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 /* IMPORTANT NOTE: Everything whose name begins with an underscore is for
12 * internal core Perl use only. */
14 #ifndef HANDY_H /* Guard against nested #inclusion */
17 #if !defined(__STDC__)
25 # define Null(type) ((type)NULL)
30 =for apidoc AmU||Nullch
31 Null character pointer. (No longer available when C<PERL_CORE> is
34 =for apidoc AmU||Nullsv
35 Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
40 # define Nullch Null(char*)
41 # define Nullfp Null(PerlIO*)
42 # define Nullsv Null(SV*)
54 /* The MUTABLE_*() macros cast pointers to the types shown, in such a way
55 * (compiler permitting) that casting away const-ness will give a warning;
59 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
60 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
63 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
64 # define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
66 # define MUTABLE_PTR(p) ((void *) (p))
69 #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
70 #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
71 #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
72 #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
73 #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
74 #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
76 #if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR)
83 /* bool is built-in for g++-2.6.3 and later, which might be used
84 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
85 be sure _G_config.h will be included before this file. _G_config.h
86 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
87 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
88 g++ can be identified by __GNUG__.
89 Andy Dougherty February 2000
91 #ifdef __GNUG__ /* GNU g++ has bool built-in */
92 # ifndef PERL_BOOL_AS_CHAR
99 /* The NeXT dynamic loader headers will not build with the bool macro
100 So declare them now to clear confusion.
102 #if defined(NeXT) || defined(__NeXT__)
105 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
109 # endif /* !HAS_BOOL */
110 #endif /* NeXT || __NeXT__ */
120 /* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
121 * defined as char for example, then the cast from int is
122 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
124 #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
126 /* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
127 * XXX Should really be a Configure probe, with HAS__FUNCTION__
128 * and FUNCTION__ as results.
129 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
130 #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
131 # define FUNCTION__ __func__
133 # if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
134 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
135 # define FUNCTION__ ""
137 # define FUNCTION__ __FUNCTION__ /* Common extension. */
141 /* XXX A note on the perl source internal type system. The
142 original intent was that I32 be *exactly* 32 bits.
144 Currently, we only guarantee that I32 is *at least* 32 bits.
145 Specifically, if int is 64 bits, then so is I32. (This is the case
146 for the Cray.) This has the advantage of meshing nicely with
147 standard library calls (where we pass an I32 and the library is
148 expecting an int), but the disadvantage that an I32 is not 32 bits.
149 Andy Dougherty August 1996
151 There is no guarantee that there is *any* integral type with
152 exactly 32 bits. It is perfectly legal for a system to have
153 sizeof(short) == sizeof(int) == sizeof(long) == 8.
155 Similarly, there is no guarantee that I16 and U16 have exactly 16
158 For dealing with issues that may arise from various 32/64-bit
159 systems, we will ask Configure to check out
161 SHORTSIZE == sizeof(short)
162 INTSIZE == sizeof(int)
163 LONGSIZE == sizeof(long)
164 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
165 PTRSIZE == sizeof(void *)
166 DOUBLESIZE == sizeof(double)
167 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
171 #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
172 # include <inttypes.h>
173 # ifdef INT32_MIN_BROKEN
175 # define INT32_MIN (-2147483647-1)
177 # ifdef INT64_MIN_BROKEN
179 # define INT64_MIN (-9223372036854775807LL-1)
194 #endif /* PERL_CORE */
196 /* INT64_C/UINT64_C are C99 from <stdint.h> (so they will not be
197 * available in strict C89 mode), but they are nice, so let's define
198 * them if necessary. */
199 #if defined(HAS_QUAD)
201 # undef PeRl_UINT64_C
202 /* Prefer the native integer types (int and long) over long long
203 * (which is not C89) and Win32-specific __int64. */
204 # if QUADKIND == QUAD_IS_INT && INTSIZE == 8
205 # define PeRl_INT64_C(c) (c)
206 # define PeRl_UINT64_C(c) CAT2(c,U)
208 # if QUADKIND == QUAD_IS_LONG && LONGSIZE == 8
209 # define PeRl_INT64_C(c) CAT2(c,L)
210 # define PeRl_UINT64_C(c) CAT2(c,UL)
212 # if QUADKIND == QUAD_IS_LONG_LONG && defined(HAS_LONG_LONG)
213 # define PeRl_INT64_C(c) CAT2(c,LL)
214 # define PeRl_UINT64_C(c) CAT2(c,ULL)
216 # if QUADKIND == QUAD_IS___INT64
217 # define PeRl_INT64_C(c) CAT2(c,I64)
218 # define PeRl_UINT64_C(c) CAT2(c,UI64)
220 # ifndef PeRl_INT64_C
221 # define PeRl_INT64_C(c) ((I64TYPE)(c)) /* last resort */
222 # define PeRl_UINT64_C(c) ((U64TYPE)(c))
224 /* In OS X the INT64_C/UINT64_C are defined with LL/ULL, which will
225 * not fly with C89-pedantic gcc, so let's undefine them first so that
226 * we can redefine them with our native integer preferring versions. */
227 # if defined(PERL_DARWIN) && defined(PERL_GCC_PEDANTIC)
232 # define INT64_C(c) PeRl_INT64_C(c)
235 # define UINT64_C(c) PeRl_UINT64_C(c)
239 #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
241 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
242 Please search CHAR_MAX in perl.h for further details. */
243 #define U8_MAX UINT8_MAX
244 #define U8_MIN UINT8_MIN
246 #define I16_MAX INT16_MAX
247 #define I16_MIN INT16_MIN
248 #define U16_MAX UINT16_MAX
249 #define U16_MIN UINT16_MIN
251 #define I32_MAX INT32_MAX
252 #define I32_MIN INT32_MIN
253 #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
254 # define U32_MAX UINT32_MAX
256 # define U32_MAX 4294967295U
258 #define U32_MIN UINT32_MIN
262 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
263 Please search CHAR_MAX in perl.h for further details. */
264 #define U8_MAX PERL_UCHAR_MAX
265 #define U8_MIN PERL_UCHAR_MIN
267 #define I16_MAX PERL_SHORT_MAX
268 #define I16_MIN PERL_SHORT_MIN
269 #define U16_MAX PERL_USHORT_MAX
270 #define U16_MIN PERL_USHORT_MIN
273 # define I32_MAX PERL_INT_MAX
274 # define I32_MIN PERL_INT_MIN
275 # define U32_MAX PERL_UINT_MAX
276 # define U32_MIN PERL_UINT_MIN
278 # define I32_MAX PERL_LONG_MAX
279 # define I32_MIN PERL_LONG_MIN
280 # define U32_MAX PERL_ULONG_MAX
281 # define U32_MIN PERL_ULONG_MIN
286 /* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
287 #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
288 #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
289 #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
291 #define Ctl(ch) ((ch) & 037)
293 /* This is a helper macro to avoid preprocessor issues, replaced by nothing
294 * unless under DEBUGGING, where it expands to an assert of its argument,
295 * followed by a comma (hence the comma operator). If we just used a straight
296 * assert(), we would get a comma with nothing before it when not DEBUGGING */
298 # define __ASSERT_(statement) assert(statement),
300 # define __ASSERT_(statement)
304 =head1 SV-Body Allocation
306 =for apidoc Ama|SV*|newSVpvs|const char* s
307 Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
309 =for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
310 Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
313 =for apidoc Ama|SV*|newSVpvs_share|const char* s
314 Like C<newSVpvn_share>, but takes a literal string instead of a string/length
315 pair and omits the hash parameter.
317 =for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
318 Like C<sv_catpvn_flags>, but takes a literal string instead of a
321 =for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
322 Like C<sv_catpvn_nomg>, but takes a literal string instead of a
325 =for apidoc Am|void|sv_catpvs|SV* sv|const char* s
326 Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
328 =for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
329 Like C<sv_catpvn_mg>, but takes a literal string instead of a
332 =for apidoc Am|void|sv_setpvs|SV* sv|const char* s
333 Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
335 =for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
336 Like C<sv_setpvn_mg>, but takes a literal string instead of a
339 =for apidoc Am|SV *|sv_setref_pvs|const char* s
340 Like C<sv_setref_pvn>, but takes a literal string instead of a
343 =head1 Memory Management
345 =for apidoc Ama|char*|savepvs|const char* s
346 Like C<savepvn>, but takes a literal string instead of a string/length pair.
348 =for apidoc Ama|char*|savesharedpvs|const char* s
349 A version of C<savepvs()> which allocates the duplicate string in memory
350 which is shared between threads.
354 =for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
355 Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
357 =head1 Hash Manipulation Functions
359 =for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
360 Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
362 =for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
363 Like C<hv_store>, but takes a literal string instead of a string/length pair
364 and omits the hash parameter.
366 =head1 Lexer interface
368 =for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
370 Like L</lex_stuff_pvn>, but takes a literal string instead of a
376 /* concatenating with "" ensures that only literal strings are accepted as
378 #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
380 /* note that STR_WITH_LEN() can't be used as argument to macros or functions
381 * that under some configurations might be macros, which means that it requires
382 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
385 /* STR_WITH_LEN() shortcuts */
386 #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
387 #define newSVpvs_flags(str,flags) \
388 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
389 #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
390 #define sv_catpvs_flags(sv, str, flags) \
391 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
392 #define sv_catpvs_nomg(sv, str) \
393 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
394 #define sv_catpvs(sv, str) \
395 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
396 #define sv_catpvs_mg(sv, str) \
397 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
398 #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
399 #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
400 #define sv_setref_pvs(rv, classname, str) \
401 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
402 #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
403 #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
404 #define gv_stashpvs(str, create) \
405 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
406 #define gv_fetchpvs(namebeg, add, sv_type) \
407 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
408 #define gv_fetchpvn(namebeg, len, add, sv_type) \
409 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
410 #define sv_catxmlpvs(dsv, str, utf8) \
411 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
412 #define hv_fetchs(hv,key,lval) \
413 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
414 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
415 : HV_FETCH_JUST_SV, NULL, 0))
417 #define hv_stores(hv,key,val) \
418 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
419 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
421 #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
423 #define get_cvs(str, flags) \
424 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
427 =head1 Miscellaneous Functions
429 =for apidoc Am|bool|strNE|char* s1|char* s2
430 Test two strings to see if they are different. Returns true or
433 =for apidoc Am|bool|strEQ|char* s1|char* s2
434 Test two strings to see if they are equal. Returns true or false.
436 =for apidoc Am|bool|strLT|char* s1|char* s2
437 Test two strings to see if the first, C<s1>, is less than the second,
438 C<s2>. Returns true or false.
440 =for apidoc Am|bool|strLE|char* s1|char* s2
441 Test two strings to see if the first, C<s1>, is less than or equal to the
442 second, C<s2>. Returns true or false.
444 =for apidoc Am|bool|strGT|char* s1|char* s2
445 Test two strings to see if the first, C<s1>, is greater than the second,
446 C<s2>. Returns true or false.
448 =for apidoc Am|bool|strGE|char* s1|char* s2
449 Test two strings to see if the first, C<s1>, is greater than or equal to
450 the second, C<s2>. Returns true or false.
452 =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
453 Test two strings to see if they are different. The C<len> parameter
454 indicates the number of bytes to compare. Returns true or false. (A
455 wrapper for C<strncmp>).
457 =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
458 Test two strings to see if they are equal. The C<len> parameter indicates
459 the number of bytes to compare. Returns true or false. (A wrapper for
465 #define strNE(s1,s2) (strcmp(s1,s2))
466 #define strEQ(s1,s2) (!strcmp(s1,s2))
467 #define strLT(s1,s2) (strcmp(s1,s2) < 0)
468 #define strLE(s1,s2) (strcmp(s1,s2) <= 0)
469 #define strGT(s1,s2) (strcmp(s1,s2) > 0)
470 #define strGE(s1,s2) (strcmp(s1,s2) >= 0)
471 #define strnNE(s1,s2,l) (strncmp(s1,s2,l))
472 #define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
475 # define memNE(s1,s2,l) (memcmp(s1,s2,l))
476 # define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
478 # define memNE(s1,s2,l) (bcmp(s1,s2,l))
479 # define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
482 #define memEQs(s1, l, s2) \
483 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
484 #define memNEs(s1, l, s2) !memEQs(s1, l, s2)
489 * Unfortunately, the introduction of locales means that we
490 * can't trust isupper(), etc. to tell the truth. And when
491 * it comes to /\w+/ with tainting enabled, we *must* be able
492 * to trust our character classes.
494 * Therefore, the default tests in the text of Perl will be
495 * independent of locale. Any code that wants to depend on
496 * the current locale will use the tests that begin with "lc".
499 #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
507 =head1 Character classes
508 This section is about functions (really macros) that classify characters
509 into types, such as punctuation versus alphabetic, etc. Most of these are
510 analogous to regular expression character classes. (See
511 L<perlrecharclass/POSIX Character Classes>.) There are several variants for
512 each class. (Not all macros have all variants; each item below lists the
513 ones valid for it.) None are affected by C<use bytes>, and only the ones
514 with C<LC> in the name are affected by the current locale.
516 The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
517 C<U8>) as input and returns a boolean as to whether or not the character
518 represented by that octet is (or on non-ASCII platforms, corresponds to) an
519 ASCII character in the named class based on platform, Unicode, and Perl rules.
520 If the input is a number that doesn't fit in an octet, FALSE is returned.
522 Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
523 with no suffix C<"_A">.
525 Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
526 onto the platform. That is, the code points that are ASCII are unaffected,
527 since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
528 as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
529 true when called with the code point 0xDF, which is a word character in both
530 ASCII and EBCDIC (though it represents different characters in each).
532 Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
533 point as input. If the code point is larger than 255, Unicode rules are used
534 to determine if it is in the character class. For example,
535 C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
536 WITH MACRON in Unicode, and is a word character.
538 Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
539 (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
540 classification of just the first (possibly multi-byte) character in the string
543 Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but the
544 result is based on the current locale, which is what C<LC> in the name stands
545 for. If Perl can determine that the current locale is a UTF-8 locale, it uses
546 the published Unicode rules; otherwise, it uses the C library function that
547 gives the named classification. For example, C<isDIGIT_LC()> when not in a
548 UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always
549 returned if the input won't fit into an octet.
551 Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
552 returns the same as C<isFOO_LC> for input code points less than 256, and
553 returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
555 Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
556 (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
557 classification of just the first (possibly multi-byte) character in the string
560 =for apidoc Am|bool|isALPHA|char ch
561 Returns a boolean indicating whether the specified character is an
562 alphabetic character, analogous to C<m/[[:alpha:]]/>.
563 See the L<top of this section|/Character classes> for an explanation of variants
564 C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
565 C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
567 =for apidoc Am|bool|isALPHANUMERIC|char ch
568 Returns a boolean indicating whether the specified character is a either an
569 alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
570 See the L<top of this section|/Character classes> for an explanation of variants
571 C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
572 C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
573 C<isALPHANUMERIC_LC_utf8>.
575 =for apidoc Am|bool|isASCII|char ch
576 Returns a boolean indicating whether the specified character is one of the 128
577 characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
578 On non-ASCII platforms, it returns TRUE iff this
579 character corresponds to an ASCII character. Variants C<isASCII_A()> and
580 C<isASCII_L1()> are identical to C<isASCII()>.
581 See the L<top of this section|/Character classes> for an explanation of variants
582 C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
583 C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
584 library routine C<isascii()>. In these cases, the variants whose names contain
585 C<LC> are the same as the corresponding ones without.
587 Also note, that because all ASCII characters are UTF-8 invariant (meaning they
588 have the exact same representation (always a single byte) whether encoded in
589 UTF-8 or not), C<isASCII> will give the correct results when called with any
590 byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
591 work properly on any string encoded or not in UTF-8.
593 =for apidoc Am|bool|isBLANK|char ch
594 Returns a boolean indicating whether the specified character is a
595 character considered to be a blank, analogous to C<m/[[:blank:]]/>.
596 See the L<top of this section|/Character classes> for an explanation of variants
597 C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
598 C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
599 platforms do not have the C library routine C<isblank()>. In these cases, the
600 variants whose names contain C<LC> are the same as the corresponding ones
603 =for apidoc Am|bool|isCNTRL|char ch
604 Returns a boolean indicating whether the specified character is a
605 control character, analogous to C<m/[[:cntrl:]]/>.
606 See the L<top of this section|/Character classes> for an explanation of variants
607 C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
608 C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
609 On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
611 =for apidoc Am|bool|isDIGIT|char ch
612 Returns a boolean indicating whether the specified character is a
613 digit, analogous to C<m/[[:digit:]]/>.
614 Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
615 See the L<top of this section|/Character classes> for an explanation of variants
616 C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
619 =for apidoc Am|bool|isGRAPH|char ch
620 Returns a boolean indicating whether the specified character is a
621 graphic character, analogous to C<m/[[:graph:]]/>.
622 See the L<top of this section|/Character classes> for an explanation of variants
623 C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
624 C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
626 =for apidoc Am|bool|isLOWER|char ch
627 Returns a boolean indicating whether the specified character is a
628 lowercase character, analogous to C<m/[[:lower:]]/>.
629 See the L<top of this section|/Character classes> for an explanation of variants
630 C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
631 C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
633 =for apidoc Am|bool|isOCTAL|char ch
634 Returns a boolean indicating whether the specified character is an
636 The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
639 =for apidoc Am|bool|isPUNCT|char ch
640 Returns a boolean indicating whether the specified character is a
641 punctuation character, analogous to C<m/[[:punct:]]/>.
642 Note that the definition of what is punctuation isn't as
643 straightforward as one might desire. See L<perlrecharclass/POSIX Character
644 Classes> for details.
645 See the L<top of this section|/Character classes> for an explanation of variants
646 C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
647 C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
649 =for apidoc Am|bool|isSPACE|char ch
650 Returns a boolean indicating whether the specified character is a
651 whitespace character. This is analogous
652 to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
653 (experimentally), this also matches what C<m/[[:space:]]/> does.
654 ("Experimentally" means that this change may be backed out in 5.22 if
655 field experience indicates that it was unwise.) Prior to 5.18, only the
656 locale forms of this macro (the ones with C<LC> in their names) matched
657 precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
658 in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
659 (See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
660 See the L<top of this section|/Character classes> for an explanation of variants
661 C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
662 C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
664 =for apidoc Am|bool|isPSXSPC|char ch
665 (short for Posix Space)
666 Starting in 5.18, this is identical (experimentally) in all its forms to the
667 corresponding C<isSPACE()> macros. ("Experimentally" means that this change
668 may be backed out in 5.22 if field experience indicates that it
670 The locale forms of this macro are identical to their corresponding
671 C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
672 non-locale forms differ from their C<isSPACE()> forms only in that the
673 C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
674 Otherwise they are identical. Thus this macro is analogous to what
675 C<m/[[:space:]]/> matches in a regular expression.
676 See the L<top of this section|/Character classes> for an explanation of variants
677 C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
678 C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
680 =for apidoc Am|bool|isUPPER|char ch
681 Returns a boolean indicating whether the specified character is an
682 uppercase character, analogous to C<m/[[:upper:]]/>.
683 See the L<top of this section|/Character classes> for an explanation of variants
684 C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
685 C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
687 =for apidoc Am|bool|isPRINT|char ch
688 Returns a boolean indicating whether the specified character is a
689 printable character, analogous to C<m/[[:print:]]/>.
690 See the L<top of this section|/Character classes> for an explanation of variants
691 C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
692 C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
694 =for apidoc Am|bool|isWORDCHAR|char ch
695 Returns a boolean indicating whether the specified character is a character
696 that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
697 in a regular expression. A word character is an alphabetic character, a
698 decimal digit, a connecting punctuation character (such as an underscore), or
699 a "mark" character that attaches to one of those (like some sort of accent).
700 C<isALNUM()> is a synonym provided for backward compatibility, even though a
701 word character includes more than the standard C language meaning of
703 See the L<top of this section|/Character classes> for an explanation of variants
704 C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
705 C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
707 =for apidoc Am|bool|isXDIGIT|char ch
708 Returns a boolean indicating whether the specified character is a hexadecimal
709 digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
710 and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
711 See the L<top of this section|/Character classes> for an explanation of variants
712 C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
715 =for apidoc Am|bool|isIDFIRST|char ch
716 Returns a boolean indicating whether the specified character can be the first
717 character of an identifier. This is very close to, but not quite the same as
718 the official Unicode property C<XID_Start>. The difference is that this
719 returns true only if the input character also matches L</isWORDCHAR>.
720 See the L<top of this section|/Character classes> for an explanation of variants
721 C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
722 C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
724 =for apidoc Am|bool|isIDCONT|char ch
725 Returns a boolean indicating whether the specified character can be the
726 second or succeeding character of an identifier. This is very close to, but
727 not quite the same as the official Unicode property C<XID_Continue>. The
728 difference is that this returns true only if the input character also matches
729 L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
730 explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
731 C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
734 =head1 Miscellaneous Functions
736 =for apidoc Am|U8|READ_XDIGIT|char str*
737 Returns the value of an ASCII-range hex digit and advances the string pointer.
738 Behaviour is only well defined when isXDIGIT(*str) is true.
740 =head1 Character case changing
742 =for apidoc Am|U8|toUPPER|U8 ch
743 Converts the specified character to uppercase. If the input is anything but an
744 ASCII lowercase character, that input character itself is returned. Variant
745 C<toUPPER_A> is equivalent.
747 =for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
748 Converts the Unicode code point C<cp> to its uppercase version, and
749 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
750 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
751 bytes since the uppercase version may be longer than the original character.
753 The first code point of the uppercased version is returned
754 (but note, as explained just above, that there may be more.)
756 =for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
757 Converts the UTF-8 encoded character at C<p> to its uppercase version, and
758 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
759 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
760 bytes since the uppercase version may be longer than the original character.
762 The first code point of the uppercased version is returned
763 (but note, as explained just above, that there may be more.)
765 The input character at C<p> is assumed to be well-formed.
767 =for apidoc Am|U8|toFOLD|U8 ch
768 Converts the specified character to foldcase. If the input is anything but an
769 ASCII uppercase character, that input character itself is returned. Variant
770 C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
771 Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
773 =for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
774 Converts the Unicode code point C<cp> to its foldcase version, and
775 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
776 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
777 bytes since the foldcase version may be longer than the original character.
779 The first code point of the foldcased version is returned
780 (but note, as explained just above, that there may be more.)
782 =for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
783 Converts the UTF-8 encoded character at C<p> to its foldcase version, and
784 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
785 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
786 bytes since the foldcase version may be longer than the original character.
788 The first code point of the foldcased version is returned
789 (but note, as explained just above, that there may be more.)
791 The input character at C<p> is assumed to be well-formed.
793 =for apidoc Am|U8|toLOWER|U8 ch
794 Converts the specified character to lowercase. If the input is anything but an
795 ASCII uppercase character, that input character itself is returned. Variant
796 C<toLOWER_A> is equivalent.
798 =for apidoc Am|U8|toLOWER_L1|U8 ch
799 Converts the specified Latin1 character to lowercase. The results are undefined if
800 the input doesn't fit in a byte.
802 =for apidoc Am|U8|toLOWER_LC|U8 ch
803 Converts the specified character to lowercase using the current locale's rules,
804 if possible; otherwise returns the input character itself.
806 =for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
807 Converts the Unicode code point C<cp> to its lowercase version, and
808 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
809 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
810 bytes since the lowercase version may be longer than the original character.
812 The first code point of the lowercased version is returned
813 (but note, as explained just above, that there may be more.)
815 =for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
816 Converts the UTF-8 encoded character at C<p> to its lowercase version, and
817 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
818 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
819 bytes since the lowercase version may be longer than the original character.
821 The first code point of the lowercased version is returned
822 (but note, as explained just above, that there may be more.)
824 The input character at C<p> is assumed to be well-formed.
826 =for apidoc Am|U8|toLOWER_LC|U8 ch
827 Converts the specified character to lowercase using the current locale's rules,
828 if possible; otherwise returns the input character itself.
830 =for apidoc Am|U8|toTITLE|U8 ch
831 Converts the specified character to titlecase. If the input is anything but an
832 ASCII lowercase character, that input character itself is returned. Variant
833 C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
834 as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
835 concept used in locale handling, so there is no functionality for that.)
837 =for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
838 Converts the Unicode code point C<cp> to its titlecase version, and
839 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
840 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
841 bytes since the titlecase version may be longer than the original character.
843 The first code point of the titlecased version is returned
844 (but note, as explained just above, that there may be more.)
846 =for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
847 Converts the UTF-8 encoded character at C<p> to its titlecase version, and
848 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
849 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
850 bytes since the titlecase version may be longer than the original character.
852 The first code point of the titlecased version is returned
853 (but note, as explained just above, that there may be more.)
855 The input character at C<p> is assumed to be well-formed.
859 XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
860 really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
862 Note that these macros are repeated in Devel::PPPort, so should also be
863 patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
867 /* Specify the widest unsigned type on the platform. Use U64TYPE because U64
868 * is known only in the perl core, and this macro can be called from outside
871 # define WIDEST_UTYPE U64TYPE
873 # define WIDEST_UTYPE U32
876 /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
877 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
878 * bits of information are lost even on a 64-bit machine, but to get the
879 * compiler to optimize it out if possible. This is because Configure makes
880 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
881 * sizeof() guarantees that this evaluates to a constant true at compile time.
883 #define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
887 /* This returns the wrong results on at least z/OS unless this is
889 # error _ALL_SOURCE should probably be defined
892 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
893 * likely not defined, and so we use the native function */
894 # define isASCII(c) cBOOL(isascii(c))
896 # define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
899 #define isASCII_A(c) isASCII(c)
900 #define isASCII_L1(c) isASCII(c)
902 /* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
903 * and the 8 possible permutations of those bits exactly comprise the 8 octal
905 #define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
907 /* ASCII range only */
908 #ifdef H_PERL /* If have access to perl.h, lookup in its table */
910 /* Character class numbers. For internal core Perl use only. The ones less
911 * than 32 are used in PL_charclass[] and the ones up through the one that
912 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
913 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
914 * their actual definitions are here. If that file has a name not used here,
917 * The first group of these is ordered in what I (khw) estimate to be the
918 * frequency of their use. This gives a slight edge to exiting a loop earlier
919 * (in reginclass() in regexec.c) */
920 # define _CC_WORDCHAR 0 /* \w and [:word:] */
921 # define _CC_DIGIT 1 /* \d and [:digit:] */
922 # define _CC_ALPHA 2 /* [:alpha:] */
923 # define _CC_LOWER 3 /* [:lower:] */
924 # define _CC_UPPER 4 /* [:upper:] */
925 # define _CC_PUNCT 5 /* [:punct:] */
926 # define _CC_PRINT 6 /* [:print:] */
927 # define _CC_ALPHANUMERIC 7 /* [:alnum:] */
928 # define _CC_GRAPH 8 /* [:graph:] */
929 # define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
931 #define _FIRST_NON_SWASH_CC 10
932 /* The character classes above are implemented with swashes. The second group
933 * (just below) contains the ones implemented without. These are also sorted
934 * in rough order of the frequency of their use, except that \v should be last,
935 * as it isn't a real Posix character class, and some (small) inefficiencies in
936 * regular expression handling would be introduced by putting it in the middle
937 * of those that are. Also, cntrl and ascii come after the others as it may be
938 * useful to group these which have no members that match above Latin1, (or
939 * above ASCII in the latter case) */
941 # define _CC_SPACE 10 /* \s */
942 # define _CC_BLANK 11 /* [:blank:] */
943 # define _CC_XDIGIT 12 /* [:xdigit:] */
944 # define _CC_PSXSPC 13 /* [:space:] */
945 # define _CC_CNTRL 14 /* [:cntrl:] */
946 # define _CC_ASCII 15 /* [:ascii:] */
947 # define _CC_VERTSPACE 16 /* \v */
949 # define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
951 /* The members of the third group below do not need to be coordinated with data
952 * structures in regcomp.[ch] and regexec.c. But they should be added to
953 * bootstrap_ctype() */
954 # define _CC_IDFIRST 17
955 # define _CC_CHARNAME_CONT 18
956 # define _CC_NONLATIN1_FOLD 19
957 # define _CC_QUOTEMETA 20
958 # define _CC_NON_FINAL_FOLD 21
959 # define _CC_IS_IN_SOME_FOLD 22
960 # define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
962 * If more bits are needed, one could add a second word for non-64bit
963 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
964 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
965 * is used only for optimization (as of this writing), and differs in the
966 * Latin1 range from the ALPHA bit only in two relatively unimportant
967 * characters: the masculine and feminine ordinal indicators, so removing it
968 * would just cause /i regexes which match them to run less efficiently */
970 #if defined(PERL_CORE) || defined(PERL_EXT)
971 /* An enum version of the character class numbers, to help compilers
974 _CC_ENUM_ALPHA = _CC_ALPHA,
975 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
976 _CC_ENUM_ASCII = _CC_ASCII,
977 _CC_ENUM_BLANK = _CC_BLANK,
978 _CC_ENUM_CASED = _CC_CASED,
979 _CC_ENUM_CNTRL = _CC_CNTRL,
980 _CC_ENUM_DIGIT = _CC_DIGIT,
981 _CC_ENUM_GRAPH = _CC_GRAPH,
982 _CC_ENUM_LOWER = _CC_LOWER,
983 _CC_ENUM_PRINT = _CC_PRINT,
984 _CC_ENUM_PSXSPC = _CC_PSXSPC,
985 _CC_ENUM_PUNCT = _CC_PUNCT,
986 _CC_ENUM_SPACE = _CC_SPACE,
987 _CC_ENUM_UPPER = _CC_UPPER,
988 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
989 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
990 _CC_ENUM_XDIGIT = _CC_XDIGIT
991 } _char_class_number;
994 #define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
995 #define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
997 #if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
998 # if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
999 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
1000 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
1001 #error Need to adjust order of swash_property_names[]
1004 /* This is declared static in each of the few files that this is #defined for
1005 * to keep them from being publicly accessible. Hence there is a small amount
1006 * of wasted space */
1008 static const char* const swash_property_names[] = {
1023 EXTCONST U32 PL_charclass[] = {
1024 # include "l1_char_class_tab.h"
1027 # else /* ! DOINIT */
1028 EXTCONST U32 PL_charclass[];
1030 #endif /* Has perl.h */
1032 #if defined(H_PERL) && ! defined(BOOTSTRAP_CHARSET)
1034 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
1035 # define _CC_mask(classnum) (1U << (classnum))
1037 /* For internal core Perl use only: the base macro for defining macros like
1039 # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
1040 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
1042 /* The mask for the _A versions of the macros; it just adds in the bit for
1044 # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1046 /* For internal core Perl use only: the base macro for defining macros like
1047 * isALPHA_A. The foo_A version makes sure that both the desired bit and
1048 * the ASCII bit are present */
1049 # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
1050 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
1051 == _CC_mask_A(classnum)))
1053 # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
1054 # define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
1055 # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1056 # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1057 # define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
1058 # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1059 # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1060 # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1061 # define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1062 # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1063 # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1064 # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1065 # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1066 # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
1067 # define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
1068 # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1069 # define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1070 # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1072 /* continuation character for legal NAME in \N{NAME} */
1073 # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1075 # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1076 # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1077 # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1078 # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1079 # define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1080 # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1081 # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1082 # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1083 # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1084 # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
1086 /* Either participates in a fold with a character above 255, or is a
1087 * multi-char fold */
1088 # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_FOLD)))
1090 # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
1091 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1092 _generic_isCC(c, _CC_NON_FINAL_FOLD)
1093 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1094 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
1095 #else /* Either don't have perl.h or don't want to use char_class_tab.h */
1097 /* If we don't have perl.h, we are compiling a utility program. Below we
1098 * hard-code various macro definitions that wouldn't otherwise be available
1099 * to it. We can also get here if we are configured to bootstrap up Perl
1100 * on a non-ASCII platform that doesn't have a working Perl (currently only
1101 * EBCDIC). For these we currently use the native definitions to get
1102 * things going. (It should also be possible to use the translation
1103 * function NATIVE_TO_LATIN1(), but that is an extra layer of dependence on
1104 * Perl, so it is currently avoided for the macros where it's possible to
1107 /* Use the native functions. They likely will return false for all
1108 * non-ASCII values, but this makes sure */
1109 # define isALPHA_A(c) (isASCII(c) && isalpha(c))
1110 # define isALPHANUMERIC_A(c) (isASCII(c) && isalnum(c))
1111 # define isCNTRL_A(c) (isASCII(c) && iscntrl(c))
1112 # define isDIGIT_A(c) (isASCII(c) && isdigit(c))
1113 # define isGRAPH_A(c) (isASCII(c) && isgraph(c))
1114 # define isLOWER_A(c) (isASCII(c) && islower(c))
1115 # define isPRINT_A(c) (isASCII(c) && isprint(c))
1116 # define isPUNCT_A(c) (isASCII(c) && ispunct(c))
1117 # define isSPACE_A(c) (isASCII(c) && isspace(c))
1118 # define isUPPER_A(c) (isASCII(c) && isupper(c))
1119 # define isXDIGIT_A(c) (isASCII(c) && isxdigit(c))
1120 # else /* ASCII platform. These are coded based on first principals */
1121 # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
1122 # define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
1123 # define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
1124 # define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
1125 # define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
1126 # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1127 # define isPRINT_A(c) (((c) >= 32 && (c) < 127))
1128 # define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1129 # define isSPACE_A(c) ((c) == ' ' \
1135 # define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
1136 # define isXDIGIT_A(c) (isDIGIT_A(c) \
1137 || ((c) >= 'a' && (c) <= 'f') \
1138 || ((c) <= 'F' && (c) >= 'A'))
1139 # endif /* Below are common definitions for ASCII and non-ASCII */
1140 # define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1141 # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1142 # define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1144 /* The _L1 macros may be unnecessary for both the utilities and for
1145 * bootstrapping; I (khw) added them during debugging of bootstrapping, and
1146 * it seems best to keep them. */
1147 # define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1148 # define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1149 # define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1150 # define isBLANK_L1(c) (isBLANK_A(c) \
1151 || (FITS_IN_8_BITS(c) \
1152 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1153 # define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1154 # define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1155 # define isLOWER_L1(c) (isLOWER_A(c) \
1156 || (FITS_IN_8_BITS(c) \
1157 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1158 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1159 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1160 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1161 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1162 # define isPRINT_L1(c) (isPRINT_A(c) \
1163 || (FITS_IN_8_BITS(c) \
1164 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1165 # define isPSXSPC_L1(c) isSPACE_L1(c)
1166 # define isPUNCT_L1(c) (isPUNCT_A(c) \
1167 || (FITS_IN_8_BITS(c) \
1168 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1169 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1170 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1171 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1172 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1173 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1174 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1175 # define isSPACE_L1(c) (isSPACE_A(c) \
1176 || (FITS_IN_8_BITS(c) \
1177 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1178 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1179 # define isUPPER_L1(c) (isUPPER_A(c) \
1180 || (FITS_IN_8_BITS(c) \
1181 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1182 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1183 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1184 # define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1185 # define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1186 # define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1191 /* The following are not fully accurate in the above-ASCII range. I (khw)
1192 * don't think it's necessary to be so for the purposes where this gets
1194 # define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1195 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1197 /* And these aren't accurate at all. They are useful only for above
1198 * Latin1, which utilities and bootstrapping don't deal with */
1199 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1200 # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1202 /* Many of the macros later in this file are defined in terms of these. By
1203 * implementing them with a function, which converts the class number into
1204 * a call to the desired macro, all of the later ones work. However, that
1205 * function won't be actually defined when building a utility program (no
1206 * perl.h), and so a compiler error will be generated if one is attempted
1207 * to be used. And the above-Latin1 code points require Unicode tables to
1208 * be present, something unlikely to be the case when bootstrapping */
1209 # define _generic_isCC(c, classnum) \
1210 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1211 # define _generic_isCC_A(c, classnum) \
1212 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
1213 #endif /* End of no perl.h or have BOOTSTRAP_CHARSET */
1215 #define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1216 #define isALPHA(c) isALPHA_A(c)
1217 #define isBLANK(c) isBLANK_A(c)
1218 #define isCNTRL(c) isCNTRL_A(c)
1219 #define isDIGIT(c) isDIGIT_A(c)
1220 #define isGRAPH(c) isGRAPH_A(c)
1221 #define isIDFIRST(c) isIDFIRST_A(c)
1222 #define isLOWER(c) isLOWER_A(c)
1223 #define isPRINT(c) isPRINT_A(c)
1224 #define isPSXSPC(c) isPSXSPC_A(c)
1225 #define isPUNCT(c) isPUNCT_A(c)
1226 #define isSPACE(c) isSPACE_A(c)
1227 #define isUPPER(c) isUPPER_A(c)
1228 #define isWORDCHAR(c) isWORDCHAR_A(c)
1229 #define isXDIGIT(c) isXDIGIT_A(c)
1231 /* ASCII casing. These could also be written as
1232 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1233 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1234 which uses table lookup and mask instead of subtraction. (This would
1235 work because the _MOD does not apply in the ASCII range) */
1236 #define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1237 #define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
1239 /* In the ASCII range, these are equivalent to what they're here defined to be.
1240 * But by creating these definitions, other code doesn't have to be aware of
1242 #define toFOLD(c) toLOWER(c)
1243 #define toTITLE(c) toUPPER(c)
1245 #define toLOWER_A(c) toLOWER(c)
1246 #define toUPPER_A(c) toUPPER(c)
1247 #define toFOLD_A(c) toFOLD(c)
1248 #define toTITLE_A(c) toTITLE(c)
1250 /* Use table lookup for speed; returns the input itself if is out-of-range */
1251 #define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
1253 : PL_latin1_lc[ (U8) (c) ])
1254 #define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1256 /* Modified uc. Is correct uc except for three non-ascii chars which are
1257 * all mapped to one of them, and these need special handling; returns the
1258 * input itself if is out-of-range */
1259 #define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
1261 : PL_mod_latin1_uc[ (U8) (c) ])
1262 #define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale
1264 /* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
1266 /* For internal core Perl use only: the base macro for defining macros like
1267 * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point
1268 * (0-255) to check. In a UTF-8 locale, the result is the same as calling
1269 * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like
1270 * _CC_UPPER, which gives the class number for doing this. For non-UTF-8
1271 * locales, the code to actually do the test this is passed in 'non_utf8'. If
1272 * 'c' is above 255, 0 is returned. For accessing the full range of possible
1273 * code points under locale rules, use the macros based on _generic_LC_uvchr
1274 * instead of this. */
1275 #define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \
1276 (! FITS_IN_8_BITS(c) \
1278 : IN_UTF8_CTYPE_LOCALE \
1279 ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \
1282 /* For internal core Perl use only: a helper macro for defining macros like
1283 * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to
1284 * actually do this test is passed in 'non_utf8_func', which is called on 'c',
1285 * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See
1286 * _generic_LC_base for more info */
1287 #define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \
1288 _generic_LC_base(c,utf8_locale_classnum, \
1289 non_utf8_func( (_LC_CAST) (c)))
1291 /* For internal core Perl use only: like _generic_LC, but also returns TRUE if
1292 * 'c' is the platform's native underscore character */
1293 #define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \
1294 _generic_LC_base(c, utf8_locale_classnum, \
1295 (non_utf8_func( (_LC_CAST) (c)) \
1296 || (char)(c) == '_'))
1298 /* These next three are also for internal core Perl use only: case-change
1300 #define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \
1302 : (IN_UTF8_CTYPE_LOCALE) \
1303 ? PL_latin1_lc[ (U8) (c) ] \
1304 : function((cast)(c)))
1306 /* Note that the result can be larger than a byte in a UTF-8 locale. It
1307 * returns a single value, so can't adequately return the upper case of LATIN
1308 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1309 * values "SS"); instead it asserts against that under DEBUGGING, and
1310 * otherwise returns its input */
1311 #define _generic_toUPPER_LC(c, function, cast) \
1312 (! FITS_IN_8_BITS(c) \
1314 : ((! IN_UTF8_CTYPE_LOCALE) \
1315 ? function((cast)(c)) \
1316 : ((((U8)(c)) == MICRO_SIGN) \
1317 ? GREEK_CAPITAL_LETTER_MU \
1318 : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \
1319 ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \
1320 : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \
1321 ? (__ASSERT_(0) (c)) \
1322 : PL_mod_latin1_uc[ (U8) (c) ])))))
1324 /* Note that the result can be larger than a byte in a UTF-8 locale. It
1325 * returns a single value, so can't adequately return the fold case of LATIN
1326 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1327 * values "ss"); instead it asserts against that under DEBUGGING, and
1328 * otherwise returns its input */
1329 #define _generic_toFOLD_LC(c, function, cast) \
1330 ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \
1331 ? GREEK_SMALL_LETTER_MU \
1332 : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \
1333 || (c) != LATIN_SMALL_LETTER_SHARP_S) \
1334 _generic_toLOWER_LC(c, function, cast)))
1336 /* Use the libc versions for these if available. */
1337 #if defined(HAS_ISASCII) && ! defined(USE_NEXT_CTYPE)
1338 # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c)))
1340 # define isASCII_LC(c) isASCII(c)
1343 #if defined(HAS_ISBLANK) && ! defined(USE_NEXT_CTYPE)
1344 # define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank)
1345 #else /* Unlike isASCII, varies if in a UTF-8 locale */
1346 # define isBLANK_LC(c) (IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c)
1349 #ifdef USE_NEXT_CTYPE /* NeXT computers */
1351 # define _LC_CAST unsigned int /* Needed by _generic_LC. NeXT functions
1352 use this as their input type */
1354 # define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, NXIsAlpha)
1355 # define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, NXIsAlNum)
1356 # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, NXIsCntrl)
1357 # define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, NXIsDigit)
1358 # define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, NXIsGraph)
1359 # define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, NXIsAlpha)
1360 # define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, NXIsLower)
1361 # define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, NXIsPrint)
1362 # define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, NXIsPunct)
1363 # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, NXIsSpace)
1364 # define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, NXIsUpper)
1365 # define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, NXIsAlNum)
1366 # define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, NXIsXdigit)
1368 # define toLOWER_LC(c) _generic_toLOWER_LC((c), NXToLower, unsigned int)
1369 # define toUPPER_LC(c) _generic_toUPPER_LC((c), NXToUpper, unsigned int)
1370 # define toFOLD_LC(c) _generic_toFOLD_LC((c), NXToLower, unsigned int)
1372 #else /* !USE_NEXT_CTYPE */
1374 # define _LC_CAST U8
1376 # if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1377 /* For most other platforms */
1379 # define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha)
1380 # define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum)
1381 # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1382 # define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit)
1383 # define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph)
1384 # define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha)
1385 # define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower)
1386 # define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint)
1387 # define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct)
1388 # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1389 # define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper)
1390 # define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum)
1391 # define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit)
1394 # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1395 # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1396 # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
1398 # else /* The final fallback position */
1400 # define isALPHA_LC(c) (isascii(c) && isalpha(c))
1401 # define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
1402 # define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
1403 # define isDIGIT_LC(c) (isascii(c) && isdigit(c))
1404 # define isGRAPH_LC(c) (isascii(c) && isgraph(c))
1405 # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1406 # define isLOWER_LC(c) (isascii(c) && islower(c))
1407 # define isPRINT_LC(c) (isascii(c) && isprint(c))
1408 # define isPUNCT_LC(c) (isascii(c) && ispunct(c))
1409 # define isSPACE_LC(c) (isascii(c) && isspace(c))
1410 # define isUPPER_LC(c) (isascii(c) && isupper(c))
1411 # define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
1412 # define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
1414 # define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
1415 # define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
1416 # define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c))
1419 #endif /* USE_NEXT_CTYPE */
1421 #define isIDCONT(c) isWORDCHAR(c)
1422 #define isIDCONT_A(c) isWORDCHAR_A(c)
1423 #define isIDCONT_L1(c) isWORDCHAR_L1(c)
1424 #define isIDCONT_LC(c) isWORDCHAR_LC(c)
1425 #define isPSXSPC_LC(c) isSPACE_LC(c)
1427 /* For internal core Perl use only: the base macros for defining macros like
1428 * isALPHA_uni. 'c' is the code point to check. 'classnum' is the POSIX class
1429 * number defined earlier in this file. _generic_uni() is used for POSIX
1430 * classes where there is a macro or function 'above_latin1' that takes the
1431 * single argument 'c' and returns the desired value. These exist for those
1432 * classes which have simple definitions, avoiding the overhead of a hash
1433 * lookup or inversion list binary search. _generic_swash_uni() can be used
1434 * for classes where that overhead is faster than a direct lookup.
1435 * _generic_uni() won't compile if 'c' isn't unsigned, as it won't match the
1436 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so
1437 * have duplicate checks here, so could create versions of the macros that
1438 * don't, but experiments show that gcc optimizes them out anyway. */
1440 /* Note that all ignore 'use bytes' */
1441 #define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1442 ? _generic_isCC(c, classnum) \
1444 #define _generic_swash_uni(classnum, c) ((c) < 256 \
1445 ? _generic_isCC(c, classnum) \
1446 : _is_uni_FOO(classnum, c))
1447 #define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1448 #define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
1449 #define isASCII_uni(c) isASCII(c)
1450 #define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1451 #define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
1452 #define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1453 #define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
1454 #define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
1455 #define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
1456 #define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1457 #define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
1459 /* Posix and regular space are identical above Latin1 */
1460 #define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
1462 #define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
1463 #define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
1464 #define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
1465 #define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
1466 #define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
1467 #define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
1469 #define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
1470 #define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1471 #define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1472 #define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
1474 /* For internal core Perl use only: the base macros for defining macros like
1475 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code
1476 * point, not just 0-255. Like _generic_uni, there are two versions, one for
1477 * simple class definitions; the other for more complex. These are like
1478 * _generic_uni, so see it for more info. */
1479 #define _generic_LC_uvchr(latin1, above_latin1, c) \
1480 (c < 256 ? latin1(c) : above_latin1(c))
1481 #define _generic_LC_swash_uvchr(latin1, classnum, c) \
1482 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1484 #define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1485 #define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1486 _CC_ALPHANUMERIC, c)
1487 #define isASCII_LC_uvchr(c) isASCII_LC(c)
1488 #define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
1489 #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
1490 #define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1491 #define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
1492 #define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1493 _is_uni_perl_idcont, c)
1494 #define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1495 _is_uni_perl_idstart, c)
1496 #define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1497 #define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
1498 #define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1499 space under locale */
1500 #define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1501 #define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
1502 is_XPERLSPACE_cp_high, c)
1503 #define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1504 #define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1506 #define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
1508 #define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
1510 /* For internal core Perl use only: the base macros for defining macros like
1511 * isALPHA_utf8. These are like the earlier defined macros, but take an input
1512 * UTF-8 encoded string 'p'. If the input is in the Latin1 range, use
1513 * the Latin1 macro 'classnum' on 'p'. Otherwise use the value given by the
1514 * 'utf8' parameter. This relies on the fact that ASCII characters have the
1515 * same representation whether utf8 or not. Note that it assumes that the utf8
1516 * has been validated, and ignores 'use bytes' */
1517 #define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
1518 ? _generic_isCC(*(p), classnum) \
1519 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1521 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
1525 /* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1527 #define _generic_func_utf8(classnum, above_latin1, p) \
1528 _generic_utf8(classnum, p, above_latin1(p))
1529 /* Like the above, but passes classnum to _isFOO_utf8(), instead of having an
1530 * 'above_latin1' parameter */
1531 #define _generic_swash_utf8(classnum, p) \
1532 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
1534 /* Like the above, but should be used only when it is known that there are no
1535 * characters in the range 128-255 which the class is TRUE for. Hence it can
1536 * skip the tests for this range. 'above_latin1' should include its arguments */
1537 #define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
1538 (UTF8_IS_INVARIANT(*(p)) \
1539 ? _generic_isCC(*(p), classnum) \
1540 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
1544 /* NOTE that some of these macros have very similar ones in regcharclass.h.
1545 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1546 * there, differing in name only by an underscore from the one here
1547 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1548 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1549 * points; the regcharclass.h ones are implemented as a series of
1550 * "if-else-if-else ..." */
1552 #define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1553 #define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
1554 #define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1555 utf8, the non-utf8 macro works
1557 #define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1558 #define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1559 #define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1560 _is_utf8_FOO(_CC_DIGIT, p))
1561 #define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
1562 #define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
1563 _is_utf8_perl_idcont, p)
1565 /* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1566 * IDFIRST is an alnum. See
1567 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
1568 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1569 * This used to be not the XID version, but we decided to go with the more
1570 * modern Unicode definition */
1571 #define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
1572 _is_utf8_perl_idstart, p)
1574 #define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1575 #define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
1577 /* Posix and regular space are identical above Latin1 */
1578 #define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
1580 #define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
1581 #define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
1582 #define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
1583 #define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
1584 #define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
1585 #define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1588 #define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
1589 #define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1590 #define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1591 #define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
1593 /* For internal core Perl use only: the base macros for defining macros like
1594 * isALPHA_LC_utf8. These are like _generic_utf8, but if the first code point
1595 * in 'p' is within the 0-255 range, it uses locale rules from the passed-in
1596 * 'macro' parameter */
1597 #define _generic_LC_utf8(macro, p, utf8) \
1598 (UTF8_IS_INVARIANT(*(p)) \
1600 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1601 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
1604 #define _generic_LC_swash_utf8(macro, classnum, p) \
1605 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1606 #define _generic_LC_func_utf8(macro, above_latin1, p) \
1607 _generic_LC_utf8(macro, p, above_latin1(p))
1609 #define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1610 _CC_ALPHANUMERIC, p)
1611 #define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
1612 #define isASCII_LC_utf8(p) isASCII_LC(*p)
1613 #define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1614 #define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1615 #define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1616 #define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
1617 #define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
1618 #define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1619 #define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1620 #define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
1621 #define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1622 space under locale */
1623 #define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1624 #define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1625 #define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1626 #define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1628 #define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
1630 /* Macros for backwards compatibility and for completeness when the ASCII and
1631 * Latin1 values are identical */
1632 #define isALPHAU(c) isALPHA_L1(c)
1633 #define isDIGIT_L1(c) isDIGIT_A(c)
1634 #define isOCTAL(c) isOCTAL_A(c)
1635 #define isOCTAL_L1(c) isOCTAL_A(c)
1636 #define isXDIGIT_L1(c) isXDIGIT_A(c)
1637 #define isALNUM(c) isWORDCHAR(c)
1638 #define isALNUMU(c) isWORDCHAR_L1(c)
1639 #define isALNUM_LC(c) isWORDCHAR_LC(c)
1640 #define isALNUM_uni(c) isWORDCHAR_uni(c)
1641 #define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1642 #define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1643 #define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
1644 #define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1645 #define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1646 #define isALNUMC(c) isALPHANUMERIC(c)
1647 #define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1648 #define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1649 #define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1650 #define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1651 #define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
1653 /* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII,
1654 * except that they don't necessarily mean the same characters, e.g. CTRL-D is
1655 * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC.
1656 * '?' is special-cased on EBCDIC to APC, which is the control there that is
1657 * the outlier from the block that contains the other controls, just like
1658 * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0
1659 * block. If it weren't special cased, it would yield a non-control.
1660 * The conversion works both ways, so CTRL('D') is 4, and CTRL(4) is D, etc. */
1662 # define toCTRL(c) (toUPPER(c) ^ 64)
1664 # define toCTRL(c) ((c) == '?' \
1665 ? LATIN1_TO_NATIVE(0x9F) \
1666 : (c) == LATIN1_TO_NATIVE(0x9F) \
1668 : (NATIVE_TO_LATIN1(toUPPER(c)) ^ 64))
1671 /* Line numbers are unsigned, 32 bits. */
1673 #define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
1675 /* Helpful alias for version prescan */
1676 #define is_LAX_VERSION(a,b) \
1677 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1679 #define is_STRICT_VERSION(a,b) \
1680 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1682 #define BADVERSION(a,b,c) \
1688 /* Converts a character known to represent a hexadecimal digit (0-9, A-F, or
1689 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer,
1690 * which is advanced. The input is validated only by an assert() in DEBUGGING
1691 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and
1692 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */
1693 #define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \
1696 #define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \
1700 /* Converts a character known to represent an octal digit (0-7) to its numeric
1701 * value. The input is validated only by an assert() in DEBUGGING builds. In
1702 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */
1703 #define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c)))
1706 =head1 Memory Management
1708 =for apidoc Am|void|Newx|void* ptr|int nitems|type
1709 The XSUB-writer's interface to the C C<malloc> function.
1711 Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1713 In 5.9.3, Newx() and friends replace the older New() API, and drops
1714 the first parameter, I<x>, a debug aid which allowed callers to identify
1715 themselves. This aid has been superseded by a new build option,
1716 PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
1717 there for use in XS modules supporting older perls.
1719 =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
1720 The XSUB-writer's interface to the C C<malloc> function, with
1721 cast. See also C<Newx>.
1723 Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1725 =for apidoc Am|void|Newxz|void* ptr|int nitems|type
1726 The XSUB-writer's interface to the C C<malloc> function. The allocated
1727 memory is zeroed with C<memzero>. See also C<Newx>.
1729 Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1731 =for apidoc Am|void|Renew|void* ptr|int nitems|type
1732 The XSUB-writer's interface to the C C<realloc> function.
1734 Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1736 =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1737 The XSUB-writer's interface to the C C<realloc> function, with
1740 Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1742 =for apidoc Am|void|Safefree|void* ptr
1743 The XSUB-writer's interface to the C C<free> function.
1745 This should B<ONLY> be used on memory obtained using L<"Newx"> and friends.
1747 =for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1748 The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
1749 source, C<dest> is the destination, C<nitems> is the number of items, and
1750 C<type> is the type. Can do overlapping moves. See also C<Copy>.
1752 =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1753 Like C<Move> but returns dest. Useful
1754 for encouraging compilers to tail-call
1757 =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1758 The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
1759 source, C<dest> is the destination, C<nitems> is the number of items, and
1760 C<type> is the type. May fail on overlapping copies. See also C<Move>.
1762 =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1764 Like C<Copy> but returns dest. Useful
1765 for encouraging compilers to tail-call
1768 =for apidoc Am|void|Zero|void* dest|int nitems|type
1770 The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1771 destination, C<nitems> is the number of items, and C<type> is the type.
1773 =for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1775 Like C<Zero> but returns dest. Useful
1776 for encouraging compilers to tail-call
1779 =for apidoc Am|void|StructCopy|type *src|type *dest|type
1780 This is an architecture-independent macro to copy one structure to another.
1782 =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1784 Fill up memory with a byte pattern (a byte repeated over and over
1785 again) that hopefully catches attempts to access uninitialized memory.
1787 =for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1789 PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1791 =for apidoc Am|void|PoisonFree|void* dest|int nitems|type
1793 PoisonWith(0xEF) for catching access to freed memory.
1795 =for apidoc Am|void|Poison|void* dest|int nitems|type
1797 PoisonWith(0xEF) for catching access to freed memory.
1801 /* Maintained for backwards-compatibility only. Use newSV() instead. */
1803 #define NEWSV(x,len) newSV(len)
1806 #define MEM_SIZE_MAX ((MEM_SIZE)~0)
1808 /* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1809 * overly eager compilers that will bleat about e.g.
1810 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
1811 #ifdef PERL_MALLOC_WRAP
1812 #define MEM_WRAP_CHECK(n,t) \
1813 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (croak_memory_wrap(),0))
1814 #define MEM_WRAP_CHECK_1(n,t,a) \
1815 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (Perl_croak_nocontext("%s",(a)),0))
1816 #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
1818 #define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (croak_memory_wrap(),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1821 #define MEM_WRAP_CHECK(n,t)
1822 #define MEM_WRAP_CHECK_1(n,t,a)
1823 #define MEM_WRAP_CHECK_2(n,t,a,b)
1824 #define MEM_WRAP_CHECK_(n,t)
1826 #define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1832 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
1833 * go through functions, which are handy for debugging breakpoints, but
1834 * which more importantly get the immediate calling environment (file and
1835 * line number, and C function name if available) passed in. This info can
1836 * then be used for logging the calls, for which one gets a sample
1837 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
1840 * - not all memory allocs get logged, only those
1841 * that go through Newx() and derivatives (while all
1842 * Safefrees do get logged)
1843 * - __FILE__ and __LINE__ do not work everywhere
1844 * - __func__ or __FUNCTION__ even less so
1845 * - I think more goes on after the perlio frees but
1846 * the thing is that STDERR gets closed (as do all
1847 * the file descriptors)
1848 * - no deeper calling stack than the caller of the Newx()
1849 * or the kind, but do I look like a C reflection/introspection
1851 * - the function prototypes for the logging functions
1852 * probably should maybe be somewhere else than handy.h
1853 * - one could consider inlining (macrofying) the logging
1854 * for speed, but I am too lazy
1855 * - one could imagine recording the allocations in a hash,
1856 * (keyed by the allocation address?), and maintain that
1857 * through reallocs and frees, but how to do that without
1858 * any News() happening...?
1859 * - lots of -Ddefines to get useful/controllable output
1860 * - lots of ENV reads
1863 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);
1865 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);
1867 PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
1870 # ifndef PERL_MEM_LOG_NOIMPL
1879 # if defined(PERL_IN_SV_C) /* those are only used in sv.c */
1880 void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1881 void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1888 #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1889 #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
1890 #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
1893 #ifndef MEM_LOG_ALLOC
1894 #define MEM_LOG_ALLOC(n,t,a) (a)
1896 #ifndef MEM_LOG_REALLOC
1897 #define MEM_LOG_REALLOC(n,t,v,a) (a)
1899 #ifndef MEM_LOG_FREE
1900 #define MEM_LOG_FREE(a) (a)
1903 #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1904 #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1905 #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
1908 /* pre 5.9.x compatibility */
1909 #define New(x,v,n,t) Newx(v,n,t)
1910 #define Newc(x,v,n,t,c) Newxc(v,n,t,c)
1911 #define Newz(x,v,n,t) Newxz(v,n,t)
1914 #define Renew(v,n,t) \
1915 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1916 #define Renewc(v,n,t,c) \
1917 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1920 #define Safefree(d) \
1921 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
1923 #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
1926 #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1927 #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1928 #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
1930 #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1931 #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1933 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
1935 /* Using bzero(), which returns void. */
1936 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
1939 #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1940 #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1941 #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1942 #define Poison(d,n,t) PoisonFree(d,n,t)
1945 # define PERL_POISON_EXPR(x) x
1947 # define PERL_POISON_EXPR(x)
1950 #ifdef USE_STRUCT_COPY
1951 #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
1953 #define StructCopy(s,d,t) Copy(s,d,1,t)
1956 /* C_ARRAY_LENGTH is the number of elements in the C array (so you
1957 * want your zero-based indices to be less than but not equal to).
1959 * C_ARRAY_END is one past the last: half-open/half-closed range,
1960 * not last-inclusive range. */
1961 #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
1962 #define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a))
1966 # define Perl_va_copy(s, d) va_copy(d, s)
1968 # if defined(__va_copy)
1969 # define Perl_va_copy(s, d) __va_copy(d, s)
1971 # define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1976 /* convenience debug macros */
1978 #define pTHX_FORMAT "Perl interpreter: 0x%p"
1979 #define pTHX__FORMAT ", Perl interpreter: 0x%p"
1980 #define pTHX_VALUE_ (void *)my_perl,
1981 #define pTHX_VALUE (void *)my_perl
1982 #define pTHX__VALUE_ ,(void *)my_perl,
1983 #define pTHX__VALUE ,(void *)my_perl
1986 #define pTHX__FORMAT
1989 #define pTHX__VALUE_
1991 #endif /* USE_ITHREADS */
1993 /* Perl_deprecate was not part of the public API, and did not have a deprecate()
1994 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1995 CPAN::Unpack show any users outside the core. */
1997 # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2000 /* Internal macros to deal with gids and uids */
2003 # if Uid_t_size > IVSIZE
2004 # define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
2005 # define SvUID(sv) SvNV(sv)
2007 # if Uid_t_sign <= 0
2008 # define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
2009 # define SvUID(sv) SvIV(sv)
2011 # define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
2012 # define SvUID(sv) SvUV(sv)
2014 # endif /* Uid_t_size */
2016 # if Gid_t_size > IVSIZE
2017 # define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
2018 # define SvGID(sv) SvNV(sv)
2020 # if Gid_t_sign <= 0
2021 # define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
2022 # define SvGID(sv) SvIV(sv)
2024 # define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
2025 # define SvGID(sv) SvUV(sv)
2027 # endif /* Gid_t_size */
2031 #endif /* HANDY_H */
2035 * c-indentation-style: bsd
2037 * indent-tabs-mode: nil
2040 * ex: set ts=8 sts=4 sw=4 et: