This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perlapi.pod: Clarify character classification 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
6aff1f14
KW
487represented by that octet is (or on non-ASCII platforms, corresponds to) an
488ASCII character in the named class based on platform, Unicode, and Perl rules.
489If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 490
6aff1f14
KW
491Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
492with no suffix C<"_A">.
4b9734bf
KW
493
494Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
495onto the platform. That is, the code points that are ASCII are unaffected,
496since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
497as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
498true when called with the code point 0xDF, which is a word character in both
499ASCII and EBCDIC (though it represent different characters in each).
243effed
KW
500
501Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
502point as input. If the code point is larger than 255, Unicode rules are used
503to determine if it is in the character class. For example,
6aff1f14
KW
504C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
505WITH MACRON in Unicode, and is a word character.
243effed
KW
506
507Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
508(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
509classification of just the first (possibly multi-byte) character in the string
510is tested.
243effed
KW
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
6aff1f14
KW
525classification of just the first (possibly multi-byte) character in the string
526is tested.
ccfc67b7 527
954c1994 528=for apidoc Am|bool|isALPHA|char ch
2787a470 529Returns a boolean indicating whether the specified character is an
6aff1f14 530alphabetic character, analogous to C<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
6aff1f14 537alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
15861f94
KW
538See the L<top of this section|/Character classes> for an explanation of variants
539C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 540C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
541C<isALPHANUMERIC_LC_utf8>.
542
8a58bdcf
KW
543=for apidoc Am|bool|isASCII|char ch
544Returns a boolean indicating whether the specified character is one of the 128
243effed 545characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 546On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
547character corresponds to an ASCII character. Variants C<isASCII_A()> and
548C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
549See the L<top of this section|/Character classes> for an explanation of variants
550C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
551C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
552library routine C<isascii()>. In these cases, the variants whose names contain
553C<LC> are the same as the corresponding ones without.
554
555=for apidoc Am|bool|isBLANK|char ch
556Returns a boolean indicating whether the specified character is a
6aff1f14 557character considered to be a blank, analogous to C<m/[[:blank:]]/>.
243effed 558See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 559C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
560C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
561platforms do not have the C library routine C<isblank()>. In these cases, the
562variants whose names contain C<LC> are the same as the corresponding ones
563without.
564
565=for apidoc Am|bool|isCNTRL|char ch
566Returns a boolean indicating whether the specified character is a
6aff1f14 567control character, analogous to C<m/[[:cntrl:]]/>.
243effed 568See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 569C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
570C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
571On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 572
954c1994 573=for apidoc Am|bool|isDIGIT|char ch
2787a470 574Returns a boolean indicating whether the specified character is a
6aff1f14 575digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 576Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed 577See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 578C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
579C<isDIGIT_LC_utf8>.
580
581=for apidoc Am|bool|isGRAPH|char ch
582Returns a boolean indicating whether the specified character is a
6aff1f14 583graphic character, analogous to C<m/[[:graph:]]/>.
243effed 584See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 585C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 586C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 587
0c82b6df 588=for apidoc Am|bool|isLOWER|char ch
2787a470 589Returns a boolean indicating whether the specified character is a
6aff1f14 590lowercase character, analogous to C<m/[[:lower:]]/>.
8a58bdcf 591See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 592C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 593C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 594
c99e91e9 595=for apidoc Am|bool|isOCTAL|char ch
2787a470 596Returns a boolean indicating whether the specified character is an
6aff1f14 597octal digit, [0-7].
243effed
KW
598The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
599C<isOCTAL>.
600
601=for apidoc Am|bool|isPUNCT|char ch
602Returns a boolean indicating whether the specified character is a
6aff1f14
KW
603punctuation character, analogous to C<m/[[:punct:]]/>.
604Note that the definition of what is punctuation isn't as
243effed
KW
605straightforward as one might desire. See L<perlrecharclass/POSIX Character
606Classes> for details.
607See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 608C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 609C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 610
0c82b6df 611=for apidoc Am|bool|isSPACE|char ch
2787a470 612Returns a boolean indicating whether the specified character is a
6aff1f14 613whitespace character. This is analogous
398d098a
KW
614to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
615(experimentally), this also matches what C<m/[[:space:]]/> does.
616("Experimentally" means that this change may be backed out in 5.20 or 5.22 if
617field experience indicates that it was unwise.) Prior to 5.18, only the
618locale forms of this macro (the ones with C<LC> in their names) matched
619precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
620in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
621(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
8a58bdcf 622See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 623C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 624C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 625
398d098a
KW
626=for apidoc Am|bool|isPSXSPC|char ch
627(short for Posix Space)
628Starting in 5.18, this is identical (experimentally) in all its forms to the
629corresponding C<isSPACE()> macros. ("Experimentally" means that this change
630may be backed out in 5.20 or 5.22 if field experience indicates that it
631was unwise.)
632The locale forms of this macro are identical to their corresponding
633C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
634non-locale forms differ from their C<isSPACE()> forms only in that the
635C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
636Otherwise they are identical. Thus this macro is analogous to what
637C<m/[[:space:]]/> matches in a regular expression.
638See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 639C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
640C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
641
954c1994 642=for apidoc Am|bool|isUPPER|char ch
2787a470 643Returns a boolean indicating whether the specified character is an
6aff1f14 644uppercase character, analogous to C<m/[[:upper:]]/>.
8a58bdcf 645See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 646C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 647C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 648
243effed 649=for apidoc Am|bool|isPRINT|char ch
8eea39dd 650Returns a boolean indicating whether the specified character is a
6aff1f14 651printable character, analogous to C<m/[[:print:]]/>.
8a58bdcf 652See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 653C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
654C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
655
656=for apidoc Am|bool|isWORDCHAR|char ch
657Returns a boolean indicating whether the specified character is a character
658that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
659in a regular expression. A word character is an alphabetic character, a
660decimal digit, a connecting punctuation character (such as an underscore), or
661a "mark" character that attaches to one of those (like some sort of accent).
662C<isALNUM()> is a synonym provided for backward compatibility, even though a
663word character includes more than the standard C language meaning of
664alphanumeric.
665See the L<top of this section|/Character classes> for an explanation of variants
666C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
667C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
668
669=for apidoc Am|bool|isXDIGIT|char ch
670Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
671digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
672and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
673See the L<top of this section|/Character classes> for an explanation of variants
674C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
675C<isXDIGIT_LC_utf8>.
676
3c3ecf18
KW
677=for apidoc Am|bool|isIDFIRST|char ch
678Returns a boolean indicating whether the specified character can be the first
679character of an identifier. This is very close to, but not quite the same as
680the official Unicode property C<XID_Start>. The difference is that this
681returns true only if the input character also matches L</isWORDCHAR>.
682See the L<top of this section|/Character classes> for an explanation of variants
683C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
684C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
685
686=for apidoc Am|bool|isIDCONT|char ch
687Returns a boolean indicating whether the specified character can be the
688second or succeeding character of an identifier. This is very close to, but
689not quite the same as the official Unicode property C<XID_Continue>. The
690difference is that this returns true only if the input character also matches
691L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
692explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
693C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
694C<isIDCONT_LC_utf8>.
695
243effed 696=head1 Miscellaneous Functions
8eea39dd 697
95a59cab 698=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 699Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
700Behaviour is only well defined when isXDIGIT(*str) is true.
701
e7c1e6c1
KW
702=head1 Character case changing
703
954c1994 704=for apidoc Am|char|toUPPER|char ch
6aff1f14
KW
705Converts the specified character to uppercase, if possible; otherwise returns
706the input character itself.
954c1994
GS
707
708=for apidoc Am|char|toLOWER|char ch
6aff1f14
KW
709Converts the specified character to lowercase, if possible; otherwise returns
710the input character itself.
954c1994
GS
711
712=cut
353c9b6f 713
6aff1f14 714XXX Still undocumented isVERTWS_uni and _utf8, and the other toUPPER etc functions
243effed 715
8a58bdcf 716Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
717patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
718
954c1994
GS
719*/
720
7c062697
KW
721/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
722 * is known only in the perl core, and this macro can be called from outside
723 * that */
724#ifdef HAS_QUAD
725# define WIDEST_UTYPE U64TYPE
726#else
727# define WIDEST_UTYPE U32
728#endif
729
3912bc88
KW
730/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
731 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
732 * bits of information are lost even on a 64-bit machine, but to get the
733 * compiler to optimize it out if possible. This is because Configure makes
734 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
735 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
736 */
737#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 738
41f43cc2 739#ifdef EBCDIC
ff7eba81 740# define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128))
41f43cc2
KW
741#else
742# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
743#endif
744
ec79cfbb 745#define isASCII_A(c) isASCII(c)
3509bb53 746#define isASCII_L1(c) isASCII(c)
6515cb76 747
c2da0b36
KW
748/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
749 * and the 8 possible permutations of those bits exactly comprise the 8 octal
750 * digits */
3d33b3e3 751#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 752
ec79cfbb 753/* ASCII range only */
9fb1bf9d 754#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 755
a500dc72
KW
756/* Character class numbers. For internal core Perl use only. The ones less
757 * than 32 are used in PL_charclass[] and the ones up through the one that
758 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
759 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
760 * their actual definitions are here. If that file has a name not used here,
761 * it won't compile.
1709d539
KW
762 *
763 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
764 * frequency of their use. This gives a slight edge to exiting a loop earlier
765 * (in reginclass() in regexec.c) */
1709d539
KW
766# define _CC_WORDCHAR 0 /* \w and [:word:] */
767# define _CC_DIGIT 1 /* \d and [:digit:] */
768# define _CC_ALPHA 2 /* [:alpha:] */
769# define _CC_LOWER 3 /* [:lower:] */
770# define _CC_UPPER 4 /* [:upper:] */
771# define _CC_PUNCT 5 /* [:punct:] */
772# define _CC_PRINT 6 /* [:print:] */
15861f94 773# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 774# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 775# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 776
b0d691b2 777#define _FIRST_NON_SWASH_CC 10
1709d539
KW
778/* The character classes above are implemented with swashes. The second group
779 * (just below) contains the ones implemented without. These are also sorted
780 * in rough order of the frequency of their use, except that \v should be last,
781 * as it isn't a real Posix character class, and some (small) inefficiencies in
782 * regular expression handling would be introduced by putting it in the middle
783 * of those that are. Also, cntrl and ascii come after the others as it may be
784 * useful to group these which have no members that match above Latin1, (or
785 * above ASCII in the latter case) */
786
b0d691b2
KW
787# define _CC_SPACE 10 /* \s */
788# define _CC_BLANK 11 /* [:blank:] */
789# define _CC_XDIGIT 12 /* [:xdigit:] */
790# define _CC_PSXSPC 13 /* [:space:] */
791# define _CC_CNTRL 14 /* [:cntrl:] */
792# define _CC_ASCII 15 /* [:ascii:] */
793# define _CC_VERTSPACE 16 /* \v */
1709d539 794
a0947d7b
KW
795# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
796
1709d539
KW
797/* The members of the third group below do not need to be coordinated with data
798 * structures in regcomp.[ch] and regexec.c */
b0d691b2
KW
799# define _CC_IDFIRST 17
800# define _CC_CHARNAME_CONT 18
801# define _CC_NONLATIN1_FOLD 19
802# define _CC_QUOTEMETA 20
803# define _CC_NON_FINAL_FOLD 21
804# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
805# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
806/* Unused: 23-30
f4cdb42c
KW
807 * If more bits are needed, one could add a second word for non-64bit
808 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
809 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
810 * is used only for optimization (as of this writing), and differs in the
811 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
812 * characters: the masculine and feminine ordinal indicators, so removing it
813 * would just cause /i regexes which match them to run less efficiently */
96ac0975 814
3a371f2f
KW
815#if defined(PERL_CORE) || defined(PERL_EXT)
816/* An enum version of the character class numbers, to help compilers
817 * optimize */
818typedef enum {
3a371f2f 819 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
820 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
821 _CC_ENUM_ASCII = _CC_ASCII,
822 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 823 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 824 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
825 _CC_ENUM_DIGIT = _CC_DIGIT,
826 _CC_ENUM_GRAPH = _CC_GRAPH,
827 _CC_ENUM_LOWER = _CC_LOWER,
828 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 829 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 830 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 831 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 832 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 833 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 834 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 835 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
836} _char_class_number;
837#endif
838
63c61c3f 839#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 840#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 841
3018b823 842#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 843# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
844 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
845 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
846 #error Need to adjust order of swash_property_names[]
847# endif
848
b44daeeb
KW
849/* This is declared static in each of the few files that this is #defined for
850 * to keep them from being publicly accessible. Hence there is a small amount
851 * of wasted space */
852
922e8cb4
KW
853static const char* const swash_property_names[] = {
854 "XPosixWord",
855 "XPosixDigit",
856 "XPosixAlpha",
857 "XPosixLower",
858 "XPosixUpper",
859 "XPosixPunct",
860 "XPosixPrint",
861 "XPosixAlnum",
b0d691b2
KW
862 "XPosixGraph",
863 "Cased"
922e8cb4
KW
864};
865#endif
866
96ac0975
NC
867# ifdef DOINIT
868EXTCONST U32 PL_charclass[] = {
869# include "l1_char_class_tab.h"
870};
871
872# else /* ! DOINIT */
873EXTCONST U32 PL_charclass[];
874# endif
875
265c1f46 876 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70
KW
877# define _CC_mask(classnum) (1U << (classnum))
878# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
926bb54c 879 && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask(classnum)))
4eeeb416 880
f4cdb42c
KW
881 /* The mask for the _A versions of the macros; it just adds in the bit for
882 * ASCII. */
883# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
884
885 /* The _A version makes sure that both the desired bit and the ASCII bit
886 * are present */
887# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
888 && ((PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask_A(classnum)) \
889 == _CC_mask_A(classnum)))
890
f4cdb42c 891# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 892# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
893# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
894# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
895# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
896# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
897# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
898# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
899# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
900# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
901# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
902# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
903# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
904# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
905# define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST))
906
fb9bfbf7
KW
907 /* Either participates in a fold with a character above 255, or is a
908 * multi-char fold */
430b7c70
KW
909# 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)))
910
4eeeb416 911# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
912# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
913 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
914# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
915 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
9fb1bf9d 916#else /* No perl.h. */
9fb1bf9d 917# ifdef EBCDIC
9fb1bf9d 918# define isALPHA_A(c) (isASCII(c) && isALPHA(c))
15861f94 919# define isALPHANUMERIC_A(c) (isASCII(c) && isALPHANUMERIC(c))
9fb1bf9d
KW
920# define isBLANK_A(c) (isASCII(c) && isBLANK(c))
921# define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
922# define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
923# define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
924# define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
925# define isLOWER_A(c) (isASCII(c) && isLOWER(c))
926# define isPRINT_A(c) (isASCII(c) && isPRINT(c))
927# define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
928# define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
929# define isSPACE_A(c) (isASCII(c) && isSPACE(c))
930# define isUPPER_A(c) (isASCII(c) && isUPPER(c))
931# define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
932# define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
933# else /* ASCII platform, no perl.h */
9fb1bf9d 934# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
15861f94 935# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
9fb1bf9d 936# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
926bb54c 937# define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
2181da29 938# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
9fb1bf9d
KW
939# define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
940# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
941# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
942# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
943# define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
926bb54c
KW
944# define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) \
945 || ((c) >= 58 && (c) <= 64) \
946 || ((c) >= 91 && (c) <= 96) \
947 || ((c) >= 123 && (c) <= 126))
948# define isSPACE_A(c) ((c) == ' ' \
949 || (c) == '\t' \
950 || (c) == '\n' \
951 || (c) =='\r' \
952 || (c) == '\f')
2181da29 953# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
9fb1bf9d 954# define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
926bb54c
KW
955# define isXDIGIT_A(c) (isDIGIT_A(c) \
956 || ((c) >= 'a' && (c) <= 'f') \
957 || ((c) <= 'F' && (c) >= 'A'))
9fb1bf9d
KW
958# endif
959#endif /* ASCII range definitions */
ec79cfbb
KW
960
961/* Latin1 definitions */
8a58bdcf 962#ifdef H_PERL
f4cdb42c 963# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
15861f94 964# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
f4cdb42c 965# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
e5dcd934 966
8a58bdcf 967/* continuation character for legal NAME in \N{NAME} */
4eeeb416 968# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
e5dcd934 969
f4cdb42c
KW
970# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
971# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
972# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
973# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
974# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
975# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
976# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
977# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
978# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
979# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
8a58bdcf
KW
980#else /* No access to perl.h. Only a few provided here, just in case needed
981 * for backwards compatibility */
982 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
983 * >= AA test to speed up ASCII-only tests at the expense of the others */
984# define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
985 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
926bb54c 986 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
8a58bdcf
KW
987 || NATIVE_TO_UNI((U8) c) == 0xAA \
988 || NATIVE_TO_UNI((U8) c) == 0xB5 \
989 || NATIVE_TO_UNI((U8) c) == 0xBA)))
926bb54c
KW
990# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
991 || (c) == ' ' \
992 || (c) == '-' \
993 || (c) == '(' \
994 || (c) == ')' \
995 || (c) == ':' \
996 || NATIVE_TO_UNI((U8) c) == 0xA0)
8a58bdcf
KW
997#endif
998
8a58bdcf
KW
999/* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
1000 * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
1001 * therefore can use native functions. For those where C89 doesn't define a
1002 * function, use our function, assuming that the EBCDIC code page is isomorphic
1003 * with Latin1, which the three currently recognized by Perl are. Some libc's
1004 * have an isblank(), but it's not guaranteed. */
9d116dd7 1005#ifdef EBCDIC
271b5cb3 1006# define isALPHA(c) isalpha(c)
15861f94 1007# define isALPHANUMERIC(c) isalnum(c)
0bc70f7f 1008# define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
b8c5462f 1009# define isCNTRL(c) iscntrl(c)
9916188f 1010# define isDIGIT(c) isdigit(c)
b8c5462f 1011# define isGRAPH(c) isgraph(c)
8a58bdcf 1012# define isIDFIRST(c) (isALPHA(c) || (c) == '_')
7d79683e 1013# define isLOWER(c) islower(c)
9d116dd7 1014# define isPRINT(c) isprint(c)
56e4b611 1015# define isPSXSPC(c) isspace(c)
b8c5462f 1016# define isPUNCT(c) ispunct(c)
f1f70107
KW
1017# define isSPACE(c) (isPSXSPC(c) /* && (c) != '\v' (Experimentally making
1018 these macros identical) */)
7d79683e 1019# define isUPPER(c) isupper(c)
b8c5462f 1020# define isXDIGIT(c) isxdigit(c)
8a58bdcf 1021# define isWORDCHAR(c) (isalnum(c) || (c) == '_')
9d116dd7 1022# define toLOWER(c) tolower(c)
7d79683e 1023# define toUPPER(c) toupper(c)
ec79cfbb 1024#else /* Not EBCDIC: ASCII-only matching */
15861f94 1025# define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
271b5cb3 1026# define isALPHA(c) isALPHA_A(c)
ec79cfbb
KW
1027# define isBLANK(c) isBLANK_A(c)
1028# define isCNTRL(c) isCNTRL_A(c)
1029# define isDIGIT(c) isDIGIT_A(c)
1030# define isGRAPH(c) isGRAPH_A(c)
8a58bdcf 1031# define isIDFIRST(c) isIDFIRST_A(c)
ec79cfbb
KW
1032# define isLOWER(c) isLOWER_A(c)
1033# define isPRINT(c) isPRINT_A(c)
1034# define isPSXSPC(c) isPSXSPC_A(c)
1035# define isPUNCT(c) isPUNCT_A(c)
1036# define isSPACE(c) isSPACE_A(c)
1037# define isUPPER(c) isUPPER_A(c)
8a58bdcf 1038# define isWORDCHAR(c) isWORDCHAR_A(c)
ec79cfbb 1039# define isXDIGIT(c) isXDIGIT_A(c)
00f254e2 1040
8a58bdcf
KW
1041 /* ASCII casing. These could also be written as
1042 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1043 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1044 which uses table lookup and mask instead of subtraction. (This would
1045 work because the _MOD does not apply in the ASCII range) */
9d116dd7 1046# define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
7d79683e 1047# define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
9d116dd7 1048#endif
bbce6d69 1049
1a0901db
KW
1050
1051/* Use table lookup for speed; return error character for input
1052 * out-of-range */
1053#define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \
1054 ? UNI_TO_NATIVE(PL_latin1_lc[ \
1055 NATIVE_TO_UNI( (U8) (c)) ]) \
1056 : UNICODE_REPLACEMENT)
1057/* Modified uc. Is correct uc except for three non-ascii chars which are
1058 * all mapped to one of them, and these need special handling; error
1059 * character for input out-of-range */
1060#define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \
1061 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \
1062 NATIVE_TO_UNI( (U8) (c)) ]) \
1063 : UNICODE_REPLACEMENT)
1064
bbce6d69
PP
1065#ifdef USE_NEXT_CTYPE
1066
15861f94 1067# define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c))
ff68c719 1068# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 1069# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 1070# define isBLANK_LC(c) isBLANK((unsigned int)(c))
b8c5462f 1071# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
509fb054 1072# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
b8c5462f 1073# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
509fb054
KW
1074# define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
1075# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
ff68c719 1076# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 1077# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
509fb054
KW
1078# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
1079# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
1080# define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
e712593e 1081# define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
ff68c719 1082# define toLOWER_LC(c) NXToLower((unsigned int)(c))
509fb054 1083# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
bbce6d69
PP
1084
1085#else /* !USE_NEXT_CTYPE */
b8c5462f 1086
bbce6d69
PP
1087# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1088
76fde96a 1089/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
851131d3 1090
76fde96a 1091# define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
15861f94
KW
1092# define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \
1093 && isalnum((unsigned char)(c)))
07315176 1094# ifdef HAS_ISASCII
76fde96a 1095# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
07315176 1096# else
76fde96a 1097# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
07315176 1098# endif
67addccf 1099# ifdef HAS_ISBLANK
76fde96a 1100# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
67addccf 1101# else
76fde96a 1102# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
67addccf 1103# endif
76fde96a 1104# define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
509fb054 1105# define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
76fde96a 1106# define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
509fb054
KW
1107# define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1108 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1109# define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
76fde96a
KW
1110# define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1111# define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
509fb054
KW
1112# define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1113# define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1114# define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1115 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
e712593e 1116# define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
76fde96a 1117# define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? tolower((unsigned char)(c)) : (c))
509fb054 1118# define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? toupper((unsigned char)(c)) : (c))
bbce6d69
PP
1119
1120# else
1121
bbce6d69 1122# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1123# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
07315176 1124# define isASCII_LC(c) isascii(c)
67addccf
KW
1125# ifdef HAS_ISBLANK
1126# define isBLANK_LC(c) (isascii(c) && isblank(c))
1127# else
3e66ab74 1128# define isBLANK_LC(c) isBLANK_A(c)
67addccf 1129# endif
b8c5462f 1130# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1131# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1132# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1133# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1134# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1135# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1136# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1137# define isSPACE_LC(c) (isascii(c) && isspace(c))
1138# define isUPPER_LC(c) (isascii(c) && isupper(c))
1139# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1140# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
76fde96a 1141# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1142# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
bbce6d69
PP
1143
1144# endif
a0d0e21e 1145#endif /* USE_NEXT_CTYPE */
55204971 1146
eba68aa0
KW
1147#define isIDCONT(c) isWORDCHAR(c)
1148#define isIDCONT_A(c) isWORDCHAR_A(c)
1149#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1150#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1151#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1152
66c17564 1153/* For internal core Perl use only. If the input is Latin1, use the Latin1
a500dc72
KW
1154 * macro; otherwise use the function 'above_latin1'. Won't compile if 'c' isn't unsigned, as
1155 * won't match above_latin1 prototype. The macros do bounds checking, so have
66c17564
KW
1156 * duplicate checks here, so could create versions of the macros that don't,
1157 * but experiments show that gcc optimizes them out anyway. */
1158
1159/* Note that all ignore 'use bytes' */
cd500f2f
KW
1160#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1161 ? _generic_isCC(c, classnum) \
1162 : above_latin1(c))
1163#define _generic_swash_uni(classnum, c) ((c) < 256 \
1164 ? _generic_isCC(c, classnum) \
922e8cb4 1165 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1166#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1167#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1168#define isASCII_uni(c) isASCII(c)
1169#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1170#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1171#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1172#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1173#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1174#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1175#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1176#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1177
509fb054
KW
1178/* Posix and regular space are identical above Latin1 */
1179#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1180
cd500f2f 1181#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1182#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1183#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1184#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1185#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1186#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1187
b8d68ded 1188#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1189#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1190#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1191#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1192
cd500f2f
KW
1193#define _generic_LC_uvchr(latin1, above_latin1, c) \
1194 (c < 256 ? latin1(c) : above_latin1(c))
1195#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1196 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1197
1198#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1199#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1200 _CC_ALPHANUMERIC, c)
509fb054 1201#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1202#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1203#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1204#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1205#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1206#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1207 _is_uni_perl_idcont, c)
cd500f2f
KW
1208#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1209 _is_uni_perl_idstart, c)
1210#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1211#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1212#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1213 space under locale */
cd500f2f
KW
1214#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1215#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1216 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1217#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1218#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1219 _CC_WORDCHAR, c)
1220#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1221
a0ed51b3 1222
36b00f93 1223#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1224
a500dc72
KW
1225/* Everything whose name begins with an underscore is for internal core Perl
1226 * use only. */
1227
1228/* If the input is in the Latin1 range, use
252810d0
KW
1229 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1230 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1231 * fact that ASCII characters have the same representation whether utf8 or not.
1232 * Note that it assumes that the utf8 has been validated, and ignores 'use
1233 * bytes' */
cd500f2f 1234#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1235 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1236 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c
KW
1237 ? _generic_isCC( \
1238 TWO_BYTE_UTF8_TO_UNI(*(p), \
1239 *((p)+1 )), \
1240 classnum) \
252810d0 1241 : utf8)
cd500f2f
KW
1242/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1243 * can be a macro */
1244#define _generic_func_utf8(classnum, above_latin1, p) \
1245 _generic_utf8(classnum, p, above_latin1(p))
1246/* Like the above, but passes classnum to _isFOO_utf8(), instead of having a
1247 * 'above_latin1' parameter */
922e8cb4 1248#define _generic_swash_utf8(classnum, p) \
cd500f2f 1249 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1250
cc8ab7c0
KW
1251/* Like the above, but should be used only when it is known that there are no
1252 * characters in the range 128-255 which the class is TRUE for. Hence it can
cd500f2f
KW
1253 * skip the tests for this range. 'above_latin1' should include its arguments */
1254#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1255 (UTF8_IS_INVARIANT(*(p)) \
1256 ? _generic_isCC(*(p), classnum) \
1257 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
cd500f2f 1258 ? above_latin1 \
cc8ab7c0
KW
1259 : 0)
1260
84238efa
KW
1261/* NOTE that some of these macros have very similar ones in regcharclass.h.
1262 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1263 * there, differing in name only by an underscore from the one here
1264 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1265 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1266 * points; the regcharclass.h ones are implemented as a series of
1267 * "if-else-if-else ..." */
1268
922e8cb4
KW
1269#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1270#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1271#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1272 utf8, the non-utf8 macro works
1273 */
cd500f2f
KW
1274#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1275#define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1276#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1277 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1278#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1279#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1280 _is_utf8_perl_idcont, p)
e5dcd934 1281
c11ff943
KW
1282/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1283 * IDFIRST is an alnum. See
c8362b00 1284 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1285 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1286 * This used to be not the XID version, but we decided to go with the more
1287 * modern Unicode definition */
cd500f2f 1288#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1289 _is_utf8_perl_idstart, p)
e5dcd934 1290
922e8cb4
KW
1291#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1292#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1293
509fb054 1294/* Posix and regular space are identical above Latin1 */
cd500f2f 1295#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1296
922e8cb4 1297#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1298#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1299#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1300#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1301#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1302#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1303 is_XDIGIT_high(p))
a0ed51b3 1304
509fb054
KW
1305#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1306#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1307#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1308
9201a572
KW
1309/* For internal core Perl use only. If the input is in the Latin1 range, use
1310 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1311 * use the value given by the 'utf8' parameter. This relies on the fact that
1312 * ASCII characters have the same representation whether utf8 or not. Note
1313 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
cd500f2f 1314#define _generic_LC_utf8(macro, p, utf8) \
9201a572
KW
1315 (UTF8_IS_INVARIANT(*(p)) \
1316 ? macro(*(p)) \
1317 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1318 ? macro(TWO_BYTE_UTF8_TO_UNI(*(p), *((p)+1))) \
1319 : utf8)
1320
cd500f2f
KW
1321#define _generic_LC_swash_utf8(macro, classnum, p) \
1322 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1323#define _generic_LC_func_utf8(macro, above_latin1, p) \
1324 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1325
cd500f2f
KW
1326#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1327 _CC_ALPHANUMERIC, p)
1328#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1329#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1330#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1331#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1332#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1333#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1334#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1335#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1336#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1337#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1338#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1339 space under locale */
cd500f2f
KW
1340#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1341#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1342#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1343#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1344 _CC_WORDCHAR, p)
1345#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1346
fbc19f27
KW
1347/* Macros for backwards compatibility and for completeness when the ASCII and
1348 * Latin1 values are identical */
fbc19f27
KW
1349#define isALPHAU(c) isALPHA_L1(c)
1350#define isDIGIT_L1(c) isDIGIT_A(c)
1351#define isOCTAL(c) isOCTAL_A(c)
1352#define isOCTAL_L1(c) isOCTAL_A(c)
1353#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1354#define isALNUM(c) isWORDCHAR(c)
1355#define isALNUMU(c) isWORDCHAR_L1(c)
1356#define isALNUM_LC(c) isWORDCHAR_LC(c)
1357#define isALNUM_uni(c) isWORDCHAR_uni(c)
1358#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1359#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1360#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1361#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1362#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1363#define isALNUMC(c) isALPHANUMERIC(c)
1364#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1365#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1366#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1367#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1368#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1369
6ae9f32a 1370/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1371 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1372 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1373 * that is EOT on ASCII; ST on EBCDIC */
6ae9f32a 1374# define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64)
bbce6d69 1375
dea28490
JJ
1376/* Line numbers are unsigned, 32 bits. */
1377typedef U32 line_t;
e5dcd934 1378#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1379
91152fc1
DG
1380/* Helpful alias for version prescan */
1381#define is_LAX_VERSION(a,b) \
1382 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1383
1384#define is_STRICT_VERSION(a,b) \
1385 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1386
1387#define BADVERSION(a,b,c) \
1388 if (b) { \
1389 *b = c; \
1390 } \
1391 return a;
8c52afec 1392
95a59cab
YO
1393#define READ_XDIGIT(s) (isALPHA(*(s)) ? ((*(s)++ + 9) & 0xf) : (*(s)++ & 0xf))
1394
8e84507e 1395/*
ccfc67b7
JH
1396=head1 Memory Management
1397
a02a5408 1398=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1399The XSUB-writer's interface to the C C<malloc> function.
1400
c5008215
JC
1401In 5.9.3, Newx() and friends replace the older New() API, and drops
1402the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1403themselves. This aid has been superseded by a new build option,
d10b4965 1404PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1405there for use in XS modules supporting older perls.
1406
a02a5408 1407=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1408The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1409cast. See also C<Newx>.
954c1994 1410
a02a5408 1411=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1412The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1413memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1414
954c1994
GS
1415=for apidoc Am|void|Renew|void* ptr|int nitems|type
1416The XSUB-writer's interface to the C C<realloc> function.
1417
1418=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1419The XSUB-writer's interface to the C C<realloc> function, with
1420cast.
1421
49b8b560 1422=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1423The XSUB-writer's interface to the C C<free> function.
1424
1425=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1426The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1427source, C<dest> is the destination, C<nitems> is the number of items, and
1428C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1429
e90e2364
NC
1430=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1431Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1432optimise.
1433
954c1994
GS
1434=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1435The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1436source, C<dest> is the destination, C<nitems> is the number of items, and
1437C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1438
e90e2364
NC
1439=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1440
1441Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1442optimise.
1443
954c1994
GS
1444=for apidoc Am|void|Zero|void* dest|int nitems|type
1445
1446The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1447destination, C<nitems> is the number of items, and C<type> is the type.
1448
e90e2364
NC
1449=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1450
1451Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1452optimise.
1453
da5d8dbb 1454=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1455This is an architecture-independent macro to copy one structure to another.
954c1994 1456
7e337ee0
JH
1457=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1458
1459Fill up memory with a byte pattern (a byte repeated over and over
1460again) that hopefully catches attempts to access uninitialized memory.
1461
1462=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1463
1464PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1465
1c12ffb4 1466=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1467
1468PoisonWith(0xEF) for catching access to freed memory.
1469
9965345d
JH
1470=for apidoc Am|void|Poison|void* dest|int nitems|type
1471
7e337ee0 1472PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1473
1474=cut */
954c1994 1475
561b68a9
SH
1476/* Maintained for backwards-compatibility only. Use newSV() instead. */
1477#ifndef PERL_CORE
ff06c60c 1478#define NEWSV(x,len) newSV(len)
561b68a9 1479#endif
ff06c60c 1480
19a94d75
JH
1481#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1482
2720f793
JH
1483/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1484 * overly eager compilers that will bleat about e.g.
1485 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1486#ifdef PERL_MALLOC_WRAP
f019c49e 1487#define MEM_WRAP_CHECK(n,t) \
43387ee1 1488 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_memory_wrap(),0))
27d5b266 1489#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1490 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1491#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1492
43387ee1 1493#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
1494#else
1495
410319be
NC
1496#define MEM_WRAP_CHECK(n,t)
1497#define MEM_WRAP_CHECK_1(n,t,a)
1498#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1499#define MEM_WRAP_CHECK_(n,t)
1500
6fff79ce 1501#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1502
1936d2a7 1503#endif
8b44ba4c 1504
fe4f188c 1505#ifdef PERL_MEM_LOG
46c6c7e2 1506/*
9f653bb5 1507 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1508 * go through functions, which are handy for debugging breakpoints, but
1509 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1510 * line number, and C function name if available) passed in. This info can
1511 * then be used for logging the calls, for which one gets a sample
73d1d973 1512 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1513 *
46c6c7e2 1514 * Known problems:
94e892a6 1515 * - not all memory allocs get logged, only those
46c6c7e2 1516 * that go through Newx() and derivatives (while all
94e892a6 1517 * Safefrees do get logged)
46c6c7e2
JH
1518 * - __FILE__ and __LINE__ do not work everywhere
1519 * - __func__ or __FUNCTION__ even less so
1520 * - I think more goes on after the perlio frees but
1521 * the thing is that STDERR gets closed (as do all
1522 * the file descriptors)
1523 * - no deeper calling stack than the caller of the Newx()
1524 * or the kind, but do I look like a C reflection/introspection
1525 * utility to you?
1526 * - the function prototypes for the logging functions
1527 * probably should maybe be somewhere else than handy.h
1528 * - one could consider inlining (macrofying) the logging
1529 * for speed, but I am too lazy
1530 * - one could imagine recording the allocations in a hash,
1531 * (keyed by the allocation address?), and maintain that
1532 * through reallocs and frees, but how to do that without
1533 * any News() happening...?
73d1d973 1534 * - lots of -Ddefines to get useful/controllable output
b953482e 1535 * - lots of ENV reads
46c6c7e2
JH
1536 */
1537
12754f92 1538PERL_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 1539
12754f92 1540PERL_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 1541
12754f92 1542PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1543
0b0ab801 1544# ifdef PERL_CORE
73d1d973 1545# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1546enum mem_log_type {
1547 MLT_ALLOC,
1548 MLT_REALLOC,
d7a2c63c
MHM
1549 MLT_FREE,
1550 MLT_NEW_SV,
1551 MLT_DEL_SV
0b0ab801
MHM
1552};
1553# endif
12754f92 1554# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1555void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1556void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1557# endif
0b0ab801
MHM
1558# endif
1559
fe4f188c
JH
1560#endif
1561
1562#ifdef PERL_MEM_LOG
d1401ee9
MHM
1563#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1564#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1565#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1566#endif
1567
1568#ifndef MEM_LOG_ALLOC
1569#define MEM_LOG_ALLOC(n,t,a) (a)
1570#endif
1571#ifndef MEM_LOG_REALLOC
1572#define MEM_LOG_REALLOC(n,t,v,a) (a)
1573#endif
1574#ifndef MEM_LOG_FREE
1575#define MEM_LOG_FREE(a) (a)
1576#endif
1577
d1401ee9
MHM
1578#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1579#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1580#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1581
1582#ifndef PERL_CORE
a02a5408
JC
1583/* pre 5.9.x compatibility */
1584#define New(x,v,n,t) Newx(v,n,t)
1585#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1586#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1587#endif
a02a5408 1588
ff68c719 1589#define Renew(v,n,t) \
d1401ee9 1590 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1591#define Renewc(v,n,t,c) \
d1401ee9 1592 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1593
1594#ifdef PERL_POISON
1595#define Safefree(d) \
06c0cc96 1596 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1597#else
fe4f188c 1598#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1599#endif
55497cff 1600
8b44ba4c
NC
1601#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1602#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1603#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1604
8b44ba4c
NC
1605#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1606#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1607#ifdef HAS_MEMSET
8b44ba4c 1608#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1609#else
8b44ba4c
NC
1610/* Using bzero(), which returns void. */
1611#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1612#endif
1613
7e337ee0
JH
1614#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1615#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1616#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1617#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1618
2304df62 1619#ifdef USE_STRUCT_COPY
ff68c719 1620#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1621#else
1622#define StructCopy(s,d,t) Copy(s,d,1,t)
1623#endif
2cc61e15 1624
622913ab 1625#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1626#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1627
2cc61e15
DD
1628#ifdef NEED_VA_COPY
1629# ifdef va_copy
1630# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1631# else
a1866d1b
JH
1632# if defined(__va_copy)
1633# define Perl_va_copy(s, d) __va_copy(d, s)
1634# else
1635# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1636# endif
2cc61e15
DD
1637# endif
1638#endif
1639
472d47bc
SB
1640/* convenience debug macros */
1641#ifdef USE_ITHREADS
1642#define pTHX_FORMAT "Perl interpreter: 0x%p"
1643#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1644#define pTHX_VALUE_ (void *)my_perl,
1645#define pTHX_VALUE (void *)my_perl
1646#define pTHX__VALUE_ ,(void *)my_perl,
1647#define pTHX__VALUE ,(void *)my_perl
472d47bc 1648#else
3609ea0d 1649#define pTHX_FORMAT
472d47bc 1650#define pTHX__FORMAT
3609ea0d 1651#define pTHX_VALUE_
472d47bc 1652#define pTHX_VALUE
3609ea0d 1653#define pTHX__VALUE_
472d47bc
SB
1654#define pTHX__VALUE
1655#endif /* USE_ITHREADS */
3609ea0d 1656
2acdbac1
NC
1657/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1658 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1659 CPAN::Unpack show any users outside the core. */
1660#ifdef PERL_CORE
d1d15184 1661# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1662#endif
1663
9d745869
KW
1664#endif /* HANDY_H */
1665
e9a8c099
MHM
1666/*
1667 * Local variables:
1668 * c-indentation-style: bsd
1669 * c-basic-offset: 4
14d04a33 1670 * indent-tabs-mode: nil
e9a8c099
MHM
1671 * End:
1672 *
14d04a33 1673 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1674 */