This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
utf8.c: Add locale support to functions
[perl5.git] / handy.h
CommitLineData
a0d0e21e 1/* handy.h
a687059c 2 *
1129b882 3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000,
da5d8dbb 4 * 2001, 2002, 2004, 2005, 2006, 2007, 2008, 2012 by Larry Wall and others
a687059c 5 *
6e21c824
LW
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.
8d063cd8 8 *
8d063cd8
LW
9 */
10
85e6fe83 11#if !defined(__STDC__)
378cc40b
LW
12#ifdef NULL
13#undef NULL
14#endif
a687059c 15# define NULL 0
85e6fe83
LW
16#endif
17
24792b8d
NC
18#ifndef PERL_CORE
19# define Null(type) ((type)NULL)
954c1994
GS
20
21/*
ccfc67b7 22=head1 Handy Values
954c1994 23
3609ea0d 24=for apidoc AmU||Nullch
24792b8d 25Null character pointer. (No longer available when C<PERL_CORE> is defined.)
2307c6d0 26
954c1994 27=for apidoc AmU||Nullsv
24792b8d 28Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
29
30=cut
31*/
32
24792b8d
NC
33# define Nullch Null(char*)
34# define Nullfp Null(PerlIO*)
35# define Nullsv Null(SV*)
36#endif
8d063cd8 37
641d3f0b
PP
38#ifdef TRUE
39#undef TRUE
40#endif
41#ifdef FALSE
42#undef FALSE
43#endif
44#define TRUE (1)
45#define FALSE (0)
46
cf3f0ffb
DM
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;
49 * e.g.:
50 *
51 * const SV *sv = ...;
52 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
53 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
54 */
55
b1bc3f34 56#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 57# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
58#else
59# define MUTABLE_PTR(p) ((void *) (p))
60#endif
61
a062e10d 62#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 63#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 64#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 65#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 66#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 67#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 68
bd31be4b
NC
69#ifdef I_STDBOOL
70# include <stdbool.h>
71# ifndef HAS_BOOL
72# define HAS_BOOL 1
73# endif
74#endif
75
8e84507e 76/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
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
8e84507e 79 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
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
5d94fbed 83*/
3609ea0d 84#ifdef __GNUG__ /* GNU g++ has bool built-in */
5d94fbed 85# ifndef HAS_BOOL
c1d22f6b 86# define HAS_BOOL 1
5d94fbed 87# endif
5d94fbed
AD
88#endif
89
641d3f0b
PP
90/* The NeXT dynamic loader headers will not build with the bool macro
91 So declare them now to clear confusion.
92*/
8f1f23e8 93#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
94# undef FALSE
95# undef TRUE
96 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
97# define ENUM_BOOL 1
98# ifndef HAS_BOOL
99# define HAS_BOOL 1
100# endif /* !HAS_BOOL */
8f1f23e8 101#endif /* NeXT || __NeXT__ */
641d3f0b 102
5d94fbed 103#ifndef HAS_BOOL
cc3315ba 104# if defined(VMS)
5d94fbed
AD
105# define bool int
106# else
107# define bool char
108# endif
c1d22f6b 109# define HAS_BOOL 1
a687059c 110#endif
0d3e774c 111
f2338a2e
DM
112/* a simple (bool) cast may not do the right thing: if bool is defined
113 * as char for example, then the cast from int is implementation-defined
18f5643b 114 * (bool)!!(cbool) in a ternary triggers a bug in xlc on AIX
f2338a2e
DM
115 */
116
18f5643b 117#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 118
46c6c7e2 119/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
120 * XXX Should really be a Configure probe, with HAS__FUNCTION__
121 * and FUNCTION__ as results.
122 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
123#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
124# define FUNCTION__ __func__
125#else
70764c00 126# if (defined(_MSC_VER) && _MSC_VER < 1300) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 127 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
46c6c7e2
JH
128# define FUNCTION__ ""
129# else
130# define FUNCTION__ __FUNCTION__ /* Common extension. */
131# endif
132#endif
133
27d4fb96
PP
134/* XXX A note on the perl source internal type system. The
135 original intent was that I32 be *exactly* 32 bits.
136
137 Currently, we only guarantee that I32 is *at least* 32 bits.
138 Specifically, if int is 64 bits, then so is I32. (This is the case
139 for the Cray.) This has the advantage of meshing nicely with
140 standard library calls (where we pass an I32 and the library is
141 expecting an int), but the disadvantage that an I32 is not 32 bits.
142 Andy Dougherty August 1996
24fef2a7 143
dc45a647
MB
144 There is no guarantee that there is *any* integral type with
145 exactly 32 bits. It is perfectly legal for a system to have
146 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 147
dc45a647
MB
148 Similarly, there is no guarantee that I16 and U16 have exactly 16
149 bits.
693762b4 150
8e84507e
NIS
151 For dealing with issues that may arise from various 32/64-bit
152 systems, we will ask Configure to check out
8175356b 153
3609ea0d
JH
154 SHORTSIZE == sizeof(short)
155 INTSIZE == sizeof(int)
156 LONGSIZE == sizeof(long)
dc45a647 157 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 158 PTRSIZE == sizeof(void *)
dc45a647
MB
159 DOUBLESIZE == sizeof(double)
160 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 161
27d4fb96
PP
162*/
163
69512466
JH
164#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
165# include <inttypes.h>
dd0eed91
JH
166# ifdef INT32_MIN_BROKEN
167# undef INT32_MIN
168# define INT32_MIN (-2147483647-1)
169# endif
170# ifdef INT64_MIN_BROKEN
171# undef INT64_MIN
172# define INT64_MIN (-9223372036854775807LL-1)
173# endif
69512466
JH
174#endif
175
8175356b
JH
176typedef I8TYPE I8;
177typedef U8TYPE U8;
178typedef I16TYPE I16;
179typedef U16TYPE U16;
180typedef I32TYPE I32;
181typedef U32TYPE U32;
6b8eaf93
JH
182#ifdef PERL_CORE
183# ifdef HAS_QUAD
8175356b
JH
184typedef I64TYPE I64;
185typedef U64TYPE U64;
6b8eaf93
JH
186# endif
187#endif /* PERL_CORE */
8175356b 188
69512466 189#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
80524f33
JL
190# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
191# define PeRl_INT64_C(c) CAT2(c,LL)
192# define PeRl_UINT64_C(c) CAT2(c,ULL)
193# else
52865553
SH
194# if QUADKIND == QUAD_IS___INT64
195# define PeRl_INT64_C(c) CAT2(c,I64)
196# define PeRl_UINT64_C(c) CAT2(c,UI64)
69512466 197# else
52865553
SH
198# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
199# define PeRl_INT64_C(c) CAT2(c,L)
200# define PeRl_UINT64_C(c) CAT2(c,UL)
69512466 201# else
80524f33
JL
202# define PeRl_INT64_C(c) ((I64TYPE)(c))
203# define PeRl_UINT64_C(c) ((U64TYPE)(c))
69512466
JH
204# endif
205# endif
e8c95190 206# endif
80524f33
JL
207# ifndef UINT64_C
208# define UINT64_C(c) PeRl_UINT64_C(c)
209# endif
210# ifndef INT64_C
211# define INT64_C(c) PeRl_INT64_C(c)
212# endif
e8c95190 213#endif
dc750f2e 214
d8668976 215#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 216
5ff3f7a4
GS
217/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
218 Please search CHAR_MAX in perl.h for further details. */
219#define U8_MAX UINT8_MAX
220#define U8_MIN UINT8_MIN
221
5ff3f7a4
GS
222#define I16_MAX INT16_MAX
223#define I16_MIN INT16_MIN
224#define U16_MAX UINT16_MAX
225#define U16_MIN UINT16_MIN
226
5ff3f7a4
GS
227#define I32_MAX INT32_MAX
228#define I32_MIN INT32_MIN
0e983133
GS
229#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
230# define U32_MAX UINT32_MAX
231#else
232# define U32_MAX 4294967295U
233#endif
5ff3f7a4
GS
234#define U32_MIN UINT32_MIN
235
236#else
237
5c9fa16e
KA
238/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
239 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
240#define U8_MAX PERL_UCHAR_MAX
241#define U8_MIN PERL_UCHAR_MIN
79072805 242
27d4fb96
PP
243#define I16_MAX PERL_SHORT_MAX
244#define I16_MIN PERL_SHORT_MIN
245#define U16_MAX PERL_USHORT_MAX
246#define U16_MIN PERL_USHORT_MIN
79072805 247
c4f23d77 248#if LONGSIZE > 4
27d4fb96
PP
249# define I32_MAX PERL_INT_MAX
250# define I32_MIN PERL_INT_MIN
251# define U32_MAX PERL_UINT_MAX
252# define U32_MIN PERL_UINT_MIN
79072805 253#else
27d4fb96
PP
254# define I32_MAX PERL_LONG_MAX
255# define I32_MIN PERL_LONG_MIN
256# define U32_MAX PERL_ULONG_MAX
257# define U32_MIN PERL_ULONG_MIN
79072805
LW
258#endif
259
5ff3f7a4
GS
260#endif
261
58a9a5d5 262/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
263#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
264#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
265#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
266
ff68c719 267#define Ctl(ch) ((ch) & 037)
8d063cd8 268
3fe05580
MHM
269/*
270=head1 SV-Body Allocation
271
272=for apidoc Ama|SV*|newSVpvs|const char* s
273Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
274
84bafc02
NC
275=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
276Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
277pair.
278
3fe05580
MHM
279=for apidoc Ama|SV*|newSVpvs_share|const char* s
280Like C<newSVpvn_share>, but takes a literal string instead of a string/length
281pair and omits the hash parameter.
282
9dcc53ea
Z
283=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
284Like C<sv_catpvn_flags>, but takes a literal string instead of a
285string/length pair.
286
287=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
288Like C<sv_catpvn_nomg>, but takes a literal string instead of a
289string/length pair.
290
7b4e7469 291=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
292Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
293
9dcc53ea
Z
294=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
295Like C<sv_catpvn_mg>, but takes a literal string instead of a
296string/length pair.
297
7b4e7469 298=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
299Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
300
9dcc53ea
Z
301=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
302Like C<sv_setpvn_mg>, but takes a literal string instead of a
303string/length pair.
304
305=for apidoc Am|SV *|sv_setref_pvs|const char* s
306Like C<sv_setref_pvn>, but takes a literal string instead of a
307string/length pair.
308
3fe05580
MHM
309=head1 Memory Management
310
311=for apidoc Ama|char*|savepvs|const char* s
312Like C<savepvn>, but takes a literal string instead of a string/length pair.
313
9dcc53ea
Z
314=for apidoc Ama|char*|savesharedpvs|const char* s
315A version of C<savepvs()> which allocates the duplicate string in memory
316which is shared between threads.
317
3fe05580
MHM
318=head1 GV Functions
319
320=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
321Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
322
323=head1 Hash Manipulation Functions
324
325=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
326Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
327
328=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
329Like C<hv_store>, but takes a literal string instead of a string/length pair
330and omits the hash parameter.
331
510966aa
Z
332=head1 Lexer interface
333
334=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
335
336Like L</lex_stuff_pvn>, but takes a literal string instead of a
337string/length pair.
338
3fe05580
MHM
339=cut
340*/
341
926bb54c
KW
342/* concatenating with "" ensures that only literal strings are accepted as
343 * argument */
26bfd924 344#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 345
926bb54c
KW
346/* note that STR_WITH_LEN() can't be used as argument to macros or functions
347 * that under some configurations might be macros, which means that it requires
348 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
349 */
350
351/* STR_WITH_LEN() shortcuts */
352#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
353#define newSVpvs_flags(str,flags) \
354 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 355#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
356#define sv_catpvs_flags(sv, str, flags) \
357 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
358#define sv_catpvs_nomg(sv, str) \
359 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
360#define sv_catpvs(sv, str) \
361 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
362#define sv_catpvs_mg(sv, str) \
363 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 364#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
365#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
366#define sv_setref_pvs(rv, classname, str) \
367 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 368#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
369#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
370#define gv_stashpvs(str, create) \
371 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
372#define gv_fetchpvs(namebeg, add, sv_type) \
373 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
374#define gv_fetchpvn(namebeg, len, add, sv_type) \
375 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
376#define sv_catxmlpvs(dsv, str, utf8) \
377 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
378#define hv_fetchs(hv,key,lval) \
379 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
380 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
381 : HV_FETCH_JUST_SV, NULL, 0))
382
383#define hv_stores(hv,key,val) \
384 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
385 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 386
510966aa
Z
387#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
388
b96d8cd9
NC
389#define get_cvs(str, flags) \
390 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 391
954c1994 392/*
ccfc67b7
JH
393=head1 Miscellaneous Functions
394
954c1994
GS
395=for apidoc Am|bool|strNE|char* s1|char* s2
396Test two strings to see if they are different. Returns true or
397false.
398
399=for apidoc Am|bool|strEQ|char* s1|char* s2
400Test two strings to see if they are equal. Returns true or false.
401
402=for apidoc Am|bool|strLT|char* s1|char* s2
403Test two strings to see if the first, C<s1>, is less than the second,
404C<s2>. Returns true or false.
405
406=for apidoc Am|bool|strLE|char* s1|char* s2
407Test two strings to see if the first, C<s1>, is less than or equal to the
408second, C<s2>. Returns true or false.
409
410=for apidoc Am|bool|strGT|char* s1|char* s2
411Test two strings to see if the first, C<s1>, is greater than the second,
412C<s2>. Returns true or false.
413
414=for apidoc Am|bool|strGE|char* s1|char* s2
415Test two strings to see if the first, C<s1>, is greater than or equal to
416the second, C<s2>. Returns true or false.
417
418=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
419Test two strings to see if they are different. The C<len> parameter
420indicates the number of bytes to compare. Returns true or false. (A
421wrapper for C<strncmp>).
422
423=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
424Test two strings to see if they are equal. The C<len> parameter indicates
425the number of bytes to compare. Returns true or false. (A wrapper for
426C<strncmp>).
427
428=cut
429*/
430
8d063cd8
LW
431#define strNE(s1,s2) (strcmp(s1,s2))
432#define strEQ(s1,s2) (!strcmp(s1,s2))
433#define strLT(s1,s2) (strcmp(s1,s2) < 0)
434#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
435#define strGT(s1,s2) (strcmp(s1,s2) > 0)
436#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
437#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
438#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 439
36477c24
PP
440#ifdef HAS_MEMCMP
441# define memNE(s1,s2,l) (memcmp(s1,s2,l))
442# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
443#else
444# define memNE(s1,s2,l) (bcmp(s1,s2,l))
445# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
446#endif
447
568a785a 448#define memEQs(s1, l, s2) \
8e6497b6 449 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
450#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
451
bbce6d69
PP
452/*
453 * Character classes.
454 *
455 * Unfortunately, the introduction of locales means that we
456 * can't trust isupper(), etc. to tell the truth. And when
457 * it comes to /\w+/ with tainting enabled, we *must* be able
458 * to trust our character classes.
459 *
460 * Therefore, the default tests in the text of Perl will be
461 * independent of locale. Any code that wants to depend on
462 * the current locale will use the tests that begin with "lc".
463 */
464
2304df62
AD
465#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
466# ifndef CTYPE256
467# define CTYPE256
468# endif
469#endif
470
954c1994 471/*
ccfc67b7
JH
472
473=head1 Character classes
243effed
KW
474This section is about functions (really macros) that classify characters
475into types, such as punctuation versus alphabetic, etc. Most of these are
476analogous to regular expression character classes. (See
477L<perlrecharclass/POSIX Character Classes>.) There are several variants for
478each class. (Not all macros have all variants; each item below lists the
479ones valid for it.) None are affected by C<use bytes>, and only the ones
480with C<LC> in the name are affected by the current locale.
481
482The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
483C<U8>) as input and returns a boolean as to whether or not the character
484represented by that octet is in the named class based on platform, Unicode, and
485Perl rules. If the input is a number that doesn't fit in an octet, FALSE is
486always returned.
487
488Variant C<isFOO_A> (e.g., C<isALPHA_A()>) will return TRUE only if the input is
489also in the ASCII character set. For ASCII platforms, the base function with
490no suffix and the one with the C<_A> suffix are identical. On EBCDIC
491platforms, the C<_A> suffix function will not return true unless the specified
492character also has an ASCII equivalent.
493
494Variant C<isFOO_L1> operates on the full Latin1 character set. For EBCDIC
495platforms, the base function with no suffix and the one with the C<_L1> suffix
496are identical. For ASCII platforms, the C<_L1> suffix imposes the Latin-1
497character set onto the platform. That is, the code points that are ASCII are
498unaffected, since ASCII is a subset of Latin-1. But the non-ASCII code points
499are treated as if they are Latin-1 characters. For example, C<isSPACE_L1()>
500will return true when called with the code point 0xA0, which is the Latin-1
501NO-BREAK SPACE.
502
503Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
504point as input. If the code point is larger than 255, Unicode rules are used
505to determine if it is in the character class. For example,
506C<isWORDCHAR(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A WITH
507MACRON in Unicode, and is a word character.
508
509Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
510(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
511classification of just the first character in the string is tested.
512
513Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but uses
514the C library function that gives the named classification instead of
515hard-coded rules. For example, C<isDIGIT_LC()> returns the result of calling
516C<isdigit()>. This means that the result is based on the current locale, which
517is what C<LC> in the name stands for. FALSE is always returned if the input
518won't fit into an octet.
519
520Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
521returns the same as C<isFOO_LC> for input code points less than 256, and
522returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
523
524Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
525(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
526classification of just the first character in the string is tested.
ccfc67b7 527
954c1994 528=for apidoc Am|bool|isALPHA|char ch
2787a470 529Returns a boolean indicating whether the specified character is an
243effed
KW
530alphabetic character in the platform's native character set, analogous to
531C<m/[[:alpha:]]/>.
8a58bdcf 532See the L<top of this section|/Character classes> for an explanation of variants
243effed
KW
533C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>
534C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf
KW
535
536=for apidoc Am|bool|isASCII|char ch
537Returns a boolean indicating whether the specified character is one of the 128
243effed
KW
538characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
539On non-ASCII platforms, it is if this
8a58bdcf
KW
540character corresponds to an ASCII character. Variants C<isASCII_A()> and
541C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
542See the L<top of this section|/Character classes> for an explanation of variants
543C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
544C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
545library routine C<isascii()>. In these cases, the variants whose names contain
546C<LC> are the same as the corresponding ones without.
547
548=for apidoc Am|bool|isBLANK|char ch
549Returns a boolean indicating whether the specified character is a
550character considered to be a blank in the platform's native character set,
551analogous to C<m/[[:blank:]]/>.
552See the L<top of this section|/Character classes> for an explanation of variants
553C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>
554C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
555platforms do not have the C library routine C<isblank()>. In these cases, the
556variants whose names contain C<LC> are the same as the corresponding ones
557without.
558
559=for apidoc Am|bool|isCNTRL|char ch
560Returns a boolean indicating whether the specified character is a
561control character in the platform's native character set,
562analogous to C<m/[[:cntrl:]]/>.
563See the L<top of this section|/Character classes> for an explanation of variants
564C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>
565C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>.
954c1994 566
954c1994 567=for apidoc Am|bool|isDIGIT|char ch
2787a470 568Returns a boolean indicating whether the specified character is a
243effed 569digit in the platform's native character set, analogous to C<m/[[:digit:]]/>.
8a58bdcf 570Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed
KW
571See the L<top of this section|/Character classes> for an explanation of variants
572C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC> C<isDIGIT_LC_uvchr>, and
573C<isDIGIT_LC_utf8>.
574
575=for apidoc Am|bool|isGRAPH|char ch
576Returns a boolean indicating whether the specified character is a
577graphic character in the platform's native character set, analogous to
578C<m/[[:graph:]]/>.
579See the L<top of this section|/Character classes> for an explanation of variants
580C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>
581C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 582
0c82b6df 583=for apidoc Am|bool|isLOWER|char ch
2787a470 584Returns a boolean indicating whether the specified character is a
243effed
KW
585lowercase character in the platform's native character set, analogous to
586C<m/[[:lower:]]/>.
8a58bdcf 587See the L<top of this section|/Character classes> for an explanation of variants
243effed
KW
588C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>
589C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 590
c99e91e9 591=for apidoc Am|bool|isOCTAL|char ch
2787a470 592Returns a boolean indicating whether the specified character is an
06dc89af 593octal digit, [0-7] in the platform's native character set.
243effed
KW
594The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
595C<isOCTAL>.
596
597=for apidoc Am|bool|isPUNCT|char ch
598Returns a boolean indicating whether the specified character is a
599punctuation character in the platform's native character set, analogous to
600C<m/[[:punct:]]/>. Note that the definition of what is punctuation isn't as
601straightforward as one might desire. See L<perlrecharclass/POSIX Character
602Classes> for details.
603See the L<top of this section|/Character classes> for an explanation of variants
604C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>
605C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 606
0c82b6df 607=for apidoc Am|bool|isSPACE|char ch
2787a470 608Returns a boolean indicating whether the specified character is a
243effed
KW
609whitespace character in the platform's native character set. This is analogous
610to what C<m/\s/> and C<m/[[:space:]]/> match in a regular expression.
8a58bdcf 611See the L<top of this section|/Character classes> for an explanation of variants
243effed
KW
612C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>
613C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 614
954c1994 615=for apidoc Am|bool|isUPPER|char ch
2787a470 616Returns a boolean indicating whether the specified character is an
243effed
KW
617uppercase character in the platform's native character set, analogous to
618C<m/[[:upper:]]/>.
8a58bdcf 619See the L<top of this section|/Character classes> for an explanation of variants
243effed
KW
620C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>
621C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 622
243effed 623=for apidoc Am|bool|isPRINT|char ch
8eea39dd 624Returns a boolean indicating whether the specified character is a
243effed
KW
625printable character in the platform's native character set, analogous to
626C<m/[[:print:]]/>.
8a58bdcf 627See the L<top of this section|/Character classes> for an explanation of variants
243effed
KW
628C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>
629C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
630
631=for apidoc Am|bool|isWORDCHAR|char ch
632Returns a boolean indicating whether the specified character is a character
633that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
634in a regular expression. A word character is an alphabetic character, a
635decimal digit, a connecting punctuation character (such as an underscore), or
636a "mark" character that attaches to one of those (like some sort of accent).
637C<isALNUM()> is a synonym provided for backward compatibility, even though a
638word character includes more than the standard C language meaning of
639alphanumeric.
640See the L<top of this section|/Character classes> for an explanation of variants
641C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
642C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
643
644=for apidoc Am|bool|isXDIGIT|char ch
645Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
646digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
647and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
648See the L<top of this section|/Character classes> for an explanation of variants
649C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
650C<isXDIGIT_LC_utf8>.
651
652=head1 Miscellaneous Functions
8eea39dd 653
95a59cab 654=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 655Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
656Behaviour is only well defined when isXDIGIT(*str) is true.
657
e7c1e6c1
KW
658=head1 Character case changing
659
954c1994 660=for apidoc Am|char|toUPPER|char ch
06dc89af
KW
661Converts the specified character to uppercase in the platform's native
662character set, if possible; otherwise returns the input character itself.
954c1994
GS
663
664=for apidoc Am|char|toLOWER|char ch
06dc89af
KW
665Converts the specified character to lowercase in the platform's native
666character set, if possible; otherwise returns the input character itself.
954c1994
GS
667
668=cut
353c9b6f 669
243effed
KW
670Still undocumented are ALNUMC, PSXSPC, VERTSPACE, and IDFIRST, and the other
671toUPPER etc functions
672
8a58bdcf 673Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
674patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
675
954c1994
GS
676*/
677
7c062697
KW
678/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
679 * is known only in the perl core, and this macro can be called from outside
680 * that */
681#ifdef HAS_QUAD
682# define WIDEST_UTYPE U64TYPE
683#else
684# define WIDEST_UTYPE U32
685#endif
686
3912bc88
KW
687/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
688 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
689 * bits of information are lost even on a 64-bit machine, but to get the
690 * compiler to optimize it out if possible. This is because Configure makes
691 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
692 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
693 */
694#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 695
41f43cc2 696#ifdef EBCDIC
ff7eba81 697# define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128))
41f43cc2
KW
698#else
699# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
700#endif
701
ec79cfbb 702#define isASCII_A(c) isASCII(c)
3509bb53 703#define isASCII_L1(c) isASCII(c)
6515cb76 704
c2da0b36
KW
705/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
706 * and the 8 possible permutations of those bits exactly comprise the 8 octal
707 * digits */
3d33b3e3 708#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 709
ec79cfbb 710/* ASCII range only */
9fb1bf9d 711#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 712
66c17564
KW
713/* Character class numbers. For internal core Perl use only. These are used
714 * in PL_charclass[] and the ones up through the one that corresponds to
715 * <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h. These use names used
716 * in l1_char_class_tab.h but their actual definitions are here. If that has a
717 * name not used here, it won't compile. */
f4cdb42c
KW
718# define _CC_WORDCHAR 0
719# define _CC_SPACE 1
720# define _CC_DIGIT 2
721# define _CC_ALNUMC 3
722# define _CC_ALPHA 4
723# define _CC_ASCII 5
724# define _CC_CNTRL 6
725# define _CC_GRAPH 7
726# define _CC_LOWER 8
727# define _CC_PRINT 9
728# define _CC_PUNCT 10
729# define _CC_UPPER 11
730# define _CC_XDIGIT 12
731# define _CC_PSXSPC 13
732# define _CC_BLANK 14
a0947d7b
KW
733# define _CC_VERTSPACE 15
734# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
735
736# define _CC_IDFIRST 16
737# define _CC_CHARNAME_CONT 17
738# define _CC_NONLATIN1_FOLD 18
739# define _CC_QUOTEMETA 19
740# define _CC_NON_FINAL_FOLD 20
741# define _CC_IS_IN_SOME_FOLD 21
742/* Unused: 22-31
f4cdb42c
KW
743 * If more bits are needed, one could add a second word for non-64bit
744 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
745 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
746 * is used only for optimization (as of this writing), and differs in the
747 * Latin1 range from the ALPHA bit only in two relatively unimportant
748 * characters: the masculine and feminine ordinal indicators */
96ac0975
NC
749
750# ifdef DOINIT
751EXTCONST U32 PL_charclass[] = {
752# include "l1_char_class_tab.h"
753};
754
755# else /* ! DOINIT */
756EXTCONST U32 PL_charclass[];
757# endif
758
265c1f46 759 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70
KW
760# define _CC_mask(classnum) (1U << (classnum))
761# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
926bb54c 762 && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask(classnum)))
4eeeb416 763
f4cdb42c
KW
764 /* The mask for the _A versions of the macros; it just adds in the bit for
765 * ASCII. */
766# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
767
768 /* The _A version makes sure that both the desired bit and the ASCII bit
769 * are present */
770# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
771 && ((PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask_A(classnum)) \
772 == _CC_mask_A(classnum)))
773
774# define isALNUMC_A(c) _generic_isCC_A(c, _CC_ALNUMC)
775# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
776# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
777# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
778# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
779# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
780# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
781# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
782# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
783# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
784# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
785# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
786# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
787# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
788# define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST))
789
fb9bfbf7
KW
790 /* Either participates in a fold with a character above 255, or is a
791 * multi-char fold */
430b7c70
KW
792# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask(_CC_NONLATIN1_FOLD)))
793
4eeeb416 794# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
795# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
796 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
797# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
798 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
9fb1bf9d 799#else /* No perl.h. */
9fb1bf9d
KW
800# ifdef EBCDIC
801# define isALNUMC_A(c) (isASCII(c) && isALNUMC(c))
802# define isALPHA_A(c) (isASCII(c) && isALPHA(c))
803# define isBLANK_A(c) (isASCII(c) && isBLANK(c))
804# define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
805# define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
806# define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
807# define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
808# define isLOWER_A(c) (isASCII(c) && isLOWER(c))
809# define isPRINT_A(c) (isASCII(c) && isPRINT(c))
810# define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
811# define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
812# define isSPACE_A(c) (isASCII(c) && isSPACE(c))
813# define isUPPER_A(c) (isASCII(c) && isUPPER(c))
814# define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
815# define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
816# else /* ASCII platform, no perl.h */
817# define isALNUMC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
818# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
819# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
926bb54c 820# define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
2181da29 821# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
9fb1bf9d
KW
822# define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
823# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
824# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
825# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
826# define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
926bb54c
KW
827# define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) \
828 || ((c) >= 58 && (c) <= 64) \
829 || ((c) >= 91 && (c) <= 96) \
830 || ((c) >= 123 && (c) <= 126))
831# define isSPACE_A(c) ((c) == ' ' \
832 || (c) == '\t' \
833 || (c) == '\n' \
834 || (c) =='\r' \
835 || (c) == '\f')
2181da29 836# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
9fb1bf9d 837# define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
926bb54c
KW
838# define isXDIGIT_A(c) (isDIGIT_A(c) \
839 || ((c) >= 'a' && (c) <= 'f') \
840 || ((c) <= 'F' && (c) >= 'A'))
9fb1bf9d
KW
841# endif
842#endif /* ASCII range definitions */
ec79cfbb
KW
843
844/* Latin1 definitions */
8a58bdcf 845#ifdef H_PERL
f4cdb42c
KW
846# define isALNUMC_L1(c) _generic_isCC(c, _CC_ALNUMC)
847# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
848# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
e5dcd934 849
8a58bdcf 850/* continuation character for legal NAME in \N{NAME} */
4eeeb416 851# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
e5dcd934 852
f4cdb42c
KW
853# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
854# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
855# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
856# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
857# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
858# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
859# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
860# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
861# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
862# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
8a58bdcf
KW
863#else /* No access to perl.h. Only a few provided here, just in case needed
864 * for backwards compatibility */
865 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
866 * >= AA test to speed up ASCII-only tests at the expense of the others */
867# define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
868 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
926bb54c 869 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
8a58bdcf
KW
870 || NATIVE_TO_UNI((U8) c) == 0xAA \
871 || NATIVE_TO_UNI((U8) c) == 0xB5 \
872 || NATIVE_TO_UNI((U8) c) == 0xBA)))
926bb54c
KW
873# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
874 || (c) == ' ' \
875 || (c) == '-' \
876 || (c) == '(' \
877 || (c) == ')' \
878 || (c) == ':' \
879 || NATIVE_TO_UNI((U8) c) == 0xA0)
8a58bdcf
KW
880#endif
881
882/* Macros for backwards compatibility and for completeness when the ASCII and
883 * Latin1 values are identical */
ec79cfbb
KW
884#define isALNUM(c) isWORDCHAR(c)
885#define isALNUMU(c) isWORDCHAR_L1(c)
8a58bdcf
KW
886#define isALPHAU(c) isALPHA_L1(c)
887#define isDIGIT_L1(c) isDIGIT_A(c)
888#define isOCTAL(c) isOCTAL_A(c)
889#define isOCTAL_L1(c) isOCTAL_A(c)
890#define isXDIGIT_L1(c) isXDIGIT_A(c)
891
892/* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
893 * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
894 * therefore can use native functions. For those where C89 doesn't define a
895 * function, use our function, assuming that the EBCDIC code page is isomorphic
896 * with Latin1, which the three currently recognized by Perl are. Some libc's
897 * have an isblank(), but it's not guaranteed. */
9d116dd7 898#ifdef EBCDIC
b8c5462f 899# define isALNUMC(c) isalnum(c)
271b5cb3 900# define isALPHA(c) isalpha(c)
0bc70f7f 901# define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
b8c5462f 902# define isCNTRL(c) iscntrl(c)
9916188f 903# define isDIGIT(c) isdigit(c)
b8c5462f 904# define isGRAPH(c) isgraph(c)
8a58bdcf 905# define isIDFIRST(c) (isALPHA(c) || (c) == '_')
7d79683e 906# define isLOWER(c) islower(c)
9d116dd7 907# define isPRINT(c) isprint(c)
56e4b611 908# define isPSXSPC(c) isspace(c)
b8c5462f 909# define isPUNCT(c) ispunct(c)
f1f70107
KW
910# define isSPACE(c) (isPSXSPC(c) /* && (c) != '\v' (Experimentally making
911 these macros identical) */)
7d79683e 912# define isUPPER(c) isupper(c)
b8c5462f 913# define isXDIGIT(c) isxdigit(c)
8a58bdcf 914# define isWORDCHAR(c) (isalnum(c) || (c) == '_')
9d116dd7 915# define toLOWER(c) tolower(c)
7d79683e 916# define toUPPER(c) toupper(c)
ec79cfbb 917#else /* Not EBCDIC: ASCII-only matching */
926bb54c 918# define isALNUMC(c) isALNUMC_A(c) /* Mnemonic: "C's alnum" = alpha + digit */
271b5cb3 919# define isALPHA(c) isALPHA_A(c)
ec79cfbb
KW
920# define isBLANK(c) isBLANK_A(c)
921# define isCNTRL(c) isCNTRL_A(c)
922# define isDIGIT(c) isDIGIT_A(c)
923# define isGRAPH(c) isGRAPH_A(c)
8a58bdcf 924# define isIDFIRST(c) isIDFIRST_A(c)
ec79cfbb
KW
925# define isLOWER(c) isLOWER_A(c)
926# define isPRINT(c) isPRINT_A(c)
927# define isPSXSPC(c) isPSXSPC_A(c)
928# define isPUNCT(c) isPUNCT_A(c)
929# define isSPACE(c) isSPACE_A(c)
930# define isUPPER(c) isUPPER_A(c)
8a58bdcf 931# define isWORDCHAR(c) isWORDCHAR_A(c)
ec79cfbb 932# define isXDIGIT(c) isXDIGIT_A(c)
00f254e2 933
8a58bdcf
KW
934 /* ASCII casing. These could also be written as
935 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
936 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
937 which uses table lookup and mask instead of subtraction. (This would
938 work because the _MOD does not apply in the ASCII range) */
9d116dd7 939# define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
7d79683e 940# define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
9d116dd7 941#endif
bbce6d69 942
1a0901db
KW
943
944/* Use table lookup for speed; return error character for input
945 * out-of-range */
946#define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \
947 ? UNI_TO_NATIVE(PL_latin1_lc[ \
948 NATIVE_TO_UNI( (U8) (c)) ]) \
949 : UNICODE_REPLACEMENT)
950/* Modified uc. Is correct uc except for three non-ascii chars which are
951 * all mapped to one of them, and these need special handling; error
952 * character for input out-of-range */
953#define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \
954 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \
955 NATIVE_TO_UNI( (U8) (c)) ]) \
956 : UNICODE_REPLACEMENT)
957
bbce6d69
PP
958#ifdef USE_NEXT_CTYPE
959
509fb054
KW
960# define isALNUMC_LC(c) NXIsAlNum((unsigned int)(c))
961# define isALNUM_LC(c) isWORDCHAR_LC(c)
ff68c719 962# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 963# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 964# define isBLANK_LC(c) isBLANK((unsigned int)(c))
b8c5462f 965# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
509fb054 966# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
b8c5462f 967# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
509fb054
KW
968# define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
969# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
ff68c719 970# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 971# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
509fb054
KW
972# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
973# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
974# define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
e712593e 975# define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
ff68c719 976# define toLOWER_LC(c) NXToLower((unsigned int)(c))
509fb054 977# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
bbce6d69
PP
978
979#else /* !USE_NEXT_CTYPE */
b8c5462f 980
bbce6d69
PP
981# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
982
76fde96a 983/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
851131d3 984
509fb054 985# define isALNUMC_LC(c) (FITS_IN_8_BITS(c) && isalnum((unsigned char)(c)))
ca5b5914 986# define isALNUM_LC(c) isWORDCHAR_LC(c)
76fde96a 987# define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
07315176 988# ifdef HAS_ISASCII
76fde96a 989# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
07315176 990# else
76fde96a 991# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
07315176 992# endif
67addccf 993# ifdef HAS_ISBLANK
76fde96a 994# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
67addccf 995# else
76fde96a 996# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
67addccf 997# endif
76fde96a 998# define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
509fb054 999# define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
76fde96a 1000# define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
509fb054
KW
1001# define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1002 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1003# define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
76fde96a
KW
1004# define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1005# define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
509fb054
KW
1006# define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1007# define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1008# define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1009 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
e712593e 1010# define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
76fde96a 1011# define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? tolower((unsigned char)(c)) : (c))
509fb054 1012# define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? toupper((unsigned char)(c)) : (c))
bbce6d69
PP
1013
1014# else
1015
509fb054 1016# define isALNUMC_LC(c) (isascii(c) && isalnum(c))
ca5b5914 1017# define isALNUM_LC(c) isWORDCHAR_LC(c)
bbce6d69 1018# define isALPHA_LC(c) (isascii(c) && isalpha(c))
07315176 1019# define isASCII_LC(c) isascii(c)
67addccf
KW
1020# ifdef HAS_ISBLANK
1021# define isBLANK_LC(c) (isascii(c) && isblank(c))
1022# else
3e66ab74 1023# define isBLANK_LC(c) isBLANK_A(c)
67addccf 1024# endif
b8c5462f 1025# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1026# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1027# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1028# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1029# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1030# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1031# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1032# define isSPACE_LC(c) (isascii(c) && isspace(c))
1033# define isUPPER_LC(c) (isascii(c) && isupper(c))
1034# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1035# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
76fde96a 1036# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1037# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
bbce6d69
PP
1038
1039# endif
a0d0e21e 1040#endif /* USE_NEXT_CTYPE */
55204971 1041
13380643 1042#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1043
66c17564
KW
1044/* For internal core Perl use only. If the input is Latin1, use the Latin1
1045 * macro; otherwise use the function. Won't compile if 'c' isn't unsigned, as
1046 * won't match function prototype. The macros do bounds checking, so have
1047 * duplicate checks here, so could create versions of the macros that don't,
1048 * but experiments show that gcc optimizes them out anyway. */
1049
1050/* Note that all ignore 'use bytes' */
1051#define _generic_uni(classnum, function, c) ((c) < 256 \
aea0c08c
KW
1052 ? _generic_isCC(c, classnum) \
1053 : function(c))
509fb054 1054#define isALNUMC_uni(c) _generic_uni(_CC_ALNUMC, is_uni_alnumc, c)
926bb54c 1055#define isALNUM_uni(c) isWORDCHAR_uni(c)
926bb54c 1056#define isALPHA_uni(c) _generic_uni(_CC_ALPHA, is_uni_alpha, c)
509fb054
KW
1057#define isASCII_uni(c) isASCII(c)
1058#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1059#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
926bb54c 1060#define isDIGIT_uni(c) _generic_uni(_CC_DIGIT, is_uni_digit, c)
509fb054
KW
1061#define isGRAPH_uni(c) _generic_uni(_CC_GRAPH, is_uni_graph, c)
1062#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
926bb54c 1063#define isLOWER_uni(c) _generic_uni(_CC_LOWER, is_uni_lower, c)
509fb054 1064#define isPRINT_uni(c) _generic_uni(_CC_PRINT, is_uni_print, c)
aea0c08c 1065
509fb054
KW
1066/* Posix and regular space are identical above Latin1 */
1067#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1068
509fb054
KW
1069#define isPUNCT_uni(c) _generic_uni(_CC_PUNCT, is_uni_punct, c)
1070#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
1071#define isUPPER_uni(c) _generic_uni(_CC_UPPER, is_uni_upper, c)
1072#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
1073#define isWORDCHAR_uni(c) _generic_uni(_CC_WORDCHAR, is_uni_alnum, c)
1074#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1075
b8d68ded 1076#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1077#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1078#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1079#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1080
69788afd
KW
1081#define _gnrc_is_LC_uvchr(latin1, above_latin1, c) \
1082 (c < 256 ? latin1(c) : above_latin1(NATIVE_TO_UNI(c)))
509fb054 1083#define isALNUMC_LC_uvchr(c) _gnrc_is_LC_uvchr(isALNUMC_LC, is_uni_alnumc_lc, c)
ca5b5914 1084#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
69788afd 1085#define isALPHA_LC_uvchr(c) _gnrc_is_LC_uvchr(isALPHA_LC, is_uni_alpha_lc, c)
509fb054
KW
1086#define isASCII_LC_uvchr(c) isASCII_LC(c)
1087#define isBLANK_LC_uvchr(c) _gnrc_is_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1088#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
509fb054 1089#define isDIGIT_LC_uvchr(c) _gnrc_is_LC_uvchr(isDIGIT_LC, is_uni_digit_lc, c)
69788afd 1090#define isGRAPH_LC_uvchr(c) _gnrc_is_LC_uvchr(isGRAPH_LC, is_uni_graph_lc, c)
509fb054
KW
1091#define isIDFIRST_LC_uvchr(c) _gnrc_is_LC_uvchr(isIDFIRST_LC, \
1092 is_uni_idfirst_lc, c)
1093#define isLOWER_LC_uvchr(c) _gnrc_is_LC_uvchr(isLOWER_LC, is_uni_lower_lc, c)
69788afd 1094#define isPRINT_LC_uvchr(c) _gnrc_is_LC_uvchr(isPRINT_LC, is_uni_print_lc, c)
509fb054
KW
1095#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1096 space under locale */
69788afd 1097#define isPUNCT_LC_uvchr(c) _gnrc_is_LC_uvchr(isPUNCT_LC, is_uni_punct_lc, c)
509fb054
KW
1098#define isSPACE_LC_uvchr(c) _gnrc_is_LC_uvchr(isSPACE_LC, \
1099 is_XPERLSPACE_cp_high, c)
1100#define isUPPER_LC_uvchr(c) _gnrc_is_LC_uvchr(isUPPER_LC, is_uni_upper_lc, c)
1101#define isWORDCHAR_LC_uvchr(c) _gnrc_is_LC_uvchr(isWORDCHAR_LC, \
1102 is_uni_alnum_lc, c)
e712593e 1103#define isXDIGIT_LC_uvchr(c) _gnrc_is_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1104
a0ed51b3 1105
36b00f93 1106#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1107
66c17564 1108/* For internal core Perl use only. If the input is in the Latin1 range, use
252810d0
KW
1109 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1110 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1111 * fact that ASCII characters have the same representation whether utf8 or not.
1112 * Note that it assumes that the utf8 has been validated, and ignores 'use
1113 * bytes' */
1114#define _generic_utf8_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1115 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1116 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c
KW
1117 ? _generic_isCC( \
1118 TWO_BYTE_UTF8_TO_UNI(*(p), \
1119 *((p)+1 )), \
1120 classnum) \
252810d0
KW
1121 : utf8)
1122/* Like the above, but calls 'function(p)' to get the utf8 value */
1123#define _generic_utf8(classnum, function, p) \
1124 _generic_utf8_utf8(classnum, p, function(p))
c8362b00 1125
509fb054 1126#define isALNUMC_utf8(p) _generic_utf8(_CC_ALNUMC, is_utf8_alnumc, p)
252810d0 1127#define isALNUM_utf8(p) isWORDCHAR_utf8(p) /* back compat */
509fb054
KW
1128#define isALPHA_utf8(p) _generic_utf8(_CC_ALPHA, is_utf8_alpha, p)
1129#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1130 utf8, the non-utf8 macro works
1131 */
1132#define isBLANK_utf8(p) _generic_utf8(_CC_BLANK, is_HORIZWS_high, p)
1133#define isCNTRL_utf8(p) _generic_utf8_utf8(_CC_CNTRL, p, 0)
1134#define isDIGIT_utf8(p) _generic_utf8(_CC_DIGIT, is_utf8_digit, p)
1135#define isGRAPH_utf8(p) _generic_utf8(_CC_GRAPH, is_utf8_graph, p)
1136#define isIDCONT_utf8(p) _generic_utf8(_CC_WORDCHAR, is_utf8_xidcont, p)
e5dcd934 1137
c11ff943
KW
1138/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1139 * IDFIRST is an alnum. See
c8362b00 1140 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1141 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1142 * This used to be not the XID version, but we decided to go with the more
1143 * modern Unicode definition */
252810d0 1144#define isIDFIRST_utf8(p) _generic_utf8(_CC_IDFIRST, \
509fb054 1145 _is_utf8_perl_idstart, p)
e5dcd934 1146
252810d0 1147#define isLOWER_utf8(p) _generic_utf8(_CC_LOWER, is_utf8_lower, p)
509fb054 1148#define isPRINT_utf8(p) _generic_utf8(_CC_PRINT, is_utf8_print, p)
aea0c08c 1149
509fb054
KW
1150/* Posix and regular space are identical above Latin1 */
1151#define isPSXSPC_utf8(p) _generic_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1152
252810d0 1153#define isPUNCT_utf8(p) _generic_utf8(_CC_PUNCT, is_utf8_punct, p)
509fb054
KW
1154#define isSPACE_utf8(p) _generic_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
1155#define isUPPER_utf8(p) _generic_utf8(_CC_UPPER, is_utf8_upper, p)
1156#define isVERTWS_utf8(p) _generic_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
1157#define isWORDCHAR_utf8(p) _generic_utf8(_CC_WORDCHAR, is_utf8_alnum, p)
4ac6419d 1158#define isXDIGIT_utf8(p) _generic_utf8(_CC_XDIGIT, is_XDIGIT_high, p)
a0ed51b3 1159
509fb054
KW
1160#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1161#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1162#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1163
9201a572
KW
1164/* For internal core Perl use only. If the input is in the Latin1 range, use
1165 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1166 * use the value given by the 'utf8' parameter. This relies on the fact that
1167 * ASCII characters have the same representation whether utf8 or not. Note
1168 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
1169#define _generic_LC_utf8_utf8(macro, p, utf8) \
1170 (UTF8_IS_INVARIANT(*(p)) \
1171 ? macro(*(p)) \
1172 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1173 ? macro(TWO_BYTE_UTF8_TO_UNI(*(p), *((p)+1))) \
1174 : utf8)
1175
1176#define _generic_LC_utf8(macro, utf8_func, p) \
1177 _generic_LC_utf8_utf8(macro, p, utf8_func(p))
1178
509fb054 1179#define isALNUMC_LC_utf8(p) _generic_LC_utf8(isALNUMC_LC, is_utf8_alnumc, p)
ca5b5914 1180#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
9201a572 1181#define isALPHA_LC_utf8(p) _generic_LC_utf8(isALPHA_LC, is_utf8_alpha, p)
509fb054 1182#define isASCII_LC_utf8(p) isASCII_LC(*p)
9201a572 1183#define isBLANK_LC_utf8(p) _generic_LC_utf8(isBLANK_LC, is_HORIZWS_high, p)
9201a572 1184#define isCNTRL_LC_utf8(p) _generic_LC_utf8_utf8(isCNTRL_LC, p, 0)
509fb054 1185#define isDIGIT_LC_utf8(p) _generic_LC_utf8(isDIGIT_LC, is_utf8_digit, p)
9201a572 1186#define isGRAPH_LC_utf8(p) _generic_LC_utf8(isGRAPH_LC, is_utf8_graph, p)
509fb054
KW
1187#define isIDFIRST_LC_utf8(p) _generic_LC_utf8(isIDFIRST_LC, \
1188 _is_utf8_perl_idstart, p)
1189#define isLOWER_LC_utf8(p) _generic_LC_utf8(isLOWER_LC, is_utf8_lower, p)
9201a572 1190#define isPRINT_LC_utf8(p) _generic_LC_utf8(isPRINT_LC, is_utf8_print, p)
509fb054
KW
1191#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1192 space under locale */
9201a572 1193#define isPUNCT_LC_utf8(p) _generic_LC_utf8(isPUNCT_LC, is_utf8_punct, p)
509fb054
KW
1194#define isSPACE_LC_utf8(p) _generic_LC_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1195#define isUPPER_LC_utf8(p) _generic_LC_utf8(isUPPER_LC, is_utf8_upper, p)
1196#define isWORDCHAR_LC_utf8(p) _generic_LC_utf8(isWORDCHAR_LC, is_utf8_alnum, p)
e712593e 1197#define isXDIGIT_LC_utf8(p) _generic_LC_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1198
6ae9f32a 1199/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1200 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1201 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1202 * that is EOT on ASCII; ST on EBCDIC */
6ae9f32a 1203# define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64)
bbce6d69 1204
dea28490
JJ
1205/* Line numbers are unsigned, 32 bits. */
1206typedef U32 line_t;
e5dcd934 1207#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1208
91152fc1
DG
1209/* Helpful alias for version prescan */
1210#define is_LAX_VERSION(a,b) \
1211 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1212
1213#define is_STRICT_VERSION(a,b) \
1214 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1215
1216#define BADVERSION(a,b,c) \
1217 if (b) { \
1218 *b = c; \
1219 } \
1220 return a;
8c52afec 1221
95a59cab
YO
1222#define READ_XDIGIT(s) (isALPHA(*(s)) ? ((*(s)++ + 9) & 0xf) : (*(s)++ & 0xf))
1223
8e84507e 1224/*
ccfc67b7
JH
1225=head1 Memory Management
1226
a02a5408 1227=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1228The XSUB-writer's interface to the C C<malloc> function.
1229
c5008215
JC
1230In 5.9.3, Newx() and friends replace the older New() API, and drops
1231the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1232themselves. This aid has been superseded by a new build option,
d10b4965 1233PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1234there for use in XS modules supporting older perls.
1235
a02a5408 1236=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1237The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1238cast. See also C<Newx>.
954c1994 1239
a02a5408 1240=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1241The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1242memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1243
954c1994
GS
1244=for apidoc Am|void|Renew|void* ptr|int nitems|type
1245The XSUB-writer's interface to the C C<realloc> function.
1246
1247=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1248The XSUB-writer's interface to the C C<realloc> function, with
1249cast.
1250
49b8b560 1251=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1252The XSUB-writer's interface to the C C<free> function.
1253
1254=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1255The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1256source, C<dest> is the destination, C<nitems> is the number of items, and
1257C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1258
e90e2364
NC
1259=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1260Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1261optimise.
1262
954c1994
GS
1263=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1264The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1265source, C<dest> is the destination, C<nitems> is the number of items, and
1266C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1267
e90e2364
NC
1268=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1269
1270Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1271optimise.
1272
954c1994
GS
1273=for apidoc Am|void|Zero|void* dest|int nitems|type
1274
1275The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1276destination, C<nitems> is the number of items, and C<type> is the type.
1277
e90e2364
NC
1278=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1279
1280Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1281optimise.
1282
da5d8dbb 1283=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1284This is an architecture-independent macro to copy one structure to another.
954c1994 1285
7e337ee0
JH
1286=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1287
1288Fill up memory with a byte pattern (a byte repeated over and over
1289again) that hopefully catches attempts to access uninitialized memory.
1290
1291=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1292
1293PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1294
1c12ffb4 1295=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1296
1297PoisonWith(0xEF) for catching access to freed memory.
1298
9965345d
JH
1299=for apidoc Am|void|Poison|void* dest|int nitems|type
1300
7e337ee0 1301PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1302
1303=cut */
954c1994 1304
561b68a9
SH
1305/* Maintained for backwards-compatibility only. Use newSV() instead. */
1306#ifndef PERL_CORE
ff06c60c 1307#define NEWSV(x,len) newSV(len)
561b68a9 1308#endif
ff06c60c 1309
19a94d75
JH
1310#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1311
2720f793
JH
1312/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1313 * overly eager compilers that will bleat about e.g.
1314 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1315#ifdef PERL_MALLOC_WRAP
f019c49e
DD
1316#define MEM_WRAP_CHECK(n,t) \
1317 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
27d5b266 1318#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1319 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1320#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1321
f019c49e 1322#define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (croak_memory_wrap(),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266
JH
1323#else
1324
410319be
NC
1325#define MEM_WRAP_CHECK(n,t)
1326#define MEM_WRAP_CHECK_1(n,t,a)
1327#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1328#define MEM_WRAP_CHECK_(n,t)
1329
6fff79ce 1330#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1331
1936d2a7 1332#endif
8b44ba4c 1333
fe4f188c 1334#ifdef PERL_MEM_LOG
46c6c7e2 1335/*
9f653bb5 1336 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1337 * go through functions, which are handy for debugging breakpoints, but
1338 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1339 * line number, and C function name if available) passed in. This info can
1340 * then be used for logging the calls, for which one gets a sample
73d1d973 1341 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1342 *
46c6c7e2 1343 * Known problems:
94e892a6 1344 * - not all memory allocs get logged, only those
46c6c7e2 1345 * that go through Newx() and derivatives (while all
94e892a6 1346 * Safefrees do get logged)
46c6c7e2
JH
1347 * - __FILE__ and __LINE__ do not work everywhere
1348 * - __func__ or __FUNCTION__ even less so
1349 * - I think more goes on after the perlio frees but
1350 * the thing is that STDERR gets closed (as do all
1351 * the file descriptors)
1352 * - no deeper calling stack than the caller of the Newx()
1353 * or the kind, but do I look like a C reflection/introspection
1354 * utility to you?
1355 * - the function prototypes for the logging functions
1356 * probably should maybe be somewhere else than handy.h
1357 * - one could consider inlining (macrofying) the logging
1358 * for speed, but I am too lazy
1359 * - one could imagine recording the allocations in a hash,
1360 * (keyed by the allocation address?), and maintain that
1361 * through reallocs and frees, but how to do that without
1362 * any News() happening...?
73d1d973 1363 * - lots of -Ddefines to get useful/controllable output
b953482e 1364 * - lots of ENV reads
46c6c7e2
JH
1365 */
1366
12754f92 1367PERL_EXPORT_C Malloc_t Perl_mem_log_alloc(const UV n, const UV typesize, const char *type_name, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1368
12754f92 1369PERL_EXPORT_C Malloc_t Perl_mem_log_realloc(const UV n, const UV typesize, const char *type_name, Malloc_t oldalloc, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1370
12754f92 1371PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1372
0b0ab801 1373# ifdef PERL_CORE
73d1d973 1374# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1375enum mem_log_type {
1376 MLT_ALLOC,
1377 MLT_REALLOC,
d7a2c63c
MHM
1378 MLT_FREE,
1379 MLT_NEW_SV,
1380 MLT_DEL_SV
0b0ab801
MHM
1381};
1382# endif
12754f92 1383# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1384void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1385void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1386# endif
0b0ab801
MHM
1387# endif
1388
fe4f188c
JH
1389#endif
1390
1391#ifdef PERL_MEM_LOG
d1401ee9
MHM
1392#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1393#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1394#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1395#endif
1396
1397#ifndef MEM_LOG_ALLOC
1398#define MEM_LOG_ALLOC(n,t,a) (a)
1399#endif
1400#ifndef MEM_LOG_REALLOC
1401#define MEM_LOG_REALLOC(n,t,v,a) (a)
1402#endif
1403#ifndef MEM_LOG_FREE
1404#define MEM_LOG_FREE(a) (a)
1405#endif
1406
d1401ee9
MHM
1407#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1408#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1409#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1410
1411#ifndef PERL_CORE
a02a5408
JC
1412/* pre 5.9.x compatibility */
1413#define New(x,v,n,t) Newx(v,n,t)
1414#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1415#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1416#endif
a02a5408 1417
ff68c719 1418#define Renew(v,n,t) \
d1401ee9 1419 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1420#define Renewc(v,n,t,c) \
d1401ee9 1421 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1422
1423#ifdef PERL_POISON
1424#define Safefree(d) \
06c0cc96 1425 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1426#else
fe4f188c 1427#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1428#endif
55497cff 1429
8b44ba4c
NC
1430#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1431#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1432#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1433
8b44ba4c
NC
1434#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1435#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1436#ifdef HAS_MEMSET
8b44ba4c 1437#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1438#else
8b44ba4c
NC
1439/* Using bzero(), which returns void. */
1440#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1441#endif
1442
7e337ee0
JH
1443#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1444#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1445#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1446#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1447
2304df62 1448#ifdef USE_STRUCT_COPY
ff68c719 1449#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1450#else
1451#define StructCopy(s,d,t) Copy(s,d,1,t)
1452#endif
2cc61e15 1453
622913ab 1454#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1455#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1456
2cc61e15
DD
1457#ifdef NEED_VA_COPY
1458# ifdef va_copy
1459# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1460# else
a1866d1b
JH
1461# if defined(__va_copy)
1462# define Perl_va_copy(s, d) __va_copy(d, s)
1463# else
1464# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1465# endif
2cc61e15
DD
1466# endif
1467#endif
1468
472d47bc
SB
1469/* convenience debug macros */
1470#ifdef USE_ITHREADS
1471#define pTHX_FORMAT "Perl interpreter: 0x%p"
1472#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1473#define pTHX_VALUE_ (void *)my_perl,
1474#define pTHX_VALUE (void *)my_perl
1475#define pTHX__VALUE_ ,(void *)my_perl,
1476#define pTHX__VALUE ,(void *)my_perl
472d47bc 1477#else
3609ea0d 1478#define pTHX_FORMAT
472d47bc 1479#define pTHX__FORMAT
3609ea0d 1480#define pTHX_VALUE_
472d47bc 1481#define pTHX_VALUE
3609ea0d 1482#define pTHX__VALUE_
472d47bc
SB
1483#define pTHX__VALUE
1484#endif /* USE_ITHREADS */
3609ea0d 1485
2acdbac1
NC
1486/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1487 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1488 CPAN::Unpack show any users outside the core. */
1489#ifdef PERL_CORE
d1d15184 1490# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1491#endif
1492
e9a8c099
MHM
1493/*
1494 * Local variables:
1495 * c-indentation-style: bsd
1496 * c-basic-offset: 4
14d04a33 1497 * indent-tabs-mode: nil
e9a8c099
MHM
1498 * End:
1499 *
14d04a33 1500 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1501 */