This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
EBCDIC has the unicode bug too
[perl5.git] / handy.h
CommitLineData
a0d0e21e 1/* handy.h
a687059c 2 *
1129b882 3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000,
da5d8dbb 4 * 2001, 2002, 2004, 2005, 2006, 2007, 2008, 2012 by Larry Wall and others
a687059c 5 *
6e21c824
LW
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
8d063cd8 8 *
8d063cd8
LW
9 */
10
9d745869
KW
11#ifndef HANDY_H /* Guard against nested #inclusion */
12#define HANDY_H
13
85e6fe83 14#if !defined(__STDC__)
378cc40b
LW
15#ifdef NULL
16#undef NULL
17#endif
a687059c 18# define NULL 0
85e6fe83
LW
19#endif
20
24792b8d
NC
21#ifndef PERL_CORE
22# define Null(type) ((type)NULL)
954c1994
GS
23
24/*
ccfc67b7 25=head1 Handy Values
954c1994 26
3609ea0d 27=for apidoc AmU||Nullch
24792b8d 28Null character pointer. (No longer available when C<PERL_CORE> is defined.)
2307c6d0 29
954c1994 30=for apidoc AmU||Nullsv
24792b8d 31Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
32
33=cut
34*/
35
24792b8d
NC
36# define Nullch Null(char*)
37# define Nullfp Null(PerlIO*)
38# define Nullsv Null(SV*)
39#endif
8d063cd8 40
641d3f0b
PP
41#ifdef TRUE
42#undef TRUE
43#endif
44#ifdef FALSE
45#undef FALSE
46#endif
47#define TRUE (1)
48#define FALSE (0)
49
cf3f0ffb
DM
50/* The MUTABLE_*() macros cast pointers to the types shown, in such a way
51 * (compiler permitting) that casting away const-ness will give a warning;
52 * e.g.:
53 *
54 * const SV *sv = ...;
55 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
56 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
57 */
58
b1bc3f34 59#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 60# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
61#else
62# define MUTABLE_PTR(p) ((void *) (p))
63#endif
64
a062e10d 65#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 66#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 67#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 68#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 69#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 70#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 71
bd31be4b
NC
72#ifdef I_STDBOOL
73# include <stdbool.h>
74# ifndef HAS_BOOL
75# define HAS_BOOL 1
76# endif
77#endif
78
8e84507e 79/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
80 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
81 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 82 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
83 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
84 g++ can be identified by __GNUG__.
85 Andy Dougherty February 2000
5d94fbed 86*/
3609ea0d 87#ifdef __GNUG__ /* GNU g++ has bool built-in */
5d94fbed 88# ifndef HAS_BOOL
c1d22f6b 89# define HAS_BOOL 1
5d94fbed 90# endif
5d94fbed
AD
91#endif
92
641d3f0b
PP
93/* The NeXT dynamic loader headers will not build with the bool macro
94 So declare them now to clear confusion.
95*/
8f1f23e8 96#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
97# undef FALSE
98# undef TRUE
99 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
100# define ENUM_BOOL 1
101# ifndef HAS_BOOL
102# define HAS_BOOL 1
103# endif /* !HAS_BOOL */
8f1f23e8 104#endif /* NeXT || __NeXT__ */
641d3f0b 105
5d94fbed 106#ifndef HAS_BOOL
cc3315ba 107# if defined(VMS)
5d94fbed
AD
108# define bool int
109# else
110# define bool char
111# endif
c1d22f6b 112# define HAS_BOOL 1
a687059c 113#endif
0d3e774c 114
236415e2
KW
115/* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
116 * defined as char for example, then the cast from int is
117 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
118 * AIX */
18f5643b 119#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 120
46c6c7e2 121/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
122 * XXX Should really be a Configure probe, with HAS__FUNCTION__
123 * and FUNCTION__ as results.
124 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
125#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
126# define FUNCTION__ __func__
127#else
70764c00 128# if (defined(_MSC_VER) && _MSC_VER < 1300) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 129 (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
130# define FUNCTION__ ""
131# else
132# define FUNCTION__ __FUNCTION__ /* Common extension. */
133# endif
134#endif
135
27d4fb96
PP
136/* XXX A note on the perl source internal type system. The
137 original intent was that I32 be *exactly* 32 bits.
138
139 Currently, we only guarantee that I32 is *at least* 32 bits.
140 Specifically, if int is 64 bits, then so is I32. (This is the case
141 for the Cray.) This has the advantage of meshing nicely with
142 standard library calls (where we pass an I32 and the library is
143 expecting an int), but the disadvantage that an I32 is not 32 bits.
144 Andy Dougherty August 1996
24fef2a7 145
dc45a647
MB
146 There is no guarantee that there is *any* integral type with
147 exactly 32 bits. It is perfectly legal for a system to have
148 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 149
dc45a647
MB
150 Similarly, there is no guarantee that I16 and U16 have exactly 16
151 bits.
693762b4 152
8e84507e
NIS
153 For dealing with issues that may arise from various 32/64-bit
154 systems, we will ask Configure to check out
8175356b 155
3609ea0d
JH
156 SHORTSIZE == sizeof(short)
157 INTSIZE == sizeof(int)
158 LONGSIZE == sizeof(long)
dc45a647 159 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 160 PTRSIZE == sizeof(void *)
dc45a647
MB
161 DOUBLESIZE == sizeof(double)
162 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 163
27d4fb96
PP
164*/
165
69512466
JH
166#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
167# include <inttypes.h>
dd0eed91
JH
168# ifdef INT32_MIN_BROKEN
169# undef INT32_MIN
170# define INT32_MIN (-2147483647-1)
171# endif
172# ifdef INT64_MIN_BROKEN
173# undef INT64_MIN
174# define INT64_MIN (-9223372036854775807LL-1)
175# endif
69512466
JH
176#endif
177
8175356b
JH
178typedef I8TYPE I8;
179typedef U8TYPE U8;
180typedef I16TYPE I16;
181typedef U16TYPE U16;
182typedef I32TYPE I32;
183typedef U32TYPE U32;
6b8eaf93
JH
184#ifdef PERL_CORE
185# ifdef HAS_QUAD
8175356b
JH
186typedef I64TYPE I64;
187typedef U64TYPE U64;
6b8eaf93
JH
188# endif
189#endif /* PERL_CORE */
8175356b 190
69512466 191#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
80524f33
JL
192# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
193# define PeRl_INT64_C(c) CAT2(c,LL)
194# define PeRl_UINT64_C(c) CAT2(c,ULL)
195# else
52865553
SH
196# if QUADKIND == QUAD_IS___INT64
197# define PeRl_INT64_C(c) CAT2(c,I64)
198# define PeRl_UINT64_C(c) CAT2(c,UI64)
69512466 199# else
52865553
SH
200# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
201# define PeRl_INT64_C(c) CAT2(c,L)
202# define PeRl_UINT64_C(c) CAT2(c,UL)
69512466 203# else
80524f33
JL
204# define PeRl_INT64_C(c) ((I64TYPE)(c))
205# define PeRl_UINT64_C(c) ((U64TYPE)(c))
69512466
JH
206# endif
207# endif
e8c95190 208# endif
80524f33
JL
209# ifndef UINT64_C
210# define UINT64_C(c) PeRl_UINT64_C(c)
211# endif
212# ifndef INT64_C
213# define INT64_C(c) PeRl_INT64_C(c)
214# endif
e8c95190 215#endif
dc750f2e 216
d8668976 217#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 218
5ff3f7a4
GS
219/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
220 Please search CHAR_MAX in perl.h for further details. */
221#define U8_MAX UINT8_MAX
222#define U8_MIN UINT8_MIN
223
5ff3f7a4
GS
224#define I16_MAX INT16_MAX
225#define I16_MIN INT16_MIN
226#define U16_MAX UINT16_MAX
227#define U16_MIN UINT16_MIN
228
5ff3f7a4
GS
229#define I32_MAX INT32_MAX
230#define I32_MIN INT32_MIN
0e983133
GS
231#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
232# define U32_MAX UINT32_MAX
233#else
234# define U32_MAX 4294967295U
235#endif
5ff3f7a4
GS
236#define U32_MIN UINT32_MIN
237
238#else
239
5c9fa16e
KA
240/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
241 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
242#define U8_MAX PERL_UCHAR_MAX
243#define U8_MIN PERL_UCHAR_MIN
79072805 244
27d4fb96
PP
245#define I16_MAX PERL_SHORT_MAX
246#define I16_MIN PERL_SHORT_MIN
247#define U16_MAX PERL_USHORT_MAX
248#define U16_MIN PERL_USHORT_MIN
79072805 249
c4f23d77 250#if LONGSIZE > 4
27d4fb96
PP
251# define I32_MAX PERL_INT_MAX
252# define I32_MIN PERL_INT_MIN
253# define U32_MAX PERL_UINT_MAX
254# define U32_MIN PERL_UINT_MIN
79072805 255#else
27d4fb96
PP
256# define I32_MAX PERL_LONG_MAX
257# define I32_MIN PERL_LONG_MIN
258# define U32_MAX PERL_ULONG_MAX
259# define U32_MIN PERL_ULONG_MIN
79072805
LW
260#endif
261
5ff3f7a4
GS
262#endif
263
58a9a5d5 264/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
265#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
266#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
267#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
268
ff68c719 269#define Ctl(ch) ((ch) & 037)
8d063cd8 270
3fe05580
MHM
271/*
272=head1 SV-Body Allocation
273
274=for apidoc Ama|SV*|newSVpvs|const char* s
275Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
276
84bafc02
NC
277=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
278Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
279pair.
280
3fe05580
MHM
281=for apidoc Ama|SV*|newSVpvs_share|const char* s
282Like C<newSVpvn_share>, but takes a literal string instead of a string/length
283pair and omits the hash parameter.
284
9dcc53ea
Z
285=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
286Like C<sv_catpvn_flags>, but takes a literal string instead of a
287string/length pair.
288
289=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
290Like C<sv_catpvn_nomg>, but takes a literal string instead of a
291string/length pair.
292
7b4e7469 293=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
294Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
295
9dcc53ea
Z
296=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
297Like C<sv_catpvn_mg>, but takes a literal string instead of a
298string/length pair.
299
7b4e7469 300=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
301Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
302
9dcc53ea
Z
303=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
304Like C<sv_setpvn_mg>, but takes a literal string instead of a
305string/length pair.
306
307=for apidoc Am|SV *|sv_setref_pvs|const char* s
308Like C<sv_setref_pvn>, but takes a literal string instead of a
309string/length pair.
310
3fe05580
MHM
311=head1 Memory Management
312
313=for apidoc Ama|char*|savepvs|const char* s
314Like C<savepvn>, but takes a literal string instead of a string/length pair.
315
9dcc53ea
Z
316=for apidoc Ama|char*|savesharedpvs|const char* s
317A version of C<savepvs()> which allocates the duplicate string in memory
318which is shared between threads.
319
3fe05580
MHM
320=head1 GV Functions
321
322=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
323Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
324
325=head1 Hash Manipulation Functions
326
327=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
328Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
329
330=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
331Like C<hv_store>, but takes a literal string instead of a string/length pair
332and omits the hash parameter.
333
510966aa
Z
334=head1 Lexer interface
335
336=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
337
338Like L</lex_stuff_pvn>, but takes a literal string instead of a
339string/length pair.
340
3fe05580
MHM
341=cut
342*/
343
926bb54c
KW
344/* concatenating with "" ensures that only literal strings are accepted as
345 * argument */
26bfd924 346#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 347
926bb54c
KW
348/* note that STR_WITH_LEN() can't be used as argument to macros or functions
349 * that under some configurations might be macros, which means that it requires
350 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
351 */
352
353/* STR_WITH_LEN() shortcuts */
354#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
355#define newSVpvs_flags(str,flags) \
356 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 357#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
358#define sv_catpvs_flags(sv, str, flags) \
359 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
360#define sv_catpvs_nomg(sv, str) \
361 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
362#define sv_catpvs(sv, str) \
363 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
364#define sv_catpvs_mg(sv, str) \
365 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 366#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
367#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
368#define sv_setref_pvs(rv, classname, str) \
369 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 370#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
371#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
372#define gv_stashpvs(str, create) \
373 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
374#define gv_fetchpvs(namebeg, add, sv_type) \
375 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
376#define gv_fetchpvn(namebeg, len, add, sv_type) \
377 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
378#define sv_catxmlpvs(dsv, str, utf8) \
379 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
380#define hv_fetchs(hv,key,lval) \
381 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
382 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
383 : HV_FETCH_JUST_SV, NULL, 0))
384
385#define hv_stores(hv,key,val) \
386 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
387 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 388
510966aa
Z
389#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
390
b96d8cd9
NC
391#define get_cvs(str, flags) \
392 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 393
954c1994 394/*
ccfc67b7
JH
395=head1 Miscellaneous Functions
396
954c1994
GS
397=for apidoc Am|bool|strNE|char* s1|char* s2
398Test two strings to see if they are different. Returns true or
399false.
400
401=for apidoc Am|bool|strEQ|char* s1|char* s2
402Test two strings to see if they are equal. Returns true or false.
403
404=for apidoc Am|bool|strLT|char* s1|char* s2
405Test two strings to see if the first, C<s1>, is less than the second,
406C<s2>. Returns true or false.
407
408=for apidoc Am|bool|strLE|char* s1|char* s2
409Test two strings to see if the first, C<s1>, is less than or equal to the
410second, C<s2>. Returns true or false.
411
412=for apidoc Am|bool|strGT|char* s1|char* s2
413Test two strings to see if the first, C<s1>, is greater than the second,
414C<s2>. Returns true or false.
415
416=for apidoc Am|bool|strGE|char* s1|char* s2
417Test two strings to see if the first, C<s1>, is greater than or equal to
418the second, C<s2>. Returns true or false.
419
420=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
421Test two strings to see if they are different. The C<len> parameter
422indicates the number of bytes to compare. Returns true or false. (A
423wrapper for C<strncmp>).
424
425=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
426Test two strings to see if they are equal. The C<len> parameter indicates
427the number of bytes to compare. Returns true or false. (A wrapper for
428C<strncmp>).
429
430=cut
431*/
432
8d063cd8
LW
433#define strNE(s1,s2) (strcmp(s1,s2))
434#define strEQ(s1,s2) (!strcmp(s1,s2))
435#define strLT(s1,s2) (strcmp(s1,s2) < 0)
436#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
437#define strGT(s1,s2) (strcmp(s1,s2) > 0)
438#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
439#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
440#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 441
36477c24
PP
442#ifdef HAS_MEMCMP
443# define memNE(s1,s2,l) (memcmp(s1,s2,l))
444# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
445#else
446# define memNE(s1,s2,l) (bcmp(s1,s2,l))
447# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
448#endif
449
568a785a 450#define memEQs(s1, l, s2) \
8e6497b6 451 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
452#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
453
bbce6d69
PP
454/*
455 * Character classes.
456 *
457 * Unfortunately, the introduction of locales means that we
458 * can't trust isupper(), etc. to tell the truth. And when
459 * it comes to /\w+/ with tainting enabled, we *must* be able
460 * to trust our character classes.
461 *
462 * Therefore, the default tests in the text of Perl will be
463 * independent of locale. Any code that wants to depend on
464 * the current locale will use the tests that begin with "lc".
465 */
466
2304df62
AD
467#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
468# ifndef CTYPE256
469# define CTYPE256
470# endif
471#endif
472
954c1994 473/*
ccfc67b7
JH
474
475=head1 Character classes
243effed
KW
476This section is about functions (really macros) that classify characters
477into types, such as punctuation versus alphabetic, etc. Most of these are
478analogous to regular expression character classes. (See
479L<perlrecharclass/POSIX Character Classes>.) There are several variants for
480each class. (Not all macros have all variants; each item below lists the
481ones valid for it.) None are affected by C<use bytes>, and only the ones
482with C<LC> in the name are affected by the current locale.
483
484The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
485C<U8>) as input and returns a boolean as to whether or not the character
6aff1f14
KW
486represented by that octet is (or on non-ASCII platforms, corresponds to) an
487ASCII character in the named class based on platform, Unicode, and Perl rules.
488If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 489
6aff1f14
KW
490Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
491with no suffix C<"_A">.
4b9734bf
KW
492
493Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
494onto the platform. That is, the code points that are ASCII are unaffected,
495since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
496as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
497true when called with the code point 0xDF, which is a word character in both
498ASCII and EBCDIC (though it represent different characters in each).
243effed
KW
499
500Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
501point as input. If the code point is larger than 255, Unicode rules are used
502to determine if it is in the character class. For example,
6aff1f14
KW
503C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
504WITH MACRON in Unicode, and is a word character.
243effed
KW
505
506Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
507(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
508classification of just the first (possibly multi-byte) character in the string
509is tested.
243effed
KW
510
511Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but uses
512the C library function that gives the named classification instead of
513hard-coded rules. For example, C<isDIGIT_LC()> returns the result of calling
514C<isdigit()>. This means that the result is based on the current locale, which
515is what C<LC> in the name stands for. FALSE is always returned if the input
516won't fit into an octet.
517
518Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
519returns the same as C<isFOO_LC> for input code points less than 256, and
520returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
521
522Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
523(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
524classification of just the first (possibly multi-byte) character in the string
525is tested.
ccfc67b7 526
954c1994 527=for apidoc Am|bool|isALPHA|char ch
2787a470 528Returns a boolean indicating whether the specified character is an
6aff1f14 529alphabetic character, analogous to C<m/[[:alpha:]]/>.
8a58bdcf 530See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 531C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 532C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 533
15861f94
KW
534=for apidoc Am|bool|isALPHANUMERIC|char ch
535Returns a boolean indicating whether the specified character is a either an
6aff1f14 536alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
15861f94
KW
537See the L<top of this section|/Character classes> for an explanation of variants
538C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 539C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
540C<isALPHANUMERIC_LC_utf8>.
541
8a58bdcf
KW
542=for apidoc Am|bool|isASCII|char ch
543Returns a boolean indicating whether the specified character is one of the 128
243effed 544characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 545On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
546character corresponds to an ASCII character. Variants C<isASCII_A()> and
547C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
548See the L<top of this section|/Character classes> for an explanation of variants
549C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
550C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
551library routine C<isascii()>. In these cases, the variants whose names contain
552C<LC> are the same as the corresponding ones without.
553
d98532ea
KW
554Also note, that because all ASCII characters are UTF-8 invariant (meaning they
555have the exact same representation (always a single byte) whether encoded in
556UTF-8 or not), C<isASCII> will give the correct results when called with any
557byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
558work properly on any string encoded or not in UTF-8.
559
243effed
KW
560=for apidoc Am|bool|isBLANK|char ch
561Returns a boolean indicating whether the specified character is a
6aff1f14 562character considered to be a blank, analogous to C<m/[[:blank:]]/>.
243effed 563See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 564C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
565C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
566platforms do not have the C library routine C<isblank()>. In these cases, the
567variants whose names contain C<LC> are the same as the corresponding ones
568without.
569
570=for apidoc Am|bool|isCNTRL|char ch
571Returns a boolean indicating whether the specified character is a
6aff1f14 572control character, analogous to C<m/[[:cntrl:]]/>.
243effed 573See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 574C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
575C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
576On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 577
954c1994 578=for apidoc Am|bool|isDIGIT|char ch
2787a470 579Returns a boolean indicating whether the specified character is a
6aff1f14 580digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 581Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed 582See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 583C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
584C<isDIGIT_LC_utf8>.
585
586=for apidoc Am|bool|isGRAPH|char ch
587Returns a boolean indicating whether the specified character is a
6aff1f14 588graphic character, analogous to C<m/[[:graph:]]/>.
243effed 589See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 590C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 591C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 592
0c82b6df 593=for apidoc Am|bool|isLOWER|char ch
2787a470 594Returns a boolean indicating whether the specified character is a
6aff1f14 595lowercase character, analogous to C<m/[[:lower:]]/>.
8a58bdcf 596See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 597C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 598C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 599
c99e91e9 600=for apidoc Am|bool|isOCTAL|char ch
2787a470 601Returns a boolean indicating whether the specified character is an
6aff1f14 602octal digit, [0-7].
243effed
KW
603The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
604C<isOCTAL>.
605
606=for apidoc Am|bool|isPUNCT|char ch
607Returns a boolean indicating whether the specified character is a
6aff1f14
KW
608punctuation character, analogous to C<m/[[:punct:]]/>.
609Note that the definition of what is punctuation isn't as
243effed
KW
610straightforward as one might desire. See L<perlrecharclass/POSIX Character
611Classes> for details.
612See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 613C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 614C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 615
0c82b6df 616=for apidoc Am|bool|isSPACE|char ch
2787a470 617Returns a boolean indicating whether the specified character is a
6aff1f14 618whitespace character. This is analogous
398d098a
KW
619to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
620(experimentally), this also matches what C<m/[[:space:]]/> does.
621("Experimentally" means that this change may be backed out in 5.20 or 5.22 if
622field experience indicates that it was unwise.) Prior to 5.18, only the
623locale forms of this macro (the ones with C<LC> in their names) matched
624precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
625in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
626(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
8a58bdcf 627See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 628C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 629C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 630
398d098a
KW
631=for apidoc Am|bool|isPSXSPC|char ch
632(short for Posix Space)
633Starting in 5.18, this is identical (experimentally) in all its forms to the
634corresponding C<isSPACE()> macros. ("Experimentally" means that this change
635may be backed out in 5.20 or 5.22 if field experience indicates that it
636was unwise.)
637The locale forms of this macro are identical to their corresponding
638C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
639non-locale forms differ from their C<isSPACE()> forms only in that the
640C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
641Otherwise they are identical. Thus this macro is analogous to what
642C<m/[[:space:]]/> matches in a regular expression.
643See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 644C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
645C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
646
954c1994 647=for apidoc Am|bool|isUPPER|char ch
2787a470 648Returns a boolean indicating whether the specified character is an
6aff1f14 649uppercase character, analogous to C<m/[[:upper:]]/>.
8a58bdcf 650See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 651C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 652C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 653
243effed 654=for apidoc Am|bool|isPRINT|char ch
8eea39dd 655Returns a boolean indicating whether the specified character is a
6aff1f14 656printable character, analogous to C<m/[[:print:]]/>.
8a58bdcf 657See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 658C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
659C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
660
661=for apidoc Am|bool|isWORDCHAR|char ch
662Returns a boolean indicating whether the specified character is a character
663that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
664in a regular expression. A word character is an alphabetic character, a
665decimal digit, a connecting punctuation character (such as an underscore), or
666a "mark" character that attaches to one of those (like some sort of accent).
667C<isALNUM()> is a synonym provided for backward compatibility, even though a
668word character includes more than the standard C language meaning of
669alphanumeric.
670See the L<top of this section|/Character classes> for an explanation of variants
671C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
672C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
673
674=for apidoc Am|bool|isXDIGIT|char ch
675Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
676digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
677and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
678See the L<top of this section|/Character classes> for an explanation of variants
679C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
680C<isXDIGIT_LC_utf8>.
681
3c3ecf18
KW
682=for apidoc Am|bool|isIDFIRST|char ch
683Returns a boolean indicating whether the specified character can be the first
684character of an identifier. This is very close to, but not quite the same as
685the official Unicode property C<XID_Start>. The difference is that this
686returns true only if the input character also matches L</isWORDCHAR>.
687See the L<top of this section|/Character classes> for an explanation of variants
688C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
689C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
690
691=for apidoc Am|bool|isIDCONT|char ch
692Returns a boolean indicating whether the specified character can be the
693second or succeeding character of an identifier. This is very close to, but
694not quite the same as the official Unicode property C<XID_Continue>. The
695difference is that this returns true only if the input character also matches
696L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
697explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
698C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
699C<isIDCONT_LC_utf8>.
700
243effed 701=head1 Miscellaneous Functions
8eea39dd 702
95a59cab 703=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 704Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
705Behaviour is only well defined when isXDIGIT(*str) is true.
706
e7c1e6c1
KW
707=head1 Character case changing
708
1f607577
KW
709=for apidoc Am|U8|toUPPER|U8 ch
710Converts the specified character to uppercase. If the input is anything but an
711ASCII lowercase character, that input character itself is returned. Variant
c753c8d3 712C<toUPPER_A> is equivalent.
954c1994 713
1f607577
KW
714=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
715Converts the Unicode code point C<cp> to its uppercase version, and
716stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
717that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
718bytes since the uppercase version may be longer than the original character.
719
720The first code point of the uppercased version is returned
721(but note, as explained just above, that there may be more.)
722
723=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
724Converts the UTF-8 encoded character at C<p> to its uppercase version, and
725stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
726that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
727bytes since the uppercase version may be longer than the original character.
728
729The first code point of the uppercased version is returned
730(but note, as explained just above, that there may be more.)
731
732The input character at C<p> is assumed to be well-formed.
733
25200305
KW
734=for apidoc Am|U8|toFOLD|U8 ch
735Converts the specified character to foldcase. If the input is anything but an
736ASCII uppercase character, that input character itself is returned. Variant
737C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
738Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
739
1f607577
KW
740=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
741Converts the Unicode code point C<cp> to its foldcase version, and
742stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
743that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
744bytes since the foldcase version may be longer than the original character.
745
746The first code point of the foldcased version is returned
747(but note, as explained just above, that there may be more.)
748
749=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
750Converts the UTF-8 encoded character at C<p> to its foldcase version, and
751stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
752that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
753bytes since the foldcase version may be longer than the original character.
754
755The first code point of the foldcased version is returned
756(but note, as explained just above, that there may be more.)
757
758The input character at C<p> is assumed to be well-formed.
759
760=for apidoc Am|U8|toLOWER|U8 ch
761Converts the specified character to lowercase. If the input is anything but an
762ASCII uppercase character, that input character itself is returned. Variant
c753c8d3 763C<toLOWER_A> is equivalent.
954c1994 764
1f607577
KW
765=for apidoc Am|U8|toLOWER_L1|U8 ch
766Converts the specified Latin1 character to lowercase. The results are undefined if
767the input doesn't fit in a byte.
768
769=for apidoc Am|U8|toLOWER_LC|U8 ch
770Converts the specified character to lowercase using the current locale's rules,
771if possible; otherwise returns the input character itself.
772
773=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
774Converts the Unicode code point C<cp> to its lowercase version, and
775stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
776that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
777bytes since the lowercase version may be longer than the original character.
778
779The first code point of the lowercased version is returned
780(but note, as explained just above, that there may be more.)
781
782=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
783Converts the UTF-8 encoded character at C<p> to its lowercase version, and
784stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
785that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
786bytes since the lowercase version may be longer than the original character.
787
788The first code point of the lowercased version is returned
789(but note, as explained just above, that there may be more.)
790
791The input character at C<p> is assumed to be well-formed.
792
793=for apidoc Am|U8|toLOWER_LC|U8 ch
794Converts the specified character to lowercase using the current locale's rules,
795if possible; otherwise returns the input character itself.
796
25200305
KW
797=for apidoc Am|U8|toTITLE|U8 ch
798Converts the specified character to titlecase. If the input is anything but an
799ASCII lowercase character, that input character itself is returned. Variant
800C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
801as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
802concept used in locale handling, so there is no functionality for that.)
803
1f607577
KW
804=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
805Converts the Unicode code point C<cp> to its titlecase version, and
806stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
807that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
808bytes since the titlecase version may be longer than the original character.
809
810The first code point of the titlecased version is returned
811(but note, as explained just above, that there may be more.)
812
813=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
814Converts the UTF-8 encoded character at C<p> to its titlecase version, and
815stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
816that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
817bytes since the titlecase version may be longer than the original character.
818
819The first code point of the titlecased version is returned
820(but note, as explained just above, that there may be more.)
821
822The input character at C<p> is assumed to be well-formed.
823
954c1994 824=cut
353c9b6f 825
1f607577 826XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
25200305 827really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
243effed 828
8a58bdcf 829Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
830patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
831
954c1994
GS
832*/
833
7c062697
KW
834/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
835 * is known only in the perl core, and this macro can be called from outside
836 * that */
837#ifdef HAS_QUAD
838# define WIDEST_UTYPE U64TYPE
839#else
840# define WIDEST_UTYPE U32
841#endif
842
3912bc88
KW
843/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
844 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
845 * bits of information are lost even on a 64-bit machine, but to get the
846 * compiler to optimize it out if possible. This is because Configure makes
847 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
848 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
849 */
850#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 851
41f43cc2 852#ifdef EBCDIC
b6340bd0
KW
853# ifndef _ALL_SOURCE
854 /* This returns the wrong results on at least z/OS unless this is
855 * defined. */
856# error _ALL_SOURCE should probably be defined
857# endif
858
859 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
860 * likely not defined, and so we use the native function */
861# define isASCII(c) isascii(c)
41f43cc2
KW
862#else
863# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
864#endif
865
ec79cfbb 866#define isASCII_A(c) isASCII(c)
3509bb53 867#define isASCII_L1(c) isASCII(c)
6515cb76 868
c2da0b36
KW
869/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
870 * and the 8 possible permutations of those bits exactly comprise the 8 octal
871 * digits */
3d33b3e3 872#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 873
ec79cfbb 874/* ASCII range only */
9fb1bf9d 875#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 876
a500dc72
KW
877/* Character class numbers. For internal core Perl use only. The ones less
878 * than 32 are used in PL_charclass[] and the ones up through the one that
879 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
880 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
881 * their actual definitions are here. If that file has a name not used here,
882 * it won't compile.
1709d539
KW
883 *
884 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
885 * frequency of their use. This gives a slight edge to exiting a loop earlier
886 * (in reginclass() in regexec.c) */
1709d539
KW
887# define _CC_WORDCHAR 0 /* \w and [:word:] */
888# define _CC_DIGIT 1 /* \d and [:digit:] */
889# define _CC_ALPHA 2 /* [:alpha:] */
890# define _CC_LOWER 3 /* [:lower:] */
891# define _CC_UPPER 4 /* [:upper:] */
892# define _CC_PUNCT 5 /* [:punct:] */
893# define _CC_PRINT 6 /* [:print:] */
15861f94 894# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 895# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 896# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 897
b0d691b2 898#define _FIRST_NON_SWASH_CC 10
1709d539
KW
899/* The character classes above are implemented with swashes. The second group
900 * (just below) contains the ones implemented without. These are also sorted
901 * in rough order of the frequency of their use, except that \v should be last,
902 * as it isn't a real Posix character class, and some (small) inefficiencies in
903 * regular expression handling would be introduced by putting it in the middle
904 * of those that are. Also, cntrl and ascii come after the others as it may be
905 * useful to group these which have no members that match above Latin1, (or
906 * above ASCII in the latter case) */
907
b0d691b2
KW
908# define _CC_SPACE 10 /* \s */
909# define _CC_BLANK 11 /* [:blank:] */
910# define _CC_XDIGIT 12 /* [:xdigit:] */
911# define _CC_PSXSPC 13 /* [:space:] */
912# define _CC_CNTRL 14 /* [:cntrl:] */
913# define _CC_ASCII 15 /* [:ascii:] */
914# define _CC_VERTSPACE 16 /* \v */
1709d539 915
a0947d7b
KW
916# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
917
1709d539
KW
918/* The members of the third group below do not need to be coordinated with data
919 * structures in regcomp.[ch] and regexec.c */
b0d691b2
KW
920# define _CC_IDFIRST 17
921# define _CC_CHARNAME_CONT 18
922# define _CC_NONLATIN1_FOLD 19
923# define _CC_QUOTEMETA 20
924# define _CC_NON_FINAL_FOLD 21
925# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
926# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
927/* Unused: 23-30
f4cdb42c
KW
928 * If more bits are needed, one could add a second word for non-64bit
929 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
930 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
931 * is used only for optimization (as of this writing), and differs in the
932 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
933 * characters: the masculine and feminine ordinal indicators, so removing it
934 * would just cause /i regexes which match them to run less efficiently */
96ac0975 935
3a371f2f
KW
936#if defined(PERL_CORE) || defined(PERL_EXT)
937/* An enum version of the character class numbers, to help compilers
938 * optimize */
939typedef enum {
3a371f2f 940 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
941 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
942 _CC_ENUM_ASCII = _CC_ASCII,
943 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 944 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 945 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
946 _CC_ENUM_DIGIT = _CC_DIGIT,
947 _CC_ENUM_GRAPH = _CC_GRAPH,
948 _CC_ENUM_LOWER = _CC_LOWER,
949 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 950 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 951 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 952 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 953 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 954 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 955 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 956 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
957} _char_class_number;
958#endif
959
63c61c3f 960#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 961#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 962
3018b823 963#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 964# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
965 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
966 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
967 #error Need to adjust order of swash_property_names[]
968# endif
969
b44daeeb
KW
970/* This is declared static in each of the few files that this is #defined for
971 * to keep them from being publicly accessible. Hence there is a small amount
972 * of wasted space */
973
922e8cb4
KW
974static const char* const swash_property_names[] = {
975 "XPosixWord",
976 "XPosixDigit",
977 "XPosixAlpha",
978 "XPosixLower",
979 "XPosixUpper",
980 "XPosixPunct",
981 "XPosixPrint",
982 "XPosixAlnum",
b0d691b2
KW
983 "XPosixGraph",
984 "Cased"
922e8cb4
KW
985};
986#endif
987
96ac0975
NC
988# ifdef DOINIT
989EXTCONST U32 PL_charclass[] = {
990# include "l1_char_class_tab.h"
991};
992
993# else /* ! DOINIT */
994EXTCONST U32 PL_charclass[];
995# endif
996
265c1f46 997 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70
KW
998# define _CC_mask(classnum) (1U << (classnum))
999# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
f4cd282c 1000 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
4eeeb416 1001
f4cdb42c
KW
1002 /* The mask for the _A versions of the macros; it just adds in the bit for
1003 * ASCII. */
1004# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1005
1006 /* The _A version makes sure that both the desired bit and the ASCII bit
1007 * are present */
1008# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
f4cd282c 1009 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
f4cdb42c
KW
1010 == _CC_mask_A(classnum)))
1011
f4cdb42c 1012# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 1013# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
1014# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1015# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1016# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
1017# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1018# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1019# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1020# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1021# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1022# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1023# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1024# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1025# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
1026# define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST))
1027
fb9bfbf7
KW
1028 /* Either participates in a fold with a character above 255, or is a
1029 * multi-char fold */
f4cd282c 1030# 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 1031
4eeeb416 1032# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
1033# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1034 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
1035# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1036 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
9fb1bf9d 1037#else /* No perl.h. */
efa6106c
KW
1038# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1039# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1040# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1041# define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
9fb1bf9d 1042# ifdef EBCDIC
b6340bd0 1043 /* We could be called without perl.h, so the native functions are the
4f48cecf
KW
1044 * easiest to code these in. They likely will return false for all
1045 * non-ASCII values, but this makes sure */
b6340bd0
KW
1046# define isALPHA_A(c) (isASCII(c) && isalpha(c))
1047# define isALPHANUMERIC_A(c) (isASCII(c) && isalnum(c))
1048# define isCNTRL_A(c) (isASCII(c) && iscntrl(c))
1049# define isDIGIT_A(c) (isASCII(c) && isdigit(c))
1050# define isGRAPH_A(c) (isASCII(c) && isgraph(c))
1051# define isLOWER_A(c) (isASCII(c) && islower(c))
1052# define isPRINT_A(c) (isASCII(c) && isprint(c))
1053# define isPUNCT_A(c) (isASCII(c) && ispunct(c))
1054# define isSPACE_A(c) (isASCII(c) && isspace(c))
1055# define isUPPER_A(c) (isASCII(c) && isupper(c))
1056# define isXDIGIT_A(c) (isASCII(c) && isxdigit(c))
9fb1bf9d 1057# else /* ASCII platform, no perl.h */
9fb1bf9d 1058# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
15861f94 1059# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
94824115 1060# define isCNTRL_A(c) (isASCII(c) && ! isPRINT_A(c))
2181da29 1061# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
94824115 1062# define isGRAPH_A(c) (isPRINT_A(c) && (c) ! = ' ')
9fb1bf9d
KW
1063# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1064# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
94824115 1065# define isPUNCT_A(c) (isGRAPH_A(c) && ! isALPHANUMERIC(c))
926bb54c
KW
1066# define isSPACE_A(c) ((c) == ' ' \
1067 || (c) == '\t' \
1068 || (c) == '\n' \
1069 || (c) =='\r' \
1070 || (c) == '\f')
2181da29 1071# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
926bb54c
KW
1072# define isXDIGIT_A(c) (isDIGIT_A(c) \
1073 || ((c) >= 'a' && (c) <= 'f') \
1074 || ((c) <= 'F' && (c) >= 'A'))
9fb1bf9d
KW
1075# endif
1076#endif /* ASCII range definitions */
ec79cfbb
KW
1077
1078/* Latin1 definitions */
8a58bdcf 1079#ifdef H_PERL
f4cdb42c 1080# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
15861f94 1081# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
f4cdb42c 1082# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
e5dcd934 1083
8a58bdcf 1084/* continuation character for legal NAME in \N{NAME} */
4eeeb416 1085# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
e5dcd934 1086
f4cdb42c
KW
1087# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1088# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1089# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1090# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1091# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1092# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1093# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1094# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1095# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1096# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
8a58bdcf
KW
1097#else /* No access to perl.h. Only a few provided here, just in case needed
1098 * for backwards compatibility */
1099 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
1100 * >= AA test to speed up ASCII-only tests at the expense of the others */
8e7c6e7d
KW
1101# define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_LATIN1((U8) c) >= 0xAA \
1102 && ((NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1103 && NATIVE_TO_LATIN1((U8) c) != 0xD7 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1104 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1105 || NATIVE_TO_LATIN1((U8) c) == 0xB5 \
1106 || NATIVE_TO_LATIN1((U8) c) == 0xBA)))
926bb54c
KW
1107# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1108 || (c) == ' ' \
1109 || (c) == '-' \
1110 || (c) == '(' \
1111 || (c) == ')' \
1112 || (c) == ':' \
8e7c6e7d 1113 || NATIVE_TO_LATIN1((U8) c) == 0xA0)
8a58bdcf
KW
1114#endif
1115
15861f94 1116# define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
271b5cb3 1117# define isALPHA(c) isALPHA_A(c)
ec79cfbb
KW
1118# define isBLANK(c) isBLANK_A(c)
1119# define isCNTRL(c) isCNTRL_A(c)
1120# define isDIGIT(c) isDIGIT_A(c)
1121# define isGRAPH(c) isGRAPH_A(c)
8a58bdcf 1122# define isIDFIRST(c) isIDFIRST_A(c)
ec79cfbb
KW
1123# define isLOWER(c) isLOWER_A(c)
1124# define isPRINT(c) isPRINT_A(c)
1125# define isPSXSPC(c) isPSXSPC_A(c)
1126# define isPUNCT(c) isPUNCT_A(c)
1127# define isSPACE(c) isSPACE_A(c)
1128# define isUPPER(c) isUPPER_A(c)
8a58bdcf 1129# define isWORDCHAR(c) isWORDCHAR_A(c)
ec79cfbb 1130# define isXDIGIT(c) isXDIGIT_A(c)
00f254e2 1131
8a58bdcf
KW
1132 /* ASCII casing. These could also be written as
1133 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1134 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1135 which uses table lookup and mask instead of subtraction. (This would
1136 work because the _MOD does not apply in the ASCII range) */
9d116dd7 1137# define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
7d79683e 1138# define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
bbce6d69 1139
25200305
KW
1140/* In the ASCII range, these are equivalent to what they're here defined to be.
1141 * But by creating these definitions, other code doesn't have to be aware of
1142 * this detail */
1143#define toFOLD(c) toLOWER(c)
1144#define toFOLD_LC(c) toLOWER_LC(c)
1145#define toTITLE(c) toUPPER(c)
1146
c753c8d3
KW
1147#define toLOWER_A(c) toLOWER(c)
1148#define toUPPER_A(c) toUPPER(c)
25200305
KW
1149#define toFOLD_A(c) toFOLD(c)
1150#define toTITLE_A(c) toTITLE(c)
1a0901db
KW
1151
1152/* Use table lookup for speed; return error character for input
1153 * out-of-range */
b2bf251f 1154#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1155 ? (c) \
f4cd282c 1156 : PL_latin1_lc[ (U8) (c) ])
c753c8d3
KW
1157#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1158
1a0901db
KW
1159/* Modified uc. Is correct uc except for three non-ascii chars which are
1160 * all mapped to one of them, and these need special handling; error
1161 * character for input out-of-range */
b2bf251f 1162#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1163 ? (c) \
f4cd282c 1164 : PL_mod_latin1_uc[ (U8) (c) ])
bbce6d69
PP
1165#ifdef USE_NEXT_CTYPE
1166
15861f94 1167# define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c))
ff68c719 1168# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 1169# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 1170# define isBLANK_LC(c) isBLANK((unsigned int)(c))
b8c5462f 1171# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
509fb054 1172# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
b8c5462f 1173# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
509fb054
KW
1174# define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
1175# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
ff68c719 1176# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 1177# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
509fb054
KW
1178# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
1179# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
1180# define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
e712593e 1181# define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
ff68c719 1182# define toLOWER_LC(c) NXToLower((unsigned int)(c))
509fb054 1183# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
bbce6d69
PP
1184
1185#else /* !USE_NEXT_CTYPE */
b8c5462f 1186
bbce6d69
PP
1187# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1188
76fde96a 1189/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
851131d3 1190
76fde96a 1191# define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
15861f94
KW
1192# define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \
1193 && isalnum((unsigned char)(c)))
07315176 1194# ifdef HAS_ISASCII
76fde96a 1195# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
07315176 1196# else
76fde96a 1197# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
07315176 1198# endif
67addccf 1199# ifdef HAS_ISBLANK
76fde96a 1200# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
67addccf 1201# else
76fde96a 1202# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
67addccf 1203# endif
76fde96a 1204# define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
509fb054 1205# define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
76fde96a 1206# define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
509fb054
KW
1207# define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1208 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1209# define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
76fde96a
KW
1210# define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1211# define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
509fb054
KW
1212# define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1213# define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1214# define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1215 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
e712593e 1216# define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
76fde96a 1217# define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? tolower((unsigned char)(c)) : (c))
509fb054 1218# define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? toupper((unsigned char)(c)) : (c))
bbce6d69
PP
1219
1220# else
1221
bbce6d69 1222# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1223# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
07315176 1224# define isASCII_LC(c) isascii(c)
67addccf
KW
1225# ifdef HAS_ISBLANK
1226# define isBLANK_LC(c) (isascii(c) && isblank(c))
1227# else
3e66ab74 1228# define isBLANK_LC(c) isBLANK_A(c)
67addccf 1229# endif
b8c5462f 1230# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1231# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1232# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1233# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1234# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1235# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1236# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1237# define isSPACE_LC(c) (isascii(c) && isspace(c))
1238# define isUPPER_LC(c) (isascii(c) && isupper(c))
1239# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1240# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
76fde96a 1241# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1242# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
bbce6d69
PP
1243
1244# endif
a0d0e21e 1245#endif /* USE_NEXT_CTYPE */
55204971 1246
eba68aa0
KW
1247#define isIDCONT(c) isWORDCHAR(c)
1248#define isIDCONT_A(c) isWORDCHAR_A(c)
1249#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1250#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1251#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1252
66c17564 1253/* For internal core Perl use only. If the input is Latin1, use the Latin1
a500dc72
KW
1254 * macro; otherwise use the function 'above_latin1'. Won't compile if 'c' isn't unsigned, as
1255 * won't match above_latin1 prototype. The macros do bounds checking, so have
66c17564
KW
1256 * duplicate checks here, so could create versions of the macros that don't,
1257 * but experiments show that gcc optimizes them out anyway. */
1258
1259/* Note that all ignore 'use bytes' */
cd500f2f
KW
1260#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1261 ? _generic_isCC(c, classnum) \
1262 : above_latin1(c))
1263#define _generic_swash_uni(classnum, c) ((c) < 256 \
1264 ? _generic_isCC(c, classnum) \
922e8cb4 1265 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1266#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1267#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1268#define isASCII_uni(c) isASCII(c)
1269#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1270#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1271#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1272#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1273#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1274#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1275#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1276#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1277
509fb054
KW
1278/* Posix and regular space are identical above Latin1 */
1279#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1280
cd500f2f 1281#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1282#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1283#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1284#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1285#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1286#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1287
b8d68ded 1288#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1289#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1290#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1291#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1292
cd500f2f
KW
1293#define _generic_LC_uvchr(latin1, above_latin1, c) \
1294 (c < 256 ? latin1(c) : above_latin1(c))
1295#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1296 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1297
1298#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1299#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1300 _CC_ALPHANUMERIC, c)
509fb054 1301#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1302#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1303#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1304#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1305#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1306#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1307 _is_uni_perl_idcont, c)
cd500f2f
KW
1308#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1309 _is_uni_perl_idstart, c)
1310#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1311#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1312#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1313 space under locale */
cd500f2f
KW
1314#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1315#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1316 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1317#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1318#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1319 _CC_WORDCHAR, c)
1320#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1321
a0ed51b3 1322
36b00f93 1323#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1324
a500dc72
KW
1325/* Everything whose name begins with an underscore is for internal core Perl
1326 * use only. */
1327
1328/* If the input is in the Latin1 range, use
252810d0
KW
1329 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1330 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1331 * fact that ASCII characters have the same representation whether utf8 or not.
1332 * Note that it assumes that the utf8 has been validated, and ignores 'use
1333 * bytes' */
cd500f2f 1334#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1335 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1336 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c 1337 ? _generic_isCC( \
94bb8c36 1338 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
aea0c08c 1339 *((p)+1 )), \
94bb8c36 1340 classnum) \
252810d0 1341 : utf8)
cd500f2f
KW
1342/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1343 * can be a macro */
1344#define _generic_func_utf8(classnum, above_latin1, p) \
1345 _generic_utf8(classnum, p, above_latin1(p))
1346/* Like the above, but passes classnum to _isFOO_utf8(), instead of having a
1347 * 'above_latin1' parameter */
922e8cb4 1348#define _generic_swash_utf8(classnum, p) \
cd500f2f 1349 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1350
cc8ab7c0
KW
1351/* Like the above, but should be used only when it is known that there are no
1352 * characters in the range 128-255 which the class is TRUE for. Hence it can
cd500f2f
KW
1353 * skip the tests for this range. 'above_latin1' should include its arguments */
1354#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1355 (UTF8_IS_INVARIANT(*(p)) \
1356 ? _generic_isCC(*(p), classnum) \
1357 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
cd500f2f 1358 ? above_latin1 \
cc8ab7c0
KW
1359 : 0)
1360
84238efa
KW
1361/* NOTE that some of these macros have very similar ones in regcharclass.h.
1362 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1363 * there, differing in name only by an underscore from the one here
1364 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1365 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1366 * points; the regcharclass.h ones are implemented as a series of
1367 * "if-else-if-else ..." */
1368
922e8cb4
KW
1369#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1370#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1371#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1372 utf8, the non-utf8 macro works
1373 */
cd500f2f
KW
1374#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1375#define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1376#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1377 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1378#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1379#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1380 _is_utf8_perl_idcont, p)
e5dcd934 1381
c11ff943
KW
1382/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1383 * IDFIRST is an alnum. See
c8362b00 1384 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1385 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1386 * This used to be not the XID version, but we decided to go with the more
1387 * modern Unicode definition */
cd500f2f 1388#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1389 _is_utf8_perl_idstart, p)
e5dcd934 1390
922e8cb4
KW
1391#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1392#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1393
509fb054 1394/* Posix and regular space are identical above Latin1 */
cd500f2f 1395#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1396
922e8cb4 1397#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1398#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1399#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1400#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1401#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1402#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1403 is_XDIGIT_high(p))
a0ed51b3 1404
7d2b61dd 1405#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
509fb054
KW
1406#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1407#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1408#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1409
9201a572
KW
1410/* For internal core Perl use only. If the input is in the Latin1 range, use
1411 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1412 * use the value given by the 'utf8' parameter. This relies on the fact that
1413 * ASCII characters have the same representation whether utf8 or not. Note
1414 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
94bb8c36
KW
1415#define _generic_LC_utf8(macro, p, utf8) \
1416 (UTF8_IS_INVARIANT(*(p)) \
1417 ? macro(*(p)) \
1418 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1419 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
9201a572
KW
1420 : utf8)
1421
cd500f2f
KW
1422#define _generic_LC_swash_utf8(macro, classnum, p) \
1423 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1424#define _generic_LC_func_utf8(macro, above_latin1, p) \
1425 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1426
cd500f2f
KW
1427#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1428 _CC_ALPHANUMERIC, p)
1429#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1430#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1431#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1432#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1433#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1434#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1435#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1436#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1437#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1438#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1439#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1440 space under locale */
cd500f2f
KW
1441#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1442#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1443#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1444#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1445 _CC_WORDCHAR, p)
1446#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1447
fbc19f27
KW
1448/* Macros for backwards compatibility and for completeness when the ASCII and
1449 * Latin1 values are identical */
fbc19f27
KW
1450#define isALPHAU(c) isALPHA_L1(c)
1451#define isDIGIT_L1(c) isDIGIT_A(c)
1452#define isOCTAL(c) isOCTAL_A(c)
1453#define isOCTAL_L1(c) isOCTAL_A(c)
1454#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1455#define isALNUM(c) isWORDCHAR(c)
1456#define isALNUMU(c) isWORDCHAR_L1(c)
1457#define isALNUM_LC(c) isWORDCHAR_LC(c)
1458#define isALNUM_uni(c) isWORDCHAR_uni(c)
1459#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1460#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1461#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1462#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1463#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1464#define isALNUMC(c) isALPHANUMERIC(c)
1465#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1466#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1467#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1468#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1469#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1470
6ae9f32a 1471/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1472 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1473 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1474 * that is EOT on ASCII; ST on EBCDIC */
8e7c6e7d 1475# define toCTRL(c) (toUPPER(NATIVE_TO_LATIN1(c)) ^ 64)
bbce6d69 1476
dea28490
JJ
1477/* Line numbers are unsigned, 32 bits. */
1478typedef U32 line_t;
e5dcd934 1479#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1480
91152fc1
DG
1481/* Helpful alias for version prescan */
1482#define is_LAX_VERSION(a,b) \
1483 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1484
1485#define is_STRICT_VERSION(a,b) \
1486 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1487
1488#define BADVERSION(a,b,c) \
1489 if (b) { \
1490 *b = c; \
1491 } \
1492 return a;
8c52afec 1493
95a59cab
YO
1494#define READ_XDIGIT(s) (isALPHA(*(s)) ? ((*(s)++ + 9) & 0xf) : (*(s)++ & 0xf))
1495
8e84507e 1496/*
ccfc67b7
JH
1497=head1 Memory Management
1498
a02a5408 1499=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1500The XSUB-writer's interface to the C C<malloc> function.
1501
c5008215
JC
1502In 5.9.3, Newx() and friends replace the older New() API, and drops
1503the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1504themselves. This aid has been superseded by a new build option,
d10b4965 1505PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1506there for use in XS modules supporting older perls.
1507
a02a5408 1508=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1509The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1510cast. See also C<Newx>.
954c1994 1511
a02a5408 1512=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1513The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1514memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1515
954c1994
GS
1516=for apidoc Am|void|Renew|void* ptr|int nitems|type
1517The XSUB-writer's interface to the C C<realloc> function.
1518
1519=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1520The XSUB-writer's interface to the C C<realloc> function, with
1521cast.
1522
49b8b560 1523=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1524The XSUB-writer's interface to the C C<free> function.
1525
1526=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1527The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1528source, C<dest> is the destination, C<nitems> is the number of items, and
1529C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1530
e90e2364
NC
1531=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1532Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1533optimise.
1534
954c1994
GS
1535=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1536The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1537source, C<dest> is the destination, C<nitems> is the number of items, and
1538C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1539
e90e2364
NC
1540=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1541
1542Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1543optimise.
1544
954c1994
GS
1545=for apidoc Am|void|Zero|void* dest|int nitems|type
1546
1547The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1548destination, C<nitems> is the number of items, and C<type> is the type.
1549
e90e2364
NC
1550=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1551
1552Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1553optimise.
1554
da5d8dbb 1555=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1556This is an architecture-independent macro to copy one structure to another.
954c1994 1557
7e337ee0
JH
1558=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1559
1560Fill up memory with a byte pattern (a byte repeated over and over
1561again) that hopefully catches attempts to access uninitialized memory.
1562
1563=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1564
1565PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1566
1c12ffb4 1567=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1568
1569PoisonWith(0xEF) for catching access to freed memory.
1570
9965345d
JH
1571=for apidoc Am|void|Poison|void* dest|int nitems|type
1572
7e337ee0 1573PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1574
1575=cut */
954c1994 1576
561b68a9
SH
1577/* Maintained for backwards-compatibility only. Use newSV() instead. */
1578#ifndef PERL_CORE
ff06c60c 1579#define NEWSV(x,len) newSV(len)
561b68a9 1580#endif
ff06c60c 1581
19a94d75
JH
1582#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1583
2720f793
JH
1584/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1585 * overly eager compilers that will bleat about e.g.
1586 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1587#ifdef PERL_MALLOC_WRAP
f019c49e 1588#define MEM_WRAP_CHECK(n,t) \
d1decf2b 1589 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
27d5b266 1590#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1591 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1592#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1593
d1decf2b 1594#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
1595#else
1596
410319be
NC
1597#define MEM_WRAP_CHECK(n,t)
1598#define MEM_WRAP_CHECK_1(n,t,a)
1599#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1600#define MEM_WRAP_CHECK_(n,t)
1601
6fff79ce 1602#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1603
1936d2a7 1604#endif
8b44ba4c 1605
fe4f188c 1606#ifdef PERL_MEM_LOG
46c6c7e2 1607/*
9f653bb5 1608 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1609 * go through functions, which are handy for debugging breakpoints, but
1610 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1611 * line number, and C function name if available) passed in. This info can
1612 * then be used for logging the calls, for which one gets a sample
73d1d973 1613 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1614 *
46c6c7e2 1615 * Known problems:
94e892a6 1616 * - not all memory allocs get logged, only those
46c6c7e2 1617 * that go through Newx() and derivatives (while all
94e892a6 1618 * Safefrees do get logged)
46c6c7e2
JH
1619 * - __FILE__ and __LINE__ do not work everywhere
1620 * - __func__ or __FUNCTION__ even less so
1621 * - I think more goes on after the perlio frees but
1622 * the thing is that STDERR gets closed (as do all
1623 * the file descriptors)
1624 * - no deeper calling stack than the caller of the Newx()
1625 * or the kind, but do I look like a C reflection/introspection
1626 * utility to you?
1627 * - the function prototypes for the logging functions
1628 * probably should maybe be somewhere else than handy.h
1629 * - one could consider inlining (macrofying) the logging
1630 * for speed, but I am too lazy
1631 * - one could imagine recording the allocations in a hash,
1632 * (keyed by the allocation address?), and maintain that
1633 * through reallocs and frees, but how to do that without
1634 * any News() happening...?
73d1d973 1635 * - lots of -Ddefines to get useful/controllable output
b953482e 1636 * - lots of ENV reads
46c6c7e2
JH
1637 */
1638
12754f92 1639PERL_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 1640
12754f92 1641PERL_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 1642
12754f92 1643PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1644
0b0ab801 1645# ifdef PERL_CORE
73d1d973 1646# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1647enum mem_log_type {
1648 MLT_ALLOC,
1649 MLT_REALLOC,
d7a2c63c
MHM
1650 MLT_FREE,
1651 MLT_NEW_SV,
1652 MLT_DEL_SV
0b0ab801
MHM
1653};
1654# endif
12754f92 1655# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1656void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1657void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1658# endif
0b0ab801
MHM
1659# endif
1660
fe4f188c
JH
1661#endif
1662
1663#ifdef PERL_MEM_LOG
d1401ee9
MHM
1664#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1665#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1666#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1667#endif
1668
1669#ifndef MEM_LOG_ALLOC
1670#define MEM_LOG_ALLOC(n,t,a) (a)
1671#endif
1672#ifndef MEM_LOG_REALLOC
1673#define MEM_LOG_REALLOC(n,t,v,a) (a)
1674#endif
1675#ifndef MEM_LOG_FREE
1676#define MEM_LOG_FREE(a) (a)
1677#endif
1678
d1401ee9
MHM
1679#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1680#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1681#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1682
1683#ifndef PERL_CORE
a02a5408
JC
1684/* pre 5.9.x compatibility */
1685#define New(x,v,n,t) Newx(v,n,t)
1686#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1687#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1688#endif
a02a5408 1689
ff68c719 1690#define Renew(v,n,t) \
d1401ee9 1691 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1692#define Renewc(v,n,t,c) \
d1401ee9 1693 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1694
1695#ifdef PERL_POISON
1696#define Safefree(d) \
06c0cc96 1697 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1698#else
fe4f188c 1699#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1700#endif
55497cff 1701
8b44ba4c
NC
1702#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1703#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1704#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1705
8b44ba4c
NC
1706#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1707#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1708#ifdef HAS_MEMSET
8b44ba4c 1709#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1710#else
8b44ba4c
NC
1711/* Using bzero(), which returns void. */
1712#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1713#endif
1714
7e337ee0
JH
1715#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1716#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1717#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1718#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1719
2304df62 1720#ifdef USE_STRUCT_COPY
ff68c719 1721#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1722#else
1723#define StructCopy(s,d,t) Copy(s,d,1,t)
1724#endif
2cc61e15 1725
622913ab 1726#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1727#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1728
2cc61e15
DD
1729#ifdef NEED_VA_COPY
1730# ifdef va_copy
1731# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1732# else
a1866d1b
JH
1733# if defined(__va_copy)
1734# define Perl_va_copy(s, d) __va_copy(d, s)
1735# else
1736# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1737# endif
2cc61e15
DD
1738# endif
1739#endif
1740
472d47bc
SB
1741/* convenience debug macros */
1742#ifdef USE_ITHREADS
1743#define pTHX_FORMAT "Perl interpreter: 0x%p"
1744#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1745#define pTHX_VALUE_ (void *)my_perl,
1746#define pTHX_VALUE (void *)my_perl
1747#define pTHX__VALUE_ ,(void *)my_perl,
1748#define pTHX__VALUE ,(void *)my_perl
472d47bc 1749#else
3609ea0d 1750#define pTHX_FORMAT
472d47bc 1751#define pTHX__FORMAT
3609ea0d 1752#define pTHX_VALUE_
472d47bc 1753#define pTHX_VALUE
3609ea0d 1754#define pTHX__VALUE_
472d47bc
SB
1755#define pTHX__VALUE
1756#endif /* USE_ITHREADS */
3609ea0d 1757
2acdbac1
NC
1758/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1759 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1760 CPAN::Unpack show any users outside the core. */
1761#ifdef PERL_CORE
d1d15184 1762# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1763#endif
1764
dfff4baf
BF
1765/* Internal macros to deal with gids and uids */
1766#ifdef PERL_CORE
1767
1768# if Uid_t_size > IVSIZE
1769# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
1770# define SvUID(sv) SvNV(sv)
1771# else
1772# if Uid_t_sign <= 0
1773# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
1774# define SvUID(sv) SvIV(sv)
1775# else
1776# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
1777# define SvUID(sv) SvUV(sv)
1778# endif
1779# endif /* Uid_t_size */
1780
1781# if Gid_t_size > IVSIZE
1782# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
1783# define SvGID(sv) SvNV(sv)
1784# else
1785# if Gid_t_sign <= 0
1786# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
1787# define SvGID(sv) SvIV(sv)
1788# else
1789# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
1790# define SvGID(sv) SvUV(sv)
1791# endif
1792# endif /* Gid_t_size */
1793
1794#endif
1795
9d745869
KW
1796#endif /* HANDY_H */
1797
e9a8c099
MHM
1798/*
1799 * Local variables:
1800 * c-indentation-style: bsd
1801 * c-basic-offset: 4
14d04a33 1802 * indent-tabs-mode: nil
e9a8c099
MHM
1803 * End:
1804 *
14d04a33 1805 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1806 */