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 #ifndef HANDY_H /* Guard against nested #inclusion */
14 #if !defined(__STDC__)
22 # define Null(type) ((type)NULL)
27 =for apidoc AmU||Nullch
28 Null character pointer. (No longer available when C<PERL_CORE> is defined.)
30 =for apidoc AmU||Nullsv
31 Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
36 # define Nullch Null(char*)
37 # define Nullfp Null(PerlIO*)
38 # define Nullsv Null(SV*)
50 /* The MUTABLE_*() macros cast pointers to the types shown, in such a way
51 * (compiler permitting) that casting away const-ness will give a warning;
55 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
56 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
59 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
60 # define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
62 # define MUTABLE_PTR(p) ((void *) (p))
65 #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
66 #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
67 #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
68 #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
69 #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
70 #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
79 /* bool is built-in for g++-2.6.3 and later, which might be used
80 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
81 be sure _G_config.h will be included before this file. _G_config.h
82 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
83 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
84 g++ can be identified by __GNUG__.
85 Andy Dougherty February 2000
87 #ifdef __GNUG__ /* GNU g++ has bool built-in */
93 /* The NeXT dynamic loader headers will not build with the bool macro
94 So declare them now to clear confusion.
96 #if defined(NeXT) || defined(__NeXT__)
99 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
103 # endif /* !HAS_BOOL */
104 #endif /* NeXT || __NeXT__ */
115 /* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
116 * defined as char for example, then the cast from int is
117 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
119 #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
121 /* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
122 * XXX Should really be a Configure probe, with HAS__FUNCTION__
123 * and FUNCTION__ as results.
124 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
125 #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
126 # define FUNCTION__ __func__
128 # if (defined(_MSC_VER) && _MSC_VER < 1300) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
129 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
130 # define FUNCTION__ ""
132 # define FUNCTION__ __FUNCTION__ /* Common extension. */
136 /* XXX A note on the perl source internal type system. The
137 original intent was that I32 be *exactly* 32 bits.
139 Currently, we only guarantee that I32 is *at least* 32 bits.
140 Specifically, if int is 64 bits, then so is I32. (This is the case
141 for the Cray.) This has the advantage of meshing nicely with
142 standard library calls (where we pass an I32 and the library is
143 expecting an int), but the disadvantage that an I32 is not 32 bits.
144 Andy Dougherty August 1996
146 There is no guarantee that there is *any* integral type with
147 exactly 32 bits. It is perfectly legal for a system to have
148 sizeof(short) == sizeof(int) == sizeof(long) == 8.
150 Similarly, there is no guarantee that I16 and U16 have exactly 16
153 For dealing with issues that may arise from various 32/64-bit
154 systems, we will ask Configure to check out
156 SHORTSIZE == sizeof(short)
157 INTSIZE == sizeof(int)
158 LONGSIZE == sizeof(long)
159 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
160 PTRSIZE == sizeof(void *)
161 DOUBLESIZE == sizeof(double)
162 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
166 #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
167 # include <inttypes.h>
168 # ifdef INT32_MIN_BROKEN
170 # define INT32_MIN (-2147483647-1)
172 # ifdef INT64_MIN_BROKEN
174 # define INT64_MIN (-9223372036854775807LL-1)
189 #endif /* PERL_CORE */
191 #if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
192 # if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
193 # define PeRl_INT64_C(c) CAT2(c,LL)
194 # define PeRl_UINT64_C(c) CAT2(c,ULL)
196 # if QUADKIND == QUAD_IS___INT64
197 # define PeRl_INT64_C(c) CAT2(c,I64)
198 # define PeRl_UINT64_C(c) CAT2(c,UI64)
200 # if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
201 # define PeRl_INT64_C(c) CAT2(c,L)
202 # define PeRl_UINT64_C(c) CAT2(c,UL)
204 # define PeRl_INT64_C(c) ((I64TYPE)(c))
205 # define PeRl_UINT64_C(c) ((U64TYPE)(c))
210 # define UINT64_C(c) PeRl_UINT64_C(c)
213 # define INT64_C(c) PeRl_INT64_C(c)
217 #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
219 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
220 Please search CHAR_MAX in perl.h for further details. */
221 #define U8_MAX UINT8_MAX
222 #define U8_MIN UINT8_MIN
224 #define I16_MAX INT16_MAX
225 #define I16_MIN INT16_MIN
226 #define U16_MAX UINT16_MAX
227 #define U16_MIN UINT16_MIN
229 #define I32_MAX INT32_MAX
230 #define I32_MIN INT32_MIN
231 #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
232 # define U32_MAX UINT32_MAX
234 # define U32_MAX 4294967295U
236 #define U32_MIN UINT32_MIN
240 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
241 Please search CHAR_MAX in perl.h for further details. */
242 #define U8_MAX PERL_UCHAR_MAX
243 #define U8_MIN PERL_UCHAR_MIN
245 #define I16_MAX PERL_SHORT_MAX
246 #define I16_MIN PERL_SHORT_MIN
247 #define U16_MAX PERL_USHORT_MAX
248 #define U16_MIN PERL_USHORT_MIN
251 # define I32_MAX PERL_INT_MAX
252 # define I32_MIN PERL_INT_MIN
253 # define U32_MAX PERL_UINT_MAX
254 # define U32_MIN PERL_UINT_MIN
256 # define I32_MAX PERL_LONG_MAX
257 # define I32_MIN PERL_LONG_MIN
258 # define U32_MAX PERL_ULONG_MAX
259 # define U32_MIN PERL_ULONG_MIN
264 /* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
265 #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
266 #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
267 #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
269 #define Ctl(ch) ((ch) & 037)
272 =head1 SV-Body Allocation
274 =for apidoc Ama|SV*|newSVpvs|const char* s
275 Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
277 =for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
278 Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
281 =for apidoc Ama|SV*|newSVpvs_share|const char* s
282 Like C<newSVpvn_share>, but takes a literal string instead of a string/length
283 pair and omits the hash parameter.
285 =for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
286 Like C<sv_catpvn_flags>, but takes a literal string instead of a
289 =for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
290 Like C<sv_catpvn_nomg>, but takes a literal string instead of a
293 =for apidoc Am|void|sv_catpvs|SV* sv|const char* s
294 Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
296 =for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
297 Like C<sv_catpvn_mg>, but takes a literal string instead of a
300 =for apidoc Am|void|sv_setpvs|SV* sv|const char* s
301 Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
303 =for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
304 Like C<sv_setpvn_mg>, but takes a literal string instead of a
307 =for apidoc Am|SV *|sv_setref_pvs|const char* s
308 Like C<sv_setref_pvn>, but takes a literal string instead of a
311 =head1 Memory Management
313 =for apidoc Ama|char*|savepvs|const char* s
314 Like C<savepvn>, but takes a literal string instead of a string/length pair.
316 =for apidoc Ama|char*|savesharedpvs|const char* s
317 A version of C<savepvs()> which allocates the duplicate string in memory
318 which is shared between threads.
322 =for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
323 Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
325 =head1 Hash Manipulation Functions
327 =for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
328 Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
330 =for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
331 Like C<hv_store>, but takes a literal string instead of a string/length pair
332 and omits the hash parameter.
334 =head1 Lexer interface
336 =for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
338 Like L</lex_stuff_pvn>, but takes a literal string instead of a
344 /* concatenating with "" ensures that only literal strings are accepted as
346 #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
348 /* note that STR_WITH_LEN() can't be used as argument to macros or functions
349 * that under some configurations might be macros, which means that it requires
350 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
353 /* STR_WITH_LEN() shortcuts */
354 #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
355 #define newSVpvs_flags(str,flags) \
356 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
357 #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
358 #define sv_catpvs_flags(sv, str, flags) \
359 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
360 #define sv_catpvs_nomg(sv, str) \
361 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
362 #define sv_catpvs(sv, str) \
363 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
364 #define sv_catpvs_mg(sv, str) \
365 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
366 #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
367 #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
368 #define sv_setref_pvs(rv, classname, str) \
369 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
370 #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
371 #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
372 #define gv_stashpvs(str, create) \
373 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
374 #define gv_fetchpvs(namebeg, add, sv_type) \
375 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
376 #define gv_fetchpvn(namebeg, len, add, sv_type) \
377 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
378 #define sv_catxmlpvs(dsv, str, utf8) \
379 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
380 #define hv_fetchs(hv,key,lval) \
381 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
382 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
383 : HV_FETCH_JUST_SV, NULL, 0))
385 #define hv_stores(hv,key,val) \
386 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
387 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
389 #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
391 #define get_cvs(str, flags) \
392 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
395 =head1 Miscellaneous Functions
397 =for apidoc Am|bool|strNE|char* s1|char* s2
398 Test two strings to see if they are different. Returns true or
401 =for apidoc Am|bool|strEQ|char* s1|char* s2
402 Test two strings to see if they are equal. Returns true or false.
404 =for apidoc Am|bool|strLT|char* s1|char* s2
405 Test two strings to see if the first, C<s1>, is less than the second,
406 C<s2>. Returns true or false.
408 =for apidoc Am|bool|strLE|char* s1|char* s2
409 Test two strings to see if the first, C<s1>, is less than or equal to the
410 second, C<s2>. Returns true or false.
412 =for apidoc Am|bool|strGT|char* s1|char* s2
413 Test two strings to see if the first, C<s1>, is greater than the second,
414 C<s2>. Returns true or false.
416 =for apidoc Am|bool|strGE|char* s1|char* s2
417 Test two strings to see if the first, C<s1>, is greater than or equal to
418 the second, C<s2>. Returns true or false.
420 =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
421 Test two strings to see if they are different. The C<len> parameter
422 indicates the number of bytes to compare. Returns true or false. (A
423 wrapper for C<strncmp>).
425 =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
426 Test two strings to see if they are equal. The C<len> parameter indicates
427 the number of bytes to compare. Returns true or false. (A wrapper for
433 #define strNE(s1,s2) (strcmp(s1,s2))
434 #define strEQ(s1,s2) (!strcmp(s1,s2))
435 #define strLT(s1,s2) (strcmp(s1,s2) < 0)
436 #define strLE(s1,s2) (strcmp(s1,s2) <= 0)
437 #define strGT(s1,s2) (strcmp(s1,s2) > 0)
438 #define strGE(s1,s2) (strcmp(s1,s2) >= 0)
439 #define strnNE(s1,s2,l) (strncmp(s1,s2,l))
440 #define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
443 # define memNE(s1,s2,l) (memcmp(s1,s2,l))
444 # define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
446 # define memNE(s1,s2,l) (bcmp(s1,s2,l))
447 # define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
450 #define memEQs(s1, l, s2) \
451 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
452 #define memNEs(s1, l, s2) !memEQs(s1, l, s2)
457 * Unfortunately, the introduction of locales means that we
458 * can't trust isupper(), etc. to tell the truth. And when
459 * it comes to /\w+/ with tainting enabled, we *must* be able
460 * to trust our character classes.
462 * Therefore, the default tests in the text of Perl will be
463 * independent of locale. Any code that wants to depend on
464 * the current locale will use the tests that begin with "lc".
467 #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
475 =head1 Character classes
476 This section is about functions (really macros) that classify characters
477 into types, such as punctuation versus alphabetic, etc. Most of these are
478 analogous to regular expression character classes. (See
479 L<perlrecharclass/POSIX Character Classes>.) There are several variants for
480 each class. (Not all macros have all variants; each item below lists the
481 ones valid for it.) None are affected by C<use bytes>, and only the ones
482 with C<LC> in the name are affected by the current locale.
484 The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
485 C<U8>) as input and returns a boolean as to whether or not the character
486 represented by that octet is (or on non-ASCII platforms, corresponds to) an
487 ASCII character in the named class based on platform, Unicode, and Perl rules.
488 If the input is a number that doesn't fit in an octet, FALSE is returned.
490 Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
491 with no suffix C<"_A">.
493 Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
494 onto the platform. That is, the code points that are ASCII are unaffected,
495 since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
496 as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
497 true when called with the code point 0xDF, which is a word character in both
498 ASCII and EBCDIC (though it represent different characters in each).
500 Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
501 point as input. If the code point is larger than 255, Unicode rules are used
502 to determine if it is in the character class. For example,
503 C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
504 WITH MACRON in Unicode, and is a word character.
506 Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
507 (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
508 classification of just the first (possibly multi-byte) character in the string
511 Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but uses
512 the C library function that gives the named classification instead of
513 hard-coded rules. For example, C<isDIGIT_LC()> returns the result of calling
514 C<isdigit()>. This means that the result is based on the current locale, which
515 is what C<LC> in the name stands for. FALSE is always returned if the input
516 won't fit into an octet.
518 Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
519 returns the same as C<isFOO_LC> for input code points less than 256, and
520 returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
522 Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
523 (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
524 classification of just the first (possibly multi-byte) character in the string
527 =for apidoc Am|bool|isALPHA|char ch
528 Returns a boolean indicating whether the specified character is an
529 alphabetic character, analogous to C<m/[[:alpha:]]/>.
530 See the L<top of this section|/Character classes> for an explanation of variants
531 C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
532 C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
534 =for apidoc Am|bool|isALPHANUMERIC|char ch
535 Returns a boolean indicating whether the specified character is a either an
536 alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
537 See the L<top of this section|/Character classes> for an explanation of variants
538 C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
539 C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
540 C<isALPHANUMERIC_LC_utf8>.
542 =for apidoc Am|bool|isASCII|char ch
543 Returns a boolean indicating whether the specified character is one of the 128
544 characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
545 On non-ASCII platforms, it returns TRUE iff this
546 character corresponds to an ASCII character. Variants C<isASCII_A()> and
547 C<isASCII_L1()> are identical to C<isASCII()>.
548 See the L<top of this section|/Character classes> for an explanation of variants
549 C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
550 C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
551 library routine C<isascii()>. In these cases, the variants whose names contain
552 C<LC> are the same as the corresponding ones without.
554 Also note, that because all ASCII characters are UTF-8 invariant (meaning they
555 have the exact same representation (always a single byte) whether encoded in
556 UTF-8 or not), C<isASCII> will give the correct results when called with any
557 byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
558 work properly on any string encoded or not in UTF-8.
560 =for apidoc Am|bool|isBLANK|char ch
561 Returns a boolean indicating whether the specified character is a
562 character considered to be a blank, analogous to C<m/[[:blank:]]/>.
563 See the L<top of this section|/Character classes> for an explanation of variants
564 C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
565 C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
566 platforms do not have the C library routine C<isblank()>. In these cases, the
567 variants whose names contain C<LC> are the same as the corresponding ones
570 =for apidoc Am|bool|isCNTRL|char ch
571 Returns a boolean indicating whether the specified character is a
572 control character, analogous to C<m/[[:cntrl:]]/>.
573 See the L<top of this section|/Character classes> for an explanation of variants
574 C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
575 C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
576 On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
578 =for apidoc Am|bool|isDIGIT|char ch
579 Returns a boolean indicating whether the specified character is a
580 digit, analogous to C<m/[[:digit:]]/>.
581 Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
582 See the L<top of this section|/Character classes> for an explanation of variants
583 C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
586 =for apidoc Am|bool|isGRAPH|char ch
587 Returns a boolean indicating whether the specified character is a
588 graphic character, analogous to C<m/[[:graph:]]/>.
589 See the L<top of this section|/Character classes> for an explanation of variants
590 C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
591 C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
593 =for apidoc Am|bool|isLOWER|char ch
594 Returns a boolean indicating whether the specified character is a
595 lowercase character, analogous to C<m/[[:lower:]]/>.
596 See the L<top of this section|/Character classes> for an explanation of variants
597 C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
598 C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
600 =for apidoc Am|bool|isOCTAL|char ch
601 Returns a boolean indicating whether the specified character is an
603 The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
606 =for apidoc Am|bool|isPUNCT|char ch
607 Returns a boolean indicating whether the specified character is a
608 punctuation character, analogous to C<m/[[:punct:]]/>.
609 Note that the definition of what is punctuation isn't as
610 straightforward as one might desire. See L<perlrecharclass/POSIX Character
611 Classes> for details.
612 See the L<top of this section|/Character classes> for an explanation of variants
613 C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
614 C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
616 =for apidoc Am|bool|isSPACE|char ch
617 Returns a boolean indicating whether the specified character is a
618 whitespace character. This is analogous
619 to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
620 (experimentally), this also matches what C<m/[[:space:]]/> does.
621 ("Experimentally" means that this change may be backed out in 5.20 or 5.22 if
622 field experience indicates that it was unwise.) Prior to 5.18, only the
623 locale forms of this macro (the ones with C<LC> in their names) matched
624 precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
625 in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
626 (See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
627 See the L<top of this section|/Character classes> for an explanation of variants
628 C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
629 C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
631 =for apidoc Am|bool|isPSXSPC|char ch
632 (short for Posix Space)
633 Starting in 5.18, this is identical (experimentally) in all its forms to the
634 corresponding C<isSPACE()> macros. ("Experimentally" means that this change
635 may be backed out in 5.20 or 5.22 if field experience indicates that it
637 The locale forms of this macro are identical to their corresponding
638 C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
639 non-locale forms differ from their C<isSPACE()> forms only in that the
640 C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
641 Otherwise they are identical. Thus this macro is analogous to what
642 C<m/[[:space:]]/> matches in a regular expression.
643 See the L<top of this section|/Character classes> for an explanation of variants
644 C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
645 C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
647 =for apidoc Am|bool|isUPPER|char ch
648 Returns a boolean indicating whether the specified character is an
649 uppercase character, analogous to C<m/[[:upper:]]/>.
650 See the L<top of this section|/Character classes> for an explanation of variants
651 C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
652 C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
654 =for apidoc Am|bool|isPRINT|char ch
655 Returns a boolean indicating whether the specified character is a
656 printable character, analogous to C<m/[[:print:]]/>.
657 See the L<top of this section|/Character classes> for an explanation of variants
658 C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
659 C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
661 =for apidoc Am|bool|isWORDCHAR|char ch
662 Returns a boolean indicating whether the specified character is a character
663 that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
664 in a regular expression. A word character is an alphabetic character, a
665 decimal digit, a connecting punctuation character (such as an underscore), or
666 a "mark" character that attaches to one of those (like some sort of accent).
667 C<isALNUM()> is a synonym provided for backward compatibility, even though a
668 word character includes more than the standard C language meaning of
670 See the L<top of this section|/Character classes> for an explanation of variants
671 C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
672 C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
674 =for apidoc Am|bool|isXDIGIT|char ch
675 Returns a boolean indicating whether the specified character is a hexadecimal
676 digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
677 and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
678 See the L<top of this section|/Character classes> for an explanation of variants
679 C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
682 =for apidoc Am|bool|isIDFIRST|char ch
683 Returns a boolean indicating whether the specified character can be the first
684 character of an identifier. This is very close to, but not quite the same as
685 the official Unicode property C<XID_Start>. The difference is that this
686 returns true only if the input character also matches L</isWORDCHAR>.
687 See the L<top of this section|/Character classes> for an explanation of variants
688 C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
689 C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
691 =for apidoc Am|bool|isIDCONT|char ch
692 Returns a boolean indicating whether the specified character can be the
693 second or succeeding character of an identifier. This is very close to, but
694 not quite the same as the official Unicode property C<XID_Continue>. The
695 difference is that this returns true only if the input character also matches
696 L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
697 explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
698 C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
701 =head1 Miscellaneous Functions
703 =for apidoc Am|U8|READ_XDIGIT|char str*
704 Returns the value of an ASCII-range hex digit and advances the string pointer.
705 Behaviour is only well defined when isXDIGIT(*str) is true.
707 =head1 Character case changing
709 =for apidoc Am|U8|toUPPER|U8 ch
710 Converts the specified character to uppercase. If the input is anything but an
711 ASCII lowercase character, that input character itself is returned. Variant
712 C<toUPPER_A> is equivalent.
714 =for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
715 Converts the Unicode code point C<cp> to its uppercase version, and
716 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
717 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
718 bytes since the uppercase version may be longer than the original character.
720 The first code point of the uppercased version is returned
721 (but note, as explained just above, that there may be more.)
723 =for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
724 Converts the UTF-8 encoded character at C<p> to its uppercase version, and
725 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
726 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
727 bytes since the uppercase version may be longer than the original character.
729 The first code point of the uppercased version is returned
730 (but note, as explained just above, that there may be more.)
732 The input character at C<p> is assumed to be well-formed.
734 =for apidoc Am|U8|toFOLD|U8 ch
735 Converts the specified character to foldcase. If the input is anything but an
736 ASCII uppercase character, that input character itself is returned. Variant
737 C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
738 Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
740 =for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
741 Converts the Unicode code point C<cp> to its foldcase version, and
742 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
743 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
744 bytes since the foldcase version may be longer than the original character.
746 The first code point of the foldcased version is returned
747 (but note, as explained just above, that there may be more.)
749 =for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
750 Converts the UTF-8 encoded character at C<p> to its foldcase version, and
751 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
752 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
753 bytes since the foldcase version may be longer than the original character.
755 The first code point of the foldcased version is returned
756 (but note, as explained just above, that there may be more.)
758 The input character at C<p> is assumed to be well-formed.
760 =for apidoc Am|U8|toLOWER|U8 ch
761 Converts the specified character to lowercase. If the input is anything but an
762 ASCII uppercase character, that input character itself is returned. Variant
763 C<toLOWER_A> is equivalent.
765 =for apidoc Am|U8|toLOWER_L1|U8 ch
766 Converts the specified Latin1 character to lowercase. The results are undefined if
767 the input doesn't fit in a byte.
769 =for apidoc Am|U8|toLOWER_LC|U8 ch
770 Converts the specified character to lowercase using the current locale's rules,
771 if possible; otherwise returns the input character itself.
773 =for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
774 Converts the Unicode code point C<cp> to its lowercase 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 lowercase version may be longer than the original character.
779 The first code point of the lowercased version is returned
780 (but note, as explained just above, that there may be more.)
782 =for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
783 Converts the UTF-8 encoded character at C<p> to its lowercase 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 lowercase version may be longer than the original character.
788 The first code point of the lowercased 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_LC|U8 ch
794 Converts the specified character to lowercase using the current locale's rules,
795 if possible; otherwise returns the input character itself.
797 =for apidoc Am|U8|toTITLE|U8 ch
798 Converts the specified character to titlecase. If the input is anything but an
799 ASCII lowercase character, that input character itself is returned. Variant
800 C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
801 as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
802 concept used in locale handling, so there is no functionality for that.)
804 =for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
805 Converts the Unicode code point C<cp> to its titlecase version, and
806 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
807 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
808 bytes since the titlecase version may be longer than the original character.
810 The first code point of the titlecased version is returned
811 (but note, as explained just above, that there may be more.)
813 =for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
814 Converts the UTF-8 encoded character at C<p> to its titlecase version, and
815 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
816 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
817 bytes since the titlecase version may be longer than the original character.
819 The first code point of the titlecased version is returned
820 (but note, as explained just above, that there may be more.)
822 The input character at C<p> is assumed to be well-formed.
826 XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
827 really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
829 Note that these macros are repeated in Devel::PPPort, so should also be
830 patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
834 /* Specify the widest unsigned type on the platform. Use U64TYPE because U64
835 * is known only in the perl core, and this macro can be called from outside
838 # define WIDEST_UTYPE U64TYPE
840 # define WIDEST_UTYPE U32
843 /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
844 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
845 * bits of information are lost even on a 64-bit machine, but to get the
846 * compiler to optimize it out if possible. This is because Configure makes
847 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
848 * sizeof() guarantees that this evaluates to a constant true at compile time.
850 #define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
854 /* This returns the wrong results on at least z/OS unless this is
856 # error _ALL_SOURCE should probably be defined
859 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
860 * likely not defined, and so we use the native function */
861 # define isASCII(c) isascii(c)
863 # define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
866 #define isASCII_A(c) isASCII(c)
867 #define isASCII_L1(c) isASCII(c)
869 /* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
870 * and the 8 possible permutations of those bits exactly comprise the 8 octal
872 #define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
874 /* ASCII range only */
875 #ifdef H_PERL /* If have access to perl.h, lookup in its table */
877 /* Character class numbers. For internal core Perl use only. The ones less
878 * than 32 are used in PL_charclass[] and the ones up through the one that
879 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
880 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
881 * their actual definitions are here. If that file has a name not used here,
884 * The first group of these is ordered in what I (khw) estimate to be the
885 * frequency of their use. This gives a slight edge to exiting a loop earlier
886 * (in reginclass() in regexec.c) */
887 # define _CC_WORDCHAR 0 /* \w and [:word:] */
888 # define _CC_DIGIT 1 /* \d and [:digit:] */
889 # define _CC_ALPHA 2 /* [:alpha:] */
890 # define _CC_LOWER 3 /* [:lower:] */
891 # define _CC_UPPER 4 /* [:upper:] */
892 # define _CC_PUNCT 5 /* [:punct:] */
893 # define _CC_PRINT 6 /* [:print:] */
894 # define _CC_ALPHANUMERIC 7 /* [:alnum:] */
895 # define _CC_GRAPH 8 /* [:graph:] */
896 # define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
898 #define _FIRST_NON_SWASH_CC 10
899 /* The character classes above are implemented with swashes. The second group
900 * (just below) contains the ones implemented without. These are also sorted
901 * in rough order of the frequency of their use, except that \v should be last,
902 * as it isn't a real Posix character class, and some (small) inefficiencies in
903 * regular expression handling would be introduced by putting it in the middle
904 * of those that are. Also, cntrl and ascii come after the others as it may be
905 * useful to group these which have no members that match above Latin1, (or
906 * above ASCII in the latter case) */
908 # define _CC_SPACE 10 /* \s */
909 # define _CC_BLANK 11 /* [:blank:] */
910 # define _CC_XDIGIT 12 /* [:xdigit:] */
911 # define _CC_PSXSPC 13 /* [:space:] */
912 # define _CC_CNTRL 14 /* [:cntrl:] */
913 # define _CC_ASCII 15 /* [:ascii:] */
914 # define _CC_VERTSPACE 16 /* \v */
916 # define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
918 /* The members of the third group below do not need to be coordinated with data
919 * structures in regcomp.[ch] and regexec.c */
920 # define _CC_IDFIRST 17
921 # define _CC_CHARNAME_CONT 18
922 # define _CC_NONLATIN1_FOLD 19
923 # define _CC_QUOTEMETA 20
924 # define _CC_NON_FINAL_FOLD 21
925 # define _CC_IS_IN_SOME_FOLD 22
926 # define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
928 * If more bits are needed, one could add a second word for non-64bit
929 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
930 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
931 * is used only for optimization (as of this writing), and differs in the
932 * Latin1 range from the ALPHA bit only in two relatively unimportant
933 * characters: the masculine and feminine ordinal indicators, so removing it
934 * would just cause /i regexes which match them to run less efficiently */
936 #if defined(PERL_CORE) || defined(PERL_EXT)
937 /* An enum version of the character class numbers, to help compilers
940 _CC_ENUM_ALPHA = _CC_ALPHA,
941 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
942 _CC_ENUM_ASCII = _CC_ASCII,
943 _CC_ENUM_BLANK = _CC_BLANK,
944 _CC_ENUM_CASED = _CC_CASED,
945 _CC_ENUM_CNTRL = _CC_CNTRL,
946 _CC_ENUM_DIGIT = _CC_DIGIT,
947 _CC_ENUM_GRAPH = _CC_GRAPH,
948 _CC_ENUM_LOWER = _CC_LOWER,
949 _CC_ENUM_PRINT = _CC_PRINT,
950 _CC_ENUM_PSXSPC = _CC_PSXSPC,
951 _CC_ENUM_PUNCT = _CC_PUNCT,
952 _CC_ENUM_SPACE = _CC_SPACE,
953 _CC_ENUM_UPPER = _CC_UPPER,
954 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
955 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
956 _CC_ENUM_XDIGIT = _CC_XDIGIT
957 } _char_class_number;
960 #define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
961 #define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
963 #if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
964 # if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
965 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
966 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
967 #error Need to adjust order of swash_property_names[]
970 /* This is declared static in each of the few files that this is #defined for
971 * to keep them from being publicly accessible. Hence there is a small amount
974 static const char* const swash_property_names[] = {
989 EXTCONST U32 PL_charclass[] = {
990 # include "l1_char_class_tab.h"
993 # else /* ! DOINIT */
994 EXTCONST U32 PL_charclass[];
997 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
998 # define _CC_mask(classnum) (1U << (classnum))
999 # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
1000 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
1002 /* The mask for the _A versions of the macros; it just adds in the bit for
1004 # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1006 /* The _A version makes sure that both the desired bit and the ASCII bit
1008 # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
1009 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
1010 == _CC_mask_A(classnum)))
1012 # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
1013 # define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
1014 # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1015 # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1016 # define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
1017 # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1018 # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1019 # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1020 # define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1021 # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1022 # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1023 # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1024 # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1025 # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
1026 # define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST))
1028 /* Either participates in a fold with a character above 255, or is a
1029 * multi-char fold */
1030 # 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)))
1032 # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
1033 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1034 _generic_isCC(c, _CC_NON_FINAL_FOLD)
1035 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1036 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
1037 #else /* No perl.h. */
1038 # define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1039 # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1040 # define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1041 # define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
1043 /* We could be called without perl.h, so the native functions are the
1044 * easiest to code these in. They likely will return false for all
1045 * non-ASCII values, but this makes sure */
1046 # define isALPHA_A(c) (isASCII(c) && isalpha(c))
1047 # define isALPHANUMERIC_A(c) (isASCII(c) && isalnum(c))
1048 # define isCNTRL_A(c) (isASCII(c) && iscntrl(c))
1049 # define isDIGIT_A(c) (isASCII(c) && isdigit(c))
1050 # define isGRAPH_A(c) (isASCII(c) && isgraph(c))
1051 # define isLOWER_A(c) (isASCII(c) && islower(c))
1052 # define isPRINT_A(c) (isASCII(c) && isprint(c))
1053 # define isPUNCT_A(c) (isASCII(c) && ispunct(c))
1054 # define isSPACE_A(c) (isASCII(c) && isspace(c))
1055 # define isUPPER_A(c) (isASCII(c) && isupper(c))
1056 # define isXDIGIT_A(c) (isASCII(c) && isxdigit(c))
1057 # else /* ASCII platform, no perl.h */
1058 # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
1059 # define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
1060 # define isCNTRL_A(c) (isASCII(c) && ! isPRINT_A(c))
1061 # define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
1062 # define isGRAPH_A(c) (isPRINT_A(c) && (c) ! = ' ')
1063 # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1064 # define isPRINT_A(c) (((c) >= 32 && (c) < 127))
1065 # define isPUNCT_A(c) (isGRAPH_A(c) && ! isALPHANUMERIC(c))
1066 # define isSPACE_A(c) ((c) == ' ' \
1071 # define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
1072 # define isXDIGIT_A(c) (isDIGIT_A(c) \
1073 || ((c) >= 'a' && (c) <= 'f') \
1074 || ((c) <= 'F' && (c) >= 'A'))
1076 #endif /* ASCII range definitions */
1078 /* Latin1 definitions */
1080 # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1081 # define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1082 # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1084 /* continuation character for legal NAME in \N{NAME} */
1085 # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1087 # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1088 # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1089 # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1090 # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1091 # define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1092 # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1093 # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1094 # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1095 # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1096 # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
1097 #else /* No access to perl.h. Only a few provided here, just in case needed
1098 * for backwards compatibility */
1099 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
1100 * >= AA test to speed up ASCII-only tests at the expense of the others */
1101 # define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_LATIN1((U8) c) >= 0xAA \
1102 && ((NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1103 && NATIVE_TO_LATIN1((U8) c) != 0xD7 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1104 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1105 || NATIVE_TO_LATIN1((U8) c) == 0xB5 \
1106 || NATIVE_TO_LATIN1((U8) c) == 0xBA)))
1107 # define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1113 || NATIVE_TO_LATIN1((U8) c) == 0xA0)
1116 #define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1117 #define isALPHA(c) isALPHA_A(c)
1118 #define isBLANK(c) isBLANK_A(c)
1119 #define isCNTRL(c) isCNTRL_A(c)
1120 #define isDIGIT(c) isDIGIT_A(c)
1121 #define isGRAPH(c) isGRAPH_A(c)
1122 #define isIDFIRST(c) isIDFIRST_A(c)
1123 #define isLOWER(c) isLOWER_A(c)
1124 #define isPRINT(c) isPRINT_A(c)
1125 #define isPSXSPC(c) isPSXSPC_A(c)
1126 #define isPUNCT(c) isPUNCT_A(c)
1127 #define isSPACE(c) isSPACE_A(c)
1128 #define isUPPER(c) isUPPER_A(c)
1129 #define isWORDCHAR(c) isWORDCHAR_A(c)
1130 #define isXDIGIT(c) isXDIGIT_A(c)
1132 /* ASCII casing. These could also be written as
1133 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1134 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1135 which uses table lookup and mask instead of subtraction. (This would
1136 work because the _MOD does not apply in the ASCII range) */
1137 #define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
1138 #define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
1140 /* In the ASCII range, these are equivalent to what they're here defined to be.
1141 * But by creating these definitions, other code doesn't have to be aware of
1143 #define toFOLD(c) toLOWER(c)
1144 #define toFOLD_LC(c) toLOWER_LC(c)
1145 #define toTITLE(c) toUPPER(c)
1147 #define toLOWER_A(c) toLOWER(c)
1148 #define toUPPER_A(c) toUPPER(c)
1149 #define toFOLD_A(c) toFOLD(c)
1150 #define toTITLE_A(c) toTITLE(c)
1152 /* Use table lookup for speed; return error character for input
1154 #define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
1156 : PL_latin1_lc[ (U8) (c) ])
1157 #define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1159 /* Modified uc. Is correct uc except for three non-ascii chars which are
1160 * all mapped to one of them, and these need special handling; error
1161 * character for input out-of-range */
1162 #define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
1164 : PL_mod_latin1_uc[ (U8) (c) ])
1165 #ifdef USE_NEXT_CTYPE
1167 # define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c))
1168 # define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
1169 # define isASCII_LC(c) isASCII((unsigned int)(c))
1170 # define isBLANK_LC(c) isBLANK((unsigned int)(c))
1171 # define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
1172 # define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
1173 # define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
1174 # define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
1175 # define isLOWER_LC(c) NXIsLower((unsigned int)(c))
1176 # define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
1177 # define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
1178 # define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
1179 # define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
1180 # define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
1181 # define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
1182 # define toLOWER_LC(c) NXToLower((unsigned int)(c))
1183 # define toUPPER_LC(c) NXToUpper((unsigned int)(c))
1185 #else /* !USE_NEXT_CTYPE */
1187 # if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1189 /* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
1191 # define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
1192 # define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \
1193 && isalnum((unsigned char)(c)))
1195 # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
1197 # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
1200 # define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
1202 # define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
1204 # define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
1205 # define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
1206 # define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
1207 # define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1208 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1209 # define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
1210 # define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1211 # define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
1212 # define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1213 # define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1214 # define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1215 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
1216 # define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
1217 # define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? tolower((unsigned char)(c)) : (c))
1218 # define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? toupper((unsigned char)(c)) : (c))
1222 # define isALPHA_LC(c) (isascii(c) && isalpha(c))
1223 # define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
1224 # define isASCII_LC(c) isascii(c)
1226 # define isBLANK_LC(c) (isascii(c) && isblank(c))
1228 # define isBLANK_LC(c) isBLANK_A(c)
1230 # define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
1231 # define isDIGIT_LC(c) (isascii(c) && isdigit(c))
1232 # define isGRAPH_LC(c) (isascii(c) && isgraph(c))
1233 # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1234 # define isLOWER_LC(c) (isascii(c) && islower(c))
1235 # define isPRINT_LC(c) (isascii(c) && isprint(c))
1236 # define isPUNCT_LC(c) (isascii(c) && ispunct(c))
1237 # define isSPACE_LC(c) (isascii(c) && isspace(c))
1238 # define isUPPER_LC(c) (isascii(c) && isupper(c))
1239 # define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
1240 # define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
1241 # define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
1242 # define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
1245 #endif /* USE_NEXT_CTYPE */
1247 #define isIDCONT(c) isWORDCHAR(c)
1248 #define isIDCONT_A(c) isWORDCHAR_A(c)
1249 #define isIDCONT_L1(c) isWORDCHAR_L1(c)
1250 #define isIDCONT_LC(c) isWORDCHAR_LC(c)
1251 #define isPSXSPC_LC(c) isSPACE_LC(c)
1253 /* For internal core Perl use only. If the input is Latin1, use the Latin1
1254 * macro; otherwise use the function 'above_latin1'. Won't compile if 'c' isn't unsigned, as
1255 * won't match above_latin1 prototype. The macros do bounds checking, so have
1256 * duplicate checks here, so could create versions of the macros that don't,
1257 * but experiments show that gcc optimizes them out anyway. */
1259 /* Note that all ignore 'use bytes' */
1260 #define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1261 ? _generic_isCC(c, classnum) \
1263 #define _generic_swash_uni(classnum, c) ((c) < 256 \
1264 ? _generic_isCC(c, classnum) \
1265 : _is_uni_FOO(classnum, c))
1266 #define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1267 #define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
1268 #define isASCII_uni(c) isASCII(c)
1269 #define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1270 #define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
1271 #define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1272 #define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
1273 #define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
1274 #define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
1275 #define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1276 #define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
1278 /* Posix and regular space are identical above Latin1 */
1279 #define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
1281 #define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
1282 #define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
1283 #define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
1284 #define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
1285 #define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
1286 #define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
1288 #define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
1289 #define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1290 #define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1291 #define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
1293 #define _generic_LC_uvchr(latin1, above_latin1, c) \
1294 (c < 256 ? latin1(c) : above_latin1(c))
1295 #define _generic_LC_swash_uvchr(latin1, classnum, c) \
1296 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1298 #define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1299 #define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1300 _CC_ALPHANUMERIC, c)
1301 #define isASCII_LC_uvchr(c) isASCII_LC(c)
1302 #define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
1303 #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
1304 #define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1305 #define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
1306 #define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1307 _is_uni_perl_idcont, c)
1308 #define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1309 _is_uni_perl_idstart, c)
1310 #define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1311 #define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
1312 #define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1313 space under locale */
1314 #define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1315 #define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
1316 is_XPERLSPACE_cp_high, c)
1317 #define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1318 #define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1320 #define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
1323 #define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
1325 /* Everything whose name begins with an underscore is for internal core Perl
1328 /* If the input is in the Latin1 range, use
1329 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1330 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1331 * fact that ASCII characters have the same representation whether utf8 or not.
1332 * Note that it assumes that the utf8 has been validated, and ignores 'use
1334 #define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
1335 ? _generic_isCC(*(p), classnum) \
1336 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1338 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
1342 /* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1344 #define _generic_func_utf8(classnum, above_latin1, p) \
1345 _generic_utf8(classnum, p, above_latin1(p))
1346 /* Like the above, but passes classnum to _isFOO_utf8(), instead of having a
1347 * 'above_latin1' parameter */
1348 #define _generic_swash_utf8(classnum, p) \
1349 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
1351 /* Like the above, but should be used only when it is known that there are no
1352 * characters in the range 128-255 which the class is TRUE for. Hence it can
1353 * skip the tests for this range. 'above_latin1' should include its arguments */
1354 #define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
1355 (UTF8_IS_INVARIANT(*(p)) \
1356 ? _generic_isCC(*(p), classnum) \
1357 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
1361 /* NOTE that some of these macros have very similar ones in regcharclass.h.
1362 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1363 * there, differing in name only by an underscore from the one here
1364 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1365 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1366 * points; the regcharclass.h ones are implemented as a series of
1367 * "if-else-if-else ..." */
1369 #define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1370 #define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
1371 #define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1372 utf8, the non-utf8 macro works
1374 #define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1375 #define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1376 #define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1377 _is_utf8_FOO(_CC_DIGIT, p))
1378 #define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
1379 #define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
1380 _is_utf8_perl_idcont, p)
1382 /* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1383 * IDFIRST is an alnum. See
1384 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
1385 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1386 * This used to be not the XID version, but we decided to go with the more
1387 * modern Unicode definition */
1388 #define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
1389 _is_utf8_perl_idstart, p)
1391 #define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1392 #define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
1394 /* Posix and regular space are identical above Latin1 */
1395 #define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
1397 #define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
1398 #define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
1399 #define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
1400 #define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
1401 #define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
1402 #define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1405 #define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
1406 #define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1407 #define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1408 #define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
1410 /* For internal core Perl use only. If the input is in the Latin1 range, use
1411 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1412 * use the value given by the 'utf8' parameter. This relies on the fact that
1413 * ASCII characters have the same representation whether utf8 or not. Note
1414 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
1415 #define _generic_LC_utf8(macro, p, utf8) \
1416 (UTF8_IS_INVARIANT(*(p)) \
1418 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1419 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
1422 #define _generic_LC_swash_utf8(macro, classnum, p) \
1423 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1424 #define _generic_LC_func_utf8(macro, above_latin1, p) \
1425 _generic_LC_utf8(macro, p, above_latin1(p))
1427 #define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1428 _CC_ALPHANUMERIC, p)
1429 #define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
1430 #define isASCII_LC_utf8(p) isASCII_LC(*p)
1431 #define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1432 #define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1433 #define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1434 #define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
1435 #define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
1436 #define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1437 #define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1438 #define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
1439 #define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1440 space under locale */
1441 #define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1442 #define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1443 #define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1444 #define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1446 #define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
1448 /* Macros for backwards compatibility and for completeness when the ASCII and
1449 * Latin1 values are identical */
1450 #define isALPHAU(c) isALPHA_L1(c)
1451 #define isDIGIT_L1(c) isDIGIT_A(c)
1452 #define isOCTAL(c) isOCTAL_A(c)
1453 #define isOCTAL_L1(c) isOCTAL_A(c)
1454 #define isXDIGIT_L1(c) isXDIGIT_A(c)
1455 #define isALNUM(c) isWORDCHAR(c)
1456 #define isALNUMU(c) isWORDCHAR_L1(c)
1457 #define isALNUM_LC(c) isWORDCHAR_LC(c)
1458 #define isALNUM_uni(c) isWORDCHAR_uni(c)
1459 #define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1460 #define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1461 #define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
1462 #define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1463 #define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1464 #define isALNUMC(c) isALPHANUMERIC(c)
1465 #define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1466 #define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1467 #define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1468 #define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1469 #define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
1471 /* This conversion works both ways, strangely enough. On EBCDIC platforms,
1472 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1473 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1474 * that is EOT on ASCII; ST on EBCDIC */
1475 # define toCTRL(c) (toUPPER(NATIVE_TO_LATIN1(c)) ^ 64)
1477 /* Line numbers are unsigned, 32 bits. */
1479 #define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
1481 /* Helpful alias for version prescan */
1482 #define is_LAX_VERSION(a,b) \
1483 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1485 #define is_STRICT_VERSION(a,b) \
1486 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1488 #define BADVERSION(a,b,c) \
1494 #define READ_XDIGIT(s) (isALPHA(*(s)) ? ((*(s)++ + 9) & 0xf) : (*(s)++ & 0xf))
1497 =head1 Memory Management
1499 =for apidoc Am|void|Newx|void* ptr|int nitems|type
1500 The XSUB-writer's interface to the C C<malloc> function.
1502 In 5.9.3, Newx() and friends replace the older New() API, and drops
1503 the first parameter, I<x>, a debug aid which allowed callers to identify
1504 themselves. This aid has been superseded by a new build option,
1505 PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
1506 there for use in XS modules supporting older perls.
1508 =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
1509 The XSUB-writer's interface to the C C<malloc> function, with
1510 cast. See also C<Newx>.
1512 =for apidoc Am|void|Newxz|void* ptr|int nitems|type
1513 The XSUB-writer's interface to the C C<malloc> function. The allocated
1514 memory is zeroed with C<memzero>. See also C<Newx>.
1516 =for apidoc Am|void|Renew|void* ptr|int nitems|type
1517 The XSUB-writer's interface to the C C<realloc> function.
1519 =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1520 The XSUB-writer's interface to the C C<realloc> function, with
1523 =for apidoc Am|void|Safefree|void* ptr
1524 The XSUB-writer's interface to the C C<free> function.
1526 =for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1527 The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
1528 source, C<dest> is the destination, C<nitems> is the number of items, and
1529 C<type> is the type. Can do overlapping moves. See also C<Copy>.
1531 =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1532 Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1535 =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1536 The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
1537 source, C<dest> is the destination, C<nitems> is the number of items, and
1538 C<type> is the type. May fail on overlapping copies. See also C<Move>.
1540 =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1542 Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1545 =for apidoc Am|void|Zero|void* dest|int nitems|type
1547 The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1548 destination, C<nitems> is the number of items, and C<type> is the type.
1550 =for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1552 Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1555 =for apidoc Am|void|StructCopy|type *src|type *dest|type
1556 This is an architecture-independent macro to copy one structure to another.
1558 =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1560 Fill up memory with a byte pattern (a byte repeated over and over
1561 again) that hopefully catches attempts to access uninitialized memory.
1563 =for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1565 PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1567 =for apidoc Am|void|PoisonFree|void* dest|int nitems|type
1569 PoisonWith(0xEF) for catching access to freed memory.
1571 =for apidoc Am|void|Poison|void* dest|int nitems|type
1573 PoisonWith(0xEF) for catching access to freed memory.
1577 /* Maintained for backwards-compatibility only. Use newSV() instead. */
1579 #define NEWSV(x,len) newSV(len)
1582 #define MEM_SIZE_MAX ((MEM_SIZE)~0)
1584 /* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1585 * overly eager compilers that will bleat about e.g.
1586 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
1587 #ifdef PERL_MALLOC_WRAP
1588 #define MEM_WRAP_CHECK(n,t) \
1589 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
1590 #define MEM_WRAP_CHECK_1(n,t,a) \
1591 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
1592 #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
1594 #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)))
1597 #define MEM_WRAP_CHECK(n,t)
1598 #define MEM_WRAP_CHECK_1(n,t,a)
1599 #define MEM_WRAP_CHECK_2(n,t,a,b)
1600 #define MEM_WRAP_CHECK_(n,t)
1602 #define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1608 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
1609 * go through functions, which are handy for debugging breakpoints, but
1610 * which more importantly get the immediate calling environment (file and
1611 * line number, and C function name if available) passed in. This info can
1612 * then be used for logging the calls, for which one gets a sample
1613 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
1616 * - not all memory allocs get logged, only those
1617 * that go through Newx() and derivatives (while all
1618 * Safefrees do get logged)
1619 * - __FILE__ and __LINE__ do not work everywhere
1620 * - __func__ or __FUNCTION__ even less so
1621 * - I think more goes on after the perlio frees but
1622 * the thing is that STDERR gets closed (as do all
1623 * the file descriptors)
1624 * - no deeper calling stack than the caller of the Newx()
1625 * or the kind, but do I look like a C reflection/introspection
1627 * - the function prototypes for the logging functions
1628 * probably should maybe be somewhere else than handy.h
1629 * - one could consider inlining (macrofying) the logging
1630 * for speed, but I am too lazy
1631 * - one could imagine recording the allocations in a hash,
1632 * (keyed by the allocation address?), and maintain that
1633 * through reallocs and frees, but how to do that without
1634 * any News() happening...?
1635 * - lots of -Ddefines to get useful/controllable output
1636 * - lots of ENV reads
1639 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);
1641 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);
1643 PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
1646 # ifndef PERL_MEM_LOG_NOIMPL
1655 # if defined(PERL_IN_SV_C) /* those are only used in sv.c */
1656 void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1657 void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1664 #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1665 #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
1666 #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
1669 #ifndef MEM_LOG_ALLOC
1670 #define MEM_LOG_ALLOC(n,t,a) (a)
1672 #ifndef MEM_LOG_REALLOC
1673 #define MEM_LOG_REALLOC(n,t,v,a) (a)
1675 #ifndef MEM_LOG_FREE
1676 #define MEM_LOG_FREE(a) (a)
1679 #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1680 #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1681 #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
1684 /* pre 5.9.x compatibility */
1685 #define New(x,v,n,t) Newx(v,n,t)
1686 #define Newc(x,v,n,t,c) Newxc(v,n,t,c)
1687 #define Newz(x,v,n,t) Newxz(v,n,t)
1690 #define Renew(v,n,t) \
1691 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1692 #define Renewc(v,n,t,c) \
1693 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1696 #define Safefree(d) \
1697 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
1699 #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
1702 #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1703 #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1704 #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
1706 #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1707 #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1709 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
1711 /* Using bzero(), which returns void. */
1712 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
1715 #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1716 #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1717 #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1718 #define Poison(d,n,t) PoisonFree(d,n,t)
1720 #ifdef USE_STRUCT_COPY
1721 #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
1723 #define StructCopy(s,d,t) Copy(s,d,1,t)
1726 #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
1727 #define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
1731 # define Perl_va_copy(s, d) va_copy(d, s)
1733 # if defined(__va_copy)
1734 # define Perl_va_copy(s, d) __va_copy(d, s)
1736 # define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1741 /* convenience debug macros */
1743 #define pTHX_FORMAT "Perl interpreter: 0x%p"
1744 #define pTHX__FORMAT ", Perl interpreter: 0x%p"
1745 #define pTHX_VALUE_ (void *)my_perl,
1746 #define pTHX_VALUE (void *)my_perl
1747 #define pTHX__VALUE_ ,(void *)my_perl,
1748 #define pTHX__VALUE ,(void *)my_perl
1751 #define pTHX__FORMAT
1754 #define pTHX__VALUE_
1756 #endif /* USE_ITHREADS */
1758 /* Perl_deprecate was not part of the public API, and did not have a deprecate()
1759 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1760 CPAN::Unpack show any users outside the core. */
1762 # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
1765 /* Internal macros to deal with gids and uids */
1768 # if Uid_t_size > IVSIZE
1769 # define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
1770 # define SvUID(sv) SvNV(sv)
1772 # if Uid_t_sign <= 0
1773 # define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
1774 # define SvUID(sv) SvIV(sv)
1776 # define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
1777 # define SvUID(sv) SvUV(sv)
1779 # endif /* Uid_t_size */
1781 # if Gid_t_size > IVSIZE
1782 # define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
1783 # define SvGID(sv) SvNV(sv)
1785 # if Gid_t_sign <= 0
1786 # define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
1787 # define SvGID(sv) SvIV(sv)
1789 # define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
1790 # define SvGID(sv) SvUV(sv)
1792 # endif /* Gid_t_size */
1796 #endif /* HANDY_H */
1800 * c-indentation-style: bsd
1802 * indent-tabs-mode: nil
1805 * ex: set ts=8 sts=4 sw=4 et:
https://perl5.git.perl.org / perl5.git / blob