This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
deal with assignment to $/ better, deprecate edge cases, and forbid others
[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
4650c663
KW
11/* IMPORTANT NOTE: Everything whose name begins with an underscore is for
12 * internal core Perl use only. */
13
9d745869
KW
14#ifndef HANDY_H /* Guard against nested #inclusion */
15#define HANDY_H
16
85e6fe83 17#if !defined(__STDC__)
378cc40b
LW
18#ifdef NULL
19#undef NULL
20#endif
a687059c 21# define NULL 0
85e6fe83
LW
22#endif
23
24792b8d
NC
24#ifndef PERL_CORE
25# define Null(type) ((type)NULL)
954c1994
GS
26
27/*
ccfc67b7 28=head1 Handy Values
954c1994 29
3609ea0d 30=for apidoc AmU||Nullch
72d33970
FC
31Null character pointer. (No longer available when C<PERL_CORE> is
32defined.)
2307c6d0 33
954c1994 34=for apidoc AmU||Nullsv
72d33970 35Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
36
37=cut
38*/
39
24792b8d
NC
40# define Nullch Null(char*)
41# define Nullfp Null(PerlIO*)
42# define Nullsv Null(SV*)
43#endif
8d063cd8 44
641d3f0b
PP
45#ifdef TRUE
46#undef TRUE
47#endif
48#ifdef FALSE
49#undef FALSE
50#endif
51#define TRUE (1)
52#define FALSE (0)
53
cf3f0ffb
DM
54/* The MUTABLE_*() macros cast pointers to the types shown, in such a way
55 * (compiler permitting) that casting away const-ness will give a warning;
56 * e.g.:
57 *
58 * const SV *sv = ...;
59 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
60 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
61 */
62
b1bc3f34 63#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 64# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
65#else
66# define MUTABLE_PTR(p) ((void *) (p))
67#endif
68
a062e10d 69#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 70#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 71#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 72#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 73#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 74#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 75
f789f6a4 76#if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR)
bd31be4b
NC
77# include <stdbool.h>
78# ifndef HAS_BOOL
79# define HAS_BOOL 1
80# endif
81#endif
82
8e84507e 83/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
84 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
85 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 86 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
87 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
88 g++ can be identified by __GNUG__.
89 Andy Dougherty February 2000
5d94fbed 90*/
3609ea0d 91#ifdef __GNUG__ /* GNU g++ has bool built-in */
f789f6a4 92# ifndef PERL_BOOL_AS_CHAR
5d94fbed 93# ifndef HAS_BOOL
c1d22f6b 94# define HAS_BOOL 1
5d94fbed 95# endif
f789f6a4 96# endif
5d94fbed
AD
97#endif
98
641d3f0b
PP
99/* The NeXT dynamic loader headers will not build with the bool macro
100 So declare them now to clear confusion.
101*/
8f1f23e8 102#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
103# undef FALSE
104# undef TRUE
105 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
106# define ENUM_BOOL 1
107# ifndef HAS_BOOL
108# define HAS_BOOL 1
109# endif /* !HAS_BOOL */
8f1f23e8 110#endif /* NeXT || __NeXT__ */
641d3f0b 111
5d94fbed 112#ifndef HAS_BOOL
f789f6a4
FC
113# ifdef bool
114# undef bool
115# endif
70d5cb32 116# define bool char
c1d22f6b 117# define HAS_BOOL 1
a687059c 118#endif
0d3e774c 119
236415e2
KW
120/* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
121 * defined as char for example, then the cast from int is
122 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
123 * AIX */
18f5643b 124#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 125
46c6c7e2 126/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
127 * XXX Should really be a Configure probe, with HAS__FUNCTION__
128 * and FUNCTION__ as results.
129 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
130#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
131# define FUNCTION__ __func__
132#else
9ea40801 133# if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 134 (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
135# define FUNCTION__ ""
136# else
137# define FUNCTION__ __FUNCTION__ /* Common extension. */
138# endif
139#endif
140
27d4fb96
PP
141/* XXX A note on the perl source internal type system. The
142 original intent was that I32 be *exactly* 32 bits.
143
144 Currently, we only guarantee that I32 is *at least* 32 bits.
145 Specifically, if int is 64 bits, then so is I32. (This is the case
146 for the Cray.) This has the advantage of meshing nicely with
147 standard library calls (where we pass an I32 and the library is
148 expecting an int), but the disadvantage that an I32 is not 32 bits.
149 Andy Dougherty August 1996
24fef2a7 150
dc45a647
MB
151 There is no guarantee that there is *any* integral type with
152 exactly 32 bits. It is perfectly legal for a system to have
153 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 154
dc45a647
MB
155 Similarly, there is no guarantee that I16 and U16 have exactly 16
156 bits.
693762b4 157
8e84507e
NIS
158 For dealing with issues that may arise from various 32/64-bit
159 systems, we will ask Configure to check out
8175356b 160
3609ea0d
JH
161 SHORTSIZE == sizeof(short)
162 INTSIZE == sizeof(int)
163 LONGSIZE == sizeof(long)
dc45a647 164 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 165 PTRSIZE == sizeof(void *)
dc45a647
MB
166 DOUBLESIZE == sizeof(double)
167 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 168
27d4fb96
PP
169*/
170
69512466
JH
171#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
172# include <inttypes.h>
dd0eed91
JH
173# ifdef INT32_MIN_BROKEN
174# undef INT32_MIN
175# define INT32_MIN (-2147483647-1)
176# endif
177# ifdef INT64_MIN_BROKEN
178# undef INT64_MIN
179# define INT64_MIN (-9223372036854775807LL-1)
180# endif
69512466
JH
181#endif
182
8175356b
JH
183typedef I8TYPE I8;
184typedef U8TYPE U8;
185typedef I16TYPE I16;
186typedef U16TYPE U16;
187typedef I32TYPE I32;
188typedef U32TYPE U32;
6b8eaf93
JH
189#ifdef PERL_CORE
190# ifdef HAS_QUAD
8175356b
JH
191typedef I64TYPE I64;
192typedef U64TYPE U64;
6b8eaf93
JH
193# endif
194#endif /* PERL_CORE */
8175356b 195
69512466 196#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
80524f33
JL
197# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
198# define PeRl_INT64_C(c) CAT2(c,LL)
199# define PeRl_UINT64_C(c) CAT2(c,ULL)
200# else
52865553
SH
201# if QUADKIND == QUAD_IS___INT64
202# define PeRl_INT64_C(c) CAT2(c,I64)
203# define PeRl_UINT64_C(c) CAT2(c,UI64)
69512466 204# else
52865553
SH
205# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
206# define PeRl_INT64_C(c) CAT2(c,L)
207# define PeRl_UINT64_C(c) CAT2(c,UL)
69512466 208# else
80524f33
JL
209# define PeRl_INT64_C(c) ((I64TYPE)(c))
210# define PeRl_UINT64_C(c) ((U64TYPE)(c))
69512466
JH
211# endif
212# endif
e8c95190 213# endif
80524f33
JL
214# ifndef UINT64_C
215# define UINT64_C(c) PeRl_UINT64_C(c)
216# endif
217# ifndef INT64_C
218# define INT64_C(c) PeRl_INT64_C(c)
219# endif
e8c95190 220#endif
dc750f2e 221
d8668976 222#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 223
5ff3f7a4
GS
224/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
225 Please search CHAR_MAX in perl.h for further details. */
226#define U8_MAX UINT8_MAX
227#define U8_MIN UINT8_MIN
228
5ff3f7a4
GS
229#define I16_MAX INT16_MAX
230#define I16_MIN INT16_MIN
231#define U16_MAX UINT16_MAX
232#define U16_MIN UINT16_MIN
233
5ff3f7a4
GS
234#define I32_MAX INT32_MAX
235#define I32_MIN INT32_MIN
0e983133
GS
236#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
237# define U32_MAX UINT32_MAX
238#else
239# define U32_MAX 4294967295U
240#endif
5ff3f7a4
GS
241#define U32_MIN UINT32_MIN
242
243#else
244
5c9fa16e
KA
245/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
246 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
247#define U8_MAX PERL_UCHAR_MAX
248#define U8_MIN PERL_UCHAR_MIN
79072805 249
27d4fb96
PP
250#define I16_MAX PERL_SHORT_MAX
251#define I16_MIN PERL_SHORT_MIN
252#define U16_MAX PERL_USHORT_MAX
253#define U16_MIN PERL_USHORT_MIN
79072805 254
c4f23d77 255#if LONGSIZE > 4
27d4fb96
PP
256# define I32_MAX PERL_INT_MAX
257# define I32_MIN PERL_INT_MIN
258# define U32_MAX PERL_UINT_MAX
259# define U32_MIN PERL_UINT_MIN
79072805 260#else
27d4fb96
PP
261# define I32_MAX PERL_LONG_MAX
262# define I32_MIN PERL_LONG_MIN
263# define U32_MAX PERL_ULONG_MAX
264# define U32_MIN PERL_ULONG_MIN
79072805
LW
265#endif
266
5ff3f7a4
GS
267#endif
268
58a9a5d5 269/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
270#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
271#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
272#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
273
ff68c719 274#define Ctl(ch) ((ch) & 037)
8d063cd8 275
4650c663
KW
276/* This is a helper macro to avoid preprocessor issues, replaced by nothing
277 * unless under DEBUGGING, where it expands to an assert of its argument,
278 * followed by a comma (hence the comma operator). If we just used a straight
279 * assert(), we would get a comma with nothing before it when not DEBUGGING */
0f092d08
KW
280#ifdef DEBUGGING
281# define __ASSERT_(statement) assert(statement),
282#else
283# define __ASSERT_(statement)
284#endif
285
3fe05580
MHM
286/*
287=head1 SV-Body Allocation
288
289=for apidoc Ama|SV*|newSVpvs|const char* s
290Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
291
84bafc02
NC
292=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
293Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
294pair.
295
3fe05580
MHM
296=for apidoc Ama|SV*|newSVpvs_share|const char* s
297Like C<newSVpvn_share>, but takes a literal string instead of a string/length
298pair and omits the hash parameter.
299
9dcc53ea
Z
300=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
301Like C<sv_catpvn_flags>, but takes a literal string instead of a
302string/length pair.
303
304=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
305Like C<sv_catpvn_nomg>, but takes a literal string instead of a
306string/length pair.
307
7b4e7469 308=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
309Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
310
9dcc53ea
Z
311=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
312Like C<sv_catpvn_mg>, but takes a literal string instead of a
313string/length pair.
314
7b4e7469 315=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
316Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
317
9dcc53ea
Z
318=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
319Like C<sv_setpvn_mg>, but takes a literal string instead of a
320string/length pair.
321
322=for apidoc Am|SV *|sv_setref_pvs|const char* s
323Like C<sv_setref_pvn>, but takes a literal string instead of a
324string/length pair.
325
3fe05580
MHM
326=head1 Memory Management
327
328=for apidoc Ama|char*|savepvs|const char* s
329Like C<savepvn>, but takes a literal string instead of a string/length pair.
330
9dcc53ea
Z
331=for apidoc Ama|char*|savesharedpvs|const char* s
332A version of C<savepvs()> which allocates the duplicate string in memory
333which is shared between threads.
334
3fe05580
MHM
335=head1 GV Functions
336
337=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
338Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
339
340=head1 Hash Manipulation Functions
341
342=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
343Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
344
345=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
346Like C<hv_store>, but takes a literal string instead of a string/length pair
347and omits the hash parameter.
348
510966aa
Z
349=head1 Lexer interface
350
351=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
352
353Like L</lex_stuff_pvn>, but takes a literal string instead of a
354string/length pair.
355
3fe05580
MHM
356=cut
357*/
358
926bb54c
KW
359/* concatenating with "" ensures that only literal strings are accepted as
360 * argument */
26bfd924 361#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 362
926bb54c
KW
363/* note that STR_WITH_LEN() can't be used as argument to macros or functions
364 * that under some configurations might be macros, which means that it requires
365 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
366 */
367
368/* STR_WITH_LEN() shortcuts */
369#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
370#define newSVpvs_flags(str,flags) \
371 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 372#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
373#define sv_catpvs_flags(sv, str, flags) \
374 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
375#define sv_catpvs_nomg(sv, str) \
376 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
377#define sv_catpvs(sv, str) \
378 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
379#define sv_catpvs_mg(sv, str) \
380 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 381#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
382#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
383#define sv_setref_pvs(rv, classname, str) \
384 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 385#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
386#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
387#define gv_stashpvs(str, create) \
388 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
389#define gv_fetchpvs(namebeg, add, sv_type) \
390 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
391#define gv_fetchpvn(namebeg, len, add, sv_type) \
392 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
393#define sv_catxmlpvs(dsv, str, utf8) \
394 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
395#define hv_fetchs(hv,key,lval) \
396 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
397 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
398 : HV_FETCH_JUST_SV, NULL, 0))
399
400#define hv_stores(hv,key,val) \
401 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
402 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 403
510966aa
Z
404#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
405
b96d8cd9
NC
406#define get_cvs(str, flags) \
407 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 408
954c1994 409/*
ccfc67b7
JH
410=head1 Miscellaneous Functions
411
954c1994
GS
412=for apidoc Am|bool|strNE|char* s1|char* s2
413Test two strings to see if they are different. Returns true or
414false.
415
416=for apidoc Am|bool|strEQ|char* s1|char* s2
417Test two strings to see if they are equal. Returns true or false.
418
419=for apidoc Am|bool|strLT|char* s1|char* s2
420Test two strings to see if the first, C<s1>, is less than the second,
421C<s2>. Returns true or false.
422
423=for apidoc Am|bool|strLE|char* s1|char* s2
424Test two strings to see if the first, C<s1>, is less than or equal to the
425second, C<s2>. Returns true or false.
426
427=for apidoc Am|bool|strGT|char* s1|char* s2
428Test two strings to see if the first, C<s1>, is greater than the second,
429C<s2>. Returns true or false.
430
431=for apidoc Am|bool|strGE|char* s1|char* s2
432Test two strings to see if the first, C<s1>, is greater than or equal to
433the second, C<s2>. Returns true or false.
434
435=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
436Test two strings to see if they are different. The C<len> parameter
72d33970 437indicates the number of bytes to compare. Returns true or false. (A
954c1994
GS
438wrapper for C<strncmp>).
439
440=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
441Test two strings to see if they are equal. The C<len> parameter indicates
72d33970 442the number of bytes to compare. Returns true or false. (A wrapper for
954c1994
GS
443C<strncmp>).
444
445=cut
446*/
447
8d063cd8
LW
448#define strNE(s1,s2) (strcmp(s1,s2))
449#define strEQ(s1,s2) (!strcmp(s1,s2))
450#define strLT(s1,s2) (strcmp(s1,s2) < 0)
451#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
452#define strGT(s1,s2) (strcmp(s1,s2) > 0)
453#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
454#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
455#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 456
36477c24
PP
457#ifdef HAS_MEMCMP
458# define memNE(s1,s2,l) (memcmp(s1,s2,l))
459# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
460#else
461# define memNE(s1,s2,l) (bcmp(s1,s2,l))
462# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
463#endif
464
568a785a 465#define memEQs(s1, l, s2) \
8e6497b6 466 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
467#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
468
bbce6d69
PP
469/*
470 * Character classes.
471 *
472 * Unfortunately, the introduction of locales means that we
473 * can't trust isupper(), etc. to tell the truth. And when
474 * it comes to /\w+/ with tainting enabled, we *must* be able
475 * to trust our character classes.
476 *
477 * Therefore, the default tests in the text of Perl will be
478 * independent of locale. Any code that wants to depend on
479 * the current locale will use the tests that begin with "lc".
480 */
481
2304df62
AD
482#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
483# ifndef CTYPE256
484# define CTYPE256
485# endif
486#endif
487
954c1994 488/*
ccfc67b7
JH
489
490=head1 Character classes
243effed
KW
491This section is about functions (really macros) that classify characters
492into types, such as punctuation versus alphabetic, etc. Most of these are
493analogous to regular expression character classes. (See
494L<perlrecharclass/POSIX Character Classes>.) There are several variants for
495each class. (Not all macros have all variants; each item below lists the
496ones valid for it.) None are affected by C<use bytes>, and only the ones
497with C<LC> in the name are affected by the current locale.
498
499The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
500C<U8>) as input and returns a boolean as to whether or not the character
6aff1f14
KW
501represented by that octet is (or on non-ASCII platforms, corresponds to) an
502ASCII character in the named class based on platform, Unicode, and Perl rules.
503If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 504
6aff1f14
KW
505Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
506with no suffix C<"_A">.
4b9734bf
KW
507
508Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
509onto the platform. That is, the code points that are ASCII are unaffected,
510since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
511as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
512true when called with the code point 0xDF, which is a word character in both
4650c663 513ASCII and EBCDIC (though it represents different characters in each).
243effed
KW
514
515Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
516point as input. If the code point is larger than 255, Unicode rules are used
517to determine if it is in the character class. For example,
6aff1f14
KW
518C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
519WITH MACRON in Unicode, and is a word character.
243effed
KW
520
521Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
522(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
523classification of just the first (possibly multi-byte) character in the string
524is tested.
243effed 525
31f05a37
KW
526Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but the
527result is based on the current locale, which is what C<LC> in the name stands
528for. If Perl can determine that the current locale is a UTF-8 locale, it uses
529the published Unicode rules; otherwise, it uses the C library function that
530gives the named classification. For example, C<isDIGIT_LC()> when not in a
531UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always
532returned if the input won't fit into an octet.
243effed
KW
533
534Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
535returns the same as C<isFOO_LC> for input code points less than 256, and
536returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
537
538Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
539(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
540classification of just the first (possibly multi-byte) character in the string
541is tested.
ccfc67b7 542
954c1994 543=for apidoc Am|bool|isALPHA|char ch
2787a470 544Returns a boolean indicating whether the specified character is an
6aff1f14 545alphabetic character, analogous to C<m/[[:alpha:]]/>.
8a58bdcf 546See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 547C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 548C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 549
15861f94
KW
550=for apidoc Am|bool|isALPHANUMERIC|char ch
551Returns a boolean indicating whether the specified character is a either an
6aff1f14 552alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
15861f94
KW
553See the L<top of this section|/Character classes> for an explanation of variants
554C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 555C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
556C<isALPHANUMERIC_LC_utf8>.
557
8a58bdcf
KW
558=for apidoc Am|bool|isASCII|char ch
559Returns a boolean indicating whether the specified character is one of the 128
243effed 560characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 561On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
562character corresponds to an ASCII character. Variants C<isASCII_A()> and
563C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
564See the L<top of this section|/Character classes> for an explanation of variants
565C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
566C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
567library routine C<isascii()>. In these cases, the variants whose names contain
568C<LC> are the same as the corresponding ones without.
569
d98532ea
KW
570Also note, that because all ASCII characters are UTF-8 invariant (meaning they
571have the exact same representation (always a single byte) whether encoded in
572UTF-8 or not), C<isASCII> will give the correct results when called with any
573byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
574work properly on any string encoded or not in UTF-8.
575
243effed
KW
576=for apidoc Am|bool|isBLANK|char ch
577Returns a boolean indicating whether the specified character is a
6aff1f14 578character considered to be a blank, analogous to C<m/[[:blank:]]/>.
243effed 579See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 580C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
581C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
582platforms do not have the C library routine C<isblank()>. In these cases, the
583variants whose names contain C<LC> are the same as the corresponding ones
584without.
585
586=for apidoc Am|bool|isCNTRL|char ch
587Returns a boolean indicating whether the specified character is a
6aff1f14 588control character, analogous to C<m/[[:cntrl:]]/>.
243effed 589See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 590C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
591C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
592On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 593
954c1994 594=for apidoc Am|bool|isDIGIT|char ch
2787a470 595Returns a boolean indicating whether the specified character is a
6aff1f14 596digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 597Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed 598See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 599C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
600C<isDIGIT_LC_utf8>.
601
602=for apidoc Am|bool|isGRAPH|char ch
603Returns a boolean indicating whether the specified character is a
6aff1f14 604graphic character, analogous to C<m/[[:graph:]]/>.
243effed 605See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 606C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 607C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 608
0c82b6df 609=for apidoc Am|bool|isLOWER|char ch
2787a470 610Returns a boolean indicating whether the specified character is a
6aff1f14 611lowercase character, analogous to C<m/[[:lower:]]/>.
8a58bdcf 612See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 613C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 614C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 615
c99e91e9 616=for apidoc Am|bool|isOCTAL|char ch
2787a470 617Returns a boolean indicating whether the specified character is an
6aff1f14 618octal digit, [0-7].
243effed
KW
619The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
620C<isOCTAL>.
621
622=for apidoc Am|bool|isPUNCT|char ch
623Returns a boolean indicating whether the specified character is a
6aff1f14
KW
624punctuation character, analogous to C<m/[[:punct:]]/>.
625Note that the definition of what is punctuation isn't as
243effed
KW
626straightforward as one might desire. See L<perlrecharclass/POSIX Character
627Classes> for details.
628See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 629C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 630C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 631
0c82b6df 632=for apidoc Am|bool|isSPACE|char ch
2787a470 633Returns a boolean indicating whether the specified character is a
6aff1f14 634whitespace character. This is analogous
398d098a
KW
635to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
636(experimentally), this also matches what C<m/[[:space:]]/> does.
637("Experimentally" means that this change may be backed out in 5.20 or 5.22 if
638field experience indicates that it was unwise.) Prior to 5.18, only the
639locale forms of this macro (the ones with C<LC> in their names) matched
640precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
641in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
642(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
8a58bdcf 643See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 644C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 645C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 646
398d098a
KW
647=for apidoc Am|bool|isPSXSPC|char ch
648(short for Posix Space)
649Starting in 5.18, this is identical (experimentally) in all its forms to the
650corresponding C<isSPACE()> macros. ("Experimentally" means that this change
651may be backed out in 5.20 or 5.22 if field experience indicates that it
652was unwise.)
653The locale forms of this macro are identical to their corresponding
654C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
655non-locale forms differ from their C<isSPACE()> forms only in that the
656C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
657Otherwise they are identical. Thus this macro is analogous to what
658C<m/[[:space:]]/> matches in a regular expression.
659See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 660C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
661C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
662
954c1994 663=for apidoc Am|bool|isUPPER|char ch
2787a470 664Returns a boolean indicating whether the specified character is an
6aff1f14 665uppercase character, analogous to C<m/[[:upper:]]/>.
8a58bdcf 666See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 667C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 668C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 669
243effed 670=for apidoc Am|bool|isPRINT|char ch
8eea39dd 671Returns a boolean indicating whether the specified character is a
6aff1f14 672printable character, analogous to C<m/[[:print:]]/>.
8a58bdcf 673See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 674C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
675C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
676
677=for apidoc Am|bool|isWORDCHAR|char ch
678Returns a boolean indicating whether the specified character is a character
679that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
680in a regular expression. A word character is an alphabetic character, a
681decimal digit, a connecting punctuation character (such as an underscore), or
682a "mark" character that attaches to one of those (like some sort of accent).
683C<isALNUM()> is a synonym provided for backward compatibility, even though a
684word character includes more than the standard C language meaning of
685alphanumeric.
686See the L<top of this section|/Character classes> for an explanation of variants
687C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
688C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
689
690=for apidoc Am|bool|isXDIGIT|char ch
691Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
692digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
693and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
694See the L<top of this section|/Character classes> for an explanation of variants
695C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
696C<isXDIGIT_LC_utf8>.
697
3c3ecf18
KW
698=for apidoc Am|bool|isIDFIRST|char ch
699Returns a boolean indicating whether the specified character can be the first
700character of an identifier. This is very close to, but not quite the same as
701the official Unicode property C<XID_Start>. The difference is that this
702returns true only if the input character also matches L</isWORDCHAR>.
703See the L<top of this section|/Character classes> for an explanation of variants
704C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
705C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
706
707=for apidoc Am|bool|isIDCONT|char ch
708Returns a boolean indicating whether the specified character can be the
709second or succeeding character of an identifier. This is very close to, but
710not quite the same as the official Unicode property C<XID_Continue>. The
711difference is that this returns true only if the input character also matches
712L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
713explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
714C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
715C<isIDCONT_LC_utf8>.
716
243effed 717=head1 Miscellaneous Functions
8eea39dd 718
95a59cab 719=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 720Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
721Behaviour is only well defined when isXDIGIT(*str) is true.
722
e7c1e6c1
KW
723=head1 Character case changing
724
1f607577
KW
725=for apidoc Am|U8|toUPPER|U8 ch
726Converts the specified character to uppercase. If the input is anything but an
727ASCII lowercase character, that input character itself is returned. Variant
c753c8d3 728C<toUPPER_A> is equivalent.
954c1994 729
1f607577
KW
730=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
731Converts the Unicode code point C<cp> to its uppercase version, and
732stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
733that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
734bytes since the uppercase version may be longer than the original character.
735
736The first code point of the uppercased version is returned
737(but note, as explained just above, that there may be more.)
738
739=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
740Converts the UTF-8 encoded character at C<p> to its uppercase version, and
741stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
742that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
743bytes since the uppercase version may be longer than the original character.
744
745The first code point of the uppercased version is returned
746(but note, as explained just above, that there may be more.)
747
748The input character at C<p> is assumed to be well-formed.
749
25200305
KW
750=for apidoc Am|U8|toFOLD|U8 ch
751Converts the specified character to foldcase. If the input is anything but an
752ASCII uppercase character, that input character itself is returned. Variant
753C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
754Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
755
1f607577
KW
756=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
757Converts the Unicode code point C<cp> to its foldcase version, and
758stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
759that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
760bytes since the foldcase version may be longer than the original character.
761
762The first code point of the foldcased version is returned
763(but note, as explained just above, that there may be more.)
764
765=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
766Converts the UTF-8 encoded character at C<p> to its foldcase version, and
767stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
768that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
769bytes since the foldcase version may be longer than the original character.
770
771The first code point of the foldcased version is returned
772(but note, as explained just above, that there may be more.)
773
774The input character at C<p> is assumed to be well-formed.
775
776=for apidoc Am|U8|toLOWER|U8 ch
777Converts the specified character to lowercase. If the input is anything but an
778ASCII uppercase character, that input character itself is returned. Variant
c753c8d3 779C<toLOWER_A> is equivalent.
954c1994 780
1f607577
KW
781=for apidoc Am|U8|toLOWER_L1|U8 ch
782Converts the specified Latin1 character to lowercase. The results are undefined if
783the input doesn't fit in a byte.
784
785=for apidoc Am|U8|toLOWER_LC|U8 ch
786Converts the specified character to lowercase using the current locale's rules,
787if possible; otherwise returns the input character itself.
788
789=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
790Converts the Unicode code point C<cp> to its lowercase version, and
791stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
792that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
793bytes since the lowercase version may be longer than the original character.
794
795The first code point of the lowercased version is returned
796(but note, as explained just above, that there may be more.)
797
798=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
799Converts the UTF-8 encoded character at C<p> to its lowercase version, and
800stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
801that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
802bytes since the lowercase version may be longer than the original character.
803
804The first code point of the lowercased version is returned
805(but note, as explained just above, that there may be more.)
806
807The input character at C<p> is assumed to be well-formed.
808
809=for apidoc Am|U8|toLOWER_LC|U8 ch
810Converts the specified character to lowercase using the current locale's rules,
811if possible; otherwise returns the input character itself.
812
25200305
KW
813=for apidoc Am|U8|toTITLE|U8 ch
814Converts the specified character to titlecase. If the input is anything but an
815ASCII lowercase character, that input character itself is returned. Variant
816C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
817as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
818concept used in locale handling, so there is no functionality for that.)
819
1f607577
KW
820=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
821Converts the Unicode code point C<cp> to its titlecase version, and
822stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
823that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
824bytes since the titlecase version may be longer than the original character.
825
826The first code point of the titlecased version is returned
827(but note, as explained just above, that there may be more.)
828
829=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
830Converts the UTF-8 encoded character at C<p> to its titlecase version, and
831stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
832that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
833bytes since the titlecase version may be longer than the original character.
834
835The first code point of the titlecased version is returned
836(but note, as explained just above, that there may be more.)
837
838The input character at C<p> is assumed to be well-formed.
839
954c1994 840=cut
353c9b6f 841
1f607577 842XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
25200305 843really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
243effed 844
8a58bdcf 845Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
846patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
847
954c1994
GS
848*/
849
7c062697
KW
850/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
851 * is known only in the perl core, and this macro can be called from outside
852 * that */
853#ifdef HAS_QUAD
854# define WIDEST_UTYPE U64TYPE
855#else
856# define WIDEST_UTYPE U32
857#endif
858
3912bc88
KW
859/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
860 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
861 * bits of information are lost even on a 64-bit machine, but to get the
862 * compiler to optimize it out if possible. This is because Configure makes
863 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
864 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
865 */
866#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 867
41f43cc2 868#ifdef EBCDIC
b6340bd0
KW
869# ifndef _ALL_SOURCE
870 /* This returns the wrong results on at least z/OS unless this is
871 * defined. */
872# error _ALL_SOURCE should probably be defined
873# endif
874
875 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
876 * likely not defined, and so we use the native function */
3cedda69 877# define isASCII(c) cBOOL(isascii(c))
41f43cc2
KW
878#else
879# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
880#endif
881
ec79cfbb 882#define isASCII_A(c) isASCII(c)
3509bb53 883#define isASCII_L1(c) isASCII(c)
6515cb76 884
c2da0b36
KW
885/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
886 * and the 8 possible permutations of those bits exactly comprise the 8 octal
887 * digits */
3d33b3e3 888#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 889
ec79cfbb 890/* ASCII range only */
9fb1bf9d 891#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 892
a500dc72
KW
893/* Character class numbers. For internal core Perl use only. The ones less
894 * than 32 are used in PL_charclass[] and the ones up through the one that
895 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
896 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
897 * their actual definitions are here. If that file has a name not used here,
898 * it won't compile.
1709d539
KW
899 *
900 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
901 * frequency of their use. This gives a slight edge to exiting a loop earlier
902 * (in reginclass() in regexec.c) */
1709d539
KW
903# define _CC_WORDCHAR 0 /* \w and [:word:] */
904# define _CC_DIGIT 1 /* \d and [:digit:] */
905# define _CC_ALPHA 2 /* [:alpha:] */
906# define _CC_LOWER 3 /* [:lower:] */
907# define _CC_UPPER 4 /* [:upper:] */
908# define _CC_PUNCT 5 /* [:punct:] */
909# define _CC_PRINT 6 /* [:print:] */
15861f94 910# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 911# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 912# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 913
b0d691b2 914#define _FIRST_NON_SWASH_CC 10
1709d539
KW
915/* The character classes above are implemented with swashes. The second group
916 * (just below) contains the ones implemented without. These are also sorted
917 * in rough order of the frequency of their use, except that \v should be last,
918 * as it isn't a real Posix character class, and some (small) inefficiencies in
919 * regular expression handling would be introduced by putting it in the middle
920 * of those that are. Also, cntrl and ascii come after the others as it may be
921 * useful to group these which have no members that match above Latin1, (or
922 * above ASCII in the latter case) */
923
b0d691b2
KW
924# define _CC_SPACE 10 /* \s */
925# define _CC_BLANK 11 /* [:blank:] */
926# define _CC_XDIGIT 12 /* [:xdigit:] */
927# define _CC_PSXSPC 13 /* [:space:] */
928# define _CC_CNTRL 14 /* [:cntrl:] */
929# define _CC_ASCII 15 /* [:ascii:] */
930# define _CC_VERTSPACE 16 /* \v */
1709d539 931
a0947d7b
KW
932# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
933
1709d539 934/* The members of the third group below do not need to be coordinated with data
3ded5eb0
KW
935 * structures in regcomp.[ch] and regexec.c. But they should be added to
936 * bootstrap_ctype() */
b0d691b2
KW
937# define _CC_IDFIRST 17
938# define _CC_CHARNAME_CONT 18
939# define _CC_NONLATIN1_FOLD 19
940# define _CC_QUOTEMETA 20
941# define _CC_NON_FINAL_FOLD 21
942# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
943# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
944/* Unused: 23-30
f4cdb42c
KW
945 * If more bits are needed, one could add a second word for non-64bit
946 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
947 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
948 * is used only for optimization (as of this writing), and differs in the
949 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
950 * characters: the masculine and feminine ordinal indicators, so removing it
951 * would just cause /i regexes which match them to run less efficiently */
96ac0975 952
3a371f2f
KW
953#if defined(PERL_CORE) || defined(PERL_EXT)
954/* An enum version of the character class numbers, to help compilers
955 * optimize */
956typedef enum {
3a371f2f 957 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
958 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
959 _CC_ENUM_ASCII = _CC_ASCII,
960 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 961 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 962 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
963 _CC_ENUM_DIGIT = _CC_DIGIT,
964 _CC_ENUM_GRAPH = _CC_GRAPH,
965 _CC_ENUM_LOWER = _CC_LOWER,
966 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 967 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 968 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 969 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 970 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 971 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 972 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 973 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
974} _char_class_number;
975#endif
976
63c61c3f 977#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 978#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 979
3018b823 980#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 981# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
982 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
983 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
984 #error Need to adjust order of swash_property_names[]
985# endif
986
b44daeeb
KW
987/* This is declared static in each of the few files that this is #defined for
988 * to keep them from being publicly accessible. Hence there is a small amount
989 * of wasted space */
990
922e8cb4
KW
991static const char* const swash_property_names[] = {
992 "XPosixWord",
993 "XPosixDigit",
994 "XPosixAlpha",
995 "XPosixLower",
996 "XPosixUpper",
997 "XPosixPunct",
998 "XPosixPrint",
999 "XPosixAlnum",
b0d691b2
KW
1000 "XPosixGraph",
1001 "Cased"
922e8cb4
KW
1002};
1003#endif
1004
96ac0975
NC
1005# ifdef DOINIT
1006EXTCONST U32 PL_charclass[] = {
1007# include "l1_char_class_tab.h"
1008};
1009
1010# else /* ! DOINIT */
1011EXTCONST U32 PL_charclass[];
1012# endif
3ded5eb0
KW
1013#endif /* Has perl.h */
1014
1015#if defined(H_PERL) && ! defined(BOOTSTRAP_CHARSET)
96ac0975 1016
265c1f46 1017 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70 1018# define _CC_mask(classnum) (1U << (classnum))
4650c663
KW
1019
1020 /* For internal core Perl use only: the base macro for defining macros like
1021 * isALPHA */
430b7c70 1022# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
f4cd282c 1023 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
4eeeb416 1024
f4cdb42c
KW
1025 /* The mask for the _A versions of the macros; it just adds in the bit for
1026 * ASCII. */
1027# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1028
4650c663
KW
1029 /* For internal core Perl use only: the base macro for defining macros like
1030 * isALPHA_A. The foo_A version makes sure that both the desired bit and
1031 * the ASCII bit are present */
3ded5eb0 1032# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
f4cd282c 1033 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
f4cdb42c
KW
1034 == _CC_mask_A(classnum)))
1035
f4cdb42c 1036# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 1037# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
1038# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1039# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1040# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
1041# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1042# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1043# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1044# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1045# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1046# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1047# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1048# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1049# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
d95f8b6a 1050# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
3ded5eb0
KW
1051# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1052# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1053# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1054
1055 /* continuation character for legal NAME in \N{NAME} */
1056# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1057
1058# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1059# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1060# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1061# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1062# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1063# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1064# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1065# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1066# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1067# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
f4cdb42c 1068
fb9bfbf7
KW
1069 /* Either participates in a fold with a character above 255, or is a
1070 * multi-char fold */
f4cd282c 1071# 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 1072
4eeeb416 1073# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
1074# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1075 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
1076# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1077 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
3ded5eb0
KW
1078#else /* Either don't have perl.h or don't want to use char_class_tab.h */
1079
1080 /* If we don't have perl.h, we are compiling a utility program. Below we
1081 * hard-code various macro definitions that wouldn't otherwise be available
1082 * to it. We can also get here if we are configured to bootstrap up Perl
1083 * on a non-ASCII platform that doesn't have a working Perl (currently only
1084 * EBCDIC). For these we currently use the native definitions to get
1085 * things going. (It should also be possible to use the translation
1086 * function NATIVE_TO_LATIN1(), but that is an extra layer of dependence on
1087 * Perl, so it is currently avoided for the macros where it's possible to
1088 * do so.) */
9fb1bf9d 1089# ifdef EBCDIC
3ded5eb0 1090 /* Use the native functions. They likely will return false for all
4f48cecf 1091 * non-ASCII values, but this makes sure */
b6340bd0
KW
1092# define isALPHA_A(c) (isASCII(c) && isalpha(c))
1093# define isALPHANUMERIC_A(c) (isASCII(c) && isalnum(c))
1094# define isCNTRL_A(c) (isASCII(c) && iscntrl(c))
1095# define isDIGIT_A(c) (isASCII(c) && isdigit(c))
1096# define isGRAPH_A(c) (isASCII(c) && isgraph(c))
1097# define isLOWER_A(c) (isASCII(c) && islower(c))
1098# define isPRINT_A(c) (isASCII(c) && isprint(c))
1099# define isPUNCT_A(c) (isASCII(c) && ispunct(c))
1100# define isSPACE_A(c) (isASCII(c) && isspace(c))
1101# define isUPPER_A(c) (isASCII(c) && isupper(c))
1102# define isXDIGIT_A(c) (isASCII(c) && isxdigit(c))
3ded5eb0 1103# else /* ASCII platform. These are coded based on first principals */
9fb1bf9d 1104# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
15861f94 1105# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
3ded5eb0 1106# define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
2181da29 1107# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
3ded5eb0 1108# define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
9fb1bf9d
KW
1109# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1110# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
3ded5eb0
KW
1111# define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1112# define isSPACE_A(c) ((c) == ' ' \
1113 || (c) == '\t' \
1114 || (c) == '\n' \
1115 || (c) == '\r' \
1116 || (c) =='\v' \
926bb54c 1117 || (c) == '\f')
2181da29 1118# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
3ded5eb0
KW
1119# define isXDIGIT_A(c) (isDIGIT_A(c) \
1120 || ((c) >= 'a' && (c) <= 'f') \
1121 || ((c) <= 'F' && (c) >= 'A'))
1122# endif /* Below are common definitions for ASCII and non-ASCII */
1123# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1124# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1125# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1126
1127 /* The _L1 macros may be unnecessary for both the utilities and for
1128 * bootstrapping; I (khw) added them during debugging of bootstrapping, and
1129 * it seems best to keep them. */
1130# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1131# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1132# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1133# define isBLANK_L1(c) (isBLANK_A(c) \
1134 || (FITS_IN_8_BITS(c) \
1135 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1136# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1137# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1138# define isLOWER_L1(c) (isLOWER_A(c) \
1139 || (FITS_IN_8_BITS(c) \
1140 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1141 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1142 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1143 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1144 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1145# define isPRINT_L1(c) (isPRINT_A(c) \
1146 || (FITS_IN_8_BITS(c) \
1147 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1148# define isPSXSPC_L1(c) isSPACE_L1(c)
1149# define isPUNCT_L1(c) (isPUNCT_A(c) \
1150 || (FITS_IN_8_BITS(c) \
1151 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1152 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1153 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1154 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1155 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1156 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1157 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1158# define isSPACE_L1(c) (isSPACE_A(c) \
1159 || (FITS_IN_8_BITS(c) \
1160 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1161 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1162# define isUPPER_L1(c) (isUPPER_A(c) \
1163 || (FITS_IN_8_BITS(c) \
1164 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1165 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1166 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1167# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1168# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1169# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1170 || isBLANK_L1(c) \
1171 || (c) == '-' \
1172 || (c) == '(' \
1173 || (c) == ')')
1174 /* The following are not fully accurate in the above-ASCII range. I (khw)
1175 * don't think it's necessary to be so for the purposes where this gets
1176 * compiled */
1177# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1178# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1179
1180 /* And these aren't accurate at all. They are useful only for above
1181 * Latin1, which utilities and bootstrapping don't deal with */
1182# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1183# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1184
1185 /* Many of the macros later in this file are defined in terms of these. By
1186 * implementing them with a function, which converts the class number into
1187 * a call to the desired macro, all of the later ones work. However, that
1188 * function won't be actually defined when building a utility program (no
1189 * perl.h), and so a compiler error will be generated if one is attempted
1190 * to be used. And the above-Latin1 code points require Unicode tables to
1191 * be present, something unlikely to be the case when bootstrapping */
1192# define _generic_isCC(c, classnum) \
1193 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1194# define _generic_isCC_A(c, classnum) \
1195 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
1196#endif /* End of no perl.h or have BOOTSTRAP_CHARSET */
8a58bdcf 1197
e66b99e9
KW
1198#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1199#define isALPHA(c) isALPHA_A(c)
1200#define isBLANK(c) isBLANK_A(c)
1201#define isCNTRL(c) isCNTRL_A(c)
1202#define isDIGIT(c) isDIGIT_A(c)
1203#define isGRAPH(c) isGRAPH_A(c)
1204#define isIDFIRST(c) isIDFIRST_A(c)
1205#define isLOWER(c) isLOWER_A(c)
1206#define isPRINT(c) isPRINT_A(c)
1207#define isPSXSPC(c) isPSXSPC_A(c)
1208#define isPUNCT(c) isPUNCT_A(c)
1209#define isSPACE(c) isSPACE_A(c)
1210#define isUPPER(c) isUPPER_A(c)
1211#define isWORDCHAR(c) isWORDCHAR_A(c)
1212#define isXDIGIT(c) isXDIGIT_A(c)
1213
1214/* ASCII casing. These could also be written as
1215 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1216 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1217 which uses table lookup and mask instead of subtraction. (This would
1218 work because the _MOD does not apply in the ASCII range) */
68067e4e
DM
1219#define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1220#define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
bbce6d69 1221
25200305
KW
1222/* In the ASCII range, these are equivalent to what they're here defined to be.
1223 * But by creating these definitions, other code doesn't have to be aware of
1224 * this detail */
1225#define toFOLD(c) toLOWER(c)
25200305
KW
1226#define toTITLE(c) toUPPER(c)
1227
c753c8d3
KW
1228#define toLOWER_A(c) toLOWER(c)
1229#define toUPPER_A(c) toUPPER(c)
25200305
KW
1230#define toFOLD_A(c) toFOLD(c)
1231#define toTITLE_A(c) toTITLE(c)
1a0901db 1232
4650c663 1233/* Use table lookup for speed; returns the input itself if is out-of-range */
b2bf251f 1234#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1235 ? (c) \
f4cd282c 1236 : PL_latin1_lc[ (U8) (c) ])
c753c8d3
KW
1237#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1238
1a0901db 1239/* Modified uc. Is correct uc except for three non-ascii chars which are
4650c663
KW
1240 * all mapped to one of them, and these need special handling; returns the
1241 * input itself if is out-of-range */
b2bf251f 1242#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1243 ? (c) \
f4cd282c 1244 : PL_mod_latin1_uc[ (U8) (c) ])
31f05a37 1245#define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale
84061b6a 1246
beab9ebe
KW
1247/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
1248
1249/* For internal core Perl use only: the base macro for defining macros like
1250 * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point
31f05a37
KW
1251 * (0-255) to check. In a UTF-8 locale, the result is the same as calling
1252 * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like
1253 * _CC_UPPER, which gives the class number for doing this. For non-UTF-8
1254 * locales, the code to actually do the test this is passed in 'non_utf8'. If
1255 * 'c' is above 255, 0 is returned. For accessing the full range of possible
1256 * code points under locale rules, use the macros based on _generic_LC_uvchr
1257 * instead of this. */
beab9ebe
KW
1258#define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \
1259 (! FITS_IN_8_BITS(c) \
1260 ? 0 \
31f05a37
KW
1261 : IN_UTF8_CTYPE_LOCALE \
1262 ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \
beab9ebe
KW
1263 : cBOOL(non_utf8))
1264
1265/* For internal core Perl use only: a helper macro for defining macros like
1266 * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to
1267 * actually do this test is passed in 'non_utf8_func', which is called on 'c',
1268 * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See
1269 * _generic_LC_base for more info */
1270#define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \
1271 _generic_LC_base(c,utf8_locale_classnum, \
1272 non_utf8_func( (_LC_CAST) (c)))
1273
1274/* For internal core Perl use only: like _generic_LC, but also returns TRUE if
1275 * 'c' is the platform's native underscore character */
1276#define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \
1277 _generic_LC_base(c, utf8_locale_classnum, \
1278 (non_utf8_func( (_LC_CAST) (c)) \
1279 || (char)(c) == '_'))
1280
1281/* These next three are also for internal core Perl use only: case-change
1282 * helper macros */
1283#define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \
1284 ? (c) \
31f05a37
KW
1285 : (IN_UTF8_CTYPE_LOCALE) \
1286 ? PL_latin1_lc[ (U8) (c) ] \
beab9ebe
KW
1287 : function((cast)(c)))
1288
31f05a37
KW
1289/* Note that the result can be larger than a byte in a UTF-8 locale. It
1290 * returns a single value, so can't adequately return the upper case of LATIN
1291 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1292 * values "SS"); instead it asserts against that under DEBUGGING, and
1293 * otherwise returns its input */
beab9ebe
KW
1294#define _generic_toUPPER_LC(c, function, cast) \
1295 (! FITS_IN_8_BITS(c) \
1296 ? (c) \
31f05a37
KW
1297 : ((! IN_UTF8_CTYPE_LOCALE) \
1298 ? function((cast)(c)) \
1299 : ((((U8)(c)) == MICRO_SIGN) \
1300 ? GREEK_CAPITAL_LETTER_MU \
1301 : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \
1302 ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \
1303 : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \
1304 ? (__ASSERT_(0) (c)) \
1305 : PL_mod_latin1_uc[ (U8) (c) ])))))
1306
1307/* Note that the result can be larger than a byte in a UTF-8 locale. It
1308 * returns a single value, so can't adequately return the fold case of LATIN
1309 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1310 * values "ss"); instead it asserts against that under DEBUGGING, and
1311 * otherwise returns its input */
beab9ebe 1312#define _generic_toFOLD_LC(c, function, cast) \
4d7db1b9 1313 ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \
31f05a37 1314 ? GREEK_SMALL_LETTER_MU \
4d7db1b9
KW
1315 : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \
1316 || (c) != LATIN_SMALL_LETTER_SHARP_S) \
1317 _generic_toLOWER_LC(c, function, cast)))
beab9ebe 1318
84061b6a
KW
1319/* Use the libc versions for these if available. */
1320#if defined(HAS_ISASCII) && ! defined(USE_NEXT_CTYPE)
1321# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c)))
1322#else
1323# define isASCII_LC(c) isASCII(c)
1324#endif
1325
1326#if defined(HAS_ISBLANK) && ! defined(USE_NEXT_CTYPE)
1327# define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank)
1328#else
1329# define isBLANK_LC(c) isBLANK(c)
1330#endif
1331
4650c663 1332#ifdef USE_NEXT_CTYPE /* NeXT computers */
bbce6d69 1333
beab9ebe
KW
1334# define _LC_CAST unsigned int /* Needed by _generic_LC. NeXT functions
1335 use this as their input type */
1336
1337# define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, NXIsAlpha)
1338# define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, NXIsAlNum)
1339# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, NXIsCntrl)
1340# define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, NXIsDigit)
1341# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, NXIsGraph)
1342# define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, NXIsAlpha)
1343# define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, NXIsLower)
1344# define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, NXIsPrint)
1345# define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, NXIsPunct)
1346# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, NXIsSpace)
1347# define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, NXIsUpper)
1348# define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, NXIsAlNum)
1349# define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, NXIsXdigit)
1350
1351# define toLOWER_LC(c) _generic_toLOWER_LC((c), NXToLower, unsigned int)
1352# define toUPPER_LC(c) _generic_toUPPER_LC((c), NXToUpper, unsigned int)
1353# define toFOLD_LC(c) _generic_toFOLD_LC((c), NXToLower, unsigned int)
bbce6d69
PP
1354
1355#else /* !USE_NEXT_CTYPE */
b8c5462f 1356
841f64f7
BF
1357# define _LC_CAST U8
1358
bbce6d69 1359# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
4650c663 1360 /* For most other platforms */
beab9ebe 1361
beab9ebe
KW
1362# define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha)
1363# define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum)
1364# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1365# define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit)
1366# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph)
1367# define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha)
1368# define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower)
1369# define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint)
1370# define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct)
1371# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1372# define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper)
1373# define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum)
1374# define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit)
1375
1376
1377# define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1378# define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1379# define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
bbce6d69 1380
4650c663 1381# else /* The final fallback position */
bbce6d69 1382
bbce6d69 1383# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1384# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
b8c5462f 1385# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1386# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1387# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1388# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1389# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1390# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1391# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1392# define isSPACE_LC(c) (isascii(c) && isspace(c))
1393# define isUPPER_LC(c) (isascii(c) && isupper(c))
1394# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1395# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
4650c663 1396
76fde96a 1397# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1398# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
beab9ebe 1399# define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c))
bbce6d69
PP
1400
1401# endif
a0d0e21e 1402#endif /* USE_NEXT_CTYPE */
55204971 1403
eba68aa0
KW
1404#define isIDCONT(c) isWORDCHAR(c)
1405#define isIDCONT_A(c) isWORDCHAR_A(c)
1406#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1407#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1408#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1409
4650c663
KW
1410/* For internal core Perl use only: the base macros for defining macros like
1411 * isALPHA_uni. 'c' is the code point to check. 'classnum' is the POSIX class
1412 * number defined earlier in this file. _generic_uni() is used for POSIX
1413 * classes where there is a macro or function 'above_latin1' that takes the
1414 * single argument 'c' and returns the desired value. These exist for those
1415 * classes which have simple definitions, avoiding the overhead of a hash
1416 * lookup or inversion list binary search. _generic_swash_uni() can be used
1417 * for classes where that overhead is faster than a direct lookup.
1418 * _generic_uni() won't compile if 'c' isn't unsigned, as it won't match the
1419 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so
1420 * have duplicate checks here, so could create versions of the macros that
1421 * don't, but experiments show that gcc optimizes them out anyway. */
66c17564
KW
1422
1423/* Note that all ignore 'use bytes' */
cd500f2f
KW
1424#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1425 ? _generic_isCC(c, classnum) \
1426 : above_latin1(c))
1427#define _generic_swash_uni(classnum, c) ((c) < 256 \
1428 ? _generic_isCC(c, classnum) \
922e8cb4 1429 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1430#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1431#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1432#define isASCII_uni(c) isASCII(c)
1433#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1434#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1435#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1436#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1437#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1438#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1439#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1440#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1441
509fb054
KW
1442/* Posix and regular space are identical above Latin1 */
1443#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1444
cd500f2f 1445#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1446#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1447#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1448#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1449#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1450#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1451
b8d68ded 1452#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1453#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1454#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1455#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1456
4650c663
KW
1457/* For internal core Perl use only: the base macros for defining macros like
1458 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code
1459 * point, not just 0-255. Like _generic_uni, there are two versions, one for
1460 * simple class definitions; the other for more complex. These are like
1461 * _generic_uni, so see it for more info. */
cd500f2f
KW
1462#define _generic_LC_uvchr(latin1, above_latin1, c) \
1463 (c < 256 ? latin1(c) : above_latin1(c))
1464#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1465 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1466
1467#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1468#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1469 _CC_ALPHANUMERIC, c)
509fb054 1470#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1471#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1472#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1473#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1474#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1475#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1476 _is_uni_perl_idcont, c)
cd500f2f
KW
1477#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1478 _is_uni_perl_idstart, c)
1479#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1480#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1481#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1482 space under locale */
cd500f2f
KW
1483#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1484#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1485 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1486#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1487#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1488 _CC_WORDCHAR, c)
1489#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1490
36b00f93 1491#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1492
4650c663
KW
1493/* For internal core Perl use only: the base macros for defining macros like
1494 * isALPHA_utf8. These are like the earlier defined macros, but take an input
1495 * UTF-8 encoded string 'p'. If the input is in the Latin1 range, use
1496 * the Latin1 macro 'classnum' on 'p'. Otherwise use the value given by the
1497 * 'utf8' parameter. This relies on the fact that ASCII characters have the
1498 * same representation whether utf8 or not. Note that it assumes that the utf8
1499 * has been validated, and ignores 'use bytes' */
cd500f2f 1500#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1501 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1502 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c 1503 ? _generic_isCC( \
94bb8c36 1504 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
aea0c08c 1505 *((p)+1 )), \
94bb8c36 1506 classnum) \
252810d0 1507 : utf8)
cd500f2f
KW
1508/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1509 * can be a macro */
1510#define _generic_func_utf8(classnum, above_latin1, p) \
1511 _generic_utf8(classnum, p, above_latin1(p))
4650c663 1512/* Like the above, but passes classnum to _isFOO_utf8(), instead of having an
cd500f2f 1513 * 'above_latin1' parameter */
922e8cb4 1514#define _generic_swash_utf8(classnum, p) \
cd500f2f 1515 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1516
cc8ab7c0
KW
1517/* Like the above, but should be used only when it is known that there are no
1518 * characters in the range 128-255 which the class is TRUE for. Hence it can
cd500f2f 1519 * skip the tests for this range. 'above_latin1' should include its arguments */
3ded5eb0 1520#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1521 (UTF8_IS_INVARIANT(*(p)) \
1522 ? _generic_isCC(*(p), classnum) \
1523 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
3ded5eb0 1524 ? above_latin1 \
cc8ab7c0
KW
1525 : 0)
1526
84238efa
KW
1527/* NOTE that some of these macros have very similar ones in regcharclass.h.
1528 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1529 * there, differing in name only by an underscore from the one here
1530 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1531 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1532 * points; the regcharclass.h ones are implemented as a series of
1533 * "if-else-if-else ..." */
1534
922e8cb4
KW
1535#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1536#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1537#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1538 utf8, the non-utf8 macro works
1539 */
cd500f2f
KW
1540#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1541#define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1542#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1543 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1544#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1545#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1546 _is_utf8_perl_idcont, p)
e5dcd934 1547
c11ff943
KW
1548/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1549 * IDFIRST is an alnum. See
c8362b00 1550 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1551 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1552 * This used to be not the XID version, but we decided to go with the more
1553 * modern Unicode definition */
cd500f2f 1554#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1555 _is_utf8_perl_idstart, p)
e5dcd934 1556
922e8cb4
KW
1557#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1558#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1559
509fb054 1560/* Posix and regular space are identical above Latin1 */
cd500f2f 1561#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1562
922e8cb4 1563#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1564#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1565#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1566#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1567#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1568#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1569 is_XDIGIT_high(p))
a0ed51b3 1570
7d2b61dd 1571#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
509fb054
KW
1572#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1573#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1574#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1575
4650c663
KW
1576/* For internal core Perl use only: the base macros for defining macros like
1577 * isALPHA_LC_utf8. These are like _generic_utf8, but if the first code point
1578 * in 'p' is within the 0-255 range, it uses locale rules from the passed-in
1579 * 'macro' parameter */
94bb8c36
KW
1580#define _generic_LC_utf8(macro, p, utf8) \
1581 (UTF8_IS_INVARIANT(*(p)) \
1582 ? macro(*(p)) \
1583 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1584 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
9201a572
KW
1585 : utf8)
1586
cd500f2f
KW
1587#define _generic_LC_swash_utf8(macro, classnum, p) \
1588 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1589#define _generic_LC_func_utf8(macro, above_latin1, p) \
4650c663 1590 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1591
cd500f2f
KW
1592#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1593 _CC_ALPHANUMERIC, p)
1594#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1595#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1596#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1597#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1598#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1599#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1600#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1601#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1602#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1603#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1604#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1605 space under locale */
cd500f2f
KW
1606#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1607#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1608#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1609#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1610 _CC_WORDCHAR, p)
1611#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1612
fbc19f27
KW
1613/* Macros for backwards compatibility and for completeness when the ASCII and
1614 * Latin1 values are identical */
fbc19f27
KW
1615#define isALPHAU(c) isALPHA_L1(c)
1616#define isDIGIT_L1(c) isDIGIT_A(c)
1617#define isOCTAL(c) isOCTAL_A(c)
1618#define isOCTAL_L1(c) isOCTAL_A(c)
1619#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1620#define isALNUM(c) isWORDCHAR(c)
1621#define isALNUMU(c) isWORDCHAR_L1(c)
1622#define isALNUM_LC(c) isWORDCHAR_LC(c)
1623#define isALNUM_uni(c) isWORDCHAR_uni(c)
1624#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1625#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1626#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1627#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1628#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1629#define isALNUMC(c) isALPHANUMERIC(c)
1630#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1631#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1632#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1633#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1634#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1635
6ae9f32a 1636/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1637 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1638 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1639 * that is EOT on ASCII; ST on EBCDIC */
8e7c6e7d 1640# define toCTRL(c) (toUPPER(NATIVE_TO_LATIN1(c)) ^ 64)
bbce6d69 1641
dea28490
JJ
1642/* Line numbers are unsigned, 32 bits. */
1643typedef U32 line_t;
e5dcd934 1644#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1645
91152fc1
DG
1646/* Helpful alias for version prescan */
1647#define is_LAX_VERSION(a,b) \
1648 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1649
1650#define is_STRICT_VERSION(a,b) \
1651 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1652
1653#define BADVERSION(a,b,c) \
1654 if (b) { \
1655 *b = c; \
1656 } \
1657 return a;
8c52afec 1658
cb27eebd
KW
1659/* Converts a character known to represent a hexadecimal digit (0-9, A-F, or
1660 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer,
1661 * which is advanced. The input is validated only by an assert() in DEBUGGING
1662 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and
1663 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */
1664#define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \
1665 ? (c) \
1666 : ((c) + 9))))
902008b8
KW
1667#define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \
1668 ? (*(s)++) \
1669 : (*(s)++ + 9))))
95a59cab 1670
cb27eebd
KW
1671/* Converts a character known to represent an octal digit (0-7) to its numeric
1672 * value. The input is validated only by an assert() in DEBUGGING builds. In
1673 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */
1674#define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c)))
1675
8e84507e 1676/*
ccfc67b7
JH
1677=head1 Memory Management
1678
a02a5408 1679=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1680The XSUB-writer's interface to the C C<malloc> function.
1681
0d7b2759
MH
1682Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1683
c5008215
JC
1684In 5.9.3, Newx() and friends replace the older New() API, and drops
1685the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1686themselves. This aid has been superseded by a new build option,
d10b4965 1687PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1688there for use in XS modules supporting older perls.
1689
a02a5408 1690=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1691The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1692cast. See also C<Newx>.
954c1994 1693
0d7b2759
MH
1694Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1695
a02a5408 1696=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1697The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1698memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1699
0d7b2759
MH
1700Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1701
954c1994
GS
1702=for apidoc Am|void|Renew|void* ptr|int nitems|type
1703The XSUB-writer's interface to the C C<realloc> function.
1704
0d7b2759
MH
1705Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1706
954c1994
GS
1707=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1708The XSUB-writer's interface to the C C<realloc> function, with
1709cast.
1710
0d7b2759
MH
1711Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1712
49b8b560 1713=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1714The XSUB-writer's interface to the C C<free> function.
1715
0d7b2759
MH
1716This should B<ONLY> be used on memory obtained using L<"Newx"> and friends.
1717
954c1994
GS
1718=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1719The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1720source, C<dest> is the destination, C<nitems> is the number of items, and
1721C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1722
e90e2364 1723=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
72d33970
FC
1724Like C<Move> but returns dest. Useful
1725for encouraging compilers to tail-call
e90e2364
NC
1726optimise.
1727
954c1994
GS
1728=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1729The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1730source, C<dest> is the destination, C<nitems> is the number of items, and
1731C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1732
e90e2364
NC
1733=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1734
72d33970
FC
1735Like C<Copy> but returns dest. Useful
1736for encouraging compilers to tail-call
e90e2364
NC
1737optimise.
1738
954c1994
GS
1739=for apidoc Am|void|Zero|void* dest|int nitems|type
1740
1741The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1742destination, C<nitems> is the number of items, and C<type> is the type.
1743
e90e2364
NC
1744=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1745
72d33970
FC
1746Like C<Zero> but returns dest. Useful
1747for encouraging compilers to tail-call
e90e2364
NC
1748optimise.
1749
da5d8dbb 1750=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1751This is an architecture-independent macro to copy one structure to another.
954c1994 1752
7e337ee0
JH
1753=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1754
1755Fill up memory with a byte pattern (a byte repeated over and over
1756again) that hopefully catches attempts to access uninitialized memory.
1757
1758=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1759
1760PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1761
1c12ffb4 1762=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1763
1764PoisonWith(0xEF) for catching access to freed memory.
1765
9965345d
JH
1766=for apidoc Am|void|Poison|void* dest|int nitems|type
1767
7e337ee0 1768PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1769
1770=cut */
954c1994 1771
561b68a9
SH
1772/* Maintained for backwards-compatibility only. Use newSV() instead. */
1773#ifndef PERL_CORE
ff06c60c 1774#define NEWSV(x,len) newSV(len)
561b68a9 1775#endif
ff06c60c 1776
19a94d75
JH
1777#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1778
2720f793
JH
1779/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1780 * overly eager compilers that will bleat about e.g.
1781 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1782#ifdef PERL_MALLOC_WRAP
f019c49e 1783#define MEM_WRAP_CHECK(n,t) \
d1decf2b 1784 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
27d5b266 1785#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1786 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1787#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1788
d1decf2b 1789#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
1790#else
1791
410319be
NC
1792#define MEM_WRAP_CHECK(n,t)
1793#define MEM_WRAP_CHECK_1(n,t,a)
1794#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1795#define MEM_WRAP_CHECK_(n,t)
1796
6fff79ce 1797#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1798
1936d2a7 1799#endif
8b44ba4c 1800
fe4f188c 1801#ifdef PERL_MEM_LOG
46c6c7e2 1802/*
9f653bb5 1803 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1804 * go through functions, which are handy for debugging breakpoints, but
1805 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1806 * line number, and C function name if available) passed in. This info can
1807 * then be used for logging the calls, for which one gets a sample
73d1d973 1808 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1809 *
46c6c7e2 1810 * Known problems:
94e892a6 1811 * - not all memory allocs get logged, only those
46c6c7e2 1812 * that go through Newx() and derivatives (while all
94e892a6 1813 * Safefrees do get logged)
46c6c7e2
JH
1814 * - __FILE__ and __LINE__ do not work everywhere
1815 * - __func__ or __FUNCTION__ even less so
1816 * - I think more goes on after the perlio frees but
1817 * the thing is that STDERR gets closed (as do all
1818 * the file descriptors)
1819 * - no deeper calling stack than the caller of the Newx()
1820 * or the kind, but do I look like a C reflection/introspection
1821 * utility to you?
1822 * - the function prototypes for the logging functions
1823 * probably should maybe be somewhere else than handy.h
1824 * - one could consider inlining (macrofying) the logging
1825 * for speed, but I am too lazy
1826 * - one could imagine recording the allocations in a hash,
1827 * (keyed by the allocation address?), and maintain that
1828 * through reallocs and frees, but how to do that without
1829 * any News() happening...?
73d1d973 1830 * - lots of -Ddefines to get useful/controllable output
b953482e 1831 * - lots of ENV reads
46c6c7e2
JH
1832 */
1833
12754f92 1834PERL_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 1835
12754f92 1836PERL_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 1837
12754f92 1838PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1839
0b0ab801 1840# ifdef PERL_CORE
73d1d973 1841# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1842enum mem_log_type {
1843 MLT_ALLOC,
1844 MLT_REALLOC,
d7a2c63c
MHM
1845 MLT_FREE,
1846 MLT_NEW_SV,
1847 MLT_DEL_SV
0b0ab801
MHM
1848};
1849# endif
12754f92 1850# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1851void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1852void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1853# endif
0b0ab801
MHM
1854# endif
1855
fe4f188c
JH
1856#endif
1857
1858#ifdef PERL_MEM_LOG
d1401ee9
MHM
1859#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1860#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1861#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1862#endif
1863
1864#ifndef MEM_LOG_ALLOC
1865#define MEM_LOG_ALLOC(n,t,a) (a)
1866#endif
1867#ifndef MEM_LOG_REALLOC
1868#define MEM_LOG_REALLOC(n,t,v,a) (a)
1869#endif
1870#ifndef MEM_LOG_FREE
1871#define MEM_LOG_FREE(a) (a)
1872#endif
1873
d1401ee9
MHM
1874#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1875#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1876#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1877
1878#ifndef PERL_CORE
a02a5408
JC
1879/* pre 5.9.x compatibility */
1880#define New(x,v,n,t) Newx(v,n,t)
1881#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1882#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1883#endif
a02a5408 1884
ff68c719 1885#define Renew(v,n,t) \
d1401ee9 1886 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1887#define Renewc(v,n,t,c) \
d1401ee9 1888 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1889
1890#ifdef PERL_POISON
1891#define Safefree(d) \
06c0cc96 1892 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1893#else
fe4f188c 1894#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1895#endif
55497cff 1896
8b44ba4c
NC
1897#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1898#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1899#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1900
8b44ba4c
NC
1901#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1902#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1903#ifdef HAS_MEMSET
8b44ba4c 1904#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1905#else
8b44ba4c
NC
1906/* Using bzero(), which returns void. */
1907#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1908#endif
1909
7e337ee0
JH
1910#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1911#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1912#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1913#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1914
caa674f3
DD
1915#ifdef PERL_POISON
1916# define PERL_POISON_EXPR(x) x
1917#else
1918# define PERL_POISON_EXPR(x)
1919#endif
1920
2304df62 1921#ifdef USE_STRUCT_COPY
ff68c719 1922#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1923#else
1924#define StructCopy(s,d,t) Copy(s,d,1,t)
1925#endif
2cc61e15 1926
622913ab 1927#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1928#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1929
2cc61e15
DD
1930#ifdef NEED_VA_COPY
1931# ifdef va_copy
1932# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1933# else
a1866d1b
JH
1934# if defined(__va_copy)
1935# define Perl_va_copy(s, d) __va_copy(d, s)
1936# else
1937# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1938# endif
2cc61e15
DD
1939# endif
1940#endif
1941
472d47bc
SB
1942/* convenience debug macros */
1943#ifdef USE_ITHREADS
1944#define pTHX_FORMAT "Perl interpreter: 0x%p"
1945#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1946#define pTHX_VALUE_ (void *)my_perl,
1947#define pTHX_VALUE (void *)my_perl
1948#define pTHX__VALUE_ ,(void *)my_perl,
1949#define pTHX__VALUE ,(void *)my_perl
472d47bc 1950#else
3609ea0d 1951#define pTHX_FORMAT
472d47bc 1952#define pTHX__FORMAT
3609ea0d 1953#define pTHX_VALUE_
472d47bc 1954#define pTHX_VALUE
3609ea0d 1955#define pTHX__VALUE_
472d47bc
SB
1956#define pTHX__VALUE
1957#endif /* USE_ITHREADS */
3609ea0d 1958
2acdbac1
NC
1959/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1960 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1961 CPAN::Unpack show any users outside the core. */
1962#ifdef PERL_CORE
d1d15184 1963# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1964#endif
1965
dfff4baf
BF
1966/* Internal macros to deal with gids and uids */
1967#ifdef PERL_CORE
1968
1969# if Uid_t_size > IVSIZE
1970# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
1971# define SvUID(sv) SvNV(sv)
1972# else
1973# if Uid_t_sign <= 0
1974# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
1975# define SvUID(sv) SvIV(sv)
1976# else
1977# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
1978# define SvUID(sv) SvUV(sv)
1979# endif
1980# endif /* Uid_t_size */
1981
1982# if Gid_t_size > IVSIZE
1983# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
1984# define SvGID(sv) SvNV(sv)
1985# else
1986# if Gid_t_sign <= 0
1987# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
1988# define SvGID(sv) SvIV(sv)
1989# else
1990# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
1991# define SvGID(sv) SvUV(sv)
1992# endif
1993# endif /* Gid_t_size */
1994
1995#endif
1996
9d745869
KW
1997#endif /* HANDY_H */
1998
e9a8c099
MHM
1999/*
2000 * Local variables:
2001 * c-indentation-style: bsd
2002 * c-basic-offset: 4
14d04a33 2003 * indent-tabs-mode: nil
e9a8c099
MHM
2004 * End:
2005 *
14d04a33 2006 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 2007 */