # define UINT64_C(c) CAT2(c,UL)
202 # if defined(_WIN64) && defined(_MSC_VER)
203 # define INT64_C(c) CAT2(c,I64)
204 # define UINT64_C(c) CAT2(c,UI64)
206 # define INT64_C(c) ((I64TYPE)(c))
207 # define UINT64_C(c) ((U64TYPE)(c))
214 #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
216 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
217 Please search CHAR_MAX in perl.h for further details. */
218 #define U8_MAX UINT8_MAX
219 #define U8_MIN UINT8_MIN
221 #define I16_MAX INT16_MAX
222 #define I16_MIN INT16_MIN
223 #define U16_MAX UINT16_MAX
224 #define U16_MIN UINT16_MIN
226 #define I32_MAX INT32_MAX
227 #define I32_MIN INT32_MIN
228 #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
229 # define U32_MAX UINT32_MAX
231 # define U32_MAX 4294967295U
233 #define U32_MIN UINT32_MIN
237 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
238 Please search CHAR_MAX in perl.h for further details. */
239 #define U8_MAX PERL_UCHAR_MAX
240 #define U8_MIN PERL_UCHAR_MIN
242 #define I16_MAX PERL_SHORT_MAX
243 #define I16_MIN PERL_SHORT_MIN
244 #define U16_MAX PERL_USHORT_MAX
245 #define U16_MIN PERL_USHORT_MIN
248 # define I32_MAX PERL_INT_MAX
249 # define I32_MIN PERL_INT_MIN
250 # define U32_MAX PERL_UINT_MAX
251 # define U32_MIN PERL_UINT_MIN
253 # define I32_MAX PERL_LONG_MAX
254 # define I32_MIN PERL_LONG_MIN
255 # define U32_MAX PERL_ULONG_MAX
256 # define U32_MIN PERL_ULONG_MIN
261 /* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
262 #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
263 #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
264 #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
266 #define Ctl(ch) ((ch) & 037)
269 =head1 SV-Body Allocation
271 =for apidoc Ama|SV*|newSVpvs|const char* s
272 Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
274 =for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
275 Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
278 =for apidoc Ama|SV*|newSVpvs_share|const char* s
279 Like C<newSVpvn_share>, but takes a literal string instead of a string/length
280 pair and omits the hash parameter.
282 =for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
283 Like C<sv_catpvn_flags>, but takes a literal string instead of a
286 =for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
287 Like C<sv_catpvn_nomg>, but takes a literal string instead of a
290 =for apidoc Am|void|sv_catpvs|SV* sv|const char* s
291 Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
293 =for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
294 Like C<sv_catpvn_mg>, but takes a literal string instead of a
297 =for apidoc Am|void|sv_setpvs|SV* sv|const char* s
298 Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
300 =for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
301 Like C<sv_setpvn_mg>, but takes a literal string instead of a
304 =for apidoc Am|SV *|sv_setref_pvs|const char* s
305 Like C<sv_setref_pvn>, but takes a literal string instead of a
308 =head1 Memory Management
310 =for apidoc Ama|char*|savepvs|const char* s
311 Like C<savepvn>, but takes a literal string instead of a string/length pair.
313 =for apidoc Ama|char*|savesharedpvs|const char* s
314 A version of C<savepvs()> which allocates the duplicate string in memory
315 which is shared between threads.
319 =for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
320 Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
322 =head1 Hash Manipulation Functions
324 =for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
325 Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
327 =for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
328 Like C<hv_store>, but takes a literal string instead of a string/length pair
329 and omits the hash parameter.
331 =head1 Lexer interface
333 =for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
335 Like L</lex_stuff_pvn>, but takes a literal string instead of a
341 /* concatenating with "" ensures that only literal strings are accepted as argument */
342 #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
344 /* note that STR_WITH_LEN() can't be used as argument to macros or functions that
345 * under some configurations might be macros, which means that it requires the full
346 * Perl_xxx(aTHX_ ...) form for any API calls where it's used.
349 /* STR_WITH_LEN() shortcuts */
350 #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
351 #define newSVpvs_flags(str,flags) \
352 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
353 #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
354 #define sv_catpvs_flags(sv, str, flags) \
355 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
356 #define sv_catpvs_nomg(sv, str) \
357 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
358 #define sv_catpvs(sv, str) \
359 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
360 #define sv_catpvs_mg(sv, str) \
361 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
362 #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
363 #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
364 #define sv_setref_pvs(rv, classname, str) \
365 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
366 #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
367 #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
368 #define gv_stashpvs(str, create) \
369 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
370 #define gv_fetchpvs(namebeg, add, sv_type) \
371 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
372 #define gv_fetchpvn(namebeg, len, add, sv_type) \
373 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
374 #define sv_catxmlpvs(dsv, str, utf8) \
375 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
376 #define hv_fetchs(hv,key,lval) \
377 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
378 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
379 : HV_FETCH_JUST_SV, NULL, 0))
381 #define hv_stores(hv,key,val) \
382 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
383 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
385 #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
387 #define get_cvs(str, flags) \
388 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
391 =head1 Miscellaneous Functions
393 =for apidoc Am|bool|strNE|char* s1|char* s2
394 Test two strings to see if they are different. Returns true or
397 =for apidoc Am|bool|strEQ|char* s1|char* s2
398 Test two strings to see if they are equal. Returns true or false.
400 =for apidoc Am|bool|strLT|char* s1|char* s2
401 Test two strings to see if the first, C<s1>, is less than the second,
402 C<s2>. Returns true or false.
404 =for apidoc Am|bool|strLE|char* s1|char* s2
405 Test two strings to see if the first, C<s1>, is less than or equal to the
406 second, C<s2>. Returns true or false.
408 =for apidoc Am|bool|strGT|char* s1|char* s2
409 Test two strings to see if the first, C<s1>, is greater than the second,
410 C<s2>. Returns true or false.
412 =for apidoc Am|bool|strGE|char* s1|char* s2
413 Test two strings to see if the first, C<s1>, is greater than or equal to
414 the second, C<s2>. Returns true or false.
416 =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
417 Test two strings to see if they are different. The C<len> parameter
418 indicates the number of bytes to compare. Returns true or false. (A
419 wrapper for C<strncmp>).
421 =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
422 Test two strings to see if they are equal. The C<len> parameter indicates
423 the number of bytes to compare. Returns true or false. (A wrapper for
429 #define strNE(s1,s2) (strcmp(s1,s2))
430 #define strEQ(s1,s2) (!strcmp(s1,s2))
431 #define strLT(s1,s2) (strcmp(s1,s2) < 0)
432 #define strLE(s1,s2) (strcmp(s1,s2) <= 0)
433 #define strGT(s1,s2) (strcmp(s1,s2) > 0)
434 #define strGE(s1,s2) (strcmp(s1,s2) >= 0)
435 #define strnNE(s1,s2,l) (strncmp(s1,s2,l))
436 #define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
439 # define memNE(s1,s2,l) (memcmp(s1,s2,l))
440 # define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
442 # define memNE(s1,s2,l) (bcmp(s1,s2,l))
443 # define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
446 #define memEQs(s1, l, s2) \
447 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
448 #define memNEs(s1, l, s2) !memEQs(s1, l, s2)
453 * Unfortunately, the introduction of locales means that we
454 * can't trust isupper(), etc. to tell the truth. And when
455 * it comes to /\w+/ with tainting enabled, we *must* be able
456 * to trust our character classes.
458 * Therefore, the default tests in the text of Perl will be
459 * independent of locale. Any code that wants to depend on
460 * the current locale will use the tests that begin with "lc".
463 #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
471 =head1 Character classes
472 There are three variants for all the functions in this section. The base ones
473 operate using the character set of the platform Perl is running on. The ones
474 with an C<_A> suffix operate on the ASCII character set, and the ones with an
475 C<_L1> suffix operate on the full Latin1 character set. All are unaffected by
476 locale and by C<use bytes>.
478 For ASCII platforms, the base function with no suffix and the one with the
479 C<_A> suffix are identical. The function with the C<_L1> suffix imposes the
480 Latin-1 character set onto the platform. That is, the code points that are
481 ASCII are unaffected, since ASCII is a subset of Latin-1. But the non-ASCII
482 code points are treated as if they are Latin-1 characters. For example,
483 C<isSPACE_L1()> will return true when called with the code point 0xA0, which is
484 the Latin-1 NO-BREAK SPACE.
486 For EBCDIC platforms, the base function with no suffix and the one with the
487 C<_L1> suffix should be identical, since, as of this writing, the EBCDIC code
488 pages that Perl knows about all are equivalent to Latin-1. The function that
489 ends in an C<_A> suffix will not return true unless the specified character also
490 has an ASCII equivalent.
492 =for apidoc Am|bool|isALPHA|char ch
493 Returns a boolean indicating whether the specified character is an
494 alphabetic character in the platform's native character set.
495 See the L<top of this section|/Character classes> for an explanation of variants
496 C<isALPHA_A> and C<isALPHA_L1>.
498 =for apidoc Am|bool|isASCII|char ch
499 Returns a boolean indicating whether the specified character is one of the 128
500 characters in the ASCII character set. On non-ASCII platforms, it is if this
501 character corresponds to an ASCII character. Variants C<isASCII_A()> and
502 C<isASCII_L1()> are identical to C<isASCII()>.
504 =for apidoc Am|bool|isDIGIT|char ch
505 Returns a boolean indicating whether the specified character is a
506 digit in the platform's native character set.
507 Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
509 =for apidoc Am|bool|isLOWER|char ch
510 Returns a boolean indicating whether the specified character is a
511 lowercase character in the platform's native character set.
512 See the L<top of this section|/Character classes> for an explanation of variants
513 C<isLOWER_A> and C<isLOWER_L1>.
515 =for apidoc Am|bool|isOCTAL|char ch
516 Returns a boolean indicating whether the specified character is an
517 octal digit, [0-7] in the platform's native character set.
518 Variants C<isOCTAL_A> and C<isOCTAL_L1> are identical to C<isOCTAL>.
520 =for apidoc Am|bool|isSPACE|char ch
521 Returns a boolean indicating whether the specified character is a
522 whitespace character in the platform's native character set. This is the same
523 as what C<\s> matches in a regular expression.
524 See the L<top of this section|/Character classes> for an explanation of variants
525 C<isSPACE_A> and C<isSPACE_L1>.
527 =for apidoc Am|bool|isUPPER|char ch
528 Returns a boolean indicating whether the specified character is an
529 uppercase character in the platform's native character set.
530 See the L<top of this section|/Character classes> for an explanation of variants
531 C<isUPPER_A> and C<isUPPER_L1>.
533 =for apidoc Am|bool|isWORDCHAR|char ch
534 Returns a boolean indicating whether the specified character is a
535 character that is any of: alphabetic, numeric, or an underscore. This is the
536 same as what C<\w> matches in a regular expression.
537 C<isALNUM()> is a synonym provided for backward compatibility. Note that it
538 does not have the standard C language meaning of alphanumeric, since it matches
539 an underscore and the standard meaning does not.
540 See the L<top of this section|/Character classes> for an explanation of variants
541 C<isWORDCHAR_A> and C<isWORDCHAR_L1>.
543 =for apidoc Am|bool|isXDIGIT|char ch
544 Returns a boolean indicating whether the specified character is a hexadecimal
545 digit, [0-9A-Fa-f]. Variants C<isXDIGIT_A()> and C<isXDIGIT_L1()> are
546 identical to C<isXDIGIT()>.
548 =head1 Character case changing
550 =for apidoc Am|char|toUPPER|char ch
551 Converts the specified character to uppercase in the platform's native
552 character set, if possible; otherwise returns the input character itself.
554 =for apidoc Am|char|toLOWER|char ch
555 Converts the specified character to lowercase in the platform's native
556 character set, if possible; otherwise returns the input character itself.
560 Note that these macros are repeated in Devel::PPPort, so should also be
561 patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
565 /* Specify the widest unsigned type on the platform. Use U64TYPE because U64
566 * is known only in the perl core, and this macro can be called from outside
569 # define WIDEST_UTYPE U64TYPE
571 # define WIDEST_UTYPE U32
574 /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
575 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
576 * bits of information are lost even on a 64-bit machine, but to get the
577 * compiler to optimize it out if possible. This is because Configure makes
578 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
579 * sizeof() guarantees that this evaluates to a constant true at compile time.
581 #define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
584 # define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128))
586 # define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
589 #define isASCII_A(c) isASCII(c)
590 #define isASCII_L1(c) isASCII(c)
592 /* ASCII range only */
593 #ifdef H_PERL /* If have access to perl.h, lookup in its table */
594 /* Bits for PL_charclass[]. These use names used in l1_char_class_tab.h but
595 * their actual definitions are here. If that has a name not used here, it
597 # define _CC_ALNUMC_A (1<<0)
598 # define _CC_ALNUMC_L1 (1<<1)
599 # define _CC_ALPHA_A (1<<2)
600 # define _CC_ALPHA_L1 (1<<3)
601 # define _CC_BLANK_A (1<<4)
602 # define _CC_BLANK_L1 (1<<5)
603 # define _CC_CHARNAME_CONT (1<<6)
604 # define _CC_CNTRL_A (1<<7)
605 # define _CC_CNTRL_L1 (1<<8)
606 # define _CC_DIGIT_A (1<<9)
607 # define _CC_GRAPH_A (1<<10)
608 # define _CC_GRAPH_L1 (1<<11)
609 # define _CC_IDFIRST_A (1<<12)
610 # define _CC_IDFIRST_L1 (1<<13)
611 # define _CC_LOWER_A (1<<14)
612 # define _CC_LOWER_L1 (1<<15)
613 # define _CC_OCTAL_A (1<<16)
614 # define _CC_PRINT_A (1<<17)
615 # define _CC_PRINT_L1 (1<<18)
616 # define _CC_PSXSPC_A (1<<19)
617 # define _CC_PSXSPC_L1 (1<<20)
618 # define _CC_PUNCT_A (1<<21)
619 # define _CC_PUNCT_L1 (1<<22)
620 # define _CC_SPACE_A (1<<23)
621 # define _CC_SPACE_L1 (1<<24)
622 # define _CC_UPPER_A (1<<25)
623 # define _CC_UPPER_L1 (1<<26)
624 # define _CC_WORDCHAR_A (1<<27)
625 # define _CC_WORDCHAR_L1 (1<<28)
626 # define _CC_XDIGIT_A (1<<29)
627 # define _CC_NONLATIN1_FOLD (1<<30)
630 * If more are needed, can give up some of the above. The first ones to go
631 * would be those that require just two tests to verify, either there are two
632 * code points, like BLANK_A, or occupy a single range like OCTAL_A, DIGIT_A,
633 * UPPER_A, and LOWER_A.
637 EXTCONST U32 PL_charclass[] = {
638 # include "l1_char_class_tab.h"
641 # else /* ! DOINIT */
642 EXTCONST U32 PL_charclass[];
645 # define isALNUMC_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALNUMC_A))
646 # define isALPHA_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALPHA_A))
647 # define isBLANK_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_BLANK_A))
648 # define isCNTRL_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CNTRL_A))
649 # define isDIGIT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_DIGIT_A))
650 # define isGRAPH_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_GRAPH_A))
651 # define isIDFIRST_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_IDFIRST_A))
652 # define isLOWER_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_LOWER_A))
653 # define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_OCTAL_A))
654 # define isPRINT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PRINT_A))
655 # define isPSXSPC_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PSXSPC_A))
656 # define isPUNCT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PUNCT_A))
657 # define isSPACE_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_SPACE_A))
658 # define isUPPER_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_UPPER_A))
659 # define isWORDCHAR_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_WORDCHAR_A))
660 # define isXDIGIT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_XDIGIT_A))
661 /* Either participates in a fold with a character above 255, or is a
663 # 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_NONLATIN1_FOLD))
664 #else /* No perl.h. */
665 # define isOCTAL_A(c) ((c) <= '7' && (c) >= '0')
667 # define isALNUMC_A(c) (isASCII(c) && isALNUMC(c))
668 # define isALPHA_A(c) (isASCII(c) && isALPHA(c))
669 # define isBLANK_A(c) (isASCII(c) && isBLANK(c))
670 # define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
671 # define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
672 # define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
673 # define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
674 # define isLOWER_A(c) (isASCII(c) && isLOWER(c))
675 # define isPRINT_A(c) (isASCII(c) && isPRINT(c))
676 # define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
677 # define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
678 # define isSPACE_A(c) (isASCII(c) && isSPACE(c))
679 # define isUPPER_A(c) (isASCII(c) && isUPPER(c))
680 # define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
681 # define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
682 # else /* ASCII platform, no perl.h */
683 # define isALNUMC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
684 # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
685 # define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
686 # define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
687 # define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
688 # define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
689 # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
690 # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
691 # define isPRINT_A(c) (((c) >= 32 && (c) < 127))
692 # define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
693 # define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) || ((c) >= 58 && (c) <= 64) || ((c) >= 91 && (c) <= 96) || ((c) >= 123 && (c) <= 126))
694 # define isSPACE_A(c) ((c) == ' ' || (c) == '\t' || (c) == '\n' || (c) =='\r' || (c) == '\f')
695 # define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
696 # define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
697 # define isXDIGIT_A(c) (isDIGIT_A(c) || ((c) >= 'a' && (c) <= 'f') || ((c) <= 'F' && (c) >= 'A'))
699 #endif /* ASCII range definitions */
701 /* Latin1 definitions */
703 # define isALNUMC_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALNUMC_L1))
704 # define isALPHA_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALPHA_L1))
705 # define isBLANK_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_BLANK_L1))
706 /* continuation character for legal NAME in \N{NAME} */
707 # define isCHARNAME_CONT(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CHARNAME_CONT))
708 # define isCNTRL_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CNTRL_L1))
709 # define isGRAPH_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_GRAPH_L1))
710 # define isIDFIRST_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_IDFIRST_L1))
711 # define isLOWER_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_LOWER_L1))
712 # define isPRINT_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PRINT_L1))
713 # define isPSXSPC_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PSXSPC_L1))
714 # define isPUNCT_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PUNCT_L1))
715 # define isSPACE_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_SPACE_L1))
716 # define isUPPER_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_UPPER_L1))
717 # define isWORDCHAR_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_WORDCHAR_L1))
718 #else /* No access to perl.h. Only a few provided here, just in case needed
719 * for backwards compatibility */
720 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
721 * >= AA test to speed up ASCII-only tests at the expense of the others */
722 # define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
723 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
724 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
725 || NATIVE_TO_UNI((U8) c) == 0xAA \
726 || NATIVE_TO_UNI((U8) c) == 0xB5 \
727 || NATIVE_TO_UNI((U8) c) == 0xBA)))
728 # define isCHARNAME_CONT(c) (isALNUM_L1(c) || (c) == ' ' || (c) == '-' || (c) == '(' || (c) == ')' || (c) == ':' || NATIVE_TO_UNI((U8) c) == 0xA0)
731 /* Macros for backwards compatibility and for completeness when the ASCII and
732 * Latin1 values are identical */
733 #define isALNUM(c) isWORDCHAR(c)
734 #define isALNUMU(c) isWORDCHAR_L1(c)
735 #define isALPHAU(c) isALPHA_L1(c)
736 #define isDIGIT_L1(c) isDIGIT_A(c)
737 #define isOCTAL(c) isOCTAL_A(c)
738 #define isOCTAL_L1(c) isOCTAL_A(c)
739 #define isXDIGIT_L1(c) isXDIGIT_A(c)
741 /* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
742 * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
743 * therefore can use native functions. For those where C89 doesn't define a
744 * function, use our function, assuming that the EBCDIC code page is isomorphic
745 * with Latin1, which the three currently recognized by Perl are. Some libc's
746 * have an isblank(), but it's not guaranteed. */
748 # define isALNUMC(c) isalnum(c)
749 # define isALPHA(c) isalpha(c)
750 # define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
751 # define isCNTRL(c) iscntrl(c)
752 # define isDIGIT(c) isdigit(c)
753 # define isGRAPH(c) isgraph(c)
754 # define isIDFIRST(c) (isALPHA(c) || (c) == '_')
755 # define isLOWER(c) islower(c)
756 # define isPRINT(c) isprint(c)
757 # define isPSXSPC(c) isspace(c)
758 # define isPUNCT(c) ispunct(c)
759 # define isSPACE(c) (isPSXSPC(c) && (c) != '\v')
760 # define isUPPER(c) isupper(c)
761 # define isXDIGIT(c) isxdigit(c)
762 # define isWORDCHAR(c) (isalnum(c) || (c) == '_')
763 # define toLOWER(c) tolower(c)
764 # define toUPPER(c) toupper(c)
765 #else /* Not EBCDIC: ASCII-only matching */
766 # define isALNUMC(c) isALNUMC_A(c)
767 # define isALPHA(c) isALPHA_A(c)
768 # define isBLANK(c) isBLANK_A(c)
769 # define isCNTRL(c) isCNTRL_A(c)
770 # define isDIGIT(c) isDIGIT_A(c)
771 # define isGRAPH(c) isGRAPH_A(c)
772 # define isIDFIRST(c) isIDFIRST_A(c)
773 # define isLOWER(c) isLOWER_A(c)
774 # define isPRINT(c) isPRINT_A(c)
775 # define isPSXSPC(c) isPSXSPC_A(c)
776 # define isPUNCT(c) isPUNCT_A(c)
777 # define isSPACE(c) isSPACE_A(c)
778 # define isUPPER(c) isUPPER_A(c)
779 # define isWORDCHAR(c) isWORDCHAR_A(c)
780 # define isXDIGIT(c) isXDIGIT_A(c)
782 /* ASCII casing. These could also be written as
783 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
784 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
785 which uses table lookup and mask instead of subtraction. (This would
786 work because the _MOD does not apply in the ASCII range) */
787 # define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
788 # define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
792 /* Use table lookup for speed; return error character for input
794 #define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \
795 ? UNI_TO_NATIVE(PL_latin1_lc[ \
796 NATIVE_TO_UNI( (U8) (c)) ]) \
797 : UNICODE_REPLACEMENT)
798 /* Modified uc. Is correct uc except for three non-ascii chars which are
799 * all mapped to one of them, and these need special handling; error
800 * character for input out-of-range */
801 #define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \
802 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \
803 NATIVE_TO_UNI( (U8) (c)) ]) \
804 : UNICODE_REPLACEMENT)
806 #ifdef USE_NEXT_CTYPE
808 # define isALNUM_LC(c) \
809 (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
810 # define isIDFIRST_LC(c) \
811 (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
812 # define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
813 # define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
814 # define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
815 # define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
816 # define isLOWER_LC(c) NXIsLower((unsigned int)(c))
817 # define isALNUMC_LC(c) NXIsAlNum((unsigned int)(c))
818 # define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
819 # define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
820 # define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
821 # define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
822 # define toUPPER_LC(c) NXToUpper((unsigned int)(c))
823 # define toLOWER_LC(c) NXToLower((unsigned int)(c))
825 #else /* !USE_NEXT_CTYPE */
827 # if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
829 # define isALNUM_LC(c) (isalnum((unsigned char)(c)) || (char)(c) == '_')
830 # define isIDFIRST_LC(c) (isalpha((unsigned char)(c)) || (char)(c) == '_')
831 # define isALPHA_LC(c) isalpha((unsigned char)(c))
832 # define isSPACE_LC(c) isspace((unsigned char)(c))
833 # define isDIGIT_LC(c) isdigit((unsigned char)(c))
834 # define isUPPER_LC(c) isupper((unsigned char)(c))
835 # define isLOWER_LC(c) islower((unsigned char)(c))
836 # define isALNUMC_LC(c) isalnum((unsigned char)(c))
837 # define isCNTRL_LC(c) iscntrl((unsigned char)(c))
838 # define isGRAPH_LC(c) isgraph((unsigned char)(c))
839 # define isPRINT_LC(c) isprint((unsigned char)(c))
840 # define isPUNCT_LC(c) ispunct((unsigned char)(c))
841 # define toUPPER_LC(c) toupper((unsigned char)(c))
842 # define toLOWER_LC(c) tolower((unsigned char)(c))
846 # define isALNUM_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
847 # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
848 # define isALPHA_LC(c) (isascii(c) && isalpha(c))
849 # define isSPACE_LC(c) (isascii(c) && isspace(c))
850 # define isDIGIT_LC(c) (isascii(c) && isdigit(c))
851 # define isUPPER_LC(c) (isascii(c) && isupper(c))
852 # define isLOWER_LC(c) (isascii(c) && islower(c))
853 # define isALNUMC_LC(c) (isascii(c) && isalnum(c))
854 # define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
855 # define isGRAPH_LC(c) (isascii(c) && isgraph(c))
856 # define isPRINT_LC(c) (isascii(c) && isprint(c))
857 # define isPUNCT_LC(c) (isascii(c) && ispunct(c))
858 # define toUPPER_LC(c) toupper(c)
859 # define toLOWER_LC(c) tolower(c)
862 #endif /* USE_NEXT_CTYPE */
864 #define isPSXSPC_LC(c) (isSPACE_LC(c) || (c) == '\v')
865 #define isBLANK_LC(c) isBLANK(c) /* could be wrong */
867 /* For use in the macros just below. If the input is Latin1, use the Latin1
868 * (_L1) version of the macro; otherwise use the function. Won't compile if
869 * 'c' isn't unsigned, as won't match function prototype. The macros do bounds
870 * checking, so have duplicate checks here, so could create versions of the
871 * macros that don't, but experiments show that gcc optimizes them out anyway.
873 #define generic_uni(macro, function, c) ((c) < 256 \
874 ? CAT2(macro, _L1)(c) \
876 /* Note that all ignore 'use bytes' */
878 #define isALNUM_uni(c) generic_uni(isWORDCHAR, is_uni_alnum, c)
879 #define isIDFIRST_uni(c) generic_uni(isIDFIRST, is_uni_idfirst, c)
880 #define isALPHA_uni(c) generic_uni(isALPHA, is_uni_alpha, c)
881 #define isSPACE_uni(c) generic_uni(isSPACE, is_uni_space, c)
882 #define isDIGIT_uni(c) generic_uni(isDIGIT, is_uni_digit, c)
883 #define isUPPER_uni(c) generic_uni(isUPPER, is_uni_upper, c)
884 #define isLOWER_uni(c) generic_uni(isLOWER, is_uni_lower, c)
885 #define isASCII_uni(c) isASCII(c)
886 /* All controls are in Latin1 */
887 #define isCNTRL_uni(c) ((c) < 256 && isCNTRL_L1(c))
888 #define isGRAPH_uni(c) generic_uni(isGRAPH, is_uni_graph, c)
889 #define isPRINT_uni(c) generic_uni(isPRINT, is_uni_print, c)
890 #define isPUNCT_uni(c) generic_uni(isPUNCT, is_uni_punct, c)
891 #define isXDIGIT_uni(c) generic_uni(isXDIGIT, is_uni_xdigit, c)
892 #define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
893 #define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
894 #define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
895 #define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
897 /* Posix and regular space differ only in U+000B, which is in Latin1 */
898 #define isPSXSPC_uni(c) ((c) < 256 ? isPSXSPC_L1(c) : isSPACE_uni(c))
899 #define isBLANK_uni(c) isBLANK(c) /* could be wrong */
901 #define isALNUM_LC_uvchr(c) (c < 256 ? isALNUM_LC(c) : is_uni_alnum_lc(c))
902 #define isIDFIRST_LC_uvchr(c) (c < 256 ? isIDFIRST_LC(c) : is_uni_idfirst_lc(c))
903 #define isALPHA_LC_uvchr(c) (c < 256 ? isALPHA_LC(c) : is_uni_alpha_lc(c))
904 #define isSPACE_LC_uvchr(c) (c < 256 ? isSPACE_LC(c) : is_uni_space_lc(c))
905 #define isDIGIT_LC_uvchr(c) (c < 256 ? isDIGIT_LC(c) : is_uni_digit_lc(c))
906 #define isUPPER_LC_uvchr(c) (c < 256 ? isUPPER_LC(c) : is_uni_upper_lc(c))
907 #define isLOWER_LC_uvchr(c) (c < 256 ? isLOWER_LC(c) : is_uni_lower_lc(c))
908 #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : is_uni_cntrl_lc(c))
909 #define isGRAPH_LC_uvchr(c) (c < 256 ? isGRAPH_LC(c) : is_uni_graph_lc(c))
910 #define isPRINT_LC_uvchr(c) (c < 256 ? isPRINT_LC(c) : is_uni_print_lc(c))
911 #define isPUNCT_LC_uvchr(c) (c < 256 ? isPUNCT_LC(c) : is_uni_punct_lc(c))
913 #define isPSXSPC_LC_uni(c) (isSPACE_LC_uni(c) ||(c) == '\f')
914 #define isBLANK_LC_uni(c) isBLANK(c) /* could be wrong */
916 /* For use in the macros just below. If the input is ASCII, use the ASCII (_A)
917 * version of the macro; if the input is in the upper Latin1 range, use the
918 * Latin1 (_L1) version of the macro, after converting from utf8; otherwise use
919 * the function. This relies on the fact that ASCII characters have the same
920 * representation whether utf8 or not */
921 #define generic_utf8(macro, function, p) (isASCII(*(p)) \
922 ? CAT2(CAT2(macro,_),A)(*(p)) \
923 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
925 (TWO_BYTE_UTF8_TO_UNI(*(p), \
929 /* Note that all ignore 'use bytes' */
931 #define isALNUM_utf8(p) generic_utf8(isWORDCHAR, is_utf8_alnum, p)
932 /* To prevent S_scan_word in toke.c from hanging, we have to make sure that
933 * IDFIRST is an alnum. See
934 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
935 * ever wanted to know about. XXX It is unclear if this should extend to
936 * isIDFIRST_uni() which it hasn't so far. (In the ASCII range, there isn't a
937 * difference.) This used to be not the XID version, but we decided to go with
938 * the more modern Unicode definition */
939 #define isIDFIRST_utf8(p) (isASCII(*(p)) \
940 ? isIDFIRST_A(*(p)) \
941 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
942 ? isIDFIRST_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \
944 : Perl__is_utf8__perl_idstart(aTHX_ p))
945 #define isIDCONT_utf8(p) generic_utf8(isWORDCHAR, is_utf8_xidcont, p)
946 #define isALPHA_utf8(p) generic_utf8(isALPHA, is_utf8_alpha, p)
947 #define isSPACE_utf8(p) generic_utf8(isSPACE, is_utf8_space, p)
948 #define isDIGIT_utf8(p) generic_utf8(isDIGIT, is_utf8_digit, p)
949 #define isUPPER_utf8(p) generic_utf8(isUPPER, is_utf8_upper, p)
950 #define isLOWER_utf8(p) generic_utf8(isLOWER, is_utf8_lower, p)
951 /* Because ASCII is invariant under utf8, the non-utf8 macro works */
952 #define isASCII_utf8(p) isASCII(p)
953 #define isCNTRL_utf8(p) generic_utf8(isCNTRL, is_utf8_cntrl, p)
954 #define isGRAPH_utf8(p) generic_utf8(isGRAPH, is_utf8_graph, p)
955 #define isPRINT_utf8(p) generic_utf8(isPRINT, is_utf8_print, p)
956 #define isPUNCT_utf8(p) generic_utf8(isPUNCT, is_utf8_punct, p)
957 #define isXDIGIT_utf8(p) generic_utf8(isXDIGIT, is_utf8_xdigit, p)
958 #define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
959 #define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
960 #define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
962 /* Posix and regular space differ only in U+000B, which is in ASCII (and hence
964 #define isPSXSPC_utf8(p) ((isASCII(*(p))) \
966 : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \
967 ? isPSXSPC_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \
970 #define isBLANK_utf8(c) isBLANK(c) /* could be wrong */
972 #define isALNUM_LC_utf8(p) isALNUM_LC_uvchr(utf8_to_uvchr(p, 0))
973 #define isIDFIRST_LC_utf8(p) isIDFIRST_LC_uvchr(utf8_to_uvchr(p, 0))
974 #define isALPHA_LC_utf8(p) isALPHA_LC_uvchr(utf8_to_uvchr(p, 0))
975 #define isSPACE_LC_utf8(p) isSPACE_LC_uvchr(utf8_to_uvchr(p, 0))
976 #define isDIGIT_LC_utf8(p) isDIGIT_LC_uvchr(utf8_to_uvchr(p, 0))
977 #define isUPPER_LC_utf8(p) isUPPER_LC_uvchr(utf8_to_uvchr(p, 0))
978 #define isLOWER_LC_utf8(p) isLOWER_LC_uvchr(utf8_to_uvchr(p, 0))
979 #define isALNUMC_LC_utf8(p) isALNUMC_LC_uvchr(utf8_to_uvchr(p, 0))
980 #define isCNTRL_LC_utf8(p) isCNTRL_LC_uvchr(utf8_to_uvchr(p, 0))
981 #define isGRAPH_LC_utf8(p) isGRAPH_LC_uvchr(utf8_to_uvchr(p, 0))
982 #define isPRINT_LC_utf8(p) isPRINT_LC_uvchr(utf8_to_uvchr(p, 0))
983 #define isPUNCT_LC_utf8(p) isPUNCT_LC_uvchr(utf8_to_uvchr(p, 0))
985 #define isPSXSPC_LC_utf8(c) (isSPACE_LC_utf8(c) ||(c) == '\f')
986 #define isBLANK_LC_utf8(c) isBLANK(c) /* could be wrong */
988 /* This conversion works both ways, strangely enough. On EBCDIC platforms,
989 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII */
990 # define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64)
992 /* Line numbers are unsigned, 32 bits. */
994 #define NOLINE ((line_t) 4294967295UL)
996 /* Helpful alias for version prescan */
997 #define is_LAX_VERSION(a,b) \
998 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1000 #define is_STRICT_VERSION(a,b) \
1001 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1003 #define BADVERSION(a,b,c) \
1010 =head1 Memory Management
1012 =for apidoc Am|void|Newx|void* ptr|int nitems|type
1013 The XSUB-writer's interface to the C C<malloc> function.
1015 In 5.9.3, Newx() and friends replace the older New() API, and drops
1016 the first parameter, I<x>, a debug aid which allowed callers to identify
1017 themselves. This aid has been superseded by a new build option,
1018 PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
1019 there for use in XS modules supporting older perls.
1021 =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
1022 The XSUB-writer's interface to the C C<malloc> function, with
1023 cast. See also C<Newx>.
1025 =for apidoc Am|void|Newxz|void* ptr|int nitems|type
1026 The XSUB-writer's interface to the C C<malloc> function. The allocated
1027 memory is zeroed with C<memzero>. See also C<Newx>.
1029 =for apidoc Am|void|Renew|void* ptr|int nitems|type
1030 The XSUB-writer's interface to the C C<realloc> function.
1032 =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1033 The XSUB-writer's interface to the C C<realloc> function, with
1036 =for apidoc Am|void|Safefree|void* ptr
1037 The XSUB-writer's interface to the C C<free> function.
1039 =for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1040 The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
1041 source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
1042 the type. Can do overlapping moves. See also C<Copy>.
1044 =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1045 Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1048 =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1049 The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
1050 source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
1051 the type. May fail on overlapping copies. See also C<Move>.
1053 =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1055 Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1058 =for apidoc Am|void|Zero|void* dest|int nitems|type
1060 The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1061 destination, C<nitems> is the number of items, and C<type> is the type.
1063 =for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1065 Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1068 =for apidoc Am|void|StructCopy|type src|type dest|type
1069 This is an architecture-independent macro to copy one structure to another.
1071 =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1073 Fill up memory with a byte pattern (a byte repeated over and over
1074 again) that hopefully catches attempts to access uninitialized memory.
1076 =for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1078 PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1080 =for apidoc Am|void|PoisonFree|void* dest|int nitems|type
1082 PoisonWith(0xEF) for catching access to freed memory.
1084 =for apidoc Am|void|Poison|void* dest|int nitems|type
1086 PoisonWith(0xEF) for catching access to freed memory.
1090 /* Maintained for backwards-compatibility only. Use newSV() instead. */
1092 #define NEWSV(x,len) newSV(len)
1095 #define MEM_SIZE_MAX ((MEM_SIZE)~0)
1097 /* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1098 * overly eager compilers that will bleat about e.g.
1099 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
1100 #ifdef PERL_MALLOC_WRAP
1101 #define MEM_WRAP_CHECK(n,t) MEM_WRAP_CHECK_1(n,t,PL_memory_wrap)
1102 #define MEM_WRAP_CHECK_1(n,t,a) \
1103 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
1104 #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
1106 #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)))
1110 #define MEM_WRAP_CHECK(n,t)
1111 #define MEM_WRAP_CHECK_1(n,t,a)
1112 #define MEM_WRAP_CHECK_2(n,t,a,b)
1113 #define MEM_WRAP_CHECK_(n,t)
1115 #define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1121 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
1122 * go through functions, which are handy for debugging breakpoints, but
1123 * which more importantly get the immediate calling environment (file and
1124 * line number, and C function name if available) passed in. This info can
1125 * then be used for logging the calls, for which one gets a sample
1126 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
1129 * - not all memory allocs get logged, only those
1130 * that go through Newx() and derivatives (while all
1131 * Safefrees do get logged)
1132 * - __FILE__ and __LINE__ do not work everywhere
1133 * - __func__ or __FUNCTION__ even less so
1134 * - I think more goes on after the perlio frees but
1135 * the thing is that STDERR gets closed (as do all
1136 * the file descriptors)
1137 * - no deeper calling stack than the caller of the Newx()
1138 * or the kind, but do I look like a C reflection/introspection
1140 * - the function prototypes for the logging functions
1141 * probably should maybe be somewhere else than handy.h
1142 * - one could consider inlining (macrofying) the logging
1143 * for speed, but I am too lazy
1144 * - one could imagine recording the allocations in a hash,
1145 * (keyed by the allocation address?), and maintain that
1146 * through reallocs and frees, but how to do that without
1147 * any News() happening...?
1148 * - lots of -Ddefines to get useful/controllable output
1149 * - lots of ENV reads
1152 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);
1154 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);
1156 PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
1159 # ifndef PERL_MEM_LOG_NOIMPL
1168 # if defined(PERL_IN_SV_C) /* those are only used in sv.c */
1169 void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1170 void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1177 #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1178 #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
1179 #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
1182 #ifndef MEM_LOG_ALLOC
1183 #define MEM_LOG_ALLOC(n,t,a) (a)
1185 #ifndef MEM_LOG_REALLOC
1186 #define MEM_LOG_REALLOC(n,t,v,a) (a)
1188 #ifndef MEM_LOG_FREE
1189 #define MEM_LOG_FREE(a) (a)
1192 #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1193 #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1194 #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
1197 /* pre 5.9.x compatibility */
1198 #define New(x,v,n,t) Newx(v,n,t)
1199 #define Newc(x,v,n,t,c) Newxc(v,n,t,c)
1200 #define Newz(x,v,n,t) Newxz(v,n,t)
1203 #define Renew(v,n,t) \
1204 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1205 #define Renewc(v,n,t,c) \
1206 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1209 #define Safefree(d) \
1210 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
1212 #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
1215 #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1216 #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1217 #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
1219 #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1220 #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1222 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
1224 /* Using bzero(), which returns void. */
1225 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
1228 #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1229 #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1230 #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1231 #define Poison(d,n,t) PoisonFree(d,n,t)
1233 #ifdef USE_STRUCT_COPY
1234 #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
1236 #define StructCopy(s,d,t) Copy(s,d,1,t)
1239 #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
1240 #define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
1244 # define Perl_va_copy(s, d) va_copy(d, s)
1246 # if defined(__va_copy)
1247 # define Perl_va_copy(s, d) __va_copy(d, s)
1249 # define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1254 /* convenience debug macros */
1256 #define pTHX_FORMAT "Perl interpreter: 0x%p"
1257 #define pTHX__FORMAT ", Perl interpreter: 0x%p"
1258 #define pTHX_VALUE_ (void *)my_perl,
1259 #define pTHX_VALUE (void *)my_perl
1260 #define pTHX__VALUE_ ,(void *)my_perl,
1261 #define pTHX__VALUE ,(void *)my_perl
1264 #define pTHX__FORMAT
1267 #define pTHX__VALUE_
1269 #endif /* USE_ITHREADS */
1271 /* Perl_deprecate was not part of the public API, and did not have a deprecate()
1272 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1273 CPAN::Unpack show any users outside the core. */
1275 # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
1280 * c-indentation-style: bsd
1282 * indent-tabs-mode: t
1285 * ex: set ts=8 sts=4 sw=4 noet: