This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
[MERGE] Handle /@a/ array expansion within regex engine
[perl5.git] / handy.h
CommitLineData
a0d0e21e 1/* handy.h
a687059c 2 *
1129b882 3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000,
da5d8dbb 4 * 2001, 2002, 2004, 2005, 2006, 2007, 2008, 2012 by Larry Wall and others
a687059c 5 *
6e21c824
LW
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
8d063cd8 8 *
8d063cd8
LW
9 */
10
9d745869
KW
11#ifndef HANDY_H /* Guard against nested #inclusion */
12#define HANDY_H
13
85e6fe83 14#if !defined(__STDC__)
378cc40b
LW
15#ifdef NULL
16#undef NULL
17#endif
a687059c 18# define NULL 0
85e6fe83
LW
19#endif
20
24792b8d
NC
21#ifndef PERL_CORE
22# define Null(type) ((type)NULL)
954c1994
GS
23
24/*
ccfc67b7 25=head1 Handy Values
954c1994 26
3609ea0d 27=for apidoc AmU||Nullch
24792b8d 28Null character pointer. (No longer available when C<PERL_CORE> is defined.)
2307c6d0 29
954c1994 30=for apidoc AmU||Nullsv
24792b8d 31Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
32
33=cut
34*/
35
24792b8d
NC
36# define Nullch Null(char*)
37# define Nullfp Null(PerlIO*)
38# define Nullsv Null(SV*)
39#endif
8d063cd8 40
641d3f0b
PP
41#ifdef TRUE
42#undef TRUE
43#endif
44#ifdef FALSE
45#undef FALSE
46#endif
47#define TRUE (1)
48#define FALSE (0)
49
cf3f0ffb
DM
50/* The MUTABLE_*() macros cast pointers to the types shown, in such a way
51 * (compiler permitting) that casting away const-ness will give a warning;
52 * e.g.:
53 *
54 * const SV *sv = ...;
55 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
56 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
57 */
58
b1bc3f34 59#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 60# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
61#else
62# define MUTABLE_PTR(p) ((void *) (p))
63#endif
64
a062e10d 65#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 66#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 67#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 68#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 69#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 70#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 71
bd31be4b
NC
72#ifdef I_STDBOOL
73# include <stdbool.h>
74# ifndef HAS_BOOL
75# define HAS_BOOL 1
76# endif
77#endif
78
8e84507e 79/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
80 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
81 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 82 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
83 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
84 g++ can be identified by __GNUG__.
85 Andy Dougherty February 2000
5d94fbed 86*/
3609ea0d 87#ifdef __GNUG__ /* GNU g++ has bool built-in */
5d94fbed 88# ifndef HAS_BOOL
c1d22f6b 89# define HAS_BOOL 1
5d94fbed 90# endif
5d94fbed
AD
91#endif
92
641d3f0b
PP
93/* The NeXT dynamic loader headers will not build with the bool macro
94 So declare them now to clear confusion.
95*/
8f1f23e8 96#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
97# undef FALSE
98# undef TRUE
99 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
100# define ENUM_BOOL 1
101# ifndef HAS_BOOL
102# define HAS_BOOL 1
103# endif /* !HAS_BOOL */
8f1f23e8 104#endif /* NeXT || __NeXT__ */
641d3f0b 105
5d94fbed 106#ifndef HAS_BOOL
cc3315ba 107# if defined(VMS)
5d94fbed
AD
108# define bool int
109# else
110# define bool char
111# endif
c1d22f6b 112# define HAS_BOOL 1
a687059c 113#endif
0d3e774c 114
f2338a2e
DM
115/* a simple (bool) cast may not do the right thing: if bool is defined
116 * as char for example, then the cast from int is implementation-defined
18f5643b 117 * (bool)!!(cbool) in a ternary triggers a bug in xlc on AIX
f2338a2e
DM
118 */
119
18f5643b 120#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 121
46c6c7e2 122/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
123 * XXX Should really be a Configure probe, with HAS__FUNCTION__
124 * and FUNCTION__ as results.
125 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
126#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
127# define FUNCTION__ __func__
128#else
70764c00 129# if (defined(_MSC_VER) && _MSC_VER < 1300) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 130 (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
131# define FUNCTION__ ""
132# else
133# define FUNCTION__ __FUNCTION__ /* Common extension. */
134# endif
135#endif
136
27d4fb96
PP
137/* XXX A note on the perl source internal type system. The
138 original intent was that I32 be *exactly* 32 bits.
139
140 Currently, we only guarantee that I32 is *at least* 32 bits.
141 Specifically, if int is 64 bits, then so is I32. (This is the case
142 for the Cray.) This has the advantage of meshing nicely with
143 standard library calls (where we pass an I32 and the library is
144 expecting an int), but the disadvantage that an I32 is not 32 bits.
145 Andy Dougherty August 1996
24fef2a7 146
dc45a647
MB
147 There is no guarantee that there is *any* integral type with
148 exactly 32 bits. It is perfectly legal for a system to have
149 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 150
dc45a647
MB
151 Similarly, there is no guarantee that I16 and U16 have exactly 16
152 bits.
693762b4 153
8e84507e
NIS
154 For dealing with issues that may arise from various 32/64-bit
155 systems, we will ask Configure to check out
8175356b 156
3609ea0d
JH
157 SHORTSIZE == sizeof(short)
158 INTSIZE == sizeof(int)
159 LONGSIZE == sizeof(long)
dc45a647 160 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 161 PTRSIZE == sizeof(void *)
dc45a647
MB
162 DOUBLESIZE == sizeof(double)
163 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 164
27d4fb96
PP
165*/
166
69512466
JH
167#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
168# include <inttypes.h>
dd0eed91
JH
169# ifdef INT32_MIN_BROKEN
170# undef INT32_MIN
171# define INT32_MIN (-2147483647-1)
172# endif
173# ifdef INT64_MIN_BROKEN
174# undef INT64_MIN
175# define INT64_MIN (-9223372036854775807LL-1)
176# endif
69512466
JH
177#endif
178
8175356b
JH
179typedef I8TYPE I8;
180typedef U8TYPE U8;
181typedef I16TYPE I16;
182typedef U16TYPE U16;
183typedef I32TYPE I32;
184typedef U32TYPE U32;
6b8eaf93
JH
185#ifdef PERL_CORE
186# ifdef HAS_QUAD
8175356b
JH
187typedef I64TYPE I64;
188typedef U64TYPE U64;
6b8eaf93
JH
189# endif
190#endif /* PERL_CORE */
8175356b 191
69512466 192#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
80524f33
JL
193# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
194# define PeRl_INT64_C(c) CAT2(c,LL)
195# define PeRl_UINT64_C(c) CAT2(c,ULL)
196# else
52865553
SH
197# if QUADKIND == QUAD_IS___INT64
198# define PeRl_INT64_C(c) CAT2(c,I64)
199# define PeRl_UINT64_C(c) CAT2(c,UI64)
69512466 200# else
52865553
SH
201# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
202# define PeRl_INT64_C(c) CAT2(c,L)
203# define PeRl_UINT64_C(c) CAT2(c,UL)
69512466 204# else
80524f33
JL
205# define PeRl_INT64_C(c) ((I64TYPE)(c))
206# define PeRl_UINT64_C(c) ((U64TYPE)(c))
69512466
JH
207# endif
208# endif
e8c95190 209# endif
80524f33
JL
210# ifndef UINT64_C
211# define UINT64_C(c) PeRl_UINT64_C(c)
212# endif
213# ifndef INT64_C
214# define INT64_C(c) PeRl_INT64_C(c)
215# endif
e8c95190 216#endif
dc750f2e 217
d8668976 218#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 219
5ff3f7a4
GS
220/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
221 Please search CHAR_MAX in perl.h for further details. */
222#define U8_MAX UINT8_MAX
223#define U8_MIN UINT8_MIN
224
5ff3f7a4
GS
225#define I16_MAX INT16_MAX
226#define I16_MIN INT16_MIN
227#define U16_MAX UINT16_MAX
228#define U16_MIN UINT16_MIN
229
5ff3f7a4
GS
230#define I32_MAX INT32_MAX
231#define I32_MIN INT32_MIN
0e983133
GS
232#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
233# define U32_MAX UINT32_MAX
234#else
235# define U32_MAX 4294967295U
236#endif
5ff3f7a4
GS
237#define U32_MIN UINT32_MIN
238
239#else
240
5c9fa16e
KA
241/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
242 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
243#define U8_MAX PERL_UCHAR_MAX
244#define U8_MIN PERL_UCHAR_MIN
79072805 245
27d4fb96
PP
246#define I16_MAX PERL_SHORT_MAX
247#define I16_MIN PERL_SHORT_MIN
248#define U16_MAX PERL_USHORT_MAX
249#define U16_MIN PERL_USHORT_MIN
79072805 250
c4f23d77 251#if LONGSIZE > 4
27d4fb96
PP
252# define I32_MAX PERL_INT_MAX
253# define I32_MIN PERL_INT_MIN
254# define U32_MAX PERL_UINT_MAX
255# define U32_MIN PERL_UINT_MIN
79072805 256#else
27d4fb96
PP
257# define I32_MAX PERL_LONG_MAX
258# define I32_MIN PERL_LONG_MIN
259# define U32_MAX PERL_ULONG_MAX
260# define U32_MIN PERL_ULONG_MIN
79072805
LW
261#endif
262
5ff3f7a4
GS
263#endif
264
58a9a5d5 265/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
266#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
267#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
268#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
269
ff68c719 270#define Ctl(ch) ((ch) & 037)
8d063cd8 271
3fe05580
MHM
272/*
273=head1 SV-Body Allocation
274
275=for apidoc Ama|SV*|newSVpvs|const char* s
276Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
277
84bafc02
NC
278=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
279Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
280pair.
281
3fe05580
MHM
282=for apidoc Ama|SV*|newSVpvs_share|const char* s
283Like C<newSVpvn_share>, but takes a literal string instead of a string/length
284pair and omits the hash parameter.
285
9dcc53ea
Z
286=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
287Like C<sv_catpvn_flags>, but takes a literal string instead of a
288string/length pair.
289
290=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
291Like C<sv_catpvn_nomg>, but takes a literal string instead of a
292string/length pair.
293
7b4e7469 294=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
295Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
296
9dcc53ea
Z
297=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
298Like C<sv_catpvn_mg>, but takes a literal string instead of a
299string/length pair.
300
7b4e7469 301=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
302Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
303
9dcc53ea
Z
304=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
305Like C<sv_setpvn_mg>, but takes a literal string instead of a
306string/length pair.
307
308=for apidoc Am|SV *|sv_setref_pvs|const char* s
309Like C<sv_setref_pvn>, but takes a literal string instead of a
310string/length pair.
311
3fe05580
MHM
312=head1 Memory Management
313
314=for apidoc Ama|char*|savepvs|const char* s
315Like C<savepvn>, but takes a literal string instead of a string/length pair.
316
9dcc53ea
Z
317=for apidoc Ama|char*|savesharedpvs|const char* s
318A version of C<savepvs()> which allocates the duplicate string in memory
319which is shared between threads.
320
3fe05580
MHM
321=head1 GV Functions
322
323=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
324Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
325
326=head1 Hash Manipulation Functions
327
328=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
329Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
330
331=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
332Like C<hv_store>, but takes a literal string instead of a string/length pair
333and omits the hash parameter.
334
510966aa
Z
335=head1 Lexer interface
336
337=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
338
339Like L</lex_stuff_pvn>, but takes a literal string instead of a
340string/length pair.
341
3fe05580
MHM
342=cut
343*/
344
926bb54c
KW
345/* concatenating with "" ensures that only literal strings are accepted as
346 * argument */
26bfd924 347#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 348
926bb54c
KW
349/* note that STR_WITH_LEN() can't be used as argument to macros or functions
350 * that under some configurations might be macros, which means that it requires
351 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
352 */
353
354/* STR_WITH_LEN() shortcuts */
355#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
356#define newSVpvs_flags(str,flags) \
357 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 358#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
359#define sv_catpvs_flags(sv, str, flags) \
360 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
361#define sv_catpvs_nomg(sv, str) \
362 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
363#define sv_catpvs(sv, str) \
364 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
365#define sv_catpvs_mg(sv, str) \
366 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 367#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
368#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
369#define sv_setref_pvs(rv, classname, str) \
370 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 371#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
372#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
373#define gv_stashpvs(str, create) \
374 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
375#define gv_fetchpvs(namebeg, add, sv_type) \
376 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
377#define gv_fetchpvn(namebeg, len, add, sv_type) \
378 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
379#define sv_catxmlpvs(dsv, str, utf8) \
380 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
381#define hv_fetchs(hv,key,lval) \
382 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
383 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
384 : HV_FETCH_JUST_SV, NULL, 0))
385
386#define hv_stores(hv,key,val) \
387 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
388 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 389
510966aa
Z
390#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
391
b96d8cd9
NC
392#define get_cvs(str, flags) \
393 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 394
954c1994 395/*
ccfc67b7
JH
396=head1 Miscellaneous Functions
397
954c1994
GS
398=for apidoc Am|bool|strNE|char* s1|char* s2
399Test two strings to see if they are different. Returns true or
400false.
401
402=for apidoc Am|bool|strEQ|char* s1|char* s2
403Test two strings to see if they are equal. Returns true or false.
404
405=for apidoc Am|bool|strLT|char* s1|char* s2
406Test two strings to see if the first, C<s1>, is less than the second,
407C<s2>. Returns true or false.
408
409=for apidoc Am|bool|strLE|char* s1|char* s2
410Test two strings to see if the first, C<s1>, is less than or equal to the
411second, C<s2>. Returns true or false.
412
413=for apidoc Am|bool|strGT|char* s1|char* s2
414Test two strings to see if the first, C<s1>, is greater than the second,
415C<s2>. Returns true or false.
416
417=for apidoc Am|bool|strGE|char* s1|char* s2
418Test two strings to see if the first, C<s1>, is greater than or equal to
419the second, C<s2>. Returns true or false.
420
421=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
422Test two strings to see if they are different. The C<len> parameter
423indicates the number of bytes to compare. Returns true or false. (A
424wrapper for C<strncmp>).
425
426=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
427Test two strings to see if they are equal. The C<len> parameter indicates
428the number of bytes to compare. Returns true or false. (A wrapper for
429C<strncmp>).
430
431=cut
432*/
433
8d063cd8
LW
434#define strNE(s1,s2) (strcmp(s1,s2))
435#define strEQ(s1,s2) (!strcmp(s1,s2))
436#define strLT(s1,s2) (strcmp(s1,s2) < 0)
437#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
438#define strGT(s1,s2) (strcmp(s1,s2) > 0)
439#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
440#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
441#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 442
36477c24
PP
443#ifdef HAS_MEMCMP
444# define memNE(s1,s2,l) (memcmp(s1,s2,l))
445# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
446#else
447# define memNE(s1,s2,l) (bcmp(s1,s2,l))
448# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
449#endif
450
568a785a 451#define memEQs(s1, l, s2) \
8e6497b6 452 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
453#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
454
bbce6d69
PP
455/*
456 * Character classes.
457 *
458 * Unfortunately, the introduction of locales means that we
459 * can't trust isupper(), etc. to tell the truth. And when
460 * it comes to /\w+/ with tainting enabled, we *must* be able
461 * to trust our character classes.
462 *
463 * Therefore, the default tests in the text of Perl will be
464 * independent of locale. Any code that wants to depend on
465 * the current locale will use the tests that begin with "lc".
466 */
467
2304df62
AD
468#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
469# ifndef CTYPE256
470# define CTYPE256
471# endif
472#endif
473
954c1994 474/*
ccfc67b7
JH
475
476=head1 Character classes
243effed
KW
477This section is about functions (really macros) that classify characters
478into types, such as punctuation versus alphabetic, etc. Most of these are
479analogous to regular expression character classes. (See
480L<perlrecharclass/POSIX Character Classes>.) There are several variants for
481each class. (Not all macros have all variants; each item below lists the
482ones valid for it.) None are affected by C<use bytes>, and only the ones
483with C<LC> in the name are affected by the current locale.
484
485The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
486C<U8>) as input and returns a boolean as to whether or not the character
487represented by that octet is in the named class based on platform, Unicode, and
488Perl rules. If the input is a number that doesn't fit in an octet, FALSE is
489always returned.
490
491Variant C<isFOO_A> (e.g., C<isALPHA_A()>) will return TRUE only if the input is
4b9734bf
KW
492also in the ASCII character set. The base function with no suffix and the one
493with the C<_A> suffix are identical.
494
495Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
496onto the platform. That is, the code points that are ASCII are unaffected,
497since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
498as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
499true when called with the code point 0xDF, which is a word character in both
500ASCII and EBCDIC (though it represent different characters in each).
243effed
KW
501
502Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
503point as input. If the code point is larger than 255, Unicode rules are used
504to determine if it is in the character class. For example,
505C<isWORDCHAR(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A WITH
506MACRON in Unicode, and is a word character.
507
508Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
509(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
510classification of just the first character in the string is tested.
511
512Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but uses
513the C library function that gives the named classification instead of
514hard-coded rules. For example, C<isDIGIT_LC()> returns the result of calling
515C<isdigit()>. This means that the result is based on the current locale, which
516is what C<LC> in the name stands for. FALSE is always returned if the input
517won't fit into an octet.
518
519Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
520returns the same as C<isFOO_LC> for input code points less than 256, and
521returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
522
523Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
524(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
525classification of just the first character in the string is tested.
ccfc67b7 526
954c1994 527=for apidoc Am|bool|isALPHA|char ch
2787a470 528Returns a boolean indicating whether the specified character is an
243effed
KW
529alphabetic character in the platform's native character set, analogous to
530C<m/[[:alpha:]]/>.
8a58bdcf 531See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 532C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 533C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 534
15861f94
KW
535=for apidoc Am|bool|isALPHANUMERIC|char ch
536Returns a boolean indicating whether the specified character is a either an
537alphabetic character or decimal digit in the platform's native character set,
538analogous to C<m/[[:alnum:]]/>.
539See the L<top of this section|/Character classes> for an explanation of variants
540C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 541C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
542C<isALPHANUMERIC_LC_utf8>.
543
8a58bdcf
KW
544=for apidoc Am|bool|isASCII|char ch
545Returns a boolean indicating whether the specified character is one of the 128
243effed 546characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 547On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
548character corresponds to an ASCII character. Variants C<isASCII_A()> and
549C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
550See the L<top of this section|/Character classes> for an explanation of variants
551C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
552C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
553library routine C<isascii()>. In these cases, the variants whose names contain
554C<LC> are the same as the corresponding ones without.
555
556=for apidoc Am|bool|isBLANK|char ch
557Returns a boolean indicating whether the specified character is a
558character considered to be a blank in the platform's native character set,
559analogous to C<m/[[:blank:]]/>.
560See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 561C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
562C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
563platforms do not have the C library routine C<isblank()>. In these cases, the
564variants whose names contain C<LC> are the same as the corresponding ones
565without.
566
567=for apidoc Am|bool|isCNTRL|char ch
568Returns a boolean indicating whether the specified character is a
569control character in the platform's native character set,
570analogous to C<m/[[:cntrl:]]/>.
571See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 572C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
243effed 573C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>.
954c1994 574
954c1994 575=for apidoc Am|bool|isDIGIT|char ch
2787a470 576Returns a boolean indicating whether the specified character is a
243effed 577digit in the platform's native character set, analogous to C<m/[[:digit:]]/>.
8a58bdcf 578Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed 579See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 580C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
581C<isDIGIT_LC_utf8>.
582
583=for apidoc Am|bool|isGRAPH|char ch
584Returns a boolean indicating whether the specified character is a
585graphic character in the platform's native character set, analogous to
586C<m/[[:graph:]]/>.
587See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 588C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 589C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 590
0c82b6df 591=for apidoc Am|bool|isLOWER|char ch
2787a470 592Returns a boolean indicating whether the specified character is a
243effed
KW
593lowercase character in the platform's native character set, analogous to
594C<m/[[:lower:]]/>.
8a58bdcf 595See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 596C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 597C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 598
c99e91e9 599=for apidoc Am|bool|isOCTAL|char ch
2787a470 600Returns a boolean indicating whether the specified character is an
06dc89af 601octal digit, [0-7] in the platform's native character set.
243effed
KW
602The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
603C<isOCTAL>.
604
605=for apidoc Am|bool|isPUNCT|char ch
606Returns a boolean indicating whether the specified character is a
607punctuation character in the platform's native character set, analogous to
608C<m/[[:punct:]]/>. Note that the definition of what is punctuation isn't as
609straightforward as one might desire. See L<perlrecharclass/POSIX Character
610Classes> for details.
611See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 612C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 613C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 614
0c82b6df 615=for apidoc Am|bool|isSPACE|char ch
2787a470 616Returns a boolean indicating whether the specified character is a
243effed 617whitespace character in the platform's native character set. This is analogous
398d098a
KW
618to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
619(experimentally), this also matches what C<m/[[:space:]]/> does.
620("Experimentally" means that this change may be backed out in 5.20 or 5.22 if
621field experience indicates that it was unwise.) Prior to 5.18, only the
622locale forms of this macro (the ones with C<LC> in their names) matched
623precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
624in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
625(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
8a58bdcf 626See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 627C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 628C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 629
398d098a
KW
630=for apidoc Am|bool|isPSXSPC|char ch
631(short for Posix Space)
632Starting in 5.18, this is identical (experimentally) in all its forms to the
633corresponding C<isSPACE()> macros. ("Experimentally" means that this change
634may be backed out in 5.20 or 5.22 if field experience indicates that it
635was unwise.)
636The locale forms of this macro are identical to their corresponding
637C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
638non-locale forms differ from their C<isSPACE()> forms only in that the
639C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
640Otherwise they are identical. Thus this macro is analogous to what
641C<m/[[:space:]]/> matches in a regular expression.
642See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 643C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
644C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
645
954c1994 646=for apidoc Am|bool|isUPPER|char ch
2787a470 647Returns a boolean indicating whether the specified character is an
243effed
KW
648uppercase character in the platform's native character set, analogous to
649C<m/[[:upper:]]/>.
8a58bdcf 650See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 651C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 652C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 653
243effed 654=for apidoc Am|bool|isPRINT|char ch
8eea39dd 655Returns a boolean indicating whether the specified character is a
243effed
KW
656printable character in the platform's native character set, analogous to
657C<m/[[:print:]]/>.
8a58bdcf 658See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 659C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
660C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
661
662=for apidoc Am|bool|isWORDCHAR|char ch
663Returns a boolean indicating whether the specified character is a character
664that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
665in a regular expression. A word character is an alphabetic character, a
666decimal digit, a connecting punctuation character (such as an underscore), or
667a "mark" character that attaches to one of those (like some sort of accent).
668C<isALNUM()> is a synonym provided for backward compatibility, even though a
669word character includes more than the standard C language meaning of
670alphanumeric.
671See the L<top of this section|/Character classes> for an explanation of variants
672C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
673C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
674
675=for apidoc Am|bool|isXDIGIT|char ch
676Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
677digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
678and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
679See the L<top of this section|/Character classes> for an explanation of variants
680C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
681C<isXDIGIT_LC_utf8>.
682
3c3ecf18
KW
683=for apidoc Am|bool|isIDFIRST|char ch
684Returns a boolean indicating whether the specified character can be the first
685character of an identifier. This is very close to, but not quite the same as
686the official Unicode property C<XID_Start>. The difference is that this
687returns true only if the input character also matches L</isWORDCHAR>.
688See the L<top of this section|/Character classes> for an explanation of variants
689C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
690C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
691
692=for apidoc Am|bool|isIDCONT|char ch
693Returns a boolean indicating whether the specified character can be the
694second or succeeding character of an identifier. This is very close to, but
695not quite the same as the official Unicode property C<XID_Continue>. The
696difference is that this returns true only if the input character also matches
697L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
698explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
699C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
700C<isIDCONT_LC_utf8>.
701
243effed 702=head1 Miscellaneous Functions
8eea39dd 703
95a59cab 704=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 705Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
706Behaviour is only well defined when isXDIGIT(*str) is true.
707
e7c1e6c1
KW
708=head1 Character case changing
709
954c1994 710=for apidoc Am|char|toUPPER|char ch
06dc89af
KW
711Converts the specified character to uppercase in the platform's native
712character set, if possible; otherwise returns the input character itself.
954c1994
GS
713
714=for apidoc Am|char|toLOWER|char ch
06dc89af
KW
715Converts the specified character to lowercase in the platform's native
716character set, if possible; otherwise returns the input character itself.
954c1994
GS
717
718=cut
353c9b6f 719
3c3ecf18 720XXX Still undocumented the other toUPPER etc functions
243effed 721
8a58bdcf 722Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
723patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
724
954c1994
GS
725*/
726
7c062697
KW
727/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
728 * is known only in the perl core, and this macro can be called from outside
729 * that */
730#ifdef HAS_QUAD
731# define WIDEST_UTYPE U64TYPE
732#else
733# define WIDEST_UTYPE U32
734#endif
735
3912bc88
KW
736/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
737 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
738 * bits of information are lost even on a 64-bit machine, but to get the
739 * compiler to optimize it out if possible. This is because Configure makes
740 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
741 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
742 */
743#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 744
41f43cc2 745#ifdef EBCDIC
ff7eba81 746# define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128))
41f43cc2
KW
747#else
748# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
749#endif
750
ec79cfbb 751#define isASCII_A(c) isASCII(c)
3509bb53 752#define isASCII_L1(c) isASCII(c)
6515cb76 753
c2da0b36
KW
754/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
755 * and the 8 possible permutations of those bits exactly comprise the 8 octal
756 * digits */
3d33b3e3 757#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 758
ec79cfbb 759/* ASCII range only */
9fb1bf9d 760#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 761
a500dc72
KW
762/* Character class numbers. For internal core Perl use only. The ones less
763 * than 32 are used in PL_charclass[] and the ones up through the one that
764 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
765 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
766 * their actual definitions are here. If that file has a name not used here,
767 * it won't compile.
1709d539
KW
768 *
769 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
770 * frequency of their use. This gives a slight edge to exiting a loop earlier
771 * (in reginclass() in regexec.c) */
1709d539
KW
772# define _CC_WORDCHAR 0 /* \w and [:word:] */
773# define _CC_DIGIT 1 /* \d and [:digit:] */
774# define _CC_ALPHA 2 /* [:alpha:] */
775# define _CC_LOWER 3 /* [:lower:] */
776# define _CC_UPPER 4 /* [:upper:] */
777# define _CC_PUNCT 5 /* [:punct:] */
778# define _CC_PRINT 6 /* [:print:] */
15861f94 779# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 780# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 781# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 782
b0d691b2 783#define _FIRST_NON_SWASH_CC 10
1709d539
KW
784/* The character classes above are implemented with swashes. The second group
785 * (just below) contains the ones implemented without. These are also sorted
786 * in rough order of the frequency of their use, except that \v should be last,
787 * as it isn't a real Posix character class, and some (small) inefficiencies in
788 * regular expression handling would be introduced by putting it in the middle
789 * of those that are. Also, cntrl and ascii come after the others as it may be
790 * useful to group these which have no members that match above Latin1, (or
791 * above ASCII in the latter case) */
792
b0d691b2
KW
793# define _CC_SPACE 10 /* \s */
794# define _CC_BLANK 11 /* [:blank:] */
795# define _CC_XDIGIT 12 /* [:xdigit:] */
796# define _CC_PSXSPC 13 /* [:space:] */
797# define _CC_CNTRL 14 /* [:cntrl:] */
798# define _CC_ASCII 15 /* [:ascii:] */
799# define _CC_VERTSPACE 16 /* \v */
1709d539 800
a0947d7b
KW
801# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
802
1709d539
KW
803/* The members of the third group below do not need to be coordinated with data
804 * structures in regcomp.[ch] and regexec.c */
b0d691b2
KW
805# define _CC_IDFIRST 17
806# define _CC_CHARNAME_CONT 18
807# define _CC_NONLATIN1_FOLD 19
808# define _CC_QUOTEMETA 20
809# define _CC_NON_FINAL_FOLD 21
810# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
811# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
812/* Unused: 23-30
f4cdb42c
KW
813 * If more bits are needed, one could add a second word for non-64bit
814 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
815 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
816 * is used only for optimization (as of this writing), and differs in the
817 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
818 * characters: the masculine and feminine ordinal indicators, so removing it
819 * would just cause /i regexes which match them to run less efficiently */
96ac0975 820
3a371f2f
KW
821#if defined(PERL_CORE) || defined(PERL_EXT)
822/* An enum version of the character class numbers, to help compilers
823 * optimize */
824typedef enum {
3a371f2f 825 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
826 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
827 _CC_ENUM_ASCII = _CC_ASCII,
828 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 829 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 830 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
831 _CC_ENUM_DIGIT = _CC_DIGIT,
832 _CC_ENUM_GRAPH = _CC_GRAPH,
833 _CC_ENUM_LOWER = _CC_LOWER,
834 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 835 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 836 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 837 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 838 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 839 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 840 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 841 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
842} _char_class_number;
843#endif
844
63c61c3f 845#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 846#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 847
3018b823 848#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 849# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
850 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
851 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
852 #error Need to adjust order of swash_property_names[]
853# endif
854
b44daeeb
KW
855/* This is declared static in each of the few files that this is #defined for
856 * to keep them from being publicly accessible. Hence there is a small amount
857 * of wasted space */
858
922e8cb4
KW
859static const char* const swash_property_names[] = {
860 "XPosixWord",
861 "XPosixDigit",
862 "XPosixAlpha",
863 "XPosixLower",
864 "XPosixUpper",
865 "XPosixPunct",
866 "XPosixPrint",
867 "XPosixAlnum",
b0d691b2
KW
868 "XPosixGraph",
869 "Cased"
922e8cb4
KW
870};
871#endif
872
96ac0975
NC
873# ifdef DOINIT
874EXTCONST U32 PL_charclass[] = {
875# include "l1_char_class_tab.h"
876};
877
878# else /* ! DOINIT */
879EXTCONST U32 PL_charclass[];
880# endif
881
265c1f46 882 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70
KW
883# define _CC_mask(classnum) (1U << (classnum))
884# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
926bb54c 885 && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask(classnum)))
4eeeb416 886
f4cdb42c
KW
887 /* The mask for the _A versions of the macros; it just adds in the bit for
888 * ASCII. */
889# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
890
891 /* The _A version makes sure that both the desired bit and the ASCII bit
892 * are present */
893# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
894 && ((PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask_A(classnum)) \
895 == _CC_mask_A(classnum)))
896
f4cdb42c 897# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 898# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
899# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
900# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
901# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
902# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
903# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
904# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
905# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
906# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
907# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
908# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
909# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
910# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
911# define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST))
912
fb9bfbf7
KW
913 /* Either participates in a fold with a character above 255, or is a
914 * multi-char fold */
430b7c70
KW
915# 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) NATIVE_TO_UNI(c)] & _CC_mask(_CC_NONLATIN1_FOLD)))
916
4eeeb416 917# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
918# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
919 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
920# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
921 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
9fb1bf9d 922#else /* No perl.h. */
9fb1bf9d 923# ifdef EBCDIC
9fb1bf9d 924# define isALPHA_A(c) (isASCII(c) && isALPHA(c))
15861f94 925# define isALPHANUMERIC_A(c) (isASCII(c) && isALPHANUMERIC(c))
9fb1bf9d
KW
926# define isBLANK_A(c) (isASCII(c) && isBLANK(c))
927# define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
928# define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
929# define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
930# define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
931# define isLOWER_A(c) (isASCII(c) && isLOWER(c))
932# define isPRINT_A(c) (isASCII(c) && isPRINT(c))
933# define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
934# define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
935# define isSPACE_A(c) (isASCII(c) && isSPACE(c))
936# define isUPPER_A(c) (isASCII(c) && isUPPER(c))
937# define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
938# define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
939# else /* ASCII platform, no perl.h */
9fb1bf9d 940# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
15861f94 941# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
9fb1bf9d 942# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
926bb54c 943# define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
2181da29 944# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
9fb1bf9d
KW
945# define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
946# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
947# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
948# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
949# define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
926bb54c
KW
950# define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) \
951 || ((c) >= 58 && (c) <= 64) \
952 || ((c) >= 91 && (c) <= 96) \
953 || ((c) >= 123 && (c) <= 126))
954# define isSPACE_A(c) ((c) == ' ' \
955 || (c) == '\t' \
956 || (c) == '\n' \
957 || (c) =='\r' \
958 || (c) == '\f')
2181da29 959# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
9fb1bf9d 960# define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
926bb54c
KW
961# define isXDIGIT_A(c) (isDIGIT_A(c) \
962 || ((c) >= 'a' && (c) <= 'f') \
963 || ((c) <= 'F' && (c) >= 'A'))
9fb1bf9d
KW
964# endif
965#endif /* ASCII range definitions */
ec79cfbb
KW
966
967/* Latin1 definitions */
8a58bdcf 968#ifdef H_PERL
f4cdb42c 969# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
15861f94 970# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
f4cdb42c 971# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
e5dcd934 972
8a58bdcf 973/* continuation character for legal NAME in \N{NAME} */
4eeeb416 974# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
e5dcd934 975
f4cdb42c
KW
976# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
977# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
978# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
979# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
980# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
981# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
982# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
983# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
984# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
985# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
8a58bdcf
KW
986#else /* No access to perl.h. Only a few provided here, just in case needed
987 * for backwards compatibility */
988 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
989 * >= AA test to speed up ASCII-only tests at the expense of the others */
990# define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
991 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
926bb54c 992 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
8a58bdcf
KW
993 || NATIVE_TO_UNI((U8) c) == 0xAA \
994 || NATIVE_TO_UNI((U8) c) == 0xB5 \
995 || NATIVE_TO_UNI((U8) c) == 0xBA)))
926bb54c
KW
996# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
997 || (c) == ' ' \
998 || (c) == '-' \
999 || (c) == '(' \
1000 || (c) == ')' \
1001 || (c) == ':' \
1002 || NATIVE_TO_UNI((U8) c) == 0xA0)
8a58bdcf
KW
1003#endif
1004
8a58bdcf
KW
1005/* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
1006 * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
1007 * therefore can use native functions. For those where C89 doesn't define a
1008 * function, use our function, assuming that the EBCDIC code page is isomorphic
1009 * with Latin1, which the three currently recognized by Perl are. Some libc's
1010 * have an isblank(), but it's not guaranteed. */
9d116dd7 1011#ifdef EBCDIC
271b5cb3 1012# define isALPHA(c) isalpha(c)
15861f94 1013# define isALPHANUMERIC(c) isalnum(c)
0bc70f7f 1014# define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
b8c5462f 1015# define isCNTRL(c) iscntrl(c)
9916188f 1016# define isDIGIT(c) isdigit(c)
b8c5462f 1017# define isGRAPH(c) isgraph(c)
8a58bdcf 1018# define isIDFIRST(c) (isALPHA(c) || (c) == '_')
7d79683e 1019# define isLOWER(c) islower(c)
9d116dd7 1020# define isPRINT(c) isprint(c)
56e4b611 1021# define isPSXSPC(c) isspace(c)
b8c5462f 1022# define isPUNCT(c) ispunct(c)
f1f70107
KW
1023# define isSPACE(c) (isPSXSPC(c) /* && (c) != '\v' (Experimentally making
1024 these macros identical) */)
7d79683e 1025# define isUPPER(c) isupper(c)
b8c5462f 1026# define isXDIGIT(c) isxdigit(c)
8a58bdcf 1027# define isWORDCHAR(c) (isalnum(c) || (c) == '_')
9d116dd7 1028# define toLOWER(c) tolower(c)
7d79683e 1029# define toUPPER(c) toupper(c)
ec79cfbb 1030#else /* Not EBCDIC: ASCII-only matching */
15861f94 1031# define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
271b5cb3 1032# define isALPHA(c) isALPHA_A(c)
ec79cfbb
KW
1033# define isBLANK(c) isBLANK_A(c)
1034# define isCNTRL(c) isCNTRL_A(c)
1035# define isDIGIT(c) isDIGIT_A(c)
1036# define isGRAPH(c) isGRAPH_A(c)
8a58bdcf 1037# define isIDFIRST(c) isIDFIRST_A(c)
ec79cfbb
KW
1038# define isLOWER(c) isLOWER_A(c)
1039# define isPRINT(c) isPRINT_A(c)
1040# define isPSXSPC(c) isPSXSPC_A(c)
1041# define isPUNCT(c) isPUNCT_A(c)
1042# define isSPACE(c) isSPACE_A(c)
1043# define isUPPER(c) isUPPER_A(c)
8a58bdcf 1044# define isWORDCHAR(c) isWORDCHAR_A(c)
ec79cfbb 1045# define isXDIGIT(c) isXDIGIT_A(c)
00f254e2 1046
8a58bdcf
KW
1047 /* ASCII casing. These could also be written as
1048 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1049 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1050 which uses table lookup and mask instead of subtraction. (This would
1051 work because the _MOD does not apply in the ASCII range) */
9d116dd7 1052# define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
7d79683e 1053# define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
9d116dd7 1054#endif
bbce6d69 1055
1a0901db
KW
1056
1057/* Use table lookup for speed; return error character for input
1058 * out-of-range */
1059#define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \
1060 ? UNI_TO_NATIVE(PL_latin1_lc[ \
1061 NATIVE_TO_UNI( (U8) (c)) ]) \
1062 : UNICODE_REPLACEMENT)
1063/* Modified uc. Is correct uc except for three non-ascii chars which are
1064 * all mapped to one of them, and these need special handling; error
1065 * character for input out-of-range */
1066#define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \
1067 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \
1068 NATIVE_TO_UNI( (U8) (c)) ]) \
1069 : UNICODE_REPLACEMENT)
1070
bbce6d69
PP
1071#ifdef USE_NEXT_CTYPE
1072
15861f94 1073# define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c))
ff68c719 1074# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 1075# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 1076# define isBLANK_LC(c) isBLANK((unsigned int)(c))
b8c5462f 1077# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
509fb054 1078# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
b8c5462f 1079# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
509fb054
KW
1080# define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
1081# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
ff68c719 1082# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 1083# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
509fb054
KW
1084# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
1085# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
1086# define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
e712593e 1087# define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
ff68c719 1088# define toLOWER_LC(c) NXToLower((unsigned int)(c))
509fb054 1089# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
bbce6d69
PP
1090
1091#else /* !USE_NEXT_CTYPE */
b8c5462f 1092
bbce6d69
PP
1093# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1094
76fde96a 1095/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
851131d3 1096
76fde96a 1097# define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
15861f94
KW
1098# define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \
1099 && isalnum((unsigned char)(c)))
07315176 1100# ifdef HAS_ISASCII
76fde96a 1101# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
07315176 1102# else
76fde96a 1103# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
07315176 1104# endif
67addccf 1105# ifdef HAS_ISBLANK
76fde96a 1106# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
67addccf 1107# else
76fde96a 1108# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
67addccf 1109# endif
76fde96a 1110# define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
509fb054 1111# define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
76fde96a 1112# define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
509fb054
KW
1113# define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1114 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1115# define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
76fde96a
KW
1116# define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1117# define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
509fb054
KW
1118# define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1119# define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1120# define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1121 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
e712593e 1122# define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
76fde96a 1123# define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? tolower((unsigned char)(c)) : (c))
509fb054 1124# define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? toupper((unsigned char)(c)) : (c))
bbce6d69
PP
1125
1126# else
1127
bbce6d69 1128# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1129# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
07315176 1130# define isASCII_LC(c) isascii(c)
67addccf
KW
1131# ifdef HAS_ISBLANK
1132# define isBLANK_LC(c) (isascii(c) && isblank(c))
1133# else
3e66ab74 1134# define isBLANK_LC(c) isBLANK_A(c)
67addccf 1135# endif
b8c5462f 1136# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1137# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1138# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1139# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1140# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1141# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1142# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1143# define isSPACE_LC(c) (isascii(c) && isspace(c))
1144# define isUPPER_LC(c) (isascii(c) && isupper(c))
1145# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1146# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
76fde96a 1147# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1148# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
bbce6d69
PP
1149
1150# endif
a0d0e21e 1151#endif /* USE_NEXT_CTYPE */
55204971 1152
eba68aa0
KW
1153#define isIDCONT(c) isWORDCHAR(c)
1154#define isIDCONT_A(c) isWORDCHAR_A(c)
1155#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1156#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1157#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1158
66c17564 1159/* For internal core Perl use only. If the input is Latin1, use the Latin1
a500dc72
KW
1160 * macro; otherwise use the function 'above_latin1'. Won't compile if 'c' isn't unsigned, as
1161 * won't match above_latin1 prototype. The macros do bounds checking, so have
66c17564
KW
1162 * duplicate checks here, so could create versions of the macros that don't,
1163 * but experiments show that gcc optimizes them out anyway. */
1164
1165/* Note that all ignore 'use bytes' */
cd500f2f
KW
1166#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1167 ? _generic_isCC(c, classnum) \
1168 : above_latin1(c))
1169#define _generic_swash_uni(classnum, c) ((c) < 256 \
1170 ? _generic_isCC(c, classnum) \
922e8cb4 1171 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1172#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1173#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1174#define isASCII_uni(c) isASCII(c)
1175#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1176#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1177#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1178#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1179#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1180#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1181#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1182#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1183
509fb054
KW
1184/* Posix and regular space are identical above Latin1 */
1185#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1186
cd500f2f 1187#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1188#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1189#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1190#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1191#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1192#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1193
b8d68ded 1194#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1195#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1196#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1197#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1198
cd500f2f
KW
1199#define _generic_LC_uvchr(latin1, above_latin1, c) \
1200 (c < 256 ? latin1(c) : above_latin1(c))
1201#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1202 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1203
1204#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1205#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1206 _CC_ALPHANUMERIC, c)
509fb054 1207#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1208#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1209#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1210#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1211#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1212#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1213 _is_uni_perl_idcont, c)
cd500f2f
KW
1214#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1215 _is_uni_perl_idstart, c)
1216#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1217#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1218#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1219 space under locale */
cd500f2f
KW
1220#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1221#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1222 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1223#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1224#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1225 _CC_WORDCHAR, c)
1226#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1227
a0ed51b3 1228
36b00f93 1229#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1230
a500dc72
KW
1231/* Everything whose name begins with an underscore is for internal core Perl
1232 * use only. */
1233
1234/* If the input is in the Latin1 range, use
252810d0
KW
1235 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1236 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1237 * fact that ASCII characters have the same representation whether utf8 or not.
1238 * Note that it assumes that the utf8 has been validated, and ignores 'use
1239 * bytes' */
cd500f2f 1240#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1241 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1242 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c
KW
1243 ? _generic_isCC( \
1244 TWO_BYTE_UTF8_TO_UNI(*(p), \
1245 *((p)+1 )), \
1246 classnum) \
252810d0 1247 : utf8)
cd500f2f
KW
1248/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1249 * can be a macro */
1250#define _generic_func_utf8(classnum, above_latin1, p) \
1251 _generic_utf8(classnum, p, above_latin1(p))
1252/* Like the above, but passes classnum to _isFOO_utf8(), instead of having a
1253 * 'above_latin1' parameter */
922e8cb4 1254#define _generic_swash_utf8(classnum, p) \
cd500f2f 1255 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1256
cc8ab7c0
KW
1257/* Like the above, but should be used only when it is known that there are no
1258 * characters in the range 128-255 which the class is TRUE for. Hence it can
cd500f2f
KW
1259 * skip the tests for this range. 'above_latin1' should include its arguments */
1260#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1261 (UTF8_IS_INVARIANT(*(p)) \
1262 ? _generic_isCC(*(p), classnum) \
1263 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
cd500f2f 1264 ? above_latin1 \
cc8ab7c0
KW
1265 : 0)
1266
84238efa
KW
1267/* NOTE that some of these macros have very similar ones in regcharclass.h.
1268 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1269 * there, differing in name only by an underscore from the one here
1270 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1271 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1272 * points; the regcharclass.h ones are implemented as a series of
1273 * "if-else-if-else ..." */
1274
922e8cb4
KW
1275#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1276#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1277#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1278 utf8, the non-utf8 macro works
1279 */
cd500f2f
KW
1280#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1281#define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1282#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1283 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1284#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1285#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1286 _is_utf8_perl_idcont, p)
e5dcd934 1287
c11ff943
KW
1288/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1289 * IDFIRST is an alnum. See
c8362b00 1290 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1291 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1292 * This used to be not the XID version, but we decided to go with the more
1293 * modern Unicode definition */
cd500f2f 1294#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1295 _is_utf8_perl_idstart, p)
e5dcd934 1296
922e8cb4
KW
1297#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1298#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1299
509fb054 1300/* Posix and regular space are identical above Latin1 */
cd500f2f 1301#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1302
922e8cb4 1303#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1304#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1305#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1306#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1307#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1308#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1309 is_XDIGIT_high(p))
a0ed51b3 1310
509fb054
KW
1311#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1312#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1313#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1314
9201a572
KW
1315/* For internal core Perl use only. If the input is in the Latin1 range, use
1316 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1317 * use the value given by the 'utf8' parameter. This relies on the fact that
1318 * ASCII characters have the same representation whether utf8 or not. Note
1319 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
cd500f2f 1320#define _generic_LC_utf8(macro, p, utf8) \
9201a572
KW
1321 (UTF8_IS_INVARIANT(*(p)) \
1322 ? macro(*(p)) \
1323 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1324 ? macro(TWO_BYTE_UTF8_TO_UNI(*(p), *((p)+1))) \
1325 : utf8)
1326
cd500f2f
KW
1327#define _generic_LC_swash_utf8(macro, classnum, p) \
1328 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1329#define _generic_LC_func_utf8(macro, above_latin1, p) \
1330 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1331
cd500f2f
KW
1332#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1333 _CC_ALPHANUMERIC, p)
1334#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1335#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1336#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1337#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1338#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1339#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1340#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1341#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1342#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1343#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1344#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1345 space under locale */
cd500f2f
KW
1346#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1347#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1348#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1349#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1350 _CC_WORDCHAR, p)
1351#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1352
fbc19f27
KW
1353/* Macros for backwards compatibility and for completeness when the ASCII and
1354 * Latin1 values are identical */
fbc19f27
KW
1355#define isALPHAU(c) isALPHA_L1(c)
1356#define isDIGIT_L1(c) isDIGIT_A(c)
1357#define isOCTAL(c) isOCTAL_A(c)
1358#define isOCTAL_L1(c) isOCTAL_A(c)
1359#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1360#define isALNUM(c) isWORDCHAR(c)
1361#define isALNUMU(c) isWORDCHAR_L1(c)
1362#define isALNUM_LC(c) isWORDCHAR_LC(c)
1363#define isALNUM_uni(c) isWORDCHAR_uni(c)
1364#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1365#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1366#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1367#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1368#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1369#define isALNUMC(c) isALPHANUMERIC(c)
1370#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1371#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1372#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1373#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1374#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1375
6ae9f32a 1376/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1377 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1378 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1379 * that is EOT on ASCII; ST on EBCDIC */
6ae9f32a 1380# define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64)
bbce6d69 1381
dea28490
JJ
1382/* Line numbers are unsigned, 32 bits. */
1383typedef U32 line_t;
e5dcd934 1384#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1385
91152fc1
DG
1386/* Helpful alias for version prescan */
1387#define is_LAX_VERSION(a,b) \
1388 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1389
1390#define is_STRICT_VERSION(a,b) \
1391 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1392
1393#define BADVERSION(a,b,c) \
1394 if (b) { \
1395 *b = c; \
1396 } \
1397 return a;
8c52afec 1398
95a59cab
YO
1399#define READ_XDIGIT(s) (isALPHA(*(s)) ? ((*(s)++ + 9) & 0xf) : (*(s)++ & 0xf))
1400
8e84507e 1401/*
ccfc67b7
JH
1402=head1 Memory Management
1403
a02a5408 1404=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1405The XSUB-writer's interface to the C C<malloc> function.
1406
c5008215
JC
1407In 5.9.3, Newx() and friends replace the older New() API, and drops
1408the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1409themselves. This aid has been superseded by a new build option,
d10b4965 1410PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1411there for use in XS modules supporting older perls.
1412
a02a5408 1413=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1414The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1415cast. See also C<Newx>.
954c1994 1416
a02a5408 1417=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1418The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1419memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1420
954c1994
GS
1421=for apidoc Am|void|Renew|void* ptr|int nitems|type
1422The XSUB-writer's interface to the C C<realloc> function.
1423
1424=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1425The XSUB-writer's interface to the C C<realloc> function, with
1426cast.
1427
49b8b560 1428=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1429The XSUB-writer's interface to the C C<free> function.
1430
1431=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1432The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1433source, C<dest> is the destination, C<nitems> is the number of items, and
1434C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1435
e90e2364
NC
1436=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1437Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1438optimise.
1439
954c1994
GS
1440=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1441The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1442source, C<dest> is the destination, C<nitems> is the number of items, and
1443C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1444
e90e2364
NC
1445=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1446
1447Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1448optimise.
1449
954c1994
GS
1450=for apidoc Am|void|Zero|void* dest|int nitems|type
1451
1452The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1453destination, C<nitems> is the number of items, and C<type> is the type.
1454
e90e2364
NC
1455=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1456
1457Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1458optimise.
1459
da5d8dbb 1460=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1461This is an architecture-independent macro to copy one structure to another.
954c1994 1462
7e337ee0
JH
1463=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1464
1465Fill up memory with a byte pattern (a byte repeated over and over
1466again) that hopefully catches attempts to access uninitialized memory.
1467
1468=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1469
1470PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1471
1c12ffb4 1472=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1473
1474PoisonWith(0xEF) for catching access to freed memory.
1475
9965345d
JH
1476=for apidoc Am|void|Poison|void* dest|int nitems|type
1477
7e337ee0 1478PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1479
1480=cut */
954c1994 1481
561b68a9
SH
1482/* Maintained for backwards-compatibility only. Use newSV() instead. */
1483#ifndef PERL_CORE
ff06c60c 1484#define NEWSV(x,len) newSV(len)
561b68a9 1485#endif
ff06c60c 1486
19a94d75
JH
1487#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1488
2720f793
JH
1489/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1490 * overly eager compilers that will bleat about e.g.
1491 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1492#ifdef PERL_MALLOC_WRAP
f019c49e 1493#define MEM_WRAP_CHECK(n,t) \
43387ee1 1494 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_memory_wrap(),0))
27d5b266 1495#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1496 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1497#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1498
43387ee1 1499#define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (Perl_croak_memory_wrap(),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266
JH
1500#else
1501
410319be
NC
1502#define MEM_WRAP_CHECK(n,t)
1503#define MEM_WRAP_CHECK_1(n,t,a)
1504#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1505#define MEM_WRAP_CHECK_(n,t)
1506
6fff79ce 1507#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1508
1936d2a7 1509#endif
8b44ba4c 1510
fe4f188c 1511#ifdef PERL_MEM_LOG
46c6c7e2 1512/*
9f653bb5 1513 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1514 * go through functions, which are handy for debugging breakpoints, but
1515 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1516 * line number, and C function name if available) passed in. This info can
1517 * then be used for logging the calls, for which one gets a sample
73d1d973 1518 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1519 *
46c6c7e2 1520 * Known problems:
94e892a6 1521 * - not all memory allocs get logged, only those
46c6c7e2 1522 * that go through Newx() and derivatives (while all
94e892a6 1523 * Safefrees do get logged)
46c6c7e2
JH
1524 * - __FILE__ and __LINE__ do not work everywhere
1525 * - __func__ or __FUNCTION__ even less so
1526 * - I think more goes on after the perlio frees but
1527 * the thing is that STDERR gets closed (as do all
1528 * the file descriptors)
1529 * - no deeper calling stack than the caller of the Newx()
1530 * or the kind, but do I look like a C reflection/introspection
1531 * utility to you?
1532 * - the function prototypes for the logging functions
1533 * probably should maybe be somewhere else than handy.h
1534 * - one could consider inlining (macrofying) the logging
1535 * for speed, but I am too lazy
1536 * - one could imagine recording the allocations in a hash,
1537 * (keyed by the allocation address?), and maintain that
1538 * through reallocs and frees, but how to do that without
1539 * any News() happening...?
73d1d973 1540 * - lots of -Ddefines to get useful/controllable output
b953482e 1541 * - lots of ENV reads
46c6c7e2
JH
1542 */
1543
12754f92 1544PERL_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 1545
12754f92 1546PERL_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 1547
12754f92 1548PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1549
0b0ab801 1550# ifdef PERL_CORE
73d1d973 1551# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1552enum mem_log_type {
1553 MLT_ALLOC,
1554 MLT_REALLOC,
d7a2c63c
MHM
1555 MLT_FREE,
1556 MLT_NEW_SV,
1557 MLT_DEL_SV
0b0ab801
MHM
1558};
1559# endif
12754f92 1560# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1561void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1562void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1563# endif
0b0ab801
MHM
1564# endif
1565
fe4f188c
JH
1566#endif
1567
1568#ifdef PERL_MEM_LOG
d1401ee9
MHM
1569#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1570#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1571#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1572#endif
1573
1574#ifndef MEM_LOG_ALLOC
1575#define MEM_LOG_ALLOC(n,t,a) (a)
1576#endif
1577#ifndef MEM_LOG_REALLOC
1578#define MEM_LOG_REALLOC(n,t,v,a) (a)
1579#endif
1580#ifndef MEM_LOG_FREE
1581#define MEM_LOG_FREE(a) (a)
1582#endif
1583
d1401ee9
MHM
1584#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1585#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1586#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1587
1588#ifndef PERL_CORE
a02a5408
JC
1589/* pre 5.9.x compatibility */
1590#define New(x,v,n,t) Newx(v,n,t)
1591#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1592#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1593#endif
a02a5408 1594
ff68c719 1595#define Renew(v,n,t) \
d1401ee9 1596 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1597#define Renewc(v,n,t,c) \
d1401ee9 1598 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1599
1600#ifdef PERL_POISON
1601#define Safefree(d) \
06c0cc96 1602 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1603#else
fe4f188c 1604#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1605#endif
55497cff 1606
8b44ba4c
NC
1607#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1608#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1609#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1610
8b44ba4c
NC
1611#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1612#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1613#ifdef HAS_MEMSET
8b44ba4c 1614#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1615#else
8b44ba4c
NC
1616/* Using bzero(), which returns void. */
1617#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1618#endif
1619
7e337ee0
JH
1620#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1621#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1622#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1623#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1624
2304df62 1625#ifdef USE_STRUCT_COPY
ff68c719 1626#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1627#else
1628#define StructCopy(s,d,t) Copy(s,d,1,t)
1629#endif
2cc61e15 1630
622913ab 1631#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1632#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1633
2cc61e15
DD
1634#ifdef NEED_VA_COPY
1635# ifdef va_copy
1636# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1637# else
a1866d1b
JH
1638# if defined(__va_copy)
1639# define Perl_va_copy(s, d) __va_copy(d, s)
1640# else
1641# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1642# endif
2cc61e15
DD
1643# endif
1644#endif
1645
472d47bc
SB
1646/* convenience debug macros */
1647#ifdef USE_ITHREADS
1648#define pTHX_FORMAT "Perl interpreter: 0x%p"
1649#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1650#define pTHX_VALUE_ (void *)my_perl,
1651#define pTHX_VALUE (void *)my_perl
1652#define pTHX__VALUE_ ,(void *)my_perl,
1653#define pTHX__VALUE ,(void *)my_perl
472d47bc 1654#else
3609ea0d 1655#define pTHX_FORMAT
472d47bc 1656#define pTHX__FORMAT
3609ea0d 1657#define pTHX_VALUE_
472d47bc 1658#define pTHX_VALUE
3609ea0d 1659#define pTHX__VALUE_
472d47bc
SB
1660#define pTHX__VALUE
1661#endif /* USE_ITHREADS */
3609ea0d 1662
2acdbac1
NC
1663/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1664 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1665 CPAN::Unpack show any users outside the core. */
1666#ifdef PERL_CORE
d1d15184 1667# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1668#endif
1669
9d745869
KW
1670#endif /* HANDY_H */
1671
e9a8c099
MHM
1672/*
1673 * Local variables:
1674 * c-indentation-style: bsd
1675 * c-basic-offset: 4
14d04a33 1676 * indent-tabs-mode: nil
e9a8c099
MHM
1677 * End:
1678 *
14d04a33 1679 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1680 */