This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Wrap various pod uses of NUL with C<>
[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
9c39b21c
JH
196/* INT64_C/UINT64_C are C99 from <stdint.h> (so they will not be
197 * available in strict C89 mode), but they are nice, so let's define
198 * them if necessary. */
0d3bd0a9 199#if defined(HAS_QUAD)
9c39b21c
JH
200# undef PeRl_INT64_C
201# undef PeRl_UINT64_C
202/* Prefer the native integer types (int and long) over long long
203 * (which is not C89) and Win32-specific __int64. */
204# if QUADKIND == QUAD_IS_INT && INTSIZE == 8
205# define PeRl_INT64_C(c) (c)
206# define PeRl_UINT64_C(c) CAT2(c,U)
207# endif
208# if QUADKIND == QUAD_IS_LONG && LONGSIZE == 8
209# define PeRl_INT64_C(c) CAT2(c,L)
210# define PeRl_UINT64_C(c) CAT2(c,UL)
211# endif
212# if QUADKIND == QUAD_IS_LONG_LONG && defined(HAS_LONG_LONG)
213# define PeRl_INT64_C(c) CAT2(c,LL)
214# define PeRl_UINT64_C(c) CAT2(c,ULL)
215# endif
216# if QUADKIND == QUAD_IS___INT64
217# define PeRl_INT64_C(c) CAT2(c,I64)
218# define PeRl_UINT64_C(c) CAT2(c,UI64)
219# endif
220# ifndef PeRl_INT64_C
221# define PeRl_INT64_C(c) ((I64TYPE)(c)) /* last resort */
222# define PeRl_UINT64_C(c) ((U64TYPE)(c))
223# endif
224/* In OS X the INT64_C/UINT64_C are defined with LL/ULL, which will
225 * not fly with C89-pedantic gcc, so let's undefine them first so that
226 * we can redefine them with our native integer preferring versions. */
8eafc6eb 227# if defined(PERL_DARWIN) && defined(PERL_GCC_PEDANTIC)
9c39b21c
JH
228# undef INT64_C
229# undef UINT64_C
230# endif
231# ifndef INT64_C
232# define INT64_C(c) PeRl_INT64_C(c)
233# endif
234# ifndef UINT64_C
235# define UINT64_C(c) PeRl_UINT64_C(c)
236# endif
e8c95190 237#endif
dc750f2e 238
d8668976 239#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 240
5ff3f7a4
GS
241/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
242 Please search CHAR_MAX in perl.h for further details. */
243#define U8_MAX UINT8_MAX
244#define U8_MIN UINT8_MIN
245
5ff3f7a4
GS
246#define I16_MAX INT16_MAX
247#define I16_MIN INT16_MIN
248#define U16_MAX UINT16_MAX
249#define U16_MIN UINT16_MIN
250
5ff3f7a4
GS
251#define I32_MAX INT32_MAX
252#define I32_MIN INT32_MIN
0e983133
GS
253#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
254# define U32_MAX UINT32_MAX
255#else
256# define U32_MAX 4294967295U
257#endif
5ff3f7a4
GS
258#define U32_MIN UINT32_MIN
259
260#else
261
5c9fa16e
KA
262/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
263 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
264#define U8_MAX PERL_UCHAR_MAX
265#define U8_MIN PERL_UCHAR_MIN
79072805 266
27d4fb96
PP
267#define I16_MAX PERL_SHORT_MAX
268#define I16_MIN PERL_SHORT_MIN
269#define U16_MAX PERL_USHORT_MAX
270#define U16_MIN PERL_USHORT_MIN
79072805 271
c4f23d77 272#if LONGSIZE > 4
27d4fb96
PP
273# define I32_MAX PERL_INT_MAX
274# define I32_MIN PERL_INT_MIN
275# define U32_MAX PERL_UINT_MAX
276# define U32_MIN PERL_UINT_MIN
79072805 277#else
27d4fb96
PP
278# define I32_MAX PERL_LONG_MAX
279# define I32_MIN PERL_LONG_MIN
280# define U32_MAX PERL_ULONG_MAX
281# define U32_MIN PERL_ULONG_MIN
79072805
LW
282#endif
283
5ff3f7a4
GS
284#endif
285
58a9a5d5 286/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
287#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
288#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
289#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
290
ff68c719 291#define Ctl(ch) ((ch) & 037)
8d063cd8 292
4650c663
KW
293/* This is a helper macro to avoid preprocessor issues, replaced by nothing
294 * unless under DEBUGGING, where it expands to an assert of its argument,
295 * followed by a comma (hence the comma operator). If we just used a straight
296 * assert(), we would get a comma with nothing before it when not DEBUGGING */
0f092d08
KW
297#ifdef DEBUGGING
298# define __ASSERT_(statement) assert(statement),
299#else
300# define __ASSERT_(statement)
301#endif
302
3fe05580
MHM
303/*
304=head1 SV-Body Allocation
305
306=for apidoc Ama|SV*|newSVpvs|const char* s
307Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
308
84bafc02
NC
309=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
310Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
311pair.
312
3fe05580
MHM
313=for apidoc Ama|SV*|newSVpvs_share|const char* s
314Like C<newSVpvn_share>, but takes a literal string instead of a string/length
315pair and omits the hash parameter.
316
9dcc53ea
Z
317=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
318Like C<sv_catpvn_flags>, but takes a literal string instead of a
319string/length pair.
320
321=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
322Like C<sv_catpvn_nomg>, but takes a literal string instead of a
323string/length pair.
324
7b4e7469 325=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
326Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
327
9dcc53ea
Z
328=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
329Like C<sv_catpvn_mg>, but takes a literal string instead of a
330string/length pair.
331
7b4e7469 332=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
333Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
334
9dcc53ea
Z
335=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
336Like C<sv_setpvn_mg>, but takes a literal string instead of a
337string/length pair.
338
339=for apidoc Am|SV *|sv_setref_pvs|const char* s
340Like C<sv_setref_pvn>, but takes a literal string instead of a
341string/length pair.
342
3fe05580
MHM
343=head1 Memory Management
344
345=for apidoc Ama|char*|savepvs|const char* s
346Like C<savepvn>, but takes a literal string instead of a string/length pair.
347
9dcc53ea
Z
348=for apidoc Ama|char*|savesharedpvs|const char* s
349A version of C<savepvs()> which allocates the duplicate string in memory
350which is shared between threads.
351
3fe05580
MHM
352=head1 GV Functions
353
354=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
355Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
356
357=head1 Hash Manipulation Functions
358
359=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
360Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
361
362=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
363Like C<hv_store>, but takes a literal string instead of a string/length pair
364and omits the hash parameter.
365
510966aa
Z
366=head1 Lexer interface
367
368=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
369
370Like L</lex_stuff_pvn>, but takes a literal string instead of a
371string/length pair.
372
3fe05580
MHM
373=cut
374*/
375
926bb54c
KW
376/* concatenating with "" ensures that only literal strings are accepted as
377 * argument */
26bfd924 378#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 379
926bb54c
KW
380/* note that STR_WITH_LEN() can't be used as argument to macros or functions
381 * that under some configurations might be macros, which means that it requires
382 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
383 */
384
385/* STR_WITH_LEN() shortcuts */
386#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
387#define newSVpvs_flags(str,flags) \
388 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 389#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
390#define sv_catpvs_flags(sv, str, flags) \
391 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
392#define sv_catpvs_nomg(sv, str) \
393 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
394#define sv_catpvs(sv, str) \
395 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
396#define sv_catpvs_mg(sv, str) \
397 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 398#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
399#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
400#define sv_setref_pvs(rv, classname, str) \
401 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 402#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
403#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
404#define gv_stashpvs(str, create) \
405 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
406#define gv_fetchpvs(namebeg, add, sv_type) \
407 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
408#define gv_fetchpvn(namebeg, len, add, sv_type) \
409 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
410#define sv_catxmlpvs(dsv, str, utf8) \
411 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
412#define hv_fetchs(hv,key,lval) \
413 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
414 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
415 : HV_FETCH_JUST_SV, NULL, 0))
416
417#define hv_stores(hv,key,val) \
418 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
419 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 420
510966aa
Z
421#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
422
b96d8cd9
NC
423#define get_cvs(str, flags) \
424 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 425
954c1994 426/*
ccfc67b7
JH
427=head1 Miscellaneous Functions
428
954c1994
GS
429=for apidoc Am|bool|strNE|char* s1|char* s2
430Test two strings to see if they are different. Returns true or
431false.
432
433=for apidoc Am|bool|strEQ|char* s1|char* s2
434Test two strings to see if they are equal. Returns true or false.
435
436=for apidoc Am|bool|strLT|char* s1|char* s2
437Test two strings to see if the first, C<s1>, is less than the second,
438C<s2>. Returns true or false.
439
440=for apidoc Am|bool|strLE|char* s1|char* s2
441Test two strings to see if the first, C<s1>, is less than or equal to the
442second, C<s2>. Returns true or false.
443
444=for apidoc Am|bool|strGT|char* s1|char* s2
445Test two strings to see if the first, C<s1>, is greater than the second,
446C<s2>. Returns true or false.
447
448=for apidoc Am|bool|strGE|char* s1|char* s2
449Test two strings to see if the first, C<s1>, is greater than or equal to
450the second, C<s2>. Returns true or false.
451
452=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
453Test two strings to see if they are different. The C<len> parameter
72d33970 454indicates the number of bytes to compare. Returns true or false. (A
954c1994
GS
455wrapper for C<strncmp>).
456
457=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
458Test two strings to see if they are equal. The C<len> parameter indicates
72d33970 459the number of bytes to compare. Returns true or false. (A wrapper for
954c1994
GS
460C<strncmp>).
461
462=cut
463*/
464
8d063cd8
LW
465#define strNE(s1,s2) (strcmp(s1,s2))
466#define strEQ(s1,s2) (!strcmp(s1,s2))
467#define strLT(s1,s2) (strcmp(s1,s2) < 0)
468#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
469#define strGT(s1,s2) (strcmp(s1,s2) > 0)
470#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
471#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
472#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 473
36477c24
PP
474#ifdef HAS_MEMCMP
475# define memNE(s1,s2,l) (memcmp(s1,s2,l))
476# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
477#else
478# define memNE(s1,s2,l) (bcmp(s1,s2,l))
479# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
480#endif
481
568a785a 482#define memEQs(s1, l, s2) \
8e6497b6 483 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
484#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
485
bbce6d69
PP
486/*
487 * Character classes.
488 *
489 * Unfortunately, the introduction of locales means that we
490 * can't trust isupper(), etc. to tell the truth. And when
491 * it comes to /\w+/ with tainting enabled, we *must* be able
492 * to trust our character classes.
493 *
494 * Therefore, the default tests in the text of Perl will be
495 * independent of locale. Any code that wants to depend on
496 * the current locale will use the tests that begin with "lc".
497 */
498
2304df62
AD
499#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
500# ifndef CTYPE256
501# define CTYPE256
502# endif
503#endif
504
954c1994 505/*
ccfc67b7
JH
506
507=head1 Character classes
243effed
KW
508This section is about functions (really macros) that classify characters
509into types, such as punctuation versus alphabetic, etc. Most of these are
510analogous to regular expression character classes. (See
511L<perlrecharclass/POSIX Character Classes>.) There are several variants for
512each class. (Not all macros have all variants; each item below lists the
513ones valid for it.) None are affected by C<use bytes>, and only the ones
514with C<LC> in the name are affected by the current locale.
515
516The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
517C<U8>) as input and returns a boolean as to whether or not the character
6aff1f14
KW
518represented by that octet is (or on non-ASCII platforms, corresponds to) an
519ASCII character in the named class based on platform, Unicode, and Perl rules.
520If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 521
6aff1f14
KW
522Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
523with no suffix C<"_A">.
4b9734bf
KW
524
525Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
526onto the platform. That is, the code points that are ASCII are unaffected,
527since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
528as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
529true when called with the code point 0xDF, which is a word character in both
4650c663 530ASCII and EBCDIC (though it represents different characters in each).
243effed
KW
531
532Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
533point as input. If the code point is larger than 255, Unicode rules are used
534to determine if it is in the character class. For example,
6aff1f14
KW
535C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
536WITH MACRON in Unicode, and is a word character.
243effed
KW
537
538Variant C<isFOO_utf8> is like C<isFOO_uni>, 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.
243effed 542
31f05a37
KW
543Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but the
544result is based on the current locale, which is what C<LC> in the name stands
545for. If Perl can determine that the current locale is a UTF-8 locale, it uses
546the published Unicode rules; otherwise, it uses the C library function that
547gives the named classification. For example, C<isDIGIT_LC()> when not in a
548UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always
549returned if the input won't fit into an octet.
243effed
KW
550
551Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
552returns the same as C<isFOO_LC> for input code points less than 256, and
553returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
554
555Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
556(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
557classification of just the first (possibly multi-byte) character in the string
558is tested.
ccfc67b7 559
954c1994 560=for apidoc Am|bool|isALPHA|char ch
2787a470 561Returns a boolean indicating whether the specified character is an
6aff1f14 562alphabetic character, analogous to C<m/[[:alpha:]]/>.
8a58bdcf 563See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 564C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 565C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 566
15861f94
KW
567=for apidoc Am|bool|isALPHANUMERIC|char ch
568Returns a boolean indicating whether the specified character is a either an
6aff1f14 569alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
15861f94
KW
570See the L<top of this section|/Character classes> for an explanation of variants
571C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 572C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
573C<isALPHANUMERIC_LC_utf8>.
574
8a58bdcf
KW
575=for apidoc Am|bool|isASCII|char ch
576Returns a boolean indicating whether the specified character is one of the 128
243effed 577characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 578On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
579character corresponds to an ASCII character. Variants C<isASCII_A()> and
580C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
581See the L<top of this section|/Character classes> for an explanation of variants
582C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
583C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
584library routine C<isascii()>. In these cases, the variants whose names contain
585C<LC> are the same as the corresponding ones without.
586
d98532ea
KW
587Also note, that because all ASCII characters are UTF-8 invariant (meaning they
588have the exact same representation (always a single byte) whether encoded in
589UTF-8 or not), C<isASCII> will give the correct results when called with any
590byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
591work properly on any string encoded or not in UTF-8.
592
243effed
KW
593=for apidoc Am|bool|isBLANK|char ch
594Returns a boolean indicating whether the specified character is a
6aff1f14 595character considered to be a blank, analogous to C<m/[[:blank:]]/>.
243effed 596See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 597C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
598C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
599platforms do not have the C library routine C<isblank()>. In these cases, the
600variants whose names contain C<LC> are the same as the corresponding ones
601without.
602
603=for apidoc Am|bool|isCNTRL|char ch
604Returns a boolean indicating whether the specified character is a
6aff1f14 605control character, analogous to C<m/[[:cntrl:]]/>.
243effed 606See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 607C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
608C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
609On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 610
954c1994 611=for apidoc Am|bool|isDIGIT|char ch
2787a470 612Returns a boolean indicating whether the specified character is a
6aff1f14 613digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 614Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed 615See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 616C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
617C<isDIGIT_LC_utf8>.
618
619=for apidoc Am|bool|isGRAPH|char ch
620Returns a boolean indicating whether the specified character is a
6aff1f14 621graphic character, analogous to C<m/[[:graph:]]/>.
243effed 622See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 623C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 624C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 625
0c82b6df 626=for apidoc Am|bool|isLOWER|char ch
2787a470 627Returns a boolean indicating whether the specified character is a
6aff1f14 628lowercase character, analogous to C<m/[[:lower:]]/>.
8a58bdcf 629See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 630C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 631C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 632
c99e91e9 633=for apidoc Am|bool|isOCTAL|char ch
2787a470 634Returns a boolean indicating whether the specified character is an
6aff1f14 635octal digit, [0-7].
243effed
KW
636The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
637C<isOCTAL>.
638
639=for apidoc Am|bool|isPUNCT|char ch
640Returns a boolean indicating whether the specified character is a
6aff1f14
KW
641punctuation character, analogous to C<m/[[:punct:]]/>.
642Note that the definition of what is punctuation isn't as
243effed
KW
643straightforward as one might desire. See L<perlrecharclass/POSIX Character
644Classes> for details.
645See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 646C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 647C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 648
0c82b6df 649=for apidoc Am|bool|isSPACE|char ch
2787a470 650Returns a boolean indicating whether the specified character is a
6aff1f14 651whitespace character. This is analogous
398d098a
KW
652to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
653(experimentally), this also matches what C<m/[[:space:]]/> does.
a04e6aad 654("Experimentally" means that this change may be backed out in 5.22 if
398d098a
KW
655field experience indicates that it was unwise.) Prior to 5.18, only the
656locale forms of this macro (the ones with C<LC> in their names) matched
657precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
658in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
659(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
8a58bdcf 660See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 661C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 662C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 663
398d098a
KW
664=for apidoc Am|bool|isPSXSPC|char ch
665(short for Posix Space)
666Starting in 5.18, this is identical (experimentally) in all its forms to the
667corresponding C<isSPACE()> macros. ("Experimentally" means that this change
a04e6aad 668may be backed out in 5.22 if field experience indicates that it
398d098a
KW
669was unwise.)
670The locale forms of this macro are identical to their corresponding
671C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
672non-locale forms differ from their C<isSPACE()> forms only in that the
673C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
674Otherwise they are identical. Thus this macro is analogous to what
675C<m/[[:space:]]/> matches in a regular expression.
676See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 677C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
678C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
679
954c1994 680=for apidoc Am|bool|isUPPER|char ch
2787a470 681Returns a boolean indicating whether the specified character is an
6aff1f14 682uppercase character, analogous to C<m/[[:upper:]]/>.
8a58bdcf 683See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 684C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 685C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 686
243effed 687=for apidoc Am|bool|isPRINT|char ch
8eea39dd 688Returns a boolean indicating whether the specified character is a
6aff1f14 689printable character, analogous to C<m/[[:print:]]/>.
8a58bdcf 690See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 691C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
692C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
693
694=for apidoc Am|bool|isWORDCHAR|char ch
695Returns a boolean indicating whether the specified character is a character
696that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
697in a regular expression. A word character is an alphabetic character, a
698decimal digit, a connecting punctuation character (such as an underscore), or
699a "mark" character that attaches to one of those (like some sort of accent).
700C<isALNUM()> is a synonym provided for backward compatibility, even though a
701word character includes more than the standard C language meaning of
702alphanumeric.
703See the L<top of this section|/Character classes> for an explanation of variants
704C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
705C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
706
707=for apidoc Am|bool|isXDIGIT|char ch
708Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
709digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
710and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
711See the L<top of this section|/Character classes> for an explanation of variants
712C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
713C<isXDIGIT_LC_utf8>.
714
3c3ecf18
KW
715=for apidoc Am|bool|isIDFIRST|char ch
716Returns a boolean indicating whether the specified character can be the first
717character of an identifier. This is very close to, but not quite the same as
718the official Unicode property C<XID_Start>. The difference is that this
719returns true only if the input character also matches L</isWORDCHAR>.
720See the L<top of this section|/Character classes> for an explanation of variants
721C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
722C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
723
724=for apidoc Am|bool|isIDCONT|char ch
725Returns a boolean indicating whether the specified character can be the
726second or succeeding character of an identifier. This is very close to, but
727not quite the same as the official Unicode property C<XID_Continue>. The
728difference is that this returns true only if the input character also matches
729L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
730explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
731C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
732C<isIDCONT_LC_utf8>.
733
243effed 734=head1 Miscellaneous Functions
8eea39dd 735
95a59cab 736=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 737Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
738Behaviour is only well defined when isXDIGIT(*str) is true.
739
e7c1e6c1
KW
740=head1 Character case changing
741
1f607577
KW
742=for apidoc Am|U8|toUPPER|U8 ch
743Converts the specified character to uppercase. If the input is anything but an
744ASCII lowercase character, that input character itself is returned. Variant
c753c8d3 745C<toUPPER_A> is equivalent.
954c1994 746
1f607577
KW
747=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
748Converts the Unicode code point C<cp> to its uppercase version, and
749stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
750that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
751bytes since the uppercase version may be longer than the original character.
752
753The first code point of the uppercased version is returned
754(but note, as explained just above, that there may be more.)
755
756=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
757Converts the UTF-8 encoded character at C<p> to its uppercase 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 uppercase version may be longer than the original character.
761
762The first code point of the uppercased version is returned
763(but note, as explained just above, that there may be more.)
764
765The input character at C<p> is assumed to be well-formed.
766
25200305
KW
767=for apidoc Am|U8|toFOLD|U8 ch
768Converts the specified character to foldcase. If the input is anything but an
769ASCII uppercase character, that input character itself is returned. Variant
770C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
771Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
772
1f607577
KW
773=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
774Converts the Unicode code point C<cp> to its foldcase version, and
775stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
776that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
777bytes since the foldcase version may be longer than the original character.
778
779The first code point of the foldcased version is returned
780(but note, as explained just above, that there may be more.)
781
782=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
783Converts the UTF-8 encoded character at C<p> to its foldcase version, and
784stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
785that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
786bytes since the foldcase version may be longer than the original character.
787
788The first code point of the foldcased version is returned
789(but note, as explained just above, that there may be more.)
790
791The input character at C<p> is assumed to be well-formed.
792
793=for apidoc Am|U8|toLOWER|U8 ch
794Converts the specified character to lowercase. If the input is anything but an
795ASCII uppercase character, that input character itself is returned. Variant
c753c8d3 796C<toLOWER_A> is equivalent.
954c1994 797
1f607577
KW
798=for apidoc Am|U8|toLOWER_L1|U8 ch
799Converts the specified Latin1 character to lowercase. The results are undefined if
800the input doesn't fit in a byte.
801
802=for apidoc Am|U8|toLOWER_LC|U8 ch
803Converts the specified character to lowercase using the current locale's rules,
804if possible; otherwise returns the input character itself.
805
806=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
807Converts the Unicode code point C<cp> to its lowercase version, and
808stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
809that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
810bytes since the lowercase version may be longer than the original character.
811
812The first code point of the lowercased version is returned
813(but note, as explained just above, that there may be more.)
814
815=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
816Converts the UTF-8 encoded character at C<p> to its lowercase version, and
817stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
818that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
819bytes since the lowercase version may be longer than the original character.
820
821The first code point of the lowercased version is returned
822(but note, as explained just above, that there may be more.)
823
824The input character at C<p> is assumed to be well-formed.
825
826=for apidoc Am|U8|toLOWER_LC|U8 ch
827Converts the specified character to lowercase using the current locale's rules,
828if possible; otherwise returns the input character itself.
829
25200305
KW
830=for apidoc Am|U8|toTITLE|U8 ch
831Converts the specified character to titlecase. If the input is anything but an
832ASCII lowercase character, that input character itself is returned. Variant
833C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
834as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
835concept used in locale handling, so there is no functionality for that.)
836
1f607577
KW
837=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
838Converts the Unicode code point C<cp> to its titlecase version, and
839stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
840that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
841bytes since the titlecase version may be longer than the original character.
842
843The first code point of the titlecased version is returned
844(but note, as explained just above, that there may be more.)
845
846=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
847Converts the UTF-8 encoded character at C<p> to its titlecase version, and
848stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
849that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
850bytes since the titlecase version may be longer than the original character.
851
852The first code point of the titlecased version is returned
853(but note, as explained just above, that there may be more.)
854
855The input character at C<p> is assumed to be well-formed.
856
954c1994 857=cut
353c9b6f 858
1f607577 859XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
25200305 860really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
243effed 861
8a58bdcf 862Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
863patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
864
954c1994
GS
865*/
866
7c062697
KW
867/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
868 * is known only in the perl core, and this macro can be called from outside
869 * that */
870#ifdef HAS_QUAD
871# define WIDEST_UTYPE U64TYPE
872#else
873# define WIDEST_UTYPE U32
874#endif
875
3912bc88
KW
876/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
877 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
878 * bits of information are lost even on a 64-bit machine, but to get the
879 * compiler to optimize it out if possible. This is because Configure makes
880 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
881 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
882 */
883#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 884
41f43cc2 885#ifdef EBCDIC
b6340bd0
KW
886# ifndef _ALL_SOURCE
887 /* This returns the wrong results on at least z/OS unless this is
888 * defined. */
889# error _ALL_SOURCE should probably be defined
890# endif
891
892 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
893 * likely not defined, and so we use the native function */
3cedda69 894# define isASCII(c) cBOOL(isascii(c))
41f43cc2
KW
895#else
896# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
897#endif
898
ec79cfbb 899#define isASCII_A(c) isASCII(c)
3509bb53 900#define isASCII_L1(c) isASCII(c)
6515cb76 901
c2da0b36
KW
902/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
903 * and the 8 possible permutations of those bits exactly comprise the 8 octal
904 * digits */
3d33b3e3 905#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 906
ec79cfbb 907/* ASCII range only */
9fb1bf9d 908#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 909
a500dc72
KW
910/* Character class numbers. For internal core Perl use only. The ones less
911 * than 32 are used in PL_charclass[] and the ones up through the one that
912 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
913 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
914 * their actual definitions are here. If that file has a name not used here,
915 * it won't compile.
1709d539
KW
916 *
917 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
918 * frequency of their use. This gives a slight edge to exiting a loop earlier
919 * (in reginclass() in regexec.c) */
1709d539
KW
920# define _CC_WORDCHAR 0 /* \w and [:word:] */
921# define _CC_DIGIT 1 /* \d and [:digit:] */
922# define _CC_ALPHA 2 /* [:alpha:] */
923# define _CC_LOWER 3 /* [:lower:] */
924# define _CC_UPPER 4 /* [:upper:] */
925# define _CC_PUNCT 5 /* [:punct:] */
926# define _CC_PRINT 6 /* [:print:] */
15861f94 927# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 928# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 929# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 930
b0d691b2 931#define _FIRST_NON_SWASH_CC 10
1709d539
KW
932/* The character classes above are implemented with swashes. The second group
933 * (just below) contains the ones implemented without. These are also sorted
934 * in rough order of the frequency of their use, except that \v should be last,
935 * as it isn't a real Posix character class, and some (small) inefficiencies in
936 * regular expression handling would be introduced by putting it in the middle
937 * of those that are. Also, cntrl and ascii come after the others as it may be
938 * useful to group these which have no members that match above Latin1, (or
939 * above ASCII in the latter case) */
940
b0d691b2
KW
941# define _CC_SPACE 10 /* \s */
942# define _CC_BLANK 11 /* [:blank:] */
943# define _CC_XDIGIT 12 /* [:xdigit:] */
944# define _CC_PSXSPC 13 /* [:space:] */
945# define _CC_CNTRL 14 /* [:cntrl:] */
946# define _CC_ASCII 15 /* [:ascii:] */
947# define _CC_VERTSPACE 16 /* \v */
1709d539 948
a0947d7b
KW
949# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
950
1709d539 951/* The members of the third group below do not need to be coordinated with data
3ded5eb0
KW
952 * structures in regcomp.[ch] and regexec.c. But they should be added to
953 * bootstrap_ctype() */
b0d691b2
KW
954# define _CC_IDFIRST 17
955# define _CC_CHARNAME_CONT 18
956# define _CC_NONLATIN1_FOLD 19
957# define _CC_QUOTEMETA 20
958# define _CC_NON_FINAL_FOLD 21
959# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
960# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
961/* Unused: 23-30
f4cdb42c
KW
962 * If more bits are needed, one could add a second word for non-64bit
963 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
964 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
965 * is used only for optimization (as of this writing), and differs in the
966 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
967 * characters: the masculine and feminine ordinal indicators, so removing it
968 * would just cause /i regexes which match them to run less efficiently */
96ac0975 969
3a371f2f
KW
970#if defined(PERL_CORE) || defined(PERL_EXT)
971/* An enum version of the character class numbers, to help compilers
972 * optimize */
973typedef enum {
3a371f2f 974 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
975 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
976 _CC_ENUM_ASCII = _CC_ASCII,
977 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 978 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 979 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
980 _CC_ENUM_DIGIT = _CC_DIGIT,
981 _CC_ENUM_GRAPH = _CC_GRAPH,
982 _CC_ENUM_LOWER = _CC_LOWER,
983 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 984 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 985 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 986 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 987 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 988 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 989 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 990 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
991} _char_class_number;
992#endif
993
63c61c3f 994#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 995#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 996
3018b823 997#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 998# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
999 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
1000 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
1001 #error Need to adjust order of swash_property_names[]
1002# endif
1003
b44daeeb
KW
1004/* This is declared static in each of the few files that this is #defined for
1005 * to keep them from being publicly accessible. Hence there is a small amount
1006 * of wasted space */
1007
922e8cb4
KW
1008static const char* const swash_property_names[] = {
1009 "XPosixWord",
1010 "XPosixDigit",
1011 "XPosixAlpha",
1012 "XPosixLower",
1013 "XPosixUpper",
1014 "XPosixPunct",
1015 "XPosixPrint",
1016 "XPosixAlnum",
b0d691b2
KW
1017 "XPosixGraph",
1018 "Cased"
922e8cb4
KW
1019};
1020#endif
1021
96ac0975
NC
1022# ifdef DOINIT
1023EXTCONST U32 PL_charclass[] = {
1024# include "l1_char_class_tab.h"
1025};
1026
1027# else /* ! DOINIT */
1028EXTCONST U32 PL_charclass[];
1029# endif
3ded5eb0
KW
1030#endif /* Has perl.h */
1031
1032#if defined(H_PERL) && ! defined(BOOTSTRAP_CHARSET)
96ac0975 1033
265c1f46 1034 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70 1035# define _CC_mask(classnum) (1U << (classnum))
4650c663
KW
1036
1037 /* For internal core Perl use only: the base macro for defining macros like
1038 * isALPHA */
ff7ecfc3 1039# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
f4cd282c 1040 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
4eeeb416 1041
f4cdb42c
KW
1042 /* The mask for the _A versions of the macros; it just adds in the bit for
1043 * ASCII. */
1044# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1045
4650c663
KW
1046 /* For internal core Perl use only: the base macro for defining macros like
1047 * isALPHA_A. The foo_A version makes sure that both the desired bit and
1048 * the ASCII bit are present */
3ded5eb0 1049# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
f4cd282c 1050 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
f4cdb42c
KW
1051 == _CC_mask_A(classnum)))
1052
f4cdb42c 1053# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 1054# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
1055# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1056# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1057# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
1058# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1059# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1060# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1061# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1062# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1063# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1064# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1065# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1066# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
d95f8b6a 1067# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
3ded5eb0
KW
1068# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1069# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1070# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1071
1072 /* continuation character for legal NAME in \N{NAME} */
1073# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1074
1075# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1076# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1077# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1078# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1079# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1080# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1081# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1082# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1083# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1084# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
f4cdb42c 1085
fb9bfbf7
KW
1086 /* Either participates in a fold with a character above 255, or is a
1087 * multi-char fold */
f4cd282c 1088# 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 1089
4eeeb416 1090# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
1091# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1092 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
1093# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1094 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
3ded5eb0
KW
1095#else /* Either don't have perl.h or don't want to use char_class_tab.h */
1096
1097 /* If we don't have perl.h, we are compiling a utility program. Below we
1098 * hard-code various macro definitions that wouldn't otherwise be available
1099 * to it. We can also get here if we are configured to bootstrap up Perl
1100 * on a non-ASCII platform that doesn't have a working Perl (currently only
1101 * EBCDIC). For these we currently use the native definitions to get
1102 * things going. (It should also be possible to use the translation
1103 * function NATIVE_TO_LATIN1(), but that is an extra layer of dependence on
1104 * Perl, so it is currently avoided for the macros where it's possible to
1105 * do so.) */
9fb1bf9d 1106# ifdef EBCDIC
3ded5eb0 1107 /* Use the native functions. They likely will return false for all
4f48cecf 1108 * non-ASCII values, but this makes sure */
b6340bd0
KW
1109# define isLOWER_A(c) (isASCII(c) && islower(c))
1110# define isPRINT_A(c) (isASCII(c) && isprint(c))
b6340bd0 1111# define isUPPER_A(c) (isASCII(c) && isupper(c))
3ded5eb0 1112# else /* ASCII platform. These are coded based on first principals */
9fb1bf9d
KW
1113# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1114# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
2181da29 1115# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
3ded5eb0 1116# endif /* Below are common definitions for ASCII and non-ASCII */
a4d7a999
KW
1117# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
1118# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
3ded5eb0 1119# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
a4d7a999
KW
1120# define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
1121# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
1122# define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
3ded5eb0 1123# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
a4d7a999
KW
1124# define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1125# define isSPACE_A(c) ((c) == ' ' \
1126 || (c) == '\t' \
1127 || (c) == '\n' \
1128 || (c) == '\r' \
1129 || (c) == '\v' \
1130 || (c) == '\f')
3ded5eb0 1131# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
a4d7a999
KW
1132# define isXDIGIT_A(c) (isDIGIT_A(c) \
1133 || ((c) >= 'a' && (c) <= 'f') \
1134 || ((c) <= 'F' && (c) >= 'A'))
3ded5eb0
KW
1135
1136 /* The _L1 macros may be unnecessary for both the utilities and for
1137 * bootstrapping; I (khw) added them during debugging of bootstrapping, and
1138 * it seems best to keep them. */
1139# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1140# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1141# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1142# define isBLANK_L1(c) (isBLANK_A(c) \
1143 || (FITS_IN_8_BITS(c) \
1144 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1145# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1146# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1147# define isLOWER_L1(c) (isLOWER_A(c) \
1148 || (FITS_IN_8_BITS(c) \
1149 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1150 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1151 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1152 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1153 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1154# define isPRINT_L1(c) (isPRINT_A(c) \
1155 || (FITS_IN_8_BITS(c) \
1156 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1157# define isPSXSPC_L1(c) isSPACE_L1(c)
1158# define isPUNCT_L1(c) (isPUNCT_A(c) \
1159 || (FITS_IN_8_BITS(c) \
1160 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1161 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1162 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1163 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1164 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1165 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1166 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1167# define isSPACE_L1(c) (isSPACE_A(c) \
1168 || (FITS_IN_8_BITS(c) \
1169 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1170 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1171# define isUPPER_L1(c) (isUPPER_A(c) \
1172 || (FITS_IN_8_BITS(c) \
1173 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1174 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1175 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1176# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1177# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1178# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1179 || isBLANK_L1(c) \
1180 || (c) == '-' \
1181 || (c) == '(' \
1182 || (c) == ')')
1183 /* The following are not fully accurate in the above-ASCII range. I (khw)
1184 * don't think it's necessary to be so for the purposes where this gets
1185 * compiled */
1186# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1187# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1188
1189 /* And these aren't accurate at all. They are useful only for above
1190 * Latin1, which utilities and bootstrapping don't deal with */
1191# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1192# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1193
1194 /* Many of the macros later in this file are defined in terms of these. By
1195 * implementing them with a function, which converts the class number into
1196 * a call to the desired macro, all of the later ones work. However, that
1197 * function won't be actually defined when building a utility program (no
1198 * perl.h), and so a compiler error will be generated if one is attempted
1199 * to be used. And the above-Latin1 code points require Unicode tables to
1200 * be present, something unlikely to be the case when bootstrapping */
1201# define _generic_isCC(c, classnum) \
1202 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1203# define _generic_isCC_A(c, classnum) \
1204 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
1205#endif /* End of no perl.h or have BOOTSTRAP_CHARSET */
8a58bdcf 1206
e66b99e9
KW
1207#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1208#define isALPHA(c) isALPHA_A(c)
1209#define isBLANK(c) isBLANK_A(c)
1210#define isCNTRL(c) isCNTRL_A(c)
1211#define isDIGIT(c) isDIGIT_A(c)
1212#define isGRAPH(c) isGRAPH_A(c)
1213#define isIDFIRST(c) isIDFIRST_A(c)
1214#define isLOWER(c) isLOWER_A(c)
1215#define isPRINT(c) isPRINT_A(c)
1216#define isPSXSPC(c) isPSXSPC_A(c)
1217#define isPUNCT(c) isPUNCT_A(c)
1218#define isSPACE(c) isSPACE_A(c)
1219#define isUPPER(c) isUPPER_A(c)
1220#define isWORDCHAR(c) isWORDCHAR_A(c)
1221#define isXDIGIT(c) isXDIGIT_A(c)
1222
1223/* ASCII casing. These could also be written as
1224 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1225 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1226 which uses table lookup and mask instead of subtraction. (This would
1227 work because the _MOD does not apply in the ASCII range) */
68067e4e
DM
1228#define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1229#define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
bbce6d69 1230
25200305
KW
1231/* In the ASCII range, these are equivalent to what they're here defined to be.
1232 * But by creating these definitions, other code doesn't have to be aware of
1233 * this detail */
1234#define toFOLD(c) toLOWER(c)
25200305
KW
1235#define toTITLE(c) toUPPER(c)
1236
c753c8d3
KW
1237#define toLOWER_A(c) toLOWER(c)
1238#define toUPPER_A(c) toUPPER(c)
25200305
KW
1239#define toFOLD_A(c) toFOLD(c)
1240#define toTITLE_A(c) toTITLE(c)
1a0901db 1241
4650c663 1242/* Use table lookup for speed; returns the input itself if is out-of-range */
b2bf251f 1243#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1244 ? (c) \
f4cd282c 1245 : PL_latin1_lc[ (U8) (c) ])
c753c8d3
KW
1246#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1247
1a0901db 1248/* Modified uc. Is correct uc except for three non-ascii chars which are
4650c663
KW
1249 * all mapped to one of them, and these need special handling; returns the
1250 * input itself if is out-of-range */
b2bf251f 1251#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1252 ? (c) \
f4cd282c 1253 : PL_mod_latin1_uc[ (U8) (c) ])
31f05a37 1254#define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale
84061b6a 1255
beab9ebe
KW
1256/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
1257
1258/* For internal core Perl use only: the base macro for defining macros like
1259 * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point
31f05a37
KW
1260 * (0-255) to check. In a UTF-8 locale, the result is the same as calling
1261 * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like
1262 * _CC_UPPER, which gives the class number for doing this. For non-UTF-8
1263 * locales, the code to actually do the test this is passed in 'non_utf8'. If
1264 * 'c' is above 255, 0 is returned. For accessing the full range of possible
1265 * code points under locale rules, use the macros based on _generic_LC_uvchr
1266 * instead of this. */
beab9ebe
KW
1267#define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \
1268 (! FITS_IN_8_BITS(c) \
1269 ? 0 \
31f05a37
KW
1270 : IN_UTF8_CTYPE_LOCALE \
1271 ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \
beab9ebe
KW
1272 : cBOOL(non_utf8))
1273
1274/* For internal core Perl use only: a helper macro for defining macros like
1275 * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to
1276 * actually do this test is passed in 'non_utf8_func', which is called on 'c',
1277 * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See
1278 * _generic_LC_base for more info */
1279#define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \
1280 _generic_LC_base(c,utf8_locale_classnum, \
1281 non_utf8_func( (_LC_CAST) (c)))
1282
1283/* For internal core Perl use only: like _generic_LC, but also returns TRUE if
1284 * 'c' is the platform's native underscore character */
1285#define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \
1286 _generic_LC_base(c, utf8_locale_classnum, \
1287 (non_utf8_func( (_LC_CAST) (c)) \
1288 || (char)(c) == '_'))
1289
1290/* These next three are also for internal core Perl use only: case-change
1291 * helper macros */
1292#define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \
1293 ? (c) \
31f05a37
KW
1294 : (IN_UTF8_CTYPE_LOCALE) \
1295 ? PL_latin1_lc[ (U8) (c) ] \
beab9ebe
KW
1296 : function((cast)(c)))
1297
31f05a37
KW
1298/* Note that the result can be larger than a byte in a UTF-8 locale. It
1299 * returns a single value, so can't adequately return the upper case of LATIN
1300 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1301 * values "SS"); instead it asserts against that under DEBUGGING, and
1302 * otherwise returns its input */
beab9ebe
KW
1303#define _generic_toUPPER_LC(c, function, cast) \
1304 (! FITS_IN_8_BITS(c) \
1305 ? (c) \
31f05a37
KW
1306 : ((! IN_UTF8_CTYPE_LOCALE) \
1307 ? function((cast)(c)) \
1308 : ((((U8)(c)) == MICRO_SIGN) \
1309 ? GREEK_CAPITAL_LETTER_MU \
1310 : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \
1311 ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \
1312 : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \
1313 ? (__ASSERT_(0) (c)) \
1314 : PL_mod_latin1_uc[ (U8) (c) ])))))
1315
1316/* Note that the result can be larger than a byte in a UTF-8 locale. It
1317 * returns a single value, so can't adequately return the fold case of LATIN
1318 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1319 * values "ss"); instead it asserts against that under DEBUGGING, and
1320 * otherwise returns its input */
beab9ebe 1321#define _generic_toFOLD_LC(c, function, cast) \
4d7db1b9 1322 ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \
31f05a37 1323 ? GREEK_SMALL_LETTER_MU \
4d7db1b9
KW
1324 : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \
1325 || (c) != LATIN_SMALL_LETTER_SHARP_S) \
1326 _generic_toLOWER_LC(c, function, cast)))
beab9ebe 1327
84061b6a
KW
1328/* Use the libc versions for these if available. */
1329#if defined(HAS_ISASCII) && ! defined(USE_NEXT_CTYPE)
1330# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c)))
1331#else
1332# define isASCII_LC(c) isASCII(c)
1333#endif
1334
1335#if defined(HAS_ISBLANK) && ! defined(USE_NEXT_CTYPE)
1336# define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank)
95f34b6f
KW
1337#else /* Unlike isASCII, varies if in a UTF-8 locale */
1338# define isBLANK_LC(c) (IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c)
84061b6a
KW
1339#endif
1340
4650c663 1341#ifdef USE_NEXT_CTYPE /* NeXT computers */
bbce6d69 1342
beab9ebe
KW
1343# define _LC_CAST unsigned int /* Needed by _generic_LC. NeXT functions
1344 use this as their input type */
1345
1346# define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, NXIsAlpha)
1347# define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, NXIsAlNum)
1348# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, NXIsCntrl)
1349# define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, NXIsDigit)
1350# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, NXIsGraph)
1351# define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, NXIsAlpha)
1352# define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, NXIsLower)
1353# define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, NXIsPrint)
1354# define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, NXIsPunct)
1355# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, NXIsSpace)
1356# define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, NXIsUpper)
1357# define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, NXIsAlNum)
1358# define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, NXIsXdigit)
1359
1360# define toLOWER_LC(c) _generic_toLOWER_LC((c), NXToLower, unsigned int)
1361# define toUPPER_LC(c) _generic_toUPPER_LC((c), NXToUpper, unsigned int)
1362# define toFOLD_LC(c) _generic_toFOLD_LC((c), NXToLower, unsigned int)
bbce6d69
PP
1363
1364#else /* !USE_NEXT_CTYPE */
b8c5462f 1365
841f64f7
BF
1366# define _LC_CAST U8
1367
bbce6d69 1368# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
4650c663 1369 /* For most other platforms */
beab9ebe 1370
beab9ebe
KW
1371# define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha)
1372# define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum)
1373# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1374# define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit)
1375# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph)
1376# define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha)
1377# define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower)
1378# define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint)
1379# define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct)
1380# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1381# define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper)
1382# define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum)
1383# define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit)
1384
1385
1386# define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1387# define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1388# define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
bbce6d69 1389
4650c663 1390# else /* The final fallback position */
bbce6d69 1391
bbce6d69 1392# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1393# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
b8c5462f 1394# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1395# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1396# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1397# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1398# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1399# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1400# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1401# define isSPACE_LC(c) (isascii(c) && isspace(c))
1402# define isUPPER_LC(c) (isascii(c) && isupper(c))
1403# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1404# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
4650c663 1405
76fde96a 1406# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1407# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
beab9ebe 1408# define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c))
bbce6d69
PP
1409
1410# endif
a0d0e21e 1411#endif /* USE_NEXT_CTYPE */
55204971 1412
eba68aa0
KW
1413#define isIDCONT(c) isWORDCHAR(c)
1414#define isIDCONT_A(c) isWORDCHAR_A(c)
1415#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1416#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1417#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1418
4650c663
KW
1419/* For internal core Perl use only: the base macros for defining macros like
1420 * isALPHA_uni. 'c' is the code point to check. 'classnum' is the POSIX class
1421 * number defined earlier in this file. _generic_uni() is used for POSIX
1422 * classes where there is a macro or function 'above_latin1' that takes the
1423 * single argument 'c' and returns the desired value. These exist for those
1424 * classes which have simple definitions, avoiding the overhead of a hash
1425 * lookup or inversion list binary search. _generic_swash_uni() can be used
1426 * for classes where that overhead is faster than a direct lookup.
1427 * _generic_uni() won't compile if 'c' isn't unsigned, as it won't match the
1428 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so
1429 * have duplicate checks here, so could create versions of the macros that
1430 * don't, but experiments show that gcc optimizes them out anyway. */
66c17564
KW
1431
1432/* Note that all ignore 'use bytes' */
cd500f2f
KW
1433#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1434 ? _generic_isCC(c, classnum) \
1435 : above_latin1(c))
1436#define _generic_swash_uni(classnum, c) ((c) < 256 \
1437 ? _generic_isCC(c, classnum) \
922e8cb4 1438 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1439#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1440#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1441#define isASCII_uni(c) isASCII(c)
1442#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1443#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1444#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1445#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1446#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1447#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1448#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1449#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1450
509fb054
KW
1451/* Posix and regular space are identical above Latin1 */
1452#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1453
cd500f2f 1454#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1455#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1456#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1457#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1458#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1459#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1460
b8d68ded 1461#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1462#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1463#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1464#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1465
4650c663
KW
1466/* For internal core Perl use only: the base macros for defining macros like
1467 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code
1468 * point, not just 0-255. Like _generic_uni, there are two versions, one for
1469 * simple class definitions; the other for more complex. These are like
1470 * _generic_uni, so see it for more info. */
cd500f2f
KW
1471#define _generic_LC_uvchr(latin1, above_latin1, c) \
1472 (c < 256 ? latin1(c) : above_latin1(c))
1473#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1474 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1475
1476#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1477#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1478 _CC_ALPHANUMERIC, c)
509fb054 1479#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1480#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1481#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1482#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1483#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1484#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1485 _is_uni_perl_idcont, c)
cd500f2f
KW
1486#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1487 _is_uni_perl_idstart, c)
1488#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1489#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1490#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1491 space under locale */
cd500f2f
KW
1492#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1493#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1494 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1495#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1496#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1497 _CC_WORDCHAR, c)
1498#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1499
36b00f93 1500#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1501
4650c663
KW
1502/* For internal core Perl use only: the base macros for defining macros like
1503 * isALPHA_utf8. These are like the earlier defined macros, but take an input
1504 * UTF-8 encoded string 'p'. If the input is in the Latin1 range, use
1505 * the Latin1 macro 'classnum' on 'p'. Otherwise use the value given by the
1506 * 'utf8' parameter. This relies on the fact that ASCII characters have the
1507 * same representation whether utf8 or not. Note that it assumes that the utf8
1508 * has been validated, and ignores 'use bytes' */
cd500f2f 1509#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1510 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1511 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c 1512 ? _generic_isCC( \
94bb8c36 1513 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
aea0c08c 1514 *((p)+1 )), \
94bb8c36 1515 classnum) \
252810d0 1516 : utf8)
cd500f2f
KW
1517/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1518 * can be a macro */
1519#define _generic_func_utf8(classnum, above_latin1, p) \
1520 _generic_utf8(classnum, p, above_latin1(p))
4650c663 1521/* Like the above, but passes classnum to _isFOO_utf8(), instead of having an
cd500f2f 1522 * 'above_latin1' parameter */
922e8cb4 1523#define _generic_swash_utf8(classnum, p) \
cd500f2f 1524 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1525
cc8ab7c0 1526/* Like the above, but should be used only when it is known that there are no
ff7ecfc3
KW
1527 * characters in the upper-Latin1 range (128-255 on ASCII platforms) which the
1528 * class is TRUE for. Hence it can skip the tests for this range.
1529 * 'above_latin1' should include its arguments */
3ded5eb0 1530#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1531 (UTF8_IS_INVARIANT(*(p)) \
1532 ? _generic_isCC(*(p), classnum) \
1533 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
3ded5eb0 1534 ? above_latin1 \
cc8ab7c0
KW
1535 : 0)
1536
84238efa
KW
1537/* NOTE that some of these macros have very similar ones in regcharclass.h.
1538 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1539 * there, differing in name only by an underscore from the one here
1540 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1541 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1542 * points; the regcharclass.h ones are implemented as a series of
1543 * "if-else-if-else ..." */
1544
922e8cb4
KW
1545#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1546#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1547#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1548 utf8, the non-utf8 macro works
1549 */
cd500f2f
KW
1550#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1551#define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1552#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1553 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1554#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1555#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1556 _is_utf8_perl_idcont, p)
e5dcd934 1557
c11ff943
KW
1558/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1559 * IDFIRST is an alnum. See
c8362b00 1560 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1561 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1562 * This used to be not the XID version, but we decided to go with the more
1563 * modern Unicode definition */
cd500f2f 1564#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1565 _is_utf8_perl_idstart, p)
e5dcd934 1566
922e8cb4
KW
1567#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1568#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1569
509fb054 1570/* Posix and regular space are identical above Latin1 */
cd500f2f 1571#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1572
922e8cb4 1573#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1574#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1575#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1576#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1577#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1578#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1579 is_XDIGIT_high(p))
a0ed51b3 1580
7d2b61dd 1581#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
509fb054
KW
1582#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1583#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1584#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1585
4650c663
KW
1586/* For internal core Perl use only: the base macros for defining macros like
1587 * isALPHA_LC_utf8. These are like _generic_utf8, but if the first code point
1588 * in 'p' is within the 0-255 range, it uses locale rules from the passed-in
1589 * 'macro' parameter */
94bb8c36
KW
1590#define _generic_LC_utf8(macro, p, utf8) \
1591 (UTF8_IS_INVARIANT(*(p)) \
1592 ? macro(*(p)) \
1593 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1594 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
9201a572
KW
1595 : utf8)
1596
cd500f2f
KW
1597#define _generic_LC_swash_utf8(macro, classnum, p) \
1598 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1599#define _generic_LC_func_utf8(macro, above_latin1, p) \
4650c663 1600 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1601
cd500f2f
KW
1602#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1603 _CC_ALPHANUMERIC, p)
1604#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1605#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1606#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1607#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1608#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1609#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1610#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1611#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1612#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1613#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1614#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1615 space under locale */
cd500f2f
KW
1616#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1617#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1618#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1619#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1620 _CC_WORDCHAR, p)
1621#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1622
fbc19f27
KW
1623/* Macros for backwards compatibility and for completeness when the ASCII and
1624 * Latin1 values are identical */
fbc19f27
KW
1625#define isALPHAU(c) isALPHA_L1(c)
1626#define isDIGIT_L1(c) isDIGIT_A(c)
1627#define isOCTAL(c) isOCTAL_A(c)
1628#define isOCTAL_L1(c) isOCTAL_A(c)
1629#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1630#define isALNUM(c) isWORDCHAR(c)
1631#define isALNUMU(c) isWORDCHAR_L1(c)
1632#define isALNUM_LC(c) isWORDCHAR_LC(c)
1633#define isALNUM_uni(c) isWORDCHAR_uni(c)
1634#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1635#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1636#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1637#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1638#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1639#define isALNUMC(c) isALPHANUMERIC(c)
1640#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1641#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1642#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1643#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1644#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1645
2bd1cbf6
KW
1646/* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII,
1647 * except that they don't necessarily mean the same characters, e.g. CTRL-D is
1648 * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC.
1649 * '?' is special-cased on EBCDIC to APC, which is the control there that is
1650 * the outlier from the block that contains the other controls, just like
1651 * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0
1652 * block. If it weren't special cased, it would yield a non-control.
1653 * The conversion works both ways, so CTRL('D') is 4, and CTRL(4) is D, etc. */
1654#ifndef EBCDIC
1655# define toCTRL(c) (toUPPER(c) ^ 64)
1656#else
1657# define toCTRL(c) ((c) == '?' \
1658 ? LATIN1_TO_NATIVE(0x9F) \
1659 : (c) == LATIN1_TO_NATIVE(0x9F) \
1660 ? '?' \
1661 : (NATIVE_TO_LATIN1(toUPPER(c)) ^ 64))
1662#endif
bbce6d69 1663
dea28490
JJ
1664/* Line numbers are unsigned, 32 bits. */
1665typedef U32 line_t;
e5dcd934 1666#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1667
91152fc1
DG
1668/* Helpful alias for version prescan */
1669#define is_LAX_VERSION(a,b) \
1670 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1671
1672#define is_STRICT_VERSION(a,b) \
1673 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1674
1675#define BADVERSION(a,b,c) \
1676 if (b) { \
1677 *b = c; \
1678 } \
1679 return a;
8c52afec 1680
cb27eebd
KW
1681/* Converts a character known to represent a hexadecimal digit (0-9, A-F, or
1682 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer,
1683 * which is advanced. The input is validated only by an assert() in DEBUGGING
1684 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and
1685 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */
1686#define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \
1687 ? (c) \
1688 : ((c) + 9))))
902008b8
KW
1689#define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \
1690 ? (*(s)++) \
1691 : (*(s)++ + 9))))
95a59cab 1692
cb27eebd
KW
1693/* Converts a character known to represent an octal digit (0-7) to its numeric
1694 * value. The input is validated only by an assert() in DEBUGGING builds. In
1695 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */
1696#define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c)))
1697
8e84507e 1698/*
ccfc67b7
JH
1699=head1 Memory Management
1700
a02a5408 1701=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1702The XSUB-writer's interface to the C C<malloc> function.
1703
0d7b2759
MH
1704Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1705
c5008215
JC
1706In 5.9.3, Newx() and friends replace the older New() API, and drops
1707the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1708themselves. This aid has been superseded by a new build option,
d10b4965 1709PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1710there for use in XS modules supporting older perls.
1711
a02a5408 1712=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1713The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1714cast. See also C<Newx>.
954c1994 1715
0d7b2759
MH
1716Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1717
a02a5408 1718=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1719The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1720memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1721
0d7b2759
MH
1722Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1723
954c1994
GS
1724=for apidoc Am|void|Renew|void* ptr|int nitems|type
1725The XSUB-writer's interface to the C C<realloc> function.
1726
0d7b2759
MH
1727Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1728
954c1994
GS
1729=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1730The XSUB-writer's interface to the C C<realloc> function, with
1731cast.
1732
0d7b2759
MH
1733Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1734
49b8b560 1735=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1736The XSUB-writer's interface to the C C<free> function.
1737
0d7b2759
MH
1738This should B<ONLY> be used on memory obtained using L<"Newx"> and friends.
1739
954c1994
GS
1740=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1741The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1742source, C<dest> is the destination, C<nitems> is the number of items, and
1743C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1744
e90e2364 1745=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
72d33970
FC
1746Like C<Move> but returns dest. Useful
1747for encouraging compilers to tail-call
e90e2364
NC
1748optimise.
1749
954c1994
GS
1750=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1751The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1752source, C<dest> is the destination, C<nitems> is the number of items, and
1753C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1754
e90e2364
NC
1755=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1756
72d33970
FC
1757Like C<Copy> but returns dest. Useful
1758for encouraging compilers to tail-call
e90e2364
NC
1759optimise.
1760
954c1994
GS
1761=for apidoc Am|void|Zero|void* dest|int nitems|type
1762
1763The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1764destination, C<nitems> is the number of items, and C<type> is the type.
1765
e90e2364
NC
1766=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1767
72d33970
FC
1768Like C<Zero> but returns dest. Useful
1769for encouraging compilers to tail-call
e90e2364
NC
1770optimise.
1771
da5d8dbb 1772=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1773This is an architecture-independent macro to copy one structure to another.
954c1994 1774
7e337ee0
JH
1775=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1776
1777Fill up memory with a byte pattern (a byte repeated over and over
1778again) that hopefully catches attempts to access uninitialized memory.
1779
1780=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1781
1782PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1783
1c12ffb4 1784=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1785
1786PoisonWith(0xEF) for catching access to freed memory.
1787
9965345d
JH
1788=for apidoc Am|void|Poison|void* dest|int nitems|type
1789
7e337ee0 1790PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1791
1792=cut */
954c1994 1793
561b68a9
SH
1794/* Maintained for backwards-compatibility only. Use newSV() instead. */
1795#ifndef PERL_CORE
ff06c60c 1796#define NEWSV(x,len) newSV(len)
561b68a9 1797#endif
ff06c60c 1798
19a94d75
JH
1799#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1800
2720f793
JH
1801/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1802 * overly eager compilers that will bleat about e.g.
1803 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1804#ifdef PERL_MALLOC_WRAP
f019c49e 1805#define MEM_WRAP_CHECK(n,t) \
5d9574c1 1806 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (croak_memory_wrap(),0))
27d5b266 1807#define MEM_WRAP_CHECK_1(n,t,a) \
5d9574c1 1808 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1809#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1810
d1decf2b 1811#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
1812#else
1813
410319be
NC
1814#define MEM_WRAP_CHECK(n,t)
1815#define MEM_WRAP_CHECK_1(n,t,a)
1816#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1817#define MEM_WRAP_CHECK_(n,t)
1818
6fff79ce 1819#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1820
1936d2a7 1821#endif
8b44ba4c 1822
fe4f188c 1823#ifdef PERL_MEM_LOG
46c6c7e2 1824/*
9f653bb5 1825 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1826 * go through functions, which are handy for debugging breakpoints, but
1827 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1828 * line number, and C function name if available) passed in. This info can
1829 * then be used for logging the calls, for which one gets a sample
73d1d973 1830 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1831 *
46c6c7e2 1832 * Known problems:
94e892a6 1833 * - not all memory allocs get logged, only those
46c6c7e2 1834 * that go through Newx() and derivatives (while all
94e892a6 1835 * Safefrees do get logged)
46c6c7e2
JH
1836 * - __FILE__ and __LINE__ do not work everywhere
1837 * - __func__ or __FUNCTION__ even less so
1838 * - I think more goes on after the perlio frees but
1839 * the thing is that STDERR gets closed (as do all
1840 * the file descriptors)
1841 * - no deeper calling stack than the caller of the Newx()
1842 * or the kind, but do I look like a C reflection/introspection
1843 * utility to you?
1844 * - the function prototypes for the logging functions
1845 * probably should maybe be somewhere else than handy.h
1846 * - one could consider inlining (macrofying) the logging
1847 * for speed, but I am too lazy
1848 * - one could imagine recording the allocations in a hash,
1849 * (keyed by the allocation address?), and maintain that
1850 * through reallocs and frees, but how to do that without
1851 * any News() happening...?
73d1d973 1852 * - lots of -Ddefines to get useful/controllable output
b953482e 1853 * - lots of ENV reads
46c6c7e2
JH
1854 */
1855
12754f92 1856PERL_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 1857
12754f92 1858PERL_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 1859
12754f92 1860PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1861
0b0ab801 1862# ifdef PERL_CORE
73d1d973 1863# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1864enum mem_log_type {
1865 MLT_ALLOC,
1866 MLT_REALLOC,
d7a2c63c
MHM
1867 MLT_FREE,
1868 MLT_NEW_SV,
1869 MLT_DEL_SV
0b0ab801
MHM
1870};
1871# endif
12754f92 1872# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1873void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1874void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1875# endif
0b0ab801
MHM
1876# endif
1877
fe4f188c
JH
1878#endif
1879
1880#ifdef PERL_MEM_LOG
d1401ee9
MHM
1881#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1882#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1883#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1884#endif
1885
1886#ifndef MEM_LOG_ALLOC
1887#define MEM_LOG_ALLOC(n,t,a) (a)
1888#endif
1889#ifndef MEM_LOG_REALLOC
1890#define MEM_LOG_REALLOC(n,t,v,a) (a)
1891#endif
1892#ifndef MEM_LOG_FREE
1893#define MEM_LOG_FREE(a) (a)
1894#endif
1895
d1401ee9
MHM
1896#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1897#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1898#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1899
1900#ifndef PERL_CORE
a02a5408
JC
1901/* pre 5.9.x compatibility */
1902#define New(x,v,n,t) Newx(v,n,t)
1903#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1904#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1905#endif
a02a5408 1906
ff68c719 1907#define Renew(v,n,t) \
d1401ee9 1908 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1909#define Renewc(v,n,t,c) \
d1401ee9 1910 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1911
1912#ifdef PERL_POISON
1913#define Safefree(d) \
06c0cc96 1914 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1915#else
fe4f188c 1916#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1917#endif
55497cff 1918
8b44ba4c
NC
1919#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1920#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1921#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1922
8b44ba4c
NC
1923#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1924#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1925#ifdef HAS_MEMSET
8b44ba4c 1926#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1927#else
8b44ba4c
NC
1928/* Using bzero(), which returns void. */
1929#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1930#endif
1931
7e337ee0
JH
1932#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1933#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1934#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1935#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1936
caa674f3
DD
1937#ifdef PERL_POISON
1938# define PERL_POISON_EXPR(x) x
1939#else
1940# define PERL_POISON_EXPR(x)
1941#endif
1942
2304df62 1943#ifdef USE_STRUCT_COPY
ff68c719 1944#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1945#else
1946#define StructCopy(s,d,t) Copy(s,d,1,t)
1947#endif
2cc61e15 1948
c3caa5c3
JH
1949/* C_ARRAY_LENGTH is the number of elements in the C array (so you
1950 * want your zero-based indices to be less than but not equal to).
1951 *
1952 * C_ARRAY_END is one past the last: half-open/half-closed range,
1953 * not last-inclusive range. */
622913ab 1954#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
c3caa5c3 1955#define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a))
622913ab 1956
2cc61e15
DD
1957#ifdef NEED_VA_COPY
1958# ifdef va_copy
1959# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1960# else
a1866d1b
JH
1961# if defined(__va_copy)
1962# define Perl_va_copy(s, d) __va_copy(d, s)
1963# else
1964# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1965# endif
2cc61e15
DD
1966# endif
1967#endif
1968
472d47bc
SB
1969/* convenience debug macros */
1970#ifdef USE_ITHREADS
1971#define pTHX_FORMAT "Perl interpreter: 0x%p"
1972#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1973#define pTHX_VALUE_ (void *)my_perl,
1974#define pTHX_VALUE (void *)my_perl
1975#define pTHX__VALUE_ ,(void *)my_perl,
1976#define pTHX__VALUE ,(void *)my_perl
472d47bc 1977#else
3609ea0d 1978#define pTHX_FORMAT
472d47bc 1979#define pTHX__FORMAT
3609ea0d 1980#define pTHX_VALUE_
472d47bc 1981#define pTHX_VALUE
3609ea0d 1982#define pTHX__VALUE_
472d47bc
SB
1983#define pTHX__VALUE
1984#endif /* USE_ITHREADS */
3609ea0d 1985
2acdbac1
NC
1986/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1987 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1988 CPAN::Unpack show any users outside the core. */
1989#ifdef PERL_CORE
d1d15184 1990# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1991#endif
1992
dfff4baf
BF
1993/* Internal macros to deal with gids and uids */
1994#ifdef PERL_CORE
1995
1996# if Uid_t_size > IVSIZE
1997# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
1998# define SvUID(sv) SvNV(sv)
1999# else
2000# if Uid_t_sign <= 0
2001# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
2002# define SvUID(sv) SvIV(sv)
2003# else
2004# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
2005# define SvUID(sv) SvUV(sv)
2006# endif
2007# endif /* Uid_t_size */
2008
2009# if Gid_t_size > IVSIZE
2010# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
2011# define SvGID(sv) SvNV(sv)
2012# else
2013# if Gid_t_sign <= 0
2014# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
2015# define SvGID(sv) SvIV(sv)
2016# else
2017# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
2018# define SvGID(sv) SvUV(sv)
2019# endif
2020# endif /* Gid_t_size */
2021
2022#endif
2023
9d745869
KW
2024#endif /* HANDY_H */
2025
e9a8c099
MHM
2026/*
2027 * Local variables:
2028 * c-indentation-style: bsd
2029 * c-basic-offset: 4
14d04a33 2030 * indent-tabs-mode: nil
e9a8c099
MHM
2031 * End:
2032 *
14d04a33 2033 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 2034 */