This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
regcomp.c: Update comment to new reality
[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
30a15352
KW
307Like C<newSVpvn>, but takes a literal C<NUL>-terminated string instead of a
308string/length pair.
3fe05580 309
84bafc02 310=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
30a15352
KW
311Like C<newSVpvn_flags>, but takes a literal C<NUL>-terminated string instead of
312a string/length pair.
84bafc02 313
3fe05580 314=for apidoc Ama|SV*|newSVpvs_share|const char* s
30a15352
KW
315Like C<newSVpvn_share>, but takes a literal C<NUL>-terminated string instead of
316a string/length pair and omits the hash parameter.
3fe05580 317
9dcc53ea 318=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
30a15352
KW
319Like C<sv_catpvn_flags>, but takes a literal C<NUL>-terminated string instead
320of a string/length pair.
9dcc53ea
Z
321
322=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
323Like C<sv_catpvn_nomg>, but takes a literal string instead of a
324string/length pair.
325
7b4e7469 326=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
327Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
328
9dcc53ea
Z
329=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
330Like C<sv_catpvn_mg>, but takes a literal string instead of a
331string/length pair.
332
7b4e7469 333=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
334Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
335
9dcc53ea
Z
336=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
337Like C<sv_setpvn_mg>, but takes a literal string instead of a
338string/length pair.
339
340=for apidoc Am|SV *|sv_setref_pvs|const char* s
341Like C<sv_setref_pvn>, but takes a literal string instead of a
342string/length pair.
343
3fe05580
MHM
344=head1 Memory Management
345
346=for apidoc Ama|char*|savepvs|const char* s
30a15352
KW
347Like C<savepvn>, but takes a literal C<NUL>-terminated string instead of a
348string/length pair.
3fe05580 349
9dcc53ea
Z
350=for apidoc Ama|char*|savesharedpvs|const char* s
351A version of C<savepvs()> which allocates the duplicate string in memory
352which is shared between threads.
353
3fe05580
MHM
354=head1 GV Functions
355
356=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
357Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
358
359=head1 Hash Manipulation Functions
360
361=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
362Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
363
364=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
365Like C<hv_store>, but takes a literal string instead of a string/length pair
366and omits the hash parameter.
367
510966aa
Z
368=head1 Lexer interface
369
370=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
371
372Like L</lex_stuff_pvn>, but takes a literal string instead of a
373string/length pair.
374
3fe05580
MHM
375=cut
376*/
377
926bb54c
KW
378/* concatenating with "" ensures that only literal strings are accepted as
379 * argument */
26bfd924 380#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 381
926bb54c
KW
382/* note that STR_WITH_LEN() can't be used as argument to macros or functions
383 * that under some configurations might be macros, which means that it requires
384 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
385 */
386
387/* STR_WITH_LEN() shortcuts */
388#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
389#define newSVpvs_flags(str,flags) \
390 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 391#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
392#define sv_catpvs_flags(sv, str, flags) \
393 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
394#define sv_catpvs_nomg(sv, str) \
395 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
396#define sv_catpvs(sv, str) \
397 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
398#define sv_catpvs_mg(sv, str) \
399 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 400#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
401#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
402#define sv_setref_pvs(rv, classname, str) \
403 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 404#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
405#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
406#define gv_stashpvs(str, create) \
407 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
408#define gv_fetchpvs(namebeg, add, sv_type) \
409 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
410#define gv_fetchpvn(namebeg, len, add, sv_type) \
411 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
412#define sv_catxmlpvs(dsv, str, utf8) \
413 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
414#define hv_fetchs(hv,key,lval) \
415 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
416 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
417 : HV_FETCH_JUST_SV, NULL, 0))
418
419#define hv_stores(hv,key,val) \
420 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
421 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 422
510966aa
Z
423#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
424
b96d8cd9
NC
425#define get_cvs(str, flags) \
426 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 427
954c1994 428/*
ccfc67b7
JH
429=head1 Miscellaneous Functions
430
954c1994
GS
431=for apidoc Am|bool|strNE|char* s1|char* s2
432Test two strings to see if they are different. Returns true or
433false.
434
435=for apidoc Am|bool|strEQ|char* s1|char* s2
436Test two strings to see if they are equal. Returns true or false.
437
438=for apidoc Am|bool|strLT|char* s1|char* s2
439Test two strings to see if the first, C<s1>, is less than the second,
440C<s2>. Returns true or false.
441
442=for apidoc Am|bool|strLE|char* s1|char* s2
443Test two strings to see if the first, C<s1>, is less than or equal to the
444second, C<s2>. Returns true or false.
445
446=for apidoc Am|bool|strGT|char* s1|char* s2
447Test two strings to see if the first, C<s1>, is greater than the second,
448C<s2>. Returns true or false.
449
450=for apidoc Am|bool|strGE|char* s1|char* s2
451Test two strings to see if the first, C<s1>, is greater than or equal to
452the second, C<s2>. Returns true or false.
453
454=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
455Test two strings to see if they are different. The C<len> parameter
72d33970 456indicates the number of bytes to compare. Returns true or false. (A
954c1994
GS
457wrapper for C<strncmp>).
458
459=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
460Test two strings to see if they are equal. The C<len> parameter indicates
72d33970 461the number of bytes to compare. Returns true or false. (A wrapper for
954c1994
GS
462C<strncmp>).
463
464=cut
465*/
466
8d063cd8
LW
467#define strNE(s1,s2) (strcmp(s1,s2))
468#define strEQ(s1,s2) (!strcmp(s1,s2))
469#define strLT(s1,s2) (strcmp(s1,s2) < 0)
470#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
471#define strGT(s1,s2) (strcmp(s1,s2) > 0)
472#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
473#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
474#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 475
36477c24
PP
476#ifdef HAS_MEMCMP
477# define memNE(s1,s2,l) (memcmp(s1,s2,l))
478# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
479#else
480# define memNE(s1,s2,l) (bcmp(s1,s2,l))
481# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
482#endif
483
568a785a 484#define memEQs(s1, l, s2) \
8e6497b6 485 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
486#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
487
bbce6d69
PP
488/*
489 * Character classes.
490 *
491 * Unfortunately, the introduction of locales means that we
492 * can't trust isupper(), etc. to tell the truth. And when
493 * it comes to /\w+/ with tainting enabled, we *must* be able
494 * to trust our character classes.
495 *
496 * Therefore, the default tests in the text of Perl will be
497 * independent of locale. Any code that wants to depend on
498 * the current locale will use the tests that begin with "lc".
499 */
500
2304df62
AD
501#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
502# ifndef CTYPE256
503# define CTYPE256
504# endif
505#endif
506
954c1994 507/*
ccfc67b7
JH
508
509=head1 Character classes
243effed
KW
510This section is about functions (really macros) that classify characters
511into types, such as punctuation versus alphabetic, etc. Most of these are
512analogous to regular expression character classes. (See
513L<perlrecharclass/POSIX Character Classes>.) There are several variants for
514each class. (Not all macros have all variants; each item below lists the
515ones valid for it.) None are affected by C<use bytes>, and only the ones
516with C<LC> in the name are affected by the current locale.
517
518The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
519C<U8>) as input and returns a boolean as to whether or not the character
6aff1f14
KW
520represented by that octet is (or on non-ASCII platforms, corresponds to) an
521ASCII character in the named class based on platform, Unicode, and Perl rules.
522If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 523
6aff1f14
KW
524Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
525with no suffix C<"_A">.
4b9734bf
KW
526
527Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
528onto the platform. That is, the code points that are ASCII are unaffected,
529since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
530as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
531true when called with the code point 0xDF, which is a word character in both
4650c663 532ASCII and EBCDIC (though it represents different characters in each).
243effed
KW
533
534Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
535point as input. If the code point is larger than 255, Unicode rules are used
536to determine if it is in the character class. For example,
6aff1f14
KW
537C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
538WITH MACRON in Unicode, and is a word character.
243effed
KW
539
540Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
541(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
542classification of just the first (possibly multi-byte) character in the string
543is tested.
243effed 544
31f05a37
KW
545Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but the
546result is based on the current locale, which is what C<LC> in the name stands
547for. If Perl can determine that the current locale is a UTF-8 locale, it uses
548the published Unicode rules; otherwise, it uses the C library function that
549gives the named classification. For example, C<isDIGIT_LC()> when not in a
550UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always
551returned if the input won't fit into an octet.
243effed
KW
552
553Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
554returns the same as C<isFOO_LC> for input code points less than 256, and
555returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
556
557Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
558(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
559classification of just the first (possibly multi-byte) character in the string
560is tested.
ccfc67b7 561
954c1994 562=for apidoc Am|bool|isALPHA|char ch
2787a470 563Returns a boolean indicating whether the specified character is an
6aff1f14 564alphabetic character, analogous to C<m/[[:alpha:]]/>.
8a58bdcf 565See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 566C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 567C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 568
15861f94
KW
569=for apidoc Am|bool|isALPHANUMERIC|char ch
570Returns a boolean indicating whether the specified character is a either an
6aff1f14 571alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
15861f94
KW
572See the L<top of this section|/Character classes> for an explanation of variants
573C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 574C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
575C<isALPHANUMERIC_LC_utf8>.
576
8a58bdcf
KW
577=for apidoc Am|bool|isASCII|char ch
578Returns a boolean indicating whether the specified character is one of the 128
243effed 579characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 580On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
581character corresponds to an ASCII character. Variants C<isASCII_A()> and
582C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
583See the L<top of this section|/Character classes> for an explanation of variants
584C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
585C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
586library routine C<isascii()>. In these cases, the variants whose names contain
587C<LC> are the same as the corresponding ones without.
588
d98532ea
KW
589Also note, that because all ASCII characters are UTF-8 invariant (meaning they
590have the exact same representation (always a single byte) whether encoded in
591UTF-8 or not), C<isASCII> will give the correct results when called with any
592byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
593work properly on any string encoded or not in UTF-8.
594
243effed
KW
595=for apidoc Am|bool|isBLANK|char ch
596Returns a boolean indicating whether the specified character is a
6aff1f14 597character considered to be a blank, analogous to C<m/[[:blank:]]/>.
243effed 598See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 599C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
600C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
601platforms do not have the C library routine C<isblank()>. In these cases, the
602variants whose names contain C<LC> are the same as the corresponding ones
603without.
604
605=for apidoc Am|bool|isCNTRL|char ch
606Returns a boolean indicating whether the specified character is a
6aff1f14 607control character, analogous to C<m/[[:cntrl:]]/>.
243effed 608See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 609C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
610C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
611On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 612
954c1994 613=for apidoc Am|bool|isDIGIT|char ch
2787a470 614Returns a boolean indicating whether the specified character is a
6aff1f14 615digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 616Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed 617See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 618C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
619C<isDIGIT_LC_utf8>.
620
621=for apidoc Am|bool|isGRAPH|char ch
622Returns a boolean indicating whether the specified character is a
6aff1f14 623graphic character, analogous to C<m/[[:graph:]]/>.
243effed 624See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 625C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 626C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 627
0c82b6df 628=for apidoc Am|bool|isLOWER|char ch
2787a470 629Returns a boolean indicating whether the specified character is a
6aff1f14 630lowercase character, analogous to C<m/[[:lower:]]/>.
8a58bdcf 631See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 632C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 633C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 634
c99e91e9 635=for apidoc Am|bool|isOCTAL|char ch
2787a470 636Returns a boolean indicating whether the specified character is an
6aff1f14 637octal digit, [0-7].
243effed
KW
638The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
639C<isOCTAL>.
640
641=for apidoc Am|bool|isPUNCT|char ch
642Returns a boolean indicating whether the specified character is a
6aff1f14
KW
643punctuation character, analogous to C<m/[[:punct:]]/>.
644Note that the definition of what is punctuation isn't as
243effed
KW
645straightforward as one might desire. See L<perlrecharclass/POSIX Character
646Classes> for details.
647See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 648C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 649C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 650
0c82b6df 651=for apidoc Am|bool|isSPACE|char ch
2787a470 652Returns a boolean indicating whether the specified character is a
6aff1f14 653whitespace character. This is analogous
398d098a
KW
654to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
655(experimentally), this also matches what C<m/[[:space:]]/> does.
a04e6aad 656("Experimentally" means that this change may be backed out in 5.22 if
398d098a
KW
657field experience indicates that it was unwise.) Prior to 5.18, only the
658locale forms of this macro (the ones with C<LC> in their names) matched
659precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
660in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
661(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
8a58bdcf 662See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 663C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 664C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 665
398d098a
KW
666=for apidoc Am|bool|isPSXSPC|char ch
667(short for Posix Space)
668Starting in 5.18, this is identical (experimentally) in all its forms to the
669corresponding C<isSPACE()> macros. ("Experimentally" means that this change
a04e6aad 670may be backed out in 5.22 if field experience indicates that it
398d098a
KW
671was unwise.)
672The locale forms of this macro are identical to their corresponding
673C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
674non-locale forms differ from their C<isSPACE()> forms only in that the
675C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
676Otherwise they are identical. Thus this macro is analogous to what
677C<m/[[:space:]]/> matches in a regular expression.
678See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 679C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
680C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
681
954c1994 682=for apidoc Am|bool|isUPPER|char ch
2787a470 683Returns a boolean indicating whether the specified character is an
6aff1f14 684uppercase character, analogous to C<m/[[:upper:]]/>.
8a58bdcf 685See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 686C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 687C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 688
243effed 689=for apidoc Am|bool|isPRINT|char ch
8eea39dd 690Returns a boolean indicating whether the specified character is a
6aff1f14 691printable character, analogous to C<m/[[:print:]]/>.
8a58bdcf 692See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 693C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
694C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
695
696=for apidoc Am|bool|isWORDCHAR|char ch
697Returns a boolean indicating whether the specified character is a character
698that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
699in a regular expression. A word character is an alphabetic character, a
700decimal digit, a connecting punctuation character (such as an underscore), or
701a "mark" character that attaches to one of those (like some sort of accent).
702C<isALNUM()> is a synonym provided for backward compatibility, even though a
703word character includes more than the standard C language meaning of
704alphanumeric.
705See the L<top of this section|/Character classes> for an explanation of variants
706C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
707C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
708
709=for apidoc Am|bool|isXDIGIT|char ch
710Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
711digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
712and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
713See the L<top of this section|/Character classes> for an explanation of variants
714C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
715C<isXDIGIT_LC_utf8>.
716
3c3ecf18
KW
717=for apidoc Am|bool|isIDFIRST|char ch
718Returns a boolean indicating whether the specified character can be the first
719character of an identifier. This is very close to, but not quite the same as
720the official Unicode property C<XID_Start>. The difference is that this
721returns true only if the input character also matches L</isWORDCHAR>.
722See the L<top of this section|/Character classes> for an explanation of variants
723C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
724C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
725
726=for apidoc Am|bool|isIDCONT|char ch
727Returns a boolean indicating whether the specified character can be the
728second or succeeding character of an identifier. This is very close to, but
729not quite the same as the official Unicode property C<XID_Continue>. The
730difference is that this returns true only if the input character also matches
731L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
732explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
733C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
734C<isIDCONT_LC_utf8>.
735
243effed 736=head1 Miscellaneous Functions
8eea39dd 737
95a59cab 738=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 739Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
740Behaviour is only well defined when isXDIGIT(*str) is true.
741
e7c1e6c1
KW
742=head1 Character case changing
743
1f607577
KW
744=for apidoc Am|U8|toUPPER|U8 ch
745Converts the specified character to uppercase. If the input is anything but an
746ASCII lowercase character, that input character itself is returned. Variant
c753c8d3 747C<toUPPER_A> is equivalent.
954c1994 748
1f607577
KW
749=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
750Converts the Unicode code point C<cp> to its uppercase version, and
751stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
752that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
753bytes since the uppercase version may be longer than the original character.
754
755The first code point of the uppercased version is returned
756(but note, as explained just above, that there may be more.)
757
758=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
759Converts the UTF-8 encoded character at C<p> to its uppercase version, and
760stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
761that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
762bytes since the uppercase version may be longer than the original character.
763
764The first code point of the uppercased version is returned
765(but note, as explained just above, that there may be more.)
766
767The input character at C<p> is assumed to be well-formed.
768
25200305
KW
769=for apidoc Am|U8|toFOLD|U8 ch
770Converts the specified character to foldcase. If the input is anything but an
771ASCII uppercase character, that input character itself is returned. Variant
772C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
773Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
774
1f607577
KW
775=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
776Converts the Unicode code point C<cp> to its foldcase version, and
777stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
778that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
779bytes since the foldcase version may be longer than the original character.
780
781The first code point of the foldcased version is returned
782(but note, as explained just above, that there may be more.)
783
784=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
785Converts the UTF-8 encoded character at C<p> to its foldcase version, and
786stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
787that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
788bytes since the foldcase version may be longer than the original character.
789
790The first code point of the foldcased version is returned
791(but note, as explained just above, that there may be more.)
792
793The input character at C<p> is assumed to be well-formed.
794
795=for apidoc Am|U8|toLOWER|U8 ch
796Converts the specified character to lowercase. If the input is anything but an
797ASCII uppercase character, that input character itself is returned. Variant
c753c8d3 798C<toLOWER_A> is equivalent.
954c1994 799
1f607577
KW
800=for apidoc Am|U8|toLOWER_L1|U8 ch
801Converts the specified Latin1 character to lowercase. The results are undefined if
802the input doesn't fit in a byte.
803
804=for apidoc Am|U8|toLOWER_LC|U8 ch
805Converts the specified character to lowercase using the current locale's rules,
806if possible; otherwise returns the input character itself.
807
808=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
809Converts the Unicode code point C<cp> to its lowercase version, and
810stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
811that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
812bytes since the lowercase version may be longer than the original character.
813
814The first code point of the lowercased version is returned
815(but note, as explained just above, that there may be more.)
816
817=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
818Converts the UTF-8 encoded character at C<p> to its lowercase version, and
819stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
820that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
821bytes since the lowercase version may be longer than the original character.
822
823The first code point of the lowercased version is returned
824(but note, as explained just above, that there may be more.)
825
826The input character at C<p> is assumed to be well-formed.
827
828=for apidoc Am|U8|toLOWER_LC|U8 ch
829Converts the specified character to lowercase using the current locale's rules,
830if possible; otherwise returns the input character itself.
831
25200305
KW
832=for apidoc Am|U8|toTITLE|U8 ch
833Converts the specified character to titlecase. If the input is anything but an
834ASCII lowercase character, that input character itself is returned. Variant
835C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
836as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
837concept used in locale handling, so there is no functionality for that.)
838
1f607577
KW
839=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
840Converts the Unicode code point C<cp> to its titlecase version, and
841stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
842that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
843bytes since the titlecase version may be longer than the original character.
844
845The first code point of the titlecased version is returned
846(but note, as explained just above, that there may be more.)
847
848=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
849Converts the UTF-8 encoded character at C<p> to its titlecase version, and
850stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
851that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
852bytes since the titlecase version may be longer than the original character.
853
854The first code point of the titlecased version is returned
855(but note, as explained just above, that there may be more.)
856
857The input character at C<p> is assumed to be well-formed.
858
954c1994 859=cut
353c9b6f 860
1f607577 861XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
25200305 862really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
243effed 863
8a58bdcf 864Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
865patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
866
954c1994
GS
867*/
868
7c062697
KW
869/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
870 * is known only in the perl core, and this macro can be called from outside
871 * that */
872#ifdef HAS_QUAD
873# define WIDEST_UTYPE U64TYPE
874#else
875# define WIDEST_UTYPE U32
876#endif
877
3912bc88
KW
878/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
879 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
880 * bits of information are lost even on a 64-bit machine, but to get the
881 * compiler to optimize it out if possible. This is because Configure makes
882 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
883 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
884 */
885#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 886
41f43cc2 887#ifdef EBCDIC
b6340bd0
KW
888# ifndef _ALL_SOURCE
889 /* This returns the wrong results on at least z/OS unless this is
890 * defined. */
891# error _ALL_SOURCE should probably be defined
892# endif
893
894 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
895 * likely not defined, and so we use the native function */
3cedda69 896# define isASCII(c) cBOOL(isascii(c))
41f43cc2
KW
897#else
898# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
899#endif
900
ec79cfbb 901#define isASCII_A(c) isASCII(c)
3509bb53 902#define isASCII_L1(c) isASCII(c)
6515cb76 903
c2da0b36
KW
904/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
905 * and the 8 possible permutations of those bits exactly comprise the 8 octal
906 * digits */
3d33b3e3 907#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 908
ec79cfbb 909/* ASCII range only */
9fb1bf9d 910#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 911
a500dc72
KW
912/* Character class numbers. For internal core Perl use only. The ones less
913 * than 32 are used in PL_charclass[] and the ones up through the one that
914 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
915 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
916 * their actual definitions are here. If that file has a name not used here,
917 * it won't compile.
1709d539
KW
918 *
919 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
920 * frequency of their use. This gives a slight edge to exiting a loop earlier
921 * (in reginclass() in regexec.c) */
1709d539
KW
922# define _CC_WORDCHAR 0 /* \w and [:word:] */
923# define _CC_DIGIT 1 /* \d and [:digit:] */
924# define _CC_ALPHA 2 /* [:alpha:] */
925# define _CC_LOWER 3 /* [:lower:] */
926# define _CC_UPPER 4 /* [:upper:] */
927# define _CC_PUNCT 5 /* [:punct:] */
928# define _CC_PRINT 6 /* [:print:] */
15861f94 929# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 930# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 931# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 932
b0d691b2 933#define _FIRST_NON_SWASH_CC 10
1709d539
KW
934/* The character classes above are implemented with swashes. The second group
935 * (just below) contains the ones implemented without. These are also sorted
936 * in rough order of the frequency of their use, except that \v should be last,
937 * as it isn't a real Posix character class, and some (small) inefficiencies in
938 * regular expression handling would be introduced by putting it in the middle
939 * of those that are. Also, cntrl and ascii come after the others as it may be
940 * useful to group these which have no members that match above Latin1, (or
941 * above ASCII in the latter case) */
942
b0d691b2
KW
943# define _CC_SPACE 10 /* \s */
944# define _CC_BLANK 11 /* [:blank:] */
945# define _CC_XDIGIT 12 /* [:xdigit:] */
946# define _CC_PSXSPC 13 /* [:space:] */
947# define _CC_CNTRL 14 /* [:cntrl:] */
948# define _CC_ASCII 15 /* [:ascii:] */
949# define _CC_VERTSPACE 16 /* \v */
1709d539 950
a0947d7b
KW
951# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
952
1709d539 953/* The members of the third group below do not need to be coordinated with data
3ffc8c70 954 * structures in regcomp.[ch] and regexec.c. */
b0d691b2
KW
955# define _CC_IDFIRST 17
956# define _CC_CHARNAME_CONT 18
957# define _CC_NONLATIN1_FOLD 19
958# define _CC_QUOTEMETA 20
959# define _CC_NON_FINAL_FOLD 21
960# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
961# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
962/* Unused: 23-30
f4cdb42c
KW
963 * If more bits are needed, one could add a second word for non-64bit
964 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
965 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
966 * is used only for optimization (as of this writing), and differs in the
967 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
968 * characters: the masculine and feminine ordinal indicators, so removing it
969 * would just cause /i regexes which match them to run less efficiently */
96ac0975 970
3a371f2f
KW
971#if defined(PERL_CORE) || defined(PERL_EXT)
972/* An enum version of the character class numbers, to help compilers
973 * optimize */
974typedef enum {
3a371f2f 975 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
976 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
977 _CC_ENUM_ASCII = _CC_ASCII,
978 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 979 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 980 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
981 _CC_ENUM_DIGIT = _CC_DIGIT,
982 _CC_ENUM_GRAPH = _CC_GRAPH,
983 _CC_ENUM_LOWER = _CC_LOWER,
984 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 985 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 986 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 987 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 988 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 989 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 990 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 991 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
992} _char_class_number;
993#endif
994
63c61c3f 995#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 996#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 997
3018b823 998#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 999# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
1000 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
1001 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
1002 #error Need to adjust order of swash_property_names[]
1003# endif
1004
b44daeeb
KW
1005/* This is declared static in each of the few files that this is #defined for
1006 * to keep them from being publicly accessible. Hence there is a small amount
1007 * of wasted space */
1008
922e8cb4
KW
1009static const char* const swash_property_names[] = {
1010 "XPosixWord",
1011 "XPosixDigit",
1012 "XPosixAlpha",
1013 "XPosixLower",
1014 "XPosixUpper",
1015 "XPosixPunct",
1016 "XPosixPrint",
1017 "XPosixAlnum",
b0d691b2
KW
1018 "XPosixGraph",
1019 "Cased"
922e8cb4
KW
1020};
1021#endif
1022
96ac0975
NC
1023# ifdef DOINIT
1024EXTCONST U32 PL_charclass[] = {
1025# include "l1_char_class_tab.h"
1026};
1027
1028# else /* ! DOINIT */
1029EXTCONST U32 PL_charclass[];
1030# endif
1031
265c1f46 1032 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70 1033# define _CC_mask(classnum) (1U << (classnum))
4650c663
KW
1034
1035 /* For internal core Perl use only: the base macro for defining macros like
1036 * isALPHA */
ff7ecfc3 1037# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
f4cd282c 1038 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
4eeeb416 1039
f4cdb42c
KW
1040 /* The mask for the _A versions of the macros; it just adds in the bit for
1041 * ASCII. */
1042# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1043
4650c663
KW
1044 /* For internal core Perl use only: the base macro for defining macros like
1045 * isALPHA_A. The foo_A version makes sure that both the desired bit and
1046 * the ASCII bit are present */
3ded5eb0 1047# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
f4cd282c 1048 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
f4cdb42c
KW
1049 == _CC_mask_A(classnum)))
1050
f4cdb42c 1051# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 1052# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
1053# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1054# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1055# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
1056# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1057# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1058# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1059# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1060# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1061# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1062# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1063# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1064# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
d95f8b6a 1065# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
3ded5eb0
KW
1066# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1067# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1068# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1069
1070 /* continuation character for legal NAME in \N{NAME} */
1071# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1072
1073# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1074# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1075# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1076# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1077# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1078# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1079# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1080# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1081# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1082# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
f4cdb42c 1083
fb9bfbf7
KW
1084 /* Either participates in a fold with a character above 255, or is a
1085 * multi-char fold */
f4cd282c 1086# 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 1087
4eeeb416 1088# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
1089# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1090 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
1091# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1092 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
3ffc8c70 1093#else /* else we don't have perl.h */
3ded5eb0
KW
1094
1095 /* If we don't have perl.h, we are compiling a utility program. Below we
1096 * hard-code various macro definitions that wouldn't otherwise be available
3ffc8c70 1097 * to it. */
9fb1bf9d 1098# ifdef EBCDIC
3ded5eb0 1099 /* Use the native functions. They likely will return false for all
4f48cecf 1100 * non-ASCII values, but this makes sure */
b6340bd0
KW
1101# define isLOWER_A(c) (isASCII(c) && islower(c))
1102# define isPRINT_A(c) (isASCII(c) && isprint(c))
b6340bd0 1103# define isUPPER_A(c) (isASCII(c) && isupper(c))
3ded5eb0 1104# else /* ASCII platform. These are coded based on first principals */
9fb1bf9d
KW
1105# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1106# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
2181da29 1107# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
3ded5eb0 1108# endif /* Below are common definitions for ASCII and non-ASCII */
a4d7a999
KW
1109# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
1110# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
3ded5eb0 1111# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
a4d7a999
KW
1112# define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
1113# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
1114# define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
3ded5eb0 1115# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
a4d7a999
KW
1116# define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1117# define isSPACE_A(c) ((c) == ' ' \
1118 || (c) == '\t' \
1119 || (c) == '\n' \
1120 || (c) == '\r' \
1121 || (c) == '\v' \
1122 || (c) == '\f')
3ded5eb0 1123# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
a4d7a999
KW
1124# define isXDIGIT_A(c) (isDIGIT_A(c) \
1125 || ((c) >= 'a' && (c) <= 'f') \
1126 || ((c) <= 'F' && (c) >= 'A'))
3ded5eb0 1127
3ffc8c70
KW
1128 /* The _L1 macros may be unnecessary for the utilities; I (khw) added them
1129 * during debugging, and it seems best to keep them. */
3ded5eb0
KW
1130# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1131# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1132# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1133# define isBLANK_L1(c) (isBLANK_A(c) \
1134 || (FITS_IN_8_BITS(c) \
1135 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1136# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1137# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1138# define isLOWER_L1(c) (isLOWER_A(c) \
1139 || (FITS_IN_8_BITS(c) \
1140 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1141 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1142 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1143 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1144 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1145# define isPRINT_L1(c) (isPRINT_A(c) \
1146 || (FITS_IN_8_BITS(c) \
1147 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1148# define isPSXSPC_L1(c) isSPACE_L1(c)
1149# define isPUNCT_L1(c) (isPUNCT_A(c) \
1150 || (FITS_IN_8_BITS(c) \
1151 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1152 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1153 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1154 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1155 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1156 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1157 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1158# define isSPACE_L1(c) (isSPACE_A(c) \
1159 || (FITS_IN_8_BITS(c) \
1160 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1161 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1162# define isUPPER_L1(c) (isUPPER_A(c) \
1163 || (FITS_IN_8_BITS(c) \
1164 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1165 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1166 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1167# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1168# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1169# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1170 || isBLANK_L1(c) \
1171 || (c) == '-' \
1172 || (c) == '(' \
1173 || (c) == ')')
1174 /* The following are not fully accurate in the above-ASCII range. I (khw)
1175 * don't think it's necessary to be so for the purposes where this gets
1176 * compiled */
1177# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1178# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1179
1180 /* And these aren't accurate at all. They are useful only for above
1181 * Latin1, which utilities and bootstrapping don't deal with */
1182# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1183# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1184
1185 /* Many of the macros later in this file are defined in terms of these. By
1186 * implementing them with a function, which converts the class number into
1187 * a call to the desired macro, all of the later ones work. However, that
1188 * function won't be actually defined when building a utility program (no
1189 * perl.h), and so a compiler error will be generated if one is attempted
1190 * to be used. And the above-Latin1 code points require Unicode tables to
1191 * be present, something unlikely to be the case when bootstrapping */
1192# define _generic_isCC(c, classnum) \
1193 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1194# define _generic_isCC_A(c, classnum) \
1195 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
3ffc8c70 1196#endif /* End of no perl.h */
8a58bdcf 1197
e66b99e9
KW
1198#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1199#define isALPHA(c) isALPHA_A(c)
1200#define isBLANK(c) isBLANK_A(c)
1201#define isCNTRL(c) isCNTRL_A(c)
1202#define isDIGIT(c) isDIGIT_A(c)
1203#define isGRAPH(c) isGRAPH_A(c)
1204#define isIDFIRST(c) isIDFIRST_A(c)
1205#define isLOWER(c) isLOWER_A(c)
1206#define isPRINT(c) isPRINT_A(c)
1207#define isPSXSPC(c) isPSXSPC_A(c)
1208#define isPUNCT(c) isPUNCT_A(c)
1209#define isSPACE(c) isSPACE_A(c)
1210#define isUPPER(c) isUPPER_A(c)
1211#define isWORDCHAR(c) isWORDCHAR_A(c)
1212#define isXDIGIT(c) isXDIGIT_A(c)
1213
1214/* ASCII casing. These could also be written as
1215 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1216 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1217 which uses table lookup and mask instead of subtraction. (This would
1218 work because the _MOD does not apply in the ASCII range) */
68067e4e
DM
1219#define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1220#define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
bbce6d69 1221
25200305
KW
1222/* In the ASCII range, these are equivalent to what they're here defined to be.
1223 * But by creating these definitions, other code doesn't have to be aware of
1224 * this detail */
1225#define toFOLD(c) toLOWER(c)
25200305
KW
1226#define toTITLE(c) toUPPER(c)
1227
c753c8d3
KW
1228#define toLOWER_A(c) toLOWER(c)
1229#define toUPPER_A(c) toUPPER(c)
25200305
KW
1230#define toFOLD_A(c) toFOLD(c)
1231#define toTITLE_A(c) toTITLE(c)
1a0901db 1232
4650c663 1233/* Use table lookup for speed; returns the input itself if is out-of-range */
b2bf251f 1234#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1235 ? (c) \
f4cd282c 1236 : PL_latin1_lc[ (U8) (c) ])
c753c8d3
KW
1237#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1238
1a0901db 1239/* Modified uc. Is correct uc except for three non-ascii chars which are
4650c663
KW
1240 * all mapped to one of them, and these need special handling; returns the
1241 * input itself if is out-of-range */
b2bf251f 1242#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1243 ? (c) \
f4cd282c 1244 : PL_mod_latin1_uc[ (U8) (c) ])
31f05a37 1245#define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale
84061b6a 1246
beab9ebe
KW
1247/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
1248
1249/* For internal core Perl use only: the base macro for defining macros like
1250 * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point
31f05a37
KW
1251 * (0-255) to check. In a UTF-8 locale, the result is the same as calling
1252 * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like
1253 * _CC_UPPER, which gives the class number for doing this. For non-UTF-8
1254 * locales, the code to actually do the test this is passed in 'non_utf8'. If
1255 * 'c' is above 255, 0 is returned. For accessing the full range of possible
1256 * code points under locale rules, use the macros based on _generic_LC_uvchr
1257 * instead of this. */
beab9ebe
KW
1258#define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \
1259 (! FITS_IN_8_BITS(c) \
1260 ? 0 \
31f05a37
KW
1261 : IN_UTF8_CTYPE_LOCALE \
1262 ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \
beab9ebe
KW
1263 : cBOOL(non_utf8))
1264
1265/* For internal core Perl use only: a helper macro for defining macros like
1266 * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to
1267 * actually do this test is passed in 'non_utf8_func', which is called on 'c',
1268 * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See
1269 * _generic_LC_base for more info */
1270#define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \
1271 _generic_LC_base(c,utf8_locale_classnum, \
1272 non_utf8_func( (_LC_CAST) (c)))
1273
1274/* For internal core Perl use only: like _generic_LC, but also returns TRUE if
1275 * 'c' is the platform's native underscore character */
1276#define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \
1277 _generic_LC_base(c, utf8_locale_classnum, \
1278 (non_utf8_func( (_LC_CAST) (c)) \
1279 || (char)(c) == '_'))
1280
1281/* These next three are also for internal core Perl use only: case-change
1282 * helper macros */
1283#define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \
1284 ? (c) \
31f05a37
KW
1285 : (IN_UTF8_CTYPE_LOCALE) \
1286 ? PL_latin1_lc[ (U8) (c) ] \
beab9ebe
KW
1287 : function((cast)(c)))
1288
31f05a37
KW
1289/* Note that the result can be larger than a byte in a UTF-8 locale. It
1290 * returns a single value, so can't adequately return the upper case of LATIN
1291 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1292 * values "SS"); instead it asserts against that under DEBUGGING, and
1293 * otherwise returns its input */
beab9ebe
KW
1294#define _generic_toUPPER_LC(c, function, cast) \
1295 (! FITS_IN_8_BITS(c) \
1296 ? (c) \
31f05a37
KW
1297 : ((! IN_UTF8_CTYPE_LOCALE) \
1298 ? function((cast)(c)) \
1299 : ((((U8)(c)) == MICRO_SIGN) \
1300 ? GREEK_CAPITAL_LETTER_MU \
1301 : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \
1302 ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \
1303 : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \
1304 ? (__ASSERT_(0) (c)) \
1305 : PL_mod_latin1_uc[ (U8) (c) ])))))
1306
1307/* Note that the result can be larger than a byte in a UTF-8 locale. It
1308 * returns a single value, so can't adequately return the fold case of LATIN
1309 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1310 * values "ss"); instead it asserts against that under DEBUGGING, and
1311 * otherwise returns its input */
beab9ebe 1312#define _generic_toFOLD_LC(c, function, cast) \
4d7db1b9 1313 ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \
31f05a37 1314 ? GREEK_SMALL_LETTER_MU \
4d7db1b9
KW
1315 : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \
1316 || (c) != LATIN_SMALL_LETTER_SHARP_S) \
1317 _generic_toLOWER_LC(c, function, cast)))
beab9ebe 1318
84061b6a
KW
1319/* Use the libc versions for these if available. */
1320#if defined(HAS_ISASCII) && ! defined(USE_NEXT_CTYPE)
1321# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c)))
1322#else
1323# define isASCII_LC(c) isASCII(c)
1324#endif
1325
1326#if defined(HAS_ISBLANK) && ! defined(USE_NEXT_CTYPE)
1327# define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank)
95f34b6f
KW
1328#else /* Unlike isASCII, varies if in a UTF-8 locale */
1329# define isBLANK_LC(c) (IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c)
84061b6a
KW
1330#endif
1331
4650c663 1332#ifdef USE_NEXT_CTYPE /* NeXT computers */
bbce6d69 1333
beab9ebe
KW
1334# define _LC_CAST unsigned int /* Needed by _generic_LC. NeXT functions
1335 use this as their input type */
1336
1337# define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, NXIsAlpha)
1338# define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, NXIsAlNum)
1339# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, NXIsCntrl)
1340# define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, NXIsDigit)
1341# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, NXIsGraph)
1342# define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, NXIsAlpha)
1343# define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, NXIsLower)
1344# define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, NXIsPrint)
1345# define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, NXIsPunct)
1346# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, NXIsSpace)
1347# define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, NXIsUpper)
1348# define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, NXIsAlNum)
1349# define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, NXIsXdigit)
1350
1351# define toLOWER_LC(c) _generic_toLOWER_LC((c), NXToLower, unsigned int)
1352# define toUPPER_LC(c) _generic_toUPPER_LC((c), NXToUpper, unsigned int)
1353# define toFOLD_LC(c) _generic_toFOLD_LC((c), NXToLower, unsigned int)
bbce6d69
PP
1354
1355#else /* !USE_NEXT_CTYPE */
b8c5462f 1356
841f64f7
BF
1357# define _LC_CAST U8
1358
bbce6d69 1359# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
4650c663 1360 /* For most other platforms */
beab9ebe 1361
beab9ebe
KW
1362# define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha)
1363# define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum)
1364# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1365# define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit)
1366# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph)
1367# define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha)
1368# define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower)
1369# define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint)
1370# define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct)
1371# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1372# define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper)
1373# define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum)
1374# define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit)
1375
1376
1377# define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1378# define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1379# define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
bbce6d69 1380
4650c663 1381# else /* The final fallback position */
bbce6d69 1382
bbce6d69 1383# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1384# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
b8c5462f 1385# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1386# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1387# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1388# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1389# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1390# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1391# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1392# define isSPACE_LC(c) (isascii(c) && isspace(c))
1393# define isUPPER_LC(c) (isascii(c) && isupper(c))
1394# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1395# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
4650c663 1396
76fde96a 1397# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1398# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
beab9ebe 1399# define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c))
bbce6d69
PP
1400
1401# endif
a0d0e21e 1402#endif /* USE_NEXT_CTYPE */
55204971 1403
eba68aa0
KW
1404#define isIDCONT(c) isWORDCHAR(c)
1405#define isIDCONT_A(c) isWORDCHAR_A(c)
1406#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1407#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1408#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1409
4650c663
KW
1410/* For internal core Perl use only: the base macros for defining macros like
1411 * isALPHA_uni. 'c' is the code point to check. 'classnum' is the POSIX class
1412 * number defined earlier in this file. _generic_uni() is used for POSIX
1413 * classes where there is a macro or function 'above_latin1' that takes the
1414 * single argument 'c' and returns the desired value. These exist for those
1415 * classes which have simple definitions, avoiding the overhead of a hash
1416 * lookup or inversion list binary search. _generic_swash_uni() can be used
1417 * for classes where that overhead is faster than a direct lookup.
1418 * _generic_uni() won't compile if 'c' isn't unsigned, as it won't match the
1419 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so
1420 * have duplicate checks here, so could create versions of the macros that
1421 * don't, but experiments show that gcc optimizes them out anyway. */
66c17564
KW
1422
1423/* Note that all ignore 'use bytes' */
cd500f2f
KW
1424#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1425 ? _generic_isCC(c, classnum) \
1426 : above_latin1(c))
1427#define _generic_swash_uni(classnum, c) ((c) < 256 \
1428 ? _generic_isCC(c, classnum) \
922e8cb4 1429 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1430#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1431#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1432#define isASCII_uni(c) isASCII(c)
1433#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1434#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1435#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1436#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1437#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1438#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1439#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1440#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1441
509fb054
KW
1442/* Posix and regular space are identical above Latin1 */
1443#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1444
cd500f2f 1445#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1446#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1447#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1448#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1449#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1450#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1451
b8d68ded 1452#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1453#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1454#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1455#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1456
4650c663
KW
1457/* For internal core Perl use only: the base macros for defining macros like
1458 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code
1459 * point, not just 0-255. Like _generic_uni, there are two versions, one for
1460 * simple class definitions; the other for more complex. These are like
1461 * _generic_uni, so see it for more info. */
cd500f2f
KW
1462#define _generic_LC_uvchr(latin1, above_latin1, c) \
1463 (c < 256 ? latin1(c) : above_latin1(c))
1464#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1465 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1466
1467#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1468#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1469 _CC_ALPHANUMERIC, c)
509fb054 1470#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1471#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1472#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1473#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1474#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1475#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1476 _is_uni_perl_idcont, c)
cd500f2f
KW
1477#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1478 _is_uni_perl_idstart, c)
1479#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1480#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1481#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1482 space under locale */
cd500f2f
KW
1483#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1484#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1485 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1486#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1487#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1488 _CC_WORDCHAR, c)
1489#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1490
36b00f93 1491#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1492
4650c663
KW
1493/* For internal core Perl use only: the base macros for defining macros like
1494 * isALPHA_utf8. These are like the earlier defined macros, but take an input
1495 * UTF-8 encoded string 'p'. If the input is in the Latin1 range, use
1496 * the Latin1 macro 'classnum' on 'p'. Otherwise use the value given by the
1497 * 'utf8' parameter. This relies on the fact that ASCII characters have the
1498 * same representation whether utf8 or not. Note that it assumes that the utf8
1499 * has been validated, and ignores 'use bytes' */
cd500f2f 1500#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1501 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1502 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c 1503 ? _generic_isCC( \
94bb8c36 1504 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
aea0c08c 1505 *((p)+1 )), \
94bb8c36 1506 classnum) \
252810d0 1507 : utf8)
cd500f2f
KW
1508/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1509 * can be a macro */
1510#define _generic_func_utf8(classnum, above_latin1, p) \
1511 _generic_utf8(classnum, p, above_latin1(p))
4650c663 1512/* Like the above, but passes classnum to _isFOO_utf8(), instead of having an
cd500f2f 1513 * 'above_latin1' parameter */
922e8cb4 1514#define _generic_swash_utf8(classnum, p) \
cd500f2f 1515 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1516
cc8ab7c0 1517/* Like the above, but should be used only when it is known that there are no
ff7ecfc3
KW
1518 * characters in the upper-Latin1 range (128-255 on ASCII platforms) which the
1519 * class is TRUE for. Hence it can skip the tests for this range.
1520 * 'above_latin1' should include its arguments */
3ded5eb0 1521#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1522 (UTF8_IS_INVARIANT(*(p)) \
1523 ? _generic_isCC(*(p), classnum) \
1524 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
3ded5eb0 1525 ? above_latin1 \
cc8ab7c0
KW
1526 : 0)
1527
84238efa
KW
1528/* NOTE that some of these macros have very similar ones in regcharclass.h.
1529 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1530 * there, differing in name only by an underscore from the one here
1531 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1532 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1533 * points; the regcharclass.h ones are implemented as a series of
1534 * "if-else-if-else ..." */
1535
922e8cb4
KW
1536#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1537#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1538#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1539 utf8, the non-utf8 macro works
1540 */
cd500f2f 1541#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
e8fa43e2
KW
1542
1543#ifdef EBCDIC
1544 /* Because all controls are UTF-8 invariants in EBCDIC, we can use this
1545 * more efficient macro instead of the more general one */
1546# define isCNTRL_utf8(p) isCNTRL_L1(p)
1547#else
1548# define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1549#endif
1550
cd500f2f
KW
1551#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1552 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1553#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1554#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1555 _is_utf8_perl_idcont, p)
e5dcd934 1556
c11ff943
KW
1557/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1558 * IDFIRST is an alnum. See
c8362b00 1559 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1560 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1561 * This used to be not the XID version, but we decided to go with the more
1562 * modern Unicode definition */
cd500f2f 1563#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1564 _is_utf8_perl_idstart, p)
e5dcd934 1565
922e8cb4
KW
1566#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1567#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1568
509fb054 1569/* Posix and regular space are identical above Latin1 */
cd500f2f 1570#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1571
922e8cb4 1572#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1573#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1574#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1575#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1576#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1577#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1578 is_XDIGIT_high(p))
a0ed51b3 1579
7d2b61dd 1580#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
509fb054
KW
1581#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1582#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1583#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1584
4650c663
KW
1585/* For internal core Perl use only: the base macros for defining macros like
1586 * isALPHA_LC_utf8. These are like _generic_utf8, but if the first code point
1587 * in 'p' is within the 0-255 range, it uses locale rules from the passed-in
1588 * 'macro' parameter */
94bb8c36
KW
1589#define _generic_LC_utf8(macro, p, utf8) \
1590 (UTF8_IS_INVARIANT(*(p)) \
1591 ? macro(*(p)) \
1592 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1593 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
9201a572
KW
1594 : utf8)
1595
cd500f2f
KW
1596#define _generic_LC_swash_utf8(macro, classnum, p) \
1597 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1598#define _generic_LC_func_utf8(macro, above_latin1, p) \
4650c663 1599 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1600
cd500f2f
KW
1601#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1602 _CC_ALPHANUMERIC, p)
1603#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1604#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1605#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1606#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1607#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1608#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1609#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1610#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1611#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1612#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1613#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1614 space under locale */
cd500f2f
KW
1615#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1616#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1617#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1618#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1619 _CC_WORDCHAR, p)
1620#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1621
fbc19f27
KW
1622/* Macros for backwards compatibility and for completeness when the ASCII and
1623 * Latin1 values are identical */
fbc19f27
KW
1624#define isALPHAU(c) isALPHA_L1(c)
1625#define isDIGIT_L1(c) isDIGIT_A(c)
1626#define isOCTAL(c) isOCTAL_A(c)
1627#define isOCTAL_L1(c) isOCTAL_A(c)
1628#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1629#define isALNUM(c) isWORDCHAR(c)
1630#define isALNUMU(c) isWORDCHAR_L1(c)
1631#define isALNUM_LC(c) isWORDCHAR_LC(c)
1632#define isALNUM_uni(c) isWORDCHAR_uni(c)
1633#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1634#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1635#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1636#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1637#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1638#define isALNUMC(c) isALPHANUMERIC(c)
1639#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1640#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1641#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1642#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1643#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1644
2bd1cbf6
KW
1645/* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII,
1646 * except that they don't necessarily mean the same characters, e.g. CTRL-D is
1647 * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC.
1648 * '?' is special-cased on EBCDIC to APC, which is the control there that is
1649 * the outlier from the block that contains the other controls, just like
1650 * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0
1651 * block. If it weren't special cased, it would yield a non-control.
1652 * The conversion works both ways, so CTRL('D') is 4, and CTRL(4) is D, etc. */
1653#ifndef EBCDIC
1654# define toCTRL(c) (toUPPER(c) ^ 64)
1655#else
0ed2b00b
KW
1656# define toCTRL(c) ((isPRINT_A(c)) \
1657 ? UNLIKELY((c) == '?') \
1658 ? QUESTION_MARK_CTRL \
1659 : (NATIVE_TO_LATIN1(toUPPER(c)) ^ 64) \
1660 : UNLIKELY((c) == QUESTION_MARK_CTRL) \
1661 ? ((c) == '?') \
1662 : (LATIN1_TO_NATIVE((c) ^ 64)))
2bd1cbf6 1663#endif
bbce6d69 1664
dea28490
JJ
1665/* Line numbers are unsigned, 32 bits. */
1666typedef U32 line_t;
e5dcd934 1667#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1668
91152fc1
DG
1669/* Helpful alias for version prescan */
1670#define is_LAX_VERSION(a,b) \
1671 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1672
1673#define is_STRICT_VERSION(a,b) \
1674 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1675
1676#define BADVERSION(a,b,c) \
1677 if (b) { \
1678 *b = c; \
1679 } \
1680 return a;
8c52afec 1681
cb27eebd
KW
1682/* Converts a character known to represent a hexadecimal digit (0-9, A-F, or
1683 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer,
1684 * which is advanced. The input is validated only by an assert() in DEBUGGING
1685 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and
1686 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */
1687#define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \
1688 ? (c) \
1689 : ((c) + 9))))
902008b8
KW
1690#define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \
1691 ? (*(s)++) \
1692 : (*(s)++ + 9))))
95a59cab 1693
cb27eebd
KW
1694/* Converts a character known to represent an octal digit (0-7) to its numeric
1695 * value. The input is validated only by an assert() in DEBUGGING builds. In
1696 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */
1697#define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c)))
1698
8e84507e 1699/*
ccfc67b7
JH
1700=head1 Memory Management
1701
a02a5408 1702=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1703The XSUB-writer's interface to the C C<malloc> function.
1704
0d7b2759
MH
1705Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1706
c5008215
JC
1707In 5.9.3, Newx() and friends replace the older New() API, and drops
1708the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1709themselves. This aid has been superseded by a new build option,
d10b4965 1710PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1711there for use in XS modules supporting older perls.
1712
a02a5408 1713=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1714The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1715cast. See also C<Newx>.
954c1994 1716
0d7b2759
MH
1717Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1718
a02a5408 1719=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1720The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1721memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1722
0d7b2759
MH
1723Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1724
954c1994
GS
1725=for apidoc Am|void|Renew|void* ptr|int nitems|type
1726The XSUB-writer's interface to the C C<realloc> function.
1727
0d7b2759
MH
1728Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1729
954c1994
GS
1730=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1731The XSUB-writer's interface to the C C<realloc> function, with
1732cast.
1733
0d7b2759
MH
1734Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1735
49b8b560 1736=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1737The XSUB-writer's interface to the C C<free> function.
1738
0d7b2759
MH
1739This should B<ONLY> be used on memory obtained using L<"Newx"> and friends.
1740
954c1994
GS
1741=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1742The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1743source, C<dest> is the destination, C<nitems> is the number of items, and
1744C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1745
e90e2364 1746=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
72d33970
FC
1747Like C<Move> but returns dest. Useful
1748for encouraging compilers to tail-call
e90e2364
NC
1749optimise.
1750
954c1994
GS
1751=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1752The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1753source, C<dest> is the destination, C<nitems> is the number of items, and
1754C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1755
e90e2364
NC
1756=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1757
72d33970
FC
1758Like C<Copy> but returns dest. Useful
1759for encouraging compilers to tail-call
e90e2364
NC
1760optimise.
1761
954c1994
GS
1762=for apidoc Am|void|Zero|void* dest|int nitems|type
1763
1764The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1765destination, C<nitems> is the number of items, and C<type> is the type.
1766
e90e2364
NC
1767=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1768
72d33970
FC
1769Like C<Zero> but returns dest. Useful
1770for encouraging compilers to tail-call
e90e2364
NC
1771optimise.
1772
da5d8dbb 1773=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1774This is an architecture-independent macro to copy one structure to another.
954c1994 1775
7e337ee0
JH
1776=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1777
1778Fill up memory with a byte pattern (a byte repeated over and over
1779again) that hopefully catches attempts to access uninitialized memory.
1780
1781=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1782
1783PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1784
1c12ffb4 1785=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1786
1787PoisonWith(0xEF) for catching access to freed memory.
1788
9965345d
JH
1789=for apidoc Am|void|Poison|void* dest|int nitems|type
1790
7e337ee0 1791PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1792
1793=cut */
954c1994 1794
561b68a9
SH
1795/* Maintained for backwards-compatibility only. Use newSV() instead. */
1796#ifndef PERL_CORE
ff06c60c 1797#define NEWSV(x,len) newSV(len)
561b68a9 1798#endif
ff06c60c 1799
19a94d75
JH
1800#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1801
2720f793
JH
1802/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1803 * overly eager compilers that will bleat about e.g.
1804 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1805#ifdef PERL_MALLOC_WRAP
f019c49e 1806#define MEM_WRAP_CHECK(n,t) \
5d9574c1 1807 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (croak_memory_wrap(),0))
27d5b266 1808#define MEM_WRAP_CHECK_1(n,t,a) \
5d9574c1 1809 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1810#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1811
d1decf2b 1812#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
1813#else
1814
410319be
NC
1815#define MEM_WRAP_CHECK(n,t)
1816#define MEM_WRAP_CHECK_1(n,t,a)
1817#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1818#define MEM_WRAP_CHECK_(n,t)
1819
6fff79ce 1820#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1821
1936d2a7 1822#endif
8b44ba4c 1823
fe4f188c 1824#ifdef PERL_MEM_LOG
46c6c7e2 1825/*
9f653bb5 1826 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1827 * go through functions, which are handy for debugging breakpoints, but
1828 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1829 * line number, and C function name if available) passed in. This info can
1830 * then be used for logging the calls, for which one gets a sample
73d1d973 1831 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1832 *
46c6c7e2 1833 * Known problems:
94e892a6 1834 * - not all memory allocs get logged, only those
46c6c7e2 1835 * that go through Newx() and derivatives (while all
94e892a6 1836 * Safefrees do get logged)
46c6c7e2
JH
1837 * - __FILE__ and __LINE__ do not work everywhere
1838 * - __func__ or __FUNCTION__ even less so
1839 * - I think more goes on after the perlio frees but
1840 * the thing is that STDERR gets closed (as do all
1841 * the file descriptors)
1842 * - no deeper calling stack than the caller of the Newx()
1843 * or the kind, but do I look like a C reflection/introspection
1844 * utility to you?
1845 * - the function prototypes for the logging functions
1846 * probably should maybe be somewhere else than handy.h
1847 * - one could consider inlining (macrofying) the logging
1848 * for speed, but I am too lazy
1849 * - one could imagine recording the allocations in a hash,
1850 * (keyed by the allocation address?), and maintain that
1851 * through reallocs and frees, but how to do that without
1852 * any News() happening...?
73d1d973 1853 * - lots of -Ddefines to get useful/controllable output
b953482e 1854 * - lots of ENV reads
46c6c7e2
JH
1855 */
1856
12754f92 1857PERL_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 1858
12754f92 1859PERL_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 1860
12754f92 1861PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1862
0b0ab801 1863# ifdef PERL_CORE
73d1d973 1864# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1865enum mem_log_type {
1866 MLT_ALLOC,
1867 MLT_REALLOC,
d7a2c63c
MHM
1868 MLT_FREE,
1869 MLT_NEW_SV,
1870 MLT_DEL_SV
0b0ab801
MHM
1871};
1872# endif
12754f92 1873# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1874void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1875void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1876# endif
0b0ab801
MHM
1877# endif
1878
fe4f188c
JH
1879#endif
1880
1881#ifdef PERL_MEM_LOG
d1401ee9
MHM
1882#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1883#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1884#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1885#endif
1886
1887#ifndef MEM_LOG_ALLOC
1888#define MEM_LOG_ALLOC(n,t,a) (a)
1889#endif
1890#ifndef MEM_LOG_REALLOC
1891#define MEM_LOG_REALLOC(n,t,v,a) (a)
1892#endif
1893#ifndef MEM_LOG_FREE
1894#define MEM_LOG_FREE(a) (a)
1895#endif
1896
d1401ee9
MHM
1897#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1898#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1899#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1900
1901#ifndef PERL_CORE
a02a5408
JC
1902/* pre 5.9.x compatibility */
1903#define New(x,v,n,t) Newx(v,n,t)
1904#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1905#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1906#endif
a02a5408 1907
ff68c719 1908#define Renew(v,n,t) \
d1401ee9 1909 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1910#define Renewc(v,n,t,c) \
d1401ee9 1911 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1912
1913#ifdef PERL_POISON
1914#define Safefree(d) \
06c0cc96 1915 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1916#else
fe4f188c 1917#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1918#endif
55497cff 1919
8b44ba4c
NC
1920#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1921#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1922#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1923
8b44ba4c
NC
1924#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1925#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1926#ifdef HAS_MEMSET
8b44ba4c 1927#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1928#else
8b44ba4c
NC
1929/* Using bzero(), which returns void. */
1930#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1931#endif
1932
7e337ee0
JH
1933#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1934#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1935#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1936#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1937
caa674f3
DD
1938#ifdef PERL_POISON
1939# define PERL_POISON_EXPR(x) x
1940#else
1941# define PERL_POISON_EXPR(x)
1942#endif
1943
2304df62 1944#ifdef USE_STRUCT_COPY
ff68c719 1945#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1946#else
1947#define StructCopy(s,d,t) Copy(s,d,1,t)
1948#endif
2cc61e15 1949
c3caa5c3
JH
1950/* C_ARRAY_LENGTH is the number of elements in the C array (so you
1951 * want your zero-based indices to be less than but not equal to).
1952 *
1953 * C_ARRAY_END is one past the last: half-open/half-closed range,
1954 * not last-inclusive range. */
622913ab 1955#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
c3caa5c3 1956#define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a))
622913ab 1957
2cc61e15
DD
1958#ifdef NEED_VA_COPY
1959# ifdef va_copy
1960# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1961# else
a1866d1b
JH
1962# if defined(__va_copy)
1963# define Perl_va_copy(s, d) __va_copy(d, s)
1964# else
1965# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1966# endif
2cc61e15
DD
1967# endif
1968#endif
1969
472d47bc
SB
1970/* convenience debug macros */
1971#ifdef USE_ITHREADS
1972#define pTHX_FORMAT "Perl interpreter: 0x%p"
1973#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1974#define pTHX_VALUE_ (void *)my_perl,
1975#define pTHX_VALUE (void *)my_perl
1976#define pTHX__VALUE_ ,(void *)my_perl,
1977#define pTHX__VALUE ,(void *)my_perl
472d47bc 1978#else
3609ea0d 1979#define pTHX_FORMAT
472d47bc 1980#define pTHX__FORMAT
3609ea0d 1981#define pTHX_VALUE_
472d47bc 1982#define pTHX_VALUE
3609ea0d 1983#define pTHX__VALUE_
472d47bc
SB
1984#define pTHX__VALUE
1985#endif /* USE_ITHREADS */
3609ea0d 1986
2acdbac1
NC
1987/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1988 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1989 CPAN::Unpack show any users outside the core. */
1990#ifdef PERL_CORE
d1d15184 1991# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1992#endif
1993
dfff4baf
BF
1994/* Internal macros to deal with gids and uids */
1995#ifdef PERL_CORE
1996
1997# if Uid_t_size > IVSIZE
1998# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
1999# define SvUID(sv) SvNV(sv)
2000# else
2001# if Uid_t_sign <= 0
2002# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
2003# define SvUID(sv) SvIV(sv)
2004# else
2005# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
2006# define SvUID(sv) SvUV(sv)
2007# endif
2008# endif /* Uid_t_size */
2009
2010# if Gid_t_size > IVSIZE
2011# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
2012# define SvGID(sv) SvNV(sv)
2013# else
2014# if Gid_t_sign <= 0
2015# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
2016# define SvGID(sv) SvIV(sv)
2017# else
2018# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
2019# define SvGID(sv) SvUV(sv)
2020# endif
2021# endif /* Gid_t_size */
2022
2023#endif
2024
9d745869
KW
2025#endif /* HANDY_H */
2026
e9a8c099
MHM
2027/*
2028 * Local variables:
2029 * c-indentation-style: bsd
2030 * c-basic-offset: 4
14d04a33 2031 * indent-tabs-mode: nil
e9a8c099
MHM
2032 * End:
2033 *
14d04a33 2034 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 2035 */