This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
regen/warnings.pl: Add comments
[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
72d33970
FC
28Null character pointer. (No longer available when C<PERL_CORE> is
29defined.)
2307c6d0 30
954c1994 31=for apidoc AmU||Nullsv
72d33970 32Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
33
34=cut
35*/
36
24792b8d
NC
37# define Nullch Null(char*)
38# define Nullfp Null(PerlIO*)
39# define Nullsv Null(SV*)
40#endif
8d063cd8 41
641d3f0b
PP
42#ifdef TRUE
43#undef TRUE
44#endif
45#ifdef FALSE
46#undef FALSE
47#endif
48#define TRUE (1)
49#define FALSE (0)
50
cf3f0ffb
DM
51/* The MUTABLE_*() macros cast pointers to the types shown, in such a way
52 * (compiler permitting) that casting away const-ness will give a warning;
53 * e.g.:
54 *
55 * const SV *sv = ...;
56 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
57 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
58 */
59
b1bc3f34 60#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 61# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
62#else
63# define MUTABLE_PTR(p) ((void *) (p))
64#endif
65
a062e10d 66#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 67#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 68#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 69#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 70#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 71#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 72
f789f6a4 73#if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR)
bd31be4b
NC
74# include <stdbool.h>
75# ifndef HAS_BOOL
76# define HAS_BOOL 1
77# endif
78#endif
79
8e84507e 80/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
81 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
82 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 83 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
84 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
85 g++ can be identified by __GNUG__.
86 Andy Dougherty February 2000
5d94fbed 87*/
3609ea0d 88#ifdef __GNUG__ /* GNU g++ has bool built-in */
f789f6a4 89# ifndef PERL_BOOL_AS_CHAR
5d94fbed 90# ifndef HAS_BOOL
c1d22f6b 91# define HAS_BOOL 1
5d94fbed 92# endif
f789f6a4 93# endif
5d94fbed
AD
94#endif
95
641d3f0b
PP
96/* The NeXT dynamic loader headers will not build with the bool macro
97 So declare them now to clear confusion.
98*/
8f1f23e8 99#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
100# undef FALSE
101# undef TRUE
102 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
103# define ENUM_BOOL 1
104# ifndef HAS_BOOL
105# define HAS_BOOL 1
106# endif /* !HAS_BOOL */
8f1f23e8 107#endif /* NeXT || __NeXT__ */
641d3f0b 108
5d94fbed 109#ifndef HAS_BOOL
f789f6a4
FC
110# ifdef bool
111# undef bool
112# endif
70d5cb32 113# define bool char
c1d22f6b 114# define HAS_BOOL 1
a687059c 115#endif
0d3e774c 116
236415e2
KW
117/* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
118 * defined as char for example, then the cast from int is
119 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
120 * AIX */
18f5643b 121#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 122
46c6c7e2 123/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
124 * XXX Should really be a Configure probe, with HAS__FUNCTION__
125 * and FUNCTION__ as results.
126 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
127#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
128# define FUNCTION__ __func__
129#else
9ea40801 130# if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 131 (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
132# define FUNCTION__ ""
133# else
134# define FUNCTION__ __FUNCTION__ /* Common extension. */
135# endif
136#endif
137
27d4fb96
PP
138/* XXX A note on the perl source internal type system. The
139 original intent was that I32 be *exactly* 32 bits.
140
141 Currently, we only guarantee that I32 is *at least* 32 bits.
142 Specifically, if int is 64 bits, then so is I32. (This is the case
143 for the Cray.) This has the advantage of meshing nicely with
144 standard library calls (where we pass an I32 and the library is
145 expecting an int), but the disadvantage that an I32 is not 32 bits.
146 Andy Dougherty August 1996
24fef2a7 147
dc45a647
MB
148 There is no guarantee that there is *any* integral type with
149 exactly 32 bits. It is perfectly legal for a system to have
150 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 151
dc45a647
MB
152 Similarly, there is no guarantee that I16 and U16 have exactly 16
153 bits.
693762b4 154
8e84507e
NIS
155 For dealing with issues that may arise from various 32/64-bit
156 systems, we will ask Configure to check out
8175356b 157
3609ea0d
JH
158 SHORTSIZE == sizeof(short)
159 INTSIZE == sizeof(int)
160 LONGSIZE == sizeof(long)
dc45a647 161 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 162 PTRSIZE == sizeof(void *)
dc45a647
MB
163 DOUBLESIZE == sizeof(double)
164 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 165
27d4fb96
PP
166*/
167
69512466
JH
168#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
169# include <inttypes.h>
dd0eed91
JH
170# ifdef INT32_MIN_BROKEN
171# undef INT32_MIN
172# define INT32_MIN (-2147483647-1)
173# endif
174# ifdef INT64_MIN_BROKEN
175# undef INT64_MIN
176# define INT64_MIN (-9223372036854775807LL-1)
177# endif
69512466
JH
178#endif
179
8175356b
JH
180typedef I8TYPE I8;
181typedef U8TYPE U8;
182typedef I16TYPE I16;
183typedef U16TYPE U16;
184typedef I32TYPE I32;
185typedef U32TYPE U32;
6b8eaf93
JH
186#ifdef PERL_CORE
187# ifdef HAS_QUAD
8175356b
JH
188typedef I64TYPE I64;
189typedef U64TYPE U64;
6b8eaf93
JH
190# endif
191#endif /* PERL_CORE */
8175356b 192
69512466 193#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
80524f33
JL
194# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
195# define PeRl_INT64_C(c) CAT2(c,LL)
196# define PeRl_UINT64_C(c) CAT2(c,ULL)
197# else
52865553
SH
198# if QUADKIND == QUAD_IS___INT64
199# define PeRl_INT64_C(c) CAT2(c,I64)
200# define PeRl_UINT64_C(c) CAT2(c,UI64)
69512466 201# else
52865553
SH
202# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
203# define PeRl_INT64_C(c) CAT2(c,L)
204# define PeRl_UINT64_C(c) CAT2(c,UL)
69512466 205# else
80524f33
JL
206# define PeRl_INT64_C(c) ((I64TYPE)(c))
207# define PeRl_UINT64_C(c) ((U64TYPE)(c))
69512466
JH
208# endif
209# endif
e8c95190 210# endif
80524f33
JL
211# ifndef UINT64_C
212# define UINT64_C(c) PeRl_UINT64_C(c)
213# endif
214# ifndef INT64_C
215# define INT64_C(c) PeRl_INT64_C(c)
216# endif
e8c95190 217#endif
dc750f2e 218
d8668976 219#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 220
5ff3f7a4
GS
221/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
222 Please search CHAR_MAX in perl.h for further details. */
223#define U8_MAX UINT8_MAX
224#define U8_MIN UINT8_MIN
225
5ff3f7a4
GS
226#define I16_MAX INT16_MAX
227#define I16_MIN INT16_MIN
228#define U16_MAX UINT16_MAX
229#define U16_MIN UINT16_MIN
230
5ff3f7a4
GS
231#define I32_MAX INT32_MAX
232#define I32_MIN INT32_MIN
0e983133
GS
233#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
234# define U32_MAX UINT32_MAX
235#else
236# define U32_MAX 4294967295U
237#endif
5ff3f7a4
GS
238#define U32_MIN UINT32_MIN
239
240#else
241
5c9fa16e
KA
242/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
243 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
244#define U8_MAX PERL_UCHAR_MAX
245#define U8_MIN PERL_UCHAR_MIN
79072805 246
27d4fb96
PP
247#define I16_MAX PERL_SHORT_MAX
248#define I16_MIN PERL_SHORT_MIN
249#define U16_MAX PERL_USHORT_MAX
250#define U16_MIN PERL_USHORT_MIN
79072805 251
c4f23d77 252#if LONGSIZE > 4
27d4fb96
PP
253# define I32_MAX PERL_INT_MAX
254# define I32_MIN PERL_INT_MIN
255# define U32_MAX PERL_UINT_MAX
256# define U32_MIN PERL_UINT_MIN
79072805 257#else
27d4fb96
PP
258# define I32_MAX PERL_LONG_MAX
259# define I32_MIN PERL_LONG_MIN
260# define U32_MAX PERL_ULONG_MAX
261# define U32_MIN PERL_ULONG_MIN
79072805
LW
262#endif
263
5ff3f7a4
GS
264#endif
265
58a9a5d5 266/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
267#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
268#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
269#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
270
ff68c719 271#define Ctl(ch) ((ch) & 037)
8d063cd8 272
3fe05580
MHM
273/*
274=head1 SV-Body Allocation
275
276=for apidoc Ama|SV*|newSVpvs|const char* s
277Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
278
84bafc02
NC
279=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
280Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
281pair.
282
3fe05580
MHM
283=for apidoc Ama|SV*|newSVpvs_share|const char* s
284Like C<newSVpvn_share>, but takes a literal string instead of a string/length
285pair and omits the hash parameter.
286
9dcc53ea
Z
287=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
288Like C<sv_catpvn_flags>, but takes a literal string instead of a
289string/length pair.
290
291=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
292Like C<sv_catpvn_nomg>, but takes a literal string instead of a
293string/length pair.
294
7b4e7469 295=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
296Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
297
9dcc53ea
Z
298=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
299Like C<sv_catpvn_mg>, but takes a literal string instead of a
300string/length pair.
301
7b4e7469 302=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
303Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
304
9dcc53ea
Z
305=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
306Like C<sv_setpvn_mg>, but takes a literal string instead of a
307string/length pair.
308
309=for apidoc Am|SV *|sv_setref_pvs|const char* s
310Like C<sv_setref_pvn>, but takes a literal string instead of a
311string/length pair.
312
3fe05580
MHM
313=head1 Memory Management
314
315=for apidoc Ama|char*|savepvs|const char* s
316Like C<savepvn>, but takes a literal string instead of a string/length pair.
317
9dcc53ea
Z
318=for apidoc Ama|char*|savesharedpvs|const char* s
319A version of C<savepvs()> which allocates the duplicate string in memory
320which is shared between threads.
321
3fe05580
MHM
322=head1 GV Functions
323
324=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
325Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
326
327=head1 Hash Manipulation Functions
328
329=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
330Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
331
332=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
333Like C<hv_store>, but takes a literal string instead of a string/length pair
334and omits the hash parameter.
335
510966aa
Z
336=head1 Lexer interface
337
338=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
339
340Like L</lex_stuff_pvn>, but takes a literal string instead of a
341string/length pair.
342
3fe05580
MHM
343=cut
344*/
345
926bb54c
KW
346/* concatenating with "" ensures that only literal strings are accepted as
347 * argument */
26bfd924 348#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 349
926bb54c
KW
350/* note that STR_WITH_LEN() can't be used as argument to macros or functions
351 * that under some configurations might be macros, which means that it requires
352 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
353 */
354
355/* STR_WITH_LEN() shortcuts */
356#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
357#define newSVpvs_flags(str,flags) \
358 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 359#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
360#define sv_catpvs_flags(sv, str, flags) \
361 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
362#define sv_catpvs_nomg(sv, str) \
363 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
364#define sv_catpvs(sv, str) \
365 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
366#define sv_catpvs_mg(sv, str) \
367 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 368#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
369#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
370#define sv_setref_pvs(rv, classname, str) \
371 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 372#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
373#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
374#define gv_stashpvs(str, create) \
375 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
376#define gv_fetchpvs(namebeg, add, sv_type) \
377 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
378#define gv_fetchpvn(namebeg, len, add, sv_type) \
379 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
380#define sv_catxmlpvs(dsv, str, utf8) \
381 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
382#define hv_fetchs(hv,key,lval) \
383 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
384 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
385 : HV_FETCH_JUST_SV, NULL, 0))
386
387#define hv_stores(hv,key,val) \
388 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
389 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 390
510966aa
Z
391#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
392
b96d8cd9
NC
393#define get_cvs(str, flags) \
394 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 395
954c1994 396/*
ccfc67b7
JH
397=head1 Miscellaneous Functions
398
954c1994
GS
399=for apidoc Am|bool|strNE|char* s1|char* s2
400Test two strings to see if they are different. Returns true or
401false.
402
403=for apidoc Am|bool|strEQ|char* s1|char* s2
404Test two strings to see if they are equal. Returns true or false.
405
406=for apidoc Am|bool|strLT|char* s1|char* s2
407Test two strings to see if the first, C<s1>, is less than the second,
408C<s2>. Returns true or false.
409
410=for apidoc Am|bool|strLE|char* s1|char* s2
411Test two strings to see if the first, C<s1>, is less than or equal to the
412second, C<s2>. Returns true or false.
413
414=for apidoc Am|bool|strGT|char* s1|char* s2
415Test two strings to see if the first, C<s1>, is greater than the second,
416C<s2>. Returns true or false.
417
418=for apidoc Am|bool|strGE|char* s1|char* s2
419Test two strings to see if the first, C<s1>, is greater than or equal to
420the second, C<s2>. Returns true or false.
421
422=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
423Test two strings to see if they are different. The C<len> parameter
72d33970 424indicates the number of bytes to compare. Returns true or false. (A
954c1994
GS
425wrapper for C<strncmp>).
426
427=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
428Test two strings to see if they are equal. The C<len> parameter indicates
72d33970 429the number of bytes to compare. Returns true or false. (A wrapper for
954c1994
GS
430C<strncmp>).
431
432=cut
433*/
434
8d063cd8
LW
435#define strNE(s1,s2) (strcmp(s1,s2))
436#define strEQ(s1,s2) (!strcmp(s1,s2))
437#define strLT(s1,s2) (strcmp(s1,s2) < 0)
438#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
439#define strGT(s1,s2) (strcmp(s1,s2) > 0)
440#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
441#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
442#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 443
36477c24
PP
444#ifdef HAS_MEMCMP
445# define memNE(s1,s2,l) (memcmp(s1,s2,l))
446# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
447#else
448# define memNE(s1,s2,l) (bcmp(s1,s2,l))
449# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
450#endif
451
568a785a 452#define memEQs(s1, l, s2) \
8e6497b6 453 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
454#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
455
bbce6d69
PP
456/*
457 * Character classes.
458 *
459 * Unfortunately, the introduction of locales means that we
460 * can't trust isupper(), etc. to tell the truth. And when
461 * it comes to /\w+/ with tainting enabled, we *must* be able
462 * to trust our character classes.
463 *
464 * Therefore, the default tests in the text of Perl will be
465 * independent of locale. Any code that wants to depend on
466 * the current locale will use the tests that begin with "lc".
467 */
468
2304df62
AD
469#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
470# ifndef CTYPE256
471# define CTYPE256
472# endif
473#endif
474
954c1994 475/*
ccfc67b7
JH
476
477=head1 Character classes
243effed
KW
478This section is about functions (really macros) that classify characters
479into types, such as punctuation versus alphabetic, etc. Most of these are
480analogous to regular expression character classes. (See
481L<perlrecharclass/POSIX Character Classes>.) There are several variants for
482each class. (Not all macros have all variants; each item below lists the
483ones valid for it.) None are affected by C<use bytes>, and only the ones
484with C<LC> in the name are affected by the current locale.
485
486The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
487C<U8>) as input and returns a boolean as to whether or not the character
6aff1f14
KW
488represented by that octet is (or on non-ASCII platforms, corresponds to) an
489ASCII character in the named class based on platform, Unicode, and Perl rules.
490If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 491
6aff1f14
KW
492Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
493with no suffix C<"_A">.
4b9734bf
KW
494
495Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
496onto the platform. That is, the code points that are ASCII are unaffected,
497since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
498as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
499true when called with the code point 0xDF, which is a word character in both
500ASCII and EBCDIC (though it represent different characters in each).
243effed
KW
501
502Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
503point as input. If the code point is larger than 255, Unicode rules are used
504to determine if it is in the character class. For example,
6aff1f14
KW
505C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
506WITH MACRON in Unicode, and is a word character.
243effed
KW
507
508Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
509(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
510classification of just the first (possibly multi-byte) character in the string
511is tested.
243effed
KW
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
6aff1f14
KW
526classification of just the first (possibly multi-byte) character in the string
527is tested.
ccfc67b7 528
954c1994 529=for apidoc Am|bool|isALPHA|char ch
2787a470 530Returns a boolean indicating whether the specified character is an
6aff1f14 531alphabetic character, analogous to C<m/[[:alpha:]]/>.
8a58bdcf 532See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 533C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 534C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 535
15861f94
KW
536=for apidoc Am|bool|isALPHANUMERIC|char ch
537Returns a boolean indicating whether the specified character is a either an
6aff1f14 538alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
15861f94
KW
539See the L<top of this section|/Character classes> for an explanation of variants
540C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 541C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
542C<isALPHANUMERIC_LC_utf8>.
543
8a58bdcf
KW
544=for apidoc Am|bool|isASCII|char ch
545Returns a boolean indicating whether the specified character is one of the 128
243effed 546characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 547On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
548character corresponds to an ASCII character. Variants C<isASCII_A()> and
549C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
550See the L<top of this section|/Character classes> for an explanation of variants
551C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
552C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
553library routine C<isascii()>. In these cases, the variants whose names contain
554C<LC> are the same as the corresponding ones without.
555
d98532ea
KW
556Also note, that because all ASCII characters are UTF-8 invariant (meaning they
557have the exact same representation (always a single byte) whether encoded in
558UTF-8 or not), C<isASCII> will give the correct results when called with any
559byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
560work properly on any string encoded or not in UTF-8.
561
243effed
KW
562=for apidoc Am|bool|isBLANK|char ch
563Returns a boolean indicating whether the specified character is a
6aff1f14 564character considered to be a blank, analogous to C<m/[[:blank:]]/>.
243effed 565See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 566C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
567C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
568platforms do not have the C library routine C<isblank()>. In these cases, the
569variants whose names contain C<LC> are the same as the corresponding ones
570without.
571
572=for apidoc Am|bool|isCNTRL|char ch
573Returns a boolean indicating whether the specified character is a
6aff1f14 574control character, analogous to C<m/[[:cntrl:]]/>.
243effed 575See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 576C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
577C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
578On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 579
954c1994 580=for apidoc Am|bool|isDIGIT|char ch
2787a470 581Returns a boolean indicating whether the specified character is a
6aff1f14 582digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 583Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed 584See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 585C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
586C<isDIGIT_LC_utf8>.
587
588=for apidoc Am|bool|isGRAPH|char ch
589Returns a boolean indicating whether the specified character is a
6aff1f14 590graphic character, analogous to C<m/[[:graph:]]/>.
243effed 591See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 592C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 593C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 594
0c82b6df 595=for apidoc Am|bool|isLOWER|char ch
2787a470 596Returns a boolean indicating whether the specified character is a
6aff1f14 597lowercase character, analogous to C<m/[[:lower:]]/>.
8a58bdcf 598See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 599C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 600C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 601
c99e91e9 602=for apidoc Am|bool|isOCTAL|char ch
2787a470 603Returns a boolean indicating whether the specified character is an
6aff1f14 604octal digit, [0-7].
243effed
KW
605The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
606C<isOCTAL>.
607
608=for apidoc Am|bool|isPUNCT|char ch
609Returns a boolean indicating whether the specified character is a
6aff1f14
KW
610punctuation character, analogous to C<m/[[:punct:]]/>.
611Note that the definition of what is punctuation isn't as
243effed
KW
612straightforward as one might desire. See L<perlrecharclass/POSIX Character
613Classes> for details.
614See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 615C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 616C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 617
0c82b6df 618=for apidoc Am|bool|isSPACE|char ch
2787a470 619Returns a boolean indicating whether the specified character is a
6aff1f14 620whitespace character. This is analogous
398d098a
KW
621to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
622(experimentally), this also matches what C<m/[[:space:]]/> does.
623("Experimentally" means that this change may be backed out in 5.20 or 5.22 if
624field experience indicates that it was unwise.) Prior to 5.18, only the
625locale forms of this macro (the ones with C<LC> in their names) matched
626precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
627in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
628(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
8a58bdcf 629See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 630C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 631C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 632
398d098a
KW
633=for apidoc Am|bool|isPSXSPC|char ch
634(short for Posix Space)
635Starting in 5.18, this is identical (experimentally) in all its forms to the
636corresponding C<isSPACE()> macros. ("Experimentally" means that this change
637may be backed out in 5.20 or 5.22 if field experience indicates that it
638was unwise.)
639The locale forms of this macro are identical to their corresponding
640C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
641non-locale forms differ from their C<isSPACE()> forms only in that the
642C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
643Otherwise they are identical. Thus this macro is analogous to what
644C<m/[[:space:]]/> matches in a regular expression.
645See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 646C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
647C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
648
954c1994 649=for apidoc Am|bool|isUPPER|char ch
2787a470 650Returns a boolean indicating whether the specified character is an
6aff1f14 651uppercase character, analogous to C<m/[[:upper:]]/>.
8a58bdcf 652See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 653C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 654C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 655
243effed 656=for apidoc Am|bool|isPRINT|char ch
8eea39dd 657Returns a boolean indicating whether the specified character is a
6aff1f14 658printable character, analogous to C<m/[[:print:]]/>.
8a58bdcf 659See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 660C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
661C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
662
663=for apidoc Am|bool|isWORDCHAR|char ch
664Returns a boolean indicating whether the specified character is a character
665that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
666in a regular expression. A word character is an alphabetic character, a
667decimal digit, a connecting punctuation character (such as an underscore), or
668a "mark" character that attaches to one of those (like some sort of accent).
669C<isALNUM()> is a synonym provided for backward compatibility, even though a
670word character includes more than the standard C language meaning of
671alphanumeric.
672See the L<top of this section|/Character classes> for an explanation of variants
673C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
674C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
675
676=for apidoc Am|bool|isXDIGIT|char ch
677Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
678digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
679and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
680See the L<top of this section|/Character classes> for an explanation of variants
681C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
682C<isXDIGIT_LC_utf8>.
683
3c3ecf18
KW
684=for apidoc Am|bool|isIDFIRST|char ch
685Returns a boolean indicating whether the specified character can be the first
686character of an identifier. This is very close to, but not quite the same as
687the official Unicode property C<XID_Start>. The difference is that this
688returns true only if the input character also matches L</isWORDCHAR>.
689See the L<top of this section|/Character classes> for an explanation of variants
690C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
691C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
692
693=for apidoc Am|bool|isIDCONT|char ch
694Returns a boolean indicating whether the specified character can be the
695second or succeeding character of an identifier. This is very close to, but
696not quite the same as the official Unicode property C<XID_Continue>. The
697difference is that this returns true only if the input character also matches
698L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
699explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
700C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
701C<isIDCONT_LC_utf8>.
702
243effed 703=head1 Miscellaneous Functions
8eea39dd 704
95a59cab 705=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 706Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
707Behaviour is only well defined when isXDIGIT(*str) is true.
708
e7c1e6c1
KW
709=head1 Character case changing
710
1f607577
KW
711=for apidoc Am|U8|toUPPER|U8 ch
712Converts the specified character to uppercase. If the input is anything but an
713ASCII lowercase character, that input character itself is returned. Variant
c753c8d3 714C<toUPPER_A> is equivalent.
954c1994 715
1f607577
KW
716=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
717Converts the Unicode code point C<cp> to its uppercase version, and
718stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
719that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
720bytes since the uppercase version may be longer than the original character.
721
722The first code point of the uppercased version is returned
723(but note, as explained just above, that there may be more.)
724
725=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
726Converts the UTF-8 encoded character at C<p> to its uppercase version, and
727stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
728that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
729bytes since the uppercase version may be longer than the original character.
730
731The first code point of the uppercased version is returned
732(but note, as explained just above, that there may be more.)
733
734The input character at C<p> is assumed to be well-formed.
735
25200305
KW
736=for apidoc Am|U8|toFOLD|U8 ch
737Converts the specified character to foldcase. If the input is anything but an
738ASCII uppercase character, that input character itself is returned. Variant
739C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
740Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
741
1f607577
KW
742=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
743Converts the Unicode code point C<cp> to its foldcase version, and
744stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
745that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
746bytes since the foldcase version may be longer than the original character.
747
748The first code point of the foldcased version is returned
749(but note, as explained just above, that there may be more.)
750
751=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
752Converts the UTF-8 encoded character at C<p> to its foldcase version, and
753stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
754that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
755bytes since the foldcase version may be longer than the original character.
756
757The first code point of the foldcased version is returned
758(but note, as explained just above, that there may be more.)
759
760The input character at C<p> is assumed to be well-formed.
761
762=for apidoc Am|U8|toLOWER|U8 ch
763Converts the specified character to lowercase. If the input is anything but an
764ASCII uppercase character, that input character itself is returned. Variant
c753c8d3 765C<toLOWER_A> is equivalent.
954c1994 766
1f607577
KW
767=for apidoc Am|U8|toLOWER_L1|U8 ch
768Converts the specified Latin1 character to lowercase. The results are undefined if
769the input doesn't fit in a byte.
770
771=for apidoc Am|U8|toLOWER_LC|U8 ch
772Converts the specified character to lowercase using the current locale's rules,
773if possible; otherwise returns the input character itself.
774
775=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
776Converts the Unicode code point C<cp> to its lowercase version, and
777stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
778that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
779bytes since the lowercase version may be longer than the original character.
780
781The first code point of the lowercased version is returned
782(but note, as explained just above, that there may be more.)
783
784=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
785Converts the UTF-8 encoded character at C<p> to its lowercase version, and
786stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
787that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
788bytes since the lowercase version may be longer than the original character.
789
790The first code point of the lowercased version is returned
791(but note, as explained just above, that there may be more.)
792
793The input character at C<p> is assumed to be well-formed.
794
795=for apidoc Am|U8|toLOWER_LC|U8 ch
796Converts the specified character to lowercase using the current locale's rules,
797if possible; otherwise returns the input character itself.
798
25200305
KW
799=for apidoc Am|U8|toTITLE|U8 ch
800Converts the specified character to titlecase. If the input is anything but an
801ASCII lowercase character, that input character itself is returned. Variant
802C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
803as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
804concept used in locale handling, so there is no functionality for that.)
805
1f607577
KW
806=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
807Converts the Unicode code point C<cp> to its titlecase version, and
808stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
809that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
810bytes since the titlecase version may be longer than the original character.
811
812The first code point of the titlecased version is returned
813(but note, as explained just above, that there may be more.)
814
815=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
816Converts the UTF-8 encoded character at C<p> to its titlecase version, and
817stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
818that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
819bytes since the titlecase version may be longer than the original character.
820
821The first code point of the titlecased version is returned
822(but note, as explained just above, that there may be more.)
823
824The input character at C<p> is assumed to be well-formed.
825
954c1994 826=cut
353c9b6f 827
1f607577 828XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
25200305 829really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
243effed 830
8a58bdcf 831Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
832patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
833
954c1994
GS
834*/
835
7c062697
KW
836/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
837 * is known only in the perl core, and this macro can be called from outside
838 * that */
839#ifdef HAS_QUAD
840# define WIDEST_UTYPE U64TYPE
841#else
842# define WIDEST_UTYPE U32
843#endif
844
3912bc88
KW
845/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
846 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
847 * bits of information are lost even on a 64-bit machine, but to get the
848 * compiler to optimize it out if possible. This is because Configure makes
849 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
850 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
851 */
852#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 853
41f43cc2 854#ifdef EBCDIC
b6340bd0
KW
855# ifndef _ALL_SOURCE
856 /* This returns the wrong results on at least z/OS unless this is
857 * defined. */
858# error _ALL_SOURCE should probably be defined
859# endif
860
861 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
862 * likely not defined, and so we use the native function */
863# define isASCII(c) isascii(c)
41f43cc2
KW
864#else
865# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
866#endif
867
ec79cfbb 868#define isASCII_A(c) isASCII(c)
3509bb53 869#define isASCII_L1(c) isASCII(c)
6515cb76 870
c2da0b36
KW
871/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
872 * and the 8 possible permutations of those bits exactly comprise the 8 octal
873 * digits */
3d33b3e3 874#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 875
ec79cfbb 876/* ASCII range only */
9fb1bf9d 877#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 878
a500dc72
KW
879/* Character class numbers. For internal core Perl use only. The ones less
880 * than 32 are used in PL_charclass[] and the ones up through the one that
881 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
882 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
883 * their actual definitions are here. If that file has a name not used here,
884 * it won't compile.
1709d539
KW
885 *
886 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
887 * frequency of their use. This gives a slight edge to exiting a loop earlier
888 * (in reginclass() in regexec.c) */
1709d539
KW
889# define _CC_WORDCHAR 0 /* \w and [:word:] */
890# define _CC_DIGIT 1 /* \d and [:digit:] */
891# define _CC_ALPHA 2 /* [:alpha:] */
892# define _CC_LOWER 3 /* [:lower:] */
893# define _CC_UPPER 4 /* [:upper:] */
894# define _CC_PUNCT 5 /* [:punct:] */
895# define _CC_PRINT 6 /* [:print:] */
15861f94 896# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 897# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 898# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 899
b0d691b2 900#define _FIRST_NON_SWASH_CC 10
1709d539
KW
901/* The character classes above are implemented with swashes. The second group
902 * (just below) contains the ones implemented without. These are also sorted
903 * in rough order of the frequency of their use, except that \v should be last,
904 * as it isn't a real Posix character class, and some (small) inefficiencies in
905 * regular expression handling would be introduced by putting it in the middle
906 * of those that are. Also, cntrl and ascii come after the others as it may be
907 * useful to group these which have no members that match above Latin1, (or
908 * above ASCII in the latter case) */
909
b0d691b2
KW
910# define _CC_SPACE 10 /* \s */
911# define _CC_BLANK 11 /* [:blank:] */
912# define _CC_XDIGIT 12 /* [:xdigit:] */
913# define _CC_PSXSPC 13 /* [:space:] */
914# define _CC_CNTRL 14 /* [:cntrl:] */
915# define _CC_ASCII 15 /* [:ascii:] */
916# define _CC_VERTSPACE 16 /* \v */
1709d539 917
a0947d7b
KW
918# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
919
1709d539 920/* The members of the third group below do not need to be coordinated with data
3ded5eb0
KW
921 * structures in regcomp.[ch] and regexec.c. But they should be added to
922 * bootstrap_ctype() */
b0d691b2
KW
923# define _CC_IDFIRST 17
924# define _CC_CHARNAME_CONT 18
925# define _CC_NONLATIN1_FOLD 19
926# define _CC_QUOTEMETA 20
927# define _CC_NON_FINAL_FOLD 21
928# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
929# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
930/* Unused: 23-30
f4cdb42c
KW
931 * If more bits are needed, one could add a second word for non-64bit
932 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
933 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
934 * is used only for optimization (as of this writing), and differs in the
935 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
936 * characters: the masculine and feminine ordinal indicators, so removing it
937 * would just cause /i regexes which match them to run less efficiently */
96ac0975 938
3a371f2f
KW
939#if defined(PERL_CORE) || defined(PERL_EXT)
940/* An enum version of the character class numbers, to help compilers
941 * optimize */
942typedef enum {
3a371f2f 943 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
944 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
945 _CC_ENUM_ASCII = _CC_ASCII,
946 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 947 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 948 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
949 _CC_ENUM_DIGIT = _CC_DIGIT,
950 _CC_ENUM_GRAPH = _CC_GRAPH,
951 _CC_ENUM_LOWER = _CC_LOWER,
952 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 953 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 954 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 955 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 956 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 957 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 958 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 959 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
960} _char_class_number;
961#endif
962
63c61c3f 963#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 964#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 965
3018b823 966#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 967# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
968 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
969 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
970 #error Need to adjust order of swash_property_names[]
971# endif
972
b44daeeb
KW
973/* This is declared static in each of the few files that this is #defined for
974 * to keep them from being publicly accessible. Hence there is a small amount
975 * of wasted space */
976
922e8cb4
KW
977static const char* const swash_property_names[] = {
978 "XPosixWord",
979 "XPosixDigit",
980 "XPosixAlpha",
981 "XPosixLower",
982 "XPosixUpper",
983 "XPosixPunct",
984 "XPosixPrint",
985 "XPosixAlnum",
b0d691b2
KW
986 "XPosixGraph",
987 "Cased"
922e8cb4
KW
988};
989#endif
990
96ac0975
NC
991# ifdef DOINIT
992EXTCONST U32 PL_charclass[] = {
993# include "l1_char_class_tab.h"
994};
995
996# else /* ! DOINIT */
997EXTCONST U32 PL_charclass[];
998# endif
3ded5eb0
KW
999#endif /* Has perl.h */
1000
1001#if defined(H_PERL) && ! defined(BOOTSTRAP_CHARSET)
96ac0975 1002
265c1f46 1003 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70
KW
1004# define _CC_mask(classnum) (1U << (classnum))
1005# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
f4cd282c 1006 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
4eeeb416 1007
f4cdb42c
KW
1008 /* The mask for the _A versions of the macros; it just adds in the bit for
1009 * ASCII. */
1010# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1011
1012 /* The _A version makes sure that both the desired bit and the ASCII bit
1013 * are present */
3ded5eb0 1014# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
f4cd282c 1015 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
f4cdb42c
KW
1016 == _CC_mask_A(classnum)))
1017
f4cdb42c 1018# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 1019# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
1020# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1021# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1022# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
1023# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1024# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1025# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1026# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1027# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1028# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1029# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1030# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1031# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
d95f8b6a 1032# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
3ded5eb0
KW
1033# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1034# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1035# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1036
1037 /* continuation character for legal NAME in \N{NAME} */
1038# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1039
1040# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1041# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1042# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1043# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1044# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1045# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1046# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1047# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1048# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1049# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
f4cdb42c 1050
fb9bfbf7
KW
1051 /* Either participates in a fold with a character above 255, or is a
1052 * multi-char fold */
f4cd282c 1053# 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 1054
4eeeb416 1055# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
1056# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1057 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
1058# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1059 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
3ded5eb0
KW
1060#else /* Either don't have perl.h or don't want to use char_class_tab.h */
1061
1062 /* If we don't have perl.h, we are compiling a utility program. Below we
1063 * hard-code various macro definitions that wouldn't otherwise be available
1064 * to it. We can also get here if we are configured to bootstrap up Perl
1065 * on a non-ASCII platform that doesn't have a working Perl (currently only
1066 * EBCDIC). For these we currently use the native definitions to get
1067 * things going. (It should also be possible to use the translation
1068 * function NATIVE_TO_LATIN1(), but that is an extra layer of dependence on
1069 * Perl, so it is currently avoided for the macros where it's possible to
1070 * do so.) */
9fb1bf9d 1071# ifdef EBCDIC
3ded5eb0 1072 /* Use the native functions. They likely will return false for all
4f48cecf 1073 * non-ASCII values, but this makes sure */
b6340bd0
KW
1074# define isALPHA_A(c) (isASCII(c) && isalpha(c))
1075# define isALPHANUMERIC_A(c) (isASCII(c) && isalnum(c))
1076# define isCNTRL_A(c) (isASCII(c) && iscntrl(c))
1077# define isDIGIT_A(c) (isASCII(c) && isdigit(c))
1078# define isGRAPH_A(c) (isASCII(c) && isgraph(c))
1079# define isLOWER_A(c) (isASCII(c) && islower(c))
1080# define isPRINT_A(c) (isASCII(c) && isprint(c))
1081# define isPUNCT_A(c) (isASCII(c) && ispunct(c))
1082# define isSPACE_A(c) (isASCII(c) && isspace(c))
1083# define isUPPER_A(c) (isASCII(c) && isupper(c))
1084# define isXDIGIT_A(c) (isASCII(c) && isxdigit(c))
3ded5eb0 1085# else /* ASCII platform. These are coded based on first principals */
9fb1bf9d 1086# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
15861f94 1087# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
3ded5eb0 1088# define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
2181da29 1089# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
3ded5eb0 1090# define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
9fb1bf9d
KW
1091# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1092# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
3ded5eb0
KW
1093# define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1094# define isSPACE_A(c) ((c) == ' ' \
1095 || (c) == '\t' \
1096 || (c) == '\n' \
1097 || (c) == '\r' \
1098 || (c) =='\v' \
926bb54c 1099 || (c) == '\f')
2181da29 1100# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
3ded5eb0
KW
1101# define isXDIGIT_A(c) (isDIGIT_A(c) \
1102 || ((c) >= 'a' && (c) <= 'f') \
1103 || ((c) <= 'F' && (c) >= 'A'))
1104# endif /* Below are common definitions for ASCII and non-ASCII */
1105# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1106# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1107# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1108
1109 /* The _L1 macros may be unnecessary for both the utilities and for
1110 * bootstrapping; I (khw) added them during debugging of bootstrapping, and
1111 * it seems best to keep them. */
1112# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1113# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1114# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1115# define isBLANK_L1(c) (isBLANK_A(c) \
1116 || (FITS_IN_8_BITS(c) \
1117 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1118# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1119# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1120# define isLOWER_L1(c) (isLOWER_A(c) \
1121 || (FITS_IN_8_BITS(c) \
1122 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1123 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1124 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1125 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1126 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1127# define isPRINT_L1(c) (isPRINT_A(c) \
1128 || (FITS_IN_8_BITS(c) \
1129 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1130# define isPSXSPC_L1(c) isSPACE_L1(c)
1131# define isPUNCT_L1(c) (isPUNCT_A(c) \
1132 || (FITS_IN_8_BITS(c) \
1133 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1134 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1135 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1136 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1137 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1138 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1139 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1140# define isSPACE_L1(c) (isSPACE_A(c) \
1141 || (FITS_IN_8_BITS(c) \
1142 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1143 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1144# define isUPPER_L1(c) (isUPPER_A(c) \
1145 || (FITS_IN_8_BITS(c) \
1146 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1147 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1148 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1149# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1150# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1151# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1152 || isBLANK_L1(c) \
1153 || (c) == '-' \
1154 || (c) == '(' \
1155 || (c) == ')')
1156 /* The following are not fully accurate in the above-ASCII range. I (khw)
1157 * don't think it's necessary to be so for the purposes where this gets
1158 * compiled */
1159# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1160# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1161
1162 /* And these aren't accurate at all. They are useful only for above
1163 * Latin1, which utilities and bootstrapping don't deal with */
1164# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1165# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1166
1167 /* Many of the macros later in this file are defined in terms of these. By
1168 * implementing them with a function, which converts the class number into
1169 * a call to the desired macro, all of the later ones work. However, that
1170 * function won't be actually defined when building a utility program (no
1171 * perl.h), and so a compiler error will be generated if one is attempted
1172 * to be used. And the above-Latin1 code points require Unicode tables to
1173 * be present, something unlikely to be the case when bootstrapping */
1174# define _generic_isCC(c, classnum) \
1175 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1176# define _generic_isCC_A(c, classnum) \
1177 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
1178#endif /* End of no perl.h or have BOOTSTRAP_CHARSET */
8a58bdcf 1179
e66b99e9
KW
1180#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1181#define isALPHA(c) isALPHA_A(c)
1182#define isBLANK(c) isBLANK_A(c)
1183#define isCNTRL(c) isCNTRL_A(c)
1184#define isDIGIT(c) isDIGIT_A(c)
1185#define isGRAPH(c) isGRAPH_A(c)
1186#define isIDFIRST(c) isIDFIRST_A(c)
1187#define isLOWER(c) isLOWER_A(c)
1188#define isPRINT(c) isPRINT_A(c)
1189#define isPSXSPC(c) isPSXSPC_A(c)
1190#define isPUNCT(c) isPUNCT_A(c)
1191#define isSPACE(c) isSPACE_A(c)
1192#define isUPPER(c) isUPPER_A(c)
1193#define isWORDCHAR(c) isWORDCHAR_A(c)
1194#define isXDIGIT(c) isXDIGIT_A(c)
1195
1196/* ASCII casing. These could also be written as
1197 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1198 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1199 which uses table lookup and mask instead of subtraction. (This would
1200 work because the _MOD does not apply in the ASCII range) */
68067e4e
DM
1201#define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1202#define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
bbce6d69 1203
25200305
KW
1204/* In the ASCII range, these are equivalent to what they're here defined to be.
1205 * But by creating these definitions, other code doesn't have to be aware of
1206 * this detail */
1207#define toFOLD(c) toLOWER(c)
1208#define toFOLD_LC(c) toLOWER_LC(c)
1209#define toTITLE(c) toUPPER(c)
1210
c753c8d3
KW
1211#define toLOWER_A(c) toLOWER(c)
1212#define toUPPER_A(c) toUPPER(c)
25200305
KW
1213#define toFOLD_A(c) toFOLD(c)
1214#define toTITLE_A(c) toTITLE(c)
1a0901db
KW
1215
1216/* Use table lookup for speed; return error character for input
1217 * out-of-range */
b2bf251f 1218#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1219 ? (c) \
f4cd282c 1220 : PL_latin1_lc[ (U8) (c) ])
c753c8d3
KW
1221#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1222
1a0901db
KW
1223/* Modified uc. Is correct uc except for three non-ascii chars which are
1224 * all mapped to one of them, and these need special handling; error
1225 * character for input out-of-range */
b2bf251f 1226#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1227 ? (c) \
f4cd282c 1228 : PL_mod_latin1_uc[ (U8) (c) ])
bbce6d69
PP
1229#ifdef USE_NEXT_CTYPE
1230
15861f94 1231# define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c))
ff68c719 1232# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 1233# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 1234# define isBLANK_LC(c) isBLANK((unsigned int)(c))
b8c5462f 1235# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
509fb054 1236# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
b8c5462f 1237# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
509fb054
KW
1238# define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
1239# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
ff68c719 1240# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 1241# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
509fb054
KW
1242# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
1243# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
1244# define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
e712593e 1245# define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
ff68c719 1246# define toLOWER_LC(c) NXToLower((unsigned int)(c))
509fb054 1247# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
bbce6d69
PP
1248
1249#else /* !USE_NEXT_CTYPE */
b8c5462f 1250
bbce6d69
PP
1251# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1252
76fde96a 1253/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
851131d3 1254
76fde96a 1255# define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
15861f94
KW
1256# define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \
1257 && isalnum((unsigned char)(c)))
07315176 1258# ifdef HAS_ISASCII
76fde96a 1259# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
07315176 1260# else
76fde96a 1261# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
07315176 1262# endif
67addccf 1263# ifdef HAS_ISBLANK
76fde96a 1264# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
67addccf 1265# else
76fde96a 1266# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
67addccf 1267# endif
76fde96a 1268# define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
509fb054 1269# define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
76fde96a 1270# define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
509fb054
KW
1271# define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1272 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1273# define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
76fde96a
KW
1274# define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1275# define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
509fb054
KW
1276# define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1277# define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1278# define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1279 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
e712593e 1280# define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
548d896a
DM
1281# define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? (UV)tolower((unsigned char)(c)) : (c))
1282# define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? (UV)toupper((unsigned char)(c)) : (c))
bbce6d69
PP
1283
1284# else
1285
bbce6d69 1286# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1287# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
07315176 1288# define isASCII_LC(c) isascii(c)
67addccf
KW
1289# ifdef HAS_ISBLANK
1290# define isBLANK_LC(c) (isascii(c) && isblank(c))
1291# else
3e66ab74 1292# define isBLANK_LC(c) isBLANK_A(c)
67addccf 1293# endif
b8c5462f 1294# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1295# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1296# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1297# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1298# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1299# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1300# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1301# define isSPACE_LC(c) (isascii(c) && isspace(c))
1302# define isUPPER_LC(c) (isascii(c) && isupper(c))
1303# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1304# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
76fde96a 1305# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1306# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
bbce6d69
PP
1307
1308# endif
a0d0e21e 1309#endif /* USE_NEXT_CTYPE */
55204971 1310
eba68aa0
KW
1311#define isIDCONT(c) isWORDCHAR(c)
1312#define isIDCONT_A(c) isWORDCHAR_A(c)
1313#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1314#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1315#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1316
66c17564 1317/* For internal core Perl use only. If the input is Latin1, use the Latin1
a500dc72
KW
1318 * macro; otherwise use the function 'above_latin1'. Won't compile if 'c' isn't unsigned, as
1319 * won't match above_latin1 prototype. The macros do bounds checking, so have
66c17564
KW
1320 * duplicate checks here, so could create versions of the macros that don't,
1321 * but experiments show that gcc optimizes them out anyway. */
1322
1323/* Note that all ignore 'use bytes' */
cd500f2f
KW
1324#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1325 ? _generic_isCC(c, classnum) \
1326 : above_latin1(c))
1327#define _generic_swash_uni(classnum, c) ((c) < 256 \
1328 ? _generic_isCC(c, classnum) \
922e8cb4 1329 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1330#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1331#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1332#define isASCII_uni(c) isASCII(c)
1333#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1334#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1335#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1336#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1337#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1338#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1339#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1340#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1341
509fb054
KW
1342/* Posix and regular space are identical above Latin1 */
1343#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1344
cd500f2f 1345#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1346#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1347#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1348#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1349#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1350#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1351
b8d68ded 1352#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1353#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1354#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1355#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1356
cd500f2f
KW
1357#define _generic_LC_uvchr(latin1, above_latin1, c) \
1358 (c < 256 ? latin1(c) : above_latin1(c))
1359#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1360 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1361
1362#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1363#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1364 _CC_ALPHANUMERIC, c)
509fb054 1365#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1366#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1367#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1368#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1369#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1370#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1371 _is_uni_perl_idcont, c)
cd500f2f
KW
1372#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1373 _is_uni_perl_idstart, c)
1374#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1375#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1376#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1377 space under locale */
cd500f2f
KW
1378#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1379#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1380 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1381#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1382#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1383 _CC_WORDCHAR, c)
1384#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1385
36b00f93 1386#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1387
a500dc72
KW
1388/* Everything whose name begins with an underscore is for internal core Perl
1389 * use only. */
1390
1391/* If the input is in the Latin1 range, use
252810d0
KW
1392 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1393 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1394 * fact that ASCII characters have the same representation whether utf8 or not.
1395 * Note that it assumes that the utf8 has been validated, and ignores 'use
1396 * bytes' */
cd500f2f 1397#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1398 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1399 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c 1400 ? _generic_isCC( \
94bb8c36 1401 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
aea0c08c 1402 *((p)+1 )), \
94bb8c36 1403 classnum) \
252810d0 1404 : utf8)
cd500f2f
KW
1405/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1406 * can be a macro */
1407#define _generic_func_utf8(classnum, above_latin1, p) \
1408 _generic_utf8(classnum, p, above_latin1(p))
1409/* Like the above, but passes classnum to _isFOO_utf8(), instead of having a
1410 * 'above_latin1' parameter */
922e8cb4 1411#define _generic_swash_utf8(classnum, p) \
cd500f2f 1412 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1413
cc8ab7c0
KW
1414/* Like the above, but should be used only when it is known that there are no
1415 * characters in the range 128-255 which the class is TRUE for. Hence it can
cd500f2f 1416 * skip the tests for this range. 'above_latin1' should include its arguments */
3ded5eb0 1417#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1418 (UTF8_IS_INVARIANT(*(p)) \
1419 ? _generic_isCC(*(p), classnum) \
1420 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
3ded5eb0 1421 ? above_latin1 \
cc8ab7c0
KW
1422 : 0)
1423
84238efa
KW
1424/* NOTE that some of these macros have very similar ones in regcharclass.h.
1425 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1426 * there, differing in name only by an underscore from the one here
1427 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1428 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1429 * points; the regcharclass.h ones are implemented as a series of
1430 * "if-else-if-else ..." */
1431
922e8cb4
KW
1432#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1433#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1434#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1435 utf8, the non-utf8 macro works
1436 */
cd500f2f
KW
1437#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1438#define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1439#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1440 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1441#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1442#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1443 _is_utf8_perl_idcont, p)
e5dcd934 1444
c11ff943
KW
1445/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1446 * IDFIRST is an alnum. See
c8362b00 1447 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1448 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1449 * This used to be not the XID version, but we decided to go with the more
1450 * modern Unicode definition */
cd500f2f 1451#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1452 _is_utf8_perl_idstart, p)
e5dcd934 1453
922e8cb4
KW
1454#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1455#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1456
509fb054 1457/* Posix and regular space are identical above Latin1 */
cd500f2f 1458#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1459
922e8cb4 1460#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1461#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1462#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1463#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1464#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1465#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1466 is_XDIGIT_high(p))
a0ed51b3 1467
7d2b61dd 1468#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
509fb054
KW
1469#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1470#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1471#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1472
9201a572
KW
1473/* For internal core Perl use only. If the input is in the Latin1 range, use
1474 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1475 * use the value given by the 'utf8' parameter. This relies on the fact that
1476 * ASCII characters have the same representation whether utf8 or not. Note
1477 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
94bb8c36
KW
1478#define _generic_LC_utf8(macro, p, utf8) \
1479 (UTF8_IS_INVARIANT(*(p)) \
1480 ? macro(*(p)) \
1481 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1482 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
9201a572
KW
1483 : utf8)
1484
cd500f2f
KW
1485#define _generic_LC_swash_utf8(macro, classnum, p) \
1486 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1487#define _generic_LC_func_utf8(macro, above_latin1, p) \
1488 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1489
cd500f2f
KW
1490#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1491 _CC_ALPHANUMERIC, p)
1492#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1493#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1494#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1495#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1496#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1497#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1498#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1499#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1500#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1501#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1502#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1503 space under locale */
cd500f2f
KW
1504#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1505#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1506#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1507#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1508 _CC_WORDCHAR, p)
1509#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1510
fbc19f27
KW
1511/* Macros for backwards compatibility and for completeness when the ASCII and
1512 * Latin1 values are identical */
fbc19f27
KW
1513#define isALPHAU(c) isALPHA_L1(c)
1514#define isDIGIT_L1(c) isDIGIT_A(c)
1515#define isOCTAL(c) isOCTAL_A(c)
1516#define isOCTAL_L1(c) isOCTAL_A(c)
1517#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1518#define isALNUM(c) isWORDCHAR(c)
1519#define isALNUMU(c) isWORDCHAR_L1(c)
1520#define isALNUM_LC(c) isWORDCHAR_LC(c)
1521#define isALNUM_uni(c) isWORDCHAR_uni(c)
1522#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1523#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1524#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1525#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1526#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1527#define isALNUMC(c) isALPHANUMERIC(c)
1528#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1529#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1530#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1531#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1532#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1533
6ae9f32a 1534/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1535 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1536 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1537 * that is EOT on ASCII; ST on EBCDIC */
8e7c6e7d 1538# define toCTRL(c) (toUPPER(NATIVE_TO_LATIN1(c)) ^ 64)
bbce6d69 1539
dea28490
JJ
1540/* Line numbers are unsigned, 32 bits. */
1541typedef U32 line_t;
e5dcd934 1542#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1543
91152fc1
DG
1544/* Helpful alias for version prescan */
1545#define is_LAX_VERSION(a,b) \
1546 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1547
1548#define is_STRICT_VERSION(a,b) \
1549 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1550
1551#define BADVERSION(a,b,c) \
1552 if (b) { \
1553 *b = c; \
1554 } \
1555 return a;
8c52afec 1556
14f6cdfd
KW
1557/* Converts a hex digit in a string to its numeric value, advancing the
1558 * pointer. The input must be known to be 0-9, A-F, or a-f. In both ASCII and
1559 * EBCDIC the last 4 bits of the digits are 0-9; and the last 4 bits of A-F and
1560 * a-f are 1-6, so adding 9 yields 10-15 */
1561#define READ_XDIGIT(s) (0xf & (isDIGIT(*(s)) ? (*(s)++) : (*(s)++ + 9)))
95a59cab 1562
8e84507e 1563/*
ccfc67b7
JH
1564=head1 Memory Management
1565
a02a5408 1566=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1567The XSUB-writer's interface to the C C<malloc> function.
1568
0d7b2759
MH
1569Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1570
c5008215
JC
1571In 5.9.3, Newx() and friends replace the older New() API, and drops
1572the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1573themselves. This aid has been superseded by a new build option,
d10b4965 1574PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1575there for use in XS modules supporting older perls.
1576
a02a5408 1577=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1578The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1579cast. See also C<Newx>.
954c1994 1580
0d7b2759
MH
1581Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1582
a02a5408 1583=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1584The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1585memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1586
0d7b2759
MH
1587Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1588
954c1994
GS
1589=for apidoc Am|void|Renew|void* ptr|int nitems|type
1590The XSUB-writer's interface to the C C<realloc> function.
1591
0d7b2759
MH
1592Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1593
954c1994
GS
1594=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1595The XSUB-writer's interface to the C C<realloc> function, with
1596cast.
1597
0d7b2759
MH
1598Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1599
49b8b560 1600=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1601The XSUB-writer's interface to the C C<free> function.
1602
0d7b2759
MH
1603This should B<ONLY> be used on memory obtained using L<"Newx"> and friends.
1604
954c1994
GS
1605=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1606The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1607source, C<dest> is the destination, C<nitems> is the number of items, and
1608C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1609
e90e2364 1610=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
72d33970
FC
1611Like C<Move> but returns dest. Useful
1612for encouraging compilers to tail-call
e90e2364
NC
1613optimise.
1614
954c1994
GS
1615=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1616The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1617source, C<dest> is the destination, C<nitems> is the number of items, and
1618C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1619
e90e2364
NC
1620=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1621
72d33970
FC
1622Like C<Copy> but returns dest. Useful
1623for encouraging compilers to tail-call
e90e2364
NC
1624optimise.
1625
954c1994
GS
1626=for apidoc Am|void|Zero|void* dest|int nitems|type
1627
1628The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1629destination, C<nitems> is the number of items, and C<type> is the type.
1630
e90e2364
NC
1631=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1632
72d33970
FC
1633Like C<Zero> but returns dest. Useful
1634for encouraging compilers to tail-call
e90e2364
NC
1635optimise.
1636
da5d8dbb 1637=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1638This is an architecture-independent macro to copy one structure to another.
954c1994 1639
7e337ee0
JH
1640=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1641
1642Fill up memory with a byte pattern (a byte repeated over and over
1643again) that hopefully catches attempts to access uninitialized memory.
1644
1645=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1646
1647PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1648
1c12ffb4 1649=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1650
1651PoisonWith(0xEF) for catching access to freed memory.
1652
9965345d
JH
1653=for apidoc Am|void|Poison|void* dest|int nitems|type
1654
7e337ee0 1655PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1656
1657=cut */
954c1994 1658
561b68a9
SH
1659/* Maintained for backwards-compatibility only. Use newSV() instead. */
1660#ifndef PERL_CORE
ff06c60c 1661#define NEWSV(x,len) newSV(len)
561b68a9 1662#endif
ff06c60c 1663
19a94d75
JH
1664#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1665
2720f793
JH
1666/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1667 * overly eager compilers that will bleat about e.g.
1668 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1669#ifdef PERL_MALLOC_WRAP
f019c49e 1670#define MEM_WRAP_CHECK(n,t) \
d1decf2b 1671 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
27d5b266 1672#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1673 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1674#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1675
d1decf2b 1676#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
1677#else
1678
410319be
NC
1679#define MEM_WRAP_CHECK(n,t)
1680#define MEM_WRAP_CHECK_1(n,t,a)
1681#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1682#define MEM_WRAP_CHECK_(n,t)
1683
6fff79ce 1684#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1685
1936d2a7 1686#endif
8b44ba4c 1687
fe4f188c 1688#ifdef PERL_MEM_LOG
46c6c7e2 1689/*
9f653bb5 1690 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1691 * go through functions, which are handy for debugging breakpoints, but
1692 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1693 * line number, and C function name if available) passed in. This info can
1694 * then be used for logging the calls, for which one gets a sample
73d1d973 1695 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1696 *
46c6c7e2 1697 * Known problems:
94e892a6 1698 * - not all memory allocs get logged, only those
46c6c7e2 1699 * that go through Newx() and derivatives (while all
94e892a6 1700 * Safefrees do get logged)
46c6c7e2
JH
1701 * - __FILE__ and __LINE__ do not work everywhere
1702 * - __func__ or __FUNCTION__ even less so
1703 * - I think more goes on after the perlio frees but
1704 * the thing is that STDERR gets closed (as do all
1705 * the file descriptors)
1706 * - no deeper calling stack than the caller of the Newx()
1707 * or the kind, but do I look like a C reflection/introspection
1708 * utility to you?
1709 * - the function prototypes for the logging functions
1710 * probably should maybe be somewhere else than handy.h
1711 * - one could consider inlining (macrofying) the logging
1712 * for speed, but I am too lazy
1713 * - one could imagine recording the allocations in a hash,
1714 * (keyed by the allocation address?), and maintain that
1715 * through reallocs and frees, but how to do that without
1716 * any News() happening...?
73d1d973 1717 * - lots of -Ddefines to get useful/controllable output
b953482e 1718 * - lots of ENV reads
46c6c7e2
JH
1719 */
1720
12754f92 1721PERL_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 1722
12754f92 1723PERL_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 1724
12754f92 1725PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1726
0b0ab801 1727# ifdef PERL_CORE
73d1d973 1728# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1729enum mem_log_type {
1730 MLT_ALLOC,
1731 MLT_REALLOC,
d7a2c63c
MHM
1732 MLT_FREE,
1733 MLT_NEW_SV,
1734 MLT_DEL_SV
0b0ab801
MHM
1735};
1736# endif
12754f92 1737# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1738void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1739void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1740# endif
0b0ab801
MHM
1741# endif
1742
fe4f188c
JH
1743#endif
1744
1745#ifdef PERL_MEM_LOG
d1401ee9
MHM
1746#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1747#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1748#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1749#endif
1750
1751#ifndef MEM_LOG_ALLOC
1752#define MEM_LOG_ALLOC(n,t,a) (a)
1753#endif
1754#ifndef MEM_LOG_REALLOC
1755#define MEM_LOG_REALLOC(n,t,v,a) (a)
1756#endif
1757#ifndef MEM_LOG_FREE
1758#define MEM_LOG_FREE(a) (a)
1759#endif
1760
d1401ee9
MHM
1761#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1762#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1763#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1764
1765#ifndef PERL_CORE
a02a5408
JC
1766/* pre 5.9.x compatibility */
1767#define New(x,v,n,t) Newx(v,n,t)
1768#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1769#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1770#endif
a02a5408 1771
ff68c719 1772#define Renew(v,n,t) \
d1401ee9 1773 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1774#define Renewc(v,n,t,c) \
d1401ee9 1775 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1776
1777#ifdef PERL_POISON
1778#define Safefree(d) \
06c0cc96 1779 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1780#else
fe4f188c 1781#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1782#endif
55497cff 1783
8b44ba4c
NC
1784#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1785#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1786#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1787
8b44ba4c
NC
1788#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1789#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1790#ifdef HAS_MEMSET
8b44ba4c 1791#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1792#else
8b44ba4c
NC
1793/* Using bzero(), which returns void. */
1794#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1795#endif
1796
7e337ee0
JH
1797#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1798#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1799#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1800#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1801
caa674f3
DD
1802#ifdef PERL_POISON
1803# define PERL_POISON_EXPR(x) x
1804#else
1805# define PERL_POISON_EXPR(x)
1806#endif
1807
2304df62 1808#ifdef USE_STRUCT_COPY
ff68c719 1809#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1810#else
1811#define StructCopy(s,d,t) Copy(s,d,1,t)
1812#endif
2cc61e15 1813
622913ab 1814#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1815#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1816
2cc61e15
DD
1817#ifdef NEED_VA_COPY
1818# ifdef va_copy
1819# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1820# else
a1866d1b
JH
1821# if defined(__va_copy)
1822# define Perl_va_copy(s, d) __va_copy(d, s)
1823# else
1824# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1825# endif
2cc61e15
DD
1826# endif
1827#endif
1828
472d47bc
SB
1829/* convenience debug macros */
1830#ifdef USE_ITHREADS
1831#define pTHX_FORMAT "Perl interpreter: 0x%p"
1832#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1833#define pTHX_VALUE_ (void *)my_perl,
1834#define pTHX_VALUE (void *)my_perl
1835#define pTHX__VALUE_ ,(void *)my_perl,
1836#define pTHX__VALUE ,(void *)my_perl
472d47bc 1837#else
3609ea0d 1838#define pTHX_FORMAT
472d47bc 1839#define pTHX__FORMAT
3609ea0d 1840#define pTHX_VALUE_
472d47bc 1841#define pTHX_VALUE
3609ea0d 1842#define pTHX__VALUE_
472d47bc
SB
1843#define pTHX__VALUE
1844#endif /* USE_ITHREADS */
3609ea0d 1845
2acdbac1
NC
1846/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1847 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1848 CPAN::Unpack show any users outside the core. */
1849#ifdef PERL_CORE
d1d15184 1850# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1851#endif
1852
dfff4baf
BF
1853/* Internal macros to deal with gids and uids */
1854#ifdef PERL_CORE
1855
1856# if Uid_t_size > IVSIZE
1857# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
1858# define SvUID(sv) SvNV(sv)
1859# else
1860# if Uid_t_sign <= 0
1861# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
1862# define SvUID(sv) SvIV(sv)
1863# else
1864# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
1865# define SvUID(sv) SvUV(sv)
1866# endif
1867# endif /* Uid_t_size */
1868
1869# if Gid_t_size > IVSIZE
1870# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
1871# define SvGID(sv) SvNV(sv)
1872# else
1873# if Gid_t_sign <= 0
1874# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
1875# define SvGID(sv) SvIV(sv)
1876# else
1877# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
1878# define SvGID(sv) SvUV(sv)
1879# endif
1880# endif /* Gid_t_size */
1881
1882#endif
1883
9d745869
KW
1884#endif /* HANDY_H */
1885
e9a8c099
MHM
1886/*
1887 * Local variables:
1888 * c-indentation-style: bsd
1889 * c-basic-offset: 4
14d04a33 1890 * indent-tabs-mode: nil
e9a8c099
MHM
1891 * End:
1892 *
14d04a33 1893 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1894 */