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 PERL_HANDY_H_ /* Guard against nested #inclusion */
18 # define Null(type) ((type)NULL)
23 =for apidoc AmnU||Nullch
24 Null character pointer. (No longer available when C<PERL_CORE> is
27 =for apidoc AmnU||Nullsv
28 Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
33 # define Nullch Null(char*)
34 # define Nullfp Null(PerlIO*)
35 # define Nullsv Null(SV*)
47 /* The MUTABLE_*() macros cast pointers to the types shown, in such a way
48 * (compiler permitting) that casting away const-ness will give a warning;
52 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
53 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
56 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
57 # define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
59 # define MUTABLE_PTR(p) ((void *) (p))
62 #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
63 #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
64 #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
65 #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
66 #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
67 #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
69 #if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR)
76 /* bool is built-in for g++-2.6.3 and later, which might be used
77 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
78 be sure _G_config.h will be included before this file. _G_config.h
79 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
80 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
81 g++ can be identified by __GNUG__.
82 Andy Dougherty February 2000
84 #ifdef __GNUG__ /* GNU g++ has bool built-in */
85 # ifndef PERL_BOOL_AS_CHAR
101 =for apidoc Am|bool|cBOOL|bool expr
103 Cast-to-bool. A simple S<C<(bool) I<expr>>> cast may not do the right thing:
104 if C<bool> is defined as C<char>, for example, then the cast from C<int> is
105 implementation-defined.
107 C<(bool)!!(cbool)> in a ternary triggers a bug in xlc on AIX
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__
119 #elif (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
120 # define FUNCTION__ ""
122 # define FUNCTION__ __FUNCTION__ /* Common extension. */
125 /* XXX A note on the perl source internal type system. The
126 original intent was that I32 be *exactly* 32 bits.
128 Currently, we only guarantee that I32 is *at least* 32 bits.
129 Specifically, if int is 64 bits, then so is I32. (This is the case
130 for the Cray.) This has the advantage of meshing nicely with
131 standard library calls (where we pass an I32 and the library is
132 expecting an int), but the disadvantage that an I32 is not 32 bits.
133 Andy Dougherty August 1996
135 There is no guarantee that there is *any* integral type with
136 exactly 32 bits. It is perfectly legal for a system to have
137 sizeof(short) == sizeof(int) == sizeof(long) == 8.
139 Similarly, there is no guarantee that I16 and U16 have exactly 16
142 For dealing with issues that may arise from various 32/64-bit
143 systems, we will ask Configure to check out
145 SHORTSIZE == sizeof(short)
146 INTSIZE == sizeof(int)
147 LONGSIZE == sizeof(long)
148 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
149 PTRSIZE == sizeof(void *)
150 DOUBLESIZE == sizeof(double)
151 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
155 #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
156 # include <inttypes.h>
157 # ifdef INT32_MIN_BROKEN
159 # define INT32_MIN (-2147483647-1)
161 # ifdef INT64_MIN_BROKEN
163 # define INT64_MIN (-9223372036854775807LL-1)
179 #if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
181 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
182 Please search CHAR_MAX in perl.h for further details. */
183 #define U8_MAX UINT8_MAX
184 #define U8_MIN UINT8_MIN
186 #define I16_MAX INT16_MAX
187 #define I16_MIN INT16_MIN
188 #define U16_MAX UINT16_MAX
189 #define U16_MIN UINT16_MIN
191 #define I32_MAX INT32_MAX
192 #define I32_MIN INT32_MIN
193 #ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
194 # define U32_MAX UINT32_MAX
196 # define U32_MAX 4294967295U
198 #define U32_MIN UINT32_MIN
202 /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
203 Please search CHAR_MAX in perl.h for further details. */
204 #define U8_MAX PERL_UCHAR_MAX
205 #define U8_MIN PERL_UCHAR_MIN
207 #define I16_MAX PERL_SHORT_MAX
208 #define I16_MIN PERL_SHORT_MIN
209 #define U16_MAX PERL_USHORT_MAX
210 #define U16_MIN PERL_USHORT_MIN
213 # define I32_MAX PERL_INT_MAX
214 # define I32_MIN PERL_INT_MIN
215 # define U32_MAX PERL_UINT_MAX
216 # define U32_MIN PERL_UINT_MIN
218 # define I32_MAX PERL_LONG_MAX
219 # define I32_MIN PERL_LONG_MIN
220 # define U32_MAX PERL_ULONG_MAX
221 # define U32_MIN PERL_ULONG_MIN
226 /* These C99 typedefs are useful sometimes for, say, loop variables whose
227 * maximum values are small, but for which speed trumps size. If we have a C99
228 * compiler, use that. Otherwise, a plain 'int' should be good enough.
230 * Restrict these to core for now until we are more certain this is a good
232 #if defined(PERL_CORE) || defined(PERL_EXT)
234 typedef int_fast8_t PERL_INT_FAST8_T;
235 typedef uint_fast8_t PERL_UINT_FAST8_T;
236 typedef int_fast16_t PERL_INT_FAST16_T;
237 typedef uint_fast16_t PERL_UINT_FAST16_T;
239 typedef int PERL_INT_FAST8_T;
240 typedef unsigned int PERL_UINT_FAST8_T;
241 typedef int PERL_INT_FAST16_T;
242 typedef unsigned int PERL_UINT_FAST16_T;
246 /* log(2) (i.e., log base 10 of 2) is pretty close to 0.30103, just in case
247 * anyone is grepping for it */
248 #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log10(2) =~ 146/485 */
249 #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
250 #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
252 /* Unused by core; should be deprecated */
253 #define Ctl(ch) ((ch) & 037)
255 #if defined(PERL_CORE) || defined(PERL_EXT)
257 # define MIN(a,b) ((a) < (b) ? (a) : (b))
260 # define MAX(a,b) ((a) > (b) ? (a) : (b))
264 /* Returns a boolean as to whether the input unsigned number is a power of 2
265 * (2**0, 2**1, etc). In other words if it has just a single bit set.
266 * If not, subtracting 1 would leave the uppermost bit set, so the & would
268 #if defined(PERL_CORE) || defined(PERL_EXT)
269 # define isPOWER_OF_2(n) ((n) && ((n) & ((n)-1)) == 0)
273 =for apidoc Am|void|__ASSERT_|bool expr
275 This is a helper macro to avoid preprocessor issues, replaced by nothing
276 unless under DEBUGGING, where it expands to an assert of its argument,
277 followed by a comma (hence the comma operator). If we just used a straight
278 assert(), we would get a comma with nothing before it when not DEBUGGING.
282 We also use empty definition under Coverity since the __ASSERT__
283 checks often check for things that Really Cannot Happen, and Coverity
284 detects that and gets all excited. */
286 #if defined(DEBUGGING) && !defined(__COVERITY__)
287 # define __ASSERT_(statement) assert(statement),
289 # define __ASSERT_(statement)
293 =head1 SV Manipulation Functions
295 =for apidoc Ama|SV*|newSVpvs|"literal string"
296 Like C<newSVpvn>, but takes a literal string instead of a
299 =for apidoc Ama|SV*|newSVpvs_flags|"literal string"|U32 flags
300 Like C<newSVpvn_flags>, but takes a literal string instead of
301 a string/length pair.
303 =for apidoc Ama|SV*|newSVpvs_share|"literal string"
304 Like C<newSVpvn_share>, but takes a literal string instead of
305 a string/length pair and omits the hash parameter.
307 =for apidoc Am|void|sv_catpvs_flags|SV* sv|"literal string"|I32 flags
308 Like C<sv_catpvn_flags>, but takes a literal string instead
309 of a string/length pair.
311 =for apidoc Am|void|sv_catpvs_nomg|SV* sv|"literal string"
312 Like C<sv_catpvn_nomg>, but takes a literal string instead of
313 a string/length pair.
315 =for apidoc Am|void|sv_catpvs|SV* sv|"literal string"
316 Like C<sv_catpvn>, but takes a literal string instead of a
319 =for apidoc Am|void|sv_catpvs_mg|SV* sv|"literal string"
320 Like C<sv_catpvn_mg>, but takes a literal string instead of a
323 =for apidoc Am|void|sv_setpvs|SV* sv|"literal string"
324 Like C<sv_setpvn>, but takes a literal string instead of a
327 =for apidoc Am|void|sv_setpvs_mg|SV* sv|"literal string"
328 Like C<sv_setpvn_mg>, but takes a literal string instead of a
331 =for apidoc Am|SV *|sv_setref_pvs|SV *const rv|const char *const classname|"literal string"
332 Like C<sv_setref_pvn>, but takes a literal string instead of
333 a string/length pair.
335 =head1 Memory Management
337 =for apidoc Ama|char*|savepvs|"literal string"
338 Like C<savepvn>, but takes a literal string instead of a
341 =for apidoc Ama|char*|savesharedpvs|"literal string"
342 A version of C<savepvs()> which allocates the duplicate string in memory
343 which is shared between threads.
347 =for apidoc Am|HV*|gv_stashpvs|"name"|I32 create
348 Like C<gv_stashpvn>, but takes a literal string instead of a
351 =head1 Hash Manipulation Functions
353 =for apidoc Am|SV**|hv_fetchs|HV* tb|"key"|I32 lval
354 Like C<hv_fetch>, but takes a literal string instead of a
357 =for apidoc Am|SV**|hv_stores|HV* tb|"key"|SV* val
358 Like C<hv_store>, but takes a literal string instead of a
360 and omits the hash parameter.
362 =head1 Lexer interface
364 =for apidoc Amx|void|lex_stuff_pvs|"pv"|U32 flags
366 Like L</lex_stuff_pvn>, but takes a literal string instead of
367 a string/length pair.
375 =for apidoc Amu|pair|STR_WITH_LEN|"literal string"
377 Returns two comma separated tokens of the input literal string, and its length.
378 This is convenience macro which helps out in some API calls.
379 Note that it can't be used as an argument to macros or functions that under
380 some configurations might be macros, which means that it requires the full
381 Perl_xxx(aTHX_ ...) form for any API calls where it's used.
387 #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
389 /* STR_WITH_LEN() shortcuts */
390 #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
391 #define newSVpvs_flags(str,flags) \
392 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
393 #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
394 #define sv_catpvs_flags(sv, str, flags) \
395 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
396 #define sv_catpvs_nomg(sv, str) \
397 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
398 #define sv_catpvs(sv, str) \
399 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
400 #define sv_catpvs_mg(sv, str) \
401 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
402 #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
403 #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
404 #define sv_setref_pvs(rv, classname, str) \
405 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
406 #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
407 #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
408 #define gv_stashpvs(str, create) \
409 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
410 #define gv_fetchpvs(namebeg, add, sv_type) \
411 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
412 #define gv_fetchpvn(namebeg, len, add, sv_type) \
413 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
414 #define sv_catxmlpvs(dsv, str, utf8) \
415 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
418 #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
420 #define get_cvs(str, flags) \
421 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
424 =head1 Miscellaneous Functions
426 =for apidoc Am|bool|strNE|char* s1|char* s2
427 Test two C<NUL>-terminated strings to see if they are different. Returns true
430 =for apidoc Am|bool|strEQ|char* s1|char* s2
431 Test two C<NUL>-terminated strings to see if they are equal. Returns true or
434 =for apidoc Am|bool|strLT|char* s1|char* s2
435 Test two C<NUL>-terminated strings to see if the first, C<s1>, is less than the
436 second, C<s2>. Returns true or false.
438 =for apidoc Am|bool|strLE|char* s1|char* s2
439 Test two C<NUL>-terminated strings to see if the first, C<s1>, is less than or
440 equal to the second, C<s2>. Returns true or false.
442 =for apidoc Am|bool|strGT|char* s1|char* s2
443 Test two C<NUL>-terminated strings to see if the first, C<s1>, is greater than
444 the second, C<s2>. Returns true or false.
446 =for apidoc Am|bool|strGE|char* s1|char* s2
447 Test two C<NUL>-terminated strings to see if the first, C<s1>, is greater than
448 or equal to the second, C<s2>. Returns true or false.
450 =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
451 Test two C<NUL>-terminated strings to see if they are different. The C<len>
452 parameter indicates the number of bytes to compare. Returns true or false. (A
453 wrapper for C<strncmp>).
455 =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
456 Test two C<NUL>-terminated strings to see if they are equal. The C<len>
457 parameter indicates the number of bytes to compare. Returns true or false. (A
458 wrapper for C<strncmp>).
460 =for apidoc Am|bool|memEQ|char* s1|char* s2|STRLEN len
461 Test two buffers (which may contain embedded C<NUL> characters, to see if they
462 are equal. The C<len> parameter indicates the number of bytes to compare.
463 Returns zero if equal, or non-zero if non-equal.
465 =for apidoc Am|bool|memEQs|char* s1|STRLEN l1|"s2"
466 Like L</memEQ>, but the second string is a literal enclosed in double quotes,
467 C<l1> gives the number of bytes in C<s1>.
468 Returns zero if equal, or non-zero if non-equal.
470 =for apidoc Am|bool|memNE|char* s1|char* s2|STRLEN len
471 Test two buffers (which may contain embedded C<NUL> characters, to see if they
472 are not equal. The C<len> parameter indicates the number of bytes to compare.
473 Returns zero if non-equal, or non-zero if equal.
475 =for apidoc Am|bool|memNEs|char* s1|STRLEN l1|"s2"
476 Like L</memNE>, but the second string is a literal enclosed in double quotes,
477 C<l1> gives the number of bytes in C<s1>.
478 Returns zero if non-equal, or zero if non-equal.
482 New macros should use the following conventions for their names (which are
483 based on the underlying C library functions):
485 (mem | str n? ) (EQ | NE | LT | GT | GE | (( BEGIN | END ) P? )) l? s?
487 Each has two main parameters, string-like operands that are compared
488 against each other, as specified by the macro name. Some macros may
489 additionally have one or potentially even two length parameters. If a length
490 parameter applies to both string parameters, it will be positioned third;
491 otherwise any length parameter immediately follows the string parameter it
494 If the prefix to the name is 'str', the string parameter is a pointer to a C
495 language string. Such a string does not contain embedded NUL bytes; its
496 length may be unknown, but can be calculated by C<strlen()>, since it is
497 terminated by a NUL, which isn't included in its length.
499 The optional 'n' following 'str' means that that there is a third parameter,
500 giving the maximum number of bytes to look at in each string. Even if both
501 strings are longer than the length parameter, those extra bytes will be
504 The 's' suffix means that the 2nd byte string parameter is a literal C
505 double-quoted string. Its length will automatically be calculated by the
506 macro, so no length parameter will ever be needed for it.
508 If the prefix is 'mem', the string parameters don't have to be C strings;
509 they may contain embedded NUL bytes, do not necessarily have a terminating
510 NUL, and their lengths can be known only through other means, which in
511 practice are additional parameter(s) passed to the function. All 'mem'
512 functions have at least one length parameter. Barring any 'l' or 's' suffix,
513 there is a single length parameter, in position 3, which applies to both
514 string parameters. The 's' suffix means, as described above, that the 2nd
515 string is a literal double-quoted C string (hence its length is calculated by
516 the macro, and the length parameter to the function applies just to the first
517 string parameter, and hence is positioned just after it). An 'l' suffix
518 means that the 2nd string parameter has its own length parameter, and the
519 signature will look like memFOOl(s1, l1, s2, l2).
521 BEGIN (and END) are for testing if the 2nd string is an initial (or final)
522 substring of the 1st string. 'P' if present indicates that the substring
523 must be a "proper" one in tha mathematical sense that the first one must be
524 strictly larger than the 2nd.
529 #define strNE(s1,s2) (strcmp(s1,s2) != 0)
530 #define strEQ(s1,s2) (strcmp(s1,s2) == 0)
531 #define strLT(s1,s2) (strcmp(s1,s2) < 0)
532 #define strLE(s1,s2) (strcmp(s1,s2) <= 0)
533 #define strGT(s1,s2) (strcmp(s1,s2) > 0)
534 #define strGE(s1,s2) (strcmp(s1,s2) >= 0)
536 #define strnNE(s1,s2,l) (strncmp(s1,s2,l) != 0)
537 #define strnEQ(s1,s2,l) (strncmp(s1,s2,l) == 0)
539 #define memEQ(s1,s2,l) (memcmp(((const void *) (s1)), ((const void *) (s2)), l) == 0)
540 #define memNE(s1,s2,l) (! memEQ(s1,s2,l))
542 /* memEQ and memNE where second comparand is a string constant */
543 #define memEQs(s1, l, s2) \
544 (((sizeof(s2)-1) == (l)) && memEQ((s1), ("" s2 ""), (sizeof(s2)-1)))
545 #define memNEs(s1, l, s2) (! memEQs(s1, l, s2))
547 /* Keep these private until we decide it was a good idea */
548 #if defined(PERL_CORE) || defined(PERL_EXT) || defined(PERL_EXT_POSIX)
550 #define strBEGINs(s1,s2) (strncmp(s1,"" s2 "", sizeof(s2)-1) == 0)
552 #define memBEGINs(s1, l, s2) \
553 ( (Ptrdiff_t) (l) >= (Ptrdiff_t) sizeof(s2) - 1 \
554 && memEQ(s1, "" s2 "", sizeof(s2)-1))
555 #define memBEGINPs(s1, l, s2) \
556 ( (Ptrdiff_t) (l) > (Ptrdiff_t) sizeof(s2) - 1 \
557 && memEQ(s1, "" s2 "", sizeof(s2)-1))
558 #define memENDs(s1, l, s2) \
559 ( (Ptrdiff_t) (l) >= (Ptrdiff_t) sizeof(s2) - 1 \
560 && memEQ(s1 + (l) - (sizeof(s2) - 1), "" s2 "", sizeof(s2)-1))
561 #define memENDPs(s1, l, s2) \
562 ( (Ptrdiff_t) (l) > (Ptrdiff_t) sizeof(s2) \
563 && memEQ(s1 + (l) - (sizeof(s2) - 1), "" s2 "", sizeof(s2)-1))
564 #endif /* End of making macros private */
566 #define memLT(s1,s2,l) (memcmp(s1,s2,l) < 0)
567 #define memLE(s1,s2,l) (memcmp(s1,s2,l) <= 0)
568 #define memGT(s1,s2,l) (memcmp(s1,s2,l) > 0)
569 #define memGE(s1,s2,l) (memcmp(s1,s2,l) >= 0)
574 * Unfortunately, the introduction of locales means that we
575 * can't trust isupper(), etc. to tell the truth. And when
576 * it comes to /\w+/ with tainting enabled, we *must* be able
577 * to trust our character classes.
579 * Therefore, the default tests in the text of Perl will be
580 * independent of locale. Any code that wants to depend on
581 * the current locale will use the tests that begin with "lc".
584 #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
592 =head1 Character classification
593 This section is about functions (really macros) that classify characters
594 into types, such as punctuation versus alphabetic, etc. Most of these are
595 analogous to regular expression character classes. (See
596 L<perlrecharclass/POSIX Character Classes>.) There are several variants for
597 each class. (Not all macros have all variants; each item below lists the
598 ones valid for it.) None are affected by C<use bytes>, and only the ones
599 with C<LC> in the name are affected by the current locale.
601 The base function, e.g., C<isALPHA()>, takes any signed or unsigned value,
602 treating it as a code point, and returns a boolean as to whether or not the
603 character represented by it is (or on non-ASCII platforms, corresponds to) an
604 ASCII character in the named class based on platform, Unicode, and Perl rules.
605 If the input is a number that doesn't fit in an octet, FALSE is returned.
607 Variant C<isI<FOO>_A> (e.g., C<isALPHA_A()>) is identical to the base function
608 with no suffix C<"_A">. This variant is used to emphasize by its name that
609 only ASCII-range characters can return TRUE.
611 Variant C<isI<FOO>_L1> imposes the Latin-1 (or EBCDIC equivalent) character set
612 onto the platform. That is, the code points that are ASCII are unaffected,
613 since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
614 as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
615 true when called with the code point 0xDF, which is a word character in both
616 ASCII and EBCDIC (though it represents different characters in each).
617 If the input is a number that doesn't fit in an octet, FALSE is returned.
618 (Perl's documentation uses a colloquial definition of Latin-1, to include all
619 code points below 256.)
621 Variant C<isI<FOO>_uvchr> is exactly like the C<isI<FOO>_L1> variant, for
622 inputs below 256, but if the code point is larger than 255, Unicode rules are
623 used to determine if it is in the character class. For example,
624 C<isWORDCHAR_uvchr(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
625 WITH MACRON in Unicode, and is a word character.
627 Variants C<isI<FOO>_utf8> and C<isI<FOO>_utf8_safe> are like C<isI<FOO>_uvchr>,
628 but are used for UTF-8 encoded strings. The two forms are different names for
629 the same thing. Each call to one of these classifies the first character of
630 the string starting at C<p>. The second parameter, C<e>, points to anywhere in
631 the string beyond the first character, up to one byte past the end of the
632 entire string. Although both variants are identical, the suffix C<_safe> in
633 one name emphasizes that it will not attempt to read beyond S<C<e - 1>>,
634 provided that the constraint S<C<s E<lt> e>> is true (this is asserted for in
635 C<-DDEBUGGING> builds). If the UTF-8 for the input character is malformed in
636 some way, the program may croak, or the function may return FALSE, at the
637 discretion of the implementation, and subject to change in future releases.
639 Variant C<isI<FOO>_LC> is like the C<isI<FOO>_A> and C<isI<FOO>_L1> variants,
640 but the result is based on the current locale, which is what C<LC> in the name
641 stands for. If Perl can determine that the current locale is a UTF-8 locale,
642 it uses the published Unicode rules; otherwise, it uses the C library function
643 that gives the named classification. For example, C<isDIGIT_LC()> when not in
644 a UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always
645 returned if the input won't fit into an octet. On some platforms where the C
646 library function is known to be defective, Perl changes its result to follow
647 the POSIX standard's rules.
649 Variant C<isI<FOO>_LC_uvchr> acts exactly like C<isI<FOO>_LC> for inputs less
650 than 256, but for larger ones it returns the Unicode classification of the code
653 Variants C<isI<FOO>_LC_utf8> and C<isI<FOO>_LC_utf8_safe> are like
654 C<isI<FOO>_LC_uvchr>, but are used for UTF-8 encoded strings. The two forms
655 are different names for the same thing. Each call to one of these classifies
656 the first character of the string starting at C<p>. The second parameter,
657 C<e>, points to anywhere in the string beyond the first character, up to one
658 byte past the end of the entire string. Although both variants are identical,
659 the suffix C<_safe> in one name emphasizes that it will not attempt to read
660 beyond S<C<e - 1>>, provided that the constraint S<C<s E<lt> e>> is true (this
661 is asserted for in C<-DDEBUGGING> builds). If the UTF-8 for the input
662 character is malformed in some way, the program may croak, or the function may
663 return FALSE, at the discretion of the implementation, and subject to change in
666 =for apidoc Am|bool|isALPHA|int ch
667 Returns a boolean indicating whether the specified input is one of C<[A-Za-z]>,
668 analogous to C<m/[[:alpha:]]/>.
669 See the L<top of this section|/Character classification> for an explanation of
671 C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uvchr>, C<isALPHA_utf8>,
672 C<isALPHA_utf8_safe>, C<isALPHA_LC>, C<isALPHA_LC_uvchr>, C<isALPHA_LC_utf8>,
673 and C<isALPHA_LC_utf8_safe>.
677 Here and below, we add the protoypes of these macros for downstream programs
678 that would be interested in them, such as Devel::PPPort
680 =for apidoc Amh|bool|isALPHA_A|int ch
681 =for apidoc Amh|bool|isALPHA_L1|int ch
682 =for apidoc Amh|bool|isALPHA_uvchr|int ch
683 =for apidoc Amh|bool|isALPHA_utf8_safe|U8 * s|U8 * end
684 =for apidoc Amh|bool|isALPHA_utf8|U8 * s|U8 * end
685 =for apidoc Amh|bool|isALPHA_LC|int ch
686 =for apidoc Amh|bool|isALPHA_LC_uvchr|int ch
687 =for apidoc Amh|bool|isALPHA_LC_utf8_safe|U8 * s| U8 *end
689 =for apidoc Am|bool|isALPHANUMERIC|int ch
690 Returns a boolean indicating whether the specified character is one of
691 C<[A-Za-z0-9]>, analogous to C<m/[[:alnum:]]/>.
692 See the L<top of this section|/Character classification> for an explanation of
694 C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uvchr>,
695 C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_utf8_safe>, C<isALPHANUMERIC_LC>,
696 C<isALPHANUMERIC_LC_uvchr>, C<isALPHANUMERIC_LC_utf8>, and
697 C<isALPHANUMERIC_LC_utf8_safe>.
699 A (discouraged from use) synonym is C<isALNUMC> (where the C<C> suffix means
700 this corresponds to the C language alphanumeric definition). Also
701 there are the variants
702 C<isALNUMC_A>, C<isALNUMC_L1>
703 C<isALNUMC_LC>, and C<isALNUMC_LC_uvchr>.
705 =for apidoc Amh|bool|isALPHANUMERIC_A|int ch
706 =for apidoc Amh|bool|isALPHANUMERIC_L1|int ch
707 =for apidoc Amh|bool|isALPHANUMERIC_uvchr|int ch
708 =for apidoc Amh|bool|isALPHANUMERIC_utf8_safe|U8 * s|U8 * end
709 =for apidoc Amh|bool|isALPHANUMERIC_utf8|U8 * s|U8 * end
710 =for apidoc Amh|bool|isALPHANUMERIC_LC|int ch
711 =for apidoc Amh|bool|isALPHANUMERIC_LC_uvchr|int ch
712 =for apidoc Amh|bool|isALPHANUMERIC_LC_utf8_safe|U8 * s| U8 *end
713 =for apidoc Amh|bool|isALNUMC|int ch
714 =for apidoc Amh|bool|isALNUMC_A|int ch
715 =for apidoc Amh|bool|isALNUMC_L1|int ch
716 =for apidoc Amh|bool|isALNUMC_LC|int ch
717 =for apidoc Amh|bool|isALNUMC_LC_uvchr|int ch
719 =for apidoc Am|bool|isASCII|int ch
720 Returns a boolean indicating whether the specified character is one of the 128
721 characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
722 On non-ASCII platforms, it returns TRUE iff this
723 character corresponds to an ASCII character. Variants C<isASCII_A()> and
724 C<isASCII_L1()> are identical to C<isASCII()>.
725 See the L<top of this section|/Character classification> for an explanation of
727 C<isASCII_uvchr>, C<isASCII_utf8>, C<isASCII_utf8_safe>, C<isASCII_LC>,
728 C<isASCII_LC_uvchr>, C<isASCII_LC_utf8>, and C<isASCII_LC_utf8_safe>.
729 Note, however, that some platforms do not have the C library routine
730 C<isascii()>. In these cases, the variants whose names contain C<LC> are the
731 same as the corresponding ones without.
733 =for apidoc Amh|bool|isASCII_A|int ch
734 =for apidoc Amh|bool|isASCII_L1|int ch
735 =for apidoc Amh|bool|isASCII_uvchr|int ch
736 =for apidoc Amh|bool|isASCII_utf8_safe|U8 * s|U8 * end
737 =for apidoc Amh|bool|isASCII_utf8|U8 * s|U8 * end
738 =for apidoc Amh|bool|isASCII_LC|int ch
739 =for apidoc Amh|bool|isASCII_LC_uvchr|int ch
740 =for apidoc Amh|bool|isASCII_LC_utf8_safe|U8 * s| U8 *end
742 Also note, that because all ASCII characters are UTF-8 invariant (meaning they
743 have the exact same representation (always a single byte) whether encoded in
744 UTF-8 or not), C<isASCII> will give the correct results when called with any
745 byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> and
746 C<isASCII_utf8_safe> will work properly on any string encoded or not in UTF-8.
748 =for apidoc Am|bool|isBLANK|char ch
749 Returns a boolean indicating whether the specified character is a
750 character considered to be a blank, analogous to C<m/[[:blank:]]/>.
751 See the L<top of this section|/Character classification> for an explanation of
753 C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uvchr>, C<isBLANK_utf8>,
754 C<isBLANK_utf8_safe>, C<isBLANK_LC>, C<isBLANK_LC_uvchr>, C<isBLANK_LC_utf8>,
755 and C<isBLANK_LC_utf8_safe>. Note,
756 however, that some platforms do not have the C library routine
757 C<isblank()>. In these cases, the variants whose names contain C<LC> are
758 the same as the corresponding ones without.
760 =for apidoc Amh|bool|isBLANK_A|int ch
761 =for apidoc Amh|bool|isBLANK_L1|int ch
762 =for apidoc Amh|bool|isBLANK_uvchr|int ch
763 =for apidoc Amh|bool|isBLANK_utf8_safe|U8 * s|U8 * end
764 =for apidoc Amh|bool|isBLANK_utf8|U8 * s|U8 * end
765 =for apidoc Amh|bool|isBLANK_LC|int ch
766 =for apidoc Amh|bool|isBLANK_LC_uvchr|int ch
767 =for apidoc Amh|bool|isBLANK_LC_utf8_safe|U8 * s| U8 *end
769 =for apidoc Am|bool|isCNTRL|char ch
770 Returns a boolean indicating whether the specified character is a
771 control character, analogous to C<m/[[:cntrl:]]/>.
772 See the L<top of this section|/Character classification> for an explanation of
774 C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uvchr>, C<isCNTRL_utf8>,
775 C<isCNTRL_utf8_safe>, C<isCNTRL_LC>, C<isCNTRL_LC_uvchr>, C<isCNTRL_LC_utf8>
776 and C<isCNTRL_LC_utf8_safe>. On EBCDIC
777 platforms, you almost always want to use the C<isCNTRL_L1> variant.
779 =for apidoc Amh|bool|isCNTRL_A|int ch
780 =for apidoc Amh|bool|isCNTRL_L1|int ch
781 =for apidoc Amh|bool|isCNTRL_uvchr|int ch
782 =for apidoc Amh|bool|isCNTRL_utf8_safe|U8 * s|U8 * end
783 =for apidoc Amh|bool|isCNTRL_utf8|U8 * s|U8 * end
784 =for apidoc Amh|bool|isCNTRL_LC|int ch
785 =for apidoc Amh|bool|isCNTRL_LC_uvchr|int ch
786 =for apidoc Amh|bool|isCNTRL_LC_utf8_safe|U8 * s| U8 *end
788 =for apidoc Am|bool|isDIGIT|char ch
789 Returns a boolean indicating whether the specified character is a
790 digit, analogous to C<m/[[:digit:]]/>.
791 Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
792 See the L<top of this section|/Character classification> for an explanation of
794 C<isDIGIT_uvchr>, C<isDIGIT_utf8>, C<isDIGIT_utf8_safe>, C<isDIGIT_LC>,
795 C<isDIGIT_LC_uvchr>, C<isDIGIT_LC_utf8>, and C<isDIGIT_LC_utf8_safe>.
797 =for apidoc Amh|bool|isDIGIT_A|int ch
798 =for apidoc Amh|bool|isDIGIT_L1|int ch
799 =for apidoc Amh|bool|isDIGIT_uvchr|int ch
800 =for apidoc Amh|bool|isDIGIT_utf8_safe|U8 * s|U8 * end
801 =for apidoc Amh|bool|isDIGIT_utf8|U8 * s|U8 * end
802 =for apidoc Amh|bool|isDIGIT_LC|int ch
803 =for apidoc Amh|bool|isDIGIT_LC_uvchr|int ch
804 =for apidoc Amh|bool|isDIGIT_LC_utf8_safe|U8 * s| U8 *end
806 =for apidoc Am|bool|isGRAPH|char ch
807 Returns a boolean indicating whether the specified character is a
808 graphic character, analogous to C<m/[[:graph:]]/>.
809 See the L<top of this section|/Character classification> for an explanation of
810 variants C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uvchr>, C<isGRAPH_utf8>,
811 C<isGRAPH_utf8_safe>, C<isGRAPH_LC>, C<isGRAPH_LC_uvchr>,
812 C<isGRAPH_LC_utf8_safe>, and C<isGRAPH_LC_utf8_safe>.
814 =for apidoc Amh|bool|isGRAPH_A|int ch
815 =for apidoc Amh|bool|isGRAPH_L1|int ch
816 =for apidoc Amh|bool|isGRAPH_uvchr|int ch
817 =for apidoc Amh|bool|isGRAPH_utf8_safe|U8 * s|U8 * end
818 =for apidoc Amh|bool|isGRAPH_utf8|U8 * s|U8 * end
819 =for apidoc Amh|bool|isGRAPH_LC|int ch
820 =for apidoc Amh|bool|isGRAPH_LC_uvchr|int ch
821 =for apidoc Amh|bool|isGRAPH_LC_utf8_safe|U8 * s| U8 *end
823 =for apidoc Am|bool|isLOWER|char ch
824 Returns a boolean indicating whether the specified character is a
825 lowercase character, analogous to C<m/[[:lower:]]/>.
826 See the L<top of this section|/Character classification> for an explanation of
828 C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uvchr>, C<isLOWER_utf8>,
829 C<isLOWER_utf8_safe>, C<isLOWER_LC>, C<isLOWER_LC_uvchr>, C<isLOWER_LC_utf8>,
830 and C<isLOWER_LC_utf8_safe>.
832 =for apidoc Amh|bool|isLOWER_A|int ch
833 =for apidoc Amh|bool|isLOWER_L1|int ch
834 =for apidoc Amh|bool|isLOWER_uvchr|int ch
835 =for apidoc Amh|bool|isLOWER_utf8_safe|U8 * s|U8 * end
836 =for apidoc Amh|bool|isLOWER_utf8|U8 * s|U8 * end
837 =for apidoc Amh|bool|isLOWER_LC|int ch
838 =for apidoc Amh|bool|isLOWER_LC_uvchr|int ch
839 =for apidoc Amh|bool|isLOWER_LC_utf8_safe|U8 * s| U8 *end
841 =for apidoc Am|bool|isOCTAL|char ch
842 Returns a boolean indicating whether the specified character is an
844 The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
847 =for apidoc Amh|bool|isOCTAL_A|int ch
848 =for apidoc Amh|bool|isOCTAL_L1|int ch
850 =for apidoc Am|bool|isPUNCT|char ch
851 Returns a boolean indicating whether the specified character is a
852 punctuation character, analogous to C<m/[[:punct:]]/>.
853 Note that the definition of what is punctuation isn't as
854 straightforward as one might desire. See L<perlrecharclass/POSIX Character
855 Classes> for details.
856 See the L<top of this section|/Character classification> for an explanation of
857 variants C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uvchr>, C<isPUNCT_utf8>,
858 C<isPUNCT_utf8_safe>, C<isPUNCT_LC>, C<isPUNCT_LC_uvchr>, C<isPUNCT_LC_utf8>,
859 and C<isPUNCT_LC_utf8_safe>.
861 =for apidoc Amh|bool|isPUNCT_A|int ch
862 =for apidoc Amh|bool|isPUNCT_L1|int ch
863 =for apidoc Amh|bool|isPUNCT_uvchr|int ch
864 =for apidoc Amh|bool|isPUNCT_utf8_safe|U8 * s|U8 * end
865 =for apidoc Amh|bool|isPUNCT_utf8|U8 * s|U8 * end
866 =for apidoc Amh|bool|isPUNCT_LC|int ch
867 =for apidoc Amh|bool|isPUNCT_LC_uvchr|int ch
868 =for apidoc Amh|bool|isPUNCT_LC_utf8_safe|U8 * s| U8 *end
870 =for apidoc Am|bool|isSPACE|char ch
871 Returns a boolean indicating whether the specified character is a
872 whitespace character. This is analogous
873 to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
874 this also matches what C<m/[[:space:]]/> does. Prior to 5.18, only the
875 locale forms of this macro (the ones with C<LC> in their names) matched
876 precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
877 in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
878 (See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
879 See the L<top of this section|/Character classification> for an explanation of
881 C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uvchr>, C<isSPACE_utf8>,
882 C<isSPACE_utf8_safe>, C<isSPACE_LC>, C<isSPACE_LC_uvchr>, C<isSPACE_LC_utf8>,
883 and C<isSPACE_LC_utf8_safe>.
885 =for apidoc Amh|bool|isSPACE_A|int ch
886 =for apidoc Amh|bool|isSPACE_L1|int ch
887 =for apidoc Amh|bool|isSPACE_uvchr|int ch
888 =for apidoc Amh|bool|isSPACE_utf8_safe|U8 * s|U8 * end
889 =for apidoc Amh|bool|isSPACE_utf8|U8 * s|U8 * end
890 =for apidoc Amh|bool|isSPACE_LC|int ch
891 =for apidoc Amh|bool|isSPACE_LC_uvchr|int ch
892 =for apidoc Amh|bool|isSPACE_LC_utf8_safe|U8 * s| U8 *end
894 =for apidoc Am|bool|isPSXSPC|char ch
895 (short for Posix Space)
896 Starting in 5.18, this is identical in all its forms to the
897 corresponding C<isSPACE()> macros.
898 The locale forms of this macro are identical to their corresponding
899 C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
900 non-locale forms differ from their C<isSPACE()> forms only in that the
901 C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
902 Otherwise they are identical. Thus this macro is analogous to what
903 C<m/[[:space:]]/> matches in a regular expression.
904 See the L<top of this section|/Character classification> for an explanation of
905 variants C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uvchr>, C<isPSXSPC_utf8>,
906 C<isPSXSPC_utf8_safe>, C<isPSXSPC_LC>, C<isPSXSPC_LC_uvchr>,
907 C<isPSXSPC_LC_utf8>, and C<isPSXSPC_LC_utf8_safe>.
909 =for apidoc Amh|bool|isPSXSPC_A|int ch
910 =for apidoc Amh|bool|isPSXSPC_L1|int ch
911 =for apidoc Amh|bool|isPSXSPC_uvchr|int ch
912 =for apidoc Amh|bool|isPSXSPC_utf8_safe|U8 * s|U8 * end
913 =for apidoc Amh|bool|isPSXSPC_utf8|U8 * s|U8 * end
914 =for apidoc Amh|bool|isPSXSPC_LC|int ch
915 =for apidoc Amh|bool|isPSXSPC_LC_uvchr|int ch
916 =for apidoc Amh|bool|isPSXSPC_LC_utf8_safe|U8 * s| U8 *end
918 =for apidoc Am|bool|isUPPER|char ch
919 Returns a boolean indicating whether the specified character is an
920 uppercase character, analogous to C<m/[[:upper:]]/>.
921 See the L<top of this section|/Character classification> for an explanation of
922 variants C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uvchr>, C<isUPPER_utf8>,
923 C<isUPPER_utf8_safe>, C<isUPPER_LC>, C<isUPPER_LC_uvchr>, C<isUPPER_LC_utf8>,
924 and C<isUPPER_LC_utf8_safe>.
926 =for apidoc Amh|bool|isUPPER_A|int ch
927 =for apidoc Amh|bool|isUPPER_L1|int ch
928 =for apidoc Amh|bool|isUPPER_uvchr|int ch
929 =for apidoc Amh|bool|isUPPER_utf8_safe|U8 * s|U8 * end
930 =for apidoc Amh|bool|isUPPER_utf8|U8 * s|U8 * end
931 =for apidoc Amh|bool|isUPPER_LC|int ch
932 =for apidoc Amh|bool|isUPPER_LC_uvchr|int ch
933 =for apidoc Amh|bool|isUPPER_LC_utf8_safe|U8 * s| U8 *end
935 =for apidoc Am|bool|isPRINT|char ch
936 Returns a boolean indicating whether the specified character is a
937 printable character, analogous to C<m/[[:print:]]/>.
938 See the L<top of this section|/Character classification> for an explanation of
940 C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uvchr>, C<isPRINT_utf8>,
941 C<isPRINT_utf8_safe>, C<isPRINT_LC>, C<isPRINT_LC_uvchr>, C<isPRINT_LC_utf8>,
942 and C<isPRINT_LC_utf8_safe>.
944 =for apidoc Amh|bool|isPRINT_A|int ch
945 =for apidoc Amh|bool|isPRINT_L1|int ch
946 =for apidoc Amh|bool|isPRINT_uvchr|int ch
947 =for apidoc Amh|bool|isPRINT_utf8_safe|U8 * s|U8 * end
948 =for apidoc Amh|bool|isPRINT_utf8|U8 * s|U8 * end
949 =for apidoc Amh|bool|isPRINT_LC|int ch
950 =for apidoc Amh|bool|isPRINT_LC_uvchr|int ch
951 =for apidoc Amh|bool|isPRINT_LC_utf8_safe|U8 * s| U8 *end
953 =for apidoc Am|bool|isWORDCHAR|char ch
954 Returns a boolean indicating whether the specified character is a character
955 that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
956 in a regular expression. A word character is an alphabetic character, a
957 decimal digit, a connecting punctuation character (such as an underscore), or
958 a "mark" character that attaches to one of those (like some sort of accent).
959 C<isALNUM()> is a synonym provided for backward compatibility, even though a
960 word character includes more than the standard C language meaning of
962 See the L<top of this section|/Character classification> for an explanation of
963 variants C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uvchr>,
964 C<isWORDCHAR_utf8>, and C<isWORDCHAR_utf8_safe>. C<isWORDCHAR_LC>,
965 C<isWORDCHAR_LC_uvchr>, C<isWORDCHAR_LC_utf8>, and C<isWORDCHAR_LC_utf8_safe>
966 are also as described there, but additionally include the platform's native
969 =for apidoc Amh|bool|isWORDCHAR_A|int ch
970 =for apidoc Amh|bool|isWORDCHAR_L1|int ch
971 =for apidoc Amh|bool|isWORDCHAR_uvchr|int ch
972 =for apidoc Amh|bool|isWORDCHAR_utf8_safe|U8 * s|U8 * end
973 =for apidoc Amh|bool|isWORDCHAR_utf8|U8 * s|U8 * end
974 =for apidoc Amh|bool|isWORDCHAR_LC|int ch
975 =for apidoc Amh|bool|isWORDCHAR_LC_uvchr|int ch
976 =for apidoc Amh|bool|isWORDCHAR_LC_utf8_safe|U8 * s| U8 *end
977 =for apidoc Amh|bool|isALNUM|int ch
978 =for apidoc Amh|bool|isALNUM_A|int ch
979 =for apidoc Amh|bool|isALNUM_LC|int ch
980 =for apidoc Amh|bool|isALNUM_LC_uvchr|int ch
982 =for apidoc Am|bool|isXDIGIT|char ch
983 Returns a boolean indicating whether the specified character is a hexadecimal
984 digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
985 and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
986 See the L<top of this section|/Character classification> for an explanation of
988 C<isXDIGIT_uvchr>, C<isXDIGIT_utf8>, C<isXDIGIT_utf8_safe>, C<isXDIGIT_LC>,
989 C<isXDIGIT_LC_uvchr>, C<isXDIGIT_LC_utf8>, and C<isXDIGIT_LC_utf8_safe>.
991 =for apidoc Amh|bool|isXDIGIT_A|int ch
992 =for apidoc Amh|bool|isXDIGIT_L1|int ch
993 =for apidoc Amh|bool|isXDIGIT_uvchr|int ch
994 =for apidoc Amh|bool|isXDIGIT_utf8_safe|U8 * s|U8 * end
995 =for apidoc Amh|bool|isXDIGIT_utf8|U8 * s|U8 * end
996 =for apidoc Amh|bool|isXDIGIT_LC|int ch
997 =for apidoc Amh|bool|isXDIGIT_LC_uvchr|int ch
998 =for apidoc Amh|bool|isXDIGIT_LC_utf8_safe|U8 * s| U8 *end
1000 =for apidoc Am|bool|isIDFIRST|char ch
1001 Returns a boolean indicating whether the specified character can be the first
1002 character of an identifier. This is very close to, but not quite the same as
1003 the official Unicode property C<XID_Start>. The difference is that this
1004 returns true only if the input character also matches L</isWORDCHAR>.
1005 See the L<top of this section|/Character classification> for an explanation of
1007 C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uvchr>, C<isIDFIRST_utf8>,
1008 C<isIDFIRST_utf8_safe>, C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>,
1009 C<isIDFIRST_LC_utf8>, and C<isIDFIRST_LC_utf8_safe>.
1011 =for apidoc Amh|bool|isIDFIRST_A|int ch
1012 =for apidoc Amh|bool|isIDFIRST_L1|int ch
1013 =for apidoc Amh|bool|isIDFIRST_uvchr|int ch
1014 =for apidoc Amh|bool|isIDFIRST_utf8_safe|U8 * s|U8 * end
1015 =for apidoc Amh|bool|isIDFIRST_utf8|U8 * s|U8 * end
1016 =for apidoc Amh|bool|isIDFIRST_LC|int ch
1017 =for apidoc Amh|bool|isIDFIRST_LC_uvchr|int ch
1018 =for apidoc Amh|bool|isIDFIRST_LC_utf8_safe|U8 * s| U8 *end
1020 =for apidoc Am|bool|isIDCONT|char ch
1021 Returns a boolean indicating whether the specified character can be the
1022 second or succeeding character of an identifier. This is very close to, but
1023 not quite the same as the official Unicode property C<XID_Continue>. The
1024 difference is that this returns true only if the input character also matches
1025 L</isWORDCHAR>. See the L<top of this section|/Character classification> for
1026 an explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uvchr>,
1027 C<isIDCONT_utf8>, C<isIDCONT_utf8_safe>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>,
1028 C<isIDCONT_LC_utf8>, and C<isIDCONT_LC_utf8_safe>.
1030 =for apidoc Amh|bool|isIDCONT_A|int ch
1031 =for apidoc Amh|bool|isIDCONT_L1|int ch
1032 =for apidoc Amh|bool|isIDCONT_uvchr|int ch
1033 =for apidoc Amh|bool|isIDCONT_utf8_safe|U8 * s|U8 * end
1034 =for apidoc Amh|bool|isIDCONT_utf8|U8 * s|U8 * end
1035 =for apidoc Amh|bool|isIDCONT_LC|int ch
1036 =for apidoc Amh|bool|isIDCONT_LC_uvchr|int ch
1037 =for apidoc Amh|bool|isIDCONT_LC_utf8_safe|U8 * s| U8 *end
1039 =head1 Miscellaneous Functions
1041 =for apidoc Am|U8|READ_XDIGIT|char str*
1042 Returns the value of an ASCII-range hex digit and advances the string pointer.
1043 Behaviour is only well defined when isXDIGIT(*str) is true.
1045 =head1 Character case changing
1046 Perl uses "full" Unicode case mappings. This means that converting a single
1047 character to another case may result in a sequence of more than one character.
1048 For example, the uppercase of C<E<223>> (LATIN SMALL LETTER SHARP S) is the two
1049 character sequence C<SS>. This presents some complications The lowercase of
1050 all characters in the range 0..255 is a single character, and thus
1051 C<L</toLOWER_L1>> is furnished. But, C<toUPPER_L1> can't exist, as it couldn't
1052 return a valid result for all legal inputs. Instead C<L</toUPPER_uvchr>> has
1053 an API that does allow every possible legal result to be returned.) Likewise
1054 no other function that is crippled by not being able to give the correct
1055 results for the full range of possible inputs has been implemented here.
1057 =for apidoc Am|U8|toUPPER|int ch
1058 Converts the specified character to uppercase. If the input is anything but an
1059 ASCII lowercase character, that input character itself is returned. Variant
1060 C<toUPPER_A> is equivalent.
1062 =for apidoc Am|UV|toUPPER_uvchr|UV cp|U8* s|STRLEN* lenp
1063 Converts the code point C<cp> to its uppercase version, and
1064 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code
1065 point is interpreted as native if less than 256; otherwise as Unicode. Note
1066 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
1067 bytes since the uppercase version may be longer than the original character.
1069 The first code point of the uppercased version is returned
1070 (but note, as explained at L<the top of this section|/Character case
1071 changing>, that there may be more.)
1073 =for apidoc Am|UV|toUPPER_utf8|U8* p|U8* e|U8* s|STRLEN* lenp
1074 Converts the first UTF-8 encoded character in the sequence starting at C<p> and
1075 extending no further than S<C<e - 1>> to its uppercase version, and
1076 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
1077 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
1078 bytes since the uppercase version may be longer than the original character.
1080 The first code point of the uppercased version is returned
1081 (but note, as explained at L<the top of this section|/Character case
1082 changing>, that there may be more).
1084 It will not attempt to read beyond S<C<e - 1>>, provided that the constraint
1085 S<C<s E<lt> e>> is true (this is asserted for in C<-DDEBUGGING> builds). If
1086 the UTF-8 for the input character is malformed in some way, the program may
1087 croak, or the function may return the REPLACEMENT CHARACTER, at the discretion
1088 of the implementation, and subject to change in future releases.
1090 =for apidoc Am|UV|toUPPER_utf8_safe|U8* p|U8* e|U8* s|STRLEN* lenp
1091 Same as L</toUPPER_utf8>.
1093 =for apidoc Am|U8|toFOLD|U8 ch
1094 Converts the specified character to foldcase. If the input is anything but an
1095 ASCII uppercase character, that input character itself is returned. Variant
1096 C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
1097 Latin1 range, as the full generality of L</toFOLD_uvchr> is needed there.)
1099 =for apidoc Am|UV|toFOLD_uvchr|UV cp|U8* s|STRLEN* lenp
1100 Converts the code point C<cp> to its foldcase version, and
1101 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code
1102 point is interpreted as native if less than 256; otherwise as Unicode. Note
1103 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
1104 bytes since the foldcase version may be longer than the original character.
1106 The first code point of the foldcased version is returned
1107 (but note, as explained at L<the top of this section|/Character case
1108 changing>, that there may be more).
1110 =for apidoc Am|UV|toFOLD_utf8|U8* p|U8* e|U8* s|STRLEN* lenp
1111 Converts the first UTF-8 encoded character in the sequence starting at C<p> and
1112 extending no further than S<C<e - 1>> to its foldcase version, and
1113 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
1114 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
1115 bytes since the foldcase version may be longer than the original character.
1117 The first code point of the foldcased version is returned
1118 (but note, as explained at L<the top of this section|/Character case
1119 changing>, that there may be more).
1122 to read beyond S<C<e - 1>>, provided that the constraint S<C<s E<lt> e>> is
1123 true (this is asserted for in C<-DDEBUGGING> builds). If the UTF-8 for the
1124 input character is malformed in some way, the program may croak, or the
1125 function may return the REPLACEMENT CHARACTER, at the discretion of the
1126 implementation, and subject to change in future releases.
1128 =for apidoc Am|UV|toFOLD_utf8_safe|U8* p|U8* e|U8* s|STRLEN* lenp
1129 Same as L</toFOLD_utf8>.
1131 =for apidoc Am|U8|toLOWER|U8 ch
1132 Converts the specified character to lowercase. If the input is anything but an
1133 ASCII uppercase character, that input character itself is returned. Variant
1134 C<toLOWER_A> is equivalent.
1136 =for apidoc Am|U8|toLOWER_L1|U8 ch
1137 Converts the specified Latin1 character to lowercase. The results are
1138 undefined if the input doesn't fit in a byte.
1140 =for apidoc Am|U8|toLOWER_LC|U8 ch
1141 Converts the specified character to lowercase using the current locale's rules,
1142 if possible; otherwise returns the input character itself.
1144 =for apidoc Am|UV|toLOWER_uvchr|UV cp|U8* s|STRLEN* lenp
1145 Converts the code point C<cp> to its lowercase version, and
1146 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code
1147 point is interpreted as native if less than 256; otherwise as Unicode. Note
1148 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
1149 bytes since the lowercase version may be longer than the original character.
1151 The first code point of the lowercased version is returned
1152 (but note, as explained at L<the top of this section|/Character case
1153 changing>, that there may be more).
1155 =for apidoc Am|UV|toLOWER_utf8|U8* p|U8* e|U8* s|STRLEN* lenp
1156 Converts the first UTF-8 encoded character in the sequence starting at C<p> and
1157 extending no further than S<C<e - 1>> to its lowercase version, and
1158 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
1159 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
1160 bytes since the lowercase version may be longer than the original character.
1162 The first code point of the lowercased version is returned
1163 (but note, as explained at L<the top of this section|/Character case
1164 changing>, that there may be more).
1165 It will not attempt to read beyond S<C<e - 1>>, provided that the constraint
1166 S<C<s E<lt> e>> is true (this is asserted for in C<-DDEBUGGING> builds). If
1167 the UTF-8 for the input character is malformed in some way, the program may
1168 croak, or the function may return the REPLACEMENT CHARACTER, at the discretion
1169 of the implementation, and subject to change in future releases.
1171 =for apidoc Am|UV|toLOWER_utf8_safe|U8* p|U8* e|U8* s|STRLEN* lenp
1172 Same as L</toLOWER_utf8>.
1174 =for apidoc Am|U8|toTITLE|U8 ch
1175 Converts the specified character to titlecase. If the input is anything but an
1176 ASCII lowercase character, that input character itself is returned. Variant
1177 C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1
1178 range, as the full generality of L</toTITLE_uvchr> is needed there. Titlecase is
1179 not a concept used in locale handling, so there is no functionality for that.)
1181 =for apidoc Am|UV|toTITLE_uvchr|UV cp|U8* s|STRLEN* lenp
1182 Converts the code point C<cp> to its titlecase version, and
1183 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code
1184 point is interpreted as native if less than 256; otherwise as Unicode. Note
1185 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
1186 bytes since the titlecase version may be longer than the original character.
1188 The first code point of the titlecased version is returned
1189 (but note, as explained at L<the top of this section|/Character case
1190 changing>, that there may be more).
1192 =for apidoc Am|UV|toTITLE_utf8|U8* p|U8* e|U8* s|STRLEN* lenp
1193 Converts the first UTF-8 encoded character in the sequence starting at C<p> and
1194 extending no further than S<C<e - 1>> to its titlecase version, and
1195 stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
1196 that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
1197 bytes since the titlecase version may be longer than the original character.
1199 The first code point of the titlecased version is returned
1200 (but note, as explained at L<the top of this section|/Character case
1201 changing>, that there may be more).
1204 to read beyond S<C<e - 1>>, provided that the constraint S<C<s E<lt> e>> is
1205 true (this is asserted for in C<-DDEBUGGING> builds). If the UTF-8 for the
1206 input character is malformed in some way, the program may croak, or the
1207 function may return the REPLACEMENT CHARACTER, at the discretion of the
1208 implementation, and subject to change in future releases.
1210 =for apidoc Am|UV|toTITLE_utf8_safe|U8* p|U8* e|U8* s|STRLEN* lenp
1211 Same as L</toTITLE_utf8>.
1215 XXX Still undocumented isVERTWS_uvchr and _utf8; it's unclear what their names
1216 really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change,
1217 and aren't general purpose as they don't work on U+DF, and assert against that.
1219 Note that these macros are repeated in Devel::PPPort, so should also be
1220 patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
1225 void below because that's the best fit, and works for Devel::PPPort
1226 =for apidoc AmnU|void|WIDEST_UTYPE
1228 Yields the widest unsigned integer type on the platform, currently either
1229 C<U32> or C<64>. This can be used in declarations such as
1235 my_uv = (WIDEST_UTYPE) val;
1241 # define WIDEST_UTYPE U64
1243 # define WIDEST_UTYPE U32
1246 /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
1247 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
1248 * bits of information are lost even on a 64-bit machine, but to get the
1249 * compiler to optimize it out if possible. This is because Configure makes
1250 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
1251 * sizeof() guarantees that this evaluates to a constant true at compile time.
1253 * For Coverity, be always true, because otherwise Coverity thinks
1254 * it finds several expressions that are always true, independent
1255 * of operands. Well, they are, but that is kind of the point.
1257 #ifndef __COVERITY__
1258 /* The '| 0' part ensures a compiler error if c is not integer (like e.g., a
1260 #define FITS_IN_8_BITS(c) ( (sizeof(c) == 1) \
1261 || !(((WIDEST_UTYPE)((c) | 0)) & ~0xFF))
1263 #define FITS_IN_8_BITS(c) (1)
1266 /* Returns true if l <= c <= (l + n), where 'l' and 'n' are non-negative
1267 * Written this way so that after optimization, only one conditional test is
1268 * needed. (The NV casts stop any warnings about comparison always being true
1269 * if called with an unsigned. The cast preserves the sign, which is all we
1271 #define withinCOUNT(c, l, n) (__ASSERT_((NV) (l) >= 0) \
1272 __ASSERT_((NV) (n) >= 0) \
1273 (((WIDEST_UTYPE) (((c)) - ((l) | 0))) <= (((WIDEST_UTYPE) ((n) | 0)))))
1275 /* Returns true if c is in the range l..u, where 'l' is non-negative
1276 * Written this way so that after optimization, only one conditional test is
1278 #define inRANGE(c, l, u) (__ASSERT_((u) >= (l)) \
1279 ( (sizeof(c) == sizeof(U8)) ? withinCOUNT(((U8) (c)), (l), ((u) - (l))) \
1280 : (sizeof(c) == sizeof(U32)) ? withinCOUNT(((U32) (c)), (l), ((u) - (l))) \
1281 : (__ASSERT_(sizeof(c) == sizeof(WIDEST_UTYPE)) \
1282 withinCOUNT(((WIDEST_UTYPE) (c)), (l), ((u) - (l))))))
1285 # ifndef _ALL_SOURCE
1286 /* The native libc isascii() et.al. functions return the wrong results
1287 * on at least z/OS unless this is defined. */
1288 # error _ALL_SOURCE should probably be defined
1291 /* There is a simple definition of ASCII for ASCII platforms. But the
1292 * EBCDIC one isn't so simple, so is defined using table look-up like the
1293 * other macros below.
1295 * The cast here is used instead of '(c) >= 0', because some compilers emit
1296 * a warning that that test is always true when the parameter is an
1297 * unsigned type. khw supposes that it could be written as
1298 * && ((c) == '\0' || (c) > 0)
1299 * to avoid the message, but the cast will likely avoid extra branches even
1300 * with stupid compilers.
1302 * The '| 0' part ensures a compiler error if c is not integer (like e.g.,
1304 # define isASCII(c) ((WIDEST_UTYPE)((c) | 0) < 128)
1307 /* Take the eight possible bit patterns of the lower 3 bits and you get the
1308 * lower 3 bits of the 8 octal digits, in both ASCII and EBCDIC, so those bits
1309 * can be ignored. If the rest match '0', we have an octal */
1310 #define isOCTAL_A(c) (((WIDEST_UTYPE)((c) | 0) & ~7) == '0')
1312 #ifdef H_PERL /* If have access to perl.h, lookup in its table */
1314 /* Character class numbers. For internal core Perl use only. The ones less
1315 * than 32 are used in PL_charclass[] and the ones up through the one that
1316 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
1317 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
1318 * their actual definitions are here. If that file has a name not used here,
1321 * The first group of these is ordered in what I (khw) estimate to be the
1322 * frequency of their use. This gives a slight edge to exiting a loop earlier
1323 * (in reginclass() in regexec.c). Except \v should be last, as it isn't a
1324 * real Posix character class, and some (small) inefficiencies in regular
1325 * expression handling would be introduced by putting it in the middle of those
1326 * that are. Also, cntrl and ascii come after the others as it may be useful
1327 * to group these which have no members that match above Latin1, (or above
1328 * ASCII in the latter case) */
1330 # define _CC_WORDCHAR 0 /* \w and [:word:] */
1331 # define _CC_DIGIT 1 /* \d and [:digit:] */
1332 # define _CC_ALPHA 2 /* [:alpha:] */
1333 # define _CC_LOWER 3 /* [:lower:] */
1334 # define _CC_UPPER 4 /* [:upper:] */
1335 # define _CC_PUNCT 5 /* [:punct:] */
1336 # define _CC_PRINT 6 /* [:print:] */
1337 # define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1338 # define _CC_GRAPH 8 /* [:graph:] */
1339 # define _CC_CASED 9 /* [:lower:] or [:upper:] under /i */
1340 # define _CC_SPACE 10 /* \s, [:space:] */
1341 # define _CC_BLANK 11 /* [:blank:] */
1342 # define _CC_XDIGIT 12 /* [:xdigit:] */
1343 # define _CC_CNTRL 13 /* [:cntrl:] */
1344 # define _CC_ASCII 14 /* [:ascii:] */
1345 # define _CC_VERTSPACE 15 /* \v */
1347 # define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
1349 /* The members of the third group below do not need to be coordinated with data
1350 * structures in regcomp.[ch] and regexec.c. */
1351 # define _CC_IDFIRST 16
1352 # define _CC_CHARNAME_CONT 17
1353 # define _CC_NONLATIN1_FOLD 18
1354 # define _CC_NONLATIN1_SIMPLE_FOLD 19
1355 # define _CC_QUOTEMETA 20
1356 # define _CC_NON_FINAL_FOLD 21
1357 # define _CC_IS_IN_SOME_FOLD 22
1358 # define _CC_MNEMONIC_CNTRL 23
1360 /* This next group is only used on EBCDIC platforms, so theoretically could be
1361 * shared with something entirely different that's only on ASCII platforms */
1362 # define _CC_UTF8_START_BYTE_IS_FOR_AT_LEAST_SURROGATE 31
1364 * If more bits are needed, one could add a second word for non-64bit
1365 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
1366 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
1367 * is used only for optimization (as of this writing), and differs in the
1368 * Latin1 range from the ALPHA bit only in two relatively unimportant
1369 * characters: the masculine and feminine ordinal indicators, so removing it
1370 * would just cause /i regexes which match them to run less efficiently.
1371 * Similarly the EBCDIC-only bits are used just for speed, and could be
1372 * replaced by other means */
1374 #if defined(PERL_CORE) || defined(PERL_EXT)
1375 /* An enum version of the character class numbers, to help compilers
1378 _CC_ENUM_ALPHA = _CC_ALPHA,
1379 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
1380 _CC_ENUM_ASCII = _CC_ASCII,
1381 _CC_ENUM_BLANK = _CC_BLANK,
1382 _CC_ENUM_CASED = _CC_CASED,
1383 _CC_ENUM_CNTRL = _CC_CNTRL,
1384 _CC_ENUM_DIGIT = _CC_DIGIT,
1385 _CC_ENUM_GRAPH = _CC_GRAPH,
1386 _CC_ENUM_LOWER = _CC_LOWER,
1387 _CC_ENUM_PRINT = _CC_PRINT,
1388 _CC_ENUM_PUNCT = _CC_PUNCT,
1389 _CC_ENUM_SPACE = _CC_SPACE,
1390 _CC_ENUM_UPPER = _CC_UPPER,
1391 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
1392 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
1393 _CC_ENUM_XDIGIT = _CC_XDIGIT
1394 } _char_class_number;
1397 #define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
1401 EXTCONST U32 PL_charclass[] = {
1402 # include "l1_char_class_tab.h"
1405 # else /* ! DOINIT */
1406 EXTCONST U32 PL_charclass[];
1410 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
1411 # define _CC_mask(classnum) (1U << (classnum))
1413 /* For internal core Perl use only: the base macro for defining macros like
1415 # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
1416 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
1418 /* The mask for the _A versions of the macros; it just adds in the bit for
1420 # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1422 /* For internal core Perl use only: the base macro for defining macros like
1423 * isALPHA_A. The foo_A version makes sure that both the desired bit and
1424 * the ASCII bit are present */
1425 # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
1426 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
1427 == _CC_mask_A(classnum)))
1429 /* On ASCII platforms certain classes form a single range. It's faster to
1430 * special case these. isDIGIT is a single range on all platforms */
1432 # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
1433 # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1434 # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1435 # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1436 # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1438 /* By folding the upper and lowercase, we can use a single range */
1439 # define isALPHA_A(c) inRANGE((~('A' ^ 'a') & (c)), 'A', 'Z')
1440 # define isGRAPH_A(c) inRANGE(c, ' ' + 1, 0x7e)
1441 # define isLOWER_A(c) inRANGE(c, 'a', 'z')
1442 # define isPRINT_A(c) inRANGE(c, ' ', 0x7e)
1443 # define isUPPER_A(c) inRANGE(c, 'A', 'Z')
1445 # define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
1446 # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1447 # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1448 # define isDIGIT_A(c) inRANGE(c, '0', '9')
1449 # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1450 # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1451 # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1452 # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT) /* No non-ASCII xdigits
1454 # define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
1455 # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1456 # define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1457 # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1459 /* continuation character for legal NAME in \N{NAME} */
1460 # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1462 # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1463 # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1464 # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1465 # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1466 # define isPSXSPC_L1(c) isSPACE_L1(c)
1467 # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1468 # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1469 # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1470 # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1471 # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
1474 # define isASCII(c) _generic_isCC(c, _CC_ASCII)
1477 /* Participates in a single-character fold with a character above 255 */
1478 # 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)))
1480 /* Like the above, but also can be part of a multi-char fold */
1481 # 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)))
1483 # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
1484 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1485 _generic_isCC(c, _CC_NON_FINAL_FOLD)
1486 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1487 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
1488 # define _IS_MNEMONIC_CNTRL_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1489 _generic_isCC(c, _CC_MNEMONIC_CNTRL)
1490 #else /* else we don't have perl.h H_PERL */
1492 /* If we don't have perl.h, we are compiling a utility program. Below we
1493 * hard-code various macro definitions that wouldn't otherwise be available
1494 * to it. Most are coded based on first principles. These are written to
1495 * avoid EBCDIC vs. ASCII #ifdef's as much as possible. */
1496 # define isDIGIT_A(c) inRANGE(c, '0', '9')
1497 # define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1498 # define isSPACE_A(c) (isBLANK_A(c) \
1503 /* On EBCDIC, there are gaps between 'i' and 'j'; 'r' and 's'. Same for
1504 * uppercase. The tests for those aren't necessary on ASCII, but hurt only
1505 * performance (if optimization isn't on), and allow the same code to be
1506 * used for both platform types */
1507 # define isLOWER_A(c) inRANGE((c), 'a', 'i') \
1508 || inRANGE((c), 'j', 'r') \
1509 || inRANGE((c), 's', 'z')
1510 # define isUPPER_A(c) inRANGE((c), 'A', 'I') \
1511 || inRANGE((c), 'J', 'R') \
1512 || inRANGE((c), 'S', 'Z')
1513 # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
1514 # define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
1515 # define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1516 # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1517 # define isXDIGIT_A(c) ( isDIGIT_A(c) \
1518 || inRANGE((c), 'a', 'f') \
1519 || inRANGE((c), 'A', 'F')
1520 # define isPUNCT_A(c) ((c) == '-' || (c) == '!' || (c) == '"' \
1521 || (c) == '#' || (c) == '$' || (c) == '%' \
1522 || (c) == '&' || (c) == '\'' || (c) == '(' \
1523 || (c) == ')' || (c) == '*' || (c) == '+' \
1524 || (c) == ',' || (c) == '.' || (c) == '/' \
1525 || (c) == ':' || (c) == ';' || (c) == '<' \
1526 || (c) == '=' || (c) == '>' || (c) == '?' \
1527 || (c) == '@' || (c) == '[' || (c) == '\\' \
1528 || (c) == ']' || (c) == '^' || (c) == '_' \
1529 || (c) == '`' || (c) == '{' || (c) == '|' \
1530 || (c) == '}' || (c) == '~')
1531 # define isGRAPH_A(c) (isALPHANUMERIC_A(c) || isPUNCT_A(c))
1532 # define isPRINT_A(c) (isGRAPH_A(c) || (c) == ' ')
1535 /* The below is accurate for the 3 EBCDIC code pages traditionally
1536 * supported by perl. The only difference between them in the controls
1537 * is the position of \n, and that is represented symbolically below */
1538 # define isCNTRL_A(c) ((c) == '\0' || (c) == '\a' || (c) == '\b' \
1539 || (c) == '\f' || (c) == '\n' || (c) == '\r' \
1540 || (c) == '\t' || (c) == '\v' \
1541 || inRANGE((c), 1, 3) /* SOH, STX, ETX */ \
1542 || (c) == 7 /* U+7F DEL */ \
1543 || inRANGE((c), 0x0E, 0x13) /* SO SI DLE \
1545 || (c) == 0x18 /* U+18 CAN */ \
1546 || (c) == 0x19 /* U+19 EOM */ \
1547 || inRANGE((c), 0x1C, 0x1F) /* [FGRU]S */ \
1548 || (c) == 0x26 /* U+17 ETB */ \
1549 || (c) == 0x27 /* U+1B ESC */ \
1550 || (c) == 0x2D /* U+05 ENQ */ \
1551 || (c) == 0x2E /* U+06 ACK */ \
1552 || (c) == 0x32 /* U+16 SYN */ \
1553 || (c) == 0x37 /* U+04 EOT */ \
1554 || (c) == 0x3C /* U+14 DC4 */ \
1555 || (c) == 0x3D /* U+15 NAK */ \
1556 || (c) == 0x3F)/* U+1A SUB */
1557 # define isASCII(c) (isCNTRL_A(c) || isPRINT_A(c))
1558 # else /* isASCII is already defined for ASCII platforms, so can use that to
1560 # define isCNTRL_A(c) (isASCII(c) && ! isPRINT_A(c))
1563 /* The _L1 macros may be unnecessary for the utilities; I (khw) added them
1564 * during debugging, and it seems best to keep them. We may be called
1565 * without NATIVE_TO_LATIN1 being defined. On ASCII platforms, it doesn't
1566 * do anything anyway, so make it not a problem */
1567 # if ! defined(EBCDIC) && ! defined(NATIVE_TO_LATIN1)
1568 # define NATIVE_TO_LATIN1(ch) (ch)
1570 # define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1571 # define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1572 # define isBLANK_L1(c) (isBLANK_A(c) \
1573 || (FITS_IN_8_BITS(c) \
1574 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1575 # define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1576 # define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1577 # define isLOWER_L1(c) (isLOWER_A(c) \
1578 || (FITS_IN_8_BITS(c) \
1579 && (( NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1580 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1581 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1582 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1583 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1584 # define isPRINT_L1(c) (isPRINT_A(c) \
1585 || (FITS_IN_8_BITS(c) \
1586 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1587 # define isPUNCT_L1(c) (isPUNCT_A(c) \
1588 || (FITS_IN_8_BITS(c) \
1589 && ( NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1590 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1591 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1592 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1593 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1594 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1595 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1596 # define isSPACE_L1(c) (isSPACE_A(c) \
1597 || (FITS_IN_8_BITS(c) \
1598 && ( NATIVE_TO_LATIN1((U8) c) == 0x85 \
1599 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1600 # define isUPPER_L1(c) (isUPPER_A(c) \
1601 || (FITS_IN_8_BITS(c) \
1602 && ( IN_RANGE(NATIVE_TO_LATIN1((U8) c), \
1604 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1605 # define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1606 # define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1607 # define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1612 /* The following are not fully accurate in the above-ASCII range. I (khw)
1613 * don't think it's necessary to be so for the purposes where this gets
1615 # define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1616 # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1618 /* And these aren't accurate at all. They are useful only for above
1619 * Latin1, which utilities and bootstrapping don't deal with */
1620 # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1621 # define _HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1622 # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1624 /* Many of the macros later in this file are defined in terms of these. By
1625 * implementing them with a function, which converts the class number into
1626 * a call to the desired macro, all of the later ones work. However, that
1627 * function won't be actually defined when building a utility program (no
1628 * perl.h), and so a compiler error will be generated if one is attempted
1629 * to be used. And the above-Latin1 code points require Unicode tables to
1630 * be present, something unlikely to be the case when bootstrapping */
1631 # define _generic_isCC(c, classnum) \
1632 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1633 # define _generic_isCC_A(c, classnum) \
1634 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
1635 #endif /* End of no perl.h H_PERL */
1637 #define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1638 #define isALPHA(c) isALPHA_A(c)
1639 #define isASCII_A(c) isASCII(c)
1640 #define isASCII_L1(c) isASCII(c)
1641 #define isBLANK(c) isBLANK_A(c)
1642 #define isCNTRL(c) isCNTRL_A(c)
1643 #define isDIGIT(c) isDIGIT_A(c)
1644 #define isGRAPH(c) isGRAPH_A(c)
1645 #define isIDFIRST(c) isIDFIRST_A(c)
1646 #define isLOWER(c) isLOWER_A(c)
1647 #define isPRINT(c) isPRINT_A(c)
1648 #define isPSXSPC_A(c) isSPACE_A(c)
1649 #define isPSXSPC(c) isPSXSPC_A(c)
1650 #define isPSXSPC_L1(c) isSPACE_L1(c)
1651 #define isPUNCT(c) isPUNCT_A(c)
1652 #define isSPACE(c) isSPACE_A(c)
1653 #define isUPPER(c) isUPPER_A(c)
1654 #define isWORDCHAR(c) isWORDCHAR_A(c)
1655 #define isXDIGIT(c) isXDIGIT_A(c)
1657 /* ASCII casing. These could also be written as
1658 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1659 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1660 which uses table lookup and mask instead of subtraction. (This would
1661 work because the _MOD does not apply in the ASCII range).
1663 These actually are UTF-8 invariant casing, not just ASCII, as any non-ASCII
1664 UTF-8 invariants are neither upper nor lower. (Only on EBCDIC platforms are
1665 there non-ASCII invariants, and all of them are controls.) */
1666 #define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1667 #define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
1669 /* In the ASCII range, these are equivalent to what they're here defined to be.
1670 * But by creating these definitions, other code doesn't have to be aware of
1671 * this detail. Actually this works for all UTF-8 invariants, not just the
1672 * ASCII range. (EBCDIC platforms can have non-ASCII invariants.) */
1673 #define toFOLD(c) toLOWER(c)
1674 #define toTITLE(c) toUPPER(c)
1676 #define toLOWER_A(c) toLOWER(c)
1677 #define toUPPER_A(c) toUPPER(c)
1678 #define toFOLD_A(c) toFOLD(c)
1679 #define toTITLE_A(c) toTITLE(c)
1681 /* Use table lookup for speed; returns the input itself if is out-of-range */
1682 #define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
1684 : PL_latin1_lc[ (U8) (c) ])
1685 #define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1687 /* Modified uc. Is correct uc except for three non-ascii chars which are
1688 * all mapped to one of them, and these need special handling; returns the
1689 * input itself if is out-of-range */
1690 #define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
1692 : PL_mod_latin1_uc[ (U8) (c) ])
1693 #define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale
1695 /* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
1697 /* For internal core Perl use only: the base macro for defining macros like
1698 * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point
1699 * (0-255) to check. In a UTF-8 locale, the result is the same as calling
1700 * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like
1701 * _CC_UPPER, which gives the class number for doing this. For non-UTF-8
1702 * locales, the code to actually do the test this is passed in 'non_utf8'. If
1703 * 'c' is above 255, 0 is returned. For accessing the full range of possible
1704 * code points under locale rules, use the macros based on _generic_LC_uvchr
1705 * instead of this. */
1706 #define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \
1707 (! FITS_IN_8_BITS(c) \
1709 : IN_UTF8_CTYPE_LOCALE \
1710 ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \
1713 /* For internal core Perl use only: a helper macro for defining macros like
1714 * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to
1715 * actually do this test is passed in 'non_utf8_func', which is called on 'c',
1716 * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See
1717 * _generic_LC_base for more info */
1718 #define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \
1719 _generic_LC_base(c,utf8_locale_classnum, \
1720 non_utf8_func( (_LC_CAST) (c)))
1722 /* For internal core Perl use only: like _generic_LC, but also returns TRUE if
1723 * 'c' is the platform's native underscore character */
1724 #define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \
1725 _generic_LC_base(c, utf8_locale_classnum, \
1726 (non_utf8_func( (_LC_CAST) (c)) \
1727 || (char)(c) == '_'))
1729 /* These next three are also for internal core Perl use only: case-change
1730 * helper macros. The reason for using the PL_latin arrays is in case the
1731 * system function is defective; it ensures uniform results that conform to the
1732 * Unicod standard. It does not handle the anomalies in UTF-8 Turkic locales */
1733 #define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \
1735 : (IN_UTF8_CTYPE_LOCALE) \
1736 ? PL_latin1_lc[ (U8) (c) ] \
1737 : (cast)function((cast)(c)))
1739 /* Note that the result can be larger than a byte in a UTF-8 locale. It
1740 * returns a single value, so can't adequately return the upper case of LATIN
1741 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1742 * values "SS"); instead it asserts against that under DEBUGGING, and
1743 * otherwise returns its input. It does not handle the anomalies in UTF-8
1744 * Turkic locales. */
1745 #define _generic_toUPPER_LC(c, function, cast) \
1746 (! FITS_IN_8_BITS(c) \
1748 : ((! IN_UTF8_CTYPE_LOCALE) \
1749 ? (cast)function((cast)(c)) \
1750 : ((((U8)(c)) == MICRO_SIGN) \
1751 ? GREEK_CAPITAL_LETTER_MU \
1752 : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \
1753 ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \
1754 : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \
1755 ? (__ASSERT_(0) (c)) \
1756 : PL_mod_latin1_uc[ (U8) (c) ])))))
1758 /* Note that the result can be larger than a byte in a UTF-8 locale. It
1759 * returns a single value, so can't adequately return the fold case of LATIN
1760 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1761 * values "ss"); instead it asserts against that under DEBUGGING, and
1762 * otherwise returns its input. It does not handle the anomalies in UTF-8
1764 #define _generic_toFOLD_LC(c, function, cast) \
1765 ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \
1766 ? GREEK_SMALL_LETTER_MU \
1767 : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \
1768 || (c) != LATIN_SMALL_LETTER_SHARP_S) \
1769 _generic_toLOWER_LC(c, function, cast)))
1771 /* Use the libc versions for these if available. */
1772 #if defined(HAS_ISASCII)
1773 # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c)))
1775 # define isASCII_LC(c) isASCII(c)
1778 #if defined(HAS_ISBLANK)
1779 # define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank)
1780 #else /* Unlike isASCII, varies if in a UTF-8 locale */
1781 # define isBLANK_LC(c) ((IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c))
1787 /* The Windows functions don't bother to follow the POSIX standard, which
1788 * for example says that something can't both be a printable and a control.
1789 * But Windows treats the \t control as a printable, and does such things
1790 * as making superscripts into both digits and punctuation. This tames
1791 * these flaws by assuming that the definitions of both controls and space
1792 * are correct, and then making sure that other definitions don't have
1793 * weirdnesses, by making sure that isalnum() isn't also ispunct(), etc.
1794 * Not all possible weirdnesses are checked for, just the ones that were
1795 * detected on actual Microsoft code pages */
1797 # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1798 # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1800 # define isALPHA_LC(c) (_generic_LC(c, _CC_ALPHA, isalpha) \
1801 && isALPHANUMERIC_LC(c))
1802 # define isALPHANUMERIC_LC(c) (_generic_LC(c, _CC_ALPHANUMERIC, isalnum) && \
1804 # define isDIGIT_LC(c) (_generic_LC(c, _CC_DIGIT, isdigit) && \
1805 isALPHANUMERIC_LC(c))
1806 # define isGRAPH_LC(c) (_generic_LC(c, _CC_GRAPH, isgraph) && isPRINT_LC(c))
1807 # define isIDFIRST_LC(c) (((c) == '_') \
1808 || (_generic_LC(c, _CC_IDFIRST, isalpha) && ! isPUNCT_LC(c)))
1809 # define isLOWER_LC(c) (_generic_LC(c, _CC_LOWER, islower) && isALPHA_LC(c))
1810 # define isPRINT_LC(c) (_generic_LC(c, _CC_PRINT, isprint) && ! isCNTRL_LC(c))
1811 # define isPUNCT_LC(c) (_generic_LC(c, _CC_PUNCT, ispunct) && ! isCNTRL_LC(c))
1812 # define isUPPER_LC(c) (_generic_LC(c, _CC_UPPER, isupper) && isALPHA_LC(c))
1813 # define isWORDCHAR_LC(c) (((c) == '_') || isALPHANUMERIC_LC(c))
1814 # define isXDIGIT_LC(c) (_generic_LC(c, _CC_XDIGIT, isxdigit) \
1815 && isALPHANUMERIC_LC(c))
1817 # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1818 # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1819 # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
1821 #elif defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1822 /* For most other platforms */
1824 # define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha)
1825 # define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum)
1826 # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1827 # define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit)
1828 # define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph)
1829 # define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha)
1830 # define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower)
1831 # define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint)
1832 # define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct)
1833 # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1834 # define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper)
1835 # define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum)
1836 # define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit)
1839 # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1840 # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1841 # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
1843 #else /* The final fallback position */
1845 # define isALPHA_LC(c) (isascii(c) && isalpha(c))
1846 # define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
1847 # define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
1848 # define isDIGIT_LC(c) (isascii(c) && isdigit(c))
1849 # define isGRAPH_LC(c) (isascii(c) && isgraph(c))
1850 # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1851 # define isLOWER_LC(c) (isascii(c) && islower(c))
1852 # define isPRINT_LC(c) (isascii(c) && isprint(c))
1853 # define isPUNCT_LC(c) (isascii(c) && ispunct(c))
1854 # define isSPACE_LC(c) (isascii(c) && isspace(c))
1855 # define isUPPER_LC(c) (isascii(c) && isupper(c))
1856 # define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
1857 # define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
1859 # define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
1860 # define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
1861 # define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c))
1865 #define isIDCONT(c) isWORDCHAR(c)
1866 #define isIDCONT_A(c) isWORDCHAR_A(c)
1867 #define isIDCONT_L1(c) isWORDCHAR_L1(c)
1868 #define isIDCONT_LC(c) isWORDCHAR_LC(c)
1869 #define isPSXSPC_LC(c) isSPACE_LC(c)
1871 /* For internal core Perl use only: the base macros for defining macros like
1872 * isALPHA_uvchr. 'c' is the code point to check. 'classnum' is the POSIX class
1873 * number defined earlier in this file. _generic_uvchr() is used for POSIX
1874 * classes where there is a macro or function 'above_latin1' that takes the
1875 * single argument 'c' and returns the desired value. These exist for those
1876 * classes which have simple definitions, avoiding the overhead of a hash
1877 * lookup or inversion list binary search. _generic_swash_uvchr() can be used
1878 * for classes where that overhead is faster than a direct lookup.
1879 * _generic_uvchr() won't compile if 'c' isn't unsigned, as it won't match the
1880 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so
1881 * have duplicate checks here, so could create versions of the macros that
1882 * don't, but experiments show that gcc optimizes them out anyway. */
1884 /* Note that all ignore 'use bytes' */
1885 #define _generic_uvchr(classnum, above_latin1, c) ((c) < 256 \
1886 ? _generic_isCC(c, classnum) \
1888 #define _generic_swash_uvchr(classnum, c) ((c) < 256 \
1889 ? _generic_isCC(c, classnum) \
1890 : _is_uni_FOO(classnum, c))
1891 #define isALPHA_uvchr(c) _generic_swash_uvchr(_CC_ALPHA, c)
1892 #define isALPHANUMERIC_uvchr(c) _generic_swash_uvchr(_CC_ALPHANUMERIC, c)
1893 #define isASCII_uvchr(c) isASCII(c)
1894 #define isBLANK_uvchr(c) _generic_uvchr(_CC_BLANK, is_HORIZWS_cp_high, c)
1895 #define isCNTRL_uvchr(c) isCNTRL_L1(c) /* All controls are in Latin1 */
1896 #define isDIGIT_uvchr(c) _generic_swash_uvchr(_CC_DIGIT, c)
1897 #define isGRAPH_uvchr(c) _generic_swash_uvchr(_CC_GRAPH, c)
1898 #define isIDCONT_uvchr(c) \
1899 _generic_uvchr(_CC_WORDCHAR, _is_uni_perl_idcont, c)
1900 #define isIDFIRST_uvchr(c) \
1901 _generic_uvchr(_CC_IDFIRST, _is_uni_perl_idstart, c)
1902 #define isLOWER_uvchr(c) _generic_swash_uvchr(_CC_LOWER, c)
1903 #define isPRINT_uvchr(c) _generic_swash_uvchr(_CC_PRINT, c)
1905 #define isPUNCT_uvchr(c) _generic_swash_uvchr(_CC_PUNCT, c)
1906 #define isSPACE_uvchr(c) _generic_uvchr(_CC_SPACE, is_XPERLSPACE_cp_high, c)
1907 #define isPSXSPC_uvchr(c) isSPACE_uvchr(c)
1909 #define isUPPER_uvchr(c) _generic_swash_uvchr(_CC_UPPER, c)
1910 #define isVERTWS_uvchr(c) _generic_uvchr(_CC_VERTSPACE, is_VERTWS_cp_high, c)
1911 #define isWORDCHAR_uvchr(c) _generic_swash_uvchr(_CC_WORDCHAR, c)
1912 #define isXDIGIT_uvchr(c) _generic_uvchr(_CC_XDIGIT, is_XDIGIT_cp_high, c)
1914 #define toFOLD_uvchr(c,s,l) to_uni_fold(c,s,l)
1915 #define toLOWER_uvchr(c,s,l) to_uni_lower(c,s,l)
1916 #define toTITLE_uvchr(c,s,l) to_uni_title(c,s,l)
1917 #define toUPPER_uvchr(c,s,l) to_uni_upper(c,s,l)
1919 /* For backwards compatibility, even though '_uni' should mean official Unicode
1920 * code points, in Perl it means native for those below 256 */
1921 #define isALPHA_uni(c) isALPHA_uvchr(c)
1922 #define isALPHANUMERIC_uni(c) isALPHANUMERIC_uvchr(c)
1923 #define isASCII_uni(c) isASCII_uvchr(c)
1924 #define isBLANK_uni(c) isBLANK_uvchr(c)
1925 #define isCNTRL_uni(c) isCNTRL_uvchr(c)
1926 #define isDIGIT_uni(c) isDIGIT_uvchr(c)
1927 #define isGRAPH_uni(c) isGRAPH_uvchr(c)
1928 #define isIDCONT_uni(c) isIDCONT_uvchr(c)
1929 #define isIDFIRST_uni(c) isIDFIRST_uvchr(c)
1930 #define isLOWER_uni(c) isLOWER_uvchr(c)
1931 #define isPRINT_uni(c) isPRINT_uvchr(c)
1932 #define isPUNCT_uni(c) isPUNCT_uvchr(c)
1933 #define isSPACE_uni(c) isSPACE_uvchr(c)
1934 #define isPSXSPC_uni(c) isPSXSPC_uvchr(c)
1935 #define isUPPER_uni(c) isUPPER_uvchr(c)
1936 #define isVERTWS_uni(c) isVERTWS_uvchr(c)
1937 #define isWORDCHAR_uni(c) isWORDCHAR_uvchr(c)
1938 #define isXDIGIT_uni(c) isXDIGIT_uvchr(c)
1939 #define toFOLD_uni(c,s,l) toFOLD_uvchr(c,s,l)
1940 #define toLOWER_uni(c,s,l) toLOWER_uvchr(c,s,l)
1941 #define toTITLE_uni(c,s,l) toTITLE_uvchr(c,s,l)
1942 #define toUPPER_uni(c,s,l) toUPPER_uvchr(c,s,l)
1944 /* For internal core Perl use only: the base macros for defining macros like
1945 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code
1946 * point, not just 0-255. Like _generic_uvchr, there are two versions, one for
1947 * simple class definitions; the other for more complex. These are like
1948 * _generic_uvchr, so see it for more info. */
1949 #define _generic_LC_uvchr(latin1, above_latin1, c) \
1950 (c < 256 ? latin1(c) : above_latin1(c))
1951 #define _generic_LC_swash_uvchr(latin1, classnum, c) \
1952 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1954 #define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1955 #define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1956 _CC_ALPHANUMERIC, c)
1957 #define isASCII_LC_uvchr(c) isASCII_LC(c)
1958 #define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, \
1959 is_HORIZWS_cp_high, c)
1960 #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
1961 #define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1962 #define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
1963 #define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1964 _is_uni_perl_idcont, c)
1965 #define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1966 _is_uni_perl_idstart, c)
1967 #define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1968 #define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
1969 #define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c)
1970 #define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1971 #define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
1972 is_XPERLSPACE_cp_high, c)
1973 #define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1974 #define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1976 #define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, \
1977 is_XDIGIT_cp_high, c)
1979 #define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
1981 /* The "_safe" macros make sure that we don't attempt to read beyond 'e', but
1982 * they don't otherwise go out of their way to look for malformed UTF-8. If
1983 * they can return accurate results without knowing if the input is otherwise
1984 * malformed, they do so. For example isASCII is accurate in spite of any
1985 * non-length malformations because it looks only at a single byte. Likewise
1986 * isDIGIT looks just at the first byte for code points 0-255, as all UTF-8
1987 * variant ones return FALSE. But, if the input has to be well-formed in order
1988 * for the results to be accurate, the macros will test and if malformed will
1989 * call a routine to die
1991 * Except for toke.c, the macros do assume that e > p, asserting that on
1992 * DEBUGGING builds. Much code that calls these depends on this being true,
1993 * for other reasons. toke.c is treated specially as using the regular
1994 * assertion breaks it in many ways. All strings that these operate on there
1995 * are supposed to have an extra NUL character at the end, so that *e = \0. A
1996 * bunch of code in toke.c assumes that this is true, so the assertion allows
1998 #ifdef PERL_IN_TOKE_C
1999 # define _utf8_safe_assert(p,e) ((e) > (p) || ((e) == (p) && *(p) == '\0'))
2001 # define _utf8_safe_assert(p,e) ((e) > (p))
2004 #define _generic_utf8_safe(classnum, p, e, above_latin1) \
2005 ((! _utf8_safe_assert(p, e)) \
2006 ? (_force_out_malformed_utf8_message((U8 *) (p), (U8 *) (e), 0, 1), 0)\
2007 : (UTF8_IS_INVARIANT(*(p))) \
2008 ? _generic_isCC(*(p), classnum) \
2009 : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \
2010 ? ((LIKELY((e) - (p) > 1 && UTF8_IS_CONTINUATION(*((p)+1)))) \
2011 ? _generic_isCC(EIGHT_BIT_UTF8_TO_NATIVE(*(p), *((p)+1 )), \
2013 : (_force_out_malformed_utf8_message( \
2014 (U8 *) (p), (U8 *) (e), 0, 1), 0)) \
2016 /* Like the above, but calls 'above_latin1(p)' to get the utf8 value.
2017 * 'above_latin1' can be a macro */
2018 #define _generic_func_utf8_safe(classnum, above_latin1, p, e) \
2019 _generic_utf8_safe(classnum, p, e, above_latin1(p, e))
2020 #define _generic_non_swash_utf8_safe(classnum, above_latin1, p, e) \
2021 _generic_utf8_safe(classnum, p, e, \
2022 (UNLIKELY((e) - (p) < UTF8SKIP(p)) \
2023 ? (_force_out_malformed_utf8_message( \
2024 (U8 *) (p), (U8 *) (e), 0, 1), 0) \
2026 /* Like the above, but passes classnum to _isFOO_utf8_with_len(), instead of
2027 * having an 'above_latin1' parameter */
2028 #define _generic_swash_utf8_safe(classnum, p, e) \
2029 _generic_utf8_safe(classnum, p, e, _is_utf8_FOO_with_len(classnum, p, e))
2031 /* Like the above, but should be used only when it is known that there are no
2032 * characters in the upper-Latin1 range (128-255 on ASCII platforms) which the
2033 * class is TRUE for. Hence it can skip the tests for this range.
2034 * 'above_latin1' should include its arguments */
2035 #define _generic_utf8_safe_no_upper_latin1(classnum, p, e, above_latin1) \
2036 (__ASSERT_(_utf8_safe_assert(p, e)) \
2037 (UTF8_IS_INVARIANT(*(p))) \
2038 ? _generic_isCC(*(p), classnum) \
2039 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
2040 ? 0 /* Note that doesn't check validity for latin1 */ \
2044 #define isALPHA_utf8(p, e) isALPHA_utf8_safe(p, e)
2045 #define isALPHANUMERIC_utf8(p, e) isALPHANUMERIC_utf8_safe(p, e)
2046 #define isASCII_utf8(p, e) isASCII_utf8_safe(p, e)
2047 #define isBLANK_utf8(p, e) isBLANK_utf8_safe(p, e)
2048 #define isCNTRL_utf8(p, e) isCNTRL_utf8_safe(p, e)
2049 #define isDIGIT_utf8(p, e) isDIGIT_utf8_safe(p, e)
2050 #define isGRAPH_utf8(p, e) isGRAPH_utf8_safe(p, e)
2051 #define isIDCONT_utf8(p, e) isIDCONT_utf8_safe(p, e)
2052 #define isIDFIRST_utf8(p, e) isIDFIRST_utf8_safe(p, e)
2053 #define isLOWER_utf8(p, e) isLOWER_utf8_safe(p, e)
2054 #define isPRINT_utf8(p, e) isPRINT_utf8_safe(p, e)
2055 #define isPSXSPC_utf8(p, e) isPSXSPC_utf8_safe(p, e)
2056 #define isPUNCT_utf8(p, e) isPUNCT_utf8_safe(p, e)
2057 #define isSPACE_utf8(p, e) isSPACE_utf8_safe(p, e)
2058 #define isUPPER_utf8(p, e) isUPPER_utf8_safe(p, e)
2059 #define isVERTWS_utf8(p, e) isVERTWS_utf8_safe(p, e)
2060 #define isWORDCHAR_utf8(p, e) isWORDCHAR_utf8_safe(p, e)
2061 #define isXDIGIT_utf8(p, e) isXDIGIT_utf8_safe(p, e)
2063 #define isALPHA_utf8_safe(p, e) _generic_swash_utf8_safe(_CC_ALPHA, p, e)
2064 #define isALPHANUMERIC_utf8_safe(p, e) \
2065 _generic_swash_utf8_safe(_CC_ALPHANUMERIC, p, e)
2066 #define isASCII_utf8_safe(p, e) \
2067 /* Because ASCII is invariant under utf8, the non-utf8 macro \
2069 (__ASSERT_(_utf8_safe_assert(p, e)) isASCII(*(p)))
2070 #define isBLANK_utf8_safe(p, e) \
2071 _generic_non_swash_utf8_safe(_CC_BLANK, is_HORIZWS_high, p, e)
2074 /* Because all controls are UTF-8 invariants in EBCDIC, we can use this
2075 * more efficient macro instead of the more general one */
2076 # define isCNTRL_utf8_safe(p, e) \
2077 (__ASSERT_(_utf8_safe_assert(p, e)) isCNTRL_L1(*(p)))
2079 # define isCNTRL_utf8_safe(p, e) _generic_utf8_safe(_CC_CNTRL, p, e, 0)
2082 #define isDIGIT_utf8_safe(p, e) \
2083 _generic_utf8_safe_no_upper_latin1(_CC_DIGIT, p, e, \
2084 _is_utf8_FOO_with_len(_CC_DIGIT, p, e))
2085 #define isGRAPH_utf8_safe(p, e) _generic_swash_utf8_safe(_CC_GRAPH, p, e)
2086 #define isIDCONT_utf8_safe(p, e) _generic_func_utf8_safe(_CC_WORDCHAR, \
2087 _is_utf8_perl_idcont_with_len, p, e)
2089 /* To prevent S_scan_word in toke.c from hanging, we have to make sure that
2090 * IDFIRST is an alnum. See
2091 * https://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
2092 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
2093 * This used to be not the XID version, but we decided to go with the more
2094 * modern Unicode definition */
2095 #define isIDFIRST_utf8_safe(p, e) \
2096 _generic_func_utf8_safe(_CC_IDFIRST, \
2097 _is_utf8_perl_idstart_with_len, (U8 *) (p), (U8 *) (e))
2099 #define isLOWER_utf8_safe(p, e) _generic_swash_utf8_safe(_CC_LOWER, p, e)
2100 #define isPRINT_utf8_safe(p, e) _generic_swash_utf8_safe(_CC_PRINT, p, e)
2101 #define isPSXSPC_utf8_safe(p, e) isSPACE_utf8_safe(p, e)
2102 #define isPUNCT_utf8_safe(p, e) _generic_swash_utf8_safe(_CC_PUNCT, p, e)
2103 #define isSPACE_utf8_safe(p, e) \
2104 _generic_non_swash_utf8_safe(_CC_SPACE, is_XPERLSPACE_high, p, e)
2105 #define isUPPER_utf8_safe(p, e) _generic_swash_utf8_safe(_CC_UPPER, p, e)
2106 #define isVERTWS_utf8_safe(p, e) \
2107 _generic_non_swash_utf8_safe(_CC_VERTSPACE, is_VERTWS_high, p, e)
2108 #define isWORDCHAR_utf8_safe(p, e) \
2109 _generic_swash_utf8_safe(_CC_WORDCHAR, p, e)
2110 #define isXDIGIT_utf8_safe(p, e) \
2111 _generic_utf8_safe_no_upper_latin1(_CC_XDIGIT, p, e, \
2112 (UNLIKELY((e) - (p) < UTF8SKIP(p)) \
2113 ? (_force_out_malformed_utf8_message( \
2114 (U8 *) (p), (U8 *) (e), 0, 1), 0) \
2115 : is_XDIGIT_high(p)))
2117 #define toFOLD_utf8(p,e,s,l) toFOLD_utf8_safe(p,e,s,l)
2118 #define toLOWER_utf8(p,e,s,l) toLOWER_utf8_safe(p,e,s,l)
2119 #define toTITLE_utf8(p,e,s,l) toTITLE_utf8_safe(p,e,s,l)
2120 #define toUPPER_utf8(p,e,s,l) toUPPER_utf8_safe(p,e,s,l)
2122 /* For internal core use only, subject to change */
2123 #define _toFOLD_utf8_flags(p,e,s,l,f) _to_utf8_fold_flags (p,e,s,l,f)
2124 #define _toLOWER_utf8_flags(p,e,s,l,f) _to_utf8_lower_flags(p,e,s,l,f)
2125 #define _toTITLE_utf8_flags(p,e,s,l,f) _to_utf8_title_flags(p,e,s,l,f)
2126 #define _toUPPER_utf8_flags(p,e,s,l,f) _to_utf8_upper_flags(p,e,s,l,f)
2128 #define toFOLD_utf8_safe(p,e,s,l) _toFOLD_utf8_flags(p,e,s,l, FOLD_FLAGS_FULL)
2129 #define toLOWER_utf8_safe(p,e,s,l) _toLOWER_utf8_flags(p,e,s,l, 0)
2130 #define toTITLE_utf8_safe(p,e,s,l) _toTITLE_utf8_flags(p,e,s,l, 0)
2131 #define toUPPER_utf8_safe(p,e,s,l) _toUPPER_utf8_flags(p,e,s,l, 0)
2133 #define isALPHA_LC_utf8(p, e) isALPHA_LC_utf8_safe(p, e)
2134 #define isALPHANUMERIC_LC_utf8(p, e) isALPHANUMERIC_LC_utf8_safe(p, e)
2135 #define isASCII_LC_utf8(p, e) isASCII_LC_utf8_safe(p, e)
2136 #define isBLANK_LC_utf8(p, e) isBLANK_LC_utf8_safe(p, e)
2137 #define isCNTRL_LC_utf8(p, e) isCNTRL_LC_utf8_safe(p, e)
2138 #define isDIGIT_LC_utf8(p, e) isDIGIT_LC_utf8_safe(p, e)
2139 #define isGRAPH_LC_utf8(p, e) isGRAPH_LC_utf8_safe(p, e)
2140 #define isIDCONT_LC_utf8(p, e) isIDCONT_LC_utf8_safe(p, e)
2141 #define isIDFIRST_LC_utf8(p, e) isIDFIRST_LC_utf8_safe(p, e)
2142 #define isLOWER_LC_utf8(p, e) isLOWER_LC_utf8_safe(p, e)
2143 #define isPRINT_LC_utf8(p, e) isPRINT_LC_utf8_safe(p, e)
2144 #define isPSXSPC_LC_utf8(p, e) isPSXSPC_LC_utf8_safe(p, e)
2145 #define isPUNCT_LC_utf8(p, e) isPUNCT_LC_utf8_safe(p, e)
2146 #define isSPACE_LC_utf8(p, e) isSPACE_LC_utf8_safe(p, e)
2147 #define isUPPER_LC_utf8(p, e) isUPPER_LC_utf8_safe(p, e)
2148 #define isWORDCHAR_LC_utf8(p, e) isWORDCHAR_LC_utf8_safe(p, e)
2149 #define isXDIGIT_LC_utf8(p, e) isXDIGIT_LC_utf8_safe(p, e)
2151 /* For internal core Perl use only: the base macros for defining macros like
2152 * isALPHA_LC_utf8_safe. These are like _generic_utf8, but if the first code
2153 * point in 'p' is within the 0-255 range, it uses locale rules from the
2154 * passed-in 'macro' parameter */
2155 #define _generic_LC_utf8_safe(macro, p, e, above_latin1) \
2156 (__ASSERT_(_utf8_safe_assert(p, e)) \
2157 (UTF8_IS_INVARIANT(*(p))) \
2159 : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \
2160 ? ((LIKELY((e) - (p) > 1 && UTF8_IS_CONTINUATION(*((p)+1)))) \
2161 ? macro(EIGHT_BIT_UTF8_TO_NATIVE(*(p), *((p)+1))) \
2162 : (_force_out_malformed_utf8_message( \
2163 (U8 *) (p), (U8 *) (e), 0, 1), 0)) \
2166 #define _generic_LC_swash_utf8_safe(macro, classnum, p, e) \
2167 _generic_LC_utf8_safe(macro, p, e, \
2168 _is_utf8_FOO_with_len(classnum, p, e))
2170 #define _generic_LC_func_utf8_safe(macro, above_latin1, p, e) \
2171 _generic_LC_utf8_safe(macro, p, e, above_latin1(p, e))
2173 #define _generic_LC_non_swash_utf8_safe(classnum, above_latin1, p, e) \
2174 _generic_LC_utf8_safe(classnum, p, e, \
2175 (UNLIKELY((e) - (p) < UTF8SKIP(p)) \
2176 ? (_force_out_malformed_utf8_message( \
2177 (U8 *) (p), (U8 *) (e), 0, 1), 0) \
2180 #define isALPHANUMERIC_LC_utf8_safe(p, e) \
2181 _generic_LC_swash_utf8_safe(isALPHANUMERIC_LC, \
2182 _CC_ALPHANUMERIC, p, e)
2183 #define isALPHA_LC_utf8_safe(p, e) \
2184 _generic_LC_swash_utf8_safe(isALPHA_LC, _CC_ALPHA, p, e)
2185 #define isASCII_LC_utf8_safe(p, e) \
2186 (__ASSERT_(_utf8_safe_assert(p, e)) isASCII_LC(*(p)))
2187 #define isBLANK_LC_utf8_safe(p, e) \
2188 _generic_LC_non_swash_utf8_safe(isBLANK_LC, is_HORIZWS_high, p, e)
2189 #define isCNTRL_LC_utf8_safe(p, e) \
2190 _generic_LC_utf8_safe(isCNTRL_LC, p, e, 0)
2191 #define isDIGIT_LC_utf8_safe(p, e) \
2192 _generic_LC_swash_utf8_safe(isDIGIT_LC, _CC_DIGIT, p, e)
2193 #define isGRAPH_LC_utf8_safe(p, e) \
2194 _generic_LC_swash_utf8_safe(isGRAPH_LC, _CC_GRAPH, p, e)
2195 #define isIDCONT_LC_utf8_safe(p, e) \
2196 _generic_LC_func_utf8_safe(isIDCONT_LC, \
2197 _is_utf8_perl_idcont_with_len, p, e)
2198 #define isIDFIRST_LC_utf8_safe(p, e) \
2199 _generic_LC_func_utf8_safe(isIDFIRST_LC, \
2200 _is_utf8_perl_idstart_with_len, p, e)
2201 #define isLOWER_LC_utf8_safe(p, e) \
2202 _generic_LC_swash_utf8_safe(isLOWER_LC, _CC_LOWER, p, e)
2203 #define isPRINT_LC_utf8_safe(p, e) \
2204 _generic_LC_swash_utf8_safe(isPRINT_LC, _CC_PRINT, p, e)
2205 #define isPSXSPC_LC_utf8_safe(p, e) isSPACE_LC_utf8_safe(p, e)
2206 #define isPUNCT_LC_utf8_safe(p, e) \
2207 _generic_LC_swash_utf8_safe(isPUNCT_LC, _CC_PUNCT, p, e)
2208 #define isSPACE_LC_utf8_safe(p, e) \
2209 _generic_LC_non_swash_utf8_safe(isSPACE_LC, is_XPERLSPACE_high, p, e)
2210 #define isUPPER_LC_utf8_safe(p, e) \
2211 _generic_LC_swash_utf8_safe(isUPPER_LC, _CC_UPPER, p, e)
2212 #define isWORDCHAR_LC_utf8_safe(p, e) \
2213 _generic_LC_swash_utf8_safe(isWORDCHAR_LC, _CC_WORDCHAR, p, e)
2214 #define isXDIGIT_LC_utf8_safe(p, e) \
2215 _generic_LC_non_swash_utf8_safe(isXDIGIT_LC, is_XDIGIT_high, p, e)
2217 /* Macros for backwards compatibility and for completeness when the ASCII and
2218 * Latin1 values are identical */
2219 #define isALPHAU(c) isALPHA_L1(c)
2220 #define isDIGIT_L1(c) isDIGIT_A(c)
2221 #define isOCTAL(c) isOCTAL_A(c)
2222 #define isOCTAL_L1(c) isOCTAL_A(c)
2223 #define isXDIGIT_L1(c) isXDIGIT_A(c)
2224 #define isALNUM(c) isWORDCHAR(c)
2225 #define isALNUM_A(c) isALNUM(c)
2226 #define isALNUMU(c) isWORDCHAR_L1(c)
2227 #define isALNUM_LC(c) isWORDCHAR_LC(c)
2228 #define isALNUM_uni(c) isWORDCHAR_uni(c)
2229 #define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
2230 #define isALNUM_utf8(p,e) isWORDCHAR_utf8(p,e)
2231 #define isALNUM_LC_utf8(p,e)isWORDCHAR_LC_utf8(p,e)
2232 #define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
2233 #define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
2234 #define isALNUMC(c) isALPHANUMERIC(c)
2235 #define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
2236 #define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
2237 #define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
2238 #define isALNUMC_utf8(p,e) isALPHANUMERIC_utf8(p,e)
2239 #define isALNUMC_LC_utf8(p,e) isALPHANUMERIC_LC_utf8(p,e)
2241 /* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII,
2242 * except that they don't necessarily mean the same characters, e.g. CTRL-D is
2243 * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC.
2244 * '?' is special-cased on EBCDIC to APC, which is the control there that is
2245 * the outlier from the block that contains the other controls, just like
2246 * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0
2247 * block. If it weren't special cased, it would yield a non-control.
2248 * The conversion works both ways, so toCTRL('D') is 4, and toCTRL(4) is D,
2251 # define toCTRL(c) (__ASSERT_(FITS_IN_8_BITS(c)) toUPPER(((U8)(c))) ^ 64)
2253 # define toCTRL(c) (__ASSERT_(FITS_IN_8_BITS(c)) \
2255 ? (UNLIKELY((c) == '?') \
2256 ? QUESTION_MARK_CTRL \
2257 : (NATIVE_TO_LATIN1(toUPPER((U8) (c))) ^ 64)) \
2258 : (UNLIKELY((c) == QUESTION_MARK_CTRL) \
2260 : (LATIN1_TO_NATIVE(((U8) (c)) ^ 64)))))
2263 /* Line numbers are unsigned, 32 bits. */
2265 #define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
2267 /* Helpful alias for version prescan */
2268 #define is_LAX_VERSION(a,b) \
2269 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
2271 #define is_STRICT_VERSION(a,b) \
2272 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
2274 #define BADVERSION(a,b,c) \
2280 /* Converts a character known to represent a hexadecimal digit (0-9, A-F, or
2281 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer,
2282 * which is advanced. The input is validated only by an assert() in DEBUGGING
2283 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and
2284 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */
2285 #define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \
2288 #define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \
2292 /* Converts a character known to represent an octal digit (0-7) to its numeric
2293 * value. The input is validated only by an assert() in DEBUGGING builds. In
2294 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */
2295 #define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c)))
2297 /* Efficiently returns a boolean as to if two native characters are equivalent
2298 * case-insenstively. At least one of the characters must be one of [A-Za-z];
2299 * the ALPHA in the name is to remind you of that. This is asserted() in
2300 * DEBUGGING builds. Because [A-Za-z] are invariant under UTF-8, this macro
2301 * works (on valid input) for both non- and UTF-8-encoded bytes.
2303 * When one of the inputs is a compile-time constant and gets folded by the
2304 * compiler, this reduces to an AND and a TEST. On both EBCDIC and ASCII
2305 * machines, 'A' and 'a' differ by a single bit; the same with the upper and
2306 * lower case of all other ASCII-range alphabetics. On ASCII platforms, they
2307 * are 32 apart; on EBCDIC, they are 64. At compile time, this uses an
2308 * exclusive 'or' to find that bit and then inverts it to form a mask, with
2309 * just a single 0, in the bit position where the upper- and lowercase differ.
2311 #define isALPHA_FOLD_EQ(c1, c2) \
2312 (__ASSERT_(isALPHA_A(c1) || isALPHA_A(c2)) \
2313 ((c1) & ~('A' ^ 'a')) == ((c2) & ~('A' ^ 'a')))
2314 #define isALPHA_FOLD_NE(c1, c2) (! isALPHA_FOLD_EQ((c1), (c2)))
2317 =head1 Memory Management
2319 =for apidoc Am|void|Newx|void* ptr|int nitems|type
2320 The XSUB-writer's interface to the C C<malloc> function.
2322 Memory obtained by this should B<ONLY> be freed with L</"Safefree">.
2324 In 5.9.3, Newx() and friends replace the older New() API, and drops
2325 the first parameter, I<x>, a debug aid which allowed callers to identify
2326 themselves. This aid has been superseded by a new build option,
2327 PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
2328 there for use in XS modules supporting older perls.
2330 =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
2331 The XSUB-writer's interface to the C C<malloc> function, with
2332 cast. See also C<L</Newx>>.
2334 Memory obtained by this should B<ONLY> be freed with L</"Safefree">.
2336 =for apidoc Am|void|Newxz|void* ptr|int nitems|type
2337 The XSUB-writer's interface to the C C<malloc> function. The allocated
2338 memory is zeroed with C<memzero>. See also C<L</Newx>>.
2340 Memory obtained by this should B<ONLY> be freed with L</"Safefree">.
2342 =for apidoc Am|void|Renew|void* ptr|int nitems|type
2343 The XSUB-writer's interface to the C C<realloc> function.
2345 Memory obtained by this should B<ONLY> be freed with L</"Safefree">.
2347 =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
2348 The XSUB-writer's interface to the C C<realloc> function, with
2351 Memory obtained by this should B<ONLY> be freed with L</"Safefree">.
2353 =for apidoc Am|void|Safefree|void* ptr
2354 The XSUB-writer's interface to the C C<free> function.
2356 This should B<ONLY> be used on memory obtained using L</"Newx"> and friends.
2358 =for apidoc Am|void|Move|void* src|void* dest|int nitems|type
2359 The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
2360 source, C<dest> is the destination, C<nitems> is the number of items, and
2361 C<type> is the type. Can do overlapping moves. See also C<L</Copy>>.
2363 =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
2364 Like C<Move> but returns C<dest>. Useful
2365 for encouraging compilers to tail-call
2368 =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
2369 The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
2370 source, C<dest> is the destination, C<nitems> is the number of items, and
2371 C<type> is the type. May fail on overlapping copies. See also C<L</Move>>.
2373 =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
2375 Like C<Copy> but returns C<dest>. Useful
2376 for encouraging compilers to tail-call
2379 =for apidoc Am|void|Zero|void* dest|int nitems|type
2381 The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
2382 destination, C<nitems> is the number of items, and C<type> is the type.
2384 =for apidoc Am|void *|ZeroD|void* dest|int nitems|type
2386 Like C<Zero> but returns dest. Useful
2387 for encouraging compilers to tail-call
2390 =for apidoc Am|void|StructCopy|type *src|type *dest|type
2391 This is an architecture-independent macro to copy one structure to another.
2393 =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
2395 Fill up memory with a byte pattern (a byte repeated over and over
2396 again) that hopefully catches attempts to access uninitialized memory.
2398 =for apidoc Am|void|PoisonNew|void* dest|int nitems|type
2400 PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
2402 =for apidoc Am|void|PoisonFree|void* dest|int nitems|type
2404 PoisonWith(0xEF) for catching access to freed memory.
2406 =for apidoc Am|void|Poison|void* dest|int nitems|type
2408 PoisonWith(0xEF) for catching access to freed memory.
2412 /* Maintained for backwards-compatibility only. Use newSV() instead. */
2414 #define NEWSV(x,len) newSV(len)
2417 #define MEM_SIZE_MAX ((MEM_SIZE)-1)
2419 #define _PERL_STRLEN_ROUNDUP_UNCHECKED(n) (((n) - 1 + PERL_STRLEN_ROUNDUP_QUANTUM) & ~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM - 1))
2421 #ifdef PERL_MALLOC_WRAP
2423 /* This expression will be constant-folded at compile time. It checks
2424 * whether or not the type of the count n is so small (e.g. U8 or U16, or
2425 * U32 on 64-bit systems) that there's no way a wrap-around could occur.
2426 * As well as avoiding the need for a run-time check in some cases, it's
2427 * designed to avoid compiler warnings like:
2428 * comparison is always false due to limited range of data type
2429 * It's mathematically equivalent to
2430 * max(n) * sizeof(t) > MEM_SIZE_MAX
2433 # define _MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) \
2434 ( sizeof(MEM_SIZE) < sizeof(n) \
2435 || sizeof(t) > ((MEM_SIZE)1 << 8*(sizeof(MEM_SIZE) - sizeof(n))))
2437 /* This is written in a slightly odd way to avoid various spurious
2438 * compiler warnings. We *want* to write the expression as
2439 * _MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) && (n > C)
2440 * (for some compile-time constant C), but even when the LHS
2441 * constant-folds to false at compile-time, g++ insists on emitting
2442 * warnings about the RHS (e.g. "comparison is always false"), so instead
2445 * (cond ? n : X) > C
2447 * where X is a constant with X > C always false. Choosing a value for X
2448 * is tricky. If 0, some compilers will complain about 0 > C always being
2449 * false; if 1, Coverity complains when n happens to be the constant value
2450 * '1', that cond ? 1 : 1 has the same value on both branches; so use C
2451 * for X and hope that nothing else whines.
2454 # define _MEM_WRAP_WILL_WRAP(n,t) \
2455 ((_MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) ? (MEM_SIZE)(n) : \
2456 MEM_SIZE_MAX/sizeof(t)) > MEM_SIZE_MAX/sizeof(t))
2458 # define MEM_WRAP_CHECK(n,t) \
2459 (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \
2460 && (croak_memory_wrap(),0))
2462 # define MEM_WRAP_CHECK_1(n,t,a) \
2463 (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \
2464 && (Perl_croak_nocontext("%s",(a)),0))
2466 /* "a" arg must be a string literal */
2467 # define MEM_WRAP_CHECK_s(n,t,a) \
2468 (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \
2469 && (Perl_croak_nocontext("" a ""),0))
2471 #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
2473 #define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (croak_memory_wrap(),0) : 0), _PERL_STRLEN_ROUNDUP_UNCHECKED(n))
2476 #define MEM_WRAP_CHECK(n,t)
2477 #define MEM_WRAP_CHECK_1(n,t,a)
2478 #define MEM_WRAP_CHECK_s(n,t,a)
2479 #define MEM_WRAP_CHECK_(n,t)
2481 #define PERL_STRLEN_ROUNDUP(n) _PERL_STRLEN_ROUNDUP_UNCHECKED(n)
2487 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
2488 * go through functions, which are handy for debugging breakpoints, but
2489 * which more importantly get the immediate calling environment (file and
2490 * line number, and C function name if available) passed in. This info can
2491 * then be used for logging the calls, for which one gets a sample
2492 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
2495 * - not all memory allocs get logged, only those
2496 * that go through Newx() and derivatives (while all
2497 * Safefrees do get logged)
2498 * - __FILE__ and __LINE__ do not work everywhere
2499 * - __func__ or __FUNCTION__ even less so
2500 * - I think more goes on after the perlio frees but
2501 * the thing is that STDERR gets closed (as do all
2502 * the file descriptors)
2503 * - no deeper calling stack than the caller of the Newx()
2504 * or the kind, but do I look like a C reflection/introspection
2506 * - the function prototypes for the logging functions
2507 * probably should maybe be somewhere else than handy.h
2508 * - one could consider inlining (macrofying) the logging
2509 * for speed, but I am too lazy
2510 * - one could imagine recording the allocations in a hash,
2511 * (keyed by the allocation address?), and maintain that
2512 * through reallocs and frees, but how to do that without
2513 * any News() happening...?
2514 * - lots of -Ddefines to get useful/controllable output
2515 * - lots of ENV reads
2519 # ifndef PERL_MEM_LOG_NOIMPL
2528 # if defined(PERL_IN_SV_C) /* those are only used in sv.c */
2529 void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
2530 void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
2537 #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
2538 #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
2539 #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
2542 #ifndef MEM_LOG_ALLOC
2543 #define MEM_LOG_ALLOC(n,t,a) (a)
2545 #ifndef MEM_LOG_REALLOC
2546 #define MEM_LOG_REALLOC(n,t,v,a) (a)
2548 #ifndef MEM_LOG_FREE
2549 #define MEM_LOG_FREE(a) (a)
2552 #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
2553 #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
2554 #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
2557 /* pre 5.9.x compatibility */
2558 #define New(x,v,n,t) Newx(v,n,t)
2559 #define Newc(x,v,n,t,c) Newxc(v,n,t,c)
2560 #define Newz(x,v,n,t) Newxz(v,n,t)
2563 #define Renew(v,n,t) \
2564 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
2565 #define Renewc(v,n,t,c) \
2566 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
2569 #define Safefree(d) \
2570 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
2572 #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
2575 /* assert that a valid ptr has been supplied - use this instead of assert(ptr) *
2576 * as it handles cases like constant string arguments without throwing warnings *
2577 * the cast is required, as is the inequality check, to avoid warnings */
2578 #define perl_assert_ptr(p) assert( ((void*)(p)) != 0 )
2581 #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
2582 #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
2583 #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), (void)memzero((char*)(d), (n) * sizeof(t)))
2585 /* Like above, but returns a pointer to 'd' */
2586 #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
2587 #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
2588 #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), memzero((char*)(d), (n) * sizeof(t)))
2590 #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
2591 #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
2592 #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
2593 #define Poison(d,n,t) PoisonFree(d,n,t)
2596 # define PERL_POISON_EXPR(x) x
2598 # define PERL_POISON_EXPR(x)
2601 #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
2606 =for apidoc Am|STRLEN|C_ARRAY_LENGTH|void *a
2608 Returns the number of elements in the input C array (so you want your
2609 zero-based indices to be less than but not equal to).
2611 =for apidoc Am|void *|C_ARRAY_END|void *a
2613 Returns a pointer to one element past the final element of the input C array.
2617 C_ARRAY_END is one past the last: half-open/half-closed range, not
2618 last-inclusive range.
2620 #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
2621 #define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a))
2625 # define Perl_va_copy(s, d) va_copy(d, s)
2626 # elif defined(__va_copy)
2627 # define Perl_va_copy(s, d) __va_copy(d, s)
2629 # define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
2633 /* convenience debug macros */
2635 #define pTHX_FORMAT "Perl interpreter: 0x%p"
2636 #define pTHX__FORMAT ", Perl interpreter: 0x%p"
2637 #define pTHX_VALUE_ (void *)my_perl,
2638 #define pTHX_VALUE (void *)my_perl
2639 #define pTHX__VALUE_ ,(void *)my_perl,
2640 #define pTHX__VALUE ,(void *)my_perl
2643 #define pTHX__FORMAT
2646 #define pTHX__VALUE_
2648 #endif /* USE_ITHREADS */
2650 /* Perl_deprecate was not part of the public API, and did not have a deprecate()
2651 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
2652 CPAN::Unpack show any users outside the core. */
2654 # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \
2655 "Use of " s " is deprecated")
2656 # define deprecate_disappears_in(when,message) \
2657 Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \
2658 message ", and will disappear in Perl " when)
2659 # define deprecate_fatal_in(when,message) \
2660 Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \
2661 message ". Its use will be fatal in Perl " when)
2664 /* Internal macros to deal with gids and uids */
2667 # if Uid_t_size > IVSIZE
2668 # define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
2669 # define SvUID(sv) SvNV(sv)
2670 # elif Uid_t_sign <= 0
2671 # define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
2672 # define SvUID(sv) SvIV(sv)
2674 # define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
2675 # define SvUID(sv) SvUV(sv)
2676 # endif /* Uid_t_size */
2678 # if Gid_t_size > IVSIZE
2679 # define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
2680 # define SvGID(sv) SvNV(sv)
2681 # elif Gid_t_sign <= 0
2682 # define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
2683 # define SvGID(sv) SvIV(sv)
2685 # define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
2686 # define SvGID(sv) SvUV(sv)
2687 # endif /* Gid_t_size */
2691 #endif /* PERL_HANDY_H_ */
2694 * ex: set ts=8 sts=4 sw=4 et:
https://perl5.git.perl.org / perl5.git / blob