This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perlapi: Document some macros
[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
702=for apidoc Am|bool|isVERTWS|char ch
703Returns a boolean indicating whether the specified character is considered
704to be vertical white space, such as C<"\n"> or C<"\f">. See the L<top of this
705section|/Character classes> for an explanation of variants
706C<isVERTWS_uni>, and C<isVERTWS_utf8>.
707
243effed 708=head1 Miscellaneous Functions
8eea39dd 709
95a59cab 710=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 711Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
712Behaviour is only well defined when isXDIGIT(*str) is true.
713
e7c1e6c1
KW
714=head1 Character case changing
715
954c1994 716=for apidoc Am|char|toUPPER|char ch
06dc89af
KW
717Converts the specified character to uppercase in the platform's native
718character set, if possible; otherwise returns the input character itself.
954c1994
GS
719
720=for apidoc Am|char|toLOWER|char ch
06dc89af
KW
721Converts the specified character to lowercase in the platform's native
722character set, if possible; otherwise returns the input character itself.
954c1994
GS
723
724=cut
353c9b6f 725
3c3ecf18 726XXX Still undocumented the other toUPPER etc functions
243effed 727
8a58bdcf 728Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
729patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
730
954c1994
GS
731*/
732
7c062697
KW
733/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
734 * is known only in the perl core, and this macro can be called from outside
735 * that */
736#ifdef HAS_QUAD
737# define WIDEST_UTYPE U64TYPE
738#else
739# define WIDEST_UTYPE U32
740#endif
741
3912bc88
KW
742/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
743 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
744 * bits of information are lost even on a 64-bit machine, but to get the
745 * compiler to optimize it out if possible. This is because Configure makes
746 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
747 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
748 */
749#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 750
41f43cc2 751#ifdef EBCDIC
ff7eba81 752# define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128))
41f43cc2
KW
753#else
754# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
755#endif
756
ec79cfbb 757#define isASCII_A(c) isASCII(c)
3509bb53 758#define isASCII_L1(c) isASCII(c)
6515cb76 759
c2da0b36
KW
760/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
761 * and the 8 possible permutations of those bits exactly comprise the 8 octal
762 * digits */
3d33b3e3 763#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 764
ec79cfbb 765/* ASCII range only */
9fb1bf9d 766#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 767
a500dc72
KW
768/* Character class numbers. For internal core Perl use only. The ones less
769 * than 32 are used in PL_charclass[] and the ones up through the one that
770 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
771 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
772 * their actual definitions are here. If that file has a name not used here,
773 * it won't compile.
1709d539
KW
774 *
775 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
776 * frequency of their use. This gives a slight edge to exiting a loop earlier
777 * (in reginclass() in regexec.c) */
1709d539
KW
778# define _CC_WORDCHAR 0 /* \w and [:word:] */
779# define _CC_DIGIT 1 /* \d and [:digit:] */
780# define _CC_ALPHA 2 /* [:alpha:] */
781# define _CC_LOWER 3 /* [:lower:] */
782# define _CC_UPPER 4 /* [:upper:] */
783# define _CC_PUNCT 5 /* [:punct:] */
784# define _CC_PRINT 6 /* [:print:] */
15861f94 785# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 786# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 787# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 788
b0d691b2 789#define _FIRST_NON_SWASH_CC 10
1709d539
KW
790/* The character classes above are implemented with swashes. The second group
791 * (just below) contains the ones implemented without. These are also sorted
792 * in rough order of the frequency of their use, except that \v should be last,
793 * as it isn't a real Posix character class, and some (small) inefficiencies in
794 * regular expression handling would be introduced by putting it in the middle
795 * of those that are. Also, cntrl and ascii come after the others as it may be
796 * useful to group these which have no members that match above Latin1, (or
797 * above ASCII in the latter case) */
798
b0d691b2
KW
799# define _CC_SPACE 10 /* \s */
800# define _CC_BLANK 11 /* [:blank:] */
801# define _CC_XDIGIT 12 /* [:xdigit:] */
802# define _CC_PSXSPC 13 /* [:space:] */
803# define _CC_CNTRL 14 /* [:cntrl:] */
804# define _CC_ASCII 15 /* [:ascii:] */
805# define _CC_VERTSPACE 16 /* \v */
1709d539 806
a0947d7b
KW
807# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
808
1709d539
KW
809/* The members of the third group below do not need to be coordinated with data
810 * structures in regcomp.[ch] and regexec.c */
b0d691b2
KW
811# define _CC_IDFIRST 17
812# define _CC_CHARNAME_CONT 18
813# define _CC_NONLATIN1_FOLD 19
814# define _CC_QUOTEMETA 20
815# define _CC_NON_FINAL_FOLD 21
816# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
817# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
818/* Unused: 23-30
f4cdb42c
KW
819 * If more bits are needed, one could add a second word for non-64bit
820 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
821 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
822 * is used only for optimization (as of this writing), and differs in the
823 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
824 * characters: the masculine and feminine ordinal indicators, so removing it
825 * would just cause /i regexes which match them to run less efficiently */
96ac0975 826
3a371f2f
KW
827#if defined(PERL_CORE) || defined(PERL_EXT)
828/* An enum version of the character class numbers, to help compilers
829 * optimize */
830typedef enum {
3a371f2f 831 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
832 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
833 _CC_ENUM_ASCII = _CC_ASCII,
834 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 835 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 836 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
837 _CC_ENUM_DIGIT = _CC_DIGIT,
838 _CC_ENUM_GRAPH = _CC_GRAPH,
839 _CC_ENUM_LOWER = _CC_LOWER,
840 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 841 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 842 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 843 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 844 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 845 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 846 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 847 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
848} _char_class_number;
849#endif
850
63c61c3f 851#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 852#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 853
3018b823 854#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 855# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
856 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
857 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
858 #error Need to adjust order of swash_property_names[]
859# endif
860
b44daeeb
KW
861/* This is declared static in each of the few files that this is #defined for
862 * to keep them from being publicly accessible. Hence there is a small amount
863 * of wasted space */
864
922e8cb4
KW
865static const char* const swash_property_names[] = {
866 "XPosixWord",
867 "XPosixDigit",
868 "XPosixAlpha",
869 "XPosixLower",
870 "XPosixUpper",
871 "XPosixPunct",
872 "XPosixPrint",
873 "XPosixAlnum",
b0d691b2
KW
874 "XPosixGraph",
875 "Cased"
922e8cb4
KW
876};
877#endif
878
96ac0975
NC
879# ifdef DOINIT
880EXTCONST U32 PL_charclass[] = {
881# include "l1_char_class_tab.h"
882};
883
884# else /* ! DOINIT */
885EXTCONST U32 PL_charclass[];
886# endif
887
265c1f46 888 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70
KW
889# define _CC_mask(classnum) (1U << (classnum))
890# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
926bb54c 891 && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask(classnum)))
4eeeb416 892
f4cdb42c
KW
893 /* The mask for the _A versions of the macros; it just adds in the bit for
894 * ASCII. */
895# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
896
897 /* The _A version makes sure that both the desired bit and the ASCII bit
898 * are present */
899# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
900 && ((PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask_A(classnum)) \
901 == _CC_mask_A(classnum)))
902
f4cdb42c 903# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 904# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
905# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
906# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
907# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
908# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
909# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
910# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
911# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
912# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
913# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
914# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
915# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
916# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
917# define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST))
918
fb9bfbf7
KW
919 /* Either participates in a fold with a character above 255, or is a
920 * multi-char fold */
430b7c70
KW
921# 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)))
922
4eeeb416 923# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
924# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
925 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
926# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
927 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
9fb1bf9d 928#else /* No perl.h. */
9fb1bf9d 929# ifdef EBCDIC
9fb1bf9d 930# define isALPHA_A(c) (isASCII(c) && isALPHA(c))
15861f94 931# define isALPHANUMERIC_A(c) (isASCII(c) && isALPHANUMERIC(c))
9fb1bf9d
KW
932# define isBLANK_A(c) (isASCII(c) && isBLANK(c))
933# define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
934# define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
935# define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
936# define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
937# define isLOWER_A(c) (isASCII(c) && isLOWER(c))
938# define isPRINT_A(c) (isASCII(c) && isPRINT(c))
939# define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
940# define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
941# define isSPACE_A(c) (isASCII(c) && isSPACE(c))
942# define isUPPER_A(c) (isASCII(c) && isUPPER(c))
943# define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
944# define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
945# else /* ASCII platform, no perl.h */
9fb1bf9d 946# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
15861f94 947# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
9fb1bf9d 948# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
926bb54c 949# define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
2181da29 950# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
9fb1bf9d
KW
951# define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
952# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
953# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
954# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
955# define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
926bb54c
KW
956# define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) \
957 || ((c) >= 58 && (c) <= 64) \
958 || ((c) >= 91 && (c) <= 96) \
959 || ((c) >= 123 && (c) <= 126))
960# define isSPACE_A(c) ((c) == ' ' \
961 || (c) == '\t' \
962 || (c) == '\n' \
963 || (c) =='\r' \
964 || (c) == '\f')
2181da29 965# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
9fb1bf9d 966# define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
926bb54c
KW
967# define isXDIGIT_A(c) (isDIGIT_A(c) \
968 || ((c) >= 'a' && (c) <= 'f') \
969 || ((c) <= 'F' && (c) >= 'A'))
9fb1bf9d
KW
970# endif
971#endif /* ASCII range definitions */
ec79cfbb
KW
972
973/* Latin1 definitions */
8a58bdcf 974#ifdef H_PERL
f4cdb42c 975# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
15861f94 976# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
f4cdb42c 977# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
e5dcd934 978
8a58bdcf 979/* continuation character for legal NAME in \N{NAME} */
4eeeb416 980# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
e5dcd934 981
f4cdb42c
KW
982# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
983# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
984# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
985# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
986# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
987# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
988# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
989# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
990# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
991# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
8a58bdcf
KW
992#else /* No access to perl.h. Only a few provided here, just in case needed
993 * for backwards compatibility */
994 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
995 * >= AA test to speed up ASCII-only tests at the expense of the others */
996# define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
997 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
926bb54c 998 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
8a58bdcf
KW
999 || NATIVE_TO_UNI((U8) c) == 0xAA \
1000 || NATIVE_TO_UNI((U8) c) == 0xB5 \
1001 || NATIVE_TO_UNI((U8) c) == 0xBA)))
926bb54c
KW
1002# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1003 || (c) == ' ' \
1004 || (c) == '-' \
1005 || (c) == '(' \
1006 || (c) == ')' \
1007 || (c) == ':' \
1008 || NATIVE_TO_UNI((U8) c) == 0xA0)
8a58bdcf
KW
1009#endif
1010
8a58bdcf
KW
1011/* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
1012 * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
1013 * therefore can use native functions. For those where C89 doesn't define a
1014 * function, use our function, assuming that the EBCDIC code page is isomorphic
1015 * with Latin1, which the three currently recognized by Perl are. Some libc's
1016 * have an isblank(), but it's not guaranteed. */
9d116dd7 1017#ifdef EBCDIC
271b5cb3 1018# define isALPHA(c) isalpha(c)
15861f94 1019# define isALPHANUMERIC(c) isalnum(c)
0bc70f7f 1020# define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
b8c5462f 1021# define isCNTRL(c) iscntrl(c)
9916188f 1022# define isDIGIT(c) isdigit(c)
b8c5462f 1023# define isGRAPH(c) isgraph(c)
8a58bdcf 1024# define isIDFIRST(c) (isALPHA(c) || (c) == '_')
7d79683e 1025# define isLOWER(c) islower(c)
9d116dd7 1026# define isPRINT(c) isprint(c)
56e4b611 1027# define isPSXSPC(c) isspace(c)
b8c5462f 1028# define isPUNCT(c) ispunct(c)
f1f70107
KW
1029# define isSPACE(c) (isPSXSPC(c) /* && (c) != '\v' (Experimentally making
1030 these macros identical) */)
7d79683e 1031# define isUPPER(c) isupper(c)
b8c5462f 1032# define isXDIGIT(c) isxdigit(c)
8a58bdcf 1033# define isWORDCHAR(c) (isalnum(c) || (c) == '_')
9d116dd7 1034# define toLOWER(c) tolower(c)
7d79683e 1035# define toUPPER(c) toupper(c)
ec79cfbb 1036#else /* Not EBCDIC: ASCII-only matching */
15861f94 1037# define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
271b5cb3 1038# define isALPHA(c) isALPHA_A(c)
ec79cfbb
KW
1039# define isBLANK(c) isBLANK_A(c)
1040# define isCNTRL(c) isCNTRL_A(c)
1041# define isDIGIT(c) isDIGIT_A(c)
1042# define isGRAPH(c) isGRAPH_A(c)
8a58bdcf 1043# define isIDFIRST(c) isIDFIRST_A(c)
ec79cfbb
KW
1044# define isLOWER(c) isLOWER_A(c)
1045# define isPRINT(c) isPRINT_A(c)
1046# define isPSXSPC(c) isPSXSPC_A(c)
1047# define isPUNCT(c) isPUNCT_A(c)
1048# define isSPACE(c) isSPACE_A(c)
1049# define isUPPER(c) isUPPER_A(c)
8a58bdcf 1050# define isWORDCHAR(c) isWORDCHAR_A(c)
ec79cfbb 1051# define isXDIGIT(c) isXDIGIT_A(c)
00f254e2 1052
8a58bdcf
KW
1053 /* ASCII casing. These could also be written as
1054 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1055 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1056 which uses table lookup and mask instead of subtraction. (This would
1057 work because the _MOD does not apply in the ASCII range) */
9d116dd7 1058# define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
7d79683e 1059# define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
9d116dd7 1060#endif
bbce6d69 1061
1a0901db
KW
1062
1063/* Use table lookup for speed; return error character for input
1064 * out-of-range */
1065#define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \
1066 ? UNI_TO_NATIVE(PL_latin1_lc[ \
1067 NATIVE_TO_UNI( (U8) (c)) ]) \
1068 : UNICODE_REPLACEMENT)
1069/* Modified uc. Is correct uc except for three non-ascii chars which are
1070 * all mapped to one of them, and these need special handling; error
1071 * character for input out-of-range */
1072#define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \
1073 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \
1074 NATIVE_TO_UNI( (U8) (c)) ]) \
1075 : UNICODE_REPLACEMENT)
1076
bbce6d69
PP
1077#ifdef USE_NEXT_CTYPE
1078
15861f94 1079# define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c))
ff68c719 1080# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 1081# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 1082# define isBLANK_LC(c) isBLANK((unsigned int)(c))
b8c5462f 1083# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
509fb054 1084# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
b8c5462f 1085# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
509fb054
KW
1086# define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
1087# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
ff68c719 1088# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 1089# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
509fb054
KW
1090# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
1091# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
1092# define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
e712593e 1093# define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
ff68c719 1094# define toLOWER_LC(c) NXToLower((unsigned int)(c))
509fb054 1095# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
bbce6d69
PP
1096
1097#else /* !USE_NEXT_CTYPE */
b8c5462f 1098
bbce6d69
PP
1099# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1100
76fde96a 1101/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
851131d3 1102
76fde96a 1103# define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
15861f94
KW
1104# define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \
1105 && isalnum((unsigned char)(c)))
07315176 1106# ifdef HAS_ISASCII
76fde96a 1107# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
07315176 1108# else
76fde96a 1109# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
07315176 1110# endif
67addccf 1111# ifdef HAS_ISBLANK
76fde96a 1112# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
67addccf 1113# else
76fde96a 1114# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
67addccf 1115# endif
76fde96a 1116# define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
509fb054 1117# define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
76fde96a 1118# define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
509fb054
KW
1119# define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1120 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1121# define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
76fde96a
KW
1122# define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1123# define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
509fb054
KW
1124# define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1125# define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1126# define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1127 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
e712593e 1128# define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
76fde96a 1129# define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? tolower((unsigned char)(c)) : (c))
509fb054 1130# define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? toupper((unsigned char)(c)) : (c))
bbce6d69
PP
1131
1132# else
1133
bbce6d69 1134# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1135# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
07315176 1136# define isASCII_LC(c) isascii(c)
67addccf
KW
1137# ifdef HAS_ISBLANK
1138# define isBLANK_LC(c) (isascii(c) && isblank(c))
1139# else
3e66ab74 1140# define isBLANK_LC(c) isBLANK_A(c)
67addccf 1141# endif
b8c5462f 1142# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1143# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1144# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1145# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1146# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1147# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1148# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1149# define isSPACE_LC(c) (isascii(c) && isspace(c))
1150# define isUPPER_LC(c) (isascii(c) && isupper(c))
1151# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1152# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
76fde96a 1153# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1154# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
bbce6d69
PP
1155
1156# endif
a0d0e21e 1157#endif /* USE_NEXT_CTYPE */
55204971 1158
eba68aa0
KW
1159#define isIDCONT(c) isWORDCHAR(c)
1160#define isIDCONT_A(c) isWORDCHAR_A(c)
1161#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1162#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1163#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1164
66c17564 1165/* For internal core Perl use only. If the input is Latin1, use the Latin1
a500dc72
KW
1166 * macro; otherwise use the function 'above_latin1'. Won't compile if 'c' isn't unsigned, as
1167 * won't match above_latin1 prototype. The macros do bounds checking, so have
66c17564
KW
1168 * duplicate checks here, so could create versions of the macros that don't,
1169 * but experiments show that gcc optimizes them out anyway. */
1170
1171/* Note that all ignore 'use bytes' */
cd500f2f
KW
1172#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1173 ? _generic_isCC(c, classnum) \
1174 : above_latin1(c))
1175#define _generic_swash_uni(classnum, c) ((c) < 256 \
1176 ? _generic_isCC(c, classnum) \
922e8cb4 1177 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1178#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1179#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1180#define isASCII_uni(c) isASCII(c)
1181#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1182#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1183#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1184#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1185#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1186#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1187#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1188#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1189
509fb054
KW
1190/* Posix and regular space are identical above Latin1 */
1191#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1192
cd500f2f 1193#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1194#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1195#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1196#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1197#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1198#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1199
b8d68ded 1200#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1201#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1202#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1203#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1204
cd500f2f
KW
1205#define _generic_LC_uvchr(latin1, above_latin1, c) \
1206 (c < 256 ? latin1(c) : above_latin1(c))
1207#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1208 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1209
1210#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1211#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1212 _CC_ALPHANUMERIC, c)
509fb054 1213#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1214#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1215#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1216#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1217#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1218#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1219 _is_uni_perl_idcont, c)
cd500f2f
KW
1220#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1221 _is_uni_perl_idstart, c)
1222#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1223#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1224#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1225 space under locale */
cd500f2f
KW
1226#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1227#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1228 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1229#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1230#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1231 _CC_WORDCHAR, c)
1232#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1233
a0ed51b3 1234
36b00f93 1235#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1236
a500dc72
KW
1237/* Everything whose name begins with an underscore is for internal core Perl
1238 * use only. */
1239
1240/* If the input is in the Latin1 range, use
252810d0
KW
1241 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1242 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1243 * fact that ASCII characters have the same representation whether utf8 or not.
1244 * Note that it assumes that the utf8 has been validated, and ignores 'use
1245 * bytes' */
cd500f2f 1246#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1247 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1248 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c
KW
1249 ? _generic_isCC( \
1250 TWO_BYTE_UTF8_TO_UNI(*(p), \
1251 *((p)+1 )), \
1252 classnum) \
252810d0 1253 : utf8)
cd500f2f
KW
1254/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1255 * can be a macro */
1256#define _generic_func_utf8(classnum, above_latin1, p) \
1257 _generic_utf8(classnum, p, above_latin1(p))
1258/* Like the above, but passes classnum to _isFOO_utf8(), instead of having a
1259 * 'above_latin1' parameter */
922e8cb4 1260#define _generic_swash_utf8(classnum, p) \
cd500f2f 1261 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1262
cc8ab7c0
KW
1263/* Like the above, but should be used only when it is known that there are no
1264 * characters in the range 128-255 which the class is TRUE for. Hence it can
cd500f2f
KW
1265 * skip the tests for this range. 'above_latin1' should include its arguments */
1266#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1267 (UTF8_IS_INVARIANT(*(p)) \
1268 ? _generic_isCC(*(p), classnum) \
1269 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
cd500f2f 1270 ? above_latin1 \
cc8ab7c0
KW
1271 : 0)
1272
84238efa
KW
1273/* NOTE that some of these macros have very similar ones in regcharclass.h.
1274 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1275 * there, differing in name only by an underscore from the one here
1276 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1277 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1278 * points; the regcharclass.h ones are implemented as a series of
1279 * "if-else-if-else ..." */
1280
922e8cb4
KW
1281#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1282#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1283#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1284 utf8, the non-utf8 macro works
1285 */
cd500f2f
KW
1286#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1287#define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1288#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1289 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1290#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1291#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1292 _is_utf8_perl_idcont, p)
e5dcd934 1293
c11ff943
KW
1294/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1295 * IDFIRST is an alnum. See
c8362b00 1296 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1297 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1298 * This used to be not the XID version, but we decided to go with the more
1299 * modern Unicode definition */
cd500f2f 1300#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1301 _is_utf8_perl_idstart, p)
e5dcd934 1302
922e8cb4
KW
1303#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1304#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1305
509fb054 1306/* Posix and regular space are identical above Latin1 */
cd500f2f 1307#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1308
922e8cb4 1309#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1310#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1311#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1312#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1313#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1314#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1315 is_XDIGIT_high(p))
a0ed51b3 1316
509fb054
KW
1317#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1318#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1319#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1320
9201a572
KW
1321/* For internal core Perl use only. If the input is in the Latin1 range, use
1322 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1323 * use the value given by the 'utf8' parameter. This relies on the fact that
1324 * ASCII characters have the same representation whether utf8 or not. Note
1325 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
cd500f2f 1326#define _generic_LC_utf8(macro, p, utf8) \
9201a572
KW
1327 (UTF8_IS_INVARIANT(*(p)) \
1328 ? macro(*(p)) \
1329 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1330 ? macro(TWO_BYTE_UTF8_TO_UNI(*(p), *((p)+1))) \
1331 : utf8)
1332
cd500f2f
KW
1333#define _generic_LC_swash_utf8(macro, classnum, p) \
1334 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1335#define _generic_LC_func_utf8(macro, above_latin1, p) \
1336 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1337
cd500f2f
KW
1338#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1339 _CC_ALPHANUMERIC, p)
1340#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1341#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1342#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1343#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1344#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1345#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1346#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1347#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1348#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1349#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1350#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1351 space under locale */
cd500f2f
KW
1352#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1353#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1354#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1355#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1356 _CC_WORDCHAR, p)
1357#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1358
fbc19f27
KW
1359/* Macros for backwards compatibility and for completeness when the ASCII and
1360 * Latin1 values are identical */
fbc19f27
KW
1361#define isALPHAU(c) isALPHA_L1(c)
1362#define isDIGIT_L1(c) isDIGIT_A(c)
1363#define isOCTAL(c) isOCTAL_A(c)
1364#define isOCTAL_L1(c) isOCTAL_A(c)
1365#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1366#define isALNUM(c) isWORDCHAR(c)
1367#define isALNUMU(c) isWORDCHAR_L1(c)
1368#define isALNUM_LC(c) isWORDCHAR_LC(c)
1369#define isALNUM_uni(c) isWORDCHAR_uni(c)
1370#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1371#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1372#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1373#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1374#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1375#define isALNUMC(c) isALPHANUMERIC(c)
1376#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1377#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1378#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1379#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1380#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1381
6ae9f32a 1382/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1383 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1384 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1385 * that is EOT on ASCII; ST on EBCDIC */
6ae9f32a 1386# define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64)
bbce6d69 1387
dea28490
JJ
1388/* Line numbers are unsigned, 32 bits. */
1389typedef U32 line_t;
e5dcd934 1390#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1391
91152fc1
DG
1392/* Helpful alias for version prescan */
1393#define is_LAX_VERSION(a,b) \
1394 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1395
1396#define is_STRICT_VERSION(a,b) \
1397 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1398
1399#define BADVERSION(a,b,c) \
1400 if (b) { \
1401 *b = c; \
1402 } \
1403 return a;
8c52afec 1404
95a59cab
YO
1405#define READ_XDIGIT(s) (isALPHA(*(s)) ? ((*(s)++ + 9) & 0xf) : (*(s)++ & 0xf))
1406
8e84507e 1407/*
ccfc67b7
JH
1408=head1 Memory Management
1409
a02a5408 1410=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1411The XSUB-writer's interface to the C C<malloc> function.
1412
c5008215
JC
1413In 5.9.3, Newx() and friends replace the older New() API, and drops
1414the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1415themselves. This aid has been superseded by a new build option,
d10b4965 1416PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1417there for use in XS modules supporting older perls.
1418
a02a5408 1419=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1420The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1421cast. See also C<Newx>.
954c1994 1422
a02a5408 1423=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1424The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1425memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1426
954c1994
GS
1427=for apidoc Am|void|Renew|void* ptr|int nitems|type
1428The XSUB-writer's interface to the C C<realloc> function.
1429
1430=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1431The XSUB-writer's interface to the C C<realloc> function, with
1432cast.
1433
49b8b560 1434=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1435The XSUB-writer's interface to the C C<free> function.
1436
1437=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1438The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1439source, C<dest> is the destination, C<nitems> is the number of items, and
1440C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1441
e90e2364
NC
1442=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1443Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1444optimise.
1445
954c1994
GS
1446=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1447The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1448source, C<dest> is the destination, C<nitems> is the number of items, and
1449C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1450
e90e2364
NC
1451=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1452
1453Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1454optimise.
1455
954c1994
GS
1456=for apidoc Am|void|Zero|void* dest|int nitems|type
1457
1458The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1459destination, C<nitems> is the number of items, and C<type> is the type.
1460
e90e2364
NC
1461=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1462
1463Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1464optimise.
1465
da5d8dbb 1466=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1467This is an architecture-independent macro to copy one structure to another.
954c1994 1468
7e337ee0
JH
1469=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1470
1471Fill up memory with a byte pattern (a byte repeated over and over
1472again) that hopefully catches attempts to access uninitialized memory.
1473
1474=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1475
1476PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1477
1c12ffb4 1478=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1479
1480PoisonWith(0xEF) for catching access to freed memory.
1481
9965345d
JH
1482=for apidoc Am|void|Poison|void* dest|int nitems|type
1483
7e337ee0 1484PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1485
1486=cut */
954c1994 1487
561b68a9
SH
1488/* Maintained for backwards-compatibility only. Use newSV() instead. */
1489#ifndef PERL_CORE
ff06c60c 1490#define NEWSV(x,len) newSV(len)
561b68a9 1491#endif
ff06c60c 1492
19a94d75
JH
1493#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1494
2720f793
JH
1495/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1496 * overly eager compilers that will bleat about e.g.
1497 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1498#ifdef PERL_MALLOC_WRAP
f019c49e
DD
1499#define MEM_WRAP_CHECK(n,t) \
1500 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
27d5b266 1501#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1502 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1503#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1504
f019c49e 1505#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
1506#else
1507
410319be
NC
1508#define MEM_WRAP_CHECK(n,t)
1509#define MEM_WRAP_CHECK_1(n,t,a)
1510#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1511#define MEM_WRAP_CHECK_(n,t)
1512
6fff79ce 1513#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1514
1936d2a7 1515#endif
8b44ba4c 1516
fe4f188c 1517#ifdef PERL_MEM_LOG
46c6c7e2 1518/*
9f653bb5 1519 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1520 * go through functions, which are handy for debugging breakpoints, but
1521 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1522 * line number, and C function name if available) passed in. This info can
1523 * then be used for logging the calls, for which one gets a sample
73d1d973 1524 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1525 *
46c6c7e2 1526 * Known problems:
94e892a6 1527 * - not all memory allocs get logged, only those
46c6c7e2 1528 * that go through Newx() and derivatives (while all
94e892a6 1529 * Safefrees do get logged)
46c6c7e2
JH
1530 * - __FILE__ and __LINE__ do not work everywhere
1531 * - __func__ or __FUNCTION__ even less so
1532 * - I think more goes on after the perlio frees but
1533 * the thing is that STDERR gets closed (as do all
1534 * the file descriptors)
1535 * - no deeper calling stack than the caller of the Newx()
1536 * or the kind, but do I look like a C reflection/introspection
1537 * utility to you?
1538 * - the function prototypes for the logging functions
1539 * probably should maybe be somewhere else than handy.h
1540 * - one could consider inlining (macrofying) the logging
1541 * for speed, but I am too lazy
1542 * - one could imagine recording the allocations in a hash,
1543 * (keyed by the allocation address?), and maintain that
1544 * through reallocs and frees, but how to do that without
1545 * any News() happening...?
73d1d973 1546 * - lots of -Ddefines to get useful/controllable output
b953482e 1547 * - lots of ENV reads
46c6c7e2
JH
1548 */
1549
12754f92 1550PERL_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 1551
12754f92 1552PERL_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 1553
12754f92 1554PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1555
0b0ab801 1556# ifdef PERL_CORE
73d1d973 1557# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1558enum mem_log_type {
1559 MLT_ALLOC,
1560 MLT_REALLOC,
d7a2c63c
MHM
1561 MLT_FREE,
1562 MLT_NEW_SV,
1563 MLT_DEL_SV
0b0ab801
MHM
1564};
1565# endif
12754f92 1566# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1567void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1568void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1569# endif
0b0ab801
MHM
1570# endif
1571
fe4f188c
JH
1572#endif
1573
1574#ifdef PERL_MEM_LOG
d1401ee9
MHM
1575#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1576#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1577#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1578#endif
1579
1580#ifndef MEM_LOG_ALLOC
1581#define MEM_LOG_ALLOC(n,t,a) (a)
1582#endif
1583#ifndef MEM_LOG_REALLOC
1584#define MEM_LOG_REALLOC(n,t,v,a) (a)
1585#endif
1586#ifndef MEM_LOG_FREE
1587#define MEM_LOG_FREE(a) (a)
1588#endif
1589
d1401ee9
MHM
1590#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1591#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1592#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1593
1594#ifndef PERL_CORE
a02a5408
JC
1595/* pre 5.9.x compatibility */
1596#define New(x,v,n,t) Newx(v,n,t)
1597#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1598#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1599#endif
a02a5408 1600
ff68c719 1601#define Renew(v,n,t) \
d1401ee9 1602 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1603#define Renewc(v,n,t,c) \
d1401ee9 1604 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1605
1606#ifdef PERL_POISON
1607#define Safefree(d) \
06c0cc96 1608 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1609#else
fe4f188c 1610#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1611#endif
55497cff 1612
8b44ba4c
NC
1613#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1614#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1615#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1616
8b44ba4c
NC
1617#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1618#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1619#ifdef HAS_MEMSET
8b44ba4c 1620#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1621#else
8b44ba4c
NC
1622/* Using bzero(), which returns void. */
1623#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1624#endif
1625
7e337ee0
JH
1626#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1627#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1628#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1629#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1630
2304df62 1631#ifdef USE_STRUCT_COPY
ff68c719 1632#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1633#else
1634#define StructCopy(s,d,t) Copy(s,d,1,t)
1635#endif
2cc61e15 1636
622913ab 1637#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1638#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1639
2cc61e15
DD
1640#ifdef NEED_VA_COPY
1641# ifdef va_copy
1642# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1643# else
a1866d1b
JH
1644# if defined(__va_copy)
1645# define Perl_va_copy(s, d) __va_copy(d, s)
1646# else
1647# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1648# endif
2cc61e15
DD
1649# endif
1650#endif
1651
472d47bc
SB
1652/* convenience debug macros */
1653#ifdef USE_ITHREADS
1654#define pTHX_FORMAT "Perl interpreter: 0x%p"
1655#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1656#define pTHX_VALUE_ (void *)my_perl,
1657#define pTHX_VALUE (void *)my_perl
1658#define pTHX__VALUE_ ,(void *)my_perl,
1659#define pTHX__VALUE ,(void *)my_perl
472d47bc 1660#else
3609ea0d 1661#define pTHX_FORMAT
472d47bc 1662#define pTHX__FORMAT
3609ea0d 1663#define pTHX_VALUE_
472d47bc 1664#define pTHX_VALUE
3609ea0d 1665#define pTHX__VALUE_
472d47bc
SB
1666#define pTHX__VALUE
1667#endif /* USE_ITHREADS */
3609ea0d 1668
2acdbac1
NC
1669/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1670 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1671 CPAN::Unpack show any users outside the core. */
1672#ifdef PERL_CORE
d1d15184 1673# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1674#endif
1675
9d745869
KW
1676#endif /* HANDY_H */
1677
e9a8c099
MHM
1678/*
1679 * Local variables:
1680 * c-indentation-style: bsd
1681 * c-basic-offset: 4
14d04a33 1682 * indent-tabs-mode: nil
e9a8c099
MHM
1683 * End:
1684 *
14d04a33 1685 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1686 */