This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Revert part of 2efab60d9
[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
9d745869
KW
11#ifndef HANDY_H /* Guard against nested #inclusion */
12#define HANDY_H
13
85e6fe83 14#if !defined(__STDC__)
378cc40b
LW
15#ifdef NULL
16#undef NULL
17#endif
a687059c 18# define NULL 0
85e6fe83
LW
19#endif
20
24792b8d
NC
21#ifndef PERL_CORE
22# define Null(type) ((type)NULL)
954c1994
GS
23
24/*
ccfc67b7 25=head1 Handy Values
954c1994 26
3609ea0d 27=for apidoc AmU||Nullch
24792b8d 28Null character pointer. (No longer available when C<PERL_CORE> is defined.)
2307c6d0 29
954c1994 30=for apidoc AmU||Nullsv
24792b8d 31Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
32
33=cut
34*/
35
24792b8d
NC
36# define Nullch Null(char*)
37# define Nullfp Null(PerlIO*)
38# define Nullsv Null(SV*)
39#endif
8d063cd8 40
641d3f0b
PP
41#ifdef TRUE
42#undef TRUE
43#endif
44#ifdef FALSE
45#undef FALSE
46#endif
47#define TRUE (1)
48#define FALSE (0)
49
cf3f0ffb
DM
50/* The MUTABLE_*() macros cast pointers to the types shown, in such a way
51 * (compiler permitting) that casting away const-ness will give a warning;
52 * e.g.:
53 *
54 * const SV *sv = ...;
55 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
56 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
57 */
58
b1bc3f34 59#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 60# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
61#else
62# define MUTABLE_PTR(p) ((void *) (p))
63#endif
64
a062e10d 65#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 66#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 67#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 68#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 69#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 70#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 71
f789f6a4 72#if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR)
bd31be4b
NC
73# include <stdbool.h>
74# ifndef HAS_BOOL
75# define HAS_BOOL 1
76# endif
77#endif
78
8e84507e 79/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
80 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
81 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 82 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
83 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
84 g++ can be identified by __GNUG__.
85 Andy Dougherty February 2000
5d94fbed 86*/
3609ea0d 87#ifdef __GNUG__ /* GNU g++ has bool built-in */
f789f6a4 88# ifndef PERL_BOOL_AS_CHAR
5d94fbed 89# ifndef HAS_BOOL
c1d22f6b 90# define HAS_BOOL 1
5d94fbed 91# endif
f789f6a4 92# endif
5d94fbed
AD
93#endif
94
641d3f0b
PP
95/* The NeXT dynamic loader headers will not build with the bool macro
96 So declare them now to clear confusion.
97*/
8f1f23e8 98#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
99# undef FALSE
100# undef TRUE
101 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
102# define ENUM_BOOL 1
103# ifndef HAS_BOOL
104# define HAS_BOOL 1
105# endif /* !HAS_BOOL */
8f1f23e8 106#endif /* NeXT || __NeXT__ */
641d3f0b 107
5d94fbed 108#ifndef HAS_BOOL
f789f6a4
FC
109# ifdef bool
110# undef bool
111# endif
70d5cb32 112# define bool char
c1d22f6b 113# define HAS_BOOL 1
a687059c 114#endif
0d3e774c 115
236415e2
KW
116/* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
117 * defined as char for example, then the cast from int is
118 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
119 * AIX */
18f5643b 120#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 121
46c6c7e2 122/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
123 * XXX Should really be a Configure probe, with HAS__FUNCTION__
124 * and FUNCTION__ as results.
125 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
126#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
127# define FUNCTION__ __func__
128#else
9ea40801 129# if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 130 (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
131# define FUNCTION__ ""
132# else
133# define FUNCTION__ __FUNCTION__ /* Common extension. */
134# endif
135#endif
136
27d4fb96
PP
137/* XXX A note on the perl source internal type system. The
138 original intent was that I32 be *exactly* 32 bits.
139
140 Currently, we only guarantee that I32 is *at least* 32 bits.
141 Specifically, if int is 64 bits, then so is I32. (This is the case
142 for the Cray.) This has the advantage of meshing nicely with
143 standard library calls (where we pass an I32 and the library is
144 expecting an int), but the disadvantage that an I32 is not 32 bits.
145 Andy Dougherty August 1996
24fef2a7 146
dc45a647
MB
147 There is no guarantee that there is *any* integral type with
148 exactly 32 bits. It is perfectly legal for a system to have
149 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 150
dc45a647
MB
151 Similarly, there is no guarantee that I16 and U16 have exactly 16
152 bits.
693762b4 153
8e84507e
NIS
154 For dealing with issues that may arise from various 32/64-bit
155 systems, we will ask Configure to check out
8175356b 156
3609ea0d
JH
157 SHORTSIZE == sizeof(short)
158 INTSIZE == sizeof(int)
159 LONGSIZE == sizeof(long)
dc45a647 160 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 161 PTRSIZE == sizeof(void *)
dc45a647
MB
162 DOUBLESIZE == sizeof(double)
163 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 164
27d4fb96
PP
165*/
166
69512466
JH
167#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
168# include <inttypes.h>
dd0eed91
JH
169# ifdef INT32_MIN_BROKEN
170# undef INT32_MIN
171# define INT32_MIN (-2147483647-1)
172# endif
173# ifdef INT64_MIN_BROKEN
174# undef INT64_MIN
175# define INT64_MIN (-9223372036854775807LL-1)
176# endif
69512466
JH
177#endif
178
8175356b
JH
179typedef I8TYPE I8;
180typedef U8TYPE U8;
181typedef I16TYPE I16;
182typedef U16TYPE U16;
183typedef I32TYPE I32;
184typedef U32TYPE U32;
6b8eaf93
JH
185#ifdef PERL_CORE
186# ifdef HAS_QUAD
8175356b
JH
187typedef I64TYPE I64;
188typedef U64TYPE U64;
6b8eaf93
JH
189# endif
190#endif /* PERL_CORE */
8175356b 191
69512466 192#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
80524f33
JL
193# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
194# define PeRl_INT64_C(c) CAT2(c,LL)
195# define PeRl_UINT64_C(c) CAT2(c,ULL)
196# else
52865553
SH
197# if QUADKIND == QUAD_IS___INT64
198# define PeRl_INT64_C(c) CAT2(c,I64)
199# define PeRl_UINT64_C(c) CAT2(c,UI64)
69512466 200# else
52865553
SH
201# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
202# define PeRl_INT64_C(c) CAT2(c,L)
203# define PeRl_UINT64_C(c) CAT2(c,UL)
69512466 204# else
80524f33
JL
205# define PeRl_INT64_C(c) ((I64TYPE)(c))
206# define PeRl_UINT64_C(c) ((U64TYPE)(c))
69512466
JH
207# endif
208# endif
e8c95190 209# endif
80524f33
JL
210# ifndef UINT64_C
211# define UINT64_C(c) PeRl_UINT64_C(c)
212# endif
213# ifndef INT64_C
214# define INT64_C(c) PeRl_INT64_C(c)
215# endif
e8c95190 216#endif
dc750f2e 217
d8668976 218#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 219
5ff3f7a4
GS
220/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
221 Please search CHAR_MAX in perl.h for further details. */
222#define U8_MAX UINT8_MAX
223#define U8_MIN UINT8_MIN
224
5ff3f7a4
GS
225#define I16_MAX INT16_MAX
226#define I16_MIN INT16_MIN
227#define U16_MAX UINT16_MAX
228#define U16_MIN UINT16_MIN
229
5ff3f7a4
GS
230#define I32_MAX INT32_MAX
231#define I32_MIN INT32_MIN
0e983133
GS
232#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
233# define U32_MAX UINT32_MAX
234#else
235# define U32_MAX 4294967295U
236#endif
5ff3f7a4
GS
237#define U32_MIN UINT32_MIN
238
239#else
240
5c9fa16e
KA
241/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
242 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
243#define U8_MAX PERL_UCHAR_MAX
244#define U8_MIN PERL_UCHAR_MIN
79072805 245
27d4fb96
PP
246#define I16_MAX PERL_SHORT_MAX
247#define I16_MIN PERL_SHORT_MIN
248#define U16_MAX PERL_USHORT_MAX
249#define U16_MIN PERL_USHORT_MIN
79072805 250
c4f23d77 251#if LONGSIZE > 4
27d4fb96
PP
252# define I32_MAX PERL_INT_MAX
253# define I32_MIN PERL_INT_MIN
254# define U32_MAX PERL_UINT_MAX
255# define U32_MIN PERL_UINT_MIN
79072805 256#else
27d4fb96
PP
257# define I32_MAX PERL_LONG_MAX
258# define I32_MIN PERL_LONG_MIN
259# define U32_MAX PERL_ULONG_MAX
260# define U32_MIN PERL_ULONG_MIN
79072805
LW
261#endif
262
5ff3f7a4
GS
263#endif
264
58a9a5d5 265/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
266#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
267#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
268#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
269
ff68c719 270#define Ctl(ch) ((ch) & 037)
8d063cd8 271
3fe05580
MHM
272/*
273=head1 SV-Body Allocation
274
275=for apidoc Ama|SV*|newSVpvs|const char* s
276Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
277
84bafc02
NC
278=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
279Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
280pair.
281
3fe05580
MHM
282=for apidoc Ama|SV*|newSVpvs_share|const char* s
283Like C<newSVpvn_share>, but takes a literal string instead of a string/length
284pair and omits the hash parameter.
285
9dcc53ea
Z
286=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
287Like C<sv_catpvn_flags>, but takes a literal string instead of a
288string/length pair.
289
290=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
291Like C<sv_catpvn_nomg>, but takes a literal string instead of a
292string/length pair.
293
7b4e7469 294=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
295Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
296
9dcc53ea
Z
297=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
298Like C<sv_catpvn_mg>, but takes a literal string instead of a
299string/length pair.
300
7b4e7469 301=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
302Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
303
9dcc53ea
Z
304=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
305Like C<sv_setpvn_mg>, but takes a literal string instead of a
306string/length pair.
307
308=for apidoc Am|SV *|sv_setref_pvs|const char* s
309Like C<sv_setref_pvn>, but takes a literal string instead of a
310string/length pair.
311
3fe05580
MHM
312=head1 Memory Management
313
314=for apidoc Ama|char*|savepvs|const char* s
315Like C<savepvn>, but takes a literal string instead of a string/length pair.
316
9dcc53ea
Z
317=for apidoc Ama|char*|savesharedpvs|const char* s
318A version of C<savepvs()> which allocates the duplicate string in memory
319which is shared between threads.
320
3fe05580
MHM
321=head1 GV Functions
322
323=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
324Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
325
326=head1 Hash Manipulation Functions
327
328=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
329Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
330
331=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
332Like C<hv_store>, but takes a literal string instead of a string/length pair
333and omits the hash parameter.
334
510966aa
Z
335=head1 Lexer interface
336
337=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
338
339Like L</lex_stuff_pvn>, but takes a literal string instead of a
340string/length pair.
341
3fe05580
MHM
342=cut
343*/
344
926bb54c
KW
345/* concatenating with "" ensures that only literal strings are accepted as
346 * argument */
26bfd924 347#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 348
926bb54c
KW
349/* note that STR_WITH_LEN() can't be used as argument to macros or functions
350 * that under some configurations might be macros, which means that it requires
351 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
352 */
353
354/* STR_WITH_LEN() shortcuts */
355#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
356#define newSVpvs_flags(str,flags) \
357 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 358#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
359#define sv_catpvs_flags(sv, str, flags) \
360 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
361#define sv_catpvs_nomg(sv, str) \
362 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
363#define sv_catpvs(sv, str) \
364 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
365#define sv_catpvs_mg(sv, str) \
366 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 367#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
368#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
369#define sv_setref_pvs(rv, classname, str) \
370 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 371#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
372#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
373#define gv_stashpvs(str, create) \
374 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
375#define gv_fetchpvs(namebeg, add, sv_type) \
376 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
377#define gv_fetchpvn(namebeg, len, add, sv_type) \
378 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
379#define sv_catxmlpvs(dsv, str, utf8) \
380 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
381#define hv_fetchs(hv,key,lval) \
382 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
383 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
384 : HV_FETCH_JUST_SV, NULL, 0))
385
386#define hv_stores(hv,key,val) \
387 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
388 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 389
510966aa
Z
390#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
391
b96d8cd9
NC
392#define get_cvs(str, flags) \
393 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 394
954c1994 395/*
ccfc67b7
JH
396=head1 Miscellaneous Functions
397
954c1994
GS
398=for apidoc Am|bool|strNE|char* s1|char* s2
399Test two strings to see if they are different. Returns true or
400false.
401
402=for apidoc Am|bool|strEQ|char* s1|char* s2
403Test two strings to see if they are equal. Returns true or false.
404
405=for apidoc Am|bool|strLT|char* s1|char* s2
406Test two strings to see if the first, C<s1>, is less than the second,
407C<s2>. Returns true or false.
408
409=for apidoc Am|bool|strLE|char* s1|char* s2
410Test two strings to see if the first, C<s1>, is less than or equal to the
411second, C<s2>. Returns true or false.
412
413=for apidoc Am|bool|strGT|char* s1|char* s2
414Test two strings to see if the first, C<s1>, is greater than the second,
415C<s2>. Returns true or false.
416
417=for apidoc Am|bool|strGE|char* s1|char* s2
418Test two strings to see if the first, C<s1>, is greater than or equal to
419the second, C<s2>. Returns true or false.
420
421=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
422Test two strings to see if they are different. The C<len> parameter
423indicates the number of bytes to compare. Returns true or false. (A
424wrapper for C<strncmp>).
425
426=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
427Test two strings to see if they are equal. The C<len> parameter indicates
428the number of bytes to compare. Returns true or false. (A wrapper for
429C<strncmp>).
430
431=cut
432*/
433
8d063cd8
LW
434#define strNE(s1,s2) (strcmp(s1,s2))
435#define strEQ(s1,s2) (!strcmp(s1,s2))
436#define strLT(s1,s2) (strcmp(s1,s2) < 0)
437#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
438#define strGT(s1,s2) (strcmp(s1,s2) > 0)
439#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
440#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
441#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 442
36477c24
PP
443#ifdef HAS_MEMCMP
444# define memNE(s1,s2,l) (memcmp(s1,s2,l))
445# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
446#else
447# define memNE(s1,s2,l) (bcmp(s1,s2,l))
448# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
449#endif
450
568a785a 451#define memEQs(s1, l, s2) \
8e6497b6 452 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
453#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
454
bbce6d69
PP
455/*
456 * Character classes.
457 *
458 * Unfortunately, the introduction of locales means that we
459 * can't trust isupper(), etc. to tell the truth. And when
460 * it comes to /\w+/ with tainting enabled, we *must* be able
461 * to trust our character classes.
462 *
463 * Therefore, the default tests in the text of Perl will be
464 * independent of locale. Any code that wants to depend on
465 * the current locale will use the tests that begin with "lc".
466 */
467
2304df62
AD
468#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
469# ifndef CTYPE256
470# define CTYPE256
471# endif
472#endif
473
954c1994 474/*
ccfc67b7
JH
475
476=head1 Character classes
243effed
KW
477This section is about functions (really macros) that classify characters
478into types, such as punctuation versus alphabetic, etc. Most of these are
479analogous to regular expression character classes. (See
480L<perlrecharclass/POSIX Character Classes>.) There are several variants for
481each class. (Not all macros have all variants; each item below lists the
482ones valid for it.) None are affected by C<use bytes>, and only the ones
483with C<LC> in the name are affected by the current locale.
484
485The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
486C<U8>) as input and returns a boolean as to whether or not the character
6aff1f14
KW
487represented by that octet is (or on non-ASCII platforms, corresponds to) an
488ASCII character in the named class based on platform, Unicode, and Perl rules.
489If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 490
6aff1f14
KW
491Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
492with no suffix C<"_A">.
4b9734bf
KW
493
494Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
495onto the platform. That is, the code points that are ASCII are unaffected,
496since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
497as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
498true when called with the code point 0xDF, which is a word character in both
499ASCII and EBCDIC (though it represent different characters in each).
243effed
KW
500
501Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
502point as input. If the code point is larger than 255, Unicode rules are used
503to determine if it is in the character class. For example,
6aff1f14
KW
504C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
505WITH MACRON in Unicode, and is a word character.
243effed
KW
506
507Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
508(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
509classification of just the first (possibly multi-byte) character in the string
510is tested.
243effed
KW
511
512Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but uses
513the C library function that gives the named classification instead of
514hard-coded rules. For example, C<isDIGIT_LC()> returns the result of calling
515C<isdigit()>. This means that the result is based on the current locale, which
516is what C<LC> in the name stands for. FALSE is always returned if the input
517won't fit into an octet.
518
519Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
520returns the same as C<isFOO_LC> for input code points less than 256, and
521returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
522
523Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
524(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
525classification of just the first (possibly multi-byte) character in the string
526is tested.
ccfc67b7 527
954c1994 528=for apidoc Am|bool|isALPHA|char ch
2787a470 529Returns a boolean indicating whether the specified character is an
6aff1f14 530alphabetic character, analogous to C<m/[[:alpha:]]/>.
8a58bdcf 531See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 532C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 533C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 534
15861f94
KW
535=for apidoc Am|bool|isALPHANUMERIC|char ch
536Returns a boolean indicating whether the specified character is a either an
6aff1f14 537alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
15861f94
KW
538See the L<top of this section|/Character classes> for an explanation of variants
539C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 540C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
541C<isALPHANUMERIC_LC_utf8>.
542
8a58bdcf
KW
543=for apidoc Am|bool|isASCII|char ch
544Returns a boolean indicating whether the specified character is one of the 128
243effed 545characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 546On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
547character corresponds to an ASCII character. Variants C<isASCII_A()> and
548C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
549See the L<top of this section|/Character classes> for an explanation of variants
550C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
551C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
552library routine C<isascii()>. In these cases, the variants whose names contain
553C<LC> are the same as the corresponding ones without.
554
d98532ea
KW
555Also note, that because all ASCII characters are UTF-8 invariant (meaning they
556have the exact same representation (always a single byte) whether encoded in
557UTF-8 or not), C<isASCII> will give the correct results when called with any
558byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
559work properly on any string encoded or not in UTF-8.
560
243effed
KW
561=for apidoc Am|bool|isBLANK|char ch
562Returns a boolean indicating whether the specified character is a
6aff1f14 563character considered to be a blank, analogous to C<m/[[:blank:]]/>.
243effed 564See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 565C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
566C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
567platforms do not have the C library routine C<isblank()>. In these cases, the
568variants whose names contain C<LC> are the same as the corresponding ones
569without.
570
571=for apidoc Am|bool|isCNTRL|char ch
572Returns a boolean indicating whether the specified character is a
6aff1f14 573control character, analogous to C<m/[[:cntrl:]]/>.
243effed 574See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 575C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
576C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
577On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 578
954c1994 579=for apidoc Am|bool|isDIGIT|char ch
2787a470 580Returns a boolean indicating whether the specified character is a
6aff1f14 581digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 582Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed 583See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 584C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
585C<isDIGIT_LC_utf8>.
586
587=for apidoc Am|bool|isGRAPH|char ch
588Returns a boolean indicating whether the specified character is a
6aff1f14 589graphic character, analogous to C<m/[[:graph:]]/>.
243effed 590See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 591C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 592C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 593
0c82b6df 594=for apidoc Am|bool|isLOWER|char ch
2787a470 595Returns a boolean indicating whether the specified character is a
6aff1f14 596lowercase character, analogous to C<m/[[:lower:]]/>.
8a58bdcf 597See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 598C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 599C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 600
c99e91e9 601=for apidoc Am|bool|isOCTAL|char ch
2787a470 602Returns a boolean indicating whether the specified character is an
6aff1f14 603octal digit, [0-7].
243effed
KW
604The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
605C<isOCTAL>.
606
607=for apidoc Am|bool|isPUNCT|char ch
608Returns a boolean indicating whether the specified character is a
6aff1f14
KW
609punctuation character, analogous to C<m/[[:punct:]]/>.
610Note that the definition of what is punctuation isn't as
243effed
KW
611straightforward as one might desire. See L<perlrecharclass/POSIX Character
612Classes> for details.
613See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 614C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 615C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 616
0c82b6df 617=for apidoc Am|bool|isSPACE|char ch
2787a470 618Returns a boolean indicating whether the specified character is a
6aff1f14 619whitespace character. This is analogous
398d098a
KW
620to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
621(experimentally), this also matches what C<m/[[:space:]]/> does.
622("Experimentally" means that this change may be backed out in 5.20 or 5.22 if
623field experience indicates that it was unwise.) Prior to 5.18, only the
624locale forms of this macro (the ones with C<LC> in their names) matched
625precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
626in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
627(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
8a58bdcf 628See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 629C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 630C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 631
398d098a
KW
632=for apidoc Am|bool|isPSXSPC|char ch
633(short for Posix Space)
634Starting in 5.18, this is identical (experimentally) in all its forms to the
635corresponding C<isSPACE()> macros. ("Experimentally" means that this change
636may be backed out in 5.20 or 5.22 if field experience indicates that it
637was unwise.)
638The locale forms of this macro are identical to their corresponding
639C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
640non-locale forms differ from their C<isSPACE()> forms only in that the
641C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
642Otherwise they are identical. Thus this macro is analogous to what
643C<m/[[:space:]]/> matches in a regular expression.
644See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 645C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
646C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
647
954c1994 648=for apidoc Am|bool|isUPPER|char ch
2787a470 649Returns a boolean indicating whether the specified character is an
6aff1f14 650uppercase character, analogous to C<m/[[:upper:]]/>.
8a58bdcf 651See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 652C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 653C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 654
243effed 655=for apidoc Am|bool|isPRINT|char ch
8eea39dd 656Returns a boolean indicating whether the specified character is a
6aff1f14 657printable character, analogous to C<m/[[:print:]]/>.
8a58bdcf 658See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 659C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
660C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
661
662=for apidoc Am|bool|isWORDCHAR|char ch
663Returns a boolean indicating whether the specified character is a character
664that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
665in a regular expression. A word character is an alphabetic character, a
666decimal digit, a connecting punctuation character (such as an underscore), or
667a "mark" character that attaches to one of those (like some sort of accent).
668C<isALNUM()> is a synonym provided for backward compatibility, even though a
669word character includes more than the standard C language meaning of
670alphanumeric.
671See the L<top of this section|/Character classes> for an explanation of variants
672C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
673C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
674
675=for apidoc Am|bool|isXDIGIT|char ch
676Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
677digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
678and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
679See the L<top of this section|/Character classes> for an explanation of variants
680C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
681C<isXDIGIT_LC_utf8>.
682
3c3ecf18
KW
683=for apidoc Am|bool|isIDFIRST|char ch
684Returns a boolean indicating whether the specified character can be the first
685character of an identifier. This is very close to, but not quite the same as
686the official Unicode property C<XID_Start>. The difference is that this
687returns true only if the input character also matches L</isWORDCHAR>.
688See the L<top of this section|/Character classes> for an explanation of variants
689C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
690C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
691
692=for apidoc Am|bool|isIDCONT|char ch
693Returns a boolean indicating whether the specified character can be the
694second or succeeding character of an identifier. This is very close to, but
695not quite the same as the official Unicode property C<XID_Continue>. The
696difference is that this returns true only if the input character also matches
697L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
698explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
699C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
700C<isIDCONT_LC_utf8>.
701
243effed 702=head1 Miscellaneous Functions
8eea39dd 703
95a59cab 704=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 705Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
706Behaviour is only well defined when isXDIGIT(*str) is true.
707
e7c1e6c1
KW
708=head1 Character case changing
709
1f607577
KW
710=for apidoc Am|U8|toUPPER|U8 ch
711Converts the specified character to uppercase. If the input is anything but an
712ASCII lowercase character, that input character itself is returned. Variant
c753c8d3 713C<toUPPER_A> is equivalent.
954c1994 714
1f607577
KW
715=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
716Converts the Unicode code point C<cp> to its uppercase version, and
717stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
718that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
719bytes since the uppercase version may be longer than the original character.
720
721The first code point of the uppercased version is returned
722(but note, as explained just above, that there may be more.)
723
724=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
725Converts the UTF-8 encoded character at C<p> to its uppercase version, and
726stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
727that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
728bytes since the uppercase version may be longer than the original character.
729
730The first code point of the uppercased version is returned
731(but note, as explained just above, that there may be more.)
732
733The input character at C<p> is assumed to be well-formed.
734
25200305
KW
735=for apidoc Am|U8|toFOLD|U8 ch
736Converts the specified character to foldcase. If the input is anything but an
737ASCII uppercase character, that input character itself is returned. Variant
738C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
739Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
740
1f607577
KW
741=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
742Converts the Unicode code point C<cp> to its foldcase version, and
743stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
744that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
745bytes since the foldcase version may be longer than the original character.
746
747The first code point of the foldcased version is returned
748(but note, as explained just above, that there may be more.)
749
750=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
751Converts the UTF-8 encoded character at C<p> to its foldcase version, and
752stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
753that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
754bytes since the foldcase version may be longer than the original character.
755
756The first code point of the foldcased version is returned
757(but note, as explained just above, that there may be more.)
758
759The input character at C<p> is assumed to be well-formed.
760
761=for apidoc Am|U8|toLOWER|U8 ch
762Converts the specified character to lowercase. If the input is anything but an
763ASCII uppercase character, that input character itself is returned. Variant
c753c8d3 764C<toLOWER_A> is equivalent.
954c1994 765
1f607577
KW
766=for apidoc Am|U8|toLOWER_L1|U8 ch
767Converts the specified Latin1 character to lowercase. The results are undefined if
768the input doesn't fit in a byte.
769
770=for apidoc Am|U8|toLOWER_LC|U8 ch
771Converts the specified character to lowercase using the current locale's rules,
772if possible; otherwise returns the input character itself.
773
774=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
775Converts the Unicode code point C<cp> to its lowercase version, and
776stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
777that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
778bytes since the lowercase version may be longer than the original character.
779
780The first code point of the lowercased version is returned
781(but note, as explained just above, that there may be more.)
782
783=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
784Converts the UTF-8 encoded character at C<p> to its lowercase version, and
785stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
786that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
787bytes since the lowercase version may be longer than the original character.
788
789The first code point of the lowercased version is returned
790(but note, as explained just above, that there may be more.)
791
792The input character at C<p> is assumed to be well-formed.
793
794=for apidoc Am|U8|toLOWER_LC|U8 ch
795Converts the specified character to lowercase using the current locale's rules,
796if possible; otherwise returns the input character itself.
797
25200305
KW
798=for apidoc Am|U8|toTITLE|U8 ch
799Converts the specified character to titlecase. If the input is anything but an
800ASCII lowercase character, that input character itself is returned. Variant
801C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
802as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
803concept used in locale handling, so there is no functionality for that.)
804
1f607577
KW
805=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
806Converts the Unicode code point C<cp> to its titlecase version, and
807stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
808that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
809bytes since the titlecase version may be longer than the original character.
810
811The first code point of the titlecased version is returned
812(but note, as explained just above, that there may be more.)
813
814=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
815Converts the UTF-8 encoded character at C<p> to its titlecase version, and
816stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
817that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
818bytes since the titlecase version may be longer than the original character.
819
820The first code point of the titlecased version is returned
821(but note, as explained just above, that there may be more.)
822
823The input character at C<p> is assumed to be well-formed.
824
954c1994 825=cut
353c9b6f 826
1f607577 827XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
25200305 828really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
243effed 829
8a58bdcf 830Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
831patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
832
954c1994
GS
833*/
834
7c062697
KW
835/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
836 * is known only in the perl core, and this macro can be called from outside
837 * that */
838#ifdef HAS_QUAD
839# define WIDEST_UTYPE U64TYPE
840#else
841# define WIDEST_UTYPE U32
842#endif
843
3912bc88
KW
844/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
845 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
846 * bits of information are lost even on a 64-bit machine, but to get the
847 * compiler to optimize it out if possible. This is because Configure makes
848 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
849 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
850 */
851#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 852
41f43cc2 853#ifdef EBCDIC
b6340bd0
KW
854# ifndef _ALL_SOURCE
855 /* This returns the wrong results on at least z/OS unless this is
856 * defined. */
857# error _ALL_SOURCE should probably be defined
858# endif
859
860 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
861 * likely not defined, and so we use the native function */
862# define isASCII(c) isascii(c)
41f43cc2
KW
863#else
864# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
865#endif
866
ec79cfbb 867#define isASCII_A(c) isASCII(c)
3509bb53 868#define isASCII_L1(c) isASCII(c)
6515cb76 869
c2da0b36
KW
870/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
871 * and the 8 possible permutations of those bits exactly comprise the 8 octal
872 * digits */
3d33b3e3 873#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 874
ec79cfbb 875/* ASCII range only */
9fb1bf9d 876#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 877
a500dc72
KW
878/* Character class numbers. For internal core Perl use only. The ones less
879 * than 32 are used in PL_charclass[] and the ones up through the one that
880 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
881 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
882 * their actual definitions are here. If that file has a name not used here,
883 * it won't compile.
1709d539
KW
884 *
885 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
886 * frequency of their use. This gives a slight edge to exiting a loop earlier
887 * (in reginclass() in regexec.c) */
1709d539
KW
888# define _CC_WORDCHAR 0 /* \w and [:word:] */
889# define _CC_DIGIT 1 /* \d and [:digit:] */
890# define _CC_ALPHA 2 /* [:alpha:] */
891# define _CC_LOWER 3 /* [:lower:] */
892# define _CC_UPPER 4 /* [:upper:] */
893# define _CC_PUNCT 5 /* [:punct:] */
894# define _CC_PRINT 6 /* [:print:] */
15861f94 895# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 896# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 897# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 898
b0d691b2 899#define _FIRST_NON_SWASH_CC 10
1709d539
KW
900/* The character classes above are implemented with swashes. The second group
901 * (just below) contains the ones implemented without. These are also sorted
902 * in rough order of the frequency of their use, except that \v should be last,
903 * as it isn't a real Posix character class, and some (small) inefficiencies in
904 * regular expression handling would be introduced by putting it in the middle
905 * of those that are. Also, cntrl and ascii come after the others as it may be
906 * useful to group these which have no members that match above Latin1, (or
907 * above ASCII in the latter case) */
908
b0d691b2
KW
909# define _CC_SPACE 10 /* \s */
910# define _CC_BLANK 11 /* [:blank:] */
911# define _CC_XDIGIT 12 /* [:xdigit:] */
912# define _CC_PSXSPC 13 /* [:space:] */
913# define _CC_CNTRL 14 /* [:cntrl:] */
914# define _CC_ASCII 15 /* [:ascii:] */
915# define _CC_VERTSPACE 16 /* \v */
1709d539 916
a0947d7b
KW
917# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
918
1709d539 919/* The members of the third group below do not need to be coordinated with data
3ded5eb0
KW
920 * structures in regcomp.[ch] and regexec.c. But they should be added to
921 * bootstrap_ctype() */
b0d691b2
KW
922# define _CC_IDFIRST 17
923# define _CC_CHARNAME_CONT 18
924# define _CC_NONLATIN1_FOLD 19
925# define _CC_QUOTEMETA 20
926# define _CC_NON_FINAL_FOLD 21
927# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
928# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
929/* Unused: 23-30
f4cdb42c
KW
930 * If more bits are needed, one could add a second word for non-64bit
931 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
932 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
933 * is used only for optimization (as of this writing), and differs in the
934 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
935 * characters: the masculine and feminine ordinal indicators, so removing it
936 * would just cause /i regexes which match them to run less efficiently */
96ac0975 937
3a371f2f
KW
938#if defined(PERL_CORE) || defined(PERL_EXT)
939/* An enum version of the character class numbers, to help compilers
940 * optimize */
941typedef enum {
3a371f2f 942 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
943 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
944 _CC_ENUM_ASCII = _CC_ASCII,
945 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 946 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 947 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
948 _CC_ENUM_DIGIT = _CC_DIGIT,
949 _CC_ENUM_GRAPH = _CC_GRAPH,
950 _CC_ENUM_LOWER = _CC_LOWER,
951 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 952 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 953 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 954 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 955 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 956 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 957 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 958 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
959} _char_class_number;
960#endif
961
63c61c3f 962#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 963#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 964
3018b823 965#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 966# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
967 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
968 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
969 #error Need to adjust order of swash_property_names[]
970# endif
971
b44daeeb
KW
972/* This is declared static in each of the few files that this is #defined for
973 * to keep them from being publicly accessible. Hence there is a small amount
974 * of wasted space */
975
922e8cb4
KW
976static const char* const swash_property_names[] = {
977 "XPosixWord",
978 "XPosixDigit",
979 "XPosixAlpha",
980 "XPosixLower",
981 "XPosixUpper",
982 "XPosixPunct",
983 "XPosixPrint",
984 "XPosixAlnum",
b0d691b2
KW
985 "XPosixGraph",
986 "Cased"
922e8cb4
KW
987};
988#endif
989
96ac0975
NC
990# ifdef DOINIT
991EXTCONST U32 PL_charclass[] = {
992# include "l1_char_class_tab.h"
993};
994
995# else /* ! DOINIT */
996EXTCONST U32 PL_charclass[];
997# endif
3ded5eb0
KW
998#endif /* Has perl.h */
999
1000#if defined(H_PERL) && ! defined(BOOTSTRAP_CHARSET)
96ac0975 1001
265c1f46 1002 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70
KW
1003# define _CC_mask(classnum) (1U << (classnum))
1004# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
f4cd282c 1005 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
4eeeb416 1006
f4cdb42c
KW
1007 /* The mask for the _A versions of the macros; it just adds in the bit for
1008 * ASCII. */
1009# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1010
1011 /* The _A version makes sure that both the desired bit and the ASCII bit
1012 * are present */
3ded5eb0 1013# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
f4cd282c 1014 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
f4cdb42c
KW
1015 == _CC_mask_A(classnum)))
1016
f4cdb42c 1017# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 1018# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
1019# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1020# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1021# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
1022# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1023# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1024# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1025# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1026# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1027# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1028# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1029# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1030# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
d95f8b6a 1031# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
3ded5eb0
KW
1032# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1033# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1034# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1035
1036 /* continuation character for legal NAME in \N{NAME} */
1037# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1038
1039# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1040# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1041# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1042# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1043# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1044# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1045# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1046# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1047# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1048# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
f4cdb42c 1049
fb9bfbf7
KW
1050 /* Either participates in a fold with a character above 255, or is a
1051 * multi-char fold */
f4cd282c 1052# 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)))
430b7c70 1053
4eeeb416 1054# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
1055# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1056 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
1057# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1058 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
3ded5eb0
KW
1059#else /* Either don't have perl.h or don't want to use char_class_tab.h */
1060
1061 /* If we don't have perl.h, we are compiling a utility program. Below we
1062 * hard-code various macro definitions that wouldn't otherwise be available
1063 * to it. We can also get here if we are configured to bootstrap up Perl
1064 * on a non-ASCII platform that doesn't have a working Perl (currently only
1065 * EBCDIC). For these we currently use the native definitions to get
1066 * things going. (It should also be possible to use the translation
1067 * function NATIVE_TO_LATIN1(), but that is an extra layer of dependence on
1068 * Perl, so it is currently avoided for the macros where it's possible to
1069 * do so.) */
9fb1bf9d 1070# ifdef EBCDIC
3ded5eb0 1071 /* Use the native functions. They likely will return false for all
4f48cecf 1072 * non-ASCII values, but this makes sure */
b6340bd0
KW
1073# define isALPHA_A(c) (isASCII(c) && isalpha(c))
1074# define isALPHANUMERIC_A(c) (isASCII(c) && isalnum(c))
1075# define isCNTRL_A(c) (isASCII(c) && iscntrl(c))
1076# define isDIGIT_A(c) (isASCII(c) && isdigit(c))
1077# define isGRAPH_A(c) (isASCII(c) && isgraph(c))
1078# define isLOWER_A(c) (isASCII(c) && islower(c))
1079# define isPRINT_A(c) (isASCII(c) && isprint(c))
1080# define isPUNCT_A(c) (isASCII(c) && ispunct(c))
1081# define isSPACE_A(c) (isASCII(c) && isspace(c))
1082# define isUPPER_A(c) (isASCII(c) && isupper(c))
1083# define isXDIGIT_A(c) (isASCII(c) && isxdigit(c))
3ded5eb0 1084# else /* ASCII platform. These are coded based on first principals */
9fb1bf9d 1085# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
15861f94 1086# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
3ded5eb0 1087# define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
2181da29 1088# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
3ded5eb0 1089# define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
9fb1bf9d
KW
1090# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1091# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
3ded5eb0
KW
1092# define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1093# define isSPACE_A(c) ((c) == ' ' \
1094 || (c) == '\t' \
1095 || (c) == '\n' \
1096 || (c) == '\r' \
1097 || (c) =='\v' \
926bb54c 1098 || (c) == '\f')
2181da29 1099# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
3ded5eb0
KW
1100# define isXDIGIT_A(c) (isDIGIT_A(c) \
1101 || ((c) >= 'a' && (c) <= 'f') \
1102 || ((c) <= 'F' && (c) >= 'A'))
1103# endif /* Below are common definitions for ASCII and non-ASCII */
1104# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1105# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1106# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1107
1108 /* The _L1 macros may be unnecessary for both the utilities and for
1109 * bootstrapping; I (khw) added them during debugging of bootstrapping, and
1110 * it seems best to keep them. */
1111# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1112# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1113# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1114# define isBLANK_L1(c) (isBLANK_A(c) \
1115 || (FITS_IN_8_BITS(c) \
1116 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1117# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1118# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1119# define isLOWER_L1(c) (isLOWER_A(c) \
1120 || (FITS_IN_8_BITS(c) \
1121 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1122 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1123 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1124 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1125 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1126# define isPRINT_L1(c) (isPRINT_A(c) \
1127 || (FITS_IN_8_BITS(c) \
1128 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1129# define isPSXSPC_L1(c) isSPACE_L1(c)
1130# define isPUNCT_L1(c) (isPUNCT_A(c) \
1131 || (FITS_IN_8_BITS(c) \
1132 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1133 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1134 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1135 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1136 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1137 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1138 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1139# define isSPACE_L1(c) (isSPACE_A(c) \
1140 || (FITS_IN_8_BITS(c) \
1141 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1142 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1143# define isUPPER_L1(c) (isUPPER_A(c) \
1144 || (FITS_IN_8_BITS(c) \
1145 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1146 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1147 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1148# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1149# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1150# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1151 || isBLANK_L1(c) \
1152 || (c) == '-' \
1153 || (c) == '(' \
1154 || (c) == ')')
1155 /* The following are not fully accurate in the above-ASCII range. I (khw)
1156 * don't think it's necessary to be so for the purposes where this gets
1157 * compiled */
1158# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1159# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1160
1161 /* And these aren't accurate at all. They are useful only for above
1162 * Latin1, which utilities and bootstrapping don't deal with */
1163# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1164# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1165
1166 /* Many of the macros later in this file are defined in terms of these. By
1167 * implementing them with a function, which converts the class number into
1168 * a call to the desired macro, all of the later ones work. However, that
1169 * function won't be actually defined when building a utility program (no
1170 * perl.h), and so a compiler error will be generated if one is attempted
1171 * to be used. And the above-Latin1 code points require Unicode tables to
1172 * be present, something unlikely to be the case when bootstrapping */
1173# define _generic_isCC(c, classnum) \
1174 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1175# define _generic_isCC_A(c, classnum) \
1176 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
1177#endif /* End of no perl.h or have BOOTSTRAP_CHARSET */
8a58bdcf 1178
e66b99e9
KW
1179#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1180#define isALPHA(c) isALPHA_A(c)
1181#define isBLANK(c) isBLANK_A(c)
1182#define isCNTRL(c) isCNTRL_A(c)
1183#define isDIGIT(c) isDIGIT_A(c)
1184#define isGRAPH(c) isGRAPH_A(c)
1185#define isIDFIRST(c) isIDFIRST_A(c)
1186#define isLOWER(c) isLOWER_A(c)
1187#define isPRINT(c) isPRINT_A(c)
1188#define isPSXSPC(c) isPSXSPC_A(c)
1189#define isPUNCT(c) isPUNCT_A(c)
1190#define isSPACE(c) isSPACE_A(c)
1191#define isUPPER(c) isUPPER_A(c)
1192#define isWORDCHAR(c) isWORDCHAR_A(c)
1193#define isXDIGIT(c) isXDIGIT_A(c)
1194
1195/* ASCII casing. These could also be written as
1196 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1197 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1198 which uses table lookup and mask instead of subtraction. (This would
1199 work because the _MOD does not apply in the ASCII range) */
1200#define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
1201#define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
bbce6d69 1202
25200305
KW
1203/* In the ASCII range, these are equivalent to what they're here defined to be.
1204 * But by creating these definitions, other code doesn't have to be aware of
1205 * this detail */
1206#define toFOLD(c) toLOWER(c)
1207#define toFOLD_LC(c) toLOWER_LC(c)
1208#define toTITLE(c) toUPPER(c)
1209
c753c8d3
KW
1210#define toLOWER_A(c) toLOWER(c)
1211#define toUPPER_A(c) toUPPER(c)
25200305
KW
1212#define toFOLD_A(c) toFOLD(c)
1213#define toTITLE_A(c) toTITLE(c)
1a0901db
KW
1214
1215/* Use table lookup for speed; return error character for input
1216 * out-of-range */
b2bf251f 1217#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1218 ? (c) \
f4cd282c 1219 : PL_latin1_lc[ (U8) (c) ])
c753c8d3
KW
1220#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1221
1a0901db
KW
1222/* Modified uc. Is correct uc except for three non-ascii chars which are
1223 * all mapped to one of them, and these need special handling; error
1224 * character for input out-of-range */
b2bf251f 1225#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1226 ? (c) \
f4cd282c 1227 : PL_mod_latin1_uc[ (U8) (c) ])
bbce6d69
PP
1228#ifdef USE_NEXT_CTYPE
1229
15861f94 1230# define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c))
ff68c719 1231# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 1232# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 1233# define isBLANK_LC(c) isBLANK((unsigned int)(c))
b8c5462f 1234# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
509fb054 1235# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
b8c5462f 1236# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
509fb054
KW
1237# define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
1238# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
ff68c719 1239# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 1240# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
509fb054
KW
1241# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
1242# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
1243# define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
e712593e 1244# define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
ff68c719 1245# define toLOWER_LC(c) NXToLower((unsigned int)(c))
509fb054 1246# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
bbce6d69
PP
1247
1248#else /* !USE_NEXT_CTYPE */
b8c5462f 1249
bbce6d69
PP
1250# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1251
76fde96a 1252/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
851131d3 1253
76fde96a 1254# define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
15861f94
KW
1255# define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \
1256 && isalnum((unsigned char)(c)))
07315176 1257# ifdef HAS_ISASCII
76fde96a 1258# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
07315176 1259# else
76fde96a 1260# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
07315176 1261# endif
67addccf 1262# ifdef HAS_ISBLANK
76fde96a 1263# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
67addccf 1264# else
76fde96a 1265# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
67addccf 1266# endif
76fde96a 1267# define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
509fb054 1268# define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
76fde96a 1269# define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
509fb054
KW
1270# define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1271 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1272# define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
76fde96a
KW
1273# define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1274# define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
509fb054
KW
1275# define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1276# define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1277# define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1278 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
e712593e 1279# define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
548d896a
DM
1280# define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? (UV)tolower((unsigned char)(c)) : (c))
1281# define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? (UV)toupper((unsigned char)(c)) : (c))
bbce6d69
PP
1282
1283# else
1284
bbce6d69 1285# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1286# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
07315176 1287# define isASCII_LC(c) isascii(c)
67addccf
KW
1288# ifdef HAS_ISBLANK
1289# define isBLANK_LC(c) (isascii(c) && isblank(c))
1290# else
3e66ab74 1291# define isBLANK_LC(c) isBLANK_A(c)
67addccf 1292# endif
b8c5462f 1293# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1294# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1295# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1296# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1297# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1298# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1299# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1300# define isSPACE_LC(c) (isascii(c) && isspace(c))
1301# define isUPPER_LC(c) (isascii(c) && isupper(c))
1302# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1303# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
76fde96a 1304# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1305# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
bbce6d69
PP
1306
1307# endif
a0d0e21e 1308#endif /* USE_NEXT_CTYPE */
55204971 1309
eba68aa0
KW
1310#define isIDCONT(c) isWORDCHAR(c)
1311#define isIDCONT_A(c) isWORDCHAR_A(c)
1312#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1313#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1314#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1315
66c17564 1316/* For internal core Perl use only. If the input is Latin1, use the Latin1
a500dc72
KW
1317 * macro; otherwise use the function 'above_latin1'. Won't compile if 'c' isn't unsigned, as
1318 * won't match above_latin1 prototype. The macros do bounds checking, so have
66c17564
KW
1319 * duplicate checks here, so could create versions of the macros that don't,
1320 * but experiments show that gcc optimizes them out anyway. */
1321
1322/* Note that all ignore 'use bytes' */
cd500f2f
KW
1323#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1324 ? _generic_isCC(c, classnum) \
1325 : above_latin1(c))
1326#define _generic_swash_uni(classnum, c) ((c) < 256 \
1327 ? _generic_isCC(c, classnum) \
922e8cb4 1328 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1329#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1330#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1331#define isASCII_uni(c) isASCII(c)
1332#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1333#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1334#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1335#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1336#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1337#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1338#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1339#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1340
509fb054
KW
1341/* Posix and regular space are identical above Latin1 */
1342#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1343
cd500f2f 1344#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1345#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1346#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1347#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1348#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1349#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1350
b8d68ded 1351#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1352#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1353#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1354#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1355
cd500f2f
KW
1356#define _generic_LC_uvchr(latin1, above_latin1, c) \
1357 (c < 256 ? latin1(c) : above_latin1(c))
1358#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1359 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1360
1361#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1362#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1363 _CC_ALPHANUMERIC, c)
509fb054 1364#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1365#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1366#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1367#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1368#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1369#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1370 _is_uni_perl_idcont, c)
cd500f2f
KW
1371#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1372 _is_uni_perl_idstart, c)
1373#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1374#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1375#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1376 space under locale */
cd500f2f
KW
1377#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1378#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1379 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1380#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1381#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1382 _CC_WORDCHAR, c)
1383#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1384
36b00f93 1385#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1386
a500dc72
KW
1387/* Everything whose name begins with an underscore is for internal core Perl
1388 * use only. */
1389
1390/* If the input is in the Latin1 range, use
252810d0
KW
1391 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1392 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1393 * fact that ASCII characters have the same representation whether utf8 or not.
1394 * Note that it assumes that the utf8 has been validated, and ignores 'use
1395 * bytes' */
cd500f2f 1396#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1397 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1398 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c 1399 ? _generic_isCC( \
94bb8c36 1400 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
aea0c08c 1401 *((p)+1 )), \
94bb8c36 1402 classnum) \
252810d0 1403 : utf8)
cd500f2f
KW
1404/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1405 * can be a macro */
1406#define _generic_func_utf8(classnum, above_latin1, p) \
1407 _generic_utf8(classnum, p, above_latin1(p))
1408/* Like the above, but passes classnum to _isFOO_utf8(), instead of having a
1409 * 'above_latin1' parameter */
922e8cb4 1410#define _generic_swash_utf8(classnum, p) \
cd500f2f 1411 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1412
cc8ab7c0
KW
1413/* Like the above, but should be used only when it is known that there are no
1414 * characters in the range 128-255 which the class is TRUE for. Hence it can
cd500f2f 1415 * skip the tests for this range. 'above_latin1' should include its arguments */
3ded5eb0 1416#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1417 (UTF8_IS_INVARIANT(*(p)) \
1418 ? _generic_isCC(*(p), classnum) \
1419 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
3ded5eb0 1420 ? above_latin1 \
cc8ab7c0
KW
1421 : 0)
1422
84238efa
KW
1423/* NOTE that some of these macros have very similar ones in regcharclass.h.
1424 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1425 * there, differing in name only by an underscore from the one here
1426 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1427 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1428 * points; the regcharclass.h ones are implemented as a series of
1429 * "if-else-if-else ..." */
1430
922e8cb4
KW
1431#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1432#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1433#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1434 utf8, the non-utf8 macro works
1435 */
cd500f2f
KW
1436#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1437#define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1438#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1439 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1440#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1441#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1442 _is_utf8_perl_idcont, p)
e5dcd934 1443
c11ff943
KW
1444/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1445 * IDFIRST is an alnum. See
c8362b00 1446 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1447 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1448 * This used to be not the XID version, but we decided to go with the more
1449 * modern Unicode definition */
cd500f2f 1450#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1451 _is_utf8_perl_idstart, p)
e5dcd934 1452
922e8cb4
KW
1453#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1454#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1455
509fb054 1456/* Posix and regular space are identical above Latin1 */
cd500f2f 1457#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1458
922e8cb4 1459#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1460#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1461#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1462#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1463#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1464#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1465 is_XDIGIT_high(p))
a0ed51b3 1466
7d2b61dd 1467#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
509fb054
KW
1468#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1469#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1470#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1471
9201a572
KW
1472/* For internal core Perl use only. If the input is in the Latin1 range, use
1473 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1474 * use the value given by the 'utf8' parameter. This relies on the fact that
1475 * ASCII characters have the same representation whether utf8 or not. Note
1476 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
94bb8c36
KW
1477#define _generic_LC_utf8(macro, p, utf8) \
1478 (UTF8_IS_INVARIANT(*(p)) \
1479 ? macro(*(p)) \
1480 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1481 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
9201a572
KW
1482 : utf8)
1483
cd500f2f
KW
1484#define _generic_LC_swash_utf8(macro, classnum, p) \
1485 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1486#define _generic_LC_func_utf8(macro, above_latin1, p) \
1487 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1488
cd500f2f
KW
1489#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1490 _CC_ALPHANUMERIC, p)
1491#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1492#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1493#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1494#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1495#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1496#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1497#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1498#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1499#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1500#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1501#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1502 space under locale */
cd500f2f
KW
1503#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1504#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1505#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1506#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1507 _CC_WORDCHAR, p)
1508#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1509
fbc19f27
KW
1510/* Macros for backwards compatibility and for completeness when the ASCII and
1511 * Latin1 values are identical */
fbc19f27
KW
1512#define isALPHAU(c) isALPHA_L1(c)
1513#define isDIGIT_L1(c) isDIGIT_A(c)
1514#define isOCTAL(c) isOCTAL_A(c)
1515#define isOCTAL_L1(c) isOCTAL_A(c)
1516#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1517#define isALNUM(c) isWORDCHAR(c)
1518#define isALNUMU(c) isWORDCHAR_L1(c)
1519#define isALNUM_LC(c) isWORDCHAR_LC(c)
1520#define isALNUM_uni(c) isWORDCHAR_uni(c)
1521#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1522#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1523#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1524#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1525#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1526#define isALNUMC(c) isALPHANUMERIC(c)
1527#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1528#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1529#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1530#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1531#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1532
6ae9f32a 1533/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1534 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1535 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1536 * that is EOT on ASCII; ST on EBCDIC */
8e7c6e7d 1537# define toCTRL(c) (toUPPER(NATIVE_TO_LATIN1(c)) ^ 64)
bbce6d69 1538
dea28490
JJ
1539/* Line numbers are unsigned, 32 bits. */
1540typedef U32 line_t;
e5dcd934 1541#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1542
91152fc1
DG
1543/* Helpful alias for version prescan */
1544#define is_LAX_VERSION(a,b) \
1545 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1546
1547#define is_STRICT_VERSION(a,b) \
1548 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1549
1550#define BADVERSION(a,b,c) \
1551 if (b) { \
1552 *b = c; \
1553 } \
1554 return a;
8c52afec 1555
14f6cdfd
KW
1556/* Converts a hex digit in a string to its numeric value, advancing the
1557 * pointer. The input must be known to be 0-9, A-F, or a-f. In both ASCII and
1558 * EBCDIC the last 4 bits of the digits are 0-9; and the last 4 bits of A-F and
1559 * a-f are 1-6, so adding 9 yields 10-15 */
1560#define READ_XDIGIT(s) (0xf & (isDIGIT(*(s)) ? (*(s)++) : (*(s)++ + 9)))
95a59cab 1561
8e84507e 1562/*
ccfc67b7
JH
1563=head1 Memory Management
1564
a02a5408 1565=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1566The XSUB-writer's interface to the C C<malloc> function.
1567
c5008215
JC
1568In 5.9.3, Newx() and friends replace the older New() API, and drops
1569the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1570themselves. This aid has been superseded by a new build option,
d10b4965 1571PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1572there for use in XS modules supporting older perls.
1573
a02a5408 1574=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1575The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1576cast. See also C<Newx>.
954c1994 1577
a02a5408 1578=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1579The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1580memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1581
954c1994
GS
1582=for apidoc Am|void|Renew|void* ptr|int nitems|type
1583The XSUB-writer's interface to the C C<realloc> function.
1584
1585=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1586The XSUB-writer's interface to the C C<realloc> function, with
1587cast.
1588
49b8b560 1589=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1590The XSUB-writer's interface to the C C<free> function.
1591
1592=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1593The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1594source, C<dest> is the destination, C<nitems> is the number of items, and
1595C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1596
e90e2364
NC
1597=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1598Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1599optimise.
1600
954c1994
GS
1601=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1602The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1603source, C<dest> is the destination, C<nitems> is the number of items, and
1604C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1605
e90e2364
NC
1606=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1607
1608Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1609optimise.
1610
954c1994
GS
1611=for apidoc Am|void|Zero|void* dest|int nitems|type
1612
1613The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1614destination, C<nitems> is the number of items, and C<type> is the type.
1615
e90e2364
NC
1616=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1617
1618Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1619optimise.
1620
da5d8dbb 1621=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1622This is an architecture-independent macro to copy one structure to another.
954c1994 1623
7e337ee0
JH
1624=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1625
1626Fill up memory with a byte pattern (a byte repeated over and over
1627again) that hopefully catches attempts to access uninitialized memory.
1628
1629=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1630
1631PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1632
1c12ffb4 1633=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1634
1635PoisonWith(0xEF) for catching access to freed memory.
1636
9965345d
JH
1637=for apidoc Am|void|Poison|void* dest|int nitems|type
1638
7e337ee0 1639PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1640
1641=cut */
954c1994 1642
561b68a9
SH
1643/* Maintained for backwards-compatibility only. Use newSV() instead. */
1644#ifndef PERL_CORE
ff06c60c 1645#define NEWSV(x,len) newSV(len)
561b68a9 1646#endif
ff06c60c 1647
19a94d75
JH
1648#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1649
2720f793
JH
1650/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1651 * overly eager compilers that will bleat about e.g.
1652 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1653#ifdef PERL_MALLOC_WRAP
f019c49e 1654#define MEM_WRAP_CHECK(n,t) \
d1decf2b 1655 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
27d5b266 1656#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1657 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1658#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1659
d1decf2b 1660#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
1661#else
1662
410319be
NC
1663#define MEM_WRAP_CHECK(n,t)
1664#define MEM_WRAP_CHECK_1(n,t,a)
1665#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1666#define MEM_WRAP_CHECK_(n,t)
1667
6fff79ce 1668#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1669
1936d2a7 1670#endif
8b44ba4c 1671
fe4f188c 1672#ifdef PERL_MEM_LOG
46c6c7e2 1673/*
9f653bb5 1674 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1675 * go through functions, which are handy for debugging breakpoints, but
1676 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1677 * line number, and C function name if available) passed in. This info can
1678 * then be used for logging the calls, for which one gets a sample
73d1d973 1679 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1680 *
46c6c7e2 1681 * Known problems:
94e892a6 1682 * - not all memory allocs get logged, only those
46c6c7e2 1683 * that go through Newx() and derivatives (while all
94e892a6 1684 * Safefrees do get logged)
46c6c7e2
JH
1685 * - __FILE__ and __LINE__ do not work everywhere
1686 * - __func__ or __FUNCTION__ even less so
1687 * - I think more goes on after the perlio frees but
1688 * the thing is that STDERR gets closed (as do all
1689 * the file descriptors)
1690 * - no deeper calling stack than the caller of the Newx()
1691 * or the kind, but do I look like a C reflection/introspection
1692 * utility to you?
1693 * - the function prototypes for the logging functions
1694 * probably should maybe be somewhere else than handy.h
1695 * - one could consider inlining (macrofying) the logging
1696 * for speed, but I am too lazy
1697 * - one could imagine recording the allocations in a hash,
1698 * (keyed by the allocation address?), and maintain that
1699 * through reallocs and frees, but how to do that without
1700 * any News() happening...?
73d1d973 1701 * - lots of -Ddefines to get useful/controllable output
b953482e 1702 * - lots of ENV reads
46c6c7e2
JH
1703 */
1704
12754f92 1705PERL_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 1706
12754f92 1707PERL_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 1708
12754f92 1709PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1710
0b0ab801 1711# ifdef PERL_CORE
73d1d973 1712# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1713enum mem_log_type {
1714 MLT_ALLOC,
1715 MLT_REALLOC,
d7a2c63c
MHM
1716 MLT_FREE,
1717 MLT_NEW_SV,
1718 MLT_DEL_SV
0b0ab801
MHM
1719};
1720# endif
12754f92 1721# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1722void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1723void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1724# endif
0b0ab801
MHM
1725# endif
1726
fe4f188c
JH
1727#endif
1728
1729#ifdef PERL_MEM_LOG
d1401ee9
MHM
1730#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1731#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1732#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1733#endif
1734
1735#ifndef MEM_LOG_ALLOC
1736#define MEM_LOG_ALLOC(n,t,a) (a)
1737#endif
1738#ifndef MEM_LOG_REALLOC
1739#define MEM_LOG_REALLOC(n,t,v,a) (a)
1740#endif
1741#ifndef MEM_LOG_FREE
1742#define MEM_LOG_FREE(a) (a)
1743#endif
1744
d1401ee9
MHM
1745#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1746#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1747#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1748
1749#ifndef PERL_CORE
a02a5408
JC
1750/* pre 5.9.x compatibility */
1751#define New(x,v,n,t) Newx(v,n,t)
1752#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1753#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1754#endif
a02a5408 1755
ff68c719 1756#define Renew(v,n,t) \
d1401ee9 1757 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1758#define Renewc(v,n,t,c) \
d1401ee9 1759 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1760
1761#ifdef PERL_POISON
1762#define Safefree(d) \
06c0cc96 1763 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1764#else
fe4f188c 1765#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1766#endif
55497cff 1767
8b44ba4c
NC
1768#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1769#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1770#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1771
8b44ba4c
NC
1772#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1773#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1774#ifdef HAS_MEMSET
8b44ba4c 1775#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1776#else
8b44ba4c
NC
1777/* Using bzero(), which returns void. */
1778#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1779#endif
1780
7e337ee0
JH
1781#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1782#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1783#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1784#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1785
caa674f3
DD
1786#ifdef PERL_POISON
1787# define PERL_POISON_EXPR(x) x
1788#else
1789# define PERL_POISON_EXPR(x)
1790#endif
1791
2304df62 1792#ifdef USE_STRUCT_COPY
ff68c719 1793#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1794#else
1795#define StructCopy(s,d,t) Copy(s,d,1,t)
1796#endif
2cc61e15 1797
622913ab 1798#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1799#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1800
2cc61e15
DD
1801#ifdef NEED_VA_COPY
1802# ifdef va_copy
1803# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1804# else
a1866d1b
JH
1805# if defined(__va_copy)
1806# define Perl_va_copy(s, d) __va_copy(d, s)
1807# else
1808# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1809# endif
2cc61e15
DD
1810# endif
1811#endif
1812
472d47bc
SB
1813/* convenience debug macros */
1814#ifdef USE_ITHREADS
1815#define pTHX_FORMAT "Perl interpreter: 0x%p"
1816#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1817#define pTHX_VALUE_ (void *)my_perl,
1818#define pTHX_VALUE (void *)my_perl
1819#define pTHX__VALUE_ ,(void *)my_perl,
1820#define pTHX__VALUE ,(void *)my_perl
472d47bc 1821#else
3609ea0d 1822#define pTHX_FORMAT
472d47bc 1823#define pTHX__FORMAT
3609ea0d 1824#define pTHX_VALUE_
472d47bc 1825#define pTHX_VALUE
3609ea0d 1826#define pTHX__VALUE_
472d47bc
SB
1827#define pTHX__VALUE
1828#endif /* USE_ITHREADS */
3609ea0d 1829
2acdbac1
NC
1830/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1831 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1832 CPAN::Unpack show any users outside the core. */
1833#ifdef PERL_CORE
d1d15184 1834# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1835#endif
1836
dfff4baf
BF
1837/* Internal macros to deal with gids and uids */
1838#ifdef PERL_CORE
1839
1840# if Uid_t_size > IVSIZE
1841# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
1842# define SvUID(sv) SvNV(sv)
1843# else
1844# if Uid_t_sign <= 0
1845# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
1846# define SvUID(sv) SvIV(sv)
1847# else
1848# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
1849# define SvUID(sv) SvUV(sv)
1850# endif
1851# endif /* Uid_t_size */
1852
1853# if Gid_t_size > IVSIZE
1854# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
1855# define SvGID(sv) SvNV(sv)
1856# else
1857# if Gid_t_sign <= 0
1858# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
1859# define SvGID(sv) SvIV(sv)
1860# else
1861# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
1862# define SvGID(sv) SvUV(sv)
1863# endif
1864# endif /* Gid_t_size */
1865
1866#endif
1867
9d745869
KW
1868#endif /* HANDY_H */
1869
e9a8c099
MHM
1870/*
1871 * Local variables:
1872 * c-indentation-style: bsd
1873 * c-basic-offset: 4
14d04a33 1874 * indent-tabs-mode: nil
e9a8c099
MHM
1875 * End:
1876 *
14d04a33 1877 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1878 */