This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perlre: Expand and clarify /x and (?# comment)
[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
70d5cb32 107# define bool char
c1d22f6b 108# define HAS_BOOL 1
a687059c 109#endif
0d3e774c 110
236415e2
KW
111/* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
112 * defined as char for example, then the cast from int is
113 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
114 * AIX */
18f5643b 115#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 116
46c6c7e2 117/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
118 * XXX Should really be a Configure probe, with HAS__FUNCTION__
119 * and FUNCTION__ as results.
120 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
121#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
122# define FUNCTION__ __func__
123#else
9ea40801 124# if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 125 (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
126# define FUNCTION__ ""
127# else
128# define FUNCTION__ __FUNCTION__ /* Common extension. */
129# endif
130#endif
131
27d4fb96
PP
132/* XXX A note on the perl source internal type system. The
133 original intent was that I32 be *exactly* 32 bits.
134
135 Currently, we only guarantee that I32 is *at least* 32 bits.
136 Specifically, if int is 64 bits, then so is I32. (This is the case
137 for the Cray.) This has the advantage of meshing nicely with
138 standard library calls (where we pass an I32 and the library is
139 expecting an int), but the disadvantage that an I32 is not 32 bits.
140 Andy Dougherty August 1996
24fef2a7 141
dc45a647
MB
142 There is no guarantee that there is *any* integral type with
143 exactly 32 bits. It is perfectly legal for a system to have
144 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 145
dc45a647
MB
146 Similarly, there is no guarantee that I16 and U16 have exactly 16
147 bits.
693762b4 148
8e84507e
NIS
149 For dealing with issues that may arise from various 32/64-bit
150 systems, we will ask Configure to check out
8175356b 151
3609ea0d
JH
152 SHORTSIZE == sizeof(short)
153 INTSIZE == sizeof(int)
154 LONGSIZE == sizeof(long)
dc45a647 155 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 156 PTRSIZE == sizeof(void *)
dc45a647
MB
157 DOUBLESIZE == sizeof(double)
158 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 159
27d4fb96
PP
160*/
161
69512466
JH
162#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
163# include <inttypes.h>
dd0eed91
JH
164# ifdef INT32_MIN_BROKEN
165# undef INT32_MIN
166# define INT32_MIN (-2147483647-1)
167# endif
168# ifdef INT64_MIN_BROKEN
169# undef INT64_MIN
170# define INT64_MIN (-9223372036854775807LL-1)
171# endif
69512466
JH
172#endif
173
8175356b
JH
174typedef I8TYPE I8;
175typedef U8TYPE U8;
176typedef I16TYPE I16;
177typedef U16TYPE U16;
178typedef I32TYPE I32;
179typedef U32TYPE U32;
6b8eaf93
JH
180#ifdef PERL_CORE
181# ifdef HAS_QUAD
8175356b
JH
182typedef I64TYPE I64;
183typedef U64TYPE U64;
6b8eaf93
JH
184# endif
185#endif /* PERL_CORE */
8175356b 186
69512466 187#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
80524f33
JL
188# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
189# define PeRl_INT64_C(c) CAT2(c,LL)
190# define PeRl_UINT64_C(c) CAT2(c,ULL)
191# else
52865553
SH
192# if QUADKIND == QUAD_IS___INT64
193# define PeRl_INT64_C(c) CAT2(c,I64)
194# define PeRl_UINT64_C(c) CAT2(c,UI64)
69512466 195# else
52865553
SH
196# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
197# define PeRl_INT64_C(c) CAT2(c,L)
198# define PeRl_UINT64_C(c) CAT2(c,UL)
69512466 199# else
80524f33
JL
200# define PeRl_INT64_C(c) ((I64TYPE)(c))
201# define PeRl_UINT64_C(c) ((U64TYPE)(c))
69512466
JH
202# endif
203# endif
e8c95190 204# endif
80524f33
JL
205# ifndef UINT64_C
206# define UINT64_C(c) PeRl_UINT64_C(c)
207# endif
208# ifndef INT64_C
209# define INT64_C(c) PeRl_INT64_C(c)
210# endif
e8c95190 211#endif
dc750f2e 212
d8668976 213#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 214
5ff3f7a4
GS
215/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
216 Please search CHAR_MAX in perl.h for further details. */
217#define U8_MAX UINT8_MAX
218#define U8_MIN UINT8_MIN
219
5ff3f7a4
GS
220#define I16_MAX INT16_MAX
221#define I16_MIN INT16_MIN
222#define U16_MAX UINT16_MAX
223#define U16_MIN UINT16_MIN
224
5ff3f7a4
GS
225#define I32_MAX INT32_MAX
226#define I32_MIN INT32_MIN
0e983133
GS
227#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
228# define U32_MAX UINT32_MAX
229#else
230# define U32_MAX 4294967295U
231#endif
5ff3f7a4
GS
232#define U32_MIN UINT32_MIN
233
234#else
235
5c9fa16e
KA
236/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
237 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
238#define U8_MAX PERL_UCHAR_MAX
239#define U8_MIN PERL_UCHAR_MIN
79072805 240
27d4fb96
PP
241#define I16_MAX PERL_SHORT_MAX
242#define I16_MIN PERL_SHORT_MIN
243#define U16_MAX PERL_USHORT_MAX
244#define U16_MIN PERL_USHORT_MIN
79072805 245
c4f23d77 246#if LONGSIZE > 4
27d4fb96
PP
247# define I32_MAX PERL_INT_MAX
248# define I32_MIN PERL_INT_MIN
249# define U32_MAX PERL_UINT_MAX
250# define U32_MIN PERL_UINT_MIN
79072805 251#else
27d4fb96
PP
252# define I32_MAX PERL_LONG_MAX
253# define I32_MIN PERL_LONG_MIN
254# define U32_MAX PERL_ULONG_MAX
255# define U32_MIN PERL_ULONG_MIN
79072805
LW
256#endif
257
5ff3f7a4
GS
258#endif
259
58a9a5d5 260/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
261#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
262#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
263#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
264
ff68c719 265#define Ctl(ch) ((ch) & 037)
8d063cd8 266
3fe05580
MHM
267/*
268=head1 SV-Body Allocation
269
270=for apidoc Ama|SV*|newSVpvs|const char* s
271Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
272
84bafc02
NC
273=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
274Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
275pair.
276
3fe05580
MHM
277=for apidoc Ama|SV*|newSVpvs_share|const char* s
278Like C<newSVpvn_share>, but takes a literal string instead of a string/length
279pair and omits the hash parameter.
280
9dcc53ea
Z
281=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
282Like C<sv_catpvn_flags>, but takes a literal string instead of a
283string/length pair.
284
285=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
286Like C<sv_catpvn_nomg>, but takes a literal string instead of a
287string/length pair.
288
7b4e7469 289=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
290Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
291
9dcc53ea
Z
292=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
293Like C<sv_catpvn_mg>, but takes a literal string instead of a
294string/length pair.
295
7b4e7469 296=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
297Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
298
9dcc53ea
Z
299=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
300Like C<sv_setpvn_mg>, but takes a literal string instead of a
301string/length pair.
302
303=for apidoc Am|SV *|sv_setref_pvs|const char* s
304Like C<sv_setref_pvn>, but takes a literal string instead of a
305string/length pair.
306
3fe05580
MHM
307=head1 Memory Management
308
309=for apidoc Ama|char*|savepvs|const char* s
310Like C<savepvn>, but takes a literal string instead of a string/length pair.
311
9dcc53ea
Z
312=for apidoc Ama|char*|savesharedpvs|const char* s
313A version of C<savepvs()> which allocates the duplicate string in memory
314which is shared between threads.
315
3fe05580
MHM
316=head1 GV Functions
317
318=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
319Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
320
321=head1 Hash Manipulation Functions
322
323=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
324Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
325
326=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
327Like C<hv_store>, but takes a literal string instead of a string/length pair
328and omits the hash parameter.
329
510966aa
Z
330=head1 Lexer interface
331
332=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
333
334Like L</lex_stuff_pvn>, but takes a literal string instead of a
335string/length pair.
336
3fe05580
MHM
337=cut
338*/
339
926bb54c
KW
340/* concatenating with "" ensures that only literal strings are accepted as
341 * argument */
26bfd924 342#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 343
926bb54c
KW
344/* note that STR_WITH_LEN() can't be used as argument to macros or functions
345 * that under some configurations might be macros, which means that it requires
346 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
347 */
348
349/* STR_WITH_LEN() shortcuts */
350#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
351#define newSVpvs_flags(str,flags) \
352 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 353#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
354#define sv_catpvs_flags(sv, str, flags) \
355 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
356#define sv_catpvs_nomg(sv, str) \
357 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
358#define sv_catpvs(sv, str) \
359 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
360#define sv_catpvs_mg(sv, str) \
361 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 362#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
363#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
364#define sv_setref_pvs(rv, classname, str) \
365 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 366#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
367#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
368#define gv_stashpvs(str, create) \
369 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
370#define gv_fetchpvs(namebeg, add, sv_type) \
371 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
372#define gv_fetchpvn(namebeg, len, add, sv_type) \
373 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
374#define sv_catxmlpvs(dsv, str, utf8) \
375 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
376#define hv_fetchs(hv,key,lval) \
377 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
378 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
379 : HV_FETCH_JUST_SV, NULL, 0))
380
381#define hv_stores(hv,key,val) \
382 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
383 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 384
510966aa
Z
385#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
386
b96d8cd9
NC
387#define get_cvs(str, flags) \
388 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 389
954c1994 390/*
ccfc67b7
JH
391=head1 Miscellaneous Functions
392
954c1994
GS
393=for apidoc Am|bool|strNE|char* s1|char* s2
394Test two strings to see if they are different. Returns true or
395false.
396
397=for apidoc Am|bool|strEQ|char* s1|char* s2
398Test two strings to see if they are equal. Returns true or false.
399
400=for apidoc Am|bool|strLT|char* s1|char* s2
401Test two strings to see if the first, C<s1>, is less than the second,
402C<s2>. Returns true or false.
403
404=for apidoc Am|bool|strLE|char* s1|char* s2
405Test two strings to see if the first, C<s1>, is less than or equal to the
406second, C<s2>. Returns true or false.
407
408=for apidoc Am|bool|strGT|char* s1|char* s2
409Test two strings to see if the first, C<s1>, is greater than the second,
410C<s2>. Returns true or false.
411
412=for apidoc Am|bool|strGE|char* s1|char* s2
413Test two strings to see if the first, C<s1>, is greater than or equal to
414the second, C<s2>. Returns true or false.
415
416=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
417Test two strings to see if they are different. The C<len> parameter
418indicates the number of bytes to compare. Returns true or false. (A
419wrapper for C<strncmp>).
420
421=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
422Test two strings to see if they are equal. The C<len> parameter indicates
423the number of bytes to compare. Returns true or false. (A wrapper for
424C<strncmp>).
425
426=cut
427*/
428
8d063cd8
LW
429#define strNE(s1,s2) (strcmp(s1,s2))
430#define strEQ(s1,s2) (!strcmp(s1,s2))
431#define strLT(s1,s2) (strcmp(s1,s2) < 0)
432#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
433#define strGT(s1,s2) (strcmp(s1,s2) > 0)
434#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
435#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
436#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 437
36477c24
PP
438#ifdef HAS_MEMCMP
439# define memNE(s1,s2,l) (memcmp(s1,s2,l))
440# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
441#else
442# define memNE(s1,s2,l) (bcmp(s1,s2,l))
443# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
444#endif
445
568a785a 446#define memEQs(s1, l, s2) \
8e6497b6 447 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
448#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
449
bbce6d69
PP
450/*
451 * Character classes.
452 *
453 * Unfortunately, the introduction of locales means that we
454 * can't trust isupper(), etc. to tell the truth. And when
455 * it comes to /\w+/ with tainting enabled, we *must* be able
456 * to trust our character classes.
457 *
458 * Therefore, the default tests in the text of Perl will be
459 * independent of locale. Any code that wants to depend on
460 * the current locale will use the tests that begin with "lc".
461 */
462
2304df62
AD
463#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
464# ifndef CTYPE256
465# define CTYPE256
466# endif
467#endif
468
954c1994 469/*
ccfc67b7
JH
470
471=head1 Character classes
243effed
KW
472This section is about functions (really macros) that classify characters
473into types, such as punctuation versus alphabetic, etc. Most of these are
474analogous to regular expression character classes. (See
475L<perlrecharclass/POSIX Character Classes>.) There are several variants for
476each class. (Not all macros have all variants; each item below lists the
477ones valid for it.) None are affected by C<use bytes>, and only the ones
478with C<LC> in the name are affected by the current locale.
479
480The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
481C<U8>) as input and returns a boolean as to whether or not the character
6aff1f14
KW
482represented by that octet is (or on non-ASCII platforms, corresponds to) an
483ASCII character in the named class based on platform, Unicode, and Perl rules.
484If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 485
6aff1f14
KW
486Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
487with no suffix C<"_A">.
4b9734bf
KW
488
489Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
490onto the platform. That is, the code points that are ASCII are unaffected,
491since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
492as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
493true when called with the code point 0xDF, which is a word character in both
494ASCII and EBCDIC (though it represent different characters in each).
243effed
KW
495
496Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
497point as input. If the code point is larger than 255, Unicode rules are used
498to determine if it is in the character class. For example,
6aff1f14
KW
499C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
500WITH MACRON in Unicode, and is a word character.
243effed
KW
501
502Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
503(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
504classification of just the first (possibly multi-byte) character in the string
505is tested.
243effed
KW
506
507Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but uses
508the C library function that gives the named classification instead of
509hard-coded rules. For example, C<isDIGIT_LC()> returns the result of calling
510C<isdigit()>. This means that the result is based on the current locale, which
511is what C<LC> in the name stands for. FALSE is always returned if the input
512won't fit into an octet.
513
514Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
515returns the same as C<isFOO_LC> for input code points less than 256, and
516returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
517
518Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
519(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
520classification of just the first (possibly multi-byte) character in the string
521is tested.
ccfc67b7 522
954c1994 523=for apidoc Am|bool|isALPHA|char ch
2787a470 524Returns a boolean indicating whether the specified character is an
6aff1f14 525alphabetic character, analogous to C<m/[[:alpha:]]/>.
8a58bdcf 526See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 527C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 528C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 529
15861f94
KW
530=for apidoc Am|bool|isALPHANUMERIC|char ch
531Returns a boolean indicating whether the specified character is a either an
6aff1f14 532alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
15861f94
KW
533See the L<top of this section|/Character classes> for an explanation of variants
534C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 535C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
536C<isALPHANUMERIC_LC_utf8>.
537
8a58bdcf
KW
538=for apidoc Am|bool|isASCII|char ch
539Returns a boolean indicating whether the specified character is one of the 128
243effed 540characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 541On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
542character corresponds to an ASCII character. Variants C<isASCII_A()> and
543C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
544See the L<top of this section|/Character classes> for an explanation of variants
545C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
546C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
547library routine C<isascii()>. In these cases, the variants whose names contain
548C<LC> are the same as the corresponding ones without.
549
d98532ea
KW
550Also note, that because all ASCII characters are UTF-8 invariant (meaning they
551have the exact same representation (always a single byte) whether encoded in
552UTF-8 or not), C<isASCII> will give the correct results when called with any
553byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
554work properly on any string encoded or not in UTF-8.
555
243effed
KW
556=for apidoc Am|bool|isBLANK|char ch
557Returns a boolean indicating whether the specified character is a
6aff1f14 558character considered to be a blank, analogous to C<m/[[:blank:]]/>.
243effed 559See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 560C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
561C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
562platforms do not have the C library routine C<isblank()>. In these cases, the
563variants whose names contain C<LC> are the same as the corresponding ones
564without.
565
566=for apidoc Am|bool|isCNTRL|char ch
567Returns a boolean indicating whether the specified character is a
6aff1f14 568control character, analogous to C<m/[[:cntrl:]]/>.
243effed 569See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 570C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
571C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
572On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 573
954c1994 574=for apidoc Am|bool|isDIGIT|char ch
2787a470 575Returns a boolean indicating whether the specified character is a
6aff1f14 576digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 577Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed 578See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 579C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
580C<isDIGIT_LC_utf8>.
581
582=for apidoc Am|bool|isGRAPH|char ch
583Returns a boolean indicating whether the specified character is a
6aff1f14 584graphic character, analogous to C<m/[[:graph:]]/>.
243effed 585See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 586C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 587C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 588
0c82b6df 589=for apidoc Am|bool|isLOWER|char ch
2787a470 590Returns a boolean indicating whether the specified character is a
6aff1f14 591lowercase character, analogous to C<m/[[:lower:]]/>.
8a58bdcf 592See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 593C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 594C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 595
c99e91e9 596=for apidoc Am|bool|isOCTAL|char ch
2787a470 597Returns a boolean indicating whether the specified character is an
6aff1f14 598octal digit, [0-7].
243effed
KW
599The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
600C<isOCTAL>.
601
602=for apidoc Am|bool|isPUNCT|char ch
603Returns a boolean indicating whether the specified character is a
6aff1f14
KW
604punctuation character, analogous to C<m/[[:punct:]]/>.
605Note that the definition of what is punctuation isn't as
243effed
KW
606straightforward as one might desire. See L<perlrecharclass/POSIX Character
607Classes> for details.
608See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 609C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 610C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 611
0c82b6df 612=for apidoc Am|bool|isSPACE|char ch
2787a470 613Returns a boolean indicating whether the specified character is a
6aff1f14 614whitespace character. This is analogous
398d098a
KW
615to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
616(experimentally), this also matches what C<m/[[:space:]]/> does.
617("Experimentally" means that this change may be backed out in 5.20 or 5.22 if
618field experience indicates that it was unwise.) Prior to 5.18, only the
619locale forms of this macro (the ones with C<LC> in their names) matched
620precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
621in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
622(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
8a58bdcf 623See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 624C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 625C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 626
398d098a
KW
627=for apidoc Am|bool|isPSXSPC|char ch
628(short for Posix Space)
629Starting in 5.18, this is identical (experimentally) in all its forms to the
630corresponding C<isSPACE()> macros. ("Experimentally" means that this change
631may be backed out in 5.20 or 5.22 if field experience indicates that it
632was unwise.)
633The locale forms of this macro are identical to their corresponding
634C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
635non-locale forms differ from their C<isSPACE()> forms only in that the
636C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
637Otherwise they are identical. Thus this macro is analogous to what
638C<m/[[:space:]]/> matches in a regular expression.
639See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 640C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
641C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
642
954c1994 643=for apidoc Am|bool|isUPPER|char ch
2787a470 644Returns a boolean indicating whether the specified character is an
6aff1f14 645uppercase character, analogous to C<m/[[:upper:]]/>.
8a58bdcf 646See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 647C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 648C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 649
243effed 650=for apidoc Am|bool|isPRINT|char ch
8eea39dd 651Returns a boolean indicating whether the specified character is a
6aff1f14 652printable character, analogous to C<m/[[:print:]]/>.
8a58bdcf 653See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 654C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
655C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
656
657=for apidoc Am|bool|isWORDCHAR|char ch
658Returns a boolean indicating whether the specified character is a character
659that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
660in a regular expression. A word character is an alphabetic character, a
661decimal digit, a connecting punctuation character (such as an underscore), or
662a "mark" character that attaches to one of those (like some sort of accent).
663C<isALNUM()> is a synonym provided for backward compatibility, even though a
664word character includes more than the standard C language meaning of
665alphanumeric.
666See the L<top of this section|/Character classes> for an explanation of variants
667C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
668C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
669
670=for apidoc Am|bool|isXDIGIT|char ch
671Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
672digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
673and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
674See the L<top of this section|/Character classes> for an explanation of variants
675C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
676C<isXDIGIT_LC_utf8>.
677
3c3ecf18
KW
678=for apidoc Am|bool|isIDFIRST|char ch
679Returns a boolean indicating whether the specified character can be the first
680character of an identifier. This is very close to, but not quite the same as
681the official Unicode property C<XID_Start>. The difference is that this
682returns true only if the input character also matches L</isWORDCHAR>.
683See the L<top of this section|/Character classes> for an explanation of variants
684C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
685C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
686
687=for apidoc Am|bool|isIDCONT|char ch
688Returns a boolean indicating whether the specified character can be the
689second or succeeding character of an identifier. This is very close to, but
690not quite the same as the official Unicode property C<XID_Continue>. The
691difference is that this returns true only if the input character also matches
692L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
693explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
694C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
695C<isIDCONT_LC_utf8>.
696
243effed 697=head1 Miscellaneous Functions
8eea39dd 698
95a59cab 699=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 700Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
701Behaviour is only well defined when isXDIGIT(*str) is true.
702
e7c1e6c1
KW
703=head1 Character case changing
704
1f607577
KW
705=for apidoc Am|U8|toUPPER|U8 ch
706Converts the specified character to uppercase. If the input is anything but an
707ASCII lowercase character, that input character itself is returned. Variant
c753c8d3 708C<toUPPER_A> is equivalent.
954c1994 709
1f607577
KW
710=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
711Converts the Unicode code point C<cp> to its uppercase version, and
712stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
713that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
714bytes since the uppercase version may be longer than the original character.
715
716The first code point of the uppercased version is returned
717(but note, as explained just above, that there may be more.)
718
719=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
720Converts the UTF-8 encoded character at C<p> to its uppercase version, and
721stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
722that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
723bytes since the uppercase version may be longer than the original character.
724
725The first code point of the uppercased version is returned
726(but note, as explained just above, that there may be more.)
727
728The input character at C<p> is assumed to be well-formed.
729
25200305
KW
730=for apidoc Am|U8|toFOLD|U8 ch
731Converts the specified character to foldcase. If the input is anything but an
732ASCII uppercase character, that input character itself is returned. Variant
733C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
734Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
735
1f607577
KW
736=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
737Converts the Unicode code point C<cp> to its foldcase version, and
738stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
739that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
740bytes since the foldcase version may be longer than the original character.
741
742The first code point of the foldcased version is returned
743(but note, as explained just above, that there may be more.)
744
745=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
746Converts the UTF-8 encoded character at C<p> to its foldcase version, and
747stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
748that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
749bytes since the foldcase version may be longer than the original character.
750
751The first code point of the foldcased version is returned
752(but note, as explained just above, that there may be more.)
753
754The input character at C<p> is assumed to be well-formed.
755
756=for apidoc Am|U8|toLOWER|U8 ch
757Converts the specified character to lowercase. If the input is anything but an
758ASCII uppercase character, that input character itself is returned. Variant
c753c8d3 759C<toLOWER_A> is equivalent.
954c1994 760
1f607577
KW
761=for apidoc Am|U8|toLOWER_L1|U8 ch
762Converts the specified Latin1 character to lowercase. The results are undefined if
763the input doesn't fit in a byte.
764
765=for apidoc Am|U8|toLOWER_LC|U8 ch
766Converts the specified character to lowercase using the current locale's rules,
767if possible; otherwise returns the input character itself.
768
769=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
770Converts the Unicode code point C<cp> to its lowercase version, and
771stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
772that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
773bytes since the lowercase version may be longer than the original character.
774
775The first code point of the lowercased version is returned
776(but note, as explained just above, that there may be more.)
777
778=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
779Converts the UTF-8 encoded character at C<p> to its lowercase version, and
780stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
781that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
782bytes since the lowercase version may be longer than the original character.
783
784The first code point of the lowercased version is returned
785(but note, as explained just above, that there may be more.)
786
787The input character at C<p> is assumed to be well-formed.
788
789=for apidoc Am|U8|toLOWER_LC|U8 ch
790Converts the specified character to lowercase using the current locale's rules,
791if possible; otherwise returns the input character itself.
792
25200305
KW
793=for apidoc Am|U8|toTITLE|U8 ch
794Converts the specified character to titlecase. If the input is anything but an
795ASCII lowercase character, that input character itself is returned. Variant
796C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
797as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
798concept used in locale handling, so there is no functionality for that.)
799
1f607577
KW
800=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
801Converts the Unicode code point C<cp> to its titlecase version, and
802stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
803that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
804bytes since the titlecase version may be longer than the original character.
805
806The first code point of the titlecased version is returned
807(but note, as explained just above, that there may be more.)
808
809=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
810Converts the UTF-8 encoded character at C<p> to its titlecase version, and
811stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
812that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
813bytes since the titlecase version may be longer than the original character.
814
815The first code point of the titlecased version is returned
816(but note, as explained just above, that there may be more.)
817
818The input character at C<p> is assumed to be well-formed.
819
954c1994 820=cut
353c9b6f 821
1f607577 822XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
25200305 823really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
243effed 824
8a58bdcf 825Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
826patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
827
954c1994
GS
828*/
829
7c062697
KW
830/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
831 * is known only in the perl core, and this macro can be called from outside
832 * that */
833#ifdef HAS_QUAD
834# define WIDEST_UTYPE U64TYPE
835#else
836# define WIDEST_UTYPE U32
837#endif
838
3912bc88
KW
839/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
840 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
841 * bits of information are lost even on a 64-bit machine, but to get the
842 * compiler to optimize it out if possible. This is because Configure makes
843 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
844 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
845 */
846#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 847
41f43cc2 848#ifdef EBCDIC
b6340bd0
KW
849# ifndef _ALL_SOURCE
850 /* This returns the wrong results on at least z/OS unless this is
851 * defined. */
852# error _ALL_SOURCE should probably be defined
853# endif
854
855 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
856 * likely not defined, and so we use the native function */
857# define isASCII(c) isascii(c)
41f43cc2
KW
858#else
859# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
860#endif
861
ec79cfbb 862#define isASCII_A(c) isASCII(c)
3509bb53 863#define isASCII_L1(c) isASCII(c)
6515cb76 864
c2da0b36
KW
865/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
866 * and the 8 possible permutations of those bits exactly comprise the 8 octal
867 * digits */
3d33b3e3 868#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 869
ec79cfbb 870/* ASCII range only */
9fb1bf9d 871#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 872
a500dc72
KW
873/* Character class numbers. For internal core Perl use only. The ones less
874 * than 32 are used in PL_charclass[] and the ones up through the one that
875 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
876 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
877 * their actual definitions are here. If that file has a name not used here,
878 * it won't compile.
1709d539
KW
879 *
880 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
881 * frequency of their use. This gives a slight edge to exiting a loop earlier
882 * (in reginclass() in regexec.c) */
1709d539
KW
883# define _CC_WORDCHAR 0 /* \w and [:word:] */
884# define _CC_DIGIT 1 /* \d and [:digit:] */
885# define _CC_ALPHA 2 /* [:alpha:] */
886# define _CC_LOWER 3 /* [:lower:] */
887# define _CC_UPPER 4 /* [:upper:] */
888# define _CC_PUNCT 5 /* [:punct:] */
889# define _CC_PRINT 6 /* [:print:] */
15861f94 890# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 891# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 892# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 893
b0d691b2 894#define _FIRST_NON_SWASH_CC 10
1709d539
KW
895/* The character classes above are implemented with swashes. The second group
896 * (just below) contains the ones implemented without. These are also sorted
897 * in rough order of the frequency of their use, except that \v should be last,
898 * as it isn't a real Posix character class, and some (small) inefficiencies in
899 * regular expression handling would be introduced by putting it in the middle
900 * of those that are. Also, cntrl and ascii come after the others as it may be
901 * useful to group these which have no members that match above Latin1, (or
902 * above ASCII in the latter case) */
903
b0d691b2
KW
904# define _CC_SPACE 10 /* \s */
905# define _CC_BLANK 11 /* [:blank:] */
906# define _CC_XDIGIT 12 /* [:xdigit:] */
907# define _CC_PSXSPC 13 /* [:space:] */
908# define _CC_CNTRL 14 /* [:cntrl:] */
909# define _CC_ASCII 15 /* [:ascii:] */
910# define _CC_VERTSPACE 16 /* \v */
1709d539 911
a0947d7b
KW
912# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
913
1709d539 914/* The members of the third group below do not need to be coordinated with data
3ded5eb0
KW
915 * structures in regcomp.[ch] and regexec.c. But they should be added to
916 * bootstrap_ctype() */
b0d691b2
KW
917# define _CC_IDFIRST 17
918# define _CC_CHARNAME_CONT 18
919# define _CC_NONLATIN1_FOLD 19
920# define _CC_QUOTEMETA 20
921# define _CC_NON_FINAL_FOLD 21
922# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
923# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
924/* Unused: 23-30
f4cdb42c
KW
925 * If more bits are needed, one could add a second word for non-64bit
926 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
927 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
928 * is used only for optimization (as of this writing), and differs in the
929 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
930 * characters: the masculine and feminine ordinal indicators, so removing it
931 * would just cause /i regexes which match them to run less efficiently */
96ac0975 932
3a371f2f
KW
933#if defined(PERL_CORE) || defined(PERL_EXT)
934/* An enum version of the character class numbers, to help compilers
935 * optimize */
936typedef enum {
3a371f2f 937 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
938 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
939 _CC_ENUM_ASCII = _CC_ASCII,
940 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 941 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 942 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
943 _CC_ENUM_DIGIT = _CC_DIGIT,
944 _CC_ENUM_GRAPH = _CC_GRAPH,
945 _CC_ENUM_LOWER = _CC_LOWER,
946 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 947 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 948 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 949 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 950 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 951 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 952 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 953 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
954} _char_class_number;
955#endif
956
63c61c3f 957#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 958#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 959
3018b823 960#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 961# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
962 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
963 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
964 #error Need to adjust order of swash_property_names[]
965# endif
966
b44daeeb
KW
967/* This is declared static in each of the few files that this is #defined for
968 * to keep them from being publicly accessible. Hence there is a small amount
969 * of wasted space */
970
922e8cb4
KW
971static const char* const swash_property_names[] = {
972 "XPosixWord",
973 "XPosixDigit",
974 "XPosixAlpha",
975 "XPosixLower",
976 "XPosixUpper",
977 "XPosixPunct",
978 "XPosixPrint",
979 "XPosixAlnum",
b0d691b2
KW
980 "XPosixGraph",
981 "Cased"
922e8cb4
KW
982};
983#endif
984
96ac0975
NC
985# ifdef DOINIT
986EXTCONST U32 PL_charclass[] = {
987# include "l1_char_class_tab.h"
988};
989
990# else /* ! DOINIT */
991EXTCONST U32 PL_charclass[];
992# endif
3ded5eb0
KW
993#endif /* Has perl.h */
994
995#if defined(H_PERL) && ! defined(BOOTSTRAP_CHARSET)
96ac0975 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 */
3ded5eb0 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)
d95f8b6a 1026# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
3ded5eb0
KW
1027# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1028# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1029# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1030
1031 /* continuation character for legal NAME in \N{NAME} */
1032# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1033
1034# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1035# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1036# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1037# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1038# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1039# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1040# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1041# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1042# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1043# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
f4cdb42c 1044
fb9bfbf7
KW
1045 /* Either participates in a fold with a character above 255, or is a
1046 * multi-char fold */
f4cd282c 1047# 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 1048
4eeeb416 1049# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
1050# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1051 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
1052# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1053 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
3ded5eb0
KW
1054#else /* Either don't have perl.h or don't want to use char_class_tab.h */
1055
1056 /* If we don't have perl.h, we are compiling a utility program. Below we
1057 * hard-code various macro definitions that wouldn't otherwise be available
1058 * to it. We can also get here if we are configured to bootstrap up Perl
1059 * on a non-ASCII platform that doesn't have a working Perl (currently only
1060 * EBCDIC). For these we currently use the native definitions to get
1061 * things going. (It should also be possible to use the translation
1062 * function NATIVE_TO_LATIN1(), but that is an extra layer of dependence on
1063 * Perl, so it is currently avoided for the macros where it's possible to
1064 * do so.) */
9fb1bf9d 1065# ifdef EBCDIC
3ded5eb0 1066 /* Use the native functions. They likely will return false for all
4f48cecf 1067 * non-ASCII values, but this makes sure */
b6340bd0 1068# define isALPHA_A(c) (isASCII(c) && isalpha(c))
3ded5eb0 1069# define isALPHA_A(c) (isASCII(c) && isalpha(c))
b6340bd0
KW
1070# define isALPHANUMERIC_A(c) (isASCII(c) && isalnum(c))
1071# define isCNTRL_A(c) (isASCII(c) && iscntrl(c))
1072# define isDIGIT_A(c) (isASCII(c) && isdigit(c))
1073# define isGRAPH_A(c) (isASCII(c) && isgraph(c))
1074# define isLOWER_A(c) (isASCII(c) && islower(c))
1075# define isPRINT_A(c) (isASCII(c) && isprint(c))
1076# define isPUNCT_A(c) (isASCII(c) && ispunct(c))
1077# define isSPACE_A(c) (isASCII(c) && isspace(c))
1078# define isUPPER_A(c) (isASCII(c) && isupper(c))
1079# define isXDIGIT_A(c) (isASCII(c) && isxdigit(c))
3ded5eb0 1080# else /* ASCII platform. These are coded based on first principals */
9fb1bf9d 1081# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
15861f94 1082# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
3ded5eb0 1083# define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
2181da29 1084# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
3ded5eb0 1085# define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
9fb1bf9d
KW
1086# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1087# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
3ded5eb0
KW
1088# define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1089# define isSPACE_A(c) ((c) == ' ' \
1090 || (c) == '\t' \
1091 || (c) == '\n' \
1092 || (c) == '\r' \
1093 || (c) =='\v' \
926bb54c 1094 || (c) == '\f')
2181da29 1095# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
3ded5eb0
KW
1096# define isXDIGIT_A(c) (isDIGIT_A(c) \
1097 || ((c) >= 'a' && (c) <= 'f') \
1098 || ((c) <= 'F' && (c) >= 'A'))
1099# endif /* Below are common definitions for ASCII and non-ASCII */
1100# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1101# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1102# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1103
1104 /* The _L1 macros may be unnecessary for both the utilities and for
1105 * bootstrapping; I (khw) added them during debugging of bootstrapping, and
1106 * it seems best to keep them. */
1107# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1108# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1109# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1110# define isBLANK_L1(c) (isBLANK_A(c) \
1111 || (FITS_IN_8_BITS(c) \
1112 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1113# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1114# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1115# define isLOWER_L1(c) (isLOWER_A(c) \
1116 || (FITS_IN_8_BITS(c) \
1117 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1118 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1119 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1120 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1121 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1122# define isPRINT_L1(c) (isPRINT_A(c) \
1123 || (FITS_IN_8_BITS(c) \
1124 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1125# define isPSXSPC_L1(c) isSPACE_L1(c)
1126# define isPUNCT_L1(c) (isPUNCT_A(c) \
1127 || (FITS_IN_8_BITS(c) \
1128 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1129 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1130 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1131 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1132 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1133 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1134 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1135# define isSPACE_L1(c) (isSPACE_A(c) \
1136 || (FITS_IN_8_BITS(c) \
1137 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1138 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1139# define isUPPER_L1(c) (isUPPER_A(c) \
1140 || (FITS_IN_8_BITS(c) \
1141 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1142 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1143 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1144# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1145# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1146# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1147 || isBLANK_L1(c) \
1148 || (c) == '-' \
1149 || (c) == '(' \
1150 || (c) == ')')
1151 /* The following are not fully accurate in the above-ASCII range. I (khw)
1152 * don't think it's necessary to be so for the purposes where this gets
1153 * compiled */
1154# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1155# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1156
1157 /* And these aren't accurate at all. They are useful only for above
1158 * Latin1, which utilities and bootstrapping don't deal with */
1159# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1160# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1161
1162 /* Many of the macros later in this file are defined in terms of these. By
1163 * implementing them with a function, which converts the class number into
1164 * a call to the desired macro, all of the later ones work. However, that
1165 * function won't be actually defined when building a utility program (no
1166 * perl.h), and so a compiler error will be generated if one is attempted
1167 * to be used. And the above-Latin1 code points require Unicode tables to
1168 * be present, something unlikely to be the case when bootstrapping */
1169# define _generic_isCC(c, classnum) \
1170 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1171# define _generic_isCC_A(c, classnum) \
1172 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
1173#endif /* End of no perl.h or have BOOTSTRAP_CHARSET */
8a58bdcf 1174
e66b99e9
KW
1175#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1176#define isALPHA(c) isALPHA_A(c)
1177#define isBLANK(c) isBLANK_A(c)
1178#define isCNTRL(c) isCNTRL_A(c)
1179#define isDIGIT(c) isDIGIT_A(c)
1180#define isGRAPH(c) isGRAPH_A(c)
1181#define isIDFIRST(c) isIDFIRST_A(c)
1182#define isLOWER(c) isLOWER_A(c)
1183#define isPRINT(c) isPRINT_A(c)
1184#define isPSXSPC(c) isPSXSPC_A(c)
1185#define isPUNCT(c) isPUNCT_A(c)
1186#define isSPACE(c) isSPACE_A(c)
1187#define isUPPER(c) isUPPER_A(c)
1188#define isWORDCHAR(c) isWORDCHAR_A(c)
1189#define isXDIGIT(c) isXDIGIT_A(c)
1190
1191/* ASCII casing. These could also be written as
1192 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1193 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1194 which uses table lookup and mask instead of subtraction. (This would
1195 work because the _MOD does not apply in the ASCII range) */
1196#define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
1197#define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
bbce6d69 1198
25200305
KW
1199/* In the ASCII range, these are equivalent to what they're here defined to be.
1200 * But by creating these definitions, other code doesn't have to be aware of
1201 * this detail */
1202#define toFOLD(c) toLOWER(c)
1203#define toFOLD_LC(c) toLOWER_LC(c)
1204#define toTITLE(c) toUPPER(c)
1205
c753c8d3
KW
1206#define toLOWER_A(c) toLOWER(c)
1207#define toUPPER_A(c) toUPPER(c)
25200305
KW
1208#define toFOLD_A(c) toFOLD(c)
1209#define toTITLE_A(c) toTITLE(c)
1a0901db
KW
1210
1211/* Use table lookup for speed; return error character for input
1212 * out-of-range */
b2bf251f 1213#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1214 ? (c) \
f4cd282c 1215 : PL_latin1_lc[ (U8) (c) ])
c753c8d3
KW
1216#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1217
1a0901db
KW
1218/* Modified uc. Is correct uc except for three non-ascii chars which are
1219 * all mapped to one of them, and these need special handling; error
1220 * character for input out-of-range */
b2bf251f 1221#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1222 ? (c) \
f4cd282c 1223 : PL_mod_latin1_uc[ (U8) (c) ])
bbce6d69
PP
1224#ifdef USE_NEXT_CTYPE
1225
15861f94 1226# define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c))
ff68c719 1227# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 1228# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 1229# define isBLANK_LC(c) isBLANK((unsigned int)(c))
b8c5462f 1230# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
509fb054 1231# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
b8c5462f 1232# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
509fb054
KW
1233# define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
1234# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
ff68c719 1235# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 1236# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
509fb054
KW
1237# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
1238# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
1239# define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
e712593e 1240# define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
ff68c719 1241# define toLOWER_LC(c) NXToLower((unsigned int)(c))
509fb054 1242# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
bbce6d69
PP
1243
1244#else /* !USE_NEXT_CTYPE */
b8c5462f 1245
bbce6d69
PP
1246# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1247
76fde96a 1248/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
851131d3 1249
76fde96a 1250# define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
15861f94
KW
1251# define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \
1252 && isalnum((unsigned char)(c)))
07315176 1253# ifdef HAS_ISASCII
76fde96a 1254# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
07315176 1255# else
76fde96a 1256# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
07315176 1257# endif
67addccf 1258# ifdef HAS_ISBLANK
76fde96a 1259# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
67addccf 1260# else
76fde96a 1261# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
67addccf 1262# endif
76fde96a 1263# define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
509fb054 1264# define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
76fde96a 1265# define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
509fb054
KW
1266# define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1267 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1268# define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
76fde96a
KW
1269# define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1270# define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
509fb054
KW
1271# define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1272# define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1273# define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1274 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
e712593e 1275# define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
76fde96a 1276# define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? tolower((unsigned char)(c)) : (c))
509fb054 1277# define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? toupper((unsigned char)(c)) : (c))
bbce6d69
PP
1278
1279# else
1280
bbce6d69 1281# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1282# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
07315176 1283# define isASCII_LC(c) isascii(c)
67addccf
KW
1284# ifdef HAS_ISBLANK
1285# define isBLANK_LC(c) (isascii(c) && isblank(c))
1286# else
3e66ab74 1287# define isBLANK_LC(c) isBLANK_A(c)
67addccf 1288# endif
b8c5462f 1289# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1290# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1291# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1292# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1293# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1294# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1295# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1296# define isSPACE_LC(c) (isascii(c) && isspace(c))
1297# define isUPPER_LC(c) (isascii(c) && isupper(c))
1298# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1299# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
76fde96a 1300# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1301# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
bbce6d69
PP
1302
1303# endif
a0d0e21e 1304#endif /* USE_NEXT_CTYPE */
55204971 1305
eba68aa0
KW
1306#define isIDCONT(c) isWORDCHAR(c)
1307#define isIDCONT_A(c) isWORDCHAR_A(c)
1308#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1309#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1310#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1311
66c17564 1312/* For internal core Perl use only. If the input is Latin1, use the Latin1
a500dc72
KW
1313 * macro; otherwise use the function 'above_latin1'. Won't compile if 'c' isn't unsigned, as
1314 * won't match above_latin1 prototype. The macros do bounds checking, so have
66c17564
KW
1315 * duplicate checks here, so could create versions of the macros that don't,
1316 * but experiments show that gcc optimizes them out anyway. */
1317
1318/* Note that all ignore 'use bytes' */
cd500f2f
KW
1319#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1320 ? _generic_isCC(c, classnum) \
1321 : above_latin1(c))
1322#define _generic_swash_uni(classnum, c) ((c) < 256 \
1323 ? _generic_isCC(c, classnum) \
922e8cb4 1324 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1325#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1326#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1327#define isASCII_uni(c) isASCII(c)
1328#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1329#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1330#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1331#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1332#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1333#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1334#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1335#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1336
509fb054
KW
1337/* Posix and regular space are identical above Latin1 */
1338#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1339
cd500f2f 1340#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1341#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1342#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1343#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1344#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1345#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1346
b8d68ded 1347#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1348#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1349#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1350#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1351
cd500f2f
KW
1352#define _generic_LC_uvchr(latin1, above_latin1, c) \
1353 (c < 256 ? latin1(c) : above_latin1(c))
1354#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1355 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1356
1357#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1358#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1359 _CC_ALPHANUMERIC, c)
509fb054 1360#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1361#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1362#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1363#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1364#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1365#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1366 _is_uni_perl_idcont, c)
cd500f2f
KW
1367#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1368 _is_uni_perl_idstart, c)
1369#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1370#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1371#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1372 space under locale */
cd500f2f
KW
1373#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1374#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1375 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1376#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1377#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1378 _CC_WORDCHAR, c)
1379#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1380
36b00f93 1381#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1382
a500dc72
KW
1383/* Everything whose name begins with an underscore is for internal core Perl
1384 * use only. */
1385
1386/* If the input is in the Latin1 range, use
252810d0
KW
1387 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1388 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1389 * fact that ASCII characters have the same representation whether utf8 or not.
1390 * Note that it assumes that the utf8 has been validated, and ignores 'use
1391 * bytes' */
cd500f2f 1392#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1393 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1394 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c 1395 ? _generic_isCC( \
94bb8c36 1396 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
aea0c08c 1397 *((p)+1 )), \
94bb8c36 1398 classnum) \
252810d0 1399 : utf8)
cd500f2f
KW
1400/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1401 * can be a macro */
1402#define _generic_func_utf8(classnum, above_latin1, p) \
1403 _generic_utf8(classnum, p, above_latin1(p))
1404/* Like the above, but passes classnum to _isFOO_utf8(), instead of having a
1405 * 'above_latin1' parameter */
922e8cb4 1406#define _generic_swash_utf8(classnum, p) \
cd500f2f 1407 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1408
cc8ab7c0
KW
1409/* Like the above, but should be used only when it is known that there are no
1410 * characters in the range 128-255 which the class is TRUE for. Hence it can
cd500f2f 1411 * skip the tests for this range. 'above_latin1' should include its arguments */
3ded5eb0 1412#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1413 (UTF8_IS_INVARIANT(*(p)) \
1414 ? _generic_isCC(*(p), classnum) \
1415 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
3ded5eb0 1416 ? above_latin1 \
cc8ab7c0
KW
1417 : 0)
1418
84238efa
KW
1419/* NOTE that some of these macros have very similar ones in regcharclass.h.
1420 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1421 * there, differing in name only by an underscore from the one here
1422 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1423 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1424 * points; the regcharclass.h ones are implemented as a series of
1425 * "if-else-if-else ..." */
1426
922e8cb4
KW
1427#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1428#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1429#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1430 utf8, the non-utf8 macro works
1431 */
cd500f2f
KW
1432#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1433#define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1434#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1435 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1436#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1437#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1438 _is_utf8_perl_idcont, p)
e5dcd934 1439
c11ff943
KW
1440/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1441 * IDFIRST is an alnum. See
c8362b00 1442 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1443 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1444 * This used to be not the XID version, but we decided to go with the more
1445 * modern Unicode definition */
cd500f2f 1446#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1447 _is_utf8_perl_idstart, p)
e5dcd934 1448
922e8cb4
KW
1449#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1450#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1451
509fb054 1452/* Posix and regular space are identical above Latin1 */
cd500f2f 1453#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1454
922e8cb4 1455#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1456#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1457#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1458#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1459#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1460#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1461 is_XDIGIT_high(p))
a0ed51b3 1462
7d2b61dd 1463#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
509fb054
KW
1464#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1465#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1466#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1467
9201a572
KW
1468/* For internal core Perl use only. If the input is in the Latin1 range, use
1469 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1470 * use the value given by the 'utf8' parameter. This relies on the fact that
1471 * ASCII characters have the same representation whether utf8 or not. Note
1472 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
94bb8c36
KW
1473#define _generic_LC_utf8(macro, p, utf8) \
1474 (UTF8_IS_INVARIANT(*(p)) \
1475 ? macro(*(p)) \
1476 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1477 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
9201a572
KW
1478 : utf8)
1479
cd500f2f
KW
1480#define _generic_LC_swash_utf8(macro, classnum, p) \
1481 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1482#define _generic_LC_func_utf8(macro, above_latin1, p) \
1483 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1484
cd500f2f
KW
1485#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1486 _CC_ALPHANUMERIC, p)
1487#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1488#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1489#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1490#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1491#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1492#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1493#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1494#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1495#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1496#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1497#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1498 space under locale */
cd500f2f
KW
1499#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1500#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1501#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1502#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1503 _CC_WORDCHAR, p)
1504#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1505
fbc19f27
KW
1506/* Macros for backwards compatibility and for completeness when the ASCII and
1507 * Latin1 values are identical */
fbc19f27
KW
1508#define isALPHAU(c) isALPHA_L1(c)
1509#define isDIGIT_L1(c) isDIGIT_A(c)
1510#define isOCTAL(c) isOCTAL_A(c)
1511#define isOCTAL_L1(c) isOCTAL_A(c)
1512#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1513#define isALNUM(c) isWORDCHAR(c)
1514#define isALNUMU(c) isWORDCHAR_L1(c)
1515#define isALNUM_LC(c) isWORDCHAR_LC(c)
1516#define isALNUM_uni(c) isWORDCHAR_uni(c)
1517#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1518#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1519#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1520#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1521#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1522#define isALNUMC(c) isALPHANUMERIC(c)
1523#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1524#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1525#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1526#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1527#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1528
6ae9f32a 1529/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1530 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1531 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1532 * that is EOT on ASCII; ST on EBCDIC */
8e7c6e7d 1533# define toCTRL(c) (toUPPER(NATIVE_TO_LATIN1(c)) ^ 64)
bbce6d69 1534
dea28490
JJ
1535/* Line numbers are unsigned, 32 bits. */
1536typedef U32 line_t;
e5dcd934 1537#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1538
91152fc1
DG
1539/* Helpful alias for version prescan */
1540#define is_LAX_VERSION(a,b) \
1541 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1542
1543#define is_STRICT_VERSION(a,b) \
1544 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1545
1546#define BADVERSION(a,b,c) \
1547 if (b) { \
1548 *b = c; \
1549 } \
1550 return a;
8c52afec 1551
95a59cab
YO
1552#define READ_XDIGIT(s) (isALPHA(*(s)) ? ((*(s)++ + 9) & 0xf) : (*(s)++ & 0xf))
1553
8e84507e 1554/*
ccfc67b7
JH
1555=head1 Memory Management
1556
a02a5408 1557=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1558The XSUB-writer's interface to the C C<malloc> function.
1559
c5008215
JC
1560In 5.9.3, Newx() and friends replace the older New() API, and drops
1561the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1562themselves. This aid has been superseded by a new build option,
d10b4965 1563PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1564there for use in XS modules supporting older perls.
1565
a02a5408 1566=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1567The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1568cast. See also C<Newx>.
954c1994 1569
a02a5408 1570=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1571The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1572memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1573
954c1994
GS
1574=for apidoc Am|void|Renew|void* ptr|int nitems|type
1575The XSUB-writer's interface to the C C<realloc> function.
1576
1577=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1578The XSUB-writer's interface to the C C<realloc> function, with
1579cast.
1580
49b8b560 1581=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1582The XSUB-writer's interface to the C C<free> function.
1583
1584=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1585The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1586source, C<dest> is the destination, C<nitems> is the number of items, and
1587C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1588
e90e2364
NC
1589=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1590Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1591optimise.
1592
954c1994
GS
1593=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1594The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1595source, C<dest> is the destination, C<nitems> is the number of items, and
1596C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1597
e90e2364
NC
1598=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1599
1600Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1601optimise.
1602
954c1994
GS
1603=for apidoc Am|void|Zero|void* dest|int nitems|type
1604
1605The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1606destination, C<nitems> is the number of items, and C<type> is the type.
1607
e90e2364
NC
1608=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1609
1610Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1611optimise.
1612
da5d8dbb 1613=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1614This is an architecture-independent macro to copy one structure to another.
954c1994 1615
7e337ee0
JH
1616=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1617
1618Fill up memory with a byte pattern (a byte repeated over and over
1619again) that hopefully catches attempts to access uninitialized memory.
1620
1621=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1622
1623PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1624
1c12ffb4 1625=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1626
1627PoisonWith(0xEF) for catching access to freed memory.
1628
9965345d
JH
1629=for apidoc Am|void|Poison|void* dest|int nitems|type
1630
7e337ee0 1631PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1632
1633=cut */
954c1994 1634
561b68a9
SH
1635/* Maintained for backwards-compatibility only. Use newSV() instead. */
1636#ifndef PERL_CORE
ff06c60c 1637#define NEWSV(x,len) newSV(len)
561b68a9 1638#endif
ff06c60c 1639
19a94d75
JH
1640#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1641
2720f793
JH
1642/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1643 * overly eager compilers that will bleat about e.g.
1644 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1645#ifdef PERL_MALLOC_WRAP
f019c49e 1646#define MEM_WRAP_CHECK(n,t) \
d1decf2b 1647 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
27d5b266 1648#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1649 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1650#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1651
d1decf2b 1652#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
1653#else
1654
410319be
NC
1655#define MEM_WRAP_CHECK(n,t)
1656#define MEM_WRAP_CHECK_1(n,t,a)
1657#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1658#define MEM_WRAP_CHECK_(n,t)
1659
6fff79ce 1660#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1661
1936d2a7 1662#endif
8b44ba4c 1663
fe4f188c 1664#ifdef PERL_MEM_LOG
46c6c7e2 1665/*
9f653bb5 1666 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1667 * go through functions, which are handy for debugging breakpoints, but
1668 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1669 * line number, and C function name if available) passed in. This info can
1670 * then be used for logging the calls, for which one gets a sample
73d1d973 1671 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1672 *
46c6c7e2 1673 * Known problems:
94e892a6 1674 * - not all memory allocs get logged, only those
46c6c7e2 1675 * that go through Newx() and derivatives (while all
94e892a6 1676 * Safefrees do get logged)
46c6c7e2
JH
1677 * - __FILE__ and __LINE__ do not work everywhere
1678 * - __func__ or __FUNCTION__ even less so
1679 * - I think more goes on after the perlio frees but
1680 * the thing is that STDERR gets closed (as do all
1681 * the file descriptors)
1682 * - no deeper calling stack than the caller of the Newx()
1683 * or the kind, but do I look like a C reflection/introspection
1684 * utility to you?
1685 * - the function prototypes for the logging functions
1686 * probably should maybe be somewhere else than handy.h
1687 * - one could consider inlining (macrofying) the logging
1688 * for speed, but I am too lazy
1689 * - one could imagine recording the allocations in a hash,
1690 * (keyed by the allocation address?), and maintain that
1691 * through reallocs and frees, but how to do that without
1692 * any News() happening...?
73d1d973 1693 * - lots of -Ddefines to get useful/controllable output
b953482e 1694 * - lots of ENV reads
46c6c7e2
JH
1695 */
1696
12754f92 1697PERL_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 1698
12754f92 1699PERL_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 1700
12754f92 1701PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1702
0b0ab801 1703# ifdef PERL_CORE
73d1d973 1704# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1705enum mem_log_type {
1706 MLT_ALLOC,
1707 MLT_REALLOC,
d7a2c63c
MHM
1708 MLT_FREE,
1709 MLT_NEW_SV,
1710 MLT_DEL_SV
0b0ab801
MHM
1711};
1712# endif
12754f92 1713# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1714void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1715void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1716# endif
0b0ab801
MHM
1717# endif
1718
fe4f188c
JH
1719#endif
1720
1721#ifdef PERL_MEM_LOG
d1401ee9
MHM
1722#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1723#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1724#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1725#endif
1726
1727#ifndef MEM_LOG_ALLOC
1728#define MEM_LOG_ALLOC(n,t,a) (a)
1729#endif
1730#ifndef MEM_LOG_REALLOC
1731#define MEM_LOG_REALLOC(n,t,v,a) (a)
1732#endif
1733#ifndef MEM_LOG_FREE
1734#define MEM_LOG_FREE(a) (a)
1735#endif
1736
d1401ee9
MHM
1737#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1738#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1739#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1740
1741#ifndef PERL_CORE
a02a5408
JC
1742/* pre 5.9.x compatibility */
1743#define New(x,v,n,t) Newx(v,n,t)
1744#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1745#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1746#endif
a02a5408 1747
ff68c719 1748#define Renew(v,n,t) \
d1401ee9 1749 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1750#define Renewc(v,n,t,c) \
d1401ee9 1751 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1752
1753#ifdef PERL_POISON
1754#define Safefree(d) \
06c0cc96 1755 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1756#else
fe4f188c 1757#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1758#endif
55497cff 1759
8b44ba4c
NC
1760#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1761#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1762#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1763
8b44ba4c
NC
1764#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1765#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1766#ifdef HAS_MEMSET
8b44ba4c 1767#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1768#else
8b44ba4c
NC
1769/* Using bzero(), which returns void. */
1770#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1771#endif
1772
7e337ee0
JH
1773#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1774#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1775#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1776#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1777
2304df62 1778#ifdef USE_STRUCT_COPY
ff68c719 1779#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1780#else
1781#define StructCopy(s,d,t) Copy(s,d,1,t)
1782#endif
2cc61e15 1783
622913ab 1784#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1785#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1786
2cc61e15
DD
1787#ifdef NEED_VA_COPY
1788# ifdef va_copy
1789# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1790# else
a1866d1b
JH
1791# if defined(__va_copy)
1792# define Perl_va_copy(s, d) __va_copy(d, s)
1793# else
1794# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1795# endif
2cc61e15
DD
1796# endif
1797#endif
1798
472d47bc
SB
1799/* convenience debug macros */
1800#ifdef USE_ITHREADS
1801#define pTHX_FORMAT "Perl interpreter: 0x%p"
1802#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1803#define pTHX_VALUE_ (void *)my_perl,
1804#define pTHX_VALUE (void *)my_perl
1805#define pTHX__VALUE_ ,(void *)my_perl,
1806#define pTHX__VALUE ,(void *)my_perl
472d47bc 1807#else
3609ea0d 1808#define pTHX_FORMAT
472d47bc 1809#define pTHX__FORMAT
3609ea0d 1810#define pTHX_VALUE_
472d47bc 1811#define pTHX_VALUE
3609ea0d 1812#define pTHX__VALUE_
472d47bc
SB
1813#define pTHX__VALUE
1814#endif /* USE_ITHREADS */
3609ea0d 1815
2acdbac1
NC
1816/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1817 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1818 CPAN::Unpack show any users outside the core. */
1819#ifdef PERL_CORE
d1d15184 1820# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1821#endif
1822
dfff4baf
BF
1823/* Internal macros to deal with gids and uids */
1824#ifdef PERL_CORE
1825
1826# if Uid_t_size > IVSIZE
1827# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
1828# define SvUID(sv) SvNV(sv)
1829# else
1830# if Uid_t_sign <= 0
1831# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
1832# define SvUID(sv) SvIV(sv)
1833# else
1834# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
1835# define SvUID(sv) SvUV(sv)
1836# endif
1837# endif /* Uid_t_size */
1838
1839# if Gid_t_size > IVSIZE
1840# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
1841# define SvGID(sv) SvNV(sv)
1842# else
1843# if Gid_t_sign <= 0
1844# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
1845# define SvGID(sv) SvIV(sv)
1846# else
1847# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
1848# define SvGID(sv) SvUV(sv)
1849# endif
1850# endif /* Gid_t_size */
1851
1852#endif
1853
9d745869
KW
1854#endif /* HANDY_H */
1855
e9a8c099
MHM
1856/*
1857 * Local variables:
1858 * c-indentation-style: bsd
1859 * c-basic-offset: 4
14d04a33 1860 * indent-tabs-mode: nil
e9a8c099
MHM
1861 * End:
1862 *
14d04a33 1863 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1864 */