This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Update Test-Harness to CPAN version 3.32
[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
4650c663
KW
11/* IMPORTANT NOTE: Everything whose name begins with an underscore is for
12 * internal core Perl use only. */
13
9d745869
KW
14#ifndef HANDY_H /* Guard against nested #inclusion */
15#define HANDY_H
16
85e6fe83 17#if !defined(__STDC__)
378cc40b
LW
18#ifdef NULL
19#undef NULL
20#endif
a687059c 21# define NULL 0
85e6fe83
LW
22#endif
23
24792b8d
NC
24#ifndef PERL_CORE
25# define Null(type) ((type)NULL)
954c1994
GS
26
27/*
ccfc67b7 28=head1 Handy Values
954c1994 29
3609ea0d 30=for apidoc AmU||Nullch
72d33970
FC
31Null character pointer. (No longer available when C<PERL_CORE> is
32defined.)
2307c6d0 33
954c1994 34=for apidoc AmU||Nullsv
72d33970 35Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
36
37=cut
38*/
39
24792b8d
NC
40# define Nullch Null(char*)
41# define Nullfp Null(PerlIO*)
42# define Nullsv Null(SV*)
43#endif
8d063cd8 44
641d3f0b
PP
45#ifdef TRUE
46#undef TRUE
47#endif
48#ifdef FALSE
49#undef FALSE
50#endif
51#define TRUE (1)
52#define FALSE (0)
53
cf3f0ffb
DM
54/* The MUTABLE_*() macros cast pointers to the types shown, in such a way
55 * (compiler permitting) that casting away const-ness will give a warning;
56 * e.g.:
57 *
58 * const SV *sv = ...;
59 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
60 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
61 */
62
b1bc3f34 63#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 64# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
65#else
66# define MUTABLE_PTR(p) ((void *) (p))
67#endif
68
a062e10d 69#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 70#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 71#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 72#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 73#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 74#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 75
f789f6a4 76#if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR)
bd31be4b
NC
77# include <stdbool.h>
78# ifndef HAS_BOOL
79# define HAS_BOOL 1
80# endif
81#endif
82
8e84507e 83/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
84 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
85 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 86 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
87 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
88 g++ can be identified by __GNUG__.
89 Andy Dougherty February 2000
5d94fbed 90*/
3609ea0d 91#ifdef __GNUG__ /* GNU g++ has bool built-in */
f789f6a4 92# ifndef PERL_BOOL_AS_CHAR
5d94fbed 93# ifndef HAS_BOOL
c1d22f6b 94# define HAS_BOOL 1
5d94fbed 95# endif
f789f6a4 96# endif
5d94fbed
AD
97#endif
98
641d3f0b
PP
99/* The NeXT dynamic loader headers will not build with the bool macro
100 So declare them now to clear confusion.
101*/
8f1f23e8 102#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
103# undef FALSE
104# undef TRUE
105 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
106# define ENUM_BOOL 1
107# ifndef HAS_BOOL
108# define HAS_BOOL 1
109# endif /* !HAS_BOOL */
8f1f23e8 110#endif /* NeXT || __NeXT__ */
641d3f0b 111
5d94fbed 112#ifndef HAS_BOOL
f789f6a4
FC
113# ifdef bool
114# undef bool
115# endif
70d5cb32 116# define bool char
c1d22f6b 117# define HAS_BOOL 1
a687059c 118#endif
0d3e774c 119
236415e2
KW
120/* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
121 * defined as char for example, then the cast from int is
122 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
123 * AIX */
18f5643b 124#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 125
46c6c7e2 126/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
127 * XXX Should really be a Configure probe, with HAS__FUNCTION__
128 * and FUNCTION__ as results.
129 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
130#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
131# define FUNCTION__ __func__
132#else
9ea40801 133# if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 134 (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
135# define FUNCTION__ ""
136# else
137# define FUNCTION__ __FUNCTION__ /* Common extension. */
138# endif
139#endif
140
27d4fb96
PP
141/* XXX A note on the perl source internal type system. The
142 original intent was that I32 be *exactly* 32 bits.
143
144 Currently, we only guarantee that I32 is *at least* 32 bits.
145 Specifically, if int is 64 bits, then so is I32. (This is the case
146 for the Cray.) This has the advantage of meshing nicely with
147 standard library calls (where we pass an I32 and the library is
148 expecting an int), but the disadvantage that an I32 is not 32 bits.
149 Andy Dougherty August 1996
24fef2a7 150
dc45a647
MB
151 There is no guarantee that there is *any* integral type with
152 exactly 32 bits. It is perfectly legal for a system to have
153 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 154
dc45a647
MB
155 Similarly, there is no guarantee that I16 and U16 have exactly 16
156 bits.
693762b4 157
8e84507e
NIS
158 For dealing with issues that may arise from various 32/64-bit
159 systems, we will ask Configure to check out
8175356b 160
3609ea0d
JH
161 SHORTSIZE == sizeof(short)
162 INTSIZE == sizeof(int)
163 LONGSIZE == sizeof(long)
dc45a647 164 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 165 PTRSIZE == sizeof(void *)
dc45a647
MB
166 DOUBLESIZE == sizeof(double)
167 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 168
27d4fb96
PP
169*/
170
69512466
JH
171#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
172# include <inttypes.h>
dd0eed91
JH
173# ifdef INT32_MIN_BROKEN
174# undef INT32_MIN
175# define INT32_MIN (-2147483647-1)
176# endif
177# ifdef INT64_MIN_BROKEN
178# undef INT64_MIN
179# define INT64_MIN (-9223372036854775807LL-1)
180# endif
69512466
JH
181#endif
182
8175356b
JH
183typedef I8TYPE I8;
184typedef U8TYPE U8;
185typedef I16TYPE I16;
186typedef U16TYPE U16;
187typedef I32TYPE I32;
188typedef U32TYPE U32;
6b8eaf93
JH
189#ifdef PERL_CORE
190# ifdef HAS_QUAD
8175356b
JH
191typedef I64TYPE I64;
192typedef U64TYPE U64;
6b8eaf93
JH
193# endif
194#endif /* PERL_CORE */
8175356b 195
9c39b21c
JH
196/* INT64_C/UINT64_C are C99 from <stdint.h> (so they will not be
197 * available in strict C89 mode), but they are nice, so let's define
198 * them if necessary. */
0d3bd0a9 199#if defined(HAS_QUAD)
9c39b21c
JH
200# undef PeRl_INT64_C
201# undef PeRl_UINT64_C
202/* Prefer the native integer types (int and long) over long long
203 * (which is not C89) and Win32-specific __int64. */
204# if QUADKIND == QUAD_IS_INT && INTSIZE == 8
205# define PeRl_INT64_C(c) (c)
206# define PeRl_UINT64_C(c) CAT2(c,U)
207# endif
208# if QUADKIND == QUAD_IS_LONG && LONGSIZE == 8
209# define PeRl_INT64_C(c) CAT2(c,L)
210# define PeRl_UINT64_C(c) CAT2(c,UL)
211# endif
212# if QUADKIND == QUAD_IS_LONG_LONG && defined(HAS_LONG_LONG)
213# define PeRl_INT64_C(c) CAT2(c,LL)
214# define PeRl_UINT64_C(c) CAT2(c,ULL)
215# endif
216# if QUADKIND == QUAD_IS___INT64
217# define PeRl_INT64_C(c) CAT2(c,I64)
218# define PeRl_UINT64_C(c) CAT2(c,UI64)
219# endif
220# ifndef PeRl_INT64_C
221# define PeRl_INT64_C(c) ((I64TYPE)(c)) /* last resort */
222# define PeRl_UINT64_C(c) ((U64TYPE)(c))
223# endif
224/* In OS X the INT64_C/UINT64_C are defined with LL/ULL, which will
225 * not fly with C89-pedantic gcc, so let's undefine them first so that
226 * we can redefine them with our native integer preferring versions. */
8eafc6eb 227# if defined(PERL_DARWIN) && defined(PERL_GCC_PEDANTIC)
9c39b21c
JH
228# undef INT64_C
229# undef UINT64_C
230# endif
231# ifndef INT64_C
232# define INT64_C(c) PeRl_INT64_C(c)
233# endif
234# ifndef UINT64_C
235# define UINT64_C(c) PeRl_UINT64_C(c)
236# endif
e8c95190 237#endif
dc750f2e 238
d8668976 239#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 240
5ff3f7a4
GS
241/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
242 Please search CHAR_MAX in perl.h for further details. */
243#define U8_MAX UINT8_MAX
244#define U8_MIN UINT8_MIN
245
5ff3f7a4
GS
246#define I16_MAX INT16_MAX
247#define I16_MIN INT16_MIN
248#define U16_MAX UINT16_MAX
249#define U16_MIN UINT16_MIN
250
5ff3f7a4
GS
251#define I32_MAX INT32_MAX
252#define I32_MIN INT32_MIN
0e983133
GS
253#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
254# define U32_MAX UINT32_MAX
255#else
256# define U32_MAX 4294967295U
257#endif
5ff3f7a4
GS
258#define U32_MIN UINT32_MIN
259
260#else
261
5c9fa16e
KA
262/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
263 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
264#define U8_MAX PERL_UCHAR_MAX
265#define U8_MIN PERL_UCHAR_MIN
79072805 266
27d4fb96
PP
267#define I16_MAX PERL_SHORT_MAX
268#define I16_MIN PERL_SHORT_MIN
269#define U16_MAX PERL_USHORT_MAX
270#define U16_MIN PERL_USHORT_MIN
79072805 271
c4f23d77 272#if LONGSIZE > 4
27d4fb96
PP
273# define I32_MAX PERL_INT_MAX
274# define I32_MIN PERL_INT_MIN
275# define U32_MAX PERL_UINT_MAX
276# define U32_MIN PERL_UINT_MIN
79072805 277#else
27d4fb96
PP
278# define I32_MAX PERL_LONG_MAX
279# define I32_MIN PERL_LONG_MIN
280# define U32_MAX PERL_ULONG_MAX
281# define U32_MIN PERL_ULONG_MIN
79072805
LW
282#endif
283
5ff3f7a4
GS
284#endif
285
58a9a5d5 286/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
287#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
288#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
289#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
290
ff68c719 291#define Ctl(ch) ((ch) & 037)
8d063cd8 292
4650c663
KW
293/* This is a helper macro to avoid preprocessor issues, replaced by nothing
294 * unless under DEBUGGING, where it expands to an assert of its argument,
295 * followed by a comma (hence the comma operator). If we just used a straight
296 * assert(), we would get a comma with nothing before it when not DEBUGGING */
0f092d08
KW
297#ifdef DEBUGGING
298# define __ASSERT_(statement) assert(statement),
299#else
300# define __ASSERT_(statement)
301#endif
302
3fe05580
MHM
303/*
304=head1 SV-Body Allocation
305
306=for apidoc Ama|SV*|newSVpvs|const char* s
30a15352
KW
307Like C<newSVpvn>, but takes a literal C<NUL>-terminated string instead of a
308string/length pair.
3fe05580 309
84bafc02 310=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
30a15352
KW
311Like C<newSVpvn_flags>, but takes a literal C<NUL>-terminated string instead of
312a string/length pair.
84bafc02 313
3fe05580 314=for apidoc Ama|SV*|newSVpvs_share|const char* s
30a15352
KW
315Like C<newSVpvn_share>, but takes a literal C<NUL>-terminated string instead of
316a string/length pair and omits the hash parameter.
3fe05580 317
9dcc53ea 318=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
30a15352
KW
319Like C<sv_catpvn_flags>, but takes a literal C<NUL>-terminated string instead
320of a string/length pair.
9dcc53ea
Z
321
322=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
323Like C<sv_catpvn_nomg>, but takes a literal string instead of a
324string/length pair.
325
7b4e7469 326=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
327Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
328
9dcc53ea
Z
329=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
330Like C<sv_catpvn_mg>, but takes a literal string instead of a
331string/length pair.
332
7b4e7469 333=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
334Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
335
9dcc53ea
Z
336=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
337Like C<sv_setpvn_mg>, but takes a literal string instead of a
338string/length pair.
339
340=for apidoc Am|SV *|sv_setref_pvs|const char* s
341Like C<sv_setref_pvn>, but takes a literal string instead of a
342string/length pair.
343
3fe05580
MHM
344=head1 Memory Management
345
346=for apidoc Ama|char*|savepvs|const char* s
30a15352
KW
347Like C<savepvn>, but takes a literal C<NUL>-terminated string instead of a
348string/length pair.
3fe05580 349
9dcc53ea
Z
350=for apidoc Ama|char*|savesharedpvs|const char* s
351A version of C<savepvs()> which allocates the duplicate string in memory
352which is shared between threads.
353
3fe05580
MHM
354=head1 GV Functions
355
356=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
357Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
358
359=head1 Hash Manipulation Functions
360
361=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
362Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
363
364=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
365Like C<hv_store>, but takes a literal string instead of a string/length pair
366and omits the hash parameter.
367
510966aa
Z
368=head1 Lexer interface
369
370=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
371
372Like L</lex_stuff_pvn>, but takes a literal string instead of a
373string/length pair.
374
3fe05580
MHM
375=cut
376*/
377
926bb54c
KW
378/* concatenating with "" ensures that only literal strings are accepted as
379 * argument */
26bfd924 380#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 381
926bb54c
KW
382/* note that STR_WITH_LEN() can't be used as argument to macros or functions
383 * that under some configurations might be macros, which means that it requires
384 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
385 */
386
387/* STR_WITH_LEN() shortcuts */
388#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
389#define newSVpvs_flags(str,flags) \
390 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 391#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
392#define sv_catpvs_flags(sv, str, flags) \
393 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
394#define sv_catpvs_nomg(sv, str) \
395 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
396#define sv_catpvs(sv, str) \
397 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
398#define sv_catpvs_mg(sv, str) \
399 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 400#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
401#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
402#define sv_setref_pvs(rv, classname, str) \
403 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 404#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
405#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
406#define gv_stashpvs(str, create) \
407 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
408#define gv_fetchpvs(namebeg, add, sv_type) \
409 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
410#define gv_fetchpvn(namebeg, len, add, sv_type) \
411 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
412#define sv_catxmlpvs(dsv, str, utf8) \
413 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
414#define hv_fetchs(hv,key,lval) \
415 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
416 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
417 : HV_FETCH_JUST_SV, NULL, 0))
418
419#define hv_stores(hv,key,val) \
420 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
421 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 422
510966aa
Z
423#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
424
b96d8cd9
NC
425#define get_cvs(str, flags) \
426 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 427
954c1994 428/*
ccfc67b7
JH
429=head1 Miscellaneous Functions
430
954c1994
GS
431=for apidoc Am|bool|strNE|char* s1|char* s2
432Test two strings to see if they are different. Returns true or
433false.
434
435=for apidoc Am|bool|strEQ|char* s1|char* s2
436Test two strings to see if they are equal. Returns true or false.
437
438=for apidoc Am|bool|strLT|char* s1|char* s2
439Test two strings to see if the first, C<s1>, is less than the second,
440C<s2>. Returns true or false.
441
442=for apidoc Am|bool|strLE|char* s1|char* s2
443Test two strings to see if the first, C<s1>, is less than or equal to the
444second, C<s2>. Returns true or false.
445
446=for apidoc Am|bool|strGT|char* s1|char* s2
447Test two strings to see if the first, C<s1>, is greater than the second,
448C<s2>. Returns true or false.
449
450=for apidoc Am|bool|strGE|char* s1|char* s2
451Test two strings to see if the first, C<s1>, is greater than or equal to
452the second, C<s2>. Returns true or false.
453
454=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
455Test two strings to see if they are different. The C<len> parameter
72d33970 456indicates the number of bytes to compare. Returns true or false. (A
954c1994
GS
457wrapper for C<strncmp>).
458
459=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
460Test two strings to see if they are equal. The C<len> parameter indicates
72d33970 461the number of bytes to compare. Returns true or false. (A wrapper for
954c1994
GS
462C<strncmp>).
463
464=cut
465*/
466
8d063cd8
LW
467#define strNE(s1,s2) (strcmp(s1,s2))
468#define strEQ(s1,s2) (!strcmp(s1,s2))
469#define strLT(s1,s2) (strcmp(s1,s2) < 0)
470#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
471#define strGT(s1,s2) (strcmp(s1,s2) > 0)
472#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
473#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
474#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 475
36477c24
PP
476#ifdef HAS_MEMCMP
477# define memNE(s1,s2,l) (memcmp(s1,s2,l))
478# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
479#else
480# define memNE(s1,s2,l) (bcmp(s1,s2,l))
481# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
482#endif
483
568a785a 484#define memEQs(s1, l, s2) \
8e6497b6 485 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
486#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
487
bbce6d69
PP
488/*
489 * Character classes.
490 *
491 * Unfortunately, the introduction of locales means that we
492 * can't trust isupper(), etc. to tell the truth. And when
493 * it comes to /\w+/ with tainting enabled, we *must* be able
494 * to trust our character classes.
495 *
496 * Therefore, the default tests in the text of Perl will be
497 * independent of locale. Any code that wants to depend on
498 * the current locale will use the tests that begin with "lc".
499 */
500
2304df62
AD
501#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
502# ifndef CTYPE256
503# define CTYPE256
504# endif
505#endif
506
954c1994 507/*
ccfc67b7 508
dcccc8ff 509=head1 Character classification
243effed
KW
510This section is about functions (really macros) that classify characters
511into types, such as punctuation versus alphabetic, etc. Most of these are
512analogous to regular expression character classes. (See
513L<perlrecharclass/POSIX Character Classes>.) There are several variants for
514each class. (Not all macros have all variants; each item below lists the
515ones valid for it.) None are affected by C<use bytes>, and only the ones
516with C<LC> in the name are affected by the current locale.
517
518The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
519C<U8>) as input and returns a boolean as to whether or not the character
6aff1f14
KW
520represented by that octet is (or on non-ASCII platforms, corresponds to) an
521ASCII character in the named class based on platform, Unicode, and Perl rules.
522If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 523
6aff1f14
KW
524Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
525with no suffix C<"_A">.
4b9734bf
KW
526
527Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
528onto the platform. That is, the code points that are ASCII are unaffected,
529since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
530as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
531true when called with the code point 0xDF, which is a word character in both
4650c663 532ASCII and EBCDIC (though it represents different characters in each).
243effed
KW
533
534Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
535point as input. If the code point is larger than 255, Unicode rules are used
536to determine if it is in the character class. For example,
6aff1f14
KW
537C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
538WITH MACRON in Unicode, and is a word character.
243effed
KW
539
540Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
541(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
542classification of just the first (possibly multi-byte) character in the string
543is tested.
243effed 544
31f05a37
KW
545Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but the
546result is based on the current locale, which is what C<LC> in the name stands
547for. If Perl can determine that the current locale is a UTF-8 locale, it uses
548the published Unicode rules; otherwise, it uses the C library function that
549gives the named classification. For example, C<isDIGIT_LC()> when not in a
550UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always
551returned if the input won't fit into an octet.
243effed
KW
552
553Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
554returns the same as C<isFOO_LC> for input code points less than 256, and
555returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
556
557Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
558(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
559classification of just the first (possibly multi-byte) character in the string
560is tested.
ccfc67b7 561
954c1994 562=for apidoc Am|bool|isALPHA|char ch
2787a470 563Returns a boolean indicating whether the specified character is an
6aff1f14 564alphabetic character, analogous to C<m/[[:alpha:]]/>.
dcccc8ff
KW
565See the L<top of this section|/Character classification> for an explanation of
566variants
e5ad6aba 567C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 568C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 569
15861f94
KW
570=for apidoc Am|bool|isALPHANUMERIC|char ch
571Returns a boolean indicating whether the specified character is a either an
6aff1f14 572alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
dcccc8ff
KW
573See the L<top of this section|/Character classification> for an explanation of
574variants
15861f94 575C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 576C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
577C<isALPHANUMERIC_LC_utf8>.
578
8a58bdcf
KW
579=for apidoc Am|bool|isASCII|char ch
580Returns a boolean indicating whether the specified character is one of the 128
243effed 581characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 582On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
583character corresponds to an ASCII character. Variants C<isASCII_A()> and
584C<isASCII_L1()> are identical to C<isASCII()>.
dcccc8ff
KW
585See the L<top of this section|/Character classification> for an explanation of
586variants
243effed
KW
587C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
588C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
589library routine C<isascii()>. In these cases, the variants whose names contain
590C<LC> are the same as the corresponding ones without.
591
d98532ea
KW
592Also note, that because all ASCII characters are UTF-8 invariant (meaning they
593have the exact same representation (always a single byte) whether encoded in
594UTF-8 or not), C<isASCII> will give the correct results when called with any
595byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
596work properly on any string encoded or not in UTF-8.
597
243effed
KW
598=for apidoc Am|bool|isBLANK|char ch
599Returns a boolean indicating whether the specified character is a
6aff1f14 600character considered to be a blank, analogous to C<m/[[:blank:]]/>.
dcccc8ff
KW
601See the L<top of this section|/Character classification> for an explanation of
602variants
e5ad6aba 603C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
604C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
605platforms do not have the C library routine C<isblank()>. In these cases, the
606variants whose names contain C<LC> are the same as the corresponding ones
607without.
608
609=for apidoc Am|bool|isCNTRL|char ch
610Returns a boolean indicating whether the specified character is a
6aff1f14 611control character, analogous to C<m/[[:cntrl:]]/>.
dcccc8ff
KW
612See the L<top of this section|/Character classification> for an explanation of
613variants
e5ad6aba 614C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
615C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
616On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 617
954c1994 618=for apidoc Am|bool|isDIGIT|char ch
2787a470 619Returns a boolean indicating whether the specified character is a
6aff1f14 620digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 621Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
dcccc8ff
KW
622See the L<top of this section|/Character classification> for an explanation of
623variants
e5ad6aba 624C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
625C<isDIGIT_LC_utf8>.
626
627=for apidoc Am|bool|isGRAPH|char ch
628Returns a boolean indicating whether the specified character is a
6aff1f14 629graphic character, analogous to C<m/[[:graph:]]/>.
dcccc8ff
KW
630See the L<top of this section|/Character classification> for an explanation of
631variants
e5ad6aba 632C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 633C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 634
0c82b6df 635=for apidoc Am|bool|isLOWER|char ch
2787a470 636Returns a boolean indicating whether the specified character is a
6aff1f14 637lowercase character, analogous to C<m/[[:lower:]]/>.
dcccc8ff
KW
638See the L<top of this section|/Character classification> for an explanation of
639variants
e5ad6aba 640C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 641C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 642
c99e91e9 643=for apidoc Am|bool|isOCTAL|char ch
2787a470 644Returns a boolean indicating whether the specified character is an
6aff1f14 645octal digit, [0-7].
243effed
KW
646The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
647C<isOCTAL>.
648
649=for apidoc Am|bool|isPUNCT|char ch
650Returns a boolean indicating whether the specified character is a
6aff1f14
KW
651punctuation character, analogous to C<m/[[:punct:]]/>.
652Note that the definition of what is punctuation isn't as
243effed
KW
653straightforward as one might desire. See L<perlrecharclass/POSIX Character
654Classes> for details.
dcccc8ff
KW
655See the L<top of this section|/Character classification> for an explanation of
656variants
e5ad6aba 657C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 658C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 659
0c82b6df 660=for apidoc Am|bool|isSPACE|char ch
2787a470 661Returns a boolean indicating whether the specified character is a
6aff1f14 662whitespace character. This is analogous
398d098a
KW
663to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
664(experimentally), this also matches what C<m/[[:space:]]/> does.
a04e6aad 665("Experimentally" means that this change may be backed out in 5.22 if
398d098a
KW
666field experience indicates that it was unwise.) Prior to 5.18, only the
667locale forms of this macro (the ones with C<LC> in their names) matched
668precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
669in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
670(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
dcccc8ff
KW
671See the L<top of this section|/Character classification> for an explanation of
672variants
e5ad6aba 673C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 674C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 675
398d098a
KW
676=for apidoc Am|bool|isPSXSPC|char ch
677(short for Posix Space)
678Starting in 5.18, this is identical (experimentally) in all its forms to the
679corresponding C<isSPACE()> macros. ("Experimentally" means that this change
a04e6aad 680may be backed out in 5.22 if field experience indicates that it
398d098a
KW
681was unwise.)
682The locale forms of this macro are identical to their corresponding
683C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
684non-locale forms differ from their C<isSPACE()> forms only in that the
685C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
686Otherwise they are identical. Thus this macro is analogous to what
687C<m/[[:space:]]/> matches in a regular expression.
dcccc8ff
KW
688See the L<top of this section|/Character classification> for an explanation of
689variants
e5ad6aba 690C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
691C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
692
954c1994 693=for apidoc Am|bool|isUPPER|char ch
2787a470 694Returns a boolean indicating whether the specified character is an
6aff1f14 695uppercase character, analogous to C<m/[[:upper:]]/>.
dcccc8ff
KW
696See the L<top of this section|/Character classification> for an explanation of
697variants
e5ad6aba 698C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 699C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 700
243effed 701=for apidoc Am|bool|isPRINT|char ch
8eea39dd 702Returns a boolean indicating whether the specified character is a
6aff1f14 703printable character, analogous to C<m/[[:print:]]/>.
dcccc8ff
KW
704See the L<top of this section|/Character classification> for an explanation of
705variants
e5ad6aba 706C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
707C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
708
709=for apidoc Am|bool|isWORDCHAR|char ch
710Returns a boolean indicating whether the specified character is a character
711that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
712in a regular expression. A word character is an alphabetic character, a
713decimal digit, a connecting punctuation character (such as an underscore), or
714a "mark" character that attaches to one of those (like some sort of accent).
715C<isALNUM()> is a synonym provided for backward compatibility, even though a
716word character includes more than the standard C language meaning of
717alphanumeric.
dcccc8ff
KW
718See the L<top of this section|/Character classification> for an explanation of
719variants
243effed
KW
720C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
721C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
722
723=for apidoc Am|bool|isXDIGIT|char ch
724Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
725digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
726and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
dcccc8ff
KW
727See the L<top of this section|/Character classification> for an explanation of
728variants
243effed
KW
729C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
730C<isXDIGIT_LC_utf8>.
731
3c3ecf18
KW
732=for apidoc Am|bool|isIDFIRST|char ch
733Returns a boolean indicating whether the specified character can be the first
734character of an identifier. This is very close to, but not quite the same as
735the official Unicode property C<XID_Start>. The difference is that this
736returns true only if the input character also matches L</isWORDCHAR>.
dcccc8ff
KW
737See the L<top of this section|/Character classification> for an explanation of
738variants
3c3ecf18
KW
739C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
740C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
741
742=for apidoc Am|bool|isIDCONT|char ch
743Returns a boolean indicating whether the specified character can be the
744second or succeeding character of an identifier. This is very close to, but
745not quite the same as the official Unicode property C<XID_Continue>. The
746difference is that this returns true only if the input character also matches
dcccc8ff
KW
747L</isWORDCHAR>. See the L<top of this section|/Character classification> for
748an
3c3ecf18
KW
749explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
750C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
751C<isIDCONT_LC_utf8>.
752
243effed 753=head1 Miscellaneous Functions
8eea39dd 754
95a59cab 755=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 756Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
757Behaviour is only well defined when isXDIGIT(*str) is true.
758
e7c1e6c1
KW
759=head1 Character case changing
760
1f607577
KW
761=for apidoc Am|U8|toUPPER|U8 ch
762Converts the specified character to uppercase. If the input is anything but an
763ASCII lowercase character, that input character itself is returned. Variant
c753c8d3 764C<toUPPER_A> is equivalent.
954c1994 765
1f607577
KW
766=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
767Converts the Unicode code point C<cp> to its uppercase version, and
768stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
769that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
770bytes since the uppercase version may be longer than the original character.
771
772The first code point of the uppercased version is returned
773(but note, as explained just above, that there may be more.)
774
775=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
776Converts the UTF-8 encoded character at C<p> to its uppercase version, and
777stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
778that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
779bytes since the uppercase version may be longer than the original character.
780
781The first code point of the uppercased version is returned
782(but note, as explained just above, that there may be more.)
783
784The input character at C<p> is assumed to be well-formed.
785
25200305
KW
786=for apidoc Am|U8|toFOLD|U8 ch
787Converts the specified character to foldcase. If the input is anything but an
788ASCII uppercase character, that input character itself is returned. Variant
789C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
790Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
791
1f607577
KW
792=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
793Converts the Unicode code point C<cp> to its foldcase version, and
794stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
795that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
796bytes since the foldcase version may be longer than the original character.
797
798The first code point of the foldcased version is returned
799(but note, as explained just above, that there may be more.)
800
801=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
802Converts the UTF-8 encoded character at C<p> to its foldcase version, and
803stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
804that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
805bytes since the foldcase version may be longer than the original character.
806
807The first code point of the foldcased version is returned
808(but note, as explained just above, that there may be more.)
809
810The input character at C<p> is assumed to be well-formed.
811
812=for apidoc Am|U8|toLOWER|U8 ch
813Converts the specified character to lowercase. If the input is anything but an
814ASCII uppercase character, that input character itself is returned. Variant
c753c8d3 815C<toLOWER_A> is equivalent.
954c1994 816
1f607577
KW
817=for apidoc Am|U8|toLOWER_L1|U8 ch
818Converts the specified Latin1 character to lowercase. The results are undefined if
819the input doesn't fit in a byte.
820
821=for apidoc Am|U8|toLOWER_LC|U8 ch
822Converts the specified character to lowercase using the current locale's rules,
823if possible; otherwise returns the input character itself.
824
825=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
826Converts the Unicode code point C<cp> to its lowercase version, and
827stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
828that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
829bytes since the lowercase version may be longer than the original character.
830
831The first code point of the lowercased version is returned
832(but note, as explained just above, that there may be more.)
833
834=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
835Converts the UTF-8 encoded character at C<p> to its lowercase version, and
836stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
837that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
838bytes since the lowercase version may be longer than the original character.
839
840The first code point of the lowercased version is returned
841(but note, as explained just above, that there may be more.)
842
843The input character at C<p> is assumed to be well-formed.
844
845=for apidoc Am|U8|toLOWER_LC|U8 ch
846Converts the specified character to lowercase using the current locale's rules,
847if possible; otherwise returns the input character itself.
848
25200305
KW
849=for apidoc Am|U8|toTITLE|U8 ch
850Converts the specified character to titlecase. If the input is anything but an
851ASCII lowercase character, that input character itself is returned. Variant
852C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
853as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
854concept used in locale handling, so there is no functionality for that.)
855
1f607577
KW
856=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
857Converts the Unicode code point C<cp> to its titlecase version, and
858stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
859that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
860bytes since the titlecase version may be longer than the original character.
861
862The first code point of the titlecased version is returned
863(but note, as explained just above, that there may be more.)
864
865=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
866Converts the UTF-8 encoded character at C<p> to its titlecase version, and
867stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
868that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
869bytes since the titlecase version may be longer than the original character.
870
871The first code point of the titlecased version is returned
872(but note, as explained just above, that there may be more.)
873
874The input character at C<p> is assumed to be well-formed.
875
954c1994 876=cut
353c9b6f 877
1f607577 878XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
25200305 879really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
243effed 880
8a58bdcf 881Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
882patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
883
954c1994
GS
884*/
885
7c062697
KW
886/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
887 * is known only in the perl core, and this macro can be called from outside
888 * that */
889#ifdef HAS_QUAD
890# define WIDEST_UTYPE U64TYPE
891#else
892# define WIDEST_UTYPE U32
893#endif
894
3912bc88
KW
895/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
896 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
897 * bits of information are lost even on a 64-bit machine, but to get the
898 * compiler to optimize it out if possible. This is because Configure makes
899 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
900 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
901 */
902#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 903
41f43cc2 904#ifdef EBCDIC
b6340bd0
KW
905# ifndef _ALL_SOURCE
906 /* This returns the wrong results on at least z/OS unless this is
907 * defined. */
908# error _ALL_SOURCE should probably be defined
909# endif
910
911 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
912 * likely not defined, and so we use the native function */
3cedda69 913# define isASCII(c) cBOOL(isascii(c))
41f43cc2
KW
914#else
915# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
916#endif
917
ec79cfbb 918#define isASCII_A(c) isASCII(c)
3509bb53 919#define isASCII_L1(c) isASCII(c)
6515cb76 920
c2da0b36
KW
921/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
922 * and the 8 possible permutations of those bits exactly comprise the 8 octal
923 * digits */
3d33b3e3 924#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 925
ec79cfbb 926/* ASCII range only */
9fb1bf9d 927#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 928
a500dc72
KW
929/* Character class numbers. For internal core Perl use only. The ones less
930 * than 32 are used in PL_charclass[] and the ones up through the one that
931 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
932 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
933 * their actual definitions are here. If that file has a name not used here,
934 * it won't compile.
1709d539
KW
935 *
936 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
937 * frequency of their use. This gives a slight edge to exiting a loop earlier
938 * (in reginclass() in regexec.c) */
1709d539
KW
939# define _CC_WORDCHAR 0 /* \w and [:word:] */
940# define _CC_DIGIT 1 /* \d and [:digit:] */
941# define _CC_ALPHA 2 /* [:alpha:] */
942# define _CC_LOWER 3 /* [:lower:] */
943# define _CC_UPPER 4 /* [:upper:] */
944# define _CC_PUNCT 5 /* [:punct:] */
945# define _CC_PRINT 6 /* [:print:] */
15861f94 946# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 947# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 948# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 949
b0d691b2 950#define _FIRST_NON_SWASH_CC 10
1709d539
KW
951/* The character classes above are implemented with swashes. The second group
952 * (just below) contains the ones implemented without. These are also sorted
953 * in rough order of the frequency of their use, except that \v should be last,
954 * as it isn't a real Posix character class, and some (small) inefficiencies in
955 * regular expression handling would be introduced by putting it in the middle
956 * of those that are. Also, cntrl and ascii come after the others as it may be
957 * useful to group these which have no members that match above Latin1, (or
958 * above ASCII in the latter case) */
959
b0d691b2
KW
960# define _CC_SPACE 10 /* \s */
961# define _CC_BLANK 11 /* [:blank:] */
962# define _CC_XDIGIT 12 /* [:xdigit:] */
963# define _CC_PSXSPC 13 /* [:space:] */
964# define _CC_CNTRL 14 /* [:cntrl:] */
965# define _CC_ASCII 15 /* [:ascii:] */
966# define _CC_VERTSPACE 16 /* \v */
1709d539 967
a0947d7b
KW
968# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
969
1709d539 970/* The members of the third group below do not need to be coordinated with data
3ffc8c70 971 * structures in regcomp.[ch] and regexec.c. */
f12c0118
KW
972# define _CC_IDFIRST 17
973# define _CC_CHARNAME_CONT 18
974# define _CC_NONLATIN1_FOLD 19
975# define _CC_NONLATIN1_SIMPLE_FOLD 20
976# define _CC_QUOTEMETA 21
977# define _CC_NON_FINAL_FOLD 22
978# define _CC_IS_IN_SOME_FOLD 23
4d68ffa0 979# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
f12c0118 980/* Unused: 24-30
f4cdb42c
KW
981 * If more bits are needed, one could add a second word for non-64bit
982 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
983 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
984 * is used only for optimization (as of this writing), and differs in the
985 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
986 * characters: the masculine and feminine ordinal indicators, so removing it
987 * would just cause /i regexes which match them to run less efficiently */
96ac0975 988
3a371f2f
KW
989#if defined(PERL_CORE) || defined(PERL_EXT)
990/* An enum version of the character class numbers, to help compilers
991 * optimize */
992typedef enum {
3a371f2f 993 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
994 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
995 _CC_ENUM_ASCII = _CC_ASCII,
996 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 997 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 998 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
999 _CC_ENUM_DIGIT = _CC_DIGIT,
1000 _CC_ENUM_GRAPH = _CC_GRAPH,
1001 _CC_ENUM_LOWER = _CC_LOWER,
1002 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 1003 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 1004 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 1005 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 1006 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 1007 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 1008 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 1009 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
1010} _char_class_number;
1011#endif
1012
63c61c3f 1013#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 1014#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 1015
3018b823 1016#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 1017# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
1018 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
1019 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
1020 #error Need to adjust order of swash_property_names[]
1021# endif
1022
b44daeeb
KW
1023/* This is declared static in each of the few files that this is #defined for
1024 * to keep them from being publicly accessible. Hence there is a small amount
1025 * of wasted space */
1026
922e8cb4
KW
1027static const char* const swash_property_names[] = {
1028 "XPosixWord",
1029 "XPosixDigit",
1030 "XPosixAlpha",
1031 "XPosixLower",
1032 "XPosixUpper",
1033 "XPosixPunct",
1034 "XPosixPrint",
1035 "XPosixAlnum",
b0d691b2
KW
1036 "XPosixGraph",
1037 "Cased"
922e8cb4
KW
1038};
1039#endif
1040
96ac0975
NC
1041# ifdef DOINIT
1042EXTCONST U32 PL_charclass[] = {
1043# include "l1_char_class_tab.h"
1044};
1045
1046# else /* ! DOINIT */
1047EXTCONST U32 PL_charclass[];
1048# endif
1049
265c1f46 1050 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70 1051# define _CC_mask(classnum) (1U << (classnum))
4650c663
KW
1052
1053 /* For internal core Perl use only: the base macro for defining macros like
1054 * isALPHA */
ff7ecfc3 1055# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
f4cd282c 1056 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
4eeeb416 1057
f4cdb42c
KW
1058 /* The mask for the _A versions of the macros; it just adds in the bit for
1059 * ASCII. */
1060# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1061
4650c663
KW
1062 /* For internal core Perl use only: the base macro for defining macros like
1063 * isALPHA_A. The foo_A version makes sure that both the desired bit and
1064 * the ASCII bit are present */
3ded5eb0 1065# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
f4cd282c 1066 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
f4cdb42c
KW
1067 == _CC_mask_A(classnum)))
1068
f4cdb42c 1069# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 1070# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
1071# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1072# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1073# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
1074# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1075# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1076# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1077# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1078# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1079# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1080# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1081# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1082# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
d95f8b6a 1083# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
3ded5eb0
KW
1084# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1085# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1086# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1087
1088 /* continuation character for legal NAME in \N{NAME} */
1089# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1090
1091# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1092# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1093# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1094# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1095# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1096# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1097# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1098# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1099# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1100# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
f4cdb42c 1101
f12c0118
KW
1102 /* Participates in a single-character fold with a character above 255 */
1103# define _HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_SIMPLE_FOLD)))
1104
1105 /* Like the above, but also can be part of a multi-char fold */
f4cd282c 1106# 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) (c)] & _CC_mask(_CC_NONLATIN1_FOLD)))
430b7c70 1107
4eeeb416 1108# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
1109# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1110 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
1111# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1112 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
3ffc8c70 1113#else /* else we don't have perl.h */
3ded5eb0
KW
1114
1115 /* If we don't have perl.h, we are compiling a utility program. Below we
1116 * hard-code various macro definitions that wouldn't otherwise be available
3ffc8c70 1117 * to it. */
9fb1bf9d 1118# ifdef EBCDIC
3ded5eb0 1119 /* Use the native functions. They likely will return false for all
4f48cecf 1120 * non-ASCII values, but this makes sure */
b6340bd0
KW
1121# define isLOWER_A(c) (isASCII(c) && islower(c))
1122# define isPRINT_A(c) (isASCII(c) && isprint(c))
b6340bd0 1123# define isUPPER_A(c) (isASCII(c) && isupper(c))
3ded5eb0 1124# else /* ASCII platform. These are coded based on first principals */
9fb1bf9d
KW
1125# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1126# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
2181da29 1127# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
3ded5eb0 1128# endif /* Below are common definitions for ASCII and non-ASCII */
a4d7a999
KW
1129# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
1130# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
3ded5eb0 1131# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
a4d7a999
KW
1132# define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
1133# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
1134# define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
3ded5eb0 1135# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
a4d7a999
KW
1136# define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1137# define isSPACE_A(c) ((c) == ' ' \
1138 || (c) == '\t' \
1139 || (c) == '\n' \
1140 || (c) == '\r' \
1141 || (c) == '\v' \
1142 || (c) == '\f')
3ded5eb0 1143# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
a4d7a999
KW
1144# define isXDIGIT_A(c) (isDIGIT_A(c) \
1145 || ((c) >= 'a' && (c) <= 'f') \
1146 || ((c) <= 'F' && (c) >= 'A'))
3ded5eb0 1147
3ffc8c70
KW
1148 /* The _L1 macros may be unnecessary for the utilities; I (khw) added them
1149 * during debugging, and it seems best to keep them. */
3ded5eb0
KW
1150# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1151# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1152# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1153# define isBLANK_L1(c) (isBLANK_A(c) \
1154 || (FITS_IN_8_BITS(c) \
1155 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1156# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1157# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1158# define isLOWER_L1(c) (isLOWER_A(c) \
1159 || (FITS_IN_8_BITS(c) \
1160 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1161 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1162 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1163 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1164 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1165# define isPRINT_L1(c) (isPRINT_A(c) \
1166 || (FITS_IN_8_BITS(c) \
1167 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1168# define isPSXSPC_L1(c) isSPACE_L1(c)
1169# define isPUNCT_L1(c) (isPUNCT_A(c) \
1170 || (FITS_IN_8_BITS(c) \
1171 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1172 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1173 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1174 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1175 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1176 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1177 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1178# define isSPACE_L1(c) (isSPACE_A(c) \
1179 || (FITS_IN_8_BITS(c) \
1180 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1181 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1182# define isUPPER_L1(c) (isUPPER_A(c) \
1183 || (FITS_IN_8_BITS(c) \
1184 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1185 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1186 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1187# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1188# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1189# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1190 || isBLANK_L1(c) \
1191 || (c) == '-' \
1192 || (c) == '(' \
1193 || (c) == ')')
1194 /* The following are not fully accurate in the above-ASCII range. I (khw)
1195 * don't think it's necessary to be so for the purposes where this gets
1196 * compiled */
1197# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1198# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1199
1200 /* And these aren't accurate at all. They are useful only for above
1201 * Latin1, which utilities and bootstrapping don't deal with */
1202# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1203# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1204
1205 /* Many of the macros later in this file are defined in terms of these. By
1206 * implementing them with a function, which converts the class number into
1207 * a call to the desired macro, all of the later ones work. However, that
1208 * function won't be actually defined when building a utility program (no
1209 * perl.h), and so a compiler error will be generated if one is attempted
1210 * to be used. And the above-Latin1 code points require Unicode tables to
1211 * be present, something unlikely to be the case when bootstrapping */
1212# define _generic_isCC(c, classnum) \
1213 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1214# define _generic_isCC_A(c, classnum) \
1215 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
3ffc8c70 1216#endif /* End of no perl.h */
8a58bdcf 1217
e66b99e9
KW
1218#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1219#define isALPHA(c) isALPHA_A(c)
1220#define isBLANK(c) isBLANK_A(c)
1221#define isCNTRL(c) isCNTRL_A(c)
1222#define isDIGIT(c) isDIGIT_A(c)
1223#define isGRAPH(c) isGRAPH_A(c)
1224#define isIDFIRST(c) isIDFIRST_A(c)
1225#define isLOWER(c) isLOWER_A(c)
1226#define isPRINT(c) isPRINT_A(c)
1227#define isPSXSPC(c) isPSXSPC_A(c)
1228#define isPUNCT(c) isPUNCT_A(c)
1229#define isSPACE(c) isSPACE_A(c)
1230#define isUPPER(c) isUPPER_A(c)
1231#define isWORDCHAR(c) isWORDCHAR_A(c)
1232#define isXDIGIT(c) isXDIGIT_A(c)
1233
1234/* ASCII casing. These could also be written as
1235 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1236 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1237 which uses table lookup and mask instead of subtraction. (This would
1238 work because the _MOD does not apply in the ASCII range) */
68067e4e
DM
1239#define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1240#define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
bbce6d69 1241
25200305
KW
1242/* In the ASCII range, these are equivalent to what they're here defined to be.
1243 * But by creating these definitions, other code doesn't have to be aware of
1244 * this detail */
1245#define toFOLD(c) toLOWER(c)
25200305
KW
1246#define toTITLE(c) toUPPER(c)
1247
c753c8d3
KW
1248#define toLOWER_A(c) toLOWER(c)
1249#define toUPPER_A(c) toUPPER(c)
25200305
KW
1250#define toFOLD_A(c) toFOLD(c)
1251#define toTITLE_A(c) toTITLE(c)
1a0901db 1252
4650c663 1253/* Use table lookup for speed; returns the input itself if is out-of-range */
b2bf251f 1254#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1255 ? (c) \
f4cd282c 1256 : PL_latin1_lc[ (U8) (c) ])
c753c8d3
KW
1257#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1258
1a0901db 1259/* Modified uc. Is correct uc except for three non-ascii chars which are
4650c663
KW
1260 * all mapped to one of them, and these need special handling; returns the
1261 * input itself if is out-of-range */
b2bf251f 1262#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1263 ? (c) \
f4cd282c 1264 : PL_mod_latin1_uc[ (U8) (c) ])
31f05a37 1265#define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale
84061b6a 1266
beab9ebe
KW
1267/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
1268
1269/* For internal core Perl use only: the base macro for defining macros like
1270 * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point
31f05a37
KW
1271 * (0-255) to check. In a UTF-8 locale, the result is the same as calling
1272 * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like
1273 * _CC_UPPER, which gives the class number for doing this. For non-UTF-8
1274 * locales, the code to actually do the test this is passed in 'non_utf8'. If
1275 * 'c' is above 255, 0 is returned. For accessing the full range of possible
1276 * code points under locale rules, use the macros based on _generic_LC_uvchr
1277 * instead of this. */
beab9ebe
KW
1278#define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \
1279 (! FITS_IN_8_BITS(c) \
1280 ? 0 \
31f05a37
KW
1281 : IN_UTF8_CTYPE_LOCALE \
1282 ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \
beab9ebe
KW
1283 : cBOOL(non_utf8))
1284
1285/* For internal core Perl use only: a helper macro for defining macros like
1286 * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to
1287 * actually do this test is passed in 'non_utf8_func', which is called on 'c',
1288 * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See
1289 * _generic_LC_base for more info */
1290#define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \
1291 _generic_LC_base(c,utf8_locale_classnum, \
1292 non_utf8_func( (_LC_CAST) (c)))
1293
1294/* For internal core Perl use only: like _generic_LC, but also returns TRUE if
1295 * 'c' is the platform's native underscore character */
1296#define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \
1297 _generic_LC_base(c, utf8_locale_classnum, \
1298 (non_utf8_func( (_LC_CAST) (c)) \
1299 || (char)(c) == '_'))
1300
1301/* These next three are also for internal core Perl use only: case-change
1302 * helper macros */
1303#define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \
1304 ? (c) \
31f05a37
KW
1305 : (IN_UTF8_CTYPE_LOCALE) \
1306 ? PL_latin1_lc[ (U8) (c) ] \
beab9ebe
KW
1307 : function((cast)(c)))
1308
31f05a37
KW
1309/* Note that the result can be larger than a byte in a UTF-8 locale. It
1310 * returns a single value, so can't adequately return the upper case of LATIN
1311 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1312 * values "SS"); instead it asserts against that under DEBUGGING, and
1313 * otherwise returns its input */
beab9ebe
KW
1314#define _generic_toUPPER_LC(c, function, cast) \
1315 (! FITS_IN_8_BITS(c) \
1316 ? (c) \
31f05a37
KW
1317 : ((! IN_UTF8_CTYPE_LOCALE) \
1318 ? function((cast)(c)) \
1319 : ((((U8)(c)) == MICRO_SIGN) \
1320 ? GREEK_CAPITAL_LETTER_MU \
1321 : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \
1322 ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \
1323 : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \
1324 ? (__ASSERT_(0) (c)) \
1325 : PL_mod_latin1_uc[ (U8) (c) ])))))
1326
1327/* Note that the result can be larger than a byte in a UTF-8 locale. It
1328 * returns a single value, so can't adequately return the fold case of LATIN
1329 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1330 * values "ss"); instead it asserts against that under DEBUGGING, and
1331 * otherwise returns its input */
beab9ebe 1332#define _generic_toFOLD_LC(c, function, cast) \
4d7db1b9 1333 ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \
31f05a37 1334 ? GREEK_SMALL_LETTER_MU \
4d7db1b9
KW
1335 : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \
1336 || (c) != LATIN_SMALL_LETTER_SHARP_S) \
1337 _generic_toLOWER_LC(c, function, cast)))
beab9ebe 1338
84061b6a
KW
1339/* Use the libc versions for these if available. */
1340#if defined(HAS_ISASCII) && ! defined(USE_NEXT_CTYPE)
1341# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c)))
1342#else
1343# define isASCII_LC(c) isASCII(c)
1344#endif
1345
1346#if defined(HAS_ISBLANK) && ! defined(USE_NEXT_CTYPE)
1347# define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank)
95f34b6f
KW
1348#else /* Unlike isASCII, varies if in a UTF-8 locale */
1349# define isBLANK_LC(c) (IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c)
84061b6a
KW
1350#endif
1351
4650c663 1352#ifdef USE_NEXT_CTYPE /* NeXT computers */
bbce6d69 1353
beab9ebe
KW
1354# define _LC_CAST unsigned int /* Needed by _generic_LC. NeXT functions
1355 use this as their input type */
1356
1357# define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, NXIsAlpha)
1358# define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, NXIsAlNum)
1359# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, NXIsCntrl)
1360# define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, NXIsDigit)
1361# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, NXIsGraph)
1362# define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, NXIsAlpha)
1363# define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, NXIsLower)
1364# define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, NXIsPrint)
1365# define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, NXIsPunct)
1366# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, NXIsSpace)
1367# define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, NXIsUpper)
1368# define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, NXIsAlNum)
1369# define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, NXIsXdigit)
1370
1371# define toLOWER_LC(c) _generic_toLOWER_LC((c), NXToLower, unsigned int)
1372# define toUPPER_LC(c) _generic_toUPPER_LC((c), NXToUpper, unsigned int)
1373# define toFOLD_LC(c) _generic_toFOLD_LC((c), NXToLower, unsigned int)
bbce6d69
PP
1374
1375#else /* !USE_NEXT_CTYPE */
b8c5462f 1376
841f64f7
BF
1377# define _LC_CAST U8
1378
375f5f06
KW
1379# ifdef WIN32
1380 /* The Windows functions don't bother to follow the POSIX standard, which
1381 * for example says that something can't both be a printable and a control.
1382 * But Windows treats the \t control as a printable, and does such things
1383 * as making superscripts into both digits and punctuation. This tames
1384 * these flaws by assuming that the definitions of both controls and space
1385 * are correct, and then making sure that other definitions don't have
1386 * weirdnesses, by making sure that isalnum() isn't also ispunct(), etc.
1387 * Not all possible weirdnesses are checked for, just the ones that were
1388 * detected on actual Microsoft code pages */
1389
1390# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1391# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1392
1393# define isALPHA_LC(c) (_generic_LC(c, _CC_ALPHA, isalpha) && isALPHANUMERIC_LC(c))
1394# define isALPHANUMERIC_LC(c) (_generic_LC(c, _CC_ALPHANUMERIC, isalnum) && ! isPUNCT_LC(c))
1395# define isDIGIT_LC(c) (_generic_LC(c, _CC_DIGIT, isdigit) && isALPHANUMERIC_LC(c))
1396# define isGRAPH_LC(c) (_generic_LC(c, _CC_GRAPH, isgraph) && isPRINT_LC(c))
1397# define isIDFIRST_LC(c) (((c) == '_') || (_generic_LC(c, _CC_IDFIRST, isalpha) && ! isPUNCT_LC(c)))
1398# define isLOWER_LC(c) (_generic_LC(c, _CC_LOWER, islower) && isALPHA_LC(c))
1399# define isPRINT_LC(c) (_generic_LC(c, _CC_PRINT, isprint) && ! isCNTRL_LC(c))
1400# define isPUNCT_LC(c) (_generic_LC(c, _CC_PUNCT, ispunct) && ! isCNTRL_LC(c))
1401# define isUPPER_LC(c) (_generic_LC(c, _CC_UPPER, isupper) && isALPHA_LC(c))
1402# define isWORDCHAR_LC(c) (((c) == '_') || isALPHANUMERIC_LC(c))
1403# define isXDIGIT_LC(c) (_generic_LC(c, _CC_XDIGIT, isxdigit) && isALPHANUMERIC_LC(c))
1404
1405# define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1406# define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1407# define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
1408
1409# elif defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
4650c663 1410 /* For most other platforms */
beab9ebe 1411
beab9ebe
KW
1412# define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha)
1413# define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum)
1414# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1415# define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit)
1416# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph)
1417# define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha)
1418# define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower)
1419# define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint)
1420# define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct)
1421# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1422# define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper)
1423# define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum)
1424# define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit)
1425
1426
1427# define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1428# define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1429# define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
bbce6d69 1430
4650c663 1431# else /* The final fallback position */
bbce6d69 1432
bbce6d69 1433# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1434# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
b8c5462f 1435# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1436# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1437# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1438# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1439# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1440# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1441# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1442# define isSPACE_LC(c) (isascii(c) && isspace(c))
1443# define isUPPER_LC(c) (isascii(c) && isupper(c))
1444# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1445# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
4650c663 1446
76fde96a 1447# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1448# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
beab9ebe 1449# define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c))
bbce6d69
PP
1450
1451# endif
a0d0e21e 1452#endif /* USE_NEXT_CTYPE */
55204971 1453
eba68aa0
KW
1454#define isIDCONT(c) isWORDCHAR(c)
1455#define isIDCONT_A(c) isWORDCHAR_A(c)
1456#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1457#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1458#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1459
4650c663
KW
1460/* For internal core Perl use only: the base macros for defining macros like
1461 * isALPHA_uni. 'c' is the code point to check. 'classnum' is the POSIX class
1462 * number defined earlier in this file. _generic_uni() is used for POSIX
1463 * classes where there is a macro or function 'above_latin1' that takes the
1464 * single argument 'c' and returns the desired value. These exist for those
1465 * classes which have simple definitions, avoiding the overhead of a hash
1466 * lookup or inversion list binary search. _generic_swash_uni() can be used
1467 * for classes where that overhead is faster than a direct lookup.
1468 * _generic_uni() won't compile if 'c' isn't unsigned, as it won't match the
1469 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so
1470 * have duplicate checks here, so could create versions of the macros that
1471 * don't, but experiments show that gcc optimizes them out anyway. */
66c17564
KW
1472
1473/* Note that all ignore 'use bytes' */
cd500f2f
KW
1474#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1475 ? _generic_isCC(c, classnum) \
1476 : above_latin1(c))
1477#define _generic_swash_uni(classnum, c) ((c) < 256 \
1478 ? _generic_isCC(c, classnum) \
922e8cb4 1479 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1480#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1481#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1482#define isASCII_uni(c) isASCII(c)
1483#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1484#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1485#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1486#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1487#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1488#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1489#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1490#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1491
509fb054
KW
1492/* Posix and regular space are identical above Latin1 */
1493#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1494
cd500f2f 1495#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1496#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1497#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1498#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1499#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1500#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1501
b8d68ded 1502#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1503#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1504#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1505#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1506
4650c663
KW
1507/* For internal core Perl use only: the base macros for defining macros like
1508 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code
1509 * point, not just 0-255. Like _generic_uni, there are two versions, one for
1510 * simple class definitions; the other for more complex. These are like
1511 * _generic_uni, so see it for more info. */
cd500f2f
KW
1512#define _generic_LC_uvchr(latin1, above_latin1, c) \
1513 (c < 256 ? latin1(c) : above_latin1(c))
1514#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1515 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1516
1517#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1518#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1519 _CC_ALPHANUMERIC, c)
509fb054 1520#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1521#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1522#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1523#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1524#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1525#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1526 _is_uni_perl_idcont, c)
cd500f2f
KW
1527#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1528 _is_uni_perl_idstart, c)
1529#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1530#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1531#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1532 space under locale */
cd500f2f
KW
1533#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1534#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1535 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1536#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1537#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1538 _CC_WORDCHAR, c)
1539#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1540
36b00f93 1541#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1542
4650c663
KW
1543/* For internal core Perl use only: the base macros for defining macros like
1544 * isALPHA_utf8. These are like the earlier defined macros, but take an input
1545 * UTF-8 encoded string 'p'. If the input is in the Latin1 range, use
1546 * the Latin1 macro 'classnum' on 'p'. Otherwise use the value given by the
1547 * 'utf8' parameter. This relies on the fact that ASCII characters have the
1548 * same representation whether utf8 or not. Note that it assumes that the utf8
1549 * has been validated, and ignores 'use bytes' */
cd500f2f 1550#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1551 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1552 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c 1553 ? _generic_isCC( \
94bb8c36 1554 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
aea0c08c 1555 *((p)+1 )), \
94bb8c36 1556 classnum) \
252810d0 1557 : utf8)
cd500f2f
KW
1558/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1559 * can be a macro */
1560#define _generic_func_utf8(classnum, above_latin1, p) \
1561 _generic_utf8(classnum, p, above_latin1(p))
4650c663 1562/* Like the above, but passes classnum to _isFOO_utf8(), instead of having an
cd500f2f 1563 * 'above_latin1' parameter */
922e8cb4 1564#define _generic_swash_utf8(classnum, p) \
cd500f2f 1565 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1566
cc8ab7c0 1567/* Like the above, but should be used only when it is known that there are no
ff7ecfc3
KW
1568 * characters in the upper-Latin1 range (128-255 on ASCII platforms) which the
1569 * class is TRUE for. Hence it can skip the tests for this range.
1570 * 'above_latin1' should include its arguments */
3ded5eb0 1571#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1572 (UTF8_IS_INVARIANT(*(p)) \
1573 ? _generic_isCC(*(p), classnum) \
1574 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
3ded5eb0 1575 ? above_latin1 \
cc8ab7c0
KW
1576 : 0)
1577
84238efa
KW
1578/* NOTE that some of these macros have very similar ones in regcharclass.h.
1579 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1580 * there, differing in name only by an underscore from the one here
1581 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1582 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1583 * points; the regcharclass.h ones are implemented as a series of
1584 * "if-else-if-else ..." */
1585
922e8cb4
KW
1586#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1587#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1588#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1589 utf8, the non-utf8 macro works
1590 */
cd500f2f 1591#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
e8fa43e2
KW
1592
1593#ifdef EBCDIC
1594 /* Because all controls are UTF-8 invariants in EBCDIC, we can use this
1595 * more efficient macro instead of the more general one */
1596# define isCNTRL_utf8(p) isCNTRL_L1(p)
1597#else
1598# define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1599#endif
1600
cd500f2f
KW
1601#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1602 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1603#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1604#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1605 _is_utf8_perl_idcont, p)
e5dcd934 1606
c11ff943
KW
1607/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1608 * IDFIRST is an alnum. See
c8362b00 1609 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1610 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1611 * This used to be not the XID version, but we decided to go with the more
1612 * modern Unicode definition */
cd500f2f 1613#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1614 _is_utf8_perl_idstart, p)
e5dcd934 1615
922e8cb4
KW
1616#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1617#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1618
509fb054 1619/* Posix and regular space are identical above Latin1 */
cd500f2f 1620#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1621
922e8cb4 1622#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1623#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1624#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1625#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1626#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1627#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1628 is_XDIGIT_high(p))
a0ed51b3 1629
7d2b61dd 1630#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
509fb054
KW
1631#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1632#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1633#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1634
4650c663
KW
1635/* For internal core Perl use only: the base macros for defining macros like
1636 * isALPHA_LC_utf8. These are like _generic_utf8, but if the first code point
1637 * in 'p' is within the 0-255 range, it uses locale rules from the passed-in
1638 * 'macro' parameter */
94bb8c36
KW
1639#define _generic_LC_utf8(macro, p, utf8) \
1640 (UTF8_IS_INVARIANT(*(p)) \
1641 ? macro(*(p)) \
1642 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1643 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
9201a572
KW
1644 : utf8)
1645
cd500f2f
KW
1646#define _generic_LC_swash_utf8(macro, classnum, p) \
1647 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1648#define _generic_LC_func_utf8(macro, above_latin1, p) \
4650c663 1649 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1650
cd500f2f
KW
1651#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1652 _CC_ALPHANUMERIC, p)
1653#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1654#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1655#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1656#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1657#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1658#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1659#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1660#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1661#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1662#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1663#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1664 space under locale */
cd500f2f
KW
1665#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1666#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1667#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1668#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1669 _CC_WORDCHAR, p)
1670#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1671
fbc19f27
KW
1672/* Macros for backwards compatibility and for completeness when the ASCII and
1673 * Latin1 values are identical */
fbc19f27
KW
1674#define isALPHAU(c) isALPHA_L1(c)
1675#define isDIGIT_L1(c) isDIGIT_A(c)
1676#define isOCTAL(c) isOCTAL_A(c)
1677#define isOCTAL_L1(c) isOCTAL_A(c)
1678#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1679#define isALNUM(c) isWORDCHAR(c)
1680#define isALNUMU(c) isWORDCHAR_L1(c)
1681#define isALNUM_LC(c) isWORDCHAR_LC(c)
1682#define isALNUM_uni(c) isWORDCHAR_uni(c)
1683#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1684#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1685#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1686#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1687#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1688#define isALNUMC(c) isALPHANUMERIC(c)
1689#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1690#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1691#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1692#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1693#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1694
2bd1cbf6
KW
1695/* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII,
1696 * except that they don't necessarily mean the same characters, e.g. CTRL-D is
1697 * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC.
1698 * '?' is special-cased on EBCDIC to APC, which is the control there that is
1699 * the outlier from the block that contains the other controls, just like
1700 * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0
1701 * block. If it weren't special cased, it would yield a non-control.
1702 * The conversion works both ways, so CTRL('D') is 4, and CTRL(4) is D, etc. */
1703#ifndef EBCDIC
1704# define toCTRL(c) (toUPPER(c) ^ 64)
1705#else
0ed2b00b
KW
1706# define toCTRL(c) ((isPRINT_A(c)) \
1707 ? UNLIKELY((c) == '?') \
1708 ? QUESTION_MARK_CTRL \
1709 : (NATIVE_TO_LATIN1(toUPPER(c)) ^ 64) \
1710 : UNLIKELY((c) == QUESTION_MARK_CTRL) \
1711 ? ((c) == '?') \
1712 : (LATIN1_TO_NATIVE((c) ^ 64)))
2bd1cbf6 1713#endif
bbce6d69 1714
dea28490
JJ
1715/* Line numbers are unsigned, 32 bits. */
1716typedef U32 line_t;
e5dcd934 1717#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1718
91152fc1
DG
1719/* Helpful alias for version prescan */
1720#define is_LAX_VERSION(a,b) \
1721 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1722
1723#define is_STRICT_VERSION(a,b) \
1724 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1725
1726#define BADVERSION(a,b,c) \
1727 if (b) { \
1728 *b = c; \
1729 } \
1730 return a;
8c52afec 1731
cb27eebd
KW
1732/* Converts a character known to represent a hexadecimal digit (0-9, A-F, or
1733 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer,
1734 * which is advanced. The input is validated only by an assert() in DEBUGGING
1735 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and
1736 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */
1737#define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \
1738 ? (c) \
1739 : ((c) + 9))))
902008b8
KW
1740#define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \
1741 ? (*(s)++) \
1742 : (*(s)++ + 9))))
95a59cab 1743
cb27eebd
KW
1744/* Converts a character known to represent an octal digit (0-7) to its numeric
1745 * value. The input is validated only by an assert() in DEBUGGING builds. In
1746 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */
1747#define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c)))
1748
8e84507e 1749/*
ccfc67b7
JH
1750=head1 Memory Management
1751
a02a5408 1752=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1753The XSUB-writer's interface to the C C<malloc> function.
1754
0d7b2759
MH
1755Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1756
c5008215
JC
1757In 5.9.3, Newx() and friends replace the older New() API, and drops
1758the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1759themselves. This aid has been superseded by a new build option,
d10b4965 1760PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1761there for use in XS modules supporting older perls.
1762
a02a5408 1763=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1764The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1765cast. See also C<Newx>.
954c1994 1766
0d7b2759
MH
1767Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1768
a02a5408 1769=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1770The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1771memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1772
0d7b2759
MH
1773Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1774
954c1994
GS
1775=for apidoc Am|void|Renew|void* ptr|int nitems|type
1776The XSUB-writer's interface to the C C<realloc> function.
1777
0d7b2759
MH
1778Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1779
954c1994
GS
1780=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1781The XSUB-writer's interface to the C C<realloc> function, with
1782cast.
1783
0d7b2759
MH
1784Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1785
49b8b560 1786=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1787The XSUB-writer's interface to the C C<free> function.
1788
0d7b2759
MH
1789This should B<ONLY> be used on memory obtained using L<"Newx"> and friends.
1790
954c1994
GS
1791=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1792The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1793source, C<dest> is the destination, C<nitems> is the number of items, and
1794C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1795
e90e2364 1796=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
72d33970
FC
1797Like C<Move> but returns dest. Useful
1798for encouraging compilers to tail-call
e90e2364
NC
1799optimise.
1800
954c1994
GS
1801=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1802The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1803source, C<dest> is the destination, C<nitems> is the number of items, and
1804C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1805
e90e2364
NC
1806=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1807
72d33970
FC
1808Like C<Copy> but returns dest. Useful
1809for encouraging compilers to tail-call
e90e2364
NC
1810optimise.
1811
954c1994
GS
1812=for apidoc Am|void|Zero|void* dest|int nitems|type
1813
1814The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1815destination, C<nitems> is the number of items, and C<type> is the type.
1816
e90e2364
NC
1817=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1818
72d33970
FC
1819Like C<Zero> but returns dest. Useful
1820for encouraging compilers to tail-call
e90e2364
NC
1821optimise.
1822
da5d8dbb 1823=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1824This is an architecture-independent macro to copy one structure to another.
954c1994 1825
7e337ee0
JH
1826=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1827
1828Fill up memory with a byte pattern (a byte repeated over and over
1829again) that hopefully catches attempts to access uninitialized memory.
1830
1831=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1832
1833PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1834
1c12ffb4 1835=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1836
1837PoisonWith(0xEF) for catching access to freed memory.
1838
9965345d
JH
1839=for apidoc Am|void|Poison|void* dest|int nitems|type
1840
7e337ee0 1841PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1842
1843=cut */
954c1994 1844
561b68a9
SH
1845/* Maintained for backwards-compatibility only. Use newSV() instead. */
1846#ifndef PERL_CORE
ff06c60c 1847#define NEWSV(x,len) newSV(len)
561b68a9 1848#endif
ff06c60c 1849
19a94d75
JH
1850#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1851
2720f793
JH
1852/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1853 * overly eager compilers that will bleat about e.g.
1854 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1855#ifdef PERL_MALLOC_WRAP
f019c49e 1856#define MEM_WRAP_CHECK(n,t) \
5d9574c1 1857 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (croak_memory_wrap(),0))
27d5b266 1858#define MEM_WRAP_CHECK_1(n,t,a) \
5d9574c1 1859 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1860#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1861
d1decf2b 1862#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
1863#else
1864
410319be
NC
1865#define MEM_WRAP_CHECK(n,t)
1866#define MEM_WRAP_CHECK_1(n,t,a)
1867#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1868#define MEM_WRAP_CHECK_(n,t)
1869
6fff79ce 1870#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1871
1936d2a7 1872#endif
8b44ba4c 1873
fe4f188c 1874#ifdef PERL_MEM_LOG
46c6c7e2 1875/*
9f653bb5 1876 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1877 * go through functions, which are handy for debugging breakpoints, but
1878 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1879 * line number, and C function name if available) passed in. This info can
1880 * then be used for logging the calls, for which one gets a sample
73d1d973 1881 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1882 *
46c6c7e2 1883 * Known problems:
94e892a6 1884 * - not all memory allocs get logged, only those
46c6c7e2 1885 * that go through Newx() and derivatives (while all
94e892a6 1886 * Safefrees do get logged)
46c6c7e2
JH
1887 * - __FILE__ and __LINE__ do not work everywhere
1888 * - __func__ or __FUNCTION__ even less so
1889 * - I think more goes on after the perlio frees but
1890 * the thing is that STDERR gets closed (as do all
1891 * the file descriptors)
1892 * - no deeper calling stack than the caller of the Newx()
1893 * or the kind, but do I look like a C reflection/introspection
1894 * utility to you?
1895 * - the function prototypes for the logging functions
1896 * probably should maybe be somewhere else than handy.h
1897 * - one could consider inlining (macrofying) the logging
1898 * for speed, but I am too lazy
1899 * - one could imagine recording the allocations in a hash,
1900 * (keyed by the allocation address?), and maintain that
1901 * through reallocs and frees, but how to do that without
1902 * any News() happening...?
73d1d973 1903 * - lots of -Ddefines to get useful/controllable output
b953482e 1904 * - lots of ENV reads
46c6c7e2
JH
1905 */
1906
12754f92 1907PERL_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 1908
12754f92 1909PERL_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 1910
12754f92 1911PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1912
0b0ab801 1913# ifdef PERL_CORE
73d1d973 1914# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1915enum mem_log_type {
1916 MLT_ALLOC,
1917 MLT_REALLOC,
d7a2c63c
MHM
1918 MLT_FREE,
1919 MLT_NEW_SV,
1920 MLT_DEL_SV
0b0ab801
MHM
1921};
1922# endif
12754f92 1923# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1924void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1925void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1926# endif
0b0ab801
MHM
1927# endif
1928
fe4f188c
JH
1929#endif
1930
1931#ifdef PERL_MEM_LOG
d1401ee9
MHM
1932#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1933#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1934#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1935#endif
1936
1937#ifndef MEM_LOG_ALLOC
1938#define MEM_LOG_ALLOC(n,t,a) (a)
1939#endif
1940#ifndef MEM_LOG_REALLOC
1941#define MEM_LOG_REALLOC(n,t,v,a) (a)
1942#endif
1943#ifndef MEM_LOG_FREE
1944#define MEM_LOG_FREE(a) (a)
1945#endif
1946
d1401ee9
MHM
1947#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1948#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1949#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1950
1951#ifndef PERL_CORE
a02a5408
JC
1952/* pre 5.9.x compatibility */
1953#define New(x,v,n,t) Newx(v,n,t)
1954#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1955#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1956#endif
a02a5408 1957
ff68c719 1958#define Renew(v,n,t) \
d1401ee9 1959 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1960#define Renewc(v,n,t,c) \
d1401ee9 1961 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1962
1963#ifdef PERL_POISON
1964#define Safefree(d) \
06c0cc96 1965 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1966#else
fe4f188c 1967#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1968#endif
55497cff 1969
8b44ba4c
NC
1970#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1971#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1972#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1973
8b44ba4c
NC
1974#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1975#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1976#ifdef HAS_MEMSET
8b44ba4c 1977#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1978#else
8b44ba4c
NC
1979/* Using bzero(), which returns void. */
1980#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1981#endif
1982
7e337ee0
JH
1983#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1984#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1985#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1986#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1987
caa674f3
DD
1988#ifdef PERL_POISON
1989# define PERL_POISON_EXPR(x) x
1990#else
1991# define PERL_POISON_EXPR(x)
1992#endif
1993
2304df62 1994#ifdef USE_STRUCT_COPY
ff68c719 1995#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1996#else
1997#define StructCopy(s,d,t) Copy(s,d,1,t)
1998#endif
2cc61e15 1999
c3caa5c3
JH
2000/* C_ARRAY_LENGTH is the number of elements in the C array (so you
2001 * want your zero-based indices to be less than but not equal to).
2002 *
2003 * C_ARRAY_END is one past the last: half-open/half-closed range,
2004 * not last-inclusive range. */
622913ab 2005#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
c3caa5c3 2006#define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a))
622913ab 2007
2cc61e15
DD
2008#ifdef NEED_VA_COPY
2009# ifdef va_copy
2010# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 2011# else
a1866d1b
JH
2012# if defined(__va_copy)
2013# define Perl_va_copy(s, d) __va_copy(d, s)
2014# else
2015# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
2016# endif
2cc61e15
DD
2017# endif
2018#endif
2019
472d47bc
SB
2020/* convenience debug macros */
2021#ifdef USE_ITHREADS
2022#define pTHX_FORMAT "Perl interpreter: 0x%p"
2023#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
2024#define pTHX_VALUE_ (void *)my_perl,
2025#define pTHX_VALUE (void *)my_perl
2026#define pTHX__VALUE_ ,(void *)my_perl,
2027#define pTHX__VALUE ,(void *)my_perl
472d47bc 2028#else
3609ea0d 2029#define pTHX_FORMAT
472d47bc 2030#define pTHX__FORMAT
3609ea0d 2031#define pTHX_VALUE_
472d47bc 2032#define pTHX_VALUE
3609ea0d 2033#define pTHX__VALUE_
472d47bc
SB
2034#define pTHX__VALUE
2035#endif /* USE_ITHREADS */
3609ea0d 2036
2acdbac1
NC
2037/* Perl_deprecate was not part of the public API, and did not have a deprecate()
2038 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
2039 CPAN::Unpack show any users outside the core. */
2040#ifdef PERL_CORE
d1d15184 2041# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
2042#endif
2043
dfff4baf
BF
2044/* Internal macros to deal with gids and uids */
2045#ifdef PERL_CORE
2046
2047# if Uid_t_size > IVSIZE
2048# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
2049# define SvUID(sv) SvNV(sv)
2050# else
2051# if Uid_t_sign <= 0
2052# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
2053# define SvUID(sv) SvIV(sv)
2054# else
2055# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
2056# define SvUID(sv) SvUV(sv)
2057# endif
2058# endif /* Uid_t_size */
2059
2060# if Gid_t_size > IVSIZE
2061# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
2062# define SvGID(sv) SvNV(sv)
2063# else
2064# if Gid_t_sign <= 0
2065# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
2066# define SvGID(sv) SvIV(sv)
2067# else
2068# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
2069# define SvGID(sv) SvUV(sv)
2070# endif
2071# endif /* Gid_t_size */
2072
2073#endif
2074
9d745869
KW
2075#endif /* HANDY_H */
2076
e9a8c099
MHM
2077/*
2078 * Local variables:
2079 * c-indentation-style: bsd
2080 * c-basic-offset: 4
14d04a33 2081 * indent-tabs-mode: nil
e9a8c099
MHM
2082 * End:
2083 *
14d04a33 2084 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 2085 */