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
107 /* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
108 * defined as char for example, then the cast from int is
109 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
111 #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
113 /* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
114 * XXX Should really be a Configure probe, with HAS__FUNCTION__
115 * and FUNCTION__ as results.
116 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
117 #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
118 # define FUNCTION__ __func__
120 # if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
121 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
122 # define FUNCTION__ ""
124 # define FUNCTION__ __FUNCTION__ /* Common extension. */
128 /* XXX A note on the perl source internal type system. The
129 original intent was that I32 be *exactly* 32 bits.
131 Currently, we only guarantee that I32 is *at least* 32 bits.
132 Specifically, if int is 64 bits, then so is I32. (This is the case
133 for the Cray.) This has the advantage of meshing nicely with
134 standard library calls (where we pass an I32 and the library is
135 expecting an int), but the disadvantage that an I32 is not 32 bits.
136 Andy Dougherty August 1996
138 There is no guarantee that there is *any* integral type with
139 exactly 32 bits. It is perfectly legal for a system to have
140 sizeof(short) == sizeof(int) == sizeof(long) == 8.
142 Similarly, there is no guarantee that I16 and U16 have exactly 16
145 For dealing with issues that may arise from various 32/64-bit
146 systems, we will ask Configure to check out
148 SHORTSIZE == sizeof(short)
149 INTSIZE == sizeof(int)
150 LONGSIZE == sizeof(long)
151 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
152 PTRSIZE == sizeof(void *)
153 DOUBLESIZE == sizeof(double)
154 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
158 #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
159 # include <inttypes.h>
160 # ifdef INT32_MIN_BROKEN
162 # define INT32_MIN (-2147483647-1)
164 # ifdef INT64_MIN_BROKEN
166 # define INT64_MIN (-9223372036854775807LL-1)
182 /* INT64_C/UINT64_C are C99 from <stdint.h> (so they will not be
183 * available in strict C89 mode), but they are nice, so let's define
184 * them if necessary. */
185 #if defined(HAS_QUAD)
187 # undef PeRl_UINT64_C
188 /* Prefer the native integer types (int and long) over long long
189 * (which is not C89) and Win32-specific __int64. */
190 # if QUADKIND == QUAD_IS_INT && INTSIZE == 8
191 # define PeRl_INT64_C(c) (c)
192 # define PeRl_UINT64_C(c) CAT2(c,U)
194 # if QUADKIND == QUAD_IS_LONG && LONGSIZE == 8
195 # define PeRl_INT64_C(c) CAT2(c,L)
196 # define PeRl_UINT64_C(c) CAT2(c,UL)
198 # if QUADKIND == QUAD_IS_LONG_LONG && defined(HAS_LONG_LONG)
199 # define PeRl_INT64_C(c) CAT2(c,LL)
200 # define PeRl_UINT64_C(c) CAT2(c,ULL)
202 # if QUADKIND == QUAD_IS___INT64
203 # define PeRl_INT64_C(c) CAT2(c,I64)
204 # define PeRl_UINT64_C(c) CAT2(c,UI64)
206 # ifndef PeRl_INT64_C
207 # define PeRl_INT64_C(c) ((I64)(c)) /* last resort */
208 # define PeRl_UINT64_C(c) ((U64)(c))
210 /* In OS X the INT64_C/UINT64_C are defined with LL/ULL, which will
211 * not fly with C89-pedantic gcc, so let's undefine them first so that
212 * we can redefine them with our native integer preferring versions. */
213 # if defined(PERL_DARWIN) && defined(PERL_GCC_PEDANTIC)
218 # define INT64_C(c) PeRl_INT64_C(c)
221 # define UINT64_C(c) PeRl_UINT64_C(c)
225 #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
227 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
228 Please search CHAR_MAX in perl.h for further details. */
229 #define U8_MAX UINT8_MAX
230 #define U8_MIN UINT8_MIN
232 #define I16_MAX INT16_MAX
233 #define I16_MIN INT16_MIN
234 #define U16_MAX UINT16_MAX
235 #define U16_MIN UINT16_MIN
237 #define I32_MAX INT32_MAX
238 #define I32_MIN INT32_MIN
239 #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
240 # define U32_MAX UINT32_MAX
242 # define U32_MAX 4294967295U
244 #define U32_MIN UINT32_MIN
248 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
249 Please search CHAR_MAX in perl.h for further details. */
250 #define U8_MAX PERL_UCHAR_MAX
251 #define U8_MIN PERL_UCHAR_MIN
253 #define I16_MAX PERL_SHORT_MAX
254 #define I16_MIN PERL_SHORT_MIN
255 #define U16_MAX PERL_USHORT_MAX
256 #define U16_MIN PERL_USHORT_MIN
259 # define I32_MAX PERL_INT_MAX
260 # define I32_MIN PERL_INT_MIN
261 # define U32_MAX PERL_UINT_MAX
262 # define U32_MIN PERL_UINT_MIN
264 # define I32_MAX PERL_LONG_MAX
265 # define I32_MIN PERL_LONG_MIN
266 # define U32_MAX PERL_ULONG_MAX
267 # define U32_MIN PERL_ULONG_MIN
272 /* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
273 #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
274 #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
275 #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
277 /* Unused by core; should be deprecated */
278 #define Ctl(ch) ((ch) & 037)
280 /* This is a helper macro to avoid preprocessor issues, replaced by nothing
281 * unless under DEBUGGING, where it expands to an assert of its argument,
282 * followed by a comma (hence the comma operator). If we just used a straight
283 * assert(), we would get a comma with nothing before it when not DEBUGGING.
285 * We also use empty definition under Coverity since the __ASSERT__
286 * checks often check for things that Really Cannot Happen, and Coverity
287 * detects that and gets all excited. */
289 #if defined(DEBUGGING) && !defined(__COVERITY__)
290 # define __ASSERT_(statement) assert(statement),
292 # define __ASSERT_(statement)
296 =head1 SV-Body Allocation
298 =for apidoc Ama|SV*|newSVpvs|const char* s
299 Like C<newSVpvn>, but takes a C<NUL>-terminated literal string instead of a
302 =for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
303 Like C<newSVpvn_flags>, but takes a C<NUL>-terminated literal string instead of
304 a string/length pair.
306 =for apidoc Ama|SV*|newSVpvs_share|const char* s
307 Like C<newSVpvn_share>, but takes a C<NUL>-terminated literal string instead of
308 a string/length pair and omits the hash parameter.
310 =for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
311 Like C<sv_catpvn_flags>, but takes a C<NUL>-terminated literal string instead
312 of a string/length pair.
314 =for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
315 Like C<sv_catpvn_nomg>, but takes a C<NUL>-terminated literal string instead of
316 a string/length pair.
318 =for apidoc Am|void|sv_catpvs|SV* sv|const char* s
319 Like C<sv_catpvn>, but takes a C<NUL>-terminated literal string instead of a
322 =for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
323 Like C<sv_catpvn_mg>, but takes a C<NUL>-terminated literal string instead of a
326 =for apidoc Am|void|sv_setpvs|SV* sv|const char* s
327 Like C<sv_setpvn>, but takes a C<NUL>-terminated literal string instead of a
330 =for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
331 Like C<sv_setpvn_mg>, but takes a C<NUL>-terminated literal string instead of a
334 =for apidoc Am|SV *|sv_setref_pvs|const char* s
335 Like C<sv_setref_pvn>, but takes a C<NUL>-terminated literal string instead of
336 a string/length pair.
338 =head1 Memory Management
340 =for apidoc Ama|char*|savepvs|const char* s
341 Like C<savepvn>, but takes a C<NUL>-terminated literal string instead of a
344 =for apidoc Ama|char*|savesharedpvs|const char* s
345 A version of C<savepvs()> which allocates the duplicate string in memory
346 which is shared between threads.
350 =for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
351 Like C<gv_stashpvn>, but takes a C<NUL>-terminated literal string instead of a
354 =head1 Hash Manipulation Functions
356 =for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
357 Like C<hv_fetch>, but takes a C<NUL>-terminated literal string instead of a
360 =for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
361 Like C<hv_store>, but takes a C<NUL>-terminated literal string instead of a
363 and omits the hash parameter.
365 =head1 Lexer interface
367 =for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
369 Like L</lex_stuff_pvn>, but takes a C<NUL>-terminated literal string instead of
370 a string/length pair.
375 /* concatenating with "" ensures that only literal strings are accepted as
377 #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
379 /* note that STR_WITH_LEN() can't be used as argument to macros or functions
380 * that under some configurations might be macros, which means that it requires
381 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
384 /* STR_WITH_LEN() shortcuts */
385 #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
386 #define newSVpvs_flags(str,flags) \
387 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
388 #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
389 #define sv_catpvs_flags(sv, str, flags) \
390 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
391 #define sv_catpvs_nomg(sv, str) \
392 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
393 #define sv_catpvs(sv, str) \
394 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
395 #define sv_catpvs_mg(sv, str) \
396 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
397 #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
398 #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
399 #define sv_setref_pvs(rv, classname, str) \
400 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
401 #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
402 #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
403 #define gv_stashpvs(str, create) \
404 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
405 #define gv_fetchpvs(namebeg, add, sv_type) \
406 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
407 #define gv_fetchpvn(namebeg, len, add, sv_type) \
408 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
409 #define sv_catxmlpvs(dsv, str, utf8) \
410 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
411 #define hv_fetchs(hv,key,lval) \
412 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
413 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
414 : HV_FETCH_JUST_SV, NULL, 0))
416 #define hv_stores(hv,key,val) \
417 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
418 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
420 #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
422 #define get_cvs(str, flags) \
423 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
426 =head1 Miscellaneous Functions
428 =for apidoc Am|bool|strNE|char* s1|char* s2
429 Test two C<NUL>-terminated strings to see if they are different. Returns true
432 =for apidoc Am|bool|strEQ|char* s1|char* s2
433 Test two C<NUL>-terminated strings to see if they are equal. Returns true or
436 =for apidoc Am|bool|strLT|char* s1|char* s2
437 Test two C<NUL>-terminated strings to see if the first, C<s1>, is less than the
438 second, C<s2>. Returns true or false.
440 =for apidoc Am|bool|strLE|char* s1|char* s2
441 Test two C<NUL>-terminated strings to see if the first, C<s1>, is less than or
442 equal to the second, C<s2>. Returns true or false.
444 =for apidoc Am|bool|strGT|char* s1|char* s2
445 Test two C<NUL>-terminated strings to see if the first, C<s1>, is greater than
446 the second, C<s2>. Returns true or false.
448 =for apidoc Am|bool|strGE|char* s1|char* s2
449 Test two C<NUL>-terminated strings to see if the first, C<s1>, is greater than
450 or equal to the second, C<s2>. Returns true or false.
452 =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
453 Test two C<NUL>-terminated strings to see if they are different. The C<len>
454 parameter 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 C<NUL>-terminated strings to see if they are equal. The C<len>
459 parameter indicates the number of bytes to compare. Returns true or false. (A
460 wrapper for C<strncmp>).
462 =for apidoc Am|bool|memEQ|char* s1|char* s2|STRLEN len
463 Test two buffers (which may contain embedded C<NUL> characters, to see if they
464 are equal. The C<len> parameter indicates the number of bytes to compare.
465 Returns zero if equal, or non-zero if non-equal.
467 =for apidoc Am|bool|memNE|char* s1|char* s2|STRLEN len
468 Test two buffers (which may contain embedded C<NUL> characters, to see if they
469 are not equal. The C<len> parameter indicates the number of bytes to compare.
470 Returns zero if non-equal, or non-zero if equal.
475 #define strNE(s1,s2) (strcmp(s1,s2))
476 #define strEQ(s1,s2) (!strcmp(s1,s2))
477 #define strLT(s1,s2) (strcmp(s1,s2) < 0)
478 #define strLE(s1,s2) (strcmp(s1,s2) <= 0)
479 #define strGT(s1,s2) (strcmp(s1,s2) > 0)
480 #define strGE(s1,s2) (strcmp(s1,s2) >= 0)
481 #define strnNE(s1,s2,l) (strncmp(s1,s2,l))
482 #define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
485 # define memNE(s1,s2,l) (memcmp(s1,s2,l))
486 # define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
488 # define memNE(s1,s2,l) (bcmp(s1,s2,l))
489 # define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
492 /* memEQ and memNE where second comparand is a string constant */
493 #define memEQs(s1, l, s2) \
494 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
495 #define memNEs(s1, l, s2) !memEQs(s1, l, s2)
497 #define memLT(s1,s2,l) (memcmp(s1,s2,l) < 0)
498 #define memLE(s1,s2,l) (memcmp(s1,s2,l) <= 0)
499 #define memGT(s1,s2,l) (memcmp(s1,s2,l) > 0)
500 #define memGE(s1,s2,l) (memcmp(s1,s2,l) >= 0)
505 * Unfortunately, the introduction of locales means that we
506 * can't trust isupper(), etc. to tell the truth. And when
507 * it comes to /\w+/ with tainting enabled, we *must* be able
508 * to trust our character classes.
510 * Therefore, the default tests in the text of Perl will be
511 * independent of locale. Any code that wants to depend on
512 * the current locale will use the tests that begin with "lc".
515 #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
523 =head1 Character classification
524 This section is about functions (really macros) that classify characters
525 into types, such as punctuation versus alphabetic, etc. Most of these are
526 analogous to regular expression character classes. (See
527 L<perlrecharclass/POSIX Character Classes>.) There are several variants for
528 each class. (Not all macros have all variants; each item below lists the
529 ones valid for it.) None are affected by C<use bytes>, and only the ones
530 with C<LC> in the name are affected by the current locale.
532 The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
533 C<U8>) as input and returns a boolean as to whether or not the character
534 represented by that octet is (or on non-ASCII platforms, corresponds to) an
535 ASCII character in the named class based on platform, Unicode, and Perl rules.
536 If the input is a number that doesn't fit in an octet, FALSE is returned.
538 Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
539 with no suffix C<"_A">.
541 Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
542 onto the platform. That is, the code points that are ASCII are unaffected,
543 since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
544 as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
545 true when called with the code point 0xDF, which is a word character in both
546 ASCII and EBCDIC (though it represents different characters in each).
548 Variant C<isFOO_uvchr> is like the C<isFOO_L1> variant, but accepts any UV code
549 point as input. If the code point is larger than 255, Unicode rules are used
550 to determine if it is in the character class. For example,
551 C<isWORDCHAR_uvchr(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
552 WITH MACRON in Unicode, and is a word character.
554 Variant C<isFOO_utf8> is like C<isFOO_uvchr>, but the input is a pointer to a
555 (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>, and
556 possibly containing embedded C<NUL> characters). The classification of just
557 the first (possibly multi-byte) character in the string is tested.
559 Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but the
560 result is based on the current locale, which is what C<LC> in the name stands
561 for. If Perl can determine that the current locale is a UTF-8 locale, it uses
562 the published Unicode rules; otherwise, it uses the C library function that
563 gives the named classification. For example, C<isDIGIT_LC()> when not in a
564 UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always
565 returned if the input won't fit into an octet. On some platforms where the C
566 library function is known to be defective, Perl changes its result to follow
567 the POSIX standard's rules.
569 Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
570 returns the same as C<isFOO_LC> for input code points less than 256, and
571 returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
573 Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer
574 to a (known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>, and
575 possibly containing embedded C<NUL> characters). The classification of just
576 the first (possibly multi-byte) character in the string is tested.
578 =for apidoc Am|bool|isALPHA|char ch
579 Returns a boolean indicating whether the specified character is an
580 alphabetic character, analogous to C<m/[[:alpha:]]/>.
581 See the L<top of this section|/Character classification> for an explanation of
583 C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uvchr>, C<isALPHA_utf8>, C<isALPHA_LC>,
584 C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
586 =for apidoc Am|bool|isALPHANUMERIC|char ch
587 Returns a boolean indicating whether the specified character is a either an
588 alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
589 See the L<top of this section|/Character classification> for an explanation of
591 C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uvchr>,
592 C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
593 C<isALPHANUMERIC_LC_utf8>.
595 =for apidoc Am|bool|isASCII|char ch
596 Returns a boolean indicating whether the specified character is one of the 128
597 characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
598 On non-ASCII platforms, it returns TRUE iff this
599 character corresponds to an ASCII character. Variants C<isASCII_A()> and
600 C<isASCII_L1()> are identical to C<isASCII()>.
601 See the L<top of this section|/Character classification> for an explanation of
603 C<isASCII_uvchr>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
604 C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
605 library routine C<isascii()>. In these cases, the variants whose names contain
606 C<LC> are the same as the corresponding ones without.
608 Also note, that because all ASCII characters are UTF-8 invariant (meaning they
609 have the exact same representation (always a single byte) whether encoded in
610 UTF-8 or not), C<isASCII> will give the correct results when called with any
611 byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
612 work properly on any string encoded or not in UTF-8.
614 =for apidoc Am|bool|isBLANK|char ch
615 Returns a boolean indicating whether the specified character is a
616 character considered to be a blank, analogous to C<m/[[:blank:]]/>.
617 See the L<top of this section|/Character classification> for an explanation of
619 C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uvchr>, C<isBLANK_utf8>, C<isBLANK_LC>,
620 C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
621 platforms do not have the C library routine C<isblank()>. In these cases, the
622 variants whose names contain C<LC> are the same as the corresponding ones
625 =for apidoc Am|bool|isCNTRL|char ch
626 Returns a boolean indicating whether the specified character is a
627 control character, analogous to C<m/[[:cntrl:]]/>.
628 See the L<top of this section|/Character classification> for an explanation of
630 C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uvchr>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
631 C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
632 On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
634 =for apidoc Am|bool|isDIGIT|char ch
635 Returns a boolean indicating whether the specified character is a
636 digit, analogous to C<m/[[:digit:]]/>.
637 Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
638 See the L<top of this section|/Character classification> for an explanation of
640 C<isDIGIT_uvchr>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
643 =for apidoc Am|bool|isGRAPH|char ch
644 Returns a boolean indicating whether the specified character is a
645 graphic character, analogous to C<m/[[:graph:]]/>.
646 See the L<top of this section|/Character classification> for an explanation of
648 C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uvchr>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
649 C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
651 =for apidoc Am|bool|isLOWER|char ch
652 Returns a boolean indicating whether the specified character is a
653 lowercase character, analogous to C<m/[[:lower:]]/>.
654 See the L<top of this section|/Character classification> for an explanation of
656 C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uvchr>, C<isLOWER_utf8>, C<isLOWER_LC>,
657 C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
659 =for apidoc Am|bool|isOCTAL|char ch
660 Returns a boolean indicating whether the specified character is an
662 The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
665 =for apidoc Am|bool|isPUNCT|char ch
666 Returns a boolean indicating whether the specified character is a
667 punctuation character, analogous to C<m/[[:punct:]]/>.
668 Note that the definition of what is punctuation isn't as
669 straightforward as one might desire. See L<perlrecharclass/POSIX Character
670 Classes> for details.
671 See the L<top of this section|/Character classification> for an explanation of
673 C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uvchr>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
674 C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
676 =for apidoc Am|bool|isSPACE|char ch
677 Returns a boolean indicating whether the specified character is a
678 whitespace character. This is analogous
679 to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
680 this also matches what C<m/[[:space:]]/> does. Prior to 5.18, only the
681 locale forms of this macro (the ones with C<LC> in their names) matched
682 precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
683 in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
684 (See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
685 See the L<top of this section|/Character classification> for an explanation of
687 C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uvchr>, C<isSPACE_utf8>, C<isSPACE_LC>,
688 C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
690 =for apidoc Am|bool|isPSXSPC|char ch
691 (short for Posix Space)
692 Starting in 5.18, this is identical in all its forms to the
693 corresponding C<isSPACE()> macros.
694 The locale forms of this macro are identical to their corresponding
695 C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
696 non-locale forms differ from their C<isSPACE()> forms only in that the
697 C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
698 Otherwise they are identical. Thus this macro is analogous to what
699 C<m/[[:space:]]/> matches in a regular expression.
700 See the L<top of this section|/Character classification> for an explanation of
701 variants C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uvchr>, C<isPSXSPC_utf8>,
702 C<isPSXSPC_LC>, C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
704 =for apidoc Am|bool|isUPPER|char ch
705 Returns a boolean indicating whether the specified character is an
706 uppercase character, analogous to C<m/[[:upper:]]/>.
707 See the L<top of this section|/Character classification> for an explanation of
709 C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uvchr>, C<isUPPER_utf8>, C<isUPPER_LC>,
710 C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
712 =for apidoc Am|bool|isPRINT|char ch
713 Returns a boolean indicating whether the specified character is a
714 printable character, analogous to C<m/[[:print:]]/>.
715 See the L<top of this section|/Character classification> for an explanation of
717 C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uvchr>, C<isPRINT_utf8>, C<isPRINT_LC>,
718 C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
720 =for apidoc Am|bool|isWORDCHAR|char ch
721 Returns a boolean indicating whether the specified character is a character
722 that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
723 in a regular expression. A word character is an alphabetic character, a
724 decimal digit, a connecting punctuation character (such as an underscore), or
725 a "mark" character that attaches to one of those (like some sort of accent).
726 C<isALNUM()> is a synonym provided for backward compatibility, even though a
727 word character includes more than the standard C language meaning of
729 See the L<top of this section|/Character classification> for an explanation of
731 C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uvchr>, and C<isWORDCHAR_utf8>.
732 C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8> are also as
733 described there, but additionally include the platform's native underscore.
735 =for apidoc Am|bool|isXDIGIT|char ch
736 Returns a boolean indicating whether the specified character is a hexadecimal
737 digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
738 and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
739 See the L<top of this section|/Character classification> for an explanation of
741 C<isXDIGIT_uvchr>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
744 =for apidoc Am|bool|isIDFIRST|char ch
745 Returns a boolean indicating whether the specified character can be the first
746 character of an identifier. This is very close to, but not quite the same as
747 the official Unicode property C<XID_Start>. The difference is that this
748 returns true only if the input character also matches L</isWORDCHAR>.
749 See the L<top of this section|/Character classification> for an explanation of
751 C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uvchr>, C<isIDFIRST_utf8>,
752 C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
754 =for apidoc Am|bool|isIDCONT|char ch
755 Returns a boolean indicating whether the specified character can be the
756 second or succeeding character of an identifier. This is very close to, but
757 not quite the same as the official Unicode property C<XID_Continue>. The
758 difference is that this returns true only if the input character also matches
759 L</isWORDCHAR>. See the L<top of this section|/Character classification> for
761 explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uvchr>,
762 C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
765 =head1 Miscellaneous Functions
767 =for apidoc Am|U8|READ_XDIGIT|char str*
768 Returns the value of an ASCII-range hex digit and advances the string pointer.
769 Behaviour is only well defined when isXDIGIT(*str) is true.
771 =head1 Character case changing
773 =for apidoc Am|U8|toUPPER|U8 ch
774 Converts the specified character to uppercase. If the input is anything but an
775 ASCII lowercase character, that input character itself is returned. Variant
776 C<toUPPER_A> is equivalent.
778 =for apidoc Am|UV|toUPPER_uvchr|UV cp|U8* s|STRLEN* lenp
779 Converts the code point C<cp> to its uppercase version, and
780 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code
781 point is interpreted as native if less than 256; otherwise as Unicode. Note
782 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
783 bytes since the uppercase version may be longer than the original character.
785 The first code point of the uppercased version is returned
786 (but note, as explained just above, that there may be more.)
788 =for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
789 Converts the UTF-8 encoded character at C<p> to its uppercase version, and
790 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
791 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
792 bytes since the uppercase version may be longer than the original character.
794 The first code point of the uppercased version is returned
795 (but note, as explained just above, that there may be more.)
797 The input character at C<p> is assumed to be well-formed.
799 =for apidoc Am|U8|toFOLD|U8 ch
800 Converts the specified character to foldcase. If the input is anything but an
801 ASCII uppercase character, that input character itself is returned. Variant
802 C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
803 Latin1 range, as the full generality of L</toFOLD_uvchr> is needed there.)
805 =for apidoc Am|UV|toFOLD_uvchr|UV cp|U8* s|STRLEN* lenp
806 Converts the code point C<cp> to its foldcase version, and
807 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code
808 point is interpreted as native if less than 256; otherwise as Unicode. Note
809 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
810 bytes since the foldcase version may be longer than the original character.
812 The first code point of the foldcased version is returned
813 (but note, as explained just above, that there may be more.)
815 =for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
816 Converts the UTF-8 encoded character at C<p> to its foldcase 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 foldcase version may be longer than the original character.
821 The first code point of the foldcased 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|U8 ch
827 Converts the specified character to lowercase. If the input is anything but an
828 ASCII uppercase character, that input character itself is returned. Variant
829 C<toLOWER_A> is equivalent.
831 =for apidoc Am|U8|toLOWER_L1|U8 ch
832 Converts the specified Latin1 character to lowercase. The results are
833 undefined if the input doesn't fit in a byte.
835 =for apidoc Am|U8|toLOWER_LC|U8 ch
836 Converts the specified character to lowercase using the current locale's rules,
837 if possible; otherwise returns the input character itself.
839 =for apidoc Am|UV|toLOWER_uvchr|UV cp|U8* s|STRLEN* lenp
840 Converts the code point C<cp> to its lowercase version, and
841 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code
842 point is interpreted as native if less than 256; otherwise as Unicode. Note
843 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
844 bytes since the lowercase version may be longer than the original character.
846 The first code point of the lowercased version is returned
847 (but note, as explained just above, that there may be more.)
849 =for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
850 Converts the UTF-8 encoded character at C<p> to its lowercase version, and
851 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
852 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
853 bytes since the lowercase version may be longer than the original character.
855 The first code point of the lowercased version is returned
856 (but note, as explained just above, that there may be more.)
858 The input character at C<p> is assumed to be well-formed.
860 =for apidoc Am|U8|toTITLE|U8 ch
861 Converts the specified character to titlecase. If the input is anything but an
862 ASCII lowercase character, that input character itself is returned. Variant
863 C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1
864 range, as the full generality of L</toTITLE_uvchr> is needed there. Titlecase is
865 not a concept used in locale handling, so there is no functionality for that.)
867 =for apidoc Am|UV|toTITLE_uvchr|UV cp|U8* s|STRLEN* lenp
868 Converts the code point C<cp> to its titlecase version, and
869 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code
870 point is interpreted as native if less than 256; otherwise as Unicode. Note
871 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
872 bytes since the titlecase version may be longer than the original character.
874 The first code point of the titlecased version is returned
875 (but note, as explained just above, that there may be more.)
877 =for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
878 Converts the UTF-8 encoded character at C<p> to its titlecase version, and
879 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
880 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
881 bytes since the titlecase version may be longer than the original character.
883 The first code point of the titlecased version is returned
884 (but note, as explained just above, that there may be more.)
886 The input character at C<p> is assumed to be well-formed.
890 XXX Still undocumented isVERTWS_uvchr and _utf8; it's unclear what their names
891 really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
893 Note that these macros are repeated in Devel::PPPort, so should also be
894 patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
898 /* Specify the widest unsigned type on the platform. Use U64TYPE because U64
899 * is known only in the perl core, and this macro can be called from outside
902 # define WIDEST_UTYPE U64TYPE
904 # define WIDEST_UTYPE U32
907 /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
908 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
909 * bits of information are lost even on a 64-bit machine, but to get the
910 * compiler to optimize it out if possible. This is because Configure makes
911 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
912 * sizeof() guarantees that this evaluates to a constant true at compile time.
914 * For Coverity, be always true, because otherwise Coverity thinks
915 * it finds several expressions that are always true, independent
916 * of operands. Well, they are, but that is kind of the point.
919 /* The '| 0' part ensures a compiler error if c is not integer (like e.g., a
921 #define FITS_IN_8_BITS(c) ( (sizeof(c) == 1) \
922 || !(((WIDEST_UTYPE)((c) | 0)) & ~0xFF))
924 #define FITS_IN_8_BITS(c) (1)
929 /* The native libc isascii() et.al. functions return the wrong results
930 * on at least z/OS unless this is defined. */
931 # error _ALL_SOURCE should probably be defined
934 /* There is a simple definition of ASCII for ASCII platforms. But the
935 * EBCDIC one isn't so simple, so is defined using table look-up like the
936 * other macros below.
938 * The cast here is used instead of '(c) >= 0', because some compilers emit
939 * a warning that that test is always true when the parameter is an
940 * unsigned type. khw supposes that it could be written as
941 * && ((c) == '\0' || (c) > 0)
942 * to avoid the message, but the cast will likely avoid extra branches even
943 * with stupid compilers.
945 * The '| 0' part ensures a compiler error if c is not integer (like e.g.,
947 # define isASCII(c) ((WIDEST_UTYPE)((c) | 0) < 128)
950 /* Take the eight possible bit patterns of the lower 3 bits and you get the
951 * lower 3 bits of the 8 octal digits, in both ASCII and EBCDIC, so those bits
952 * can be ignored. If the rest match '0', we have an octal */
953 #define isOCTAL_A(c) (((WIDEST_UTYPE)((c) | 0) & ~7) == '0')
955 #ifdef H_PERL /* If have access to perl.h, lookup in its table */
957 /* Character class numbers. For internal core Perl use only. The ones less
958 * than 32 are used in PL_charclass[] and the ones up through the one that
959 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
960 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
961 * their actual definitions are here. If that file has a name not used here,
964 * The first group of these is ordered in what I (khw) estimate to be the
965 * frequency of their use. This gives a slight edge to exiting a loop earlier
966 * (in reginclass() in regexec.c) */
967 # define _CC_WORDCHAR 0 /* \w and [:word:] */
968 # define _CC_DIGIT 1 /* \d and [:digit:] */
969 # define _CC_ALPHA 2 /* [:alpha:] */
970 # define _CC_LOWER 3 /* [:lower:] */
971 # define _CC_UPPER 4 /* [:upper:] */
972 # define _CC_PUNCT 5 /* [:punct:] */
973 # define _CC_PRINT 6 /* [:print:] */
974 # define _CC_ALPHANUMERIC 7 /* [:alnum:] */
975 # define _CC_GRAPH 8 /* [:graph:] */
976 # define _CC_CASED 9 /* [:lower:] or [:upper:] under /i */
978 #define _FIRST_NON_SWASH_CC 10
979 /* The character classes above are implemented with swashes. The second group
980 * (just below) contains the ones implemented without. These are also sorted
981 * in rough order of the frequency of their use, except that \v should be last,
982 * as it isn't a real Posix character class, and some (small) inefficiencies in
983 * regular expression handling would be introduced by putting it in the middle
984 * of those that are. Also, cntrl and ascii come after the others as it may be
985 * useful to group these which have no members that match above Latin1, (or
986 * above ASCII in the latter case) */
988 # define _CC_SPACE 10 /* \s, [:space:] */
989 # define _CC_BLANK 11 /* [:blank:] */
990 # define _CC_XDIGIT 12 /* [:xdigit:] */
991 # define _CC_CNTRL 13 /* [:cntrl:] */
992 # define _CC_ASCII 14 /* [:ascii:] */
993 # define _CC_VERTSPACE 15 /* \v */
995 # define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
997 /* The members of the third group below do not need to be coordinated with data
998 * structures in regcomp.[ch] and regexec.c. */
999 # define _CC_IDFIRST 16
1000 # define _CC_CHARNAME_CONT 17
1001 # define _CC_NONLATIN1_FOLD 18
1002 # define _CC_NONLATIN1_SIMPLE_FOLD 19
1003 # define _CC_QUOTEMETA 20
1004 # define _CC_NON_FINAL_FOLD 21
1005 # define _CC_IS_IN_SOME_FOLD 22
1006 # define _CC_MNEMONIC_CNTRL 23
1008 /* This next group is only used on EBCDIC platforms, so theoretically could be
1009 * shared with something entirely different that's only on ASCII platforms */
1010 # define _CC_UTF8_START_BYTE_IS_FOR_AT_LEAST_SURROGATE 28
1011 # define _CC_UTF8_IS_START 29
1012 # define _CC_UTF8_IS_DOWNGRADEABLE_START 30
1013 # define _CC_UTF8_IS_CONTINUATION 31
1015 * If more bits are needed, one could add a second word for non-64bit
1016 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
1017 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
1018 * is used only for optimization (as of this writing), and differs in the
1019 * Latin1 range from the ALPHA bit only in two relatively unimportant
1020 * characters: the masculine and feminine ordinal indicators, so removing it
1021 * would just cause /i regexes which match them to run less efficiently.
1022 * Similarly the EBCDIC-only bits are used just for speed, and could be
1023 * replaced by other means */
1025 #if defined(PERL_CORE) || defined(PERL_EXT)
1026 /* An enum version of the character class numbers, to help compilers
1029 _CC_ENUM_ALPHA = _CC_ALPHA,
1030 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
1031 _CC_ENUM_ASCII = _CC_ASCII,
1032 _CC_ENUM_BLANK = _CC_BLANK,
1033 _CC_ENUM_CASED = _CC_CASED,
1034 _CC_ENUM_CNTRL = _CC_CNTRL,
1035 _CC_ENUM_DIGIT = _CC_DIGIT,
1036 _CC_ENUM_GRAPH = _CC_GRAPH,
1037 _CC_ENUM_LOWER = _CC_LOWER,
1038 _CC_ENUM_PRINT = _CC_PRINT,
1039 _CC_ENUM_PUNCT = _CC_PUNCT,
1040 _CC_ENUM_SPACE = _CC_SPACE,
1041 _CC_ENUM_UPPER = _CC_UPPER,
1042 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
1043 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
1044 _CC_ENUM_XDIGIT = _CC_XDIGIT
1045 } _char_class_number;
1048 #define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
1049 #define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
1051 #if defined(PERL_IN_UTF8_C) \
1052 || defined(PERL_IN_REGCOMP_C) \
1053 || defined(PERL_IN_REGEXEC_C)
1054 # if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
1055 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
1056 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
1057 #error Need to adjust order of swash_property_names[]
1060 /* This is declared static in each of the few files that this is #defined for
1061 * to keep them from being publicly accessible. Hence there is a small amount
1062 * of wasted space */
1064 static const char* const swash_property_names[] = {
1080 EXTCONST U32 PL_charclass[] = {
1081 # include "l1_char_class_tab.h"
1084 # else /* ! DOINIT */
1085 EXTCONST U32 PL_charclass[];
1089 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
1090 # define _CC_mask(classnum) (1U << (classnum))
1092 /* For internal core Perl use only: the base macro for defining macros like
1094 # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
1095 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
1097 /* The mask for the _A versions of the macros; it just adds in the bit for
1099 # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1101 /* For internal core Perl use only: the base macro for defining macros like
1102 * isALPHA_A. The foo_A version makes sure that both the desired bit and
1103 * the ASCII bit are present */
1104 # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
1105 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
1106 == _CC_mask_A(classnum)))
1108 # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
1109 # define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
1110 # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1111 # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1112 # define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT) /* No non-ASCII digits */
1113 # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1114 # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1115 # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1116 # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1117 # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1118 # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1119 # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1120 # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT) /* No non-ASCII xdigits
1122 # define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
1123 # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1124 # define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1125 # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1127 /* continuation character for legal NAME in \N{NAME} */
1128 # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1130 # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1131 # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1132 # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1133 # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1134 # define isPSXSPC_L1(c) isSPACE_L1(c)
1135 # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1136 # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1137 # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1138 # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1139 # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
1142 # define isASCII(c) _generic_isCC(c, _CC_ASCII)
1145 /* Participates in a single-character fold with a character above 255 */
1146 # define _HAS_NONLATIN1_SIMPLE_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_SIMPLE_FOLD)))
1148 /* Like the above, but also can be part of a multi-char fold */
1149 # 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)))
1151 # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
1152 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1153 _generic_isCC(c, _CC_NON_FINAL_FOLD)
1154 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1155 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
1156 # define _IS_MNEMONIC_CNTRL_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1157 _generic_isCC(c, _CC_MNEMONIC_CNTRL)
1158 #else /* else we don't have perl.h H_PERL */
1160 /* If we don't have perl.h, we are compiling a utility program. Below we
1161 * hard-code various macro definitions that wouldn't otherwise be available
1162 * to it. Most are coded based on first principles. These are written to
1163 * avoid EBCDIC vs. ASCII #ifdef's as much as possible. */
1164 # define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
1165 # define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1166 # define isSPACE_A(c) (isBLANK_A(c) \
1171 /* On EBCDIC, there are gaps between 'i' and 'j'; 'r' and 's'. Same for
1172 * uppercase. The tests for those aren't necessary on ASCII, but hurt only
1173 * performance (if optimization isn't on), and allow the same code to be
1174 * used for both platform types */
1175 # define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z' \
1177 || ((c) >= 'j' && (c) <= 'r') \
1179 # define isUPPER_A(c) ((c) >= 'A' && (c) <= 'Z' \
1181 || ((c) >= 'J' && (c) <= 'R') \
1183 # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
1184 # define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
1185 # define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1186 # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1187 # define isXDIGIT_A(c) (isDIGIT_A(c) \
1188 || ((c) >= 'a' && (c) <= 'f') \
1189 || ((c) <= 'F' && (c) >= 'A'))
1190 # define isPUNCT_A(c) ((c) == '-' || (c) == '!' || (c) == '"' \
1191 || (c) == '#' || (c) == '$' || (c) == '%' \
1192 || (c) == '&' || (c) == '\'' || (c) == '(' \
1193 || (c) == ')' || (c) == '*' || (c) == '+' \
1194 || (c) == ',' || (c) == '.' || (c) == '/' \
1195 || (c) == ':' || (c) == ';' || (c) == '<' \
1196 || (c) == '=' || (c) == '>' || (c) == '?' \
1197 || (c) == '@' || (c) == '[' || (c) == '\\' \
1198 || (c) == ']' || (c) == '^' || (c) == '_' \
1199 || (c) == '`' || (c) == '{' || (c) == '|' \
1200 || (c) == '}' || (c) == '~')
1201 # define isGRAPH_A(c) (isALPHANUMERIC_A(c) || isPUNCT_A(c))
1202 # define isPRINT_A(c) (isGRAPH_A(c) || (c) == ' ')
1205 /* The below is accurate for the 3 EBCDIC code pages traditionally
1206 * supported by perl. The only difference between them in the controls
1207 * is the position of \n, and that is represented symbolically below */
1208 # define isCNTRL_A(c) ((c) == '\0' || (c) == '\a' || (c) == '\b' \
1209 || (c) == '\f' || (c) == '\n' || (c) == '\r' \
1210 || (c) == '\t' || (c) == '\v' \
1211 || ((c) <= 3 && (c) >= 1) /* SOH, STX, ETX */ \
1212 || (c) == 7 /* U+7F DEL */ \
1213 || ((c) <= 0x13 && (c) >= 0x0E) /* SO, SI */ \
1214 /* DLE, DC[1-3] */ \
1215 || (c) == 0x18 /* U+18 CAN */ \
1216 || (c) == 0x19 /* U+19 EOM */ \
1217 || ((c) <= 0x1F && (c) >= 0x1C) /* [FGRU]S */ \
1218 || (c) == 0x26 /* U+17 ETB */ \
1219 || (c) == 0x27 /* U+1B ESC */ \
1220 || (c) == 0x2D /* U+05 ENQ */ \
1221 || (c) == 0x2E /* U+06 ACK */ \
1222 || (c) == 0x32 /* U+16 SYN */ \
1223 || (c) == 0x37 /* U+04 EOT */ \
1224 || (c) == 0x3C /* U+14 DC4 */ \
1225 || (c) == 0x3D /* U+15 NAK */ \
1226 || (c) == 0x3F)/* U+1A SUB */
1227 # define isASCII(c) (isCNTRL_A(c) || isPRINT_A(c))
1228 # else /* isASCII is already defined for ASCII platforms, so can use that to
1230 # define isCNTRL_A(c) (isASCII(c) && ! isPRINT_A(c))
1233 /* The _L1 macros may be unnecessary for the utilities; I (khw) added them
1234 * during debugging, and it seems best to keep them. We may be called
1235 * without NATIVE_TO_LATIN1 being defined. On ASCII platforms, it doesn't
1236 * do anything anyway, so make it not a problem */
1237 # if ! defined(EBCDIC) && ! defined(NATIVE_TO_LATIN1)
1238 # define NATIVE_TO_LATIN1(ch) (ch)
1240 # define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1241 # define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1242 # define isBLANK_L1(c) (isBLANK_A(c) \
1243 || (FITS_IN_8_BITS(c) \
1244 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1245 # define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1246 # define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1247 # define isLOWER_L1(c) (isLOWER_A(c) \
1248 || (FITS_IN_8_BITS(c) \
1249 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1250 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1251 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1252 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1253 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1254 # define isPRINT_L1(c) (isPRINT_A(c) \
1255 || (FITS_IN_8_BITS(c) \
1256 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1257 # define isPUNCT_L1(c) (isPUNCT_A(c) \
1258 || (FITS_IN_8_BITS(c) \
1259 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1260 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1261 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1262 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1263 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1264 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1265 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1266 # define isSPACE_L1(c) (isSPACE_A(c) \
1267 || (FITS_IN_8_BITS(c) \
1268 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1269 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1270 # define isUPPER_L1(c) (isUPPER_A(c) \
1271 || (FITS_IN_8_BITS(c) \
1272 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1273 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1274 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1275 # define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1276 # define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1277 # define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1282 /* The following are not fully accurate in the above-ASCII range. I (khw)
1283 * don't think it's necessary to be so for the purposes where this gets
1285 # define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1286 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1288 /* And these aren't accurate at all. They are useful only for above
1289 * Latin1, which utilities and bootstrapping don't deal with */
1290 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1291 # define _HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1292 # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1294 /* Many of the macros later in this file are defined in terms of these. By
1295 * implementing them with a function, which converts the class number into
1296 * a call to the desired macro, all of the later ones work. However, that
1297 * function won't be actually defined when building a utility program (no
1298 * perl.h), and so a compiler error will be generated if one is attempted
1299 * to be used. And the above-Latin1 code points require Unicode tables to
1300 * be present, something unlikely to be the case when bootstrapping */
1301 # define _generic_isCC(c, classnum) \
1302 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1303 # define _generic_isCC_A(c, classnum) \
1304 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
1305 #endif /* End of no perl.h H_PERL */
1307 #define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1308 #define isALPHA(c) isALPHA_A(c)
1309 #define isASCII_A(c) isASCII(c)
1310 #define isASCII_L1(c) isASCII(c)
1311 #define isBLANK(c) isBLANK_A(c)
1312 #define isCNTRL(c) isCNTRL_A(c)
1313 #define isDIGIT(c) isDIGIT_A(c)
1314 #define isGRAPH(c) isGRAPH_A(c)
1315 #define isIDFIRST(c) isIDFIRST_A(c)
1316 #define isLOWER(c) isLOWER_A(c)
1317 #define isPRINT(c) isPRINT_A(c)
1318 #define isPSXSPC_A(c) isSPACE_A(c)
1319 #define isPSXSPC(c) isPSXSPC_A(c)
1320 #define isPSXSPC_L1(c) isSPACE_L1(c)
1321 #define isPUNCT(c) isPUNCT_A(c)
1322 #define isSPACE(c) isSPACE_A(c)
1323 #define isUPPER(c) isUPPER_A(c)
1324 #define isWORDCHAR(c) isWORDCHAR_A(c)
1325 #define isXDIGIT(c) isXDIGIT_A(c)
1327 /* ASCII casing. These could also be written as
1328 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1329 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1330 which uses table lookup and mask instead of subtraction. (This would
1331 work because the _MOD does not apply in the ASCII range) */
1332 #define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1333 #define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
1335 /* In the ASCII range, these are equivalent to what they're here defined to be.
1336 * But by creating these definitions, other code doesn't have to be aware of
1338 #define toFOLD(c) toLOWER(c)
1339 #define toTITLE(c) toUPPER(c)
1341 #define toLOWER_A(c) toLOWER(c)
1342 #define toUPPER_A(c) toUPPER(c)
1343 #define toFOLD_A(c) toFOLD(c)
1344 #define toTITLE_A(c) toTITLE(c)
1346 /* Use table lookup for speed; returns the input itself if is out-of-range */
1347 #define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
1349 : PL_latin1_lc[ (U8) (c) ])
1350 #define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1352 /* Modified uc. Is correct uc except for three non-ascii chars which are
1353 * all mapped to one of them, and these need special handling; returns the
1354 * input itself if is out-of-range */
1355 #define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
1357 : PL_mod_latin1_uc[ (U8) (c) ])
1358 #define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale
1360 /* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
1362 /* For internal core Perl use only: the base macro for defining macros like
1363 * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point
1364 * (0-255) to check. In a UTF-8 locale, the result is the same as calling
1365 * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like
1366 * _CC_UPPER, which gives the class number for doing this. For non-UTF-8
1367 * locales, the code to actually do the test this is passed in 'non_utf8'. If
1368 * 'c' is above 255, 0 is returned. For accessing the full range of possible
1369 * code points under locale rules, use the macros based on _generic_LC_uvchr
1370 * instead of this. */
1371 #define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \
1372 (! FITS_IN_8_BITS(c) \
1374 : IN_UTF8_CTYPE_LOCALE \
1375 ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \
1378 /* For internal core Perl use only: a helper macro for defining macros like
1379 * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to
1380 * actually do this test is passed in 'non_utf8_func', which is called on 'c',
1381 * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See
1382 * _generic_LC_base for more info */
1383 #define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \
1384 _generic_LC_base(c,utf8_locale_classnum, \
1385 non_utf8_func( (_LC_CAST) (c)))
1387 /* For internal core Perl use only: like _generic_LC, but also returns TRUE if
1388 * 'c' is the platform's native underscore character */
1389 #define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \
1390 _generic_LC_base(c, utf8_locale_classnum, \
1391 (non_utf8_func( (_LC_CAST) (c)) \
1392 || (char)(c) == '_'))
1394 /* These next three are also for internal core Perl use only: case-change
1396 #define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \
1398 : (IN_UTF8_CTYPE_LOCALE) \
1399 ? PL_latin1_lc[ (U8) (c) ] \
1400 : (cast)function((cast)(c)))
1402 /* Note that the result can be larger than a byte in a UTF-8 locale. It
1403 * returns a single value, so can't adequately return the upper case of LATIN
1404 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1405 * values "SS"); instead it asserts against that under DEBUGGING, and
1406 * otherwise returns its input */
1407 #define _generic_toUPPER_LC(c, function, cast) \
1408 (! FITS_IN_8_BITS(c) \
1410 : ((! IN_UTF8_CTYPE_LOCALE) \
1411 ? (cast)function((cast)(c)) \
1412 : ((((U8)(c)) == MICRO_SIGN) \
1413 ? GREEK_CAPITAL_LETTER_MU \
1414 : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \
1415 ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \
1416 : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \
1417 ? (__ASSERT_(0) (c)) \
1418 : PL_mod_latin1_uc[ (U8) (c) ])))))
1420 /* Note that the result can be larger than a byte in a UTF-8 locale. It
1421 * returns a single value, so can't adequately return the fold case of LATIN
1422 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1423 * values "ss"); instead it asserts against that under DEBUGGING, and
1424 * otherwise returns its input */
1425 #define _generic_toFOLD_LC(c, function, cast) \
1426 ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \
1427 ? GREEK_SMALL_LETTER_MU \
1428 : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \
1429 || (c) != LATIN_SMALL_LETTER_SHARP_S) \
1430 _generic_toLOWER_LC(c, function, cast)))
1432 /* Use the libc versions for these if available. */
1433 #if defined(HAS_ISASCII)
1434 # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c)))
1436 # define isASCII_LC(c) isASCII(c)
1439 #if defined(HAS_ISBLANK)
1440 # define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank)
1441 #else /* Unlike isASCII, varies if in a UTF-8 locale */
1442 # define isBLANK_LC(c) ((IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c))
1448 /* The Windows functions don't bother to follow the POSIX standard, which
1449 * for example says that something can't both be a printable and a control.
1450 * But Windows treats the \t control as a printable, and does such things
1451 * as making superscripts into both digits and punctuation. This tames
1452 * these flaws by assuming that the definitions of both controls and space
1453 * are correct, and then making sure that other definitions don't have
1454 * weirdnesses, by making sure that isalnum() isn't also ispunct(), etc.
1455 * Not all possible weirdnesses are checked for, just the ones that were
1456 * detected on actual Microsoft code pages */
1458 # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1459 # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1461 # define isALPHA_LC(c) (_generic_LC(c, _CC_ALPHA, isalpha) \
1462 && isALPHANUMERIC_LC(c))
1463 # define isALPHANUMERIC_LC(c) (_generic_LC(c, _CC_ALPHANUMERIC, isalnum) && \
1465 # define isDIGIT_LC(c) (_generic_LC(c, _CC_DIGIT, isdigit) && \
1466 isALPHANUMERIC_LC(c))
1467 # define isGRAPH_LC(c) (_generic_LC(c, _CC_GRAPH, isgraph) && isPRINT_LC(c))
1468 # define isIDFIRST_LC(c) (((c) == '_') \
1469 || (_generic_LC(c, _CC_IDFIRST, isalpha) && ! isPUNCT_LC(c)))
1470 # define isLOWER_LC(c) (_generic_LC(c, _CC_LOWER, islower) && isALPHA_LC(c))
1471 # define isPRINT_LC(c) (_generic_LC(c, _CC_PRINT, isprint) && ! isCNTRL_LC(c))
1472 # define isPUNCT_LC(c) (_generic_LC(c, _CC_PUNCT, ispunct) && ! isCNTRL_LC(c))
1473 # define isUPPER_LC(c) (_generic_LC(c, _CC_UPPER, isupper) && isALPHA_LC(c))
1474 # define isWORDCHAR_LC(c) (((c) == '_') || isALPHANUMERIC_LC(c))
1475 # define isXDIGIT_LC(c) (_generic_LC(c, _CC_XDIGIT, isxdigit) \
1476 && isALPHANUMERIC_LC(c))
1478 # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1479 # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1480 # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
1482 #elif defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1483 /* For most other platforms */
1485 # define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha)
1486 # define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum)
1487 # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1488 # define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit)
1489 # define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph)
1490 # define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha)
1491 # define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower)
1492 # define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint)
1493 # define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct)
1494 # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1495 # define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper)
1496 # define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum)
1497 # define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit)
1500 # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1501 # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1502 # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
1504 #else /* The final fallback position */
1506 # define isALPHA_LC(c) (isascii(c) && isalpha(c))
1507 # define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
1508 # define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
1509 # define isDIGIT_LC(c) (isascii(c) && isdigit(c))
1510 # define isGRAPH_LC(c) (isascii(c) && isgraph(c))
1511 # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1512 # define isLOWER_LC(c) (isascii(c) && islower(c))
1513 # define isPRINT_LC(c) (isascii(c) && isprint(c))
1514 # define isPUNCT_LC(c) (isascii(c) && ispunct(c))
1515 # define isSPACE_LC(c) (isascii(c) && isspace(c))
1516 # define isUPPER_LC(c) (isascii(c) && isupper(c))
1517 # define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
1518 # define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
1520 # define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
1521 # define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
1522 # define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c))
1526 #define isIDCONT(c) isWORDCHAR(c)
1527 #define isIDCONT_A(c) isWORDCHAR_A(c)
1528 #define isIDCONT_L1(c) isWORDCHAR_L1(c)
1529 #define isIDCONT_LC(c) isWORDCHAR_LC(c)
1530 #define isPSXSPC_LC(c) isSPACE_LC(c)
1532 /* For internal core Perl use only: the base macros for defining macros like
1533 * isALPHA_uvchr. 'c' is the code point to check. 'classnum' is the POSIX class
1534 * number defined earlier in this file. _generic_uvchr() is used for POSIX
1535 * classes where there is a macro or function 'above_latin1' that takes the
1536 * single argument 'c' and returns the desired value. These exist for those
1537 * classes which have simple definitions, avoiding the overhead of a hash
1538 * lookup or inversion list binary search. _generic_swash_uvchr() can be used
1539 * for classes where that overhead is faster than a direct lookup.
1540 * _generic_uvchr() won't compile if 'c' isn't unsigned, as it won't match the
1541 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so
1542 * have duplicate checks here, so could create versions of the macros that
1543 * don't, but experiments show that gcc optimizes them out anyway. */
1545 /* Note that all ignore 'use bytes' */
1546 #define _generic_uvchr(classnum, above_latin1, c) ((c) < 256 \
1547 ? _generic_isCC(c, classnum) \
1549 #define _generic_swash_uvchr(classnum, c) ((c) < 256 \
1550 ? _generic_isCC(c, classnum) \
1551 : _is_uni_FOO(classnum, c))
1552 #define isALPHA_uvchr(c) _generic_swash_uvchr(_CC_ALPHA, c)
1553 #define isALPHANUMERIC_uvchr(c) _generic_swash_uvchr(_CC_ALPHANUMERIC, c)
1554 #define isASCII_uvchr(c) isASCII(c)
1555 #define isBLANK_uvchr(c) _generic_uvchr(_CC_BLANK, is_HORIZWS_cp_high, c)
1556 #define isCNTRL_uvchr(c) isCNTRL_L1(c) /* All controls are in Latin1 */
1557 #define isDIGIT_uvchr(c) _generic_swash_uvchr(_CC_DIGIT, c)
1558 #define isGRAPH_uvchr(c) _generic_swash_uvchr(_CC_GRAPH, c)
1559 #define isIDCONT_uvchr(c) _generic_uvchr(_CC_WORDCHAR, _is_uni_perl_idcont, c)
1560 #define isIDFIRST_uvchr(c) _generic_uvchr(_CC_IDFIRST, _is_uni_perl_idstart, c)
1561 #define isLOWER_uvchr(c) _generic_swash_uvchr(_CC_LOWER, c)
1562 #define isPRINT_uvchr(c) _generic_swash_uvchr(_CC_PRINT, c)
1564 #define isPUNCT_uvchr(c) _generic_swash_uvchr(_CC_PUNCT, c)
1565 #define isSPACE_uvchr(c) _generic_uvchr(_CC_SPACE, is_XPERLSPACE_cp_high, c)
1566 #define isPSXSPC_uvchr(c) isSPACE_uvchr(c)
1568 #define isUPPER_uvchr(c) _generic_swash_uvchr(_CC_UPPER, c)
1569 #define isVERTWS_uvchr(c) _generic_uvchr(_CC_VERTSPACE, is_VERTWS_cp_high, c)
1570 #define isWORDCHAR_uvchr(c) _generic_swash_uvchr(_CC_WORDCHAR, c)
1571 #define isXDIGIT_uvchr(c) _generic_uvchr(_CC_XDIGIT, is_XDIGIT_cp_high, c)
1573 #define toFOLD_uvchr(c,s,l) to_uni_fold(c,s,l)
1574 #define toLOWER_uvchr(c,s,l) to_uni_lower(c,s,l)
1575 #define toTITLE_uvchr(c,s,l) to_uni_title(c,s,l)
1576 #define toUPPER_uvchr(c,s,l) to_uni_upper(c,s,l)
1578 /* For backwards compatibility, even though '_uni' should mean official Unicode
1579 * code points, in Perl it means native for those below 256 */
1580 #define isALPHA_uni(c) isALPHA_uvchr(c)
1581 #define isALPHANUMERIC_uni(c) isALPHANUMERIC_uvchr(c)
1582 #define isASCII_uni(c) isASCII_uvchr(c)
1583 #define isBLANK_uni(c) isBLANK_uvchr(c)
1584 #define isCNTRL_uni(c) isCNTRL_uvchr(c)
1585 #define isDIGIT_uni(c) isDIGIT_uvchr(c)
1586 #define isGRAPH_uni(c) isGRAPH_uvchr(c)
1587 #define isIDCONT_uni(c) isIDCONT_uvchr(c)
1588 #define isIDFIRST_uni(c) isIDFIRST_uvchr(c)
1589 #define isLOWER_uni(c) isLOWER_uvchr(c)
1590 #define isPRINT_uni(c) isPRINT_uvchr(c)
1591 #define isPUNCT_uni(c) isPUNCT_uvchr(c)
1592 #define isSPACE_uni(c) isSPACE_uvchr(c)
1593 #define isPSXSPC_uni(c) isPSXSPC_uvchr(c)
1594 #define isUPPER_uni(c) isUPPER_uvchr(c)
1595 #define isVERTWS_uni(c) isVERTWS_uvchr(c)
1596 #define isWORDCHAR_uni(c) isWORDCHAR_uvchr(c)
1597 #define isXDIGIT_uni(c) isXDIGIT_uvchr(c)
1598 #define toFOLD_uni(c,s,l) toFOLD_uvchr(c,s,l)
1599 #define toLOWER_uni(c,s,l) toLOWER_uvchr(c,s,l)
1600 #define toTITLE_uni(c,s,l) toTITLE_uvchr(c,s,l)
1601 #define toUPPER_uni(c,s,l) toUPPER_uvchr(c,s,l)
1603 /* For internal core Perl use only: the base macros for defining macros like
1604 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code
1605 * point, not just 0-255. Like _generic_uvchr, there are two versions, one for
1606 * simple class definitions; the other for more complex. These are like
1607 * _generic_uvchr, so see it for more info. */
1608 #define _generic_LC_uvchr(latin1, above_latin1, c) \
1609 (c < 256 ? latin1(c) : above_latin1(c))
1610 #define _generic_LC_swash_uvchr(latin1, classnum, c) \
1611 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1613 #define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1614 #define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1615 _CC_ALPHANUMERIC, c)
1616 #define isASCII_LC_uvchr(c) isASCII_LC(c)
1617 #define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, \
1618 is_HORIZWS_cp_high, c)
1619 #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
1620 #define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1621 #define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
1622 #define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1623 _is_uni_perl_idcont, c)
1624 #define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1625 _is_uni_perl_idstart, c)
1626 #define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1627 #define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
1628 #define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c)
1629 #define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1630 #define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
1631 is_XPERLSPACE_cp_high, c)
1632 #define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1633 #define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1635 #define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, \
1636 is_XDIGIT_cp_high, c)
1638 #define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
1640 /* For internal core Perl use only: the base macros for defining macros like
1641 * isALPHA_utf8. These are like the earlier defined macros, but take an input
1642 * UTF-8 encoded string 'p'. If the input is in the Latin1 range, use
1643 * the Latin1 macro 'classnum' on 'p'. Otherwise use the value given by the
1644 * 'utf8' parameter. This relies on the fact that ASCII characters have the
1645 * same representation whether utf8 or not. Note that it assumes that the utf8
1646 * has been validated, and ignores 'use bytes' */
1647 #define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
1648 ? _generic_isCC(*(p), classnum) \
1649 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1651 EIGHT_BIT_UTF8_TO_NATIVE(*(p), \
1655 /* Like the above, but calls 'above_latin1(p)' to get the utf8 value.
1656 * 'above_latin1' can be a macro */
1657 #define _generic_func_utf8(classnum, above_latin1, p) \
1658 _generic_utf8(classnum, p, above_latin1(p))
1659 /* Like the above, but passes classnum to _isFOO_utf8(), instead of having an
1660 * 'above_latin1' parameter */
1661 #define _generic_swash_utf8(classnum, p) \
1662 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
1664 /* Like the above, but should be used only when it is known that there are no
1665 * characters in the upper-Latin1 range (128-255 on ASCII platforms) which the
1666 * class is TRUE for. Hence it can skip the tests for this range.
1667 * 'above_latin1' should include its arguments */
1668 #define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
1669 (UTF8_IS_INVARIANT(*(p)) \
1670 ? _generic_isCC(*(p), classnum) \
1671 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
1675 /* NOTE that some of these macros have very similar ones in regcharclass.h.
1676 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1677 * there, differing in name only by an underscore from the one here
1678 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1679 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1680 * points; the regcharclass.h ones are implemented as a series of
1681 * "if-else-if-else ..." */
1683 #define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1684 #define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
1685 #define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1686 utf8, the non-utf8 macro works
1688 #define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1691 /* Because all controls are UTF-8 invariants in EBCDIC, we can use this
1692 * more efficient macro instead of the more general one */
1693 # define isCNTRL_utf8(p) isCNTRL_L1(*(p))
1695 # define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1698 #define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1699 _is_utf8_FOO(_CC_DIGIT, p))
1700 #define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
1701 #define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
1702 _is_utf8_perl_idcont, p)
1704 /* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1705 * IDFIRST is an alnum. See
1706 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
1707 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1708 * This used to be not the XID version, but we decided to go with the more
1709 * modern Unicode definition */
1710 #define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
1711 _is_utf8_perl_idstart, p)
1713 #define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1714 #define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
1715 #define isPSXSPC_utf8(p) isSPACE_utf8(p)
1716 #define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
1717 #define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
1718 #define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
1719 #define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
1720 #define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
1721 #define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1724 #define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
1725 #define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1726 #define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1727 #define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
1729 /* For internal core Perl use only: the base macros for defining macros like
1730 * isALPHA_LC_utf8. These are like _generic_utf8, but if the first code point
1731 * in 'p' is within the 0-255 range, it uses locale rules from the passed-in
1732 * 'macro' parameter */
1733 #define _generic_LC_utf8(macro, p, utf8) \
1734 (UTF8_IS_INVARIANT(*(p)) \
1736 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1737 ? macro(EIGHT_BIT_UTF8_TO_NATIVE(*(p), *((p)+1)))\
1740 #define _generic_LC_swash_utf8(macro, classnum, p) \
1741 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1742 #define _generic_LC_func_utf8(macro, above_latin1, p) \
1743 _generic_LC_utf8(macro, p, above_latin1(p))
1745 #define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1746 _CC_ALPHANUMERIC, p)
1747 #define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
1748 #define isASCII_LC_utf8(p) isASCII_LC(*p)
1749 #define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, \
1751 #define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1752 #define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1753 #define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
1754 #define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, \
1755 _is_utf8_perl_idcont, p)
1756 #define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, \
1757 _is_utf8_perl_idstart, p)
1758 #define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1759 #define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
1760 #define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p)
1761 #define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1762 #define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, \
1763 is_XPERLSPACE_high, p)
1764 #define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1765 #define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1767 #define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, \
1770 /* Macros for backwards compatibility and for completeness when the ASCII and
1771 * Latin1 values are identical */
1772 #define isALPHAU(c) isALPHA_L1(c)
1773 #define isDIGIT_L1(c) isDIGIT_A(c)
1774 #define isOCTAL(c) isOCTAL_A(c)
1775 #define isOCTAL_L1(c) isOCTAL_A(c)
1776 #define isXDIGIT_L1(c) isXDIGIT_A(c)
1777 #define isALNUM(c) isWORDCHAR(c)
1778 #define isALNUMU(c) isWORDCHAR_L1(c)
1779 #define isALNUM_LC(c) isWORDCHAR_LC(c)
1780 #define isALNUM_uni(c) isWORDCHAR_uni(c)
1781 #define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1782 #define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1783 #define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
1784 #define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1785 #define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1786 #define isALNUMC(c) isALPHANUMERIC(c)
1787 #define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1788 #define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1789 #define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1790 #define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1791 #define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
1793 /* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII,
1794 * except that they don't necessarily mean the same characters, e.g. CTRL-D is
1795 * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC.
1796 * '?' is special-cased on EBCDIC to APC, which is the control there that is
1797 * the outlier from the block that contains the other controls, just like
1798 * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0
1799 * block. If it weren't special cased, it would yield a non-control.
1800 * The conversion works both ways, so toCTRL('D') is 4, and toCTRL(4) is D,
1803 # define toCTRL(c) (__ASSERT_(FITS_IN_8_BITS(c)) toUPPER(((U8)(c))) ^ 64)
1805 # define toCTRL(c) (__ASSERT_(FITS_IN_8_BITS(c)) \
1807 ? (UNLIKELY((c) == '?') \
1808 ? QUESTION_MARK_CTRL \
1809 : (NATIVE_TO_LATIN1(toUPPER((U8) (c))) ^ 64)) \
1810 : (UNLIKELY((c) == QUESTION_MARK_CTRL) \
1812 : (LATIN1_TO_NATIVE(((U8) (c)) ^ 64)))))
1815 /* Line numbers are unsigned, 32 bits. */
1817 #define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
1819 /* Helpful alias for version prescan */
1820 #define is_LAX_VERSION(a,b) \
1821 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1823 #define is_STRICT_VERSION(a,b) \
1824 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1826 #define BADVERSION(a,b,c) \
1832 /* Converts a character known to represent a hexadecimal digit (0-9, A-F, or
1833 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer,
1834 * which is advanced. The input is validated only by an assert() in DEBUGGING
1835 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and
1836 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */
1837 #define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \
1840 #define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \
1844 /* Converts a character known to represent an octal digit (0-7) to its numeric
1845 * value. The input is validated only by an assert() in DEBUGGING builds. In
1846 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */
1847 #define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c)))
1849 /* Efficiently returns a boolean as to if two native characters are equivalent
1850 * case-insenstively. At least one of the characters must be one of [A-Za-z];
1851 * the ALPHA in the name is to remind you of that. This is asserted() in
1852 * DEBUGGING builds. Because [A-Za-z] are invariant under UTF-8, this macro
1853 * works (on valid input) for both non- and UTF-8-encoded bytes.
1855 * When one of the inputs is a compile-time constant and gets folded by the
1856 * compiler, this reduces to an AND and a TEST. On both EBCDIC and ASCII
1857 * machines, 'A' and 'a' differ by a single bit; the same with the upper and
1858 * lower case of all other ASCII-range alphabetics. On ASCII platforms, they
1859 * are 32 apart; on EBCDIC, they are 64. At compile time, this uses an
1860 * exclusive 'or' to find that bit and then inverts it to form a mask, with
1861 * just a single 0, in the bit position where the upper- and lowercase differ.
1863 #define isALPHA_FOLD_EQ(c1, c2) \
1864 (__ASSERT_(isALPHA_A(c1) || isALPHA_A(c2)) \
1865 ((c1) & ~('A' ^ 'a')) == ((c2) & ~('A' ^ 'a')))
1866 #define isALPHA_FOLD_NE(c1, c2) (! isALPHA_FOLD_EQ((c1), (c2)))
1869 =head1 Memory Management
1871 =for apidoc Am|void|Newx|void* ptr|int nitems|type
1872 The XSUB-writer's interface to the C C<malloc> function.
1874 Memory obtained by this should B<ONLY> be freed with L</"Safefree">.
1876 In 5.9.3, Newx() and friends replace the older New() API, and drops
1877 the first parameter, I<x>, a debug aid which allowed callers to identify
1878 themselves. This aid has been superseded by a new build option,
1879 PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
1880 there for use in XS modules supporting older perls.
1882 =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
1883 The XSUB-writer's interface to the C C<malloc> function, with
1884 cast. See also C<L</Newx>>.
1886 Memory obtained by this should B<ONLY> be freed with L</"Safefree">.
1888 =for apidoc Am|void|Newxz|void* ptr|int nitems|type
1889 The XSUB-writer's interface to the C C<malloc> function. The allocated
1890 memory is zeroed with C<memzero>. See also C<L</Newx>>.
1892 Memory obtained by this should B<ONLY> be freed with L</"Safefree">.
1894 =for apidoc Am|void|Renew|void* ptr|int nitems|type
1895 The XSUB-writer's interface to the C C<realloc> function.
1897 Memory obtained by this should B<ONLY> be freed with L</"Safefree">.
1899 =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1900 The XSUB-writer's interface to the C C<realloc> function, with
1903 Memory obtained by this should B<ONLY> be freed with L</"Safefree">.
1905 =for apidoc Am|void|Safefree|void* ptr
1906 The XSUB-writer's interface to the C C<free> function.
1908 This should B<ONLY> be used on memory obtained using L</"Newx"> and friends.
1910 =for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1911 The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
1912 source, C<dest> is the destination, C<nitems> is the number of items, and
1913 C<type> is the type. Can do overlapping moves. See also C<L</Copy>>.
1915 =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1916 Like C<Move> but returns C<dest>. Useful
1917 for encouraging compilers to tail-call
1920 =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1921 The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
1922 source, C<dest> is the destination, C<nitems> is the number of items, and
1923 C<type> is the type. May fail on overlapping copies. See also C<L</Move>>.
1925 =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1927 Like C<Copy> but returns C<dest>. Useful
1928 for encouraging compilers to tail-call
1931 =for apidoc Am|void|Zero|void* dest|int nitems|type
1933 The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1934 destination, C<nitems> is the number of items, and C<type> is the type.
1936 =for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1938 Like C<Zero> but returns dest. Useful
1939 for encouraging compilers to tail-call
1942 =for apidoc Am|void|StructCopy|type *src|type *dest|type
1943 This is an architecture-independent macro to copy one structure to another.
1945 =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1947 Fill up memory with a byte pattern (a byte repeated over and over
1948 again) that hopefully catches attempts to access uninitialized memory.
1950 =for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1952 PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1954 =for apidoc Am|void|PoisonFree|void* dest|int nitems|type
1956 PoisonWith(0xEF) for catching access to freed memory.
1958 =for apidoc Am|void|Poison|void* dest|int nitems|type
1960 PoisonWith(0xEF) for catching access to freed memory.
1964 /* Maintained for backwards-compatibility only. Use newSV() instead. */
1966 #define NEWSV(x,len) newSV(len)
1969 #define MEM_SIZE_MAX ((MEM_SIZE)~0)
1972 #ifdef PERL_MALLOC_WRAP
1974 /* This expression will be constant-folded at compile time. It checks
1975 * whether or not the type of the count n is so small (e.g. U8 or U16, or
1976 * U32 on 64-bit systems) that there's no way a wrap-around could occur.
1977 * As well as avoiding the need for a run-time check in some cases, it's
1978 * designed to avoid compiler warnings like:
1979 * comparison is always false due to limited range of data type
1980 * It's mathematically equivalent to
1981 * max(n) * sizeof(t) > MEM_SIZE_MAX
1984 # define _MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) \
1985 (8 * sizeof(n) + sizeof(t) > sizeof(MEM_SIZE))
1987 /* This is written in a slightly odd way to avoid various spurious
1988 * compiler warnings. We *want* to write the expression as
1989 * _MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) && (n > C)
1990 * (for some compile-time constant C), but even when the LHS
1991 * constant-folds to false at compile-time, g++ insists on emitting
1992 * warnings about the RHS (e.g. "comparison is always false"), so instead
1995 * (cond ? n : X) > C
1997 * where X is a constant with X > C always false. Choosing a value for X
1998 * is tricky. If 0, some compilers will complain about 0 > C always being
1999 * false; if 1, Coverity complains when n happens to be the constant value
2000 * '1', that cond ? 1 : 1 has the same value on both branches; so use C
2001 * for X and hope that nothing else whines.
2004 # define _MEM_WRAP_WILL_WRAP(n,t) \
2005 ((_MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) ? (MEM_SIZE)(n) : \
2006 MEM_SIZE_MAX/sizeof(t)) > MEM_SIZE_MAX/sizeof(t))
2008 # define MEM_WRAP_CHECK(n,t) \
2009 (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \
2010 && (croak_memory_wrap(),0))
2012 # define MEM_WRAP_CHECK_1(n,t,a) \
2013 (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \
2014 && (Perl_croak_nocontext("%s",(a)),0))
2016 #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
2018 #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)))
2021 #define MEM_WRAP_CHECK(n,t)
2022 #define MEM_WRAP_CHECK_1(n,t,a)
2023 #define MEM_WRAP_CHECK_2(n,t,a,b)
2024 #define MEM_WRAP_CHECK_(n,t)
2026 #define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
2032 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
2033 * go through functions, which are handy for debugging breakpoints, but
2034 * which more importantly get the immediate calling environment (file and
2035 * line number, and C function name if available) passed in. This info can
2036 * then be used for logging the calls, for which one gets a sample
2037 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
2040 * - not all memory allocs get logged, only those
2041 * that go through Newx() and derivatives (while all
2042 * Safefrees do get logged)
2043 * - __FILE__ and __LINE__ do not work everywhere
2044 * - __func__ or __FUNCTION__ even less so
2045 * - I think more goes on after the perlio frees but
2046 * the thing is that STDERR gets closed (as do all
2047 * the file descriptors)
2048 * - no deeper calling stack than the caller of the Newx()
2049 * or the kind, but do I look like a C reflection/introspection
2051 * - the function prototypes for the logging functions
2052 * probably should maybe be somewhere else than handy.h
2053 * - one could consider inlining (macrofying) the logging
2054 * for speed, but I am too lazy
2055 * - one could imagine recording the allocations in a hash,
2056 * (keyed by the allocation address?), and maintain that
2057 * through reallocs and frees, but how to do that without
2058 * any News() happening...?
2059 * - lots of -Ddefines to get useful/controllable output
2060 * - lots of ENV reads
2064 # ifndef PERL_MEM_LOG_NOIMPL
2073 # if defined(PERL_IN_SV_C) /* those are only used in sv.c */
2074 void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
2075 void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
2082 #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
2083 #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
2084 #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
2087 #ifndef MEM_LOG_ALLOC
2088 #define MEM_LOG_ALLOC(n,t,a) (a)
2090 #ifndef MEM_LOG_REALLOC
2091 #define MEM_LOG_REALLOC(n,t,v,a) (a)
2093 #ifndef MEM_LOG_FREE
2094 #define MEM_LOG_FREE(a) (a)
2097 #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
2098 #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
2099 #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
2102 /* pre 5.9.x compatibility */
2103 #define New(x,v,n,t) Newx(v,n,t)
2104 #define Newc(x,v,n,t,c) Newxc(v,n,t,c)
2105 #define Newz(x,v,n,t) Newxz(v,n,t)
2108 #define Renew(v,n,t) \
2109 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
2110 #define Renewc(v,n,t,c) \
2111 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
2114 #define Safefree(d) \
2115 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
2117 #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
2120 #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
2121 #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
2122 #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
2124 #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
2125 #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
2127 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
2129 /* Using bzero(), which returns void. */
2130 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
2133 #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
2134 #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
2135 #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
2136 #define Poison(d,n,t) PoisonFree(d,n,t)
2139 # define PERL_POISON_EXPR(x) x
2141 # define PERL_POISON_EXPR(x)
2144 #ifdef USE_STRUCT_COPY
2145 #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
2147 #define StructCopy(s,d,t) Copy(s,d,1,t)
2150 /* C_ARRAY_LENGTH is the number of elements in the C array (so you
2151 * want your zero-based indices to be less than but not equal to).
2153 * C_ARRAY_END is one past the last: half-open/half-closed range,
2154 * not last-inclusive range. */
2155 #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
2156 #define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a))
2160 # define Perl_va_copy(s, d) va_copy(d, s)
2162 # if defined(__va_copy)
2163 # define Perl_va_copy(s, d) __va_copy(d, s)
2165 # define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
2170 /* convenience debug macros */
2172 #define pTHX_FORMAT "Perl interpreter: 0x%p"
2173 #define pTHX__FORMAT ", Perl interpreter: 0x%p"
2174 #define pTHX_VALUE_ (void *)my_perl,
2175 #define pTHX_VALUE (void *)my_perl
2176 #define pTHX__VALUE_ ,(void *)my_perl,
2177 #define pTHX__VALUE ,(void *)my_perl
2180 #define pTHX__FORMAT
2183 #define pTHX__VALUE_
2185 #endif /* USE_ITHREADS */
2187 /* Perl_deprecate was not part of the public API, and did not have a deprecate()
2188 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
2189 CPAN::Unpack show any users outside the core. */
2191 # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \
2192 "Use of " s " is deprecated")
2195 /* Internal macros to deal with gids and uids */
2198 # if Uid_t_size > IVSIZE
2199 # define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
2200 # define SvUID(sv) SvNV(sv)
2202 # if Uid_t_sign <= 0
2203 # define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
2204 # define SvUID(sv) SvIV(sv)
2206 # define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
2207 # define SvUID(sv) SvUV(sv)
2209 # endif /* Uid_t_size */
2211 # if Gid_t_size > IVSIZE
2212 # define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
2213 # define SvGID(sv) SvNV(sv)
2215 # if Gid_t_sign <= 0
2216 # define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
2217 # define SvGID(sv) SvIV(sv)
2219 # define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
2220 # define SvGID(sv) SvUV(sv)
2222 # endif /* Gid_t_size */
2226 #endif /* HANDY_H */
2229 * ex: set ts=8 sts=4 sw=4 et:
https://perl5.git.perl.org / perl5.git / blob