This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
improve and reword README.win32
[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
99#ifndef HAS_BOOL
f789f6a4
FC
100# ifdef bool
101# undef bool
102# endif
70d5cb32 103# define bool char
c1d22f6b 104# define HAS_BOOL 1
a687059c 105#endif
0d3e774c 106
236415e2
KW
107/* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
108 * defined as char for example, then the cast from int is
109 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
110 * AIX */
18f5643b 111#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 112
46c6c7e2 113/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
114 * XXX Should really be a Configure probe, with HAS__FUNCTION__
115 * and FUNCTION__ as results.
116 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
117#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
118# define FUNCTION__ __func__
119#else
9ea40801 120# if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 121 (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
122# define FUNCTION__ ""
123# else
124# define FUNCTION__ __FUNCTION__ /* Common extension. */
125# endif
126#endif
127
27d4fb96
PP
128/* XXX A note on the perl source internal type system. The
129 original intent was that I32 be *exactly* 32 bits.
130
131 Currently, we only guarantee that I32 is *at least* 32 bits.
132 Specifically, if int is 64 bits, then so is I32. (This is the case
133 for the Cray.) This has the advantage of meshing nicely with
134 standard library calls (where we pass an I32 and the library is
135 expecting an int), but the disadvantage that an I32 is not 32 bits.
136 Andy Dougherty August 1996
24fef2a7 137
dc45a647
MB
138 There is no guarantee that there is *any* integral type with
139 exactly 32 bits. It is perfectly legal for a system to have
140 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 141
dc45a647
MB
142 Similarly, there is no guarantee that I16 and U16 have exactly 16
143 bits.
693762b4 144
8e84507e
NIS
145 For dealing with issues that may arise from various 32/64-bit
146 systems, we will ask Configure to check out
8175356b 147
3609ea0d
JH
148 SHORTSIZE == sizeof(short)
149 INTSIZE == sizeof(int)
150 LONGSIZE == sizeof(long)
dc45a647 151 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 152 PTRSIZE == sizeof(void *)
dc45a647
MB
153 DOUBLESIZE == sizeof(double)
154 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 155
27d4fb96
PP
156*/
157
69512466
JH
158#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
159# include <inttypes.h>
dd0eed91
JH
160# ifdef INT32_MIN_BROKEN
161# undef INT32_MIN
162# define INT32_MIN (-2147483647-1)
163# endif
164# ifdef INT64_MIN_BROKEN
165# undef INT64_MIN
166# define INT64_MIN (-9223372036854775807LL-1)
167# endif
69512466
JH
168#endif
169
8175356b
JH
170typedef I8TYPE I8;
171typedef U8TYPE U8;
172typedef I16TYPE I16;
173typedef U16TYPE U16;
174typedef I32TYPE I32;
175typedef U32TYPE U32;
6b8eaf93
JH
176#ifdef PERL_CORE
177# ifdef HAS_QUAD
8175356b
JH
178typedef I64TYPE I64;
179typedef U64TYPE U64;
6b8eaf93
JH
180# endif
181#endif /* PERL_CORE */
8175356b 182
9c39b21c
JH
183/* INT64_C/UINT64_C are C99 from <stdint.h> (so they will not be
184 * available in strict C89 mode), but they are nice, so let's define
185 * them if necessary. */
0d3bd0a9 186#if defined(HAS_QUAD)
9c39b21c
JH
187# undef PeRl_INT64_C
188# undef PeRl_UINT64_C
189/* Prefer the native integer types (int and long) over long long
190 * (which is not C89) and Win32-specific __int64. */
191# if QUADKIND == QUAD_IS_INT && INTSIZE == 8
192# define PeRl_INT64_C(c) (c)
193# define PeRl_UINT64_C(c) CAT2(c,U)
194# endif
195# if QUADKIND == QUAD_IS_LONG && LONGSIZE == 8
196# define PeRl_INT64_C(c) CAT2(c,L)
197# define PeRl_UINT64_C(c) CAT2(c,UL)
198# endif
199# if QUADKIND == QUAD_IS_LONG_LONG && defined(HAS_LONG_LONG)
200# define PeRl_INT64_C(c) CAT2(c,LL)
201# define PeRl_UINT64_C(c) CAT2(c,ULL)
202# endif
203# if QUADKIND == QUAD_IS___INT64
204# define PeRl_INT64_C(c) CAT2(c,I64)
205# define PeRl_UINT64_C(c) CAT2(c,UI64)
206# endif
207# ifndef PeRl_INT64_C
208# define PeRl_INT64_C(c) ((I64TYPE)(c)) /* last resort */
209# define PeRl_UINT64_C(c) ((U64TYPE)(c))
210# endif
211/* In OS X the INT64_C/UINT64_C are defined with LL/ULL, which will
212 * not fly with C89-pedantic gcc, so let's undefine them first so that
213 * we can redefine them with our native integer preferring versions. */
8eafc6eb 214# if defined(PERL_DARWIN) && defined(PERL_GCC_PEDANTIC)
9c39b21c
JH
215# undef INT64_C
216# undef UINT64_C
217# endif
218# ifndef INT64_C
219# define INT64_C(c) PeRl_INT64_C(c)
220# endif
221# ifndef UINT64_C
222# define UINT64_C(c) PeRl_UINT64_C(c)
223# endif
e8c95190 224#endif
dc750f2e 225
d8668976 226#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 227
5ff3f7a4
GS
228/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
229 Please search CHAR_MAX in perl.h for further details. */
230#define U8_MAX UINT8_MAX
231#define U8_MIN UINT8_MIN
232
5ff3f7a4
GS
233#define I16_MAX INT16_MAX
234#define I16_MIN INT16_MIN
235#define U16_MAX UINT16_MAX
236#define U16_MIN UINT16_MIN
237
5ff3f7a4
GS
238#define I32_MAX INT32_MAX
239#define I32_MIN INT32_MIN
0e983133
GS
240#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
241# define U32_MAX UINT32_MAX
242#else
243# define U32_MAX 4294967295U
244#endif
5ff3f7a4
GS
245#define U32_MIN UINT32_MIN
246
247#else
248
5c9fa16e
KA
249/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
250 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
251#define U8_MAX PERL_UCHAR_MAX
252#define U8_MIN PERL_UCHAR_MIN
79072805 253
27d4fb96
PP
254#define I16_MAX PERL_SHORT_MAX
255#define I16_MIN PERL_SHORT_MIN
256#define U16_MAX PERL_USHORT_MAX
257#define U16_MIN PERL_USHORT_MIN
79072805 258
c4f23d77 259#if LONGSIZE > 4
27d4fb96
PP
260# define I32_MAX PERL_INT_MAX
261# define I32_MIN PERL_INT_MIN
262# define U32_MAX PERL_UINT_MAX
263# define U32_MIN PERL_UINT_MIN
79072805 264#else
27d4fb96
PP
265# define I32_MAX PERL_LONG_MAX
266# define I32_MIN PERL_LONG_MIN
267# define U32_MAX PERL_ULONG_MAX
268# define U32_MIN PERL_ULONG_MIN
79072805
LW
269#endif
270
5ff3f7a4
GS
271#endif
272
58a9a5d5 273/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
274#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
275#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
276#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
277
ff68c719 278#define Ctl(ch) ((ch) & 037)
8d063cd8 279
4650c663
KW
280/* This is a helper macro to avoid preprocessor issues, replaced by nothing
281 * unless under DEBUGGING, where it expands to an assert of its argument,
282 * followed by a comma (hence the comma operator). If we just used a straight
283 * assert(), we would get a comma with nothing before it when not DEBUGGING */
0f092d08
KW
284#ifdef DEBUGGING
285# define __ASSERT_(statement) assert(statement),
286#else
287# define __ASSERT_(statement)
288#endif
289
3fe05580
MHM
290/*
291=head1 SV-Body Allocation
292
293=for apidoc Ama|SV*|newSVpvs|const char* s
30a15352
KW
294Like C<newSVpvn>, but takes a literal C<NUL>-terminated string instead of a
295string/length pair.
3fe05580 296
84bafc02 297=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
30a15352
KW
298Like C<newSVpvn_flags>, but takes a literal C<NUL>-terminated string instead of
299a string/length pair.
84bafc02 300
3fe05580 301=for apidoc Ama|SV*|newSVpvs_share|const char* s
30a15352
KW
302Like C<newSVpvn_share>, but takes a literal C<NUL>-terminated string instead of
303a string/length pair and omits the hash parameter.
3fe05580 304
9dcc53ea 305=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
30a15352
KW
306Like C<sv_catpvn_flags>, but takes a literal C<NUL>-terminated string instead
307of a string/length pair.
9dcc53ea
Z
308
309=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
310Like C<sv_catpvn_nomg>, but takes a literal string instead of a
311string/length pair.
312
7b4e7469 313=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
314Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
315
9dcc53ea
Z
316=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
317Like C<sv_catpvn_mg>, but takes a literal string instead of a
318string/length pair.
319
7b4e7469 320=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
321Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
322
9dcc53ea
Z
323=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
324Like C<sv_setpvn_mg>, but takes a literal string instead of a
325string/length pair.
326
327=for apidoc Am|SV *|sv_setref_pvs|const char* s
328Like C<sv_setref_pvn>, but takes a literal string instead of a
329string/length pair.
330
3fe05580
MHM
331=head1 Memory Management
332
333=for apidoc Ama|char*|savepvs|const char* s
30a15352
KW
334Like C<savepvn>, but takes a literal C<NUL>-terminated string instead of a
335string/length pair.
3fe05580 336
9dcc53ea
Z
337=for apidoc Ama|char*|savesharedpvs|const char* s
338A version of C<savepvs()> which allocates the duplicate string in memory
339which is shared between threads.
340
3fe05580
MHM
341=head1 GV Functions
342
343=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
344Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
345
346=head1 Hash Manipulation Functions
347
348=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
349Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
350
351=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
352Like C<hv_store>, but takes a literal string instead of a string/length pair
353and omits the hash parameter.
354
510966aa
Z
355=head1 Lexer interface
356
357=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
358
359Like L</lex_stuff_pvn>, but takes a literal string instead of a
360string/length pair.
361
3fe05580
MHM
362=cut
363*/
364
926bb54c
KW
365/* concatenating with "" ensures that only literal strings are accepted as
366 * argument */
26bfd924 367#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 368
926bb54c
KW
369/* note that STR_WITH_LEN() can't be used as argument to macros or functions
370 * that under some configurations might be macros, which means that it requires
371 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
372 */
373
374/* STR_WITH_LEN() shortcuts */
375#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
376#define newSVpvs_flags(str,flags) \
377 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 378#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
379#define sv_catpvs_flags(sv, str, flags) \
380 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
381#define sv_catpvs_nomg(sv, str) \
382 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
383#define sv_catpvs(sv, str) \
384 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
385#define sv_catpvs_mg(sv, str) \
386 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 387#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
388#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
389#define sv_setref_pvs(rv, classname, str) \
390 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 391#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
392#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
393#define gv_stashpvs(str, create) \
394 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
395#define gv_fetchpvs(namebeg, add, sv_type) \
396 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
397#define gv_fetchpvn(namebeg, len, add, sv_type) \
398 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
399#define sv_catxmlpvs(dsv, str, utf8) \
400 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
401#define hv_fetchs(hv,key,lval) \
402 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
403 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
404 : HV_FETCH_JUST_SV, NULL, 0))
405
406#define hv_stores(hv,key,val) \
407 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
408 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 409
510966aa
Z
410#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
411
b96d8cd9
NC
412#define get_cvs(str, flags) \
413 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 414
954c1994 415/*
ccfc67b7
JH
416=head1 Miscellaneous Functions
417
954c1994
GS
418=for apidoc Am|bool|strNE|char* s1|char* s2
419Test two strings to see if they are different. Returns true or
420false.
421
422=for apidoc Am|bool|strEQ|char* s1|char* s2
423Test two strings to see if they are equal. Returns true or false.
424
425=for apidoc Am|bool|strLT|char* s1|char* s2
426Test two strings to see if the first, C<s1>, is less than the second,
427C<s2>. Returns true or false.
428
429=for apidoc Am|bool|strLE|char* s1|char* s2
430Test two strings to see if the first, C<s1>, is less than or equal to the
431second, C<s2>. Returns true or false.
432
433=for apidoc Am|bool|strGT|char* s1|char* s2
434Test two strings to see if the first, C<s1>, is greater than the second,
435C<s2>. Returns true or false.
436
437=for apidoc Am|bool|strGE|char* s1|char* s2
438Test two strings to see if the first, C<s1>, is greater than or equal to
439the second, C<s2>. Returns true or false.
440
441=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
442Test two strings to see if they are different. The C<len> parameter
72d33970 443indicates the number of bytes to compare. Returns true or false. (A
954c1994
GS
444wrapper for C<strncmp>).
445
446=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
447Test two strings to see if they are equal. The C<len> parameter indicates
72d33970 448the number of bytes to compare. Returns true or false. (A wrapper for
954c1994
GS
449C<strncmp>).
450
451=cut
452*/
453
8d063cd8
LW
454#define strNE(s1,s2) (strcmp(s1,s2))
455#define strEQ(s1,s2) (!strcmp(s1,s2))
456#define strLT(s1,s2) (strcmp(s1,s2) < 0)
457#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
458#define strGT(s1,s2) (strcmp(s1,s2) > 0)
459#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
460#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
461#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 462
36477c24
PP
463#ifdef HAS_MEMCMP
464# define memNE(s1,s2,l) (memcmp(s1,s2,l))
465# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
466#else
467# define memNE(s1,s2,l) (bcmp(s1,s2,l))
468# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
469#endif
470
568a785a 471#define memEQs(s1, l, s2) \
8e6497b6 472 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
473#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
474
bbce6d69
PP
475/*
476 * Character classes.
477 *
478 * Unfortunately, the introduction of locales means that we
479 * can't trust isupper(), etc. to tell the truth. And when
480 * it comes to /\w+/ with tainting enabled, we *must* be able
481 * to trust our character classes.
482 *
483 * Therefore, the default tests in the text of Perl will be
484 * independent of locale. Any code that wants to depend on
485 * the current locale will use the tests that begin with "lc".
486 */
487
2304df62
AD
488#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
489# ifndef CTYPE256
490# define CTYPE256
491# endif
492#endif
493
954c1994 494/*
ccfc67b7 495
dcccc8ff 496=head1 Character classification
243effed
KW
497This section is about functions (really macros) that classify characters
498into types, such as punctuation versus alphabetic, etc. Most of these are
499analogous to regular expression character classes. (See
500L<perlrecharclass/POSIX Character Classes>.) There are several variants for
501each class. (Not all macros have all variants; each item below lists the
502ones valid for it.) None are affected by C<use bytes>, and only the ones
503with C<LC> in the name are affected by the current locale.
504
505The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
506C<U8>) as input and returns a boolean as to whether or not the character
6aff1f14
KW
507represented by that octet is (or on non-ASCII platforms, corresponds to) an
508ASCII character in the named class based on platform, Unicode, and Perl rules.
509If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 510
6aff1f14
KW
511Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
512with no suffix C<"_A">.
4b9734bf
KW
513
514Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
515onto the platform. That is, the code points that are ASCII are unaffected,
516since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
517as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
518true when called with the code point 0xDF, which is a word character in both
4650c663 519ASCII and EBCDIC (though it represents different characters in each).
243effed
KW
520
521Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
522point as input. If the code point is larger than 255, Unicode rules are used
523to determine if it is in the character class. For example,
6aff1f14
KW
524C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
525WITH MACRON in Unicode, and is a word character.
243effed
KW
526
527Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
528(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
529classification of just the first (possibly multi-byte) character in the string
530is tested.
243effed 531
31f05a37
KW
532Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but the
533result is based on the current locale, which is what C<LC> in the name stands
534for. If Perl can determine that the current locale is a UTF-8 locale, it uses
535the published Unicode rules; otherwise, it uses the C library function that
536gives the named classification. For example, C<isDIGIT_LC()> when not in a
537UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always
538returned if the input won't fit into an octet.
243effed
KW
539
540Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
541returns the same as C<isFOO_LC> for input code points less than 256, and
542returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
543
544Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
545(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
546classification of just the first (possibly multi-byte) character in the string
547is tested.
ccfc67b7 548
954c1994 549=for apidoc Am|bool|isALPHA|char ch
2787a470 550Returns a boolean indicating whether the specified character is an
6aff1f14 551alphabetic character, analogous to C<m/[[:alpha:]]/>.
dcccc8ff
KW
552See the L<top of this section|/Character classification> for an explanation of
553variants
e5ad6aba 554C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 555C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 556
15861f94
KW
557=for apidoc Am|bool|isALPHANUMERIC|char ch
558Returns a boolean indicating whether the specified character is a either an
6aff1f14 559alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
dcccc8ff
KW
560See the L<top of this section|/Character classification> for an explanation of
561variants
15861f94 562C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 563C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
564C<isALPHANUMERIC_LC_utf8>.
565
8a58bdcf
KW
566=for apidoc Am|bool|isASCII|char ch
567Returns a boolean indicating whether the specified character is one of the 128
243effed 568characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 569On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
570character corresponds to an ASCII character. Variants C<isASCII_A()> and
571C<isASCII_L1()> are identical to C<isASCII()>.
dcccc8ff
KW
572See the L<top of this section|/Character classification> for an explanation of
573variants
243effed
KW
574C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
575C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
576library routine C<isascii()>. In these cases, the variants whose names contain
577C<LC> are the same as the corresponding ones without.
578
d98532ea
KW
579Also note, that because all ASCII characters are UTF-8 invariant (meaning they
580have the exact same representation (always a single byte) whether encoded in
581UTF-8 or not), C<isASCII> will give the correct results when called with any
582byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
583work properly on any string encoded or not in UTF-8.
584
243effed
KW
585=for apidoc Am|bool|isBLANK|char ch
586Returns a boolean indicating whether the specified character is a
6aff1f14 587character considered to be a blank, analogous to C<m/[[:blank:]]/>.
dcccc8ff
KW
588See the L<top of this section|/Character classification> for an explanation of
589variants
e5ad6aba 590C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
591C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
592platforms do not have the C library routine C<isblank()>. In these cases, the
593variants whose names contain C<LC> are the same as the corresponding ones
594without.
595
596=for apidoc Am|bool|isCNTRL|char ch
597Returns a boolean indicating whether the specified character is a
6aff1f14 598control character, analogous to C<m/[[:cntrl:]]/>.
dcccc8ff
KW
599See the L<top of this section|/Character classification> for an explanation of
600variants
e5ad6aba 601C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
602C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
603On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 604
954c1994 605=for apidoc Am|bool|isDIGIT|char ch
2787a470 606Returns a boolean indicating whether the specified character is a
6aff1f14 607digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 608Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
dcccc8ff
KW
609See the L<top of this section|/Character classification> for an explanation of
610variants
e5ad6aba 611C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
612C<isDIGIT_LC_utf8>.
613
614=for apidoc Am|bool|isGRAPH|char ch
615Returns a boolean indicating whether the specified character is a
6aff1f14 616graphic character, analogous to C<m/[[:graph:]]/>.
dcccc8ff
KW
617See the L<top of this section|/Character classification> for an explanation of
618variants
e5ad6aba 619C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 620C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 621
0c82b6df 622=for apidoc Am|bool|isLOWER|char ch
2787a470 623Returns a boolean indicating whether the specified character is a
6aff1f14 624lowercase character, analogous to C<m/[[:lower:]]/>.
dcccc8ff
KW
625See the L<top of this section|/Character classification> for an explanation of
626variants
e5ad6aba 627C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 628C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 629
c99e91e9 630=for apidoc Am|bool|isOCTAL|char ch
2787a470 631Returns a boolean indicating whether the specified character is an
6aff1f14 632octal digit, [0-7].
243effed
KW
633The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
634C<isOCTAL>.
635
636=for apidoc Am|bool|isPUNCT|char ch
637Returns a boolean indicating whether the specified character is a
6aff1f14
KW
638punctuation character, analogous to C<m/[[:punct:]]/>.
639Note that the definition of what is punctuation isn't as
243effed
KW
640straightforward as one might desire. See L<perlrecharclass/POSIX Character
641Classes> for details.
dcccc8ff
KW
642See the L<top of this section|/Character classification> for an explanation of
643variants
e5ad6aba 644C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 645C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 646
0c82b6df 647=for apidoc Am|bool|isSPACE|char ch
2787a470 648Returns a boolean indicating whether the specified character is a
6aff1f14 649whitespace character. This is analogous
398d098a
KW
650to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
651(experimentally), this also matches what C<m/[[:space:]]/> does.
a04e6aad 652("Experimentally" means that this change may be backed out in 5.22 if
398d098a
KW
653field experience indicates that it was unwise.) Prior to 5.18, only the
654locale forms of this macro (the ones with C<LC> in their names) matched
655precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
656in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
657(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
dcccc8ff
KW
658See the L<top of this section|/Character classification> for an explanation of
659variants
e5ad6aba 660C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 661C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 662
398d098a
KW
663=for apidoc Am|bool|isPSXSPC|char ch
664(short for Posix Space)
665Starting in 5.18, this is identical (experimentally) in all its forms to the
666corresponding C<isSPACE()> macros. ("Experimentally" means that this change
a04e6aad 667may be backed out in 5.22 if field experience indicates that it
398d098a
KW
668was unwise.)
669The locale forms of this macro are identical to their corresponding
670C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
671non-locale forms differ from their C<isSPACE()> forms only in that the
672C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
673Otherwise they are identical. Thus this macro is analogous to what
674C<m/[[:space:]]/> matches in a regular expression.
dcccc8ff
KW
675See the L<top of this section|/Character classification> for an explanation of
676variants
e5ad6aba 677C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
678C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
679
954c1994 680=for apidoc Am|bool|isUPPER|char ch
2787a470 681Returns a boolean indicating whether the specified character is an
6aff1f14 682uppercase character, analogous to C<m/[[:upper:]]/>.
dcccc8ff
KW
683See the L<top of this section|/Character classification> for an explanation of
684variants
e5ad6aba 685C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 686C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 687
243effed 688=for apidoc Am|bool|isPRINT|char ch
8eea39dd 689Returns a boolean indicating whether the specified character is a
6aff1f14 690printable character, analogous to C<m/[[:print:]]/>.
dcccc8ff
KW
691See the L<top of this section|/Character classification> for an explanation of
692variants
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.
dcccc8ff
KW
705See the L<top of this section|/Character classification> for an explanation of
706variants
243effed
KW
707C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
708C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
709
710=for apidoc Am|bool|isXDIGIT|char ch
711Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
712digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
713and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
dcccc8ff
KW
714See the L<top of this section|/Character classification> for an explanation of
715variants
243effed
KW
716C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
717C<isXDIGIT_LC_utf8>.
718
3c3ecf18
KW
719=for apidoc Am|bool|isIDFIRST|char ch
720Returns a boolean indicating whether the specified character can be the first
721character of an identifier. This is very close to, but not quite the same as
722the official Unicode property C<XID_Start>. The difference is that this
723returns true only if the input character also matches L</isWORDCHAR>.
dcccc8ff
KW
724See the L<top of this section|/Character classification> for an explanation of
725variants
3c3ecf18
KW
726C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
727C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
728
729=for apidoc Am|bool|isIDCONT|char ch
730Returns a boolean indicating whether the specified character can be the
731second or succeeding character of an identifier. This is very close to, but
732not quite the same as the official Unicode property C<XID_Continue>. The
733difference is that this returns true only if the input character also matches
dcccc8ff
KW
734L</isWORDCHAR>. See the L<top of this section|/Character classification> for
735an
3c3ecf18
KW
736explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
737C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
738C<isIDCONT_LC_utf8>.
739
243effed 740=head1 Miscellaneous Functions
8eea39dd 741
95a59cab 742=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 743Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
744Behaviour is only well defined when isXDIGIT(*str) is true.
745
e7c1e6c1
KW
746=head1 Character case changing
747
1f607577
KW
748=for apidoc Am|U8|toUPPER|U8 ch
749Converts the specified character to uppercase. If the input is anything but an
750ASCII lowercase character, that input character itself is returned. Variant
c753c8d3 751C<toUPPER_A> is equivalent.
954c1994 752
1f607577
KW
753=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
754Converts the Unicode code point C<cp> to its uppercase version, and
755stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
756that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
757bytes since the uppercase version may be longer than the original character.
758
759The first code point of the uppercased version is returned
760(but note, as explained just above, that there may be more.)
761
762=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
763Converts the UTF-8 encoded character at C<p> to its uppercase version, and
764stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
765that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
766bytes since the uppercase version may be longer than the original character.
767
768The first code point of the uppercased version is returned
769(but note, as explained just above, that there may be more.)
770
771The input character at C<p> is assumed to be well-formed.
772
25200305
KW
773=for apidoc Am|U8|toFOLD|U8 ch
774Converts the specified character to foldcase. If the input is anything but an
775ASCII uppercase character, that input character itself is returned. Variant
776C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
777Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
778
1f607577
KW
779=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
780Converts the Unicode code point C<cp> to its foldcase version, and
781stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
782that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
783bytes since the foldcase version may be longer than the original character.
784
785The first code point of the foldcased version is returned
786(but note, as explained just above, that there may be more.)
787
788=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
789Converts the UTF-8 encoded character at C<p> to its foldcase version, and
790stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
791that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
792bytes since the foldcase version may be longer than the original character.
793
794The first code point of the foldcased version is returned
795(but note, as explained just above, that there may be more.)
796
797The input character at C<p> is assumed to be well-formed.
798
799=for apidoc Am|U8|toLOWER|U8 ch
800Converts the specified character to lowercase. If the input is anything but an
801ASCII uppercase character, that input character itself is returned. Variant
c753c8d3 802C<toLOWER_A> is equivalent.
954c1994 803
1f607577
KW
804=for apidoc Am|U8|toLOWER_L1|U8 ch
805Converts the specified Latin1 character to lowercase. The results are undefined if
806the input doesn't fit in a byte.
807
808=for apidoc Am|U8|toLOWER_LC|U8 ch
809Converts the specified character to lowercase using the current locale's rules,
810if possible; otherwise returns the input character itself.
811
812=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
813Converts the Unicode code point C<cp> to its lowercase version, and
814stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
815that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
816bytes since the lowercase version may be longer than the original character.
817
818The first code point of the lowercased version is returned
819(but note, as explained just above, that there may be more.)
820
821=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
822Converts the UTF-8 encoded character at C<p> to its lowercase version, and
823stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
824that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
825bytes since the lowercase version may be longer than the original character.
826
827The first code point of the lowercased version is returned
828(but note, as explained just above, that there may be more.)
829
830The input character at C<p> is assumed to be well-formed.
831
832=for apidoc Am|U8|toLOWER_LC|U8 ch
833Converts the specified character to lowercase using the current locale's rules,
834if possible; otherwise returns the input character itself.
835
25200305
KW
836=for apidoc Am|U8|toTITLE|U8 ch
837Converts the specified character to titlecase. If the input is anything but an
838ASCII lowercase character, that input character itself is returned. Variant
839C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
840as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
841concept used in locale handling, so there is no functionality for that.)
842
1f607577
KW
843=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
844Converts the Unicode code point C<cp> to its titlecase version, and
845stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
846that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
847bytes since the titlecase version may be longer than the original character.
848
849The first code point of the titlecased version is returned
850(but note, as explained just above, that there may be more.)
851
852=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
853Converts the UTF-8 encoded character at C<p> to its titlecase version, and
854stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
855that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
856bytes since the titlecase version may be longer than the original character.
857
858The first code point of the titlecased version is returned
859(but note, as explained just above, that there may be more.)
860
861The input character at C<p> is assumed to be well-formed.
862
954c1994 863=cut
353c9b6f 864
1f607577 865XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
25200305 866really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
243effed 867
8a58bdcf 868Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
869patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
870
954c1994
GS
871*/
872
7c062697
KW
873/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
874 * is known only in the perl core, and this macro can be called from outside
875 * that */
876#ifdef HAS_QUAD
877# define WIDEST_UTYPE U64TYPE
878#else
879# define WIDEST_UTYPE U32
880#endif
881
3912bc88
KW
882/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
883 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
884 * bits of information are lost even on a 64-bit machine, but to get the
885 * compiler to optimize it out if possible. This is because Configure makes
886 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
887 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
888 */
889#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 890
41f43cc2 891#ifdef EBCDIC
b6340bd0
KW
892# ifndef _ALL_SOURCE
893 /* This returns the wrong results on at least z/OS unless this is
894 * defined. */
895# error _ALL_SOURCE should probably be defined
896# endif
897
898 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
899 * likely not defined, and so we use the native function */
3cedda69 900# define isASCII(c) cBOOL(isascii(c))
41f43cc2
KW
901#else
902# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
903#endif
904
ec79cfbb 905#define isASCII_A(c) isASCII(c)
3509bb53 906#define isASCII_L1(c) isASCII(c)
6515cb76 907
c2da0b36
KW
908/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
909 * and the 8 possible permutations of those bits exactly comprise the 8 octal
910 * digits */
3d33b3e3 911#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 912
ec79cfbb 913/* ASCII range only */
9fb1bf9d 914#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 915
a500dc72
KW
916/* Character class numbers. For internal core Perl use only. The ones less
917 * than 32 are used in PL_charclass[] and the ones up through the one that
918 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
919 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
920 * their actual definitions are here. If that file has a name not used here,
921 * it won't compile.
1709d539
KW
922 *
923 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
924 * frequency of their use. This gives a slight edge to exiting a loop earlier
925 * (in reginclass() in regexec.c) */
1709d539
KW
926# define _CC_WORDCHAR 0 /* \w and [:word:] */
927# define _CC_DIGIT 1 /* \d and [:digit:] */
928# define _CC_ALPHA 2 /* [:alpha:] */
929# define _CC_LOWER 3 /* [:lower:] */
930# define _CC_UPPER 4 /* [:upper:] */
931# define _CC_PUNCT 5 /* [:punct:] */
932# define _CC_PRINT 6 /* [:print:] */
15861f94 933# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 934# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 935# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 936
b0d691b2 937#define _FIRST_NON_SWASH_CC 10
1709d539
KW
938/* The character classes above are implemented with swashes. The second group
939 * (just below) contains the ones implemented without. These are also sorted
940 * in rough order of the frequency of their use, except that \v should be last,
941 * as it isn't a real Posix character class, and some (small) inefficiencies in
942 * regular expression handling would be introduced by putting it in the middle
943 * of those that are. Also, cntrl and ascii come after the others as it may be
944 * useful to group these which have no members that match above Latin1, (or
945 * above ASCII in the latter case) */
946
b0d691b2
KW
947# define _CC_SPACE 10 /* \s */
948# define _CC_BLANK 11 /* [:blank:] */
949# define _CC_XDIGIT 12 /* [:xdigit:] */
950# define _CC_PSXSPC 13 /* [:space:] */
951# define _CC_CNTRL 14 /* [:cntrl:] */
952# define _CC_ASCII 15 /* [:ascii:] */
953# define _CC_VERTSPACE 16 /* \v */
1709d539 954
a0947d7b
KW
955# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
956
1709d539 957/* The members of the third group below do not need to be coordinated with data
3ffc8c70 958 * structures in regcomp.[ch] and regexec.c. */
f12c0118
KW
959# define _CC_IDFIRST 17
960# define _CC_CHARNAME_CONT 18
961# define _CC_NONLATIN1_FOLD 19
962# define _CC_NONLATIN1_SIMPLE_FOLD 20
963# define _CC_QUOTEMETA 21
964# define _CC_NON_FINAL_FOLD 22
965# define _CC_IS_IN_SOME_FOLD 23
8e35b056
KW
966# define _CC_MNEMONIC_CNTRL 24
967/* Unused: 25-31
f4cdb42c
KW
968 * If more bits are needed, one could add a second word for non-64bit
969 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
970 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
971 * is used only for optimization (as of this writing), and differs in the
972 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
973 * characters: the masculine and feminine ordinal indicators, so removing it
974 * would just cause /i regexes which match them to run less efficiently */
96ac0975 975
3a371f2f
KW
976#if defined(PERL_CORE) || defined(PERL_EXT)
977/* An enum version of the character class numbers, to help compilers
978 * optimize */
979typedef enum {
3a371f2f 980 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
981 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
982 _CC_ENUM_ASCII = _CC_ASCII,
983 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 984 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 985 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
986 _CC_ENUM_DIGIT = _CC_DIGIT,
987 _CC_ENUM_GRAPH = _CC_GRAPH,
988 _CC_ENUM_LOWER = _CC_LOWER,
989 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 990 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 991 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 992 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 993 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 994 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 995 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 996 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
997} _char_class_number;
998#endif
999
63c61c3f 1000#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 1001#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 1002
3018b823 1003#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 1004# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
1005 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
1006 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
1007 #error Need to adjust order of swash_property_names[]
1008# endif
1009
b44daeeb
KW
1010/* This is declared static in each of the few files that this is #defined for
1011 * to keep them from being publicly accessible. Hence there is a small amount
1012 * of wasted space */
1013
922e8cb4
KW
1014static const char* const swash_property_names[] = {
1015 "XPosixWord",
1016 "XPosixDigit",
1017 "XPosixAlpha",
1018 "XPosixLower",
1019 "XPosixUpper",
1020 "XPosixPunct",
1021 "XPosixPrint",
1022 "XPosixAlnum",
b0d691b2
KW
1023 "XPosixGraph",
1024 "Cased"
922e8cb4
KW
1025};
1026#endif
1027
96ac0975
NC
1028# ifdef DOINIT
1029EXTCONST U32 PL_charclass[] = {
1030# include "l1_char_class_tab.h"
1031};
1032
1033# else /* ! DOINIT */
1034EXTCONST U32 PL_charclass[];
1035# endif
1036
265c1f46 1037 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70 1038# define _CC_mask(classnum) (1U << (classnum))
4650c663
KW
1039
1040 /* For internal core Perl use only: the base macro for defining macros like
1041 * isALPHA */
ff7ecfc3 1042# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
f4cd282c 1043 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
4eeeb416 1044
f4cdb42c
KW
1045 /* The mask for the _A versions of the macros; it just adds in the bit for
1046 * ASCII. */
1047# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1048
4650c663
KW
1049 /* For internal core Perl use only: the base macro for defining macros like
1050 * isALPHA_A. The foo_A version makes sure that both the desired bit and
1051 * the ASCII bit are present */
3ded5eb0 1052# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
f4cd282c 1053 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
f4cdb42c
KW
1054 == _CC_mask_A(classnum)))
1055
f4cdb42c 1056# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 1057# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
1058# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1059# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
96ca48da 1060# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT) /* No non-ASCII digits */
f4cdb42c
KW
1061# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1062# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1063# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1064# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1065# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1066# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1067# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1068# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
96ca48da 1069# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT) /* No non-ASCII xdigits */
d95f8b6a 1070# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
3ded5eb0
KW
1071# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1072# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1073# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1074
1075 /* continuation character for legal NAME in \N{NAME} */
1076# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1077
1078# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1079# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1080# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1081# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1082# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1083# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1084# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1085# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1086# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1087# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
f4cdb42c 1088
f12c0118
KW
1089 /* Participates in a single-character fold with a character above 255 */
1090# define _HAS_NONLATIN1_SIMPLE_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_SIMPLE_FOLD)))
1091
1092 /* Like the above, but also can be part of a multi-char fold */
f4cd282c 1093# 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 1094
4eeeb416 1095# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
1096# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1097 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
1098# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1099 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
8e35b056
KW
1100# define _IS_MNEMONIC_CNTRL_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1101 _generic_isCC(c, _CC_MNEMONIC_CNTRL)
3ffc8c70 1102#else /* else we don't have perl.h */
3ded5eb0
KW
1103
1104 /* If we don't have perl.h, we are compiling a utility program. Below we
1105 * hard-code various macro definitions that wouldn't otherwise be available
3ffc8c70 1106 * to it. */
9fb1bf9d 1107# ifdef EBCDIC
3ded5eb0 1108 /* Use the native functions. They likely will return false for all
4f48cecf 1109 * non-ASCII values, but this makes sure */
b6340bd0
KW
1110# define isLOWER_A(c) (isASCII(c) && islower(c))
1111# define isPRINT_A(c) (isASCII(c) && isprint(c))
b6340bd0 1112# define isUPPER_A(c) (isASCII(c) && isupper(c))
3ded5eb0 1113# else /* ASCII platform. These are coded based on first principals */
9fb1bf9d
KW
1114# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1115# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
2181da29 1116# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
3ded5eb0 1117# endif /* Below are common definitions for ASCII and non-ASCII */
a4d7a999
KW
1118# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
1119# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
3ded5eb0 1120# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
a4d7a999
KW
1121# define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
1122# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
1123# define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
3ded5eb0 1124# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
a4d7a999
KW
1125# define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1126# define isSPACE_A(c) ((c) == ' ' \
1127 || (c) == '\t' \
1128 || (c) == '\n' \
1129 || (c) == '\r' \
1130 || (c) == '\v' \
1131 || (c) == '\f')
3ded5eb0 1132# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
a4d7a999
KW
1133# define isXDIGIT_A(c) (isDIGIT_A(c) \
1134 || ((c) >= 'a' && (c) <= 'f') \
1135 || ((c) <= 'F' && (c) >= 'A'))
3ded5eb0 1136
3ffc8c70
KW
1137 /* The _L1 macros may be unnecessary for the utilities; I (khw) added them
1138 * during debugging, and it seems best to keep them. */
3ded5eb0
KW
1139# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1140# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1141# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1142# define isBLANK_L1(c) (isBLANK_A(c) \
1143 || (FITS_IN_8_BITS(c) \
1144 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1145# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1146# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1147# define isLOWER_L1(c) (isLOWER_A(c) \
1148 || (FITS_IN_8_BITS(c) \
1149 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1150 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1151 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1152 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1153 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1154# define isPRINT_L1(c) (isPRINT_A(c) \
1155 || (FITS_IN_8_BITS(c) \
1156 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1157# define isPSXSPC_L1(c) isSPACE_L1(c)
1158# define isPUNCT_L1(c) (isPUNCT_A(c) \
1159 || (FITS_IN_8_BITS(c) \
1160 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1161 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1162 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1163 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1164 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1165 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1166 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1167# define isSPACE_L1(c) (isSPACE_A(c) \
1168 || (FITS_IN_8_BITS(c) \
1169 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1170 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1171# define isUPPER_L1(c) (isUPPER_A(c) \
1172 || (FITS_IN_8_BITS(c) \
1173 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1174 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1175 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1176# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1177# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1178# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1179 || isBLANK_L1(c) \
1180 || (c) == '-' \
1181 || (c) == '(' \
1182 || (c) == ')')
1183 /* The following are not fully accurate in the above-ASCII range. I (khw)
1184 * don't think it's necessary to be so for the purposes where this gets
1185 * compiled */
1186# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1187# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1188
1189 /* And these aren't accurate at all. They are useful only for above
1190 * Latin1, which utilities and bootstrapping don't deal with */
1191# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1192# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1193
1194 /* Many of the macros later in this file are defined in terms of these. By
1195 * implementing them with a function, which converts the class number into
1196 * a call to the desired macro, all of the later ones work. However, that
1197 * function won't be actually defined when building a utility program (no
1198 * perl.h), and so a compiler error will be generated if one is attempted
1199 * to be used. And the above-Latin1 code points require Unicode tables to
1200 * be present, something unlikely to be the case when bootstrapping */
1201# define _generic_isCC(c, classnum) \
1202 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1203# define _generic_isCC_A(c, classnum) \
1204 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
3ffc8c70 1205#endif /* End of no perl.h */
8a58bdcf 1206
e66b99e9
KW
1207#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1208#define isALPHA(c) isALPHA_A(c)
1209#define isBLANK(c) isBLANK_A(c)
1210#define isCNTRL(c) isCNTRL_A(c)
1211#define isDIGIT(c) isDIGIT_A(c)
1212#define isGRAPH(c) isGRAPH_A(c)
1213#define isIDFIRST(c) isIDFIRST_A(c)
1214#define isLOWER(c) isLOWER_A(c)
1215#define isPRINT(c) isPRINT_A(c)
1216#define isPSXSPC(c) isPSXSPC_A(c)
1217#define isPUNCT(c) isPUNCT_A(c)
1218#define isSPACE(c) isSPACE_A(c)
1219#define isUPPER(c) isUPPER_A(c)
1220#define isWORDCHAR(c) isWORDCHAR_A(c)
1221#define isXDIGIT(c) isXDIGIT_A(c)
1222
1223/* ASCII casing. These could also be written as
1224 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1225 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1226 which uses table lookup and mask instead of subtraction. (This would
1227 work because the _MOD does not apply in the ASCII range) */
68067e4e
DM
1228#define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1229#define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
bbce6d69 1230
25200305
KW
1231/* In the ASCII range, these are equivalent to what they're here defined to be.
1232 * But by creating these definitions, other code doesn't have to be aware of
1233 * this detail */
1234#define toFOLD(c) toLOWER(c)
25200305
KW
1235#define toTITLE(c) toUPPER(c)
1236
c753c8d3
KW
1237#define toLOWER_A(c) toLOWER(c)
1238#define toUPPER_A(c) toUPPER(c)
25200305
KW
1239#define toFOLD_A(c) toFOLD(c)
1240#define toTITLE_A(c) toTITLE(c)
1a0901db 1241
4650c663 1242/* Use table lookup for speed; returns the input itself if is out-of-range */
b2bf251f 1243#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1244 ? (c) \
f4cd282c 1245 : PL_latin1_lc[ (U8) (c) ])
c753c8d3
KW
1246#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1247
1a0901db 1248/* Modified uc. Is correct uc except for three non-ascii chars which are
4650c663
KW
1249 * all mapped to one of them, and these need special handling; returns the
1250 * input itself if is out-of-range */
b2bf251f 1251#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1252 ? (c) \
f4cd282c 1253 : PL_mod_latin1_uc[ (U8) (c) ])
31f05a37 1254#define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale
84061b6a 1255
beab9ebe
KW
1256/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
1257
1258/* For internal core Perl use only: the base macro for defining macros like
1259 * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point
31f05a37
KW
1260 * (0-255) to check. In a UTF-8 locale, the result is the same as calling
1261 * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like
1262 * _CC_UPPER, which gives the class number for doing this. For non-UTF-8
1263 * locales, the code to actually do the test this is passed in 'non_utf8'. If
1264 * 'c' is above 255, 0 is returned. For accessing the full range of possible
1265 * code points under locale rules, use the macros based on _generic_LC_uvchr
1266 * instead of this. */
beab9ebe
KW
1267#define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \
1268 (! FITS_IN_8_BITS(c) \
1269 ? 0 \
31f05a37
KW
1270 : IN_UTF8_CTYPE_LOCALE \
1271 ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \
beab9ebe
KW
1272 : cBOOL(non_utf8))
1273
1274/* For internal core Perl use only: a helper macro for defining macros like
1275 * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to
1276 * actually do this test is passed in 'non_utf8_func', which is called on 'c',
1277 * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See
1278 * _generic_LC_base for more info */
1279#define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \
1280 _generic_LC_base(c,utf8_locale_classnum, \
1281 non_utf8_func( (_LC_CAST) (c)))
1282
1283/* For internal core Perl use only: like _generic_LC, but also returns TRUE if
1284 * 'c' is the platform's native underscore character */
1285#define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \
1286 _generic_LC_base(c, utf8_locale_classnum, \
1287 (non_utf8_func( (_LC_CAST) (c)) \
1288 || (char)(c) == '_'))
1289
1290/* These next three are also for internal core Perl use only: case-change
1291 * helper macros */
1292#define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \
1293 ? (c) \
31f05a37
KW
1294 : (IN_UTF8_CTYPE_LOCALE) \
1295 ? PL_latin1_lc[ (U8) (c) ] \
beab9ebe
KW
1296 : function((cast)(c)))
1297
31f05a37
KW
1298/* Note that the result can be larger than a byte in a UTF-8 locale. It
1299 * returns a single value, so can't adequately return the upper case of LATIN
1300 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1301 * values "SS"); instead it asserts against that under DEBUGGING, and
1302 * otherwise returns its input */
beab9ebe
KW
1303#define _generic_toUPPER_LC(c, function, cast) \
1304 (! FITS_IN_8_BITS(c) \
1305 ? (c) \
31f05a37
KW
1306 : ((! IN_UTF8_CTYPE_LOCALE) \
1307 ? function((cast)(c)) \
1308 : ((((U8)(c)) == MICRO_SIGN) \
1309 ? GREEK_CAPITAL_LETTER_MU \
1310 : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \
1311 ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \
1312 : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \
1313 ? (__ASSERT_(0) (c)) \
1314 : PL_mod_latin1_uc[ (U8) (c) ])))))
1315
1316/* Note that the result can be larger than a byte in a UTF-8 locale. It
1317 * returns a single value, so can't adequately return the fold case of LATIN
1318 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1319 * values "ss"); instead it asserts against that under DEBUGGING, and
1320 * otherwise returns its input */
beab9ebe 1321#define _generic_toFOLD_LC(c, function, cast) \
4d7db1b9 1322 ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \
31f05a37 1323 ? GREEK_SMALL_LETTER_MU \
4d7db1b9
KW
1324 : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \
1325 || (c) != LATIN_SMALL_LETTER_SHARP_S) \
1326 _generic_toLOWER_LC(c, function, cast)))
beab9ebe 1327
84061b6a 1328/* Use the libc versions for these if available. */
f05550c0 1329#if defined(HAS_ISASCII)
84061b6a
KW
1330# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c)))
1331#else
1332# define isASCII_LC(c) isASCII(c)
1333#endif
1334
f05550c0 1335#if defined(HAS_ISBLANK)
84061b6a 1336# define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank)
95f34b6f
KW
1337#else /* Unlike isASCII, varies if in a UTF-8 locale */
1338# define isBLANK_LC(c) (IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c)
84061b6a
KW
1339#endif
1340
f05550c0 1341#define _LC_CAST U8
bbce6d69 1342
f05550c0 1343#ifdef WIN32
375f5f06
KW
1344 /* The Windows functions don't bother to follow the POSIX standard, which
1345 * for example says that something can't both be a printable and a control.
1346 * But Windows treats the \t control as a printable, and does such things
1347 * as making superscripts into both digits and punctuation. This tames
1348 * these flaws by assuming that the definitions of both controls and space
1349 * are correct, and then making sure that other definitions don't have
1350 * weirdnesses, by making sure that isalnum() isn't also ispunct(), etc.
1351 * Not all possible weirdnesses are checked for, just the ones that were
1352 * detected on actual Microsoft code pages */
1353
f05550c0
BF
1354# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1355# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1356
1357# define isALPHA_LC(c) (_generic_LC(c, _CC_ALPHA, isalpha) && isALPHANUMERIC_LC(c))
1358# define isALPHANUMERIC_LC(c) (_generic_LC(c, _CC_ALPHANUMERIC, isalnum) && ! isPUNCT_LC(c))
1359# define isDIGIT_LC(c) (_generic_LC(c, _CC_DIGIT, isdigit) && isALPHANUMERIC_LC(c))
1360# define isGRAPH_LC(c) (_generic_LC(c, _CC_GRAPH, isgraph) && isPRINT_LC(c))
1361# define isIDFIRST_LC(c) (((c) == '_') || (_generic_LC(c, _CC_IDFIRST, isalpha) && ! isPUNCT_LC(c)))
1362# define isLOWER_LC(c) (_generic_LC(c, _CC_LOWER, islower) && isALPHA_LC(c))
1363# define isPRINT_LC(c) (_generic_LC(c, _CC_PRINT, isprint) && ! isCNTRL_LC(c))
1364# define isPUNCT_LC(c) (_generic_LC(c, _CC_PUNCT, ispunct) && ! isCNTRL_LC(c))
1365# define isUPPER_LC(c) (_generic_LC(c, _CC_UPPER, isupper) && isALPHA_LC(c))
1366# define isWORDCHAR_LC(c) (((c) == '_') || isALPHANUMERIC_LC(c))
1367# define isXDIGIT_LC(c) (_generic_LC(c, _CC_XDIGIT, isxdigit) && isALPHANUMERIC_LC(c))
1368
1369# define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1370# define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1371# define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
1372
1373#elif defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
4650c663 1374 /* For most other platforms */
beab9ebe 1375
f05550c0
BF
1376# define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha)
1377# define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum)
1378# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1379# define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit)
1380# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph)
1381# define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha)
1382# define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower)
1383# define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint)
1384# define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct)
1385# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1386# define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper)
1387# define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum)
1388# define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit)
1389
1390
1391# define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1392# define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1393# define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
1394
1395#else /* The final fallback position */
1396
1397# define isALPHA_LC(c) (isascii(c) && isalpha(c))
1398# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
1399# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
1400# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
1401# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
1402# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1403# define isLOWER_LC(c) (isascii(c) && islower(c))
1404# define isPRINT_LC(c) (isascii(c) && isprint(c))
1405# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
1406# define isSPACE_LC(c) (isascii(c) && isspace(c))
1407# define isUPPER_LC(c) (isascii(c) && isupper(c))
1408# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
1409# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
1410
1411# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
1412# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
1413# define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c))
bbce6d69 1414
f05550c0 1415#endif
55204971 1416
eba68aa0
KW
1417#define isIDCONT(c) isWORDCHAR(c)
1418#define isIDCONT_A(c) isWORDCHAR_A(c)
1419#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1420#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1421#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1422
4650c663
KW
1423/* For internal core Perl use only: the base macros for defining macros like
1424 * isALPHA_uni. 'c' is the code point to check. 'classnum' is the POSIX class
1425 * number defined earlier in this file. _generic_uni() is used for POSIX
1426 * classes where there is a macro or function 'above_latin1' that takes the
1427 * single argument 'c' and returns the desired value. These exist for those
1428 * classes which have simple definitions, avoiding the overhead of a hash
1429 * lookup or inversion list binary search. _generic_swash_uni() can be used
1430 * for classes where that overhead is faster than a direct lookup.
1431 * _generic_uni() won't compile if 'c' isn't unsigned, as it won't match the
1432 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so
1433 * have duplicate checks here, so could create versions of the macros that
1434 * don't, but experiments show that gcc optimizes them out anyway. */
66c17564
KW
1435
1436/* Note that all ignore 'use bytes' */
cd500f2f
KW
1437#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1438 ? _generic_isCC(c, classnum) \
1439 : above_latin1(c))
1440#define _generic_swash_uni(classnum, c) ((c) < 256 \
1441 ? _generic_isCC(c, classnum) \
922e8cb4 1442 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1443#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1444#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1445#define isASCII_uni(c) isASCII(c)
1446#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1447#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1448#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1449#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1450#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1451#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1452#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1453#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1454
509fb054
KW
1455/* Posix and regular space are identical above Latin1 */
1456#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1457
cd500f2f 1458#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1459#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1460#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1461#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1462#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1463#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1464
b8d68ded 1465#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1466#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1467#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1468#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1469
4650c663
KW
1470/* For internal core Perl use only: the base macros for defining macros like
1471 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code
1472 * point, not just 0-255. Like _generic_uni, there are two versions, one for
1473 * simple class definitions; the other for more complex. These are like
1474 * _generic_uni, so see it for more info. */
cd500f2f
KW
1475#define _generic_LC_uvchr(latin1, above_latin1, c) \
1476 (c < 256 ? latin1(c) : above_latin1(c))
1477#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1478 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1479
1480#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1481#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1482 _CC_ALPHANUMERIC, c)
509fb054 1483#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1484#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1485#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1486#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1487#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1488#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1489 _is_uni_perl_idcont, c)
cd500f2f
KW
1490#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1491 _is_uni_perl_idstart, c)
1492#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1493#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1494#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1495 space under locale */
cd500f2f
KW
1496#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1497#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1498 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1499#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1500#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1501 _CC_WORDCHAR, c)
1502#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1503
36b00f93 1504#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1505
4650c663
KW
1506/* For internal core Perl use only: the base macros for defining macros like
1507 * isALPHA_utf8. These are like the earlier defined macros, but take an input
1508 * UTF-8 encoded string 'p'. If the input is in the Latin1 range, use
1509 * the Latin1 macro 'classnum' on 'p'. Otherwise use the value given by the
1510 * 'utf8' parameter. This relies on the fact that ASCII characters have the
1511 * same representation whether utf8 or not. Note that it assumes that the utf8
1512 * has been validated, and ignores 'use bytes' */
cd500f2f 1513#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1514 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1515 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c 1516 ? _generic_isCC( \
94bb8c36 1517 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
aea0c08c 1518 *((p)+1 )), \
94bb8c36 1519 classnum) \
252810d0 1520 : utf8)
cd500f2f
KW
1521/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1522 * can be a macro */
1523#define _generic_func_utf8(classnum, above_latin1, p) \
1524 _generic_utf8(classnum, p, above_latin1(p))
4650c663 1525/* Like the above, but passes classnum to _isFOO_utf8(), instead of having an
cd500f2f 1526 * 'above_latin1' parameter */
922e8cb4 1527#define _generic_swash_utf8(classnum, p) \
cd500f2f 1528 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1529
cc8ab7c0 1530/* Like the above, but should be used only when it is known that there are no
ff7ecfc3
KW
1531 * characters in the upper-Latin1 range (128-255 on ASCII platforms) which the
1532 * class is TRUE for. Hence it can skip the tests for this range.
1533 * 'above_latin1' should include its arguments */
3ded5eb0 1534#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1535 (UTF8_IS_INVARIANT(*(p)) \
1536 ? _generic_isCC(*(p), classnum) \
1537 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
3ded5eb0 1538 ? above_latin1 \
cc8ab7c0
KW
1539 : 0)
1540
84238efa
KW
1541/* NOTE that some of these macros have very similar ones in regcharclass.h.
1542 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1543 * there, differing in name only by an underscore from the one here
1544 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1545 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1546 * points; the regcharclass.h ones are implemented as a series of
1547 * "if-else-if-else ..." */
1548
922e8cb4
KW
1549#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1550#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1551#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1552 utf8, the non-utf8 macro works
1553 */
cd500f2f 1554#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
e8fa43e2
KW
1555
1556#ifdef EBCDIC
1557 /* Because all controls are UTF-8 invariants in EBCDIC, we can use this
1558 * more efficient macro instead of the more general one */
1559# define isCNTRL_utf8(p) isCNTRL_L1(p)
1560#else
1561# define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1562#endif
1563
cd500f2f
KW
1564#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1565 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1566#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1567#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1568 _is_utf8_perl_idcont, p)
e5dcd934 1569
c11ff943
KW
1570/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1571 * IDFIRST is an alnum. See
c8362b00 1572 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1573 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1574 * This used to be not the XID version, but we decided to go with the more
1575 * modern Unicode definition */
cd500f2f 1576#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1577 _is_utf8_perl_idstart, p)
e5dcd934 1578
922e8cb4
KW
1579#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1580#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1581
509fb054 1582/* Posix and regular space are identical above Latin1 */
cd500f2f 1583#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1584
922e8cb4 1585#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1586#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1587#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1588#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1589#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1590#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1591 is_XDIGIT_high(p))
a0ed51b3 1592
7d2b61dd 1593#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
509fb054
KW
1594#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1595#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1596#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1597
4650c663
KW
1598/* For internal core Perl use only: the base macros for defining macros like
1599 * isALPHA_LC_utf8. These are like _generic_utf8, but if the first code point
1600 * in 'p' is within the 0-255 range, it uses locale rules from the passed-in
1601 * 'macro' parameter */
94bb8c36
KW
1602#define _generic_LC_utf8(macro, p, utf8) \
1603 (UTF8_IS_INVARIANT(*(p)) \
1604 ? macro(*(p)) \
1605 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1606 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
9201a572
KW
1607 : utf8)
1608
cd500f2f
KW
1609#define _generic_LC_swash_utf8(macro, classnum, p) \
1610 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1611#define _generic_LC_func_utf8(macro, above_latin1, p) \
4650c663 1612 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1613
cd500f2f
KW
1614#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1615 _CC_ALPHANUMERIC, p)
1616#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1617#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1618#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1619#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1620#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1621#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1622#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1623#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1624#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1625#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1626#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1627 space under locale */
cd500f2f
KW
1628#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1629#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1630#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1631#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1632 _CC_WORDCHAR, p)
1633#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1634
fbc19f27
KW
1635/* Macros for backwards compatibility and for completeness when the ASCII and
1636 * Latin1 values are identical */
fbc19f27
KW
1637#define isALPHAU(c) isALPHA_L1(c)
1638#define isDIGIT_L1(c) isDIGIT_A(c)
1639#define isOCTAL(c) isOCTAL_A(c)
1640#define isOCTAL_L1(c) isOCTAL_A(c)
1641#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1642#define isALNUM(c) isWORDCHAR(c)
1643#define isALNUMU(c) isWORDCHAR_L1(c)
1644#define isALNUM_LC(c) isWORDCHAR_LC(c)
1645#define isALNUM_uni(c) isWORDCHAR_uni(c)
1646#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1647#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1648#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1649#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1650#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1651#define isALNUMC(c) isALPHANUMERIC(c)
1652#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1653#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1654#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1655#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1656#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1657
2bd1cbf6
KW
1658/* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII,
1659 * except that they don't necessarily mean the same characters, e.g. CTRL-D is
1660 * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC.
1661 * '?' is special-cased on EBCDIC to APC, which is the control there that is
1662 * the outlier from the block that contains the other controls, just like
1663 * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0
1664 * block. If it weren't special cased, it would yield a non-control.
1665 * The conversion works both ways, so CTRL('D') is 4, and CTRL(4) is D, etc. */
1666#ifndef EBCDIC
1667# define toCTRL(c) (toUPPER(c) ^ 64)
1668#else
0ed2b00b
KW
1669# define toCTRL(c) ((isPRINT_A(c)) \
1670 ? UNLIKELY((c) == '?') \
1671 ? QUESTION_MARK_CTRL \
1672 : (NATIVE_TO_LATIN1(toUPPER(c)) ^ 64) \
1673 : UNLIKELY((c) == QUESTION_MARK_CTRL) \
1674 ? ((c) == '?') \
1675 : (LATIN1_TO_NATIVE((c) ^ 64)))
2bd1cbf6 1676#endif
bbce6d69 1677
dea28490
JJ
1678/* Line numbers are unsigned, 32 bits. */
1679typedef U32 line_t;
e5dcd934 1680#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1681
91152fc1
DG
1682/* Helpful alias for version prescan */
1683#define is_LAX_VERSION(a,b) \
1684 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1685
1686#define is_STRICT_VERSION(a,b) \
1687 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1688
1689#define BADVERSION(a,b,c) \
1690 if (b) { \
1691 *b = c; \
1692 } \
1693 return a;
8c52afec 1694
cb27eebd
KW
1695/* Converts a character known to represent a hexadecimal digit (0-9, A-F, or
1696 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer,
1697 * which is advanced. The input is validated only by an assert() in DEBUGGING
1698 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and
1699 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */
1700#define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \
1701 ? (c) \
1702 : ((c) + 9))))
902008b8
KW
1703#define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \
1704 ? (*(s)++) \
1705 : (*(s)++ + 9))))
95a59cab 1706
cb27eebd
KW
1707/* Converts a character known to represent an octal digit (0-7) to its numeric
1708 * value. The input is validated only by an assert() in DEBUGGING builds. In
1709 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */
1710#define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c)))
1711
305b8651
KW
1712/* Efficiently returns a boolean as to if two native characters are equivalent
1713 * case-insenstively. At least one of the characters must be one of [A-Za-z];
1714 * the ALPHA in the name is to remind you of that. This is asserted() in
1715 * DEBUGGING builds. Because [A-Za-z] are invariant under UTF-8, this macro
1716 * works (on valid input) for both non- and UTF-8-encoded bytes.
1717 *
1718 * When one of the inputs is a compile-time constant and gets folded by the
1719 * compiler, this reduces to an AND and a TEST. On both EBCDIC and ASCII
1720 * machines, 'A' and 'a' differ by a single bit; the same with the upper and
1721 * lower case of all other ASCII-range alphabetics. On ASCII platforms, they
96ca48da
KW
1722 * are 32 apart; on EBCDIC, they are 64. At compile time, this uses an
1723 * exclusive 'or' to find that bit and then inverts it to form a mask, with
1724 * just a single 0, in the bit position where the upper- and lowercase differ.
1725 * */
305b8651
KW
1726#define isALPHA_FOLD_EQ(c1, c2) \
1727 (__ASSERT_(isALPHA_A(c1) || isALPHA_A(c2)) \
1728 ((c1) & ~('A' ^ 'a')) == ((c2) & ~('A' ^ 'a')))
1729#define isALPHA_FOLD_NE(c1, c2) (! isALPHA_FOLD_EQ((c1), (c2)))
1730
8e84507e 1731/*
ccfc67b7
JH
1732=head1 Memory Management
1733
a02a5408 1734=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1735The XSUB-writer's interface to the C C<malloc> function.
1736
0d7b2759
MH
1737Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1738
c5008215
JC
1739In 5.9.3, Newx() and friends replace the older New() API, and drops
1740the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1741themselves. This aid has been superseded by a new build option,
d10b4965 1742PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1743there for use in XS modules supporting older perls.
1744
a02a5408 1745=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1746The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1747cast. See also C<Newx>.
954c1994 1748
0d7b2759
MH
1749Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1750
a02a5408 1751=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1752The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1753memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1754
0d7b2759
MH
1755Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1756
954c1994
GS
1757=for apidoc Am|void|Renew|void* ptr|int nitems|type
1758The XSUB-writer's interface to the C C<realloc> function.
1759
0d7b2759
MH
1760Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1761
954c1994
GS
1762=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1763The XSUB-writer's interface to the C C<realloc> function, with
1764cast.
1765
0d7b2759
MH
1766Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1767
49b8b560 1768=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1769The XSUB-writer's interface to the C C<free> function.
1770
0d7b2759
MH
1771This should B<ONLY> be used on memory obtained using L<"Newx"> and friends.
1772
954c1994
GS
1773=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1774The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1775source, C<dest> is the destination, C<nitems> is the number of items, and
1776C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1777
e90e2364 1778=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
72d33970
FC
1779Like C<Move> but returns dest. Useful
1780for encouraging compilers to tail-call
e90e2364
NC
1781optimise.
1782
954c1994
GS
1783=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1784The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1785source, C<dest> is the destination, C<nitems> is the number of items, and
1786C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1787
e90e2364
NC
1788=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1789
72d33970
FC
1790Like C<Copy> but returns dest. Useful
1791for encouraging compilers to tail-call
e90e2364
NC
1792optimise.
1793
954c1994
GS
1794=for apidoc Am|void|Zero|void* dest|int nitems|type
1795
1796The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1797destination, C<nitems> is the number of items, and C<type> is the type.
1798
e90e2364
NC
1799=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1800
72d33970
FC
1801Like C<Zero> but returns dest. Useful
1802for encouraging compilers to tail-call
e90e2364
NC
1803optimise.
1804
da5d8dbb 1805=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1806This is an architecture-independent macro to copy one structure to another.
954c1994 1807
7e337ee0
JH
1808=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1809
1810Fill up memory with a byte pattern (a byte repeated over and over
1811again) that hopefully catches attempts to access uninitialized memory.
1812
1813=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1814
1815PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1816
1c12ffb4 1817=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1818
1819PoisonWith(0xEF) for catching access to freed memory.
1820
9965345d
JH
1821=for apidoc Am|void|Poison|void* dest|int nitems|type
1822
7e337ee0 1823PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1824
1825=cut */
954c1994 1826
561b68a9
SH
1827/* Maintained for backwards-compatibility only. Use newSV() instead. */
1828#ifndef PERL_CORE
ff06c60c 1829#define NEWSV(x,len) newSV(len)
561b68a9 1830#endif
ff06c60c 1831
19a94d75
JH
1832#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1833
2720f793
JH
1834/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1835 * overly eager compilers that will bleat about e.g.
1836 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1837#ifdef PERL_MALLOC_WRAP
f019c49e 1838#define MEM_WRAP_CHECK(n,t) \
5d9574c1 1839 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (croak_memory_wrap(),0))
27d5b266 1840#define MEM_WRAP_CHECK_1(n,t,a) \
5d9574c1 1841 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1842#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1843
d1decf2b 1844#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
1845#else
1846
410319be
NC
1847#define MEM_WRAP_CHECK(n,t)
1848#define MEM_WRAP_CHECK_1(n,t,a)
1849#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1850#define MEM_WRAP_CHECK_(n,t)
1851
6fff79ce 1852#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1853
1936d2a7 1854#endif
8b44ba4c 1855
fe4f188c 1856#ifdef PERL_MEM_LOG
46c6c7e2 1857/*
9f653bb5 1858 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1859 * go through functions, which are handy for debugging breakpoints, but
1860 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1861 * line number, and C function name if available) passed in. This info can
1862 * then be used for logging the calls, for which one gets a sample
73d1d973 1863 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1864 *
46c6c7e2 1865 * Known problems:
94e892a6 1866 * - not all memory allocs get logged, only those
46c6c7e2 1867 * that go through Newx() and derivatives (while all
94e892a6 1868 * Safefrees do get logged)
46c6c7e2
JH
1869 * - __FILE__ and __LINE__ do not work everywhere
1870 * - __func__ or __FUNCTION__ even less so
1871 * - I think more goes on after the perlio frees but
1872 * the thing is that STDERR gets closed (as do all
1873 * the file descriptors)
1874 * - no deeper calling stack than the caller of the Newx()
1875 * or the kind, but do I look like a C reflection/introspection
1876 * utility to you?
1877 * - the function prototypes for the logging functions
1878 * probably should maybe be somewhere else than handy.h
1879 * - one could consider inlining (macrofying) the logging
1880 * for speed, but I am too lazy
1881 * - one could imagine recording the allocations in a hash,
1882 * (keyed by the allocation address?), and maintain that
1883 * through reallocs and frees, but how to do that without
1884 * any News() happening...?
73d1d973 1885 * - lots of -Ddefines to get useful/controllable output
b953482e 1886 * - lots of ENV reads
46c6c7e2
JH
1887 */
1888
12754f92 1889PERL_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 1890
12754f92 1891PERL_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 1892
12754f92 1893PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1894
0b0ab801 1895# ifdef PERL_CORE
73d1d973 1896# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1897enum mem_log_type {
1898 MLT_ALLOC,
1899 MLT_REALLOC,
d7a2c63c
MHM
1900 MLT_FREE,
1901 MLT_NEW_SV,
1902 MLT_DEL_SV
0b0ab801
MHM
1903};
1904# endif
12754f92 1905# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1906void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1907void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1908# endif
0b0ab801
MHM
1909# endif
1910
fe4f188c
JH
1911#endif
1912
1913#ifdef PERL_MEM_LOG
d1401ee9
MHM
1914#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1915#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1916#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1917#endif
1918
1919#ifndef MEM_LOG_ALLOC
1920#define MEM_LOG_ALLOC(n,t,a) (a)
1921#endif
1922#ifndef MEM_LOG_REALLOC
1923#define MEM_LOG_REALLOC(n,t,v,a) (a)
1924#endif
1925#ifndef MEM_LOG_FREE
1926#define MEM_LOG_FREE(a) (a)
1927#endif
1928
d1401ee9
MHM
1929#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1930#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1931#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1932
1933#ifndef PERL_CORE
a02a5408
JC
1934/* pre 5.9.x compatibility */
1935#define New(x,v,n,t) Newx(v,n,t)
1936#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1937#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1938#endif
a02a5408 1939
ff68c719 1940#define Renew(v,n,t) \
d1401ee9 1941 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1942#define Renewc(v,n,t,c) \
d1401ee9 1943 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1944
1945#ifdef PERL_POISON
1946#define Safefree(d) \
06c0cc96 1947 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1948#else
fe4f188c 1949#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1950#endif
55497cff 1951
8b44ba4c
NC
1952#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1953#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1954#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1955
8b44ba4c
NC
1956#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1957#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1958#ifdef HAS_MEMSET
8b44ba4c 1959#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1960#else
8b44ba4c
NC
1961/* Using bzero(), which returns void. */
1962#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1963#endif
1964
7e337ee0
JH
1965#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1966#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1967#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1968#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1969
caa674f3
DD
1970#ifdef PERL_POISON
1971# define PERL_POISON_EXPR(x) x
1972#else
1973# define PERL_POISON_EXPR(x)
1974#endif
1975
2304df62 1976#ifdef USE_STRUCT_COPY
ff68c719 1977#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1978#else
1979#define StructCopy(s,d,t) Copy(s,d,1,t)
1980#endif
2cc61e15 1981
c3caa5c3
JH
1982/* C_ARRAY_LENGTH is the number of elements in the C array (so you
1983 * want your zero-based indices to be less than but not equal to).
1984 *
1985 * C_ARRAY_END is one past the last: half-open/half-closed range,
1986 * not last-inclusive range. */
622913ab 1987#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
c3caa5c3 1988#define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a))
622913ab 1989
2cc61e15
DD
1990#ifdef NEED_VA_COPY
1991# ifdef va_copy
1992# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1993# else
a1866d1b
JH
1994# if defined(__va_copy)
1995# define Perl_va_copy(s, d) __va_copy(d, s)
1996# else
1997# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1998# endif
2cc61e15
DD
1999# endif
2000#endif
2001
472d47bc
SB
2002/* convenience debug macros */
2003#ifdef USE_ITHREADS
2004#define pTHX_FORMAT "Perl interpreter: 0x%p"
2005#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
2006#define pTHX_VALUE_ (void *)my_perl,
2007#define pTHX_VALUE (void *)my_perl
2008#define pTHX__VALUE_ ,(void *)my_perl,
2009#define pTHX__VALUE ,(void *)my_perl
472d47bc 2010#else
3609ea0d 2011#define pTHX_FORMAT
472d47bc 2012#define pTHX__FORMAT
3609ea0d 2013#define pTHX_VALUE_
472d47bc 2014#define pTHX_VALUE
3609ea0d 2015#define pTHX__VALUE_
472d47bc
SB
2016#define pTHX__VALUE
2017#endif /* USE_ITHREADS */
3609ea0d 2018
2acdbac1
NC
2019/* Perl_deprecate was not part of the public API, and did not have a deprecate()
2020 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
2021 CPAN::Unpack show any users outside the core. */
2022#ifdef PERL_CORE
d1d15184 2023# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
2024#endif
2025
dfff4baf
BF
2026/* Internal macros to deal with gids and uids */
2027#ifdef PERL_CORE
2028
2029# if Uid_t_size > IVSIZE
2030# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
2031# define SvUID(sv) SvNV(sv)
2032# else
2033# if Uid_t_sign <= 0
2034# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
2035# define SvUID(sv) SvIV(sv)
2036# else
2037# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
2038# define SvUID(sv) SvUV(sv)
2039# endif
2040# endif /* Uid_t_size */
2041
2042# if Gid_t_size > IVSIZE
2043# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
2044# define SvGID(sv) SvNV(sv)
2045# else
2046# if Gid_t_sign <= 0
2047# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
2048# define SvGID(sv) SvIV(sv)
2049# else
2050# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
2051# define SvGID(sv) SvUV(sv)
2052# endif
2053# endif /* Gid_t_size */
2054
2055#endif
2056
9d745869
KW
2057#endif /* HANDY_H */
2058
e9a8c099
MHM
2059/*
2060 * Local variables:
2061 * c-indentation-style: bsd
2062 * c-basic-offset: 4
14d04a33 2063 * indent-tabs-mode: nil
e9a8c099
MHM
2064 * End:
2065 *
14d04a33 2066 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 2067 */