This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perlapi: Remove a bunch of functions from API
[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
85e6fe83 11#if !defined(__STDC__)
378cc40b
LW
12#ifdef NULL
13#undef NULL
14#endif
a687059c 15# define NULL 0
85e6fe83
LW
16#endif
17
24792b8d
NC
18#ifndef PERL_CORE
19# define Null(type) ((type)NULL)
954c1994
GS
20
21/*
ccfc67b7 22=head1 Handy Values
954c1994 23
3609ea0d 24=for apidoc AmU||Nullch
24792b8d 25Null character pointer. (No longer available when C<PERL_CORE> is defined.)
2307c6d0 26
954c1994 27=for apidoc AmU||Nullsv
24792b8d 28Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
29
30=cut
31*/
32
24792b8d
NC
33# define Nullch Null(char*)
34# define Nullfp Null(PerlIO*)
35# define Nullsv Null(SV*)
36#endif
8d063cd8 37
641d3f0b
PP
38#ifdef TRUE
39#undef TRUE
40#endif
41#ifdef FALSE
42#undef FALSE
43#endif
44#define TRUE (1)
45#define FALSE (0)
46
cf3f0ffb
DM
47/* The MUTABLE_*() macros cast pointers to the types shown, in such a way
48 * (compiler permitting) that casting away const-ness will give a warning;
49 * e.g.:
50 *
51 * const SV *sv = ...;
52 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
53 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
54 */
55
b1bc3f34 56#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 57# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
58#else
59# define MUTABLE_PTR(p) ((void *) (p))
60#endif
61
a062e10d 62#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 63#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 64#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 65#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 66#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 67#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 68
bd31be4b
NC
69#ifdef I_STDBOOL
70# include <stdbool.h>
71# ifndef HAS_BOOL
72# define HAS_BOOL 1
73# endif
74#endif
75
8e84507e 76/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
77 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
78 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 79 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
80 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
81 g++ can be identified by __GNUG__.
82 Andy Dougherty February 2000
5d94fbed 83*/
3609ea0d 84#ifdef __GNUG__ /* GNU g++ has bool built-in */
5d94fbed 85# ifndef HAS_BOOL
c1d22f6b 86# define HAS_BOOL 1
5d94fbed 87# endif
5d94fbed
AD
88#endif
89
641d3f0b
PP
90/* The NeXT dynamic loader headers will not build with the bool macro
91 So declare them now to clear confusion.
92*/
8f1f23e8 93#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
94# undef FALSE
95# undef TRUE
96 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
97# define ENUM_BOOL 1
98# ifndef HAS_BOOL
99# define HAS_BOOL 1
100# endif /* !HAS_BOOL */
8f1f23e8 101#endif /* NeXT || __NeXT__ */
641d3f0b 102
5d94fbed 103#ifndef HAS_BOOL
cc3315ba 104# if defined(VMS)
5d94fbed
AD
105# define bool int
106# else
107# define bool char
108# endif
c1d22f6b 109# define HAS_BOOL 1
a687059c 110#endif
0d3e774c 111
f2338a2e
DM
112/* a simple (bool) cast may not do the right thing: if bool is defined
113 * as char for example, then the cast from int is implementation-defined
18f5643b 114 * (bool)!!(cbool) in a ternary triggers a bug in xlc on AIX
f2338a2e
DM
115 */
116
18f5643b 117#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 118
46c6c7e2 119/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
120 * XXX Should really be a Configure probe, with HAS__FUNCTION__
121 * and FUNCTION__ as results.
122 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
123#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
124# define FUNCTION__ __func__
125#else
70764c00 126# if (defined(_MSC_VER) && _MSC_VER < 1300) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 127 (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
128# define FUNCTION__ ""
129# else
130# define FUNCTION__ __FUNCTION__ /* Common extension. */
131# endif
132#endif
133
27d4fb96
PP
134/* XXX A note on the perl source internal type system. The
135 original intent was that I32 be *exactly* 32 bits.
136
137 Currently, we only guarantee that I32 is *at least* 32 bits.
138 Specifically, if int is 64 bits, then so is I32. (This is the case
139 for the Cray.) This has the advantage of meshing nicely with
140 standard library calls (where we pass an I32 and the library is
141 expecting an int), but the disadvantage that an I32 is not 32 bits.
142 Andy Dougherty August 1996
24fef2a7 143
dc45a647
MB
144 There is no guarantee that there is *any* integral type with
145 exactly 32 bits. It is perfectly legal for a system to have
146 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 147
dc45a647
MB
148 Similarly, there is no guarantee that I16 and U16 have exactly 16
149 bits.
693762b4 150
8e84507e
NIS
151 For dealing with issues that may arise from various 32/64-bit
152 systems, we will ask Configure to check out
8175356b 153
3609ea0d
JH
154 SHORTSIZE == sizeof(short)
155 INTSIZE == sizeof(int)
156 LONGSIZE == sizeof(long)
dc45a647 157 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 158 PTRSIZE == sizeof(void *)
dc45a647
MB
159 DOUBLESIZE == sizeof(double)
160 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 161
27d4fb96
PP
162*/
163
69512466
JH
164#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
165# include <inttypes.h>
dd0eed91
JH
166# ifdef INT32_MIN_BROKEN
167# undef INT32_MIN
168# define INT32_MIN (-2147483647-1)
169# endif
170# ifdef INT64_MIN_BROKEN
171# undef INT64_MIN
172# define INT64_MIN (-9223372036854775807LL-1)
173# endif
69512466
JH
174#endif
175
8175356b
JH
176typedef I8TYPE I8;
177typedef U8TYPE U8;
178typedef I16TYPE I16;
179typedef U16TYPE U16;
180typedef I32TYPE I32;
181typedef U32TYPE U32;
6b8eaf93
JH
182#ifdef PERL_CORE
183# ifdef HAS_QUAD
8175356b
JH
184typedef I64TYPE I64;
185typedef U64TYPE U64;
6b8eaf93
JH
186# endif
187#endif /* PERL_CORE */
8175356b 188
69512466 189#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
80524f33
JL
190# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
191# define PeRl_INT64_C(c) CAT2(c,LL)
192# define PeRl_UINT64_C(c) CAT2(c,ULL)
193# else
52865553
SH
194# if QUADKIND == QUAD_IS___INT64
195# define PeRl_INT64_C(c) CAT2(c,I64)
196# define PeRl_UINT64_C(c) CAT2(c,UI64)
69512466 197# else
52865553
SH
198# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
199# define PeRl_INT64_C(c) CAT2(c,L)
200# define PeRl_UINT64_C(c) CAT2(c,UL)
69512466 201# else
80524f33
JL
202# define PeRl_INT64_C(c) ((I64TYPE)(c))
203# define PeRl_UINT64_C(c) ((U64TYPE)(c))
69512466
JH
204# endif
205# endif
e8c95190 206# endif
80524f33
JL
207# ifndef UINT64_C
208# define UINT64_C(c) PeRl_UINT64_C(c)
209# endif
210# ifndef INT64_C
211# define INT64_C(c) PeRl_INT64_C(c)
212# endif
e8c95190 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
926bb54c
KW
342/* concatenating with "" ensures that only literal strings are accepted as
343 * argument */
26bfd924 344#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 345
926bb54c
KW
346/* note that STR_WITH_LEN() can't be used as argument to macros or functions
347 * that under some configurations might be macros, which means that it requires
348 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
349 */
350
351/* STR_WITH_LEN() shortcuts */
352#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
353#define newSVpvs_flags(str,flags) \
354 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 355#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
356#define sv_catpvs_flags(sv, str, flags) \
357 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
358#define sv_catpvs_nomg(sv, str) \
359 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
360#define sv_catpvs(sv, str) \
361 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
362#define sv_catpvs_mg(sv, str) \
363 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 364#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
365#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
366#define sv_setref_pvs(rv, classname, str) \
367 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 368#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
369#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
370#define gv_stashpvs(str, create) \
371 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
372#define gv_fetchpvs(namebeg, add, sv_type) \
373 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
374#define gv_fetchpvn(namebeg, len, add, sv_type) \
375 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
376#define sv_catxmlpvs(dsv, str, utf8) \
377 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
378#define hv_fetchs(hv,key,lval) \
379 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
380 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
381 : HV_FETCH_JUST_SV, NULL, 0))
382
383#define hv_stores(hv,key,val) \
384 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
385 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 386
510966aa
Z
387#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
388
b96d8cd9
NC
389#define get_cvs(str, flags) \
390 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 391
954c1994 392/*
ccfc67b7
JH
393=head1 Miscellaneous Functions
394
954c1994
GS
395=for apidoc Am|bool|strNE|char* s1|char* s2
396Test two strings to see if they are different. Returns true or
397false.
398
399=for apidoc Am|bool|strEQ|char* s1|char* s2
400Test two strings to see if they are equal. Returns true or false.
401
402=for apidoc Am|bool|strLT|char* s1|char* s2
403Test two strings to see if the first, C<s1>, is less than the second,
404C<s2>. Returns true or false.
405
406=for apidoc Am|bool|strLE|char* s1|char* s2
407Test two strings to see if the first, C<s1>, is less than or equal to the
408second, C<s2>. Returns true or false.
409
410=for apidoc Am|bool|strGT|char* s1|char* s2
411Test two strings to see if the first, C<s1>, is greater than the second,
412C<s2>. Returns true or false.
413
414=for apidoc Am|bool|strGE|char* s1|char* s2
415Test two strings to see if the first, C<s1>, is greater than or equal to
416the second, C<s2>. Returns true or false.
417
418=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
419Test two strings to see if they are different. The C<len> parameter
420indicates the number of bytes to compare. Returns true or false. (A
421wrapper for C<strncmp>).
422
423=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
424Test two strings to see if they are equal. The C<len> parameter indicates
425the number of bytes to compare. Returns true or false. (A wrapper for
426C<strncmp>).
427
428=cut
429*/
430
8d063cd8
LW
431#define strNE(s1,s2) (strcmp(s1,s2))
432#define strEQ(s1,s2) (!strcmp(s1,s2))
433#define strLT(s1,s2) (strcmp(s1,s2) < 0)
434#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
435#define strGT(s1,s2) (strcmp(s1,s2) > 0)
436#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
437#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
438#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 439
36477c24
PP
440#ifdef HAS_MEMCMP
441# define memNE(s1,s2,l) (memcmp(s1,s2,l))
442# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
443#else
444# define memNE(s1,s2,l) (bcmp(s1,s2,l))
445# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
446#endif
447
568a785a 448#define memEQs(s1, l, s2) \
8e6497b6 449 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
450#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
451
bbce6d69
PP
452/*
453 * Character classes.
454 *
455 * Unfortunately, the introduction of locales means that we
456 * can't trust isupper(), etc. to tell the truth. And when
457 * it comes to /\w+/ with tainting enabled, we *must* be able
458 * to trust our character classes.
459 *
460 * Therefore, the default tests in the text of Perl will be
461 * independent of locale. Any code that wants to depend on
462 * the current locale will use the tests that begin with "lc".
463 */
464
2304df62
AD
465#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
466# ifndef CTYPE256
467# define CTYPE256
468# endif
469#endif
470
954c1994 471/*
ccfc67b7
JH
472
473=head1 Character classes
8a58bdcf
KW
474There are three variants for all the functions in this section. The base ones
475operate using the character set of the platform Perl is running on. The ones
476with an C<_A> suffix operate on the ASCII character set, and the ones with an
477C<_L1> suffix operate on the full Latin1 character set. All are unaffected by
181ab5fd 478locale and by C<use bytes>.
8a58bdcf
KW
479
480For ASCII platforms, the base function with no suffix and the one with the
481C<_A> suffix are identical. The function with the C<_L1> suffix imposes the
482Latin-1 character set onto the platform. That is, the code points that are
483ASCII are unaffected, since ASCII is a subset of Latin-1. But the non-ASCII
484code points are treated as if they are Latin-1 characters. For example,
485C<isSPACE_L1()> will return true when called with the code point 0xA0, which is
486the Latin-1 NO-BREAK SPACE.
487
488For EBCDIC platforms, the base function with no suffix and the one with the
489C<_L1> suffix should be identical, since, as of this writing, the EBCDIC code
490pages that Perl knows about all are equivalent to Latin-1. The function that
491ends in an C<_A> suffix will not return true unless the specified character also
492has an ASCII equivalent.
ccfc67b7 493
954c1994 494=for apidoc Am|bool|isALPHA|char ch
2787a470 495Returns a boolean indicating whether the specified character is an
06dc89af 496alphabetic character in the platform's native character set.
8a58bdcf
KW
497See the L<top of this section|/Character classes> for an explanation of variants
498C<isALPHA_A> and C<isALPHA_L1>.
499
500=for apidoc Am|bool|isASCII|char ch
501Returns a boolean indicating whether the specified character is one of the 128
502characters in the ASCII character set. On non-ASCII platforms, it is if this
503character corresponds to an ASCII character. Variants C<isASCII_A()> and
504C<isASCII_L1()> are identical to C<isASCII()>.
954c1994 505
954c1994 506=for apidoc Am|bool|isDIGIT|char ch
2787a470 507Returns a boolean indicating whether the specified character is a
06dc89af 508digit in the platform's native character set.
8a58bdcf 509Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
954c1994 510
0c82b6df 511=for apidoc Am|bool|isLOWER|char ch
2787a470 512Returns a boolean indicating whether the specified character is a
0c82b6df 513lowercase character in the platform's native character set.
8a58bdcf
KW
514See the L<top of this section|/Character classes> for an explanation of variants
515C<isLOWER_A> and C<isLOWER_L1>.
0c82b6df 516
c99e91e9 517=for apidoc Am|bool|isOCTAL|char ch
2787a470 518Returns a boolean indicating whether the specified character is an
06dc89af 519octal digit, [0-7] in the platform's native character set.
8a58bdcf 520Variants C<isOCTAL_A> and C<isOCTAL_L1> are identical to C<isOCTAL>.
c99e91e9 521
0c82b6df 522=for apidoc Am|bool|isSPACE|char ch
2787a470 523Returns a boolean indicating whether the specified character is a
8a58bdcf
KW
524whitespace character in the platform's native character set. This is the same
525as what C<\s> matches in a regular expression.
526See the L<top of this section|/Character classes> for an explanation of variants
527C<isSPACE_A> and C<isSPACE_L1>.
0c82b6df 528
954c1994 529=for apidoc Am|bool|isUPPER|char ch
2787a470 530Returns a boolean indicating whether the specified character is an
06dc89af 531uppercase character in the platform's native character set.
8a58bdcf
KW
532See the L<top of this section|/Character classes> for an explanation of variants
533C<isUPPER_A> and C<isUPPER_L1>.
954c1994 534
8eea39dd
KW
535=for apidoc Am|bool|isWORDCHAR|char ch
536Returns a boolean indicating whether the specified character is a
537character that is any of: alphabetic, numeric, or an underscore. This is the
538same as what C<\w> matches in a regular expression.
539C<isALNUM()> is a synonym provided for backward compatibility. Note that it
540does not have the standard C language meaning of alphanumeric, since it matches
541an underscore and the standard meaning does not.
8a58bdcf
KW
542See the L<top of this section|/Character classes> for an explanation of variants
543C<isWORDCHAR_A> and C<isWORDCHAR_L1>.
544
545=for apidoc Am|bool|isXDIGIT|char ch
546Returns a boolean indicating whether the specified character is a hexadecimal
547digit, [0-9A-Fa-f]. Variants C<isXDIGIT_A()> and C<isXDIGIT_L1()> are
548identical to C<isXDIGIT()>.
8eea39dd 549
95a59cab
YO
550=for apidoc Am|U8|READ_XDIGIT|char str*
551Returns the value of a hex digit and advances the string pointer.
552Behaviour is only well defined when isXDIGIT(*str) is true.
553
e7c1e6c1
KW
554=head1 Character case changing
555
954c1994 556=for apidoc Am|char|toUPPER|char ch
06dc89af
KW
557Converts the specified character to uppercase in the platform's native
558character set, if possible; otherwise returns the input character itself.
954c1994
GS
559
560=for apidoc Am|char|toLOWER|char ch
06dc89af
KW
561Converts the specified character to lowercase in the platform's native
562character set, if possible; otherwise returns the input character itself.
954c1994
GS
563
564=cut
353c9b6f 565
8a58bdcf 566Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
567patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
568
954c1994
GS
569*/
570
7c062697
KW
571/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
572 * is known only in the perl core, and this macro can be called from outside
573 * that */
574#ifdef HAS_QUAD
575# define WIDEST_UTYPE U64TYPE
576#else
577# define WIDEST_UTYPE U32
578#endif
579
3912bc88
KW
580/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
581 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
582 * bits of information are lost even on a 64-bit machine, but to get the
583 * compiler to optimize it out if possible. This is because Configure makes
584 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
585 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
586 */
587#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 588
41f43cc2 589#ifdef EBCDIC
ff7eba81 590# define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128))
41f43cc2
KW
591#else
592# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
593#endif
594
ec79cfbb 595#define isASCII_A(c) isASCII(c)
3509bb53 596#define isASCII_L1(c) isASCII(c)
6515cb76 597
c2da0b36
KW
598/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
599 * and the 8 possible permutations of those bits exactly comprise the 8 octal
600 * digits */
3d33b3e3 601#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 602
ec79cfbb 603/* ASCII range only */
9fb1bf9d 604#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 605
66c17564
KW
606/* Character class numbers. For internal core Perl use only. These are used
607 * in PL_charclass[] and the ones up through the one that corresponds to
608 * <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h. These use names used
609 * in l1_char_class_tab.h but their actual definitions are here. If that has a
610 * name not used here, it won't compile. */
f4cdb42c
KW
611# define _CC_WORDCHAR 0
612# define _CC_SPACE 1
613# define _CC_DIGIT 2
614# define _CC_ALNUMC 3
615# define _CC_ALPHA 4
616# define _CC_ASCII 5
617# define _CC_CNTRL 6
618# define _CC_GRAPH 7
619# define _CC_LOWER 8
620# define _CC_PRINT 9
621# define _CC_PUNCT 10
622# define _CC_UPPER 11
623# define _CC_XDIGIT 12
624# define _CC_PSXSPC 13
625# define _CC_BLANK 14
a0947d7b
KW
626# define _CC_VERTSPACE 15
627# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
628
629# define _CC_IDFIRST 16
630# define _CC_CHARNAME_CONT 17
631# define _CC_NONLATIN1_FOLD 18
632# define _CC_QUOTEMETA 19
633# define _CC_NON_FINAL_FOLD 20
634# define _CC_IS_IN_SOME_FOLD 21
635/* Unused: 22-31
f4cdb42c
KW
636 * If more bits are needed, one could add a second word for non-64bit
637 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
638 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
639 * is used only for optimization (as of this writing), and differs in the
640 * Latin1 range from the ALPHA bit only in two relatively unimportant
641 * characters: the masculine and feminine ordinal indicators */
96ac0975
NC
642
643# ifdef DOINIT
644EXTCONST U32 PL_charclass[] = {
645# include "l1_char_class_tab.h"
646};
647
648# else /* ! DOINIT */
649EXTCONST U32 PL_charclass[];
650# endif
651
265c1f46 652 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70
KW
653# define _CC_mask(classnum) (1U << (classnum))
654# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
926bb54c 655 && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask(classnum)))
4eeeb416 656
f4cdb42c
KW
657 /* The mask for the _A versions of the macros; it just adds in the bit for
658 * ASCII. */
659# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
660
661 /* The _A version makes sure that both the desired bit and the ASCII bit
662 * are present */
663# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
664 && ((PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_mask_A(classnum)) \
665 == _CC_mask_A(classnum)))
666
667# define isALNUMC_A(c) _generic_isCC_A(c, _CC_ALNUMC)
668# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
669# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
670# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
671# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
672# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
673# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
674# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
675# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
676# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
677# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
678# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
679# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
680# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
681# define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST))
682
fb9bfbf7
KW
683 /* Either participates in a fold with a character above 255, or is a
684 * multi-char fold */
430b7c70
KW
685# 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)))
686
4eeeb416 687# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
688# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
689 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
690# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
691 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
9fb1bf9d 692#else /* No perl.h. */
9fb1bf9d
KW
693# ifdef EBCDIC
694# define isALNUMC_A(c) (isASCII(c) && isALNUMC(c))
695# define isALPHA_A(c) (isASCII(c) && isALPHA(c))
696# define isBLANK_A(c) (isASCII(c) && isBLANK(c))
697# define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
698# define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
699# define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
700# define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
701# define isLOWER_A(c) (isASCII(c) && isLOWER(c))
702# define isPRINT_A(c) (isASCII(c) && isPRINT(c))
703# define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
704# define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
705# define isSPACE_A(c) (isASCII(c) && isSPACE(c))
706# define isUPPER_A(c) (isASCII(c) && isUPPER(c))
707# define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
708# define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
709# else /* ASCII platform, no perl.h */
710# define isALNUMC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
711# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
712# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
926bb54c 713# define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
2181da29 714# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
9fb1bf9d
KW
715# define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
716# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
717# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
718# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
719# define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
926bb54c
KW
720# define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) \
721 || ((c) >= 58 && (c) <= 64) \
722 || ((c) >= 91 && (c) <= 96) \
723 || ((c) >= 123 && (c) <= 126))
724# define isSPACE_A(c) ((c) == ' ' \
725 || (c) == '\t' \
726 || (c) == '\n' \
727 || (c) =='\r' \
728 || (c) == '\f')
2181da29 729# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
9fb1bf9d 730# define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
926bb54c
KW
731# define isXDIGIT_A(c) (isDIGIT_A(c) \
732 || ((c) >= 'a' && (c) <= 'f') \
733 || ((c) <= 'F' && (c) >= 'A'))
9fb1bf9d
KW
734# endif
735#endif /* ASCII range definitions */
ec79cfbb
KW
736
737/* Latin1 definitions */
8a58bdcf 738#ifdef H_PERL
f4cdb42c
KW
739# define isALNUMC_L1(c) _generic_isCC(c, _CC_ALNUMC)
740# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
741# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
e5dcd934 742
8a58bdcf 743/* continuation character for legal NAME in \N{NAME} */
4eeeb416 744# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
e5dcd934 745
f4cdb42c
KW
746# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
747# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
748# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
749# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
750# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
751# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
752# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
753# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
754# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
755# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
8a58bdcf
KW
756#else /* No access to perl.h. Only a few provided here, just in case needed
757 * for backwards compatibility */
758 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
759 * >= AA test to speed up ASCII-only tests at the expense of the others */
760# define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
761 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
926bb54c 762 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
8a58bdcf
KW
763 || NATIVE_TO_UNI((U8) c) == 0xAA \
764 || NATIVE_TO_UNI((U8) c) == 0xB5 \
765 || NATIVE_TO_UNI((U8) c) == 0xBA)))
926bb54c
KW
766# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
767 || (c) == ' ' \
768 || (c) == '-' \
769 || (c) == '(' \
770 || (c) == ')' \
771 || (c) == ':' \
772 || NATIVE_TO_UNI((U8) c) == 0xA0)
8a58bdcf
KW
773#endif
774
775/* Macros for backwards compatibility and for completeness when the ASCII and
776 * Latin1 values are identical */
ec79cfbb
KW
777#define isALNUM(c) isWORDCHAR(c)
778#define isALNUMU(c) isWORDCHAR_L1(c)
8a58bdcf
KW
779#define isALPHAU(c) isALPHA_L1(c)
780#define isDIGIT_L1(c) isDIGIT_A(c)
781#define isOCTAL(c) isOCTAL_A(c)
782#define isOCTAL_L1(c) isOCTAL_A(c)
783#define isXDIGIT_L1(c) isXDIGIT_A(c)
784
785/* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
786 * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
787 * therefore can use native functions. For those where C89 doesn't define a
788 * function, use our function, assuming that the EBCDIC code page is isomorphic
789 * with Latin1, which the three currently recognized by Perl are. Some libc's
790 * have an isblank(), but it's not guaranteed. */
9d116dd7 791#ifdef EBCDIC
b8c5462f 792# define isALNUMC(c) isalnum(c)
271b5cb3 793# define isALPHA(c) isalpha(c)
0bc70f7f 794# define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
b8c5462f 795# define isCNTRL(c) iscntrl(c)
9916188f 796# define isDIGIT(c) isdigit(c)
b8c5462f 797# define isGRAPH(c) isgraph(c)
8a58bdcf 798# define isIDFIRST(c) (isALPHA(c) || (c) == '_')
7d79683e 799# define isLOWER(c) islower(c)
9d116dd7 800# define isPRINT(c) isprint(c)
56e4b611 801# define isPSXSPC(c) isspace(c)
b8c5462f 802# define isPUNCT(c) ispunct(c)
f1f70107
KW
803# define isSPACE(c) (isPSXSPC(c) /* && (c) != '\v' (Experimentally making
804 these macros identical) */)
7d79683e 805# define isUPPER(c) isupper(c)
b8c5462f 806# define isXDIGIT(c) isxdigit(c)
8a58bdcf 807# define isWORDCHAR(c) (isalnum(c) || (c) == '_')
9d116dd7 808# define toLOWER(c) tolower(c)
7d79683e 809# define toUPPER(c) toupper(c)
ec79cfbb 810#else /* Not EBCDIC: ASCII-only matching */
926bb54c 811# define isALNUMC(c) isALNUMC_A(c) /* Mnemonic: "C's alnum" = alpha + digit */
271b5cb3 812# define isALPHA(c) isALPHA_A(c)
ec79cfbb
KW
813# define isBLANK(c) isBLANK_A(c)
814# define isCNTRL(c) isCNTRL_A(c)
815# define isDIGIT(c) isDIGIT_A(c)
816# define isGRAPH(c) isGRAPH_A(c)
8a58bdcf 817# define isIDFIRST(c) isIDFIRST_A(c)
ec79cfbb
KW
818# define isLOWER(c) isLOWER_A(c)
819# define isPRINT(c) isPRINT_A(c)
820# define isPSXSPC(c) isPSXSPC_A(c)
821# define isPUNCT(c) isPUNCT_A(c)
822# define isSPACE(c) isSPACE_A(c)
823# define isUPPER(c) isUPPER_A(c)
8a58bdcf 824# define isWORDCHAR(c) isWORDCHAR_A(c)
ec79cfbb 825# define isXDIGIT(c) isXDIGIT_A(c)
00f254e2 826
8a58bdcf
KW
827 /* ASCII casing. These could also be written as
828 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
829 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
830 which uses table lookup and mask instead of subtraction. (This would
831 work because the _MOD does not apply in the ASCII range) */
9d116dd7 832# define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
7d79683e 833# define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
9d116dd7 834#endif
bbce6d69 835
1a0901db
KW
836
837/* Use table lookup for speed; return error character for input
838 * out-of-range */
839#define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \
840 ? UNI_TO_NATIVE(PL_latin1_lc[ \
841 NATIVE_TO_UNI( (U8) (c)) ]) \
842 : UNICODE_REPLACEMENT)
843/* Modified uc. Is correct uc except for three non-ascii chars which are
844 * all mapped to one of them, and these need special handling; error
845 * character for input out-of-range */
846#define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \
847 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \
848 NATIVE_TO_UNI( (U8) (c)) ]) \
849 : UNICODE_REPLACEMENT)
850
bbce6d69
PP
851#ifdef USE_NEXT_CTYPE
852
509fb054
KW
853# define isALNUMC_LC(c) NXIsAlNum((unsigned int)(c))
854# define isALNUM_LC(c) isWORDCHAR_LC(c)
ff68c719 855# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 856# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 857# define isBLANK_LC(c) isBLANK((unsigned int)(c))
b8c5462f 858# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
509fb054 859# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
b8c5462f 860# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
509fb054
KW
861# define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
862# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
ff68c719 863# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 864# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
509fb054
KW
865# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
866# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
867# define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
e712593e 868# define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
ff68c719 869# define toLOWER_LC(c) NXToLower((unsigned int)(c))
509fb054 870# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
bbce6d69
PP
871
872#else /* !USE_NEXT_CTYPE */
b8c5462f 873
bbce6d69
PP
874# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
875
76fde96a 876/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
851131d3 877
509fb054 878# define isALNUMC_LC(c) (FITS_IN_8_BITS(c) && isalnum((unsigned char)(c)))
ca5b5914 879# define isALNUM_LC(c) isWORDCHAR_LC(c)
76fde96a 880# define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
07315176 881# ifdef HAS_ISASCII
76fde96a 882# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
07315176 883# else
76fde96a 884# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
07315176 885# endif
67addccf 886# ifdef HAS_ISBLANK
76fde96a 887# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
67addccf 888# else
76fde96a 889# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
67addccf 890# endif
76fde96a 891# define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
509fb054 892# define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
76fde96a 893# define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
509fb054
KW
894# define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
895 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
896# define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
76fde96a
KW
897# define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
898# define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
509fb054
KW
899# define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
900# define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
901# define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
902 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
e712593e 903# define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
76fde96a 904# define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? tolower((unsigned char)(c)) : (c))
509fb054 905# define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? toupper((unsigned char)(c)) : (c))
bbce6d69
PP
906
907# else
908
509fb054 909# define isALNUMC_LC(c) (isascii(c) && isalnum(c))
ca5b5914 910# define isALNUM_LC(c) isWORDCHAR_LC(c)
bbce6d69 911# define isALPHA_LC(c) (isascii(c) && isalpha(c))
07315176 912# define isASCII_LC(c) isascii(c)
67addccf
KW
913# ifdef HAS_ISBLANK
914# define isBLANK_LC(c) (isascii(c) && isblank(c))
915# else
3e66ab74 916# define isBLANK_LC(c) isBLANK_A(c)
67addccf 917# endif
b8c5462f 918# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 919# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 920# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
921# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
922# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 923# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 924# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
925# define isSPACE_LC(c) (isascii(c) && isspace(c))
926# define isUPPER_LC(c) (isascii(c) && isupper(c))
927# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 928# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
76fde96a 929# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 930# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
bbce6d69
PP
931
932# endif
a0d0e21e 933#endif /* USE_NEXT_CTYPE */
55204971 934
13380643 935#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 936
66c17564
KW
937/* For internal core Perl use only. If the input is Latin1, use the Latin1
938 * macro; otherwise use the function. Won't compile if 'c' isn't unsigned, as
939 * won't match function prototype. The macros do bounds checking, so have
940 * duplicate checks here, so could create versions of the macros that don't,
941 * but experiments show that gcc optimizes them out anyway. */
942
943/* Note that all ignore 'use bytes' */
944#define _generic_uni(classnum, function, c) ((c) < 256 \
aea0c08c
KW
945 ? _generic_isCC(c, classnum) \
946 : function(c))
509fb054 947#define isALNUMC_uni(c) _generic_uni(_CC_ALNUMC, is_uni_alnumc, c)
926bb54c 948#define isALNUM_uni(c) isWORDCHAR_uni(c)
926bb54c 949#define isALPHA_uni(c) _generic_uni(_CC_ALPHA, is_uni_alpha, c)
509fb054
KW
950#define isASCII_uni(c) isASCII(c)
951#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
952#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
926bb54c 953#define isDIGIT_uni(c) _generic_uni(_CC_DIGIT, is_uni_digit, c)
509fb054
KW
954#define isGRAPH_uni(c) _generic_uni(_CC_GRAPH, is_uni_graph, c)
955#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
926bb54c 956#define isLOWER_uni(c) _generic_uni(_CC_LOWER, is_uni_lower, c)
509fb054 957#define isPRINT_uni(c) _generic_uni(_CC_PRINT, is_uni_print, c)
aea0c08c 958
509fb054
KW
959/* Posix and regular space are identical above Latin1 */
960#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 961
509fb054
KW
962#define isPUNCT_uni(c) _generic_uni(_CC_PUNCT, is_uni_punct, c)
963#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
964#define isUPPER_uni(c) _generic_uni(_CC_UPPER, is_uni_upper, c)
965#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
966#define isWORDCHAR_uni(c) _generic_uni(_CC_WORDCHAR, is_uni_alnum, c)
967#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 968
b8d68ded 969#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
970#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
971#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
972#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 973
69788afd
KW
974#define _gnrc_is_LC_uvchr(latin1, above_latin1, c) \
975 (c < 256 ? latin1(c) : above_latin1(NATIVE_TO_UNI(c)))
509fb054 976#define isALNUMC_LC_uvchr(c) _gnrc_is_LC_uvchr(isALNUMC_LC, is_uni_alnumc_lc, c)
ca5b5914 977#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
69788afd 978#define isALPHA_LC_uvchr(c) _gnrc_is_LC_uvchr(isALPHA_LC, is_uni_alpha_lc, c)
509fb054
KW
979#define isASCII_LC_uvchr(c) isASCII_LC(c)
980#define isBLANK_LC_uvchr(c) _gnrc_is_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 981#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
509fb054 982#define isDIGIT_LC_uvchr(c) _gnrc_is_LC_uvchr(isDIGIT_LC, is_uni_digit_lc, c)
69788afd 983#define isGRAPH_LC_uvchr(c) _gnrc_is_LC_uvchr(isGRAPH_LC, is_uni_graph_lc, c)
509fb054
KW
984#define isIDFIRST_LC_uvchr(c) _gnrc_is_LC_uvchr(isIDFIRST_LC, \
985 is_uni_idfirst_lc, c)
986#define isLOWER_LC_uvchr(c) _gnrc_is_LC_uvchr(isLOWER_LC, is_uni_lower_lc, c)
69788afd 987#define isPRINT_LC_uvchr(c) _gnrc_is_LC_uvchr(isPRINT_LC, is_uni_print_lc, c)
509fb054
KW
988#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
989 space under locale */
69788afd 990#define isPUNCT_LC_uvchr(c) _gnrc_is_LC_uvchr(isPUNCT_LC, is_uni_punct_lc, c)
509fb054
KW
991#define isSPACE_LC_uvchr(c) _gnrc_is_LC_uvchr(isSPACE_LC, \
992 is_XPERLSPACE_cp_high, c)
993#define isUPPER_LC_uvchr(c) _gnrc_is_LC_uvchr(isUPPER_LC, is_uni_upper_lc, c)
994#define isWORDCHAR_LC_uvchr(c) _gnrc_is_LC_uvchr(isWORDCHAR_LC, \
995 is_uni_alnum_lc, c)
e712593e 996#define isXDIGIT_LC_uvchr(c) _gnrc_is_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 997
a0ed51b3 998
36b00f93 999#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1000
66c17564 1001/* For internal core Perl use only. If the input is in the Latin1 range, use
252810d0
KW
1002 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1003 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1004 * fact that ASCII characters have the same representation whether utf8 or not.
1005 * Note that it assumes that the utf8 has been validated, and ignores 'use
1006 * bytes' */
1007#define _generic_utf8_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1008 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1009 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c
KW
1010 ? _generic_isCC( \
1011 TWO_BYTE_UTF8_TO_UNI(*(p), \
1012 *((p)+1 )), \
1013 classnum) \
252810d0
KW
1014 : utf8)
1015/* Like the above, but calls 'function(p)' to get the utf8 value */
1016#define _generic_utf8(classnum, function, p) \
1017 _generic_utf8_utf8(classnum, p, function(p))
c8362b00 1018
509fb054 1019#define isALNUMC_utf8(p) _generic_utf8(_CC_ALNUMC, is_utf8_alnumc, p)
252810d0 1020#define isALNUM_utf8(p) isWORDCHAR_utf8(p) /* back compat */
509fb054
KW
1021#define isALPHA_utf8(p) _generic_utf8(_CC_ALPHA, is_utf8_alpha, p)
1022#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1023 utf8, the non-utf8 macro works
1024 */
1025#define isBLANK_utf8(p) _generic_utf8(_CC_BLANK, is_HORIZWS_high, p)
1026#define isCNTRL_utf8(p) _generic_utf8_utf8(_CC_CNTRL, p, 0)
1027#define isDIGIT_utf8(p) _generic_utf8(_CC_DIGIT, is_utf8_digit, p)
1028#define isGRAPH_utf8(p) _generic_utf8(_CC_GRAPH, is_utf8_graph, p)
1029#define isIDCONT_utf8(p) _generic_utf8(_CC_WORDCHAR, is_utf8_xidcont, p)
e5dcd934 1030
c11ff943
KW
1031/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1032 * IDFIRST is an alnum. See
c8362b00 1033 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1034 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1035 * This used to be not the XID version, but we decided to go with the more
1036 * modern Unicode definition */
252810d0 1037#define isIDFIRST_utf8(p) _generic_utf8(_CC_IDFIRST, \
509fb054 1038 _is_utf8_perl_idstart, p)
e5dcd934 1039
252810d0 1040#define isLOWER_utf8(p) _generic_utf8(_CC_LOWER, is_utf8_lower, p)
509fb054 1041#define isPRINT_utf8(p) _generic_utf8(_CC_PRINT, is_utf8_print, p)
aea0c08c 1042
509fb054
KW
1043/* Posix and regular space are identical above Latin1 */
1044#define isPSXSPC_utf8(p) _generic_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1045
252810d0 1046#define isPUNCT_utf8(p) _generic_utf8(_CC_PUNCT, is_utf8_punct, p)
509fb054
KW
1047#define isSPACE_utf8(p) _generic_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
1048#define isUPPER_utf8(p) _generic_utf8(_CC_UPPER, is_utf8_upper, p)
1049#define isVERTWS_utf8(p) _generic_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
1050#define isWORDCHAR_utf8(p) _generic_utf8(_CC_WORDCHAR, is_utf8_alnum, p)
4ac6419d 1051#define isXDIGIT_utf8(p) _generic_utf8(_CC_XDIGIT, is_XDIGIT_high, p)
a0ed51b3 1052
509fb054
KW
1053#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1054#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1055#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1056
9201a572
KW
1057/* For internal core Perl use only. If the input is in the Latin1 range, use
1058 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1059 * use the value given by the 'utf8' parameter. This relies on the fact that
1060 * ASCII characters have the same representation whether utf8 or not. Note
1061 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
1062#define _generic_LC_utf8_utf8(macro, p, utf8) \
1063 (UTF8_IS_INVARIANT(*(p)) \
1064 ? macro(*(p)) \
1065 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1066 ? macro(TWO_BYTE_UTF8_TO_UNI(*(p), *((p)+1))) \
1067 : utf8)
1068
1069#define _generic_LC_utf8(macro, utf8_func, p) \
1070 _generic_LC_utf8_utf8(macro, p, utf8_func(p))
1071
509fb054 1072#define isALNUMC_LC_utf8(p) _generic_LC_utf8(isALNUMC_LC, is_utf8_alnumc, p)
ca5b5914 1073#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
9201a572 1074#define isALPHA_LC_utf8(p) _generic_LC_utf8(isALPHA_LC, is_utf8_alpha, p)
509fb054 1075#define isASCII_LC_utf8(p) isASCII_LC(*p)
9201a572 1076#define isBLANK_LC_utf8(p) _generic_LC_utf8(isBLANK_LC, is_HORIZWS_high, p)
9201a572 1077#define isCNTRL_LC_utf8(p) _generic_LC_utf8_utf8(isCNTRL_LC, p, 0)
509fb054 1078#define isDIGIT_LC_utf8(p) _generic_LC_utf8(isDIGIT_LC, is_utf8_digit, p)
9201a572 1079#define isGRAPH_LC_utf8(p) _generic_LC_utf8(isGRAPH_LC, is_utf8_graph, p)
509fb054
KW
1080#define isIDFIRST_LC_utf8(p) _generic_LC_utf8(isIDFIRST_LC, \
1081 _is_utf8_perl_idstart, p)
1082#define isLOWER_LC_utf8(p) _generic_LC_utf8(isLOWER_LC, is_utf8_lower, p)
9201a572 1083#define isPRINT_LC_utf8(p) _generic_LC_utf8(isPRINT_LC, is_utf8_print, p)
509fb054
KW
1084#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1085 space under locale */
9201a572 1086#define isPUNCT_LC_utf8(p) _generic_LC_utf8(isPUNCT_LC, is_utf8_punct, p)
509fb054
KW
1087#define isSPACE_LC_utf8(p) _generic_LC_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1088#define isUPPER_LC_utf8(p) _generic_LC_utf8(isUPPER_LC, is_utf8_upper, p)
1089#define isWORDCHAR_LC_utf8(p) _generic_LC_utf8(isWORDCHAR_LC, is_utf8_alnum, p)
e712593e 1090#define isXDIGIT_LC_utf8(p) _generic_LC_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1091
6ae9f32a 1092/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1093 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1094 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1095 * that is EOT on ASCII; ST on EBCDIC */
6ae9f32a 1096# define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64)
bbce6d69 1097
dea28490
JJ
1098/* Line numbers are unsigned, 32 bits. */
1099typedef U32 line_t;
e5dcd934 1100#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1101
91152fc1
DG
1102/* Helpful alias for version prescan */
1103#define is_LAX_VERSION(a,b) \
1104 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1105
1106#define is_STRICT_VERSION(a,b) \
1107 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1108
1109#define BADVERSION(a,b,c) \
1110 if (b) { \
1111 *b = c; \
1112 } \
1113 return a;
8c52afec 1114
95a59cab
YO
1115#define READ_XDIGIT(s) (isALPHA(*(s)) ? ((*(s)++ + 9) & 0xf) : (*(s)++ & 0xf))
1116
8e84507e 1117/*
ccfc67b7
JH
1118=head1 Memory Management
1119
a02a5408 1120=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1121The XSUB-writer's interface to the C C<malloc> function.
1122
c5008215
JC
1123In 5.9.3, Newx() and friends replace the older New() API, and drops
1124the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1125themselves. This aid has been superseded by a new build option,
d10b4965 1126PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1127there for use in XS modules supporting older perls.
1128
a02a5408 1129=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1130The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1131cast. See also C<Newx>.
954c1994 1132
a02a5408 1133=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1134The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1135memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1136
954c1994
GS
1137=for apidoc Am|void|Renew|void* ptr|int nitems|type
1138The XSUB-writer's interface to the C C<realloc> function.
1139
1140=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1141The XSUB-writer's interface to the C C<realloc> function, with
1142cast.
1143
49b8b560 1144=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1145The XSUB-writer's interface to the C C<free> function.
1146
1147=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1148The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1149source, C<dest> is the destination, C<nitems> is the number of items, and
1150C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1151
e90e2364
NC
1152=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1153Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1154optimise.
1155
954c1994
GS
1156=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1157The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1158source, C<dest> is the destination, C<nitems> is the number of items, and
1159C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1160
e90e2364
NC
1161=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1162
1163Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1164optimise.
1165
954c1994
GS
1166=for apidoc Am|void|Zero|void* dest|int nitems|type
1167
1168The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1169destination, C<nitems> is the number of items, and C<type> is the type.
1170
e90e2364
NC
1171=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1172
1173Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1174optimise.
1175
da5d8dbb 1176=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1177This is an architecture-independent macro to copy one structure to another.
954c1994 1178
7e337ee0
JH
1179=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1180
1181Fill up memory with a byte pattern (a byte repeated over and over
1182again) that hopefully catches attempts to access uninitialized memory.
1183
1184=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1185
1186PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1187
1c12ffb4 1188=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1189
1190PoisonWith(0xEF) for catching access to freed memory.
1191
9965345d
JH
1192=for apidoc Am|void|Poison|void* dest|int nitems|type
1193
7e337ee0 1194PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1195
1196=cut */
954c1994 1197
561b68a9
SH
1198/* Maintained for backwards-compatibility only. Use newSV() instead. */
1199#ifndef PERL_CORE
ff06c60c 1200#define NEWSV(x,len) newSV(len)
561b68a9 1201#endif
ff06c60c 1202
19a94d75
JH
1203#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1204
2720f793
JH
1205/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1206 * overly eager compilers that will bleat about e.g.
1207 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1208#ifdef PERL_MALLOC_WRAP
f019c49e
DD
1209#define MEM_WRAP_CHECK(n,t) \
1210 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
27d5b266 1211#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1212 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1213#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1214
f019c49e 1215#define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (croak_memory_wrap(),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266
JH
1216#else
1217
410319be
NC
1218#define MEM_WRAP_CHECK(n,t)
1219#define MEM_WRAP_CHECK_1(n,t,a)
1220#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1221#define MEM_WRAP_CHECK_(n,t)
1222
6fff79ce 1223#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1224
1936d2a7 1225#endif
8b44ba4c 1226
fe4f188c 1227#ifdef PERL_MEM_LOG
46c6c7e2 1228/*
9f653bb5 1229 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1230 * go through functions, which are handy for debugging breakpoints, but
1231 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1232 * line number, and C function name if available) passed in. This info can
1233 * then be used for logging the calls, for which one gets a sample
73d1d973 1234 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1235 *
46c6c7e2 1236 * Known problems:
94e892a6 1237 * - not all memory allocs get logged, only those
46c6c7e2 1238 * that go through Newx() and derivatives (while all
94e892a6 1239 * Safefrees do get logged)
46c6c7e2
JH
1240 * - __FILE__ and __LINE__ do not work everywhere
1241 * - __func__ or __FUNCTION__ even less so
1242 * - I think more goes on after the perlio frees but
1243 * the thing is that STDERR gets closed (as do all
1244 * the file descriptors)
1245 * - no deeper calling stack than the caller of the Newx()
1246 * or the kind, but do I look like a C reflection/introspection
1247 * utility to you?
1248 * - the function prototypes for the logging functions
1249 * probably should maybe be somewhere else than handy.h
1250 * - one could consider inlining (macrofying) the logging
1251 * for speed, but I am too lazy
1252 * - one could imagine recording the allocations in a hash,
1253 * (keyed by the allocation address?), and maintain that
1254 * through reallocs and frees, but how to do that without
1255 * any News() happening...?
73d1d973 1256 * - lots of -Ddefines to get useful/controllable output
b953482e 1257 * - lots of ENV reads
46c6c7e2
JH
1258 */
1259
12754f92 1260PERL_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 1261
12754f92 1262PERL_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 1263
12754f92 1264PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1265
0b0ab801 1266# ifdef PERL_CORE
73d1d973 1267# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1268enum mem_log_type {
1269 MLT_ALLOC,
1270 MLT_REALLOC,
d7a2c63c
MHM
1271 MLT_FREE,
1272 MLT_NEW_SV,
1273 MLT_DEL_SV
0b0ab801
MHM
1274};
1275# endif
12754f92 1276# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1277void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1278void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1279# endif
0b0ab801
MHM
1280# endif
1281
fe4f188c
JH
1282#endif
1283
1284#ifdef PERL_MEM_LOG
d1401ee9
MHM
1285#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1286#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1287#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1288#endif
1289
1290#ifndef MEM_LOG_ALLOC
1291#define MEM_LOG_ALLOC(n,t,a) (a)
1292#endif
1293#ifndef MEM_LOG_REALLOC
1294#define MEM_LOG_REALLOC(n,t,v,a) (a)
1295#endif
1296#ifndef MEM_LOG_FREE
1297#define MEM_LOG_FREE(a) (a)
1298#endif
1299
d1401ee9
MHM
1300#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1301#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1302#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1303
1304#ifndef PERL_CORE
a02a5408
JC
1305/* pre 5.9.x compatibility */
1306#define New(x,v,n,t) Newx(v,n,t)
1307#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1308#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1309#endif
a02a5408 1310
ff68c719 1311#define Renew(v,n,t) \
d1401ee9 1312 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1313#define Renewc(v,n,t,c) \
d1401ee9 1314 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1315
1316#ifdef PERL_POISON
1317#define Safefree(d) \
06c0cc96 1318 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1319#else
fe4f188c 1320#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1321#endif
55497cff 1322
8b44ba4c
NC
1323#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1324#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1325#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1326
8b44ba4c
NC
1327#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1328#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1329#ifdef HAS_MEMSET
8b44ba4c 1330#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1331#else
8b44ba4c
NC
1332/* Using bzero(), which returns void. */
1333#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1334#endif
1335
7e337ee0
JH
1336#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1337#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1338#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1339#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1340
2304df62 1341#ifdef USE_STRUCT_COPY
ff68c719 1342#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1343#else
1344#define StructCopy(s,d,t) Copy(s,d,1,t)
1345#endif
2cc61e15 1346
622913ab 1347#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1348#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1349
2cc61e15
DD
1350#ifdef NEED_VA_COPY
1351# ifdef va_copy
1352# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1353# else
a1866d1b
JH
1354# if defined(__va_copy)
1355# define Perl_va_copy(s, d) __va_copy(d, s)
1356# else
1357# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1358# endif
2cc61e15
DD
1359# endif
1360#endif
1361
472d47bc
SB
1362/* convenience debug macros */
1363#ifdef USE_ITHREADS
1364#define pTHX_FORMAT "Perl interpreter: 0x%p"
1365#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1366#define pTHX_VALUE_ (void *)my_perl,
1367#define pTHX_VALUE (void *)my_perl
1368#define pTHX__VALUE_ ,(void *)my_perl,
1369#define pTHX__VALUE ,(void *)my_perl
472d47bc 1370#else
3609ea0d 1371#define pTHX_FORMAT
472d47bc 1372#define pTHX__FORMAT
3609ea0d 1373#define pTHX_VALUE_
472d47bc 1374#define pTHX_VALUE
3609ea0d 1375#define pTHX__VALUE_
472d47bc
SB
1376#define pTHX__VALUE
1377#endif /* USE_ITHREADS */
3609ea0d 1378
2acdbac1
NC
1379/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1380 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1381 CPAN::Unpack show any users outside the core. */
1382#ifdef PERL_CORE
d1d15184 1383# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1384#endif
1385
e9a8c099
MHM
1386/*
1387 * Local variables:
1388 * c-indentation-style: bsd
1389 * c-basic-offset: 4
14d04a33 1390 * indent-tabs-mode: nil
e9a8c099
MHM
1391 * End:
1392 *
14d04a33 1393 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1394 */