This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Let while.t run from the top level
[perl5.git] / handy.h
CommitLineData
a0d0e21e 1/* handy.h
a687059c 2 *
1129b882
NC
3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000,
4 * 2001, 2002, 2004, 2005, 2006, 2007, 2008 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
85e6fe83 11#if !defined(__STDC__)
378cc40b
LW
12#ifdef NULL
13#undef NULL
14#endif
a687059c
LW
15#ifndef I286
16# define NULL 0
17#else
18# define NULL 0L
19#endif
85e6fe83
LW
20#endif
21
24792b8d
NC
22#ifndef PERL_CORE
23# define Null(type) ((type)NULL)
954c1994
GS
24
25/*
ccfc67b7 26=head1 Handy Values
954c1994 27
3609ea0d 28=for apidoc AmU||Nullch
24792b8d 29Null character pointer. (No longer available when C<PERL_CORE> is defined.)
2307c6d0 30
954c1994 31=for apidoc AmU||Nullsv
24792b8d 32Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
33
34=cut
35*/
36
24792b8d
NC
37# define Nullch Null(char*)
38# define Nullfp Null(PerlIO*)
39# define Nullsv Null(SV*)
40#endif
8d063cd8 41
641d3f0b
PP
42#ifdef TRUE
43#undef TRUE
44#endif
45#ifdef FALSE
46#undef FALSE
47#endif
48#define TRUE (1)
49#define FALSE (0)
50
cf3f0ffb
DM
51/* The MUTABLE_*() macros cast pointers to the types shown, in such a way
52 * (compiler permitting) that casting away const-ness will give a warning;
53 * e.g.:
54 *
55 * const SV *sv = ...;
56 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
57 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
58 */
59
b1bc3f34 60#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 61# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
62#else
63# define MUTABLE_PTR(p) ((void *) (p))
64#endif
65
a062e10d 66#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 67#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 68#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 69#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 70#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 71#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 72
bd31be4b
NC
73#ifdef I_STDBOOL
74# include <stdbool.h>
75# ifndef HAS_BOOL
76# define HAS_BOOL 1
77# endif
78#endif
79
8e84507e 80/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
81 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
82 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 83 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
84 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
85 g++ can be identified by __GNUG__.
86 Andy Dougherty February 2000
5d94fbed 87*/
3609ea0d 88#ifdef __GNUG__ /* GNU g++ has bool built-in */
5d94fbed 89# ifndef HAS_BOOL
c1d22f6b 90# define HAS_BOOL 1
5d94fbed 91# endif
5d94fbed
AD
92#endif
93
641d3f0b
PP
94/* The NeXT dynamic loader headers will not build with the bool macro
95 So declare them now to clear confusion.
96*/
8f1f23e8 97#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
98# undef FALSE
99# undef TRUE
100 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
101# define ENUM_BOOL 1
102# ifndef HAS_BOOL
103# define HAS_BOOL 1
104# endif /* !HAS_BOOL */
8f1f23e8 105#endif /* NeXT || __NeXT__ */
641d3f0b 106
5d94fbed 107#ifndef HAS_BOOL
61bb5906 108# if defined(UTS) || defined(VMS)
5d94fbed
AD
109# define bool int
110# else
111# define bool char
112# endif
c1d22f6b 113# define HAS_BOOL 1
a687059c 114#endif
0d3e774c 115
f2338a2e
DM
116/* a simple (bool) cast may not do the right thing: if bool is defined
117 * as char for example, then the cast from int is implementation-defined
18f5643b 118 * (bool)!!(cbool) in a ternary triggers a bug in xlc on AIX
f2338a2e
DM
119 */
120
18f5643b 121#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 122
46c6c7e2 123/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
124 * XXX Should really be a Configure probe, with HAS__FUNCTION__
125 * and FUNCTION__ as results.
126 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
127#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
128# define FUNCTION__ __func__
129#else
130# if (defined(_MSC_VER) && _MSC_VER < 1300) || /* Pre-MSVC 7.0 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 131 (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
132# define FUNCTION__ ""
133# else
134# define FUNCTION__ __FUNCTION__ /* Common extension. */
135# endif
136#endif
137
27d4fb96
PP
138/* XXX A note on the perl source internal type system. The
139 original intent was that I32 be *exactly* 32 bits.
140
141 Currently, we only guarantee that I32 is *at least* 32 bits.
142 Specifically, if int is 64 bits, then so is I32. (This is the case
143 for the Cray.) This has the advantage of meshing nicely with
144 standard library calls (where we pass an I32 and the library is
145 expecting an int), but the disadvantage that an I32 is not 32 bits.
146 Andy Dougherty August 1996
24fef2a7 147
dc45a647
MB
148 There is no guarantee that there is *any* integral type with
149 exactly 32 bits. It is perfectly legal for a system to have
150 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 151
dc45a647
MB
152 Similarly, there is no guarantee that I16 and U16 have exactly 16
153 bits.
693762b4 154
8e84507e
NIS
155 For dealing with issues that may arise from various 32/64-bit
156 systems, we will ask Configure to check out
8175356b 157
3609ea0d
JH
158 SHORTSIZE == sizeof(short)
159 INTSIZE == sizeof(int)
160 LONGSIZE == sizeof(long)
dc45a647 161 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 162 PTRSIZE == sizeof(void *)
dc45a647
MB
163 DOUBLESIZE == sizeof(double)
164 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 165
27d4fb96
PP
166*/
167
69512466
JH
168#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
169# include <inttypes.h>
dd0eed91
JH
170# ifdef INT32_MIN_BROKEN
171# undef INT32_MIN
172# define INT32_MIN (-2147483647-1)
173# endif
174# ifdef INT64_MIN_BROKEN
175# undef INT64_MIN
176# define INT64_MIN (-9223372036854775807LL-1)
177# endif
69512466
JH
178#endif
179
8175356b
JH
180typedef I8TYPE I8;
181typedef U8TYPE U8;
182typedef I16TYPE I16;
183typedef U16TYPE U16;
184typedef I32TYPE I32;
185typedef U32TYPE U32;
6b8eaf93
JH
186#ifdef PERL_CORE
187# ifdef HAS_QUAD
8175356b
JH
188typedef I64TYPE I64;
189typedef U64TYPE U64;
6b8eaf93
JH
190# endif
191#endif /* PERL_CORE */
8175356b 192
69512466
JH
193#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
194# ifndef UINT64_C /* usually from <inttypes.h> */
195# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
196# define INT64_C(c) CAT2(c,LL)
197# define UINT64_C(c) CAT2(c,ULL)
198# else
199# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
200# define INT64_C(c) CAT2(c,L)
201# define UINT64_C(c) CAT2(c,UL)
202# else
fa58a56f
S
203# if defined(_WIN64) && defined(_MSC_VER)
204# define INT64_C(c) CAT2(c,I64)
205# define UINT64_C(c) CAT2(c,UI64)
206# else
207# define INT64_C(c) ((I64TYPE)(c))
208# define UINT64_C(c) ((U64TYPE)(c))
209# endif
69512466
JH
210# endif
211# endif
e8c95190
JH
212# endif
213#endif
dc750f2e 214
d8668976 215#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 216
5ff3f7a4
GS
217/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
218 Please search CHAR_MAX in perl.h for further details. */
219#define U8_MAX UINT8_MAX
220#define U8_MIN UINT8_MIN
221
5ff3f7a4
GS
222#define I16_MAX INT16_MAX
223#define I16_MIN INT16_MIN
224#define U16_MAX UINT16_MAX
225#define U16_MIN UINT16_MIN
226
5ff3f7a4
GS
227#define I32_MAX INT32_MAX
228#define I32_MIN INT32_MIN
0e983133
GS
229#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
230# define U32_MAX UINT32_MAX
231#else
232# define U32_MAX 4294967295U
233#endif
5ff3f7a4
GS
234#define U32_MIN UINT32_MIN
235
236#else
237
5c9fa16e
KA
238/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
239 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
240#define U8_MAX PERL_UCHAR_MAX
241#define U8_MIN PERL_UCHAR_MIN
79072805 242
27d4fb96
PP
243#define I16_MAX PERL_SHORT_MAX
244#define I16_MIN PERL_SHORT_MIN
245#define U16_MAX PERL_USHORT_MAX
246#define U16_MIN PERL_USHORT_MIN
79072805 247
c4f23d77 248#if LONGSIZE > 4
27d4fb96
PP
249# define I32_MAX PERL_INT_MAX
250# define I32_MIN PERL_INT_MIN
251# define U32_MAX PERL_UINT_MAX
252# define U32_MIN PERL_UINT_MIN
79072805 253#else
27d4fb96
PP
254# define I32_MAX PERL_LONG_MAX
255# define I32_MIN PERL_LONG_MIN
256# define U32_MAX PERL_ULONG_MAX
257# define U32_MIN PERL_ULONG_MIN
79072805
LW
258#endif
259
5ff3f7a4
GS
260#endif
261
58a9a5d5 262/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
263#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
264#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
265#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
266
ff68c719 267#define Ctl(ch) ((ch) & 037)
8d063cd8 268
3fe05580
MHM
269/*
270=head1 SV-Body Allocation
271
272=for apidoc Ama|SV*|newSVpvs|const char* s
273Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
274
84bafc02
NC
275=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
276Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
277pair.
278
3fe05580
MHM
279=for apidoc Ama|SV*|newSVpvs_share|const char* s
280Like C<newSVpvn_share>, but takes a literal string instead of a string/length
281pair and omits the hash parameter.
282
9dcc53ea
Z
283=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
284Like C<sv_catpvn_flags>, but takes a literal string instead of a
285string/length pair.
286
287=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
288Like C<sv_catpvn_nomg>, but takes a literal string instead of a
289string/length pair.
290
7b4e7469 291=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
292Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
293
9dcc53ea
Z
294=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
295Like C<sv_catpvn_mg>, but takes a literal string instead of a
296string/length pair.
297
7b4e7469 298=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
299Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
300
9dcc53ea
Z
301=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
302Like C<sv_setpvn_mg>, but takes a literal string instead of a
303string/length pair.
304
305=for apidoc Am|SV *|sv_setref_pvs|const char* s
306Like C<sv_setref_pvn>, but takes a literal string instead of a
307string/length pair.
308
3fe05580
MHM
309=head1 Memory Management
310
311=for apidoc Ama|char*|savepvs|const char* s
312Like C<savepvn>, but takes a literal string instead of a string/length pair.
313
9dcc53ea
Z
314=for apidoc Ama|char*|savesharedpvs|const char* s
315A version of C<savepvs()> which allocates the duplicate string in memory
316which is shared between threads.
317
3fe05580
MHM
318=head1 GV Functions
319
320=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
321Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
322
323=head1 Hash Manipulation Functions
324
325=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
326Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
327
328=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
329Like C<hv_store>, but takes a literal string instead of a string/length pair
330and omits the hash parameter.
331
510966aa
Z
332=head1 Lexer interface
333
334=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
335
336Like L</lex_stuff_pvn>, but takes a literal string instead of a
337string/length pair.
338
3fe05580
MHM
339=cut
340*/
341
65177ae0 342/* concatenating with "" ensures that only literal strings are accepted as argument */
26bfd924 343#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 344
ba3a79e7
GA
345/* note that STR_WITH_LEN() can't be used as argument to macros or functions that
346 * under some configurations might be macros, which means that it requires the full
347 * Perl_xxx(aTHX_ ...) form for any API calls where it's used.
348 */
349
350/* STR_WITH_LEN() shortcuts */
351#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
352#define newSVpvs_flags(str,flags) \
353 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 354#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
355#define sv_catpvs_flags(sv, str, flags) \
356 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
357#define sv_catpvs_nomg(sv, str) \
358 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
359#define sv_catpvs(sv, str) \
360 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
361#define sv_catpvs_mg(sv, str) \
362 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 363#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
364#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
365#define sv_setref_pvs(rv, classname, str) \
366 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 367#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
368#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
369#define gv_stashpvs(str, create) \
370 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
371#define gv_fetchpvs(namebeg, add, sv_type) \
372 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
373#define gv_fetchpvn(namebeg, len, add, sv_type) \
374 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
375#define sv_catxmlpvs(dsv, str, utf8) \
376 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
377#define hv_fetchs(hv,key,lval) \
378 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
379 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
380 : HV_FETCH_JUST_SV, NULL, 0))
381
382#define hv_stores(hv,key,val) \
383 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
384 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 385
510966aa
Z
386#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
387
b96d8cd9
NC
388#define get_cvs(str, flags) \
389 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 390
954c1994 391/*
ccfc67b7
JH
392=head1 Miscellaneous Functions
393
954c1994
GS
394=for apidoc Am|bool|strNE|char* s1|char* s2
395Test two strings to see if they are different. Returns true or
396false.
397
398=for apidoc Am|bool|strEQ|char* s1|char* s2
399Test two strings to see if they are equal. Returns true or false.
400
401=for apidoc Am|bool|strLT|char* s1|char* s2
402Test two strings to see if the first, C<s1>, is less than the second,
403C<s2>. Returns true or false.
404
405=for apidoc Am|bool|strLE|char* s1|char* s2
406Test two strings to see if the first, C<s1>, is less than or equal to the
407second, C<s2>. Returns true or false.
408
409=for apidoc Am|bool|strGT|char* s1|char* s2
410Test two strings to see if the first, C<s1>, is greater than the second,
411C<s2>. Returns true or false.
412
413=for apidoc Am|bool|strGE|char* s1|char* s2
414Test two strings to see if the first, C<s1>, is greater than or equal to
415the second, C<s2>. Returns true or false.
416
417=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
418Test two strings to see if they are different. The C<len> parameter
419indicates the number of bytes to compare. Returns true or false. (A
420wrapper for C<strncmp>).
421
422=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
423Test two strings to see if they are equal. The C<len> parameter indicates
424the number of bytes to compare. Returns true or false. (A wrapper for
425C<strncmp>).
426
427=cut
428*/
429
8d063cd8
LW
430#define strNE(s1,s2) (strcmp(s1,s2))
431#define strEQ(s1,s2) (!strcmp(s1,s2))
432#define strLT(s1,s2) (strcmp(s1,s2) < 0)
433#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
434#define strGT(s1,s2) (strcmp(s1,s2) > 0)
435#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
436#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
437#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 438
36477c24
PP
439#ifdef HAS_MEMCMP
440# define memNE(s1,s2,l) (memcmp(s1,s2,l))
441# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
442#else
443# define memNE(s1,s2,l) (bcmp(s1,s2,l))
444# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
445#endif
446
568a785a 447#define memEQs(s1, l, s2) \
8e6497b6 448 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
449#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
450
bbce6d69
PP
451/*
452 * Character classes.
453 *
454 * Unfortunately, the introduction of locales means that we
455 * can't trust isupper(), etc. to tell the truth. And when
456 * it comes to /\w+/ with tainting enabled, we *must* be able
457 * to trust our character classes.
458 *
459 * Therefore, the default tests in the text of Perl will be
460 * independent of locale. Any code that wants to depend on
461 * the current locale will use the tests that begin with "lc".
462 */
463
2304df62
AD
464#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
465# ifndef CTYPE256
466# define CTYPE256
467# endif
468#endif
469
954c1994 470/*
ccfc67b7
JH
471
472=head1 Character classes
8a58bdcf
KW
473There are three variants for all the functions in this section. The base ones
474operate using the character set of the platform Perl is running on. The ones
475with an C<_A> suffix operate on the ASCII character set, and the ones with an
476C<_L1> suffix operate on the full Latin1 character set. All are unaffected by
181ab5fd 477locale and by C<use bytes>.
8a58bdcf
KW
478
479For ASCII platforms, the base function with no suffix and the one with the
480C<_A> suffix are identical. The function with the C<_L1> suffix imposes the
481Latin-1 character set onto the platform. That is, the code points that are
482ASCII are unaffected, since ASCII is a subset of Latin-1. But the non-ASCII
483code points are treated as if they are Latin-1 characters. For example,
484C<isSPACE_L1()> will return true when called with the code point 0xA0, which is
485the Latin-1 NO-BREAK SPACE.
486
487For EBCDIC platforms, the base function with no suffix and the one with the
488C<_L1> suffix should be identical, since, as of this writing, the EBCDIC code
489pages that Perl knows about all are equivalent to Latin-1. The function that
490ends in an C<_A> suffix will not return true unless the specified character also
491has an ASCII equivalent.
ccfc67b7 492
954c1994 493=for apidoc Am|bool|isALPHA|char ch
2787a470 494Returns a boolean indicating whether the specified character is an
06dc89af 495alphabetic character in the platform's native character set.
8a58bdcf
KW
496See the L<top of this section|/Character classes> for an explanation of variants
497C<isALPHA_A> and C<isALPHA_L1>.
498
499=for apidoc Am|bool|isASCII|char ch
500Returns a boolean indicating whether the specified character is one of the 128
501characters in the ASCII character set. On non-ASCII platforms, it is if this
502character corresponds to an ASCII character. Variants C<isASCII_A()> and
503C<isASCII_L1()> are identical to C<isASCII()>.
954c1994 504
954c1994 505=for apidoc Am|bool|isDIGIT|char ch
2787a470 506Returns a boolean indicating whether the specified character is a
06dc89af 507digit in the platform's native character set.
8a58bdcf 508Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
954c1994 509
0c82b6df 510=for apidoc Am|bool|isLOWER|char ch
2787a470 511Returns a boolean indicating whether the specified character is a
0c82b6df 512lowercase character in the platform's native character set.
8a58bdcf
KW
513See the L<top of this section|/Character classes> for an explanation of variants
514C<isLOWER_A> and C<isLOWER_L1>.
0c82b6df 515
c99e91e9 516=for apidoc Am|bool|isOCTAL|char ch
2787a470 517Returns a boolean indicating whether the specified character is an
06dc89af 518octal digit, [0-7] in the platform's native character set.
8a58bdcf 519Variants C<isOCTAL_A> and C<isOCTAL_L1> are identical to C<isOCTAL>.
c99e91e9 520
0c82b6df 521=for apidoc Am|bool|isSPACE|char ch
2787a470 522Returns a boolean indicating whether the specified character is a
8a58bdcf
KW
523whitespace character in the platform's native character set. This is the same
524as what C<\s> matches in a regular expression.
525See the L<top of this section|/Character classes> for an explanation of variants
526C<isSPACE_A> and C<isSPACE_L1>.
0c82b6df 527
954c1994 528=for apidoc Am|bool|isUPPER|char ch
2787a470 529Returns a boolean indicating whether the specified character is an
06dc89af 530uppercase character in the platform's native character set.
8a58bdcf
KW
531See the L<top of this section|/Character classes> for an explanation of variants
532C<isUPPER_A> and C<isUPPER_L1>.
954c1994 533
8eea39dd
KW
534=for apidoc Am|bool|isWORDCHAR|char ch
535Returns a boolean indicating whether the specified character is a
536character that is any of: alphabetic, numeric, or an underscore. This is the
537same as what C<\w> matches in a regular expression.
538C<isALNUM()> is a synonym provided for backward compatibility. Note that it
539does not have the standard C language meaning of alphanumeric, since it matches
540an underscore and the standard meaning does not.
8a58bdcf
KW
541See the L<top of this section|/Character classes> for an explanation of variants
542C<isWORDCHAR_A> and C<isWORDCHAR_L1>.
543
544=for apidoc Am|bool|isXDIGIT|char ch
545Returns a boolean indicating whether the specified character is a hexadecimal
546digit, [0-9A-Fa-f]. Variants C<isXDIGIT_A()> and C<isXDIGIT_L1()> are
547identical to C<isXDIGIT()>.
8eea39dd 548
e7c1e6c1
KW
549=head1 Character case changing
550
954c1994 551=for apidoc Am|char|toUPPER|char ch
06dc89af
KW
552Converts the specified character to uppercase in the platform's native
553character set, if possible; otherwise returns the input character itself.
954c1994
GS
554
555=for apidoc Am|char|toLOWER|char ch
06dc89af
KW
556Converts the specified character to lowercase in the platform's native
557character set, if possible; otherwise returns the input character itself.
954c1994
GS
558
559=cut
353c9b6f 560
8a58bdcf 561Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
562patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
563
954c1994
GS
564*/
565
7c062697
KW
566/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
567 * is known only in the perl core, and this macro can be called from outside
568 * that */
569#ifdef HAS_QUAD
570# define WIDEST_UTYPE U64TYPE
571#else
572# define WIDEST_UTYPE U32
573#endif
574
3912bc88
KW
575/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
576 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
577 * bits of information are lost even on a 64-bit machine, but to get the
578 * compiler to optimize it out if possible. This is because Configure makes
579 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
580 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
581 */
582#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 583
41f43cc2 584#ifdef EBCDIC
ff7eba81 585# define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128))
41f43cc2
KW
586#else
587# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
588#endif
589
ec79cfbb 590#define isASCII_A(c) isASCII(c)
3509bb53 591#define isASCII_L1(c) isASCII(c)
6515cb76 592
c2da0b36
KW
593/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
594 * and the 8 possible permutations of those bits exactly comprise the 8 octal
595 * digits */
3d33b3e3 596#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36
KW
597
598
ec79cfbb 599/* ASCII range only */
9fb1bf9d 600#ifdef H_PERL /* If have access to perl.h, lookup in its table */
181ab5fd
KW
601/* Bits for PL_charclass[]. These use names used in l1_char_class_tab.h but
602 * their actual definitions are here. If that has a name not used here, it
603 * won't compile. */
96ac0975
NC
604# define _CC_ALNUMC_A (1<<0)
605# define _CC_ALNUMC_L1 (1<<1)
606# define _CC_ALPHA_A (1<<2)
607# define _CC_ALPHA_L1 (1<<3)
608# define _CC_BLANK_A (1<<4)
609# define _CC_BLANK_L1 (1<<5)
610# define _CC_CHARNAME_CONT (1<<6)
611# define _CC_CNTRL_A (1<<7)
612# define _CC_CNTRL_L1 (1<<8)
613# define _CC_DIGIT_A (1<<9)
614# define _CC_GRAPH_A (1<<10)
615# define _CC_GRAPH_L1 (1<<11)
616# define _CC_IDFIRST_A (1<<12)
617# define _CC_IDFIRST_L1 (1<<13)
618# define _CC_LOWER_A (1<<14)
619# define _CC_LOWER_L1 (1<<15)
96ac0975
NC
620# define _CC_PRINT_A (1<<17)
621# define _CC_PRINT_L1 (1<<18)
622# define _CC_PSXSPC_A (1<<19)
623# define _CC_PSXSPC_L1 (1<<20)
624# define _CC_PUNCT_A (1<<21)
625# define _CC_PUNCT_L1 (1<<22)
626# define _CC_SPACE_A (1<<23)
627# define _CC_SPACE_L1 (1<<24)
628# define _CC_UPPER_A (1<<25)
629# define _CC_UPPER_L1 (1<<26)
630# define _CC_WORDCHAR_A (1<<27)
631# define _CC_WORDCHAR_L1 (1<<28)
632# define _CC_XDIGIT_A (1<<29)
633# define _CC_NONLATIN1_FOLD (1<<30)
2703178d 634# define _CC_QUOTEMETA (1U<<31) /* 1U keeps Solaris from griping */
c2da0b36 635/* Unused: (1<<16)
98d360d4 636 * If more are needed, can give up some of the above. The first ones to go
c2da0b36
KW
637 * would be those that require just two tests to verify; either there are two
638 * code points, like BLANK_A, or it occupies a single range like DIGIT_A,
639 * UPPER_A, and LOWER_A. Also consider the ones that can be replaced with two
640 * tests and an additional mask, so
641 *
642 * #define isCNTRL_A cBOOL(FITS_IN_8_BITS(c) \
643 * && (( ! (~0x1F & NATIVE_TO_UNI(c)])) \
644 * || UNLIKELY(NATIVE_TO_UNI(c) == 0x7f)))
645 *
646 * This takes advantage of the contiguous block of these with the first one's
647 * representation having the lower order bits all zero;, except the DELETE must
648 * be tested specially. A similar pattern can be used for for isCNTRL_L1,
649 * isPRINT_A, and isPRINT_L1
96ac0975
NC
650 */
651
652# ifdef DOINIT
653EXTCONST U32 PL_charclass[] = {
654# include "l1_char_class_tab.h"
655};
656
657# else /* ! DOINIT */
658EXTCONST U32 PL_charclass[];
659# endif
660
9fb1bf9d
KW
661# define isALNUMC_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALNUMC_A))
662# define isALPHA_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALPHA_A))
663# define isBLANK_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_BLANK_A))
664# define isCNTRL_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CNTRL_A))
665# define isDIGIT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_DIGIT_A))
666# define isGRAPH_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_GRAPH_A))
667# define isIDFIRST_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_IDFIRST_A))
668# define isLOWER_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_LOWER_A))
9fb1bf9d
KW
669# define isPRINT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PRINT_A))
670# define isPSXSPC_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PSXSPC_A))
671# define isPUNCT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PUNCT_A))
672# define isSPACE_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_SPACE_A))
673# define isUPPER_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_UPPER_A))
674# define isWORDCHAR_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_WORDCHAR_A))
675# define isXDIGIT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_XDIGIT_A))
fb9bfbf7
KW
676 /* Either participates in a fold with a character above 255, or is a
677 * multi-char fold */
e2e75538 678# 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_NONLATIN1_FOLD))
25093c95 679# define _isQUOTEMETA(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_QUOTEMETA))
9fb1bf9d 680#else /* No perl.h. */
9fb1bf9d
KW
681# ifdef EBCDIC
682# define isALNUMC_A(c) (isASCII(c) && isALNUMC(c))
683# define isALPHA_A(c) (isASCII(c) && isALPHA(c))
684# define isBLANK_A(c) (isASCII(c) && isBLANK(c))
685# define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
686# define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
687# define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
688# define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
689# define isLOWER_A(c) (isASCII(c) && isLOWER(c))
690# define isPRINT_A(c) (isASCII(c) && isPRINT(c))
691# define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
692# define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
693# define isSPACE_A(c) (isASCII(c) && isSPACE(c))
694# define isUPPER_A(c) (isASCII(c) && isUPPER(c))
695# define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
696# define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
697# else /* ASCII platform, no perl.h */
698# define isALNUMC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
699# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
700# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
ff7eba81 701# define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
2181da29 702# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
9fb1bf9d
KW
703# define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
704# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
705# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
706# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
707# define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
708# define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) || ((c) >= 58 && (c) <= 64) || ((c) >= 91 && (c) <= 96) || ((c) >= 123 && (c) <= 126))
709# define isSPACE_A(c) ((c) == ' ' || (c) == '\t' || (c) == '\n' || (c) =='\r' || (c) == '\f')
2181da29 710# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
9fb1bf9d 711# define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
2181da29 712# define isXDIGIT_A(c) (isDIGIT_A(c) || ((c) >= 'a' && (c) <= 'f') || ((c) <= 'F' && (c) >= 'A'))
9fb1bf9d
KW
713# endif
714#endif /* ASCII range definitions */
ec79cfbb
KW
715
716/* Latin1 definitions */
8a58bdcf
KW
717#ifdef H_PERL
718# define isALNUMC_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALNUMC_L1))
719# define isALPHA_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALPHA_L1))
720# define isBLANK_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_BLANK_L1))
721/* continuation character for legal NAME in \N{NAME} */
722# define isCHARNAME_CONT(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CHARNAME_CONT))
723# define isCNTRL_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CNTRL_L1))
724# define isGRAPH_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_GRAPH_L1))
725# define isIDFIRST_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_IDFIRST_L1))
726# define isLOWER_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_LOWER_L1))
727# define isPRINT_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PRINT_L1))
728# define isPSXSPC_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PSXSPC_L1))
729# define isPUNCT_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PUNCT_L1))
730# define isSPACE_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_SPACE_L1))
731# define isUPPER_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_UPPER_L1))
732# define isWORDCHAR_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_WORDCHAR_L1))
733#else /* No access to perl.h. Only a few provided here, just in case needed
734 * for backwards compatibility */
735 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
736 * >= AA test to speed up ASCII-only tests at the expense of the others */
737# define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
738 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
739 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
740 || NATIVE_TO_UNI((U8) c) == 0xAA \
741 || NATIVE_TO_UNI((U8) c) == 0xB5 \
742 || NATIVE_TO_UNI((U8) c) == 0xBA)))
743# define isCHARNAME_CONT(c) (isALNUM_L1(c) || (c) == ' ' || (c) == '-' || (c) == '(' || (c) == ')' || (c) == ':' || NATIVE_TO_UNI((U8) c) == 0xA0)
744#endif
745
746/* Macros for backwards compatibility and for completeness when the ASCII and
747 * Latin1 values are identical */
ec79cfbb
KW
748#define isALNUM(c) isWORDCHAR(c)
749#define isALNUMU(c) isWORDCHAR_L1(c)
8a58bdcf
KW
750#define isALPHAU(c) isALPHA_L1(c)
751#define isDIGIT_L1(c) isDIGIT_A(c)
752#define isOCTAL(c) isOCTAL_A(c)
753#define isOCTAL_L1(c) isOCTAL_A(c)
754#define isXDIGIT_L1(c) isXDIGIT_A(c)
755
756/* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
757 * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
758 * therefore can use native functions. For those where C89 doesn't define a
759 * function, use our function, assuming that the EBCDIC code page is isomorphic
760 * with Latin1, which the three currently recognized by Perl are. Some libc's
761 * have an isblank(), but it's not guaranteed. */
9d116dd7 762#ifdef EBCDIC
b8c5462f 763# define isALNUMC(c) isalnum(c)
271b5cb3 764# define isALPHA(c) isalpha(c)
0bc70f7f 765# define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
b8c5462f 766# define isCNTRL(c) iscntrl(c)
9916188f 767# define isDIGIT(c) isdigit(c)
b8c5462f 768# define isGRAPH(c) isgraph(c)
8a58bdcf 769# define isIDFIRST(c) (isALPHA(c) || (c) == '_')
7d79683e 770# define isLOWER(c) islower(c)
9d116dd7 771# define isPRINT(c) isprint(c)
56e4b611 772# define isPSXSPC(c) isspace(c)
b8c5462f 773# define isPUNCT(c) ispunct(c)
174b5e76 774# define isSPACE(c) (isPSXSPC(c) && (c) != '\v')
7d79683e 775# define isUPPER(c) isupper(c)
b8c5462f 776# define isXDIGIT(c) isxdigit(c)
8a58bdcf 777# define isWORDCHAR(c) (isalnum(c) || (c) == '_')
9d116dd7 778# define toLOWER(c) tolower(c)
7d79683e 779# define toUPPER(c) toupper(c)
ec79cfbb 780#else /* Not EBCDIC: ASCII-only matching */
e58b1dde 781# define isALNUMC(c) isALNUMC_A(c) /* Mnemonic: "C's alnum" = alpha + digit */
271b5cb3 782# define isALPHA(c) isALPHA_A(c)
ec79cfbb
KW
783# define isBLANK(c) isBLANK_A(c)
784# define isCNTRL(c) isCNTRL_A(c)
785# define isDIGIT(c) isDIGIT_A(c)
786# define isGRAPH(c) isGRAPH_A(c)
8a58bdcf 787# define isIDFIRST(c) isIDFIRST_A(c)
ec79cfbb
KW
788# define isLOWER(c) isLOWER_A(c)
789# define isPRINT(c) isPRINT_A(c)
790# define isPSXSPC(c) isPSXSPC_A(c)
791# define isPUNCT(c) isPUNCT_A(c)
792# define isSPACE(c) isSPACE_A(c)
793# define isUPPER(c) isUPPER_A(c)
8a58bdcf 794# define isWORDCHAR(c) isWORDCHAR_A(c)
ec79cfbb 795# define isXDIGIT(c) isXDIGIT_A(c)
00f254e2 796
8a58bdcf
KW
797 /* ASCII casing. These could also be written as
798 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
799 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
800 which uses table lookup and mask instead of subtraction. (This would
801 work because the _MOD does not apply in the ASCII range) */
9d116dd7 802# define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
7d79683e 803# define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
9d116dd7 804#endif
bbce6d69 805
1a0901db
KW
806
807/* Use table lookup for speed; return error character for input
808 * out-of-range */
809#define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \
810 ? UNI_TO_NATIVE(PL_latin1_lc[ \
811 NATIVE_TO_UNI( (U8) (c)) ]) \
812 : UNICODE_REPLACEMENT)
813/* Modified uc. Is correct uc except for three non-ascii chars which are
814 * all mapped to one of them, and these need special handling; error
815 * character for input out-of-range */
816#define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \
817 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \
818 NATIVE_TO_UNI( (U8) (c)) ]) \
819 : UNICODE_REPLACEMENT)
820
bbce6d69
PP
821#ifdef USE_NEXT_CTYPE
822
823# define isALNUM_LC(c) \
37bd1396 824 (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
ff68c719
PP
825# define isIDFIRST_LC(c) \
826 (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
827# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 828# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 829# define isBLANK_LC(c) isBLANK((unsigned int)(c))
ff68c719
PP
830# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
831# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
832# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
833# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
37bd1396 834# define isALNUMC_LC(c) NXIsAlNum((unsigned int)(c))
b8c5462f
JH
835# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
836# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
ff68c719 837# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 838# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
ff68c719
PP
839# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
840# define toLOWER_LC(c) NXToLower((unsigned int)(c))
bbce6d69
PP
841
842#else /* !USE_NEXT_CTYPE */
b8c5462f 843
bbce6d69
PP
844# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
845
851131d3
KW
846/* Note that the foo_LC() macros in this case generally are defined only on
847 * code points 0-256, and give undefined, unwarned results if called with
848 * values outside that range */
849
b8c5462f 850# define isALNUM_LC(c) (isalnum((unsigned char)(c)) || (char)(c) == '_')
ff68c719 851# define isIDFIRST_LC(c) (isalpha((unsigned char)(c)) || (char)(c) == '_')
bbce6d69 852# define isALPHA_LC(c) isalpha((unsigned char)(c))
07315176
KW
853# ifdef HAS_ISASCII
854# define isASCII_LC(c) isascii((unsigned char)(c))
855# else
856# define isASCII_LC(c) isASCII((unsigned char)(c))
857# endif
67addccf
KW
858# ifdef HAS_ISBLANK
859# define isBLANK_LC(c) isblank((unsigned char)(c))
860# else
861# define isBLANK_LC(c) isBLANK((unsigned char)(c))
862# endif
bbce6d69
PP
863# define isSPACE_LC(c) isspace((unsigned char)(c))
864# define isDIGIT_LC(c) isdigit((unsigned char)(c))
865# define isUPPER_LC(c) isupper((unsigned char)(c))
866# define isLOWER_LC(c) islower((unsigned char)(c))
b8c5462f
JH
867# define isALNUMC_LC(c) isalnum((unsigned char)(c))
868# define isCNTRL_LC(c) iscntrl((unsigned char)(c))
869# define isGRAPH_LC(c) isgraph((unsigned char)(c))
bbce6d69 870# define isPRINT_LC(c) isprint((unsigned char)(c))
b8c5462f 871# define isPUNCT_LC(c) ispunct((unsigned char)(c))
bbce6d69
PP
872# define toUPPER_LC(c) toupper((unsigned char)(c))
873# define toLOWER_LC(c) tolower((unsigned char)(c))
874
875# else
876
3609ea0d 877# define isALNUM_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
bbce6d69
PP
878# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
879# define isALPHA_LC(c) (isascii(c) && isalpha(c))
07315176 880# define isASCII_LC(c) isascii(c)
67addccf
KW
881# ifdef HAS_ISBLANK
882# define isBLANK_LC(c) (isascii(c) && isblank(c))
883# else
884# define isBLANK_LC(c) isBLANK(c)
885# endif
bbce6d69
PP
886# define isSPACE_LC(c) (isascii(c) && isspace(c))
887# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
888# define isUPPER_LC(c) (isascii(c) && isupper(c))
889# define isLOWER_LC(c) (isascii(c) && islower(c))
b8c5462f
JH
890# define isALNUMC_LC(c) (isascii(c) && isalnum(c))
891# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
892# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
bbce6d69 893# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 894# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
bbce6d69
PP
895# define toUPPER_LC(c) toupper(c)
896# define toLOWER_LC(c) tolower(c)
897
898# endif
a0d0e21e 899#endif /* USE_NEXT_CTYPE */
55204971 900
aaa51d5e 901#define isPSXSPC_LC(c) (isSPACE_LC(c) || (c) == '\v')
aaa51d5e 902
eb0052d1
KW
903/* For use in the macros just below. If the input is Latin1, use the Latin1
904 * (_L1) version of the macro; otherwise use the function. Won't compile if
905 * 'c' isn't unsigned, as won't match function prototype. The macros do bounds
906 * checking, so have duplicate checks here, so could create versions of the
907 * macros that don't, but experiments show that gcc optimizes them out anyway.
908 */
909#define generic_uni(macro, function, c) ((c) < 256 \
910 ? CAT2(macro, _L1)(c) \
911 : function(c))
181ab5fd 912/* Note that all ignore 'use bytes' */
eb0052d1
KW
913
914#define isALNUM_uni(c) generic_uni(isWORDCHAR, is_uni_alnum, c)
915#define isIDFIRST_uni(c) generic_uni(isIDFIRST, is_uni_idfirst, c)
916#define isALPHA_uni(c) generic_uni(isALPHA, is_uni_alpha, c)
917#define isSPACE_uni(c) generic_uni(isSPACE, is_uni_space, c)
918#define isDIGIT_uni(c) generic_uni(isDIGIT, is_uni_digit, c)
919#define isUPPER_uni(c) generic_uni(isUPPER, is_uni_upper, c)
920#define isLOWER_uni(c) generic_uni(isLOWER, is_uni_lower, c)
141d8bad 921#define isASCII_uni(c) isASCII(c)
7b952154 922/* All controls are in Latin1 */
ff7eba81 923#define isCNTRL_uni(c) ((c) < 256 && isCNTRL_L1(c))
eb0052d1
KW
924#define isGRAPH_uni(c) generic_uni(isGRAPH, is_uni_graph, c)
925#define isPRINT_uni(c) generic_uni(isPRINT, is_uni_print, c)
926#define isPUNCT_uni(c) generic_uni(isPUNCT, is_uni_punct, c)
927#define isXDIGIT_uni(c) generic_uni(isXDIGIT, is_uni_xdigit, c)
a2a2844f
JH
928#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
929#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
930#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
b8d68ded 931#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
a0ed51b3 932
eb0052d1
KW
933/* Posix and regular space differ only in U+000B, which is in Latin1 */
934#define isPSXSPC_uni(c) ((c) < 256 ? isPSXSPC_L1(c) : isSPACE_uni(c))
aaa51d5e
JF
935#define isBLANK_uni(c) isBLANK(c) /* could be wrong */
936
9041c2e3
NIS
937#define isALNUM_LC_uvchr(c) (c < 256 ? isALNUM_LC(c) : is_uni_alnum_lc(c))
938#define isIDFIRST_LC_uvchr(c) (c < 256 ? isIDFIRST_LC(c) : is_uni_idfirst_lc(c))
939#define isALPHA_LC_uvchr(c) (c < 256 ? isALPHA_LC(c) : is_uni_alpha_lc(c))
940#define isSPACE_LC_uvchr(c) (c < 256 ? isSPACE_LC(c) : is_uni_space_lc(c))
941#define isDIGIT_LC_uvchr(c) (c < 256 ? isDIGIT_LC(c) : is_uni_digit_lc(c))
942#define isUPPER_LC_uvchr(c) (c < 256 ? isUPPER_LC(c) : is_uni_upper_lc(c))
943#define isLOWER_LC_uvchr(c) (c < 256 ? isLOWER_LC(c) : is_uni_lower_lc(c))
9041c2e3
NIS
944#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : is_uni_cntrl_lc(c))
945#define isGRAPH_LC_uvchr(c) (c < 256 ? isGRAPH_LC(c) : is_uni_graph_lc(c))
946#define isPRINT_LC_uvchr(c) (c < 256 ? isPRINT_LC(c) : is_uni_print_lc(c))
947#define isPUNCT_LC_uvchr(c) (c < 256 ? isPUNCT_LC(c) : is_uni_punct_lc(c))
a0ed51b3 948
aaa51d5e
JF
949#define isPSXSPC_LC_uni(c) (isSPACE_LC_uni(c) ||(c) == '\f')
950#define isBLANK_LC_uni(c) isBLANK(c) /* could be wrong */
951
c8362b00 952/* For use in the macros just below. If the input is ASCII, use the ASCII (_A)
7d7a6efc
KW
953 * version of the macro; if the input is in the upper Latin1 range, use the
954 * Latin1 (_L1) version of the macro, after converting from utf8; otherwise use
955 * the function. This relies on the fact that ASCII characters have the same
956 * representation whether utf8 or not */
957#define generic_utf8(macro, function, p) (isASCII(*(p)) \
57a0fb10 958 ? CAT2(CAT2(macro,_),A)(*(p)) \
7d7a6efc
KW
959 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
960 ? CAT2(macro, _L1) \
961 (TWO_BYTE_UTF8_TO_UNI(*(p), \
962 *((p)+1))) \
963 : function(p))
c8362b00 964
4b88fb76
KW
965/* Note that all assume that the utf8 has been validated, and ignore 'use
966 * bytes' */
181ab5fd 967
c8362b00 968#define isALNUM_utf8(p) generic_utf8(isWORDCHAR, is_utf8_alnum, p)
c11ff943
KW
969/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
970 * IDFIRST is an alnum. See
c8362b00 971 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
9479a769
KW
972 * ever wanted to know about. XXX It is unclear if this should extend to
973 * isIDFIRST_uni() which it hasn't so far. (In the ASCII range, there isn't a
974 * difference.) This used to be not the XID version, but we decided to go with
975 * the more modern Unicode definition */
c8362b00
KW
976#define isIDFIRST_utf8(p) (isASCII(*(p)) \
977 ? isIDFIRST_A(*(p)) \
7d7a6efc
KW
978 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
979 ? isIDFIRST_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \
980 *((p)+1)))\
4244b029 981 : Perl__is_utf8__perl_idstart(aTHX_ p))
c8362b00
KW
982#define isIDCONT_utf8(p) generic_utf8(isWORDCHAR, is_utf8_xidcont, p)
983#define isALPHA_utf8(p) generic_utf8(isALPHA, is_utf8_alpha, p)
984#define isSPACE_utf8(p) generic_utf8(isSPACE, is_utf8_space, p)
985#define isDIGIT_utf8(p) generic_utf8(isDIGIT, is_utf8_digit, p)
986#define isUPPER_utf8(p) generic_utf8(isUPPER, is_utf8_upper, p)
987#define isLOWER_utf8(p) generic_utf8(isLOWER, is_utf8_lower, p)
141d8bad
KW
988/* Because ASCII is invariant under utf8, the non-utf8 macro works */
989#define isASCII_utf8(p) isASCII(p)
c8362b00
KW
990#define isCNTRL_utf8(p) generic_utf8(isCNTRL, is_utf8_cntrl, p)
991#define isGRAPH_utf8(p) generic_utf8(isGRAPH, is_utf8_graph, p)
992#define isPRINT_utf8(p) generic_utf8(isPRINT, is_utf8_print, p)
993#define isPUNCT_utf8(p) generic_utf8(isPUNCT, is_utf8_punct, p)
994#define isXDIGIT_utf8(p) generic_utf8(isXDIGIT, is_utf8_xdigit, p)
a2a2844f
JH
995#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
996#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
997#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
a0ed51b3 998
c8362b00
KW
999/* Posix and regular space differ only in U+000B, which is in ASCII (and hence
1000 * Latin1 */
1001#define isPSXSPC_utf8(p) ((isASCII(*(p))) \
1002 ? isPSXSPC_A(*(p)) \
7d7a6efc
KW
1003 : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \
1004 ? isPSXSPC_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \
1005 *((p)+1)))\
1006 : isSPACE_utf8(p)))
aaa51d5e
JF
1007#define isBLANK_utf8(c) isBLANK(c) /* could be wrong */
1008
4b88fb76
KW
1009#define isALNUM_LC_utf8(p) isALNUM_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1010#define isIDFIRST_LC_utf8(p) isIDFIRST_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1011#define isALPHA_LC_utf8(p) isALPHA_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1012#define isSPACE_LC_utf8(p) isSPACE_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1013#define isDIGIT_LC_utf8(p) isDIGIT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1014#define isUPPER_LC_utf8(p) isUPPER_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1015#define isLOWER_LC_utf8(p) isLOWER_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1016#define isALNUMC_LC_utf8(p) isALNUMC_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1017#define isCNTRL_LC_utf8(p) isCNTRL_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1018#define isGRAPH_LC_utf8(p) isGRAPH_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1019#define isPRINT_LC_utf8(p) isPRINT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1020#define isPUNCT_LC_utf8(p) isPUNCT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
a0ed51b3 1021
aaa51d5e
JF
1022#define isPSXSPC_LC_utf8(c) (isSPACE_LC_utf8(c) ||(c) == '\f')
1023#define isBLANK_LC_utf8(c) isBLANK(c) /* could be wrong */
1024
6ae9f32a
KW
1025/* This conversion works both ways, strangely enough. On EBCDIC platforms,
1026 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII */
1027# define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64)
bbce6d69 1028
dea28490
JJ
1029/* Line numbers are unsigned, 32 bits. */
1030typedef U32 line_t;
dea28490 1031#define NOLINE ((line_t) 4294967295UL)
378cc40b 1032
91152fc1
DG
1033/* Helpful alias for version prescan */
1034#define is_LAX_VERSION(a,b) \
1035 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1036
1037#define is_STRICT_VERSION(a,b) \
1038 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1039
1040#define BADVERSION(a,b,c) \
1041 if (b) { \
1042 *b = c; \
1043 } \
1044 return a;
8c52afec 1045
8e84507e 1046/*
ccfc67b7
JH
1047=head1 Memory Management
1048
a02a5408 1049=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1050The XSUB-writer's interface to the C C<malloc> function.
1051
c5008215
JC
1052In 5.9.3, Newx() and friends replace the older New() API, and drops
1053the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1054themselves. This aid has been superseded by a new build option,
d10b4965 1055PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1056there for use in XS modules supporting older perls.
1057
a02a5408 1058=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1059The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1060cast. See also C<Newx>.
954c1994 1061
a02a5408 1062=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1063The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1064memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1065
954c1994
GS
1066=for apidoc Am|void|Renew|void* ptr|int nitems|type
1067The XSUB-writer's interface to the C C<realloc> function.
1068
1069=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1070The XSUB-writer's interface to the C C<realloc> function, with
1071cast.
1072
49b8b560 1073=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1074The XSUB-writer's interface to the C C<free> function.
1075
1076=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1077The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
1078source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
1079the type. Can do overlapping moves. See also C<Copy>.
1080
e90e2364
NC
1081=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1082Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1083optimise.
1084
954c1994
GS
1085=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1086The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
1087source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
1088the type. May fail on overlapping copies. See also C<Move>.
1089
e90e2364
NC
1090=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1091
1092Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1093optimise.
1094
954c1994
GS
1095=for apidoc Am|void|Zero|void* dest|int nitems|type
1096
1097The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1098destination, C<nitems> is the number of items, and C<type> is the type.
1099
e90e2364
NC
1100=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1101
1102Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1103optimise.
1104
954c1994 1105=for apidoc Am|void|StructCopy|type src|type dest|type
4375e838 1106This is an architecture-independent macro to copy one structure to another.
954c1994 1107
7e337ee0
JH
1108=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1109
1110Fill up memory with a byte pattern (a byte repeated over and over
1111again) that hopefully catches attempts to access uninitialized memory.
1112
1113=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1114
1115PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1116
1c12ffb4 1117=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1118
1119PoisonWith(0xEF) for catching access to freed memory.
1120
9965345d
JH
1121=for apidoc Am|void|Poison|void* dest|int nitems|type
1122
7e337ee0 1123PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1124
1125=cut */
954c1994 1126
561b68a9
SH
1127/* Maintained for backwards-compatibility only. Use newSV() instead. */
1128#ifndef PERL_CORE
ff06c60c 1129#define NEWSV(x,len) newSV(len)
561b68a9 1130#endif
ff06c60c 1131
19a94d75
JH
1132#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1133
2720f793
JH
1134/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1135 * overly eager compilers that will bleat about e.g.
1136 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1137#ifdef PERL_MALLOC_WRAP
fabdb6c0 1138#define MEM_WRAP_CHECK(n,t) MEM_WRAP_CHECK_1(n,t,PL_memory_wrap)
27d5b266 1139#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1140 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1141#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1142
f1f66076 1143#define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (Perl_croak_nocontext("%s",PL_memory_wrap),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1936d2a7 1144
27d5b266
JH
1145#else
1146
410319be
NC
1147#define MEM_WRAP_CHECK(n,t)
1148#define MEM_WRAP_CHECK_1(n,t,a)
1149#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1150#define MEM_WRAP_CHECK_(n,t)
1151
6fff79ce 1152#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1153
1936d2a7 1154#endif
8b44ba4c 1155
fe4f188c 1156#ifdef PERL_MEM_LOG
46c6c7e2 1157/*
9f653bb5 1158 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1159 * go through functions, which are handy for debugging breakpoints, but
1160 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1161 * line number, and C function name if available) passed in. This info can
1162 * then be used for logging the calls, for which one gets a sample
73d1d973 1163 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1164 *
46c6c7e2 1165 * Known problems:
94e892a6 1166 * - not all memory allocs get logged, only those
46c6c7e2 1167 * that go through Newx() and derivatives (while all
94e892a6 1168 * Safefrees do get logged)
46c6c7e2
JH
1169 * - __FILE__ and __LINE__ do not work everywhere
1170 * - __func__ or __FUNCTION__ even less so
1171 * - I think more goes on after the perlio frees but
1172 * the thing is that STDERR gets closed (as do all
1173 * the file descriptors)
1174 * - no deeper calling stack than the caller of the Newx()
1175 * or the kind, but do I look like a C reflection/introspection
1176 * utility to you?
1177 * - the function prototypes for the logging functions
1178 * probably should maybe be somewhere else than handy.h
1179 * - one could consider inlining (macrofying) the logging
1180 * for speed, but I am too lazy
1181 * - one could imagine recording the allocations in a hash,
1182 * (keyed by the allocation address?), and maintain that
1183 * through reallocs and frees, but how to do that without
1184 * any News() happening...?
73d1d973 1185 * - lots of -Ddefines to get useful/controllable output
b953482e 1186 * - lots of ENV reads
46c6c7e2
JH
1187 */
1188
12754f92 1189PERL_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 1190
12754f92 1191PERL_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 1192
12754f92 1193PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1194
0b0ab801 1195# ifdef PERL_CORE
73d1d973 1196# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1197enum mem_log_type {
1198 MLT_ALLOC,
1199 MLT_REALLOC,
d7a2c63c
MHM
1200 MLT_FREE,
1201 MLT_NEW_SV,
1202 MLT_DEL_SV
0b0ab801
MHM
1203};
1204# endif
12754f92 1205# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1206void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1207void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1208# endif
0b0ab801
MHM
1209# endif
1210
fe4f188c
JH
1211#endif
1212
1213#ifdef PERL_MEM_LOG
d1401ee9
MHM
1214#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1215#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1216#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1217#endif
1218
1219#ifndef MEM_LOG_ALLOC
1220#define MEM_LOG_ALLOC(n,t,a) (a)
1221#endif
1222#ifndef MEM_LOG_REALLOC
1223#define MEM_LOG_REALLOC(n,t,v,a) (a)
1224#endif
1225#ifndef MEM_LOG_FREE
1226#define MEM_LOG_FREE(a) (a)
1227#endif
1228
d1401ee9
MHM
1229#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1230#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1231#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1232
1233#ifndef PERL_CORE
a02a5408
JC
1234/* pre 5.9.x compatibility */
1235#define New(x,v,n,t) Newx(v,n,t)
1236#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1237#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1238#endif
a02a5408 1239
ff68c719 1240#define Renew(v,n,t) \
d1401ee9 1241 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1242#define Renewc(v,n,t,c) \
d1401ee9 1243 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1244
1245#ifdef PERL_POISON
1246#define Safefree(d) \
06c0cc96 1247 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1248#else
fe4f188c 1249#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1250#endif
55497cff 1251
8b44ba4c
NC
1252#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1253#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1254#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1255
8b44ba4c
NC
1256#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1257#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1258#ifdef HAS_MEMSET
8b44ba4c 1259#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1260#else
8b44ba4c
NC
1261/* Using bzero(), which returns void. */
1262#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1263#endif
1264
7e337ee0
JH
1265#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1266#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1267#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1268#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1269
2304df62 1270#ifdef USE_STRUCT_COPY
ff68c719 1271#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1272#else
1273#define StructCopy(s,d,t) Copy(s,d,1,t)
1274#endif
2cc61e15 1275
622913ab 1276#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1277#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1278
2cc61e15
DD
1279#ifdef NEED_VA_COPY
1280# ifdef va_copy
1281# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1282# else
a1866d1b
JH
1283# if defined(__va_copy)
1284# define Perl_va_copy(s, d) __va_copy(d, s)
1285# else
1286# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1287# endif
2cc61e15
DD
1288# endif
1289#endif
1290
472d47bc
SB
1291/* convenience debug macros */
1292#ifdef USE_ITHREADS
1293#define pTHX_FORMAT "Perl interpreter: 0x%p"
1294#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1295#define pTHX_VALUE_ (void *)my_perl,
1296#define pTHX_VALUE (void *)my_perl
1297#define pTHX__VALUE_ ,(void *)my_perl,
1298#define pTHX__VALUE ,(void *)my_perl
472d47bc 1299#else
3609ea0d 1300#define pTHX_FORMAT
472d47bc 1301#define pTHX__FORMAT
3609ea0d 1302#define pTHX_VALUE_
472d47bc 1303#define pTHX_VALUE
3609ea0d 1304#define pTHX__VALUE_
472d47bc
SB
1305#define pTHX__VALUE
1306#endif /* USE_ITHREADS */
3609ea0d 1307
2acdbac1
NC
1308/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1309 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1310 CPAN::Unpack show any users outside the core. */
1311#ifdef PERL_CORE
d1d15184 1312# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1313#endif
1314
e9a8c099
MHM
1315/*
1316 * Local variables:
1317 * c-indentation-style: bsd
1318 * c-basic-offset: 4
14d04a33 1319 * indent-tabs-mode: nil
e9a8c099
MHM
1320 * End:
1321 *
14d04a33 1322 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1323 */