This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
handy.h: Add debugging assertion
[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
72d33970
FC
28Null character pointer. (No longer available when C<PERL_CORE> is
29defined.)
2307c6d0 30
954c1994 31=for apidoc AmU||Nullsv
72d33970 32Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
33
34=cut
35*/
36
24792b8d
NC
37# define Nullch Null(char*)
38# define Nullfp Null(PerlIO*)
39# define Nullsv Null(SV*)
40#endif
8d063cd8 41
641d3f0b
PP
42#ifdef TRUE
43#undef TRUE
44#endif
45#ifdef FALSE
46#undef FALSE
47#endif
48#define TRUE (1)
49#define FALSE (0)
50
cf3f0ffb
DM
51/* The MUTABLE_*() macros cast pointers to the types shown, in such a way
52 * (compiler permitting) that casting away const-ness will give a warning;
53 * e.g.:
54 *
55 * const SV *sv = ...;
56 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
57 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
58 */
59
b1bc3f34 60#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 61# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
62#else
63# define MUTABLE_PTR(p) ((void *) (p))
64#endif
65
a062e10d 66#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 67#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 68#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 69#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 70#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 71#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 72
f789f6a4 73#if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR)
bd31be4b
NC
74# include <stdbool.h>
75# ifndef HAS_BOOL
76# define HAS_BOOL 1
77# endif
78#endif
79
8e84507e 80/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
81 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
82 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 83 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
84 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
85 g++ can be identified by __GNUG__.
86 Andy Dougherty February 2000
5d94fbed 87*/
3609ea0d 88#ifdef __GNUG__ /* GNU g++ has bool built-in */
f789f6a4 89# ifndef PERL_BOOL_AS_CHAR
5d94fbed 90# ifndef HAS_BOOL
c1d22f6b 91# define HAS_BOOL 1
5d94fbed 92# endif
f789f6a4 93# endif
5d94fbed
AD
94#endif
95
641d3f0b
PP
96/* The NeXT dynamic loader headers will not build with the bool macro
97 So declare them now to clear confusion.
98*/
8f1f23e8 99#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
100# undef FALSE
101# undef TRUE
102 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
103# define ENUM_BOOL 1
104# ifndef HAS_BOOL
105# define HAS_BOOL 1
106# endif /* !HAS_BOOL */
8f1f23e8 107#endif /* NeXT || __NeXT__ */
641d3f0b 108
5d94fbed 109#ifndef HAS_BOOL
f789f6a4
FC
110# ifdef bool
111# undef bool
112# endif
70d5cb32 113# define bool char
c1d22f6b 114# define HAS_BOOL 1
a687059c 115#endif
0d3e774c 116
236415e2
KW
117/* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
118 * defined as char for example, then the cast from int is
119 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
120 * AIX */
18f5643b 121#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 122
46c6c7e2 123/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
124 * XXX Should really be a Configure probe, with HAS__FUNCTION__
125 * and FUNCTION__ as results.
126 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
127#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
128# define FUNCTION__ __func__
129#else
9ea40801 130# if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 131 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
46c6c7e2
JH
132# define FUNCTION__ ""
133# else
134# define FUNCTION__ __FUNCTION__ /* Common extension. */
135# endif
136#endif
137
27d4fb96
PP
138/* XXX A note on the perl source internal type system. The
139 original intent was that I32 be *exactly* 32 bits.
140
141 Currently, we only guarantee that I32 is *at least* 32 bits.
142 Specifically, if int is 64 bits, then so is I32. (This is the case
143 for the Cray.) This has the advantage of meshing nicely with
144 standard library calls (where we pass an I32 and the library is
145 expecting an int), but the disadvantage that an I32 is not 32 bits.
146 Andy Dougherty August 1996
24fef2a7 147
dc45a647
MB
148 There is no guarantee that there is *any* integral type with
149 exactly 32 bits. It is perfectly legal for a system to have
150 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 151
dc45a647
MB
152 Similarly, there is no guarantee that I16 and U16 have exactly 16
153 bits.
693762b4 154
8e84507e
NIS
155 For dealing with issues that may arise from various 32/64-bit
156 systems, we will ask Configure to check out
8175356b 157
3609ea0d
JH
158 SHORTSIZE == sizeof(short)
159 INTSIZE == sizeof(int)
160 LONGSIZE == sizeof(long)
dc45a647 161 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 162 PTRSIZE == sizeof(void *)
dc45a647
MB
163 DOUBLESIZE == sizeof(double)
164 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 165
27d4fb96
PP
166*/
167
69512466
JH
168#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
169# include <inttypes.h>
dd0eed91
JH
170# ifdef INT32_MIN_BROKEN
171# undef INT32_MIN
172# define INT32_MIN (-2147483647-1)
173# endif
174# ifdef INT64_MIN_BROKEN
175# undef INT64_MIN
176# define INT64_MIN (-9223372036854775807LL-1)
177# endif
69512466
JH
178#endif
179
8175356b
JH
180typedef I8TYPE I8;
181typedef U8TYPE U8;
182typedef I16TYPE I16;
183typedef U16TYPE U16;
184typedef I32TYPE I32;
185typedef U32TYPE U32;
6b8eaf93
JH
186#ifdef PERL_CORE
187# ifdef HAS_QUAD
8175356b
JH
188typedef I64TYPE I64;
189typedef U64TYPE U64;
6b8eaf93
JH
190# endif
191#endif /* PERL_CORE */
8175356b 192
69512466 193#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
80524f33
JL
194# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
195# define PeRl_INT64_C(c) CAT2(c,LL)
196# define PeRl_UINT64_C(c) CAT2(c,ULL)
197# else
52865553
SH
198# if QUADKIND == QUAD_IS___INT64
199# define PeRl_INT64_C(c) CAT2(c,I64)
200# define PeRl_UINT64_C(c) CAT2(c,UI64)
69512466 201# else
52865553
SH
202# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
203# define PeRl_INT64_C(c) CAT2(c,L)
204# define PeRl_UINT64_C(c) CAT2(c,UL)
69512466 205# else
80524f33
JL
206# define PeRl_INT64_C(c) ((I64TYPE)(c))
207# define PeRl_UINT64_C(c) ((U64TYPE)(c))
69512466
JH
208# endif
209# endif
e8c95190 210# endif
80524f33
JL
211# ifndef UINT64_C
212# define UINT64_C(c) PeRl_UINT64_C(c)
213# endif
214# ifndef INT64_C
215# define INT64_C(c) PeRl_INT64_C(c)
216# endif
e8c95190 217#endif
dc750f2e 218
d8668976 219#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 220
5ff3f7a4
GS
221/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
222 Please search CHAR_MAX in perl.h for further details. */
223#define U8_MAX UINT8_MAX
224#define U8_MIN UINT8_MIN
225
5ff3f7a4
GS
226#define I16_MAX INT16_MAX
227#define I16_MIN INT16_MIN
228#define U16_MAX UINT16_MAX
229#define U16_MIN UINT16_MIN
230
5ff3f7a4
GS
231#define I32_MAX INT32_MAX
232#define I32_MIN INT32_MIN
0e983133
GS
233#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
234# define U32_MAX UINT32_MAX
235#else
236# define U32_MAX 4294967295U
237#endif
5ff3f7a4
GS
238#define U32_MIN UINT32_MIN
239
240#else
241
5c9fa16e
KA
242/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
243 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
244#define U8_MAX PERL_UCHAR_MAX
245#define U8_MIN PERL_UCHAR_MIN
79072805 246
27d4fb96
PP
247#define I16_MAX PERL_SHORT_MAX
248#define I16_MIN PERL_SHORT_MIN
249#define U16_MAX PERL_USHORT_MAX
250#define U16_MIN PERL_USHORT_MIN
79072805 251
c4f23d77 252#if LONGSIZE > 4
27d4fb96
PP
253# define I32_MAX PERL_INT_MAX
254# define I32_MIN PERL_INT_MIN
255# define U32_MAX PERL_UINT_MAX
256# define U32_MIN PERL_UINT_MIN
79072805 257#else
27d4fb96
PP
258# define I32_MAX PERL_LONG_MAX
259# define I32_MIN PERL_LONG_MIN
260# define U32_MAX PERL_ULONG_MAX
261# define U32_MIN PERL_ULONG_MIN
79072805
LW
262#endif
263
5ff3f7a4
GS
264#endif
265
58a9a5d5 266/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
267#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
268#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
269#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
270
ff68c719 271#define Ctl(ch) ((ch) & 037)
8d063cd8 272
0f092d08
KW
273/* This is a helper macro to avoid preprocessor issues, expanding to an
274 * assert followed by a comma under DEBUGGING (hence the comma operator). If
275 * we didn't do this, we would get a comma with nothing before it when not
276 * DEBUGGING */
277#ifdef DEBUGGING
278# define __ASSERT_(statement) assert(statement),
279#else
280# define __ASSERT_(statement)
281#endif
282
3fe05580
MHM
283/*
284=head1 SV-Body Allocation
285
286=for apidoc Ama|SV*|newSVpvs|const char* s
287Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
288
84bafc02
NC
289=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
290Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
291pair.
292
3fe05580
MHM
293=for apidoc Ama|SV*|newSVpvs_share|const char* s
294Like C<newSVpvn_share>, but takes a literal string instead of a string/length
295pair and omits the hash parameter.
296
9dcc53ea
Z
297=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
298Like C<sv_catpvn_flags>, but takes a literal string instead of a
299string/length pair.
300
301=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
302Like C<sv_catpvn_nomg>, but takes a literal string instead of a
303string/length pair.
304
7b4e7469 305=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
306Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
307
9dcc53ea
Z
308=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
309Like C<sv_catpvn_mg>, but takes a literal string instead of a
310string/length pair.
311
7b4e7469 312=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
313Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
314
9dcc53ea
Z
315=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
316Like C<sv_setpvn_mg>, but takes a literal string instead of a
317string/length pair.
318
319=for apidoc Am|SV *|sv_setref_pvs|const char* s
320Like C<sv_setref_pvn>, but takes a literal string instead of a
321string/length pair.
322
3fe05580
MHM
323=head1 Memory Management
324
325=for apidoc Ama|char*|savepvs|const char* s
326Like C<savepvn>, but takes a literal string instead of a string/length pair.
327
9dcc53ea
Z
328=for apidoc Ama|char*|savesharedpvs|const char* s
329A version of C<savepvs()> which allocates the duplicate string in memory
330which is shared between threads.
331
3fe05580
MHM
332=head1 GV Functions
333
334=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
335Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
336
337=head1 Hash Manipulation Functions
338
339=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
340Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
341
342=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
343Like C<hv_store>, but takes a literal string instead of a string/length pair
344and omits the hash parameter.
345
510966aa
Z
346=head1 Lexer interface
347
348=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
349
350Like L</lex_stuff_pvn>, but takes a literal string instead of a
351string/length pair.
352
3fe05580
MHM
353=cut
354*/
355
926bb54c
KW
356/* concatenating with "" ensures that only literal strings are accepted as
357 * argument */
26bfd924 358#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 359
926bb54c
KW
360/* note that STR_WITH_LEN() can't be used as argument to macros or functions
361 * that under some configurations might be macros, which means that it requires
362 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
363 */
364
365/* STR_WITH_LEN() shortcuts */
366#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
367#define newSVpvs_flags(str,flags) \
368 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 369#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
370#define sv_catpvs_flags(sv, str, flags) \
371 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
372#define sv_catpvs_nomg(sv, str) \
373 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
374#define sv_catpvs(sv, str) \
375 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
376#define sv_catpvs_mg(sv, str) \
377 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 378#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
379#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
380#define sv_setref_pvs(rv, classname, str) \
381 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 382#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
383#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
384#define gv_stashpvs(str, create) \
385 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
386#define gv_fetchpvs(namebeg, add, sv_type) \
387 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
388#define gv_fetchpvn(namebeg, len, add, sv_type) \
389 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
390#define sv_catxmlpvs(dsv, str, utf8) \
391 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
392#define hv_fetchs(hv,key,lval) \
393 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
394 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
395 : HV_FETCH_JUST_SV, NULL, 0))
396
397#define hv_stores(hv,key,val) \
398 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
399 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 400
510966aa
Z
401#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
402
b96d8cd9
NC
403#define get_cvs(str, flags) \
404 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 405
954c1994 406/*
ccfc67b7
JH
407=head1 Miscellaneous Functions
408
954c1994
GS
409=for apidoc Am|bool|strNE|char* s1|char* s2
410Test two strings to see if they are different. Returns true or
411false.
412
413=for apidoc Am|bool|strEQ|char* s1|char* s2
414Test two strings to see if they are equal. Returns true or false.
415
416=for apidoc Am|bool|strLT|char* s1|char* s2
417Test two strings to see if the first, C<s1>, is less than the second,
418C<s2>. Returns true or false.
419
420=for apidoc Am|bool|strLE|char* s1|char* s2
421Test two strings to see if the first, C<s1>, is less than or equal to the
422second, C<s2>. Returns true or false.
423
424=for apidoc Am|bool|strGT|char* s1|char* s2
425Test two strings to see if the first, C<s1>, is greater than the second,
426C<s2>. Returns true or false.
427
428=for apidoc Am|bool|strGE|char* s1|char* s2
429Test two strings to see if the first, C<s1>, is greater than or equal to
430the second, C<s2>. Returns true or false.
431
432=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
433Test two strings to see if they are different. The C<len> parameter
72d33970 434indicates the number of bytes to compare. Returns true or false. (A
954c1994
GS
435wrapper for C<strncmp>).
436
437=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
438Test two strings to see if they are equal. The C<len> parameter indicates
72d33970 439the number of bytes to compare. Returns true or false. (A wrapper for
954c1994
GS
440C<strncmp>).
441
442=cut
443*/
444
8d063cd8
LW
445#define strNE(s1,s2) (strcmp(s1,s2))
446#define strEQ(s1,s2) (!strcmp(s1,s2))
447#define strLT(s1,s2) (strcmp(s1,s2) < 0)
448#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
449#define strGT(s1,s2) (strcmp(s1,s2) > 0)
450#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
451#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
452#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 453
36477c24
PP
454#ifdef HAS_MEMCMP
455# define memNE(s1,s2,l) (memcmp(s1,s2,l))
456# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
457#else
458# define memNE(s1,s2,l) (bcmp(s1,s2,l))
459# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
460#endif
461
568a785a 462#define memEQs(s1, l, s2) \
8e6497b6 463 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
464#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
465
bbce6d69
PP
466/*
467 * Character classes.
468 *
469 * Unfortunately, the introduction of locales means that we
470 * can't trust isupper(), etc. to tell the truth. And when
471 * it comes to /\w+/ with tainting enabled, we *must* be able
472 * to trust our character classes.
473 *
474 * Therefore, the default tests in the text of Perl will be
475 * independent of locale. Any code that wants to depend on
476 * the current locale will use the tests that begin with "lc".
477 */
478
2304df62
AD
479#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
480# ifndef CTYPE256
481# define CTYPE256
482# endif
483#endif
484
954c1994 485/*
ccfc67b7
JH
486
487=head1 Character classes
243effed
KW
488This section is about functions (really macros) that classify characters
489into types, such as punctuation versus alphabetic, etc. Most of these are
490analogous to regular expression character classes. (See
491L<perlrecharclass/POSIX Character Classes>.) There are several variants for
492each class. (Not all macros have all variants; each item below lists the
493ones valid for it.) None are affected by C<use bytes>, and only the ones
494with C<LC> in the name are affected by the current locale.
495
496The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
497C<U8>) as input and returns a boolean as to whether or not the character
6aff1f14
KW
498represented by that octet is (or on non-ASCII platforms, corresponds to) an
499ASCII character in the named class based on platform, Unicode, and Perl rules.
500If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 501
6aff1f14
KW
502Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
503with no suffix C<"_A">.
4b9734bf
KW
504
505Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
506onto the platform. That is, the code points that are ASCII are unaffected,
507since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
508as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
509true when called with the code point 0xDF, which is a word character in both
510ASCII and EBCDIC (though it represent different characters in each).
243effed
KW
511
512Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
513point as input. If the code point is larger than 255, Unicode rules are used
514to determine if it is in the character class. For example,
6aff1f14
KW
515C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
516WITH MACRON in Unicode, and is a word character.
243effed
KW
517
518Variant C<isFOO_utf8> is like C<isFOO_uni>, 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.
243effed
KW
522
523Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but uses
524the C library function that gives the named classification instead of
525hard-coded rules. For example, C<isDIGIT_LC()> returns the result of calling
526C<isdigit()>. This means that the result is based on the current locale, which
527is what C<LC> in the name stands for. FALSE is always returned if the input
528won't fit into an octet.
529
530Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
531returns the same as C<isFOO_LC> for input code points less than 256, and
532returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
533
534Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
535(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
536classification of just the first (possibly multi-byte) character in the string
537is tested.
ccfc67b7 538
954c1994 539=for apidoc Am|bool|isALPHA|char ch
2787a470 540Returns a boolean indicating whether the specified character is an
6aff1f14 541alphabetic character, analogous to C<m/[[:alpha:]]/>.
8a58bdcf 542See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 543C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 544C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 545
15861f94
KW
546=for apidoc Am|bool|isALPHANUMERIC|char ch
547Returns a boolean indicating whether the specified character is a either an
6aff1f14 548alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
15861f94
KW
549See the L<top of this section|/Character classes> for an explanation of variants
550C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 551C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
552C<isALPHANUMERIC_LC_utf8>.
553
8a58bdcf
KW
554=for apidoc Am|bool|isASCII|char ch
555Returns a boolean indicating whether the specified character is one of the 128
243effed 556characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 557On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
558character corresponds to an ASCII character. Variants C<isASCII_A()> and
559C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
560See the L<top of this section|/Character classes> for an explanation of variants
561C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
562C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
563library routine C<isascii()>. In these cases, the variants whose names contain
564C<LC> are the same as the corresponding ones without.
565
d98532ea
KW
566Also note, that because all ASCII characters are UTF-8 invariant (meaning they
567have the exact same representation (always a single byte) whether encoded in
568UTF-8 or not), C<isASCII> will give the correct results when called with any
569byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
570work properly on any string encoded or not in UTF-8.
571
243effed
KW
572=for apidoc Am|bool|isBLANK|char ch
573Returns a boolean indicating whether the specified character is a
6aff1f14 574character considered to be a blank, analogous to C<m/[[:blank:]]/>.
243effed 575See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 576C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
577C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
578platforms do not have the C library routine C<isblank()>. In these cases, the
579variants whose names contain C<LC> are the same as the corresponding ones
580without.
581
582=for apidoc Am|bool|isCNTRL|char ch
583Returns a boolean indicating whether the specified character is a
6aff1f14 584control character, analogous to C<m/[[:cntrl:]]/>.
243effed 585See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 586C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
587C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
588On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 589
954c1994 590=for apidoc Am|bool|isDIGIT|char ch
2787a470 591Returns a boolean indicating whether the specified character is a
6aff1f14 592digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 593Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed 594See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 595C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
596C<isDIGIT_LC_utf8>.
597
598=for apidoc Am|bool|isGRAPH|char ch
599Returns a boolean indicating whether the specified character is a
6aff1f14 600graphic character, analogous to C<m/[[:graph:]]/>.
243effed 601See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 602C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 603C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 604
0c82b6df 605=for apidoc Am|bool|isLOWER|char ch
2787a470 606Returns a boolean indicating whether the specified character is a
6aff1f14 607lowercase character, analogous to C<m/[[:lower:]]/>.
8a58bdcf 608See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 609C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 610C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 611
c99e91e9 612=for apidoc Am|bool|isOCTAL|char ch
2787a470 613Returns a boolean indicating whether the specified character is an
6aff1f14 614octal digit, [0-7].
243effed
KW
615The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
616C<isOCTAL>.
617
618=for apidoc Am|bool|isPUNCT|char ch
619Returns a boolean indicating whether the specified character is a
6aff1f14
KW
620punctuation character, analogous to C<m/[[:punct:]]/>.
621Note that the definition of what is punctuation isn't as
243effed
KW
622straightforward as one might desire. See L<perlrecharclass/POSIX Character
623Classes> for details.
624See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 625C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 626C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 627
0c82b6df 628=for apidoc Am|bool|isSPACE|char ch
2787a470 629Returns a boolean indicating whether the specified character is a
6aff1f14 630whitespace character. This is analogous
398d098a
KW
631to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
632(experimentally), this also matches what C<m/[[:space:]]/> does.
633("Experimentally" means that this change may be backed out in 5.20 or 5.22 if
634field experience indicates that it was unwise.) Prior to 5.18, only the
635locale forms of this macro (the ones with C<LC> in their names) matched
636precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
637in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
638(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
8a58bdcf 639See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 640C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 641C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 642
398d098a
KW
643=for apidoc Am|bool|isPSXSPC|char ch
644(short for Posix Space)
645Starting in 5.18, this is identical (experimentally) in all its forms to the
646corresponding C<isSPACE()> macros. ("Experimentally" means that this change
647may be backed out in 5.20 or 5.22 if field experience indicates that it
648was unwise.)
649The locale forms of this macro are identical to their corresponding
650C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
651non-locale forms differ from their C<isSPACE()> forms only in that the
652C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
653Otherwise they are identical. Thus this macro is analogous to what
654C<m/[[:space:]]/> matches in a regular expression.
655See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 656C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
657C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
658
954c1994 659=for apidoc Am|bool|isUPPER|char ch
2787a470 660Returns a boolean indicating whether the specified character is an
6aff1f14 661uppercase character, analogous to C<m/[[:upper:]]/>.
8a58bdcf 662See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 663C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 664C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 665
243effed 666=for apidoc Am|bool|isPRINT|char ch
8eea39dd 667Returns a boolean indicating whether the specified character is a
6aff1f14 668printable character, analogous to C<m/[[:print:]]/>.
8a58bdcf 669See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 670C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
671C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
672
673=for apidoc Am|bool|isWORDCHAR|char ch
674Returns a boolean indicating whether the specified character is a character
675that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
676in a regular expression. A word character is an alphabetic character, a
677decimal digit, a connecting punctuation character (such as an underscore), or
678a "mark" character that attaches to one of those (like some sort of accent).
679C<isALNUM()> is a synonym provided for backward compatibility, even though a
680word character includes more than the standard C language meaning of
681alphanumeric.
682See the L<top of this section|/Character classes> for an explanation of variants
683C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
684C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
685
686=for apidoc Am|bool|isXDIGIT|char ch
687Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
688digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
689and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
690See the L<top of this section|/Character classes> for an explanation of variants
691C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
692C<isXDIGIT_LC_utf8>.
693
3c3ecf18
KW
694=for apidoc Am|bool|isIDFIRST|char ch
695Returns a boolean indicating whether the specified character can be the first
696character of an identifier. This is very close to, but not quite the same as
697the official Unicode property C<XID_Start>. The difference is that this
698returns true only if the input character also matches L</isWORDCHAR>.
699See the L<top of this section|/Character classes> for an explanation of variants
700C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
701C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
702
703=for apidoc Am|bool|isIDCONT|char ch
704Returns a boolean indicating whether the specified character can be the
705second or succeeding character of an identifier. This is very close to, but
706not quite the same as the official Unicode property C<XID_Continue>. The
707difference is that this returns true only if the input character also matches
708L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
709explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
710C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
711C<isIDCONT_LC_utf8>.
712
243effed 713=head1 Miscellaneous Functions
8eea39dd 714
95a59cab 715=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 716Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
717Behaviour is only well defined when isXDIGIT(*str) is true.
718
e7c1e6c1
KW
719=head1 Character case changing
720
1f607577
KW
721=for apidoc Am|U8|toUPPER|U8 ch
722Converts the specified character to uppercase. If the input is anything but an
723ASCII lowercase character, that input character itself is returned. Variant
c753c8d3 724C<toUPPER_A> is equivalent.
954c1994 725
1f607577
KW
726=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
727Converts the Unicode code point C<cp> to its uppercase version, and
728stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
729that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
730bytes since the uppercase version may be longer than the original character.
731
732The first code point of the uppercased version is returned
733(but note, as explained just above, that there may be more.)
734
735=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
736Converts the UTF-8 encoded character at C<p> to its uppercase version, and
737stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
738that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
739bytes since the uppercase version may be longer than the original character.
740
741The first code point of the uppercased version is returned
742(but note, as explained just above, that there may be more.)
743
744The input character at C<p> is assumed to be well-formed.
745
25200305
KW
746=for apidoc Am|U8|toFOLD|U8 ch
747Converts the specified character to foldcase. If the input is anything but an
748ASCII uppercase character, that input character itself is returned. Variant
749C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
750Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
751
1f607577
KW
752=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
753Converts the Unicode code point C<cp> to its foldcase version, and
754stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
755that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
756bytes since the foldcase version may be longer than the original character.
757
758The first code point of the foldcased version is returned
759(but note, as explained just above, that there may be more.)
760
761=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
762Converts the UTF-8 encoded character at C<p> to its foldcase version, and
763stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
764that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
765bytes since the foldcase version may be longer than the original character.
766
767The first code point of the foldcased version is returned
768(but note, as explained just above, that there may be more.)
769
770The input character at C<p> is assumed to be well-formed.
771
772=for apidoc Am|U8|toLOWER|U8 ch
773Converts the specified character to lowercase. If the input is anything but an
774ASCII uppercase character, that input character itself is returned. Variant
c753c8d3 775C<toLOWER_A> is equivalent.
954c1994 776
1f607577
KW
777=for apidoc Am|U8|toLOWER_L1|U8 ch
778Converts the specified Latin1 character to lowercase. The results are undefined if
779the input doesn't fit in a byte.
780
781=for apidoc Am|U8|toLOWER_LC|U8 ch
782Converts the specified character to lowercase using the current locale's rules,
783if possible; otherwise returns the input character itself.
784
785=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
786Converts the Unicode code point C<cp> to its lowercase version, and
787stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
788that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
789bytes since the lowercase version may be longer than the original character.
790
791The first code point of the lowercased version is returned
792(but note, as explained just above, that there may be more.)
793
794=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
795Converts the UTF-8 encoded character at C<p> to its lowercase version, and
796stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
797that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
798bytes since the lowercase version may be longer than the original character.
799
800The first code point of the lowercased version is returned
801(but note, as explained just above, that there may be more.)
802
803The input character at C<p> is assumed to be well-formed.
804
805=for apidoc Am|U8|toLOWER_LC|U8 ch
806Converts the specified character to lowercase using the current locale's rules,
807if possible; otherwise returns the input character itself.
808
25200305
KW
809=for apidoc Am|U8|toTITLE|U8 ch
810Converts the specified character to titlecase. If the input is anything but an
811ASCII lowercase character, that input character itself is returned. Variant
812C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
813as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
814concept used in locale handling, so there is no functionality for that.)
815
1f607577
KW
816=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
817Converts the Unicode code point C<cp> to its titlecase version, and
818stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
819that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
820bytes since the titlecase version may be longer than the original character.
821
822The first code point of the titlecased version is returned
823(but note, as explained just above, that there may be more.)
824
825=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
826Converts the UTF-8 encoded character at C<p> to its titlecase version, and
827stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
828that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
829bytes since the titlecase version may be longer than the original character.
830
831The first code point of the titlecased version is returned
832(but note, as explained just above, that there may be more.)
833
834The input character at C<p> is assumed to be well-formed.
835
954c1994 836=cut
353c9b6f 837
1f607577 838XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
25200305 839really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
243effed 840
8a58bdcf 841Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
842patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
843
954c1994
GS
844*/
845
7c062697
KW
846/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
847 * is known only in the perl core, and this macro can be called from outside
848 * that */
849#ifdef HAS_QUAD
850# define WIDEST_UTYPE U64TYPE
851#else
852# define WIDEST_UTYPE U32
853#endif
854
3912bc88
KW
855/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
856 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
857 * bits of information are lost even on a 64-bit machine, but to get the
858 * compiler to optimize it out if possible. This is because Configure makes
859 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
860 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
861 */
862#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 863
41f43cc2 864#ifdef EBCDIC
b6340bd0
KW
865# ifndef _ALL_SOURCE
866 /* This returns the wrong results on at least z/OS unless this is
867 * defined. */
868# error _ALL_SOURCE should probably be defined
869# endif
870
871 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
872 * likely not defined, and so we use the native function */
873# define isASCII(c) isascii(c)
41f43cc2
KW
874#else
875# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
876#endif
877
ec79cfbb 878#define isASCII_A(c) isASCII(c)
3509bb53 879#define isASCII_L1(c) isASCII(c)
6515cb76 880
c2da0b36
KW
881/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
882 * and the 8 possible permutations of those bits exactly comprise the 8 octal
883 * digits */
3d33b3e3 884#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 885
ec79cfbb 886/* ASCII range only */
9fb1bf9d 887#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 888
a500dc72
KW
889/* Character class numbers. For internal core Perl use only. The ones less
890 * than 32 are used in PL_charclass[] and the ones up through the one that
891 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
892 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
893 * their actual definitions are here. If that file has a name not used here,
894 * it won't compile.
1709d539
KW
895 *
896 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
897 * frequency of their use. This gives a slight edge to exiting a loop earlier
898 * (in reginclass() in regexec.c) */
1709d539
KW
899# define _CC_WORDCHAR 0 /* \w and [:word:] */
900# define _CC_DIGIT 1 /* \d and [:digit:] */
901# define _CC_ALPHA 2 /* [:alpha:] */
902# define _CC_LOWER 3 /* [:lower:] */
903# define _CC_UPPER 4 /* [:upper:] */
904# define _CC_PUNCT 5 /* [:punct:] */
905# define _CC_PRINT 6 /* [:print:] */
15861f94 906# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 907# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 908# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 909
b0d691b2 910#define _FIRST_NON_SWASH_CC 10
1709d539
KW
911/* The character classes above are implemented with swashes. The second group
912 * (just below) contains the ones implemented without. These are also sorted
913 * in rough order of the frequency of their use, except that \v should be last,
914 * as it isn't a real Posix character class, and some (small) inefficiencies in
915 * regular expression handling would be introduced by putting it in the middle
916 * of those that are. Also, cntrl and ascii come after the others as it may be
917 * useful to group these which have no members that match above Latin1, (or
918 * above ASCII in the latter case) */
919
b0d691b2
KW
920# define _CC_SPACE 10 /* \s */
921# define _CC_BLANK 11 /* [:blank:] */
922# define _CC_XDIGIT 12 /* [:xdigit:] */
923# define _CC_PSXSPC 13 /* [:space:] */
924# define _CC_CNTRL 14 /* [:cntrl:] */
925# define _CC_ASCII 15 /* [:ascii:] */
926# define _CC_VERTSPACE 16 /* \v */
1709d539 927
a0947d7b
KW
928# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
929
1709d539 930/* The members of the third group below do not need to be coordinated with data
3ded5eb0
KW
931 * structures in regcomp.[ch] and regexec.c. But they should be added to
932 * bootstrap_ctype() */
b0d691b2
KW
933# define _CC_IDFIRST 17
934# define _CC_CHARNAME_CONT 18
935# define _CC_NONLATIN1_FOLD 19
936# define _CC_QUOTEMETA 20
937# define _CC_NON_FINAL_FOLD 21
938# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
939# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
940/* Unused: 23-30
f4cdb42c
KW
941 * If more bits are needed, one could add a second word for non-64bit
942 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
943 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
944 * is used only for optimization (as of this writing), and differs in the
945 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
946 * characters: the masculine and feminine ordinal indicators, so removing it
947 * would just cause /i regexes which match them to run less efficiently */
96ac0975 948
3a371f2f
KW
949#if defined(PERL_CORE) || defined(PERL_EXT)
950/* An enum version of the character class numbers, to help compilers
951 * optimize */
952typedef enum {
3a371f2f 953 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
954 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
955 _CC_ENUM_ASCII = _CC_ASCII,
956 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 957 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 958 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
959 _CC_ENUM_DIGIT = _CC_DIGIT,
960 _CC_ENUM_GRAPH = _CC_GRAPH,
961 _CC_ENUM_LOWER = _CC_LOWER,
962 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 963 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 964 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 965 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 966 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 967 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 968 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 969 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
970} _char_class_number;
971#endif
972
63c61c3f 973#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 974#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 975
3018b823 976#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 977# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
978 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
979 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
980 #error Need to adjust order of swash_property_names[]
981# endif
982
b44daeeb
KW
983/* This is declared static in each of the few files that this is #defined for
984 * to keep them from being publicly accessible. Hence there is a small amount
985 * of wasted space */
986
922e8cb4
KW
987static const char* const swash_property_names[] = {
988 "XPosixWord",
989 "XPosixDigit",
990 "XPosixAlpha",
991 "XPosixLower",
992 "XPosixUpper",
993 "XPosixPunct",
994 "XPosixPrint",
995 "XPosixAlnum",
b0d691b2
KW
996 "XPosixGraph",
997 "Cased"
922e8cb4
KW
998};
999#endif
1000
96ac0975
NC
1001# ifdef DOINIT
1002EXTCONST U32 PL_charclass[] = {
1003# include "l1_char_class_tab.h"
1004};
1005
1006# else /* ! DOINIT */
1007EXTCONST U32 PL_charclass[];
1008# endif
3ded5eb0
KW
1009#endif /* Has perl.h */
1010
1011#if defined(H_PERL) && ! defined(BOOTSTRAP_CHARSET)
96ac0975 1012
265c1f46 1013 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70
KW
1014# define _CC_mask(classnum) (1U << (classnum))
1015# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
f4cd282c 1016 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
4eeeb416 1017
f4cdb42c
KW
1018 /* The mask for the _A versions of the macros; it just adds in the bit for
1019 * ASCII. */
1020# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1021
1022 /* The _A version makes sure that both the desired bit and the ASCII bit
1023 * are present */
3ded5eb0 1024# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
f4cd282c 1025 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
f4cdb42c
KW
1026 == _CC_mask_A(classnum)))
1027
f4cdb42c 1028# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 1029# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
1030# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1031# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1032# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
1033# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1034# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1035# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1036# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1037# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1038# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1039# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1040# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1041# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
d95f8b6a 1042# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
3ded5eb0
KW
1043# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1044# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1045# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1046
1047 /* continuation character for legal NAME in \N{NAME} */
1048# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1049
1050# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1051# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1052# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1053# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1054# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1055# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1056# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1057# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1058# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1059# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
f4cdb42c 1060
fb9bfbf7
KW
1061 /* Either participates in a fold with a character above 255, or is a
1062 * multi-char fold */
f4cd282c 1063# 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 1064
4eeeb416 1065# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
1066# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1067 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
1068# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1069 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
3ded5eb0
KW
1070#else /* Either don't have perl.h or don't want to use char_class_tab.h */
1071
1072 /* If we don't have perl.h, we are compiling a utility program. Below we
1073 * hard-code various macro definitions that wouldn't otherwise be available
1074 * to it. We can also get here if we are configured to bootstrap up Perl
1075 * on a non-ASCII platform that doesn't have a working Perl (currently only
1076 * EBCDIC). For these we currently use the native definitions to get
1077 * things going. (It should also be possible to use the translation
1078 * function NATIVE_TO_LATIN1(), but that is an extra layer of dependence on
1079 * Perl, so it is currently avoided for the macros where it's possible to
1080 * do so.) */
9fb1bf9d 1081# ifdef EBCDIC
3ded5eb0 1082 /* Use the native functions. They likely will return false for all
4f48cecf 1083 * non-ASCII values, but this makes sure */
b6340bd0
KW
1084# define isALPHA_A(c) (isASCII(c) && isalpha(c))
1085# define isALPHANUMERIC_A(c) (isASCII(c) && isalnum(c))
1086# define isCNTRL_A(c) (isASCII(c) && iscntrl(c))
1087# define isDIGIT_A(c) (isASCII(c) && isdigit(c))
1088# define isGRAPH_A(c) (isASCII(c) && isgraph(c))
1089# define isLOWER_A(c) (isASCII(c) && islower(c))
1090# define isPRINT_A(c) (isASCII(c) && isprint(c))
1091# define isPUNCT_A(c) (isASCII(c) && ispunct(c))
1092# define isSPACE_A(c) (isASCII(c) && isspace(c))
1093# define isUPPER_A(c) (isASCII(c) && isupper(c))
1094# define isXDIGIT_A(c) (isASCII(c) && isxdigit(c))
3ded5eb0 1095# else /* ASCII platform. These are coded based on first principals */
9fb1bf9d 1096# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
15861f94 1097# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
3ded5eb0 1098# define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
2181da29 1099# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
3ded5eb0 1100# define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
9fb1bf9d
KW
1101# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1102# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
3ded5eb0
KW
1103# define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1104# define isSPACE_A(c) ((c) == ' ' \
1105 || (c) == '\t' \
1106 || (c) == '\n' \
1107 || (c) == '\r' \
1108 || (c) =='\v' \
926bb54c 1109 || (c) == '\f')
2181da29 1110# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
3ded5eb0
KW
1111# define isXDIGIT_A(c) (isDIGIT_A(c) \
1112 || ((c) >= 'a' && (c) <= 'f') \
1113 || ((c) <= 'F' && (c) >= 'A'))
1114# endif /* Below are common definitions for ASCII and non-ASCII */
1115# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1116# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1117# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1118
1119 /* The _L1 macros may be unnecessary for both the utilities and for
1120 * bootstrapping; I (khw) added them during debugging of bootstrapping, and
1121 * it seems best to keep them. */
1122# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1123# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1124# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1125# define isBLANK_L1(c) (isBLANK_A(c) \
1126 || (FITS_IN_8_BITS(c) \
1127 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1128# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1129# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1130# define isLOWER_L1(c) (isLOWER_A(c) \
1131 || (FITS_IN_8_BITS(c) \
1132 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1133 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1134 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1135 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1136 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1137# define isPRINT_L1(c) (isPRINT_A(c) \
1138 || (FITS_IN_8_BITS(c) \
1139 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1140# define isPSXSPC_L1(c) isSPACE_L1(c)
1141# define isPUNCT_L1(c) (isPUNCT_A(c) \
1142 || (FITS_IN_8_BITS(c) \
1143 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1144 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1145 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1146 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1147 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1148 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1149 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1150# define isSPACE_L1(c) (isSPACE_A(c) \
1151 || (FITS_IN_8_BITS(c) \
1152 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1153 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1154# define isUPPER_L1(c) (isUPPER_A(c) \
1155 || (FITS_IN_8_BITS(c) \
1156 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1157 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1158 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1159# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1160# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1161# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1162 || isBLANK_L1(c) \
1163 || (c) == '-' \
1164 || (c) == '(' \
1165 || (c) == ')')
1166 /* The following are not fully accurate in the above-ASCII range. I (khw)
1167 * don't think it's necessary to be so for the purposes where this gets
1168 * compiled */
1169# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1170# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1171
1172 /* And these aren't accurate at all. They are useful only for above
1173 * Latin1, which utilities and bootstrapping don't deal with */
1174# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1175# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1176
1177 /* Many of the macros later in this file are defined in terms of these. By
1178 * implementing them with a function, which converts the class number into
1179 * a call to the desired macro, all of the later ones work. However, that
1180 * function won't be actually defined when building a utility program (no
1181 * perl.h), and so a compiler error will be generated if one is attempted
1182 * to be used. And the above-Latin1 code points require Unicode tables to
1183 * be present, something unlikely to be the case when bootstrapping */
1184# define _generic_isCC(c, classnum) \
1185 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1186# define _generic_isCC_A(c, classnum) \
1187 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
1188#endif /* End of no perl.h or have BOOTSTRAP_CHARSET */
8a58bdcf 1189
e66b99e9
KW
1190#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1191#define isALPHA(c) isALPHA_A(c)
1192#define isBLANK(c) isBLANK_A(c)
1193#define isCNTRL(c) isCNTRL_A(c)
1194#define isDIGIT(c) isDIGIT_A(c)
1195#define isGRAPH(c) isGRAPH_A(c)
1196#define isIDFIRST(c) isIDFIRST_A(c)
1197#define isLOWER(c) isLOWER_A(c)
1198#define isPRINT(c) isPRINT_A(c)
1199#define isPSXSPC(c) isPSXSPC_A(c)
1200#define isPUNCT(c) isPUNCT_A(c)
1201#define isSPACE(c) isSPACE_A(c)
1202#define isUPPER(c) isUPPER_A(c)
1203#define isWORDCHAR(c) isWORDCHAR_A(c)
1204#define isXDIGIT(c) isXDIGIT_A(c)
1205
1206/* ASCII casing. These could also be written as
1207 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1208 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1209 which uses table lookup and mask instead of subtraction. (This would
1210 work because the _MOD does not apply in the ASCII range) */
68067e4e
DM
1211#define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1212#define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
bbce6d69 1213
25200305
KW
1214/* In the ASCII range, these are equivalent to what they're here defined to be.
1215 * But by creating these definitions, other code doesn't have to be aware of
1216 * this detail */
1217#define toFOLD(c) toLOWER(c)
1218#define toFOLD_LC(c) toLOWER_LC(c)
1219#define toTITLE(c) toUPPER(c)
1220
c753c8d3
KW
1221#define toLOWER_A(c) toLOWER(c)
1222#define toUPPER_A(c) toUPPER(c)
25200305
KW
1223#define toFOLD_A(c) toFOLD(c)
1224#define toTITLE_A(c) toTITLE(c)
1a0901db
KW
1225
1226/* Use table lookup for speed; return error character for input
1227 * out-of-range */
b2bf251f 1228#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1229 ? (c) \
f4cd282c 1230 : PL_latin1_lc[ (U8) (c) ])
c753c8d3
KW
1231#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1232
1a0901db
KW
1233/* Modified uc. Is correct uc except for three non-ascii chars which are
1234 * all mapped to one of them, and these need special handling; error
1235 * character for input out-of-range */
b2bf251f 1236#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1237 ? (c) \
f4cd282c 1238 : PL_mod_latin1_uc[ (U8) (c) ])
bbce6d69
PP
1239#ifdef USE_NEXT_CTYPE
1240
15861f94 1241# define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c))
ff68c719 1242# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 1243# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 1244# define isBLANK_LC(c) isBLANK((unsigned int)(c))
b8c5462f 1245# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
509fb054 1246# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
b8c5462f 1247# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
509fb054
KW
1248# define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
1249# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
ff68c719 1250# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 1251# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
509fb054
KW
1252# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
1253# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
1254# define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
e712593e 1255# define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
ff68c719 1256# define toLOWER_LC(c) NXToLower((unsigned int)(c))
509fb054 1257# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
bbce6d69
PP
1258
1259#else /* !USE_NEXT_CTYPE */
b8c5462f 1260
bbce6d69
PP
1261# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1262
76fde96a 1263/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
851131d3 1264
76fde96a 1265# define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
15861f94
KW
1266# define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \
1267 && isalnum((unsigned char)(c)))
07315176 1268# ifdef HAS_ISASCII
76fde96a 1269# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
07315176 1270# else
76fde96a 1271# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
07315176 1272# endif
67addccf 1273# ifdef HAS_ISBLANK
76fde96a 1274# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
67addccf 1275# else
76fde96a 1276# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
67addccf 1277# endif
76fde96a 1278# define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
509fb054 1279# define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
76fde96a 1280# define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
509fb054
KW
1281# define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1282 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1283# define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
76fde96a
KW
1284# define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1285# define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
509fb054
KW
1286# define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1287# define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1288# define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1289 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
e712593e 1290# define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
548d896a
DM
1291# define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? (UV)tolower((unsigned char)(c)) : (c))
1292# define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? (UV)toupper((unsigned char)(c)) : (c))
bbce6d69
PP
1293
1294# else
1295
bbce6d69 1296# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1297# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
07315176 1298# define isASCII_LC(c) isascii(c)
67addccf
KW
1299# ifdef HAS_ISBLANK
1300# define isBLANK_LC(c) (isascii(c) && isblank(c))
1301# else
3e66ab74 1302# define isBLANK_LC(c) isBLANK_A(c)
67addccf 1303# endif
b8c5462f 1304# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1305# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1306# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1307# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1308# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1309# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1310# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1311# define isSPACE_LC(c) (isascii(c) && isspace(c))
1312# define isUPPER_LC(c) (isascii(c) && isupper(c))
1313# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1314# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
76fde96a 1315# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1316# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
bbce6d69
PP
1317
1318# endif
a0d0e21e 1319#endif /* USE_NEXT_CTYPE */
55204971 1320
eba68aa0
KW
1321#define isIDCONT(c) isWORDCHAR(c)
1322#define isIDCONT_A(c) isWORDCHAR_A(c)
1323#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1324#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1325#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1326
66c17564 1327/* For internal core Perl use only. If the input is Latin1, use the Latin1
a500dc72
KW
1328 * macro; otherwise use the function 'above_latin1'. Won't compile if 'c' isn't unsigned, as
1329 * won't match above_latin1 prototype. The macros do bounds checking, so have
66c17564
KW
1330 * duplicate checks here, so could create versions of the macros that don't,
1331 * but experiments show that gcc optimizes them out anyway. */
1332
1333/* Note that all ignore 'use bytes' */
cd500f2f
KW
1334#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1335 ? _generic_isCC(c, classnum) \
1336 : above_latin1(c))
1337#define _generic_swash_uni(classnum, c) ((c) < 256 \
1338 ? _generic_isCC(c, classnum) \
922e8cb4 1339 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1340#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1341#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1342#define isASCII_uni(c) isASCII(c)
1343#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1344#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1345#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1346#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1347#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1348#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1349#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1350#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1351
509fb054
KW
1352/* Posix and regular space are identical above Latin1 */
1353#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1354
cd500f2f 1355#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1356#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1357#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1358#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1359#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1360#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1361
b8d68ded 1362#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1363#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1364#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1365#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1366
cd500f2f
KW
1367#define _generic_LC_uvchr(latin1, above_latin1, c) \
1368 (c < 256 ? latin1(c) : above_latin1(c))
1369#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1370 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1371
1372#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1373#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1374 _CC_ALPHANUMERIC, c)
509fb054 1375#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1376#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1377#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1378#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1379#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1380#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1381 _is_uni_perl_idcont, c)
cd500f2f
KW
1382#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1383 _is_uni_perl_idstart, c)
1384#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1385#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1386#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1387 space under locale */
cd500f2f
KW
1388#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1389#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1390 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1391#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1392#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1393 _CC_WORDCHAR, c)
1394#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1395
36b00f93 1396#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1397
a500dc72
KW
1398/* Everything whose name begins with an underscore is for internal core Perl
1399 * use only. */
1400
1401/* If the input is in the Latin1 range, use
252810d0
KW
1402 * the Latin1 macro 'classnum' on 'p' which is a pointer to a UTF-8 string.
1403 * Otherwise use the value given by the 'utf8' parameter. This relies on the
1404 * fact that ASCII characters have the same representation whether utf8 or not.
1405 * Note that it assumes that the utf8 has been validated, and ignores 'use
1406 * bytes' */
cd500f2f 1407#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1408 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1409 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c 1410 ? _generic_isCC( \
94bb8c36 1411 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
aea0c08c 1412 *((p)+1 )), \
94bb8c36 1413 classnum) \
252810d0 1414 : utf8)
cd500f2f
KW
1415/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1416 * can be a macro */
1417#define _generic_func_utf8(classnum, above_latin1, p) \
1418 _generic_utf8(classnum, p, above_latin1(p))
1419/* Like the above, but passes classnum to _isFOO_utf8(), instead of having a
1420 * 'above_latin1' parameter */
922e8cb4 1421#define _generic_swash_utf8(classnum, p) \
cd500f2f 1422 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1423
cc8ab7c0
KW
1424/* Like the above, but should be used only when it is known that there are no
1425 * characters in the range 128-255 which the class is TRUE for. Hence it can
cd500f2f 1426 * skip the tests for this range. 'above_latin1' should include its arguments */
3ded5eb0 1427#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1428 (UTF8_IS_INVARIANT(*(p)) \
1429 ? _generic_isCC(*(p), classnum) \
1430 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
3ded5eb0 1431 ? above_latin1 \
cc8ab7c0
KW
1432 : 0)
1433
84238efa
KW
1434/* NOTE that some of these macros have very similar ones in regcharclass.h.
1435 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1436 * there, differing in name only by an underscore from the one here
1437 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1438 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1439 * points; the regcharclass.h ones are implemented as a series of
1440 * "if-else-if-else ..." */
1441
922e8cb4
KW
1442#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1443#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1444#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1445 utf8, the non-utf8 macro works
1446 */
cd500f2f
KW
1447#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1448#define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1449#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1450 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1451#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1452#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1453 _is_utf8_perl_idcont, p)
e5dcd934 1454
c11ff943
KW
1455/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1456 * IDFIRST is an alnum. See
c8362b00 1457 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1458 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1459 * This used to be not the XID version, but we decided to go with the more
1460 * modern Unicode definition */
cd500f2f 1461#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1462 _is_utf8_perl_idstart, p)
e5dcd934 1463
922e8cb4
KW
1464#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1465#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1466
509fb054 1467/* Posix and regular space are identical above Latin1 */
cd500f2f 1468#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1469
922e8cb4 1470#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1471#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1472#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1473#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1474#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1475#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1476 is_XDIGIT_high(p))
a0ed51b3 1477
7d2b61dd 1478#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
509fb054
KW
1479#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1480#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1481#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1482
9201a572
KW
1483/* For internal core Perl use only. If the input is in the Latin1 range, use
1484 * the macro 'macro' on 'p' which is a pointer to a UTF-8 string. Otherwise
1485 * use the value given by the 'utf8' parameter. This relies on the fact that
1486 * ASCII characters have the same representation whether utf8 or not. Note
1487 * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
94bb8c36
KW
1488#define _generic_LC_utf8(macro, p, utf8) \
1489 (UTF8_IS_INVARIANT(*(p)) \
1490 ? macro(*(p)) \
1491 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1492 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
9201a572
KW
1493 : utf8)
1494
cd500f2f
KW
1495#define _generic_LC_swash_utf8(macro, classnum, p) \
1496 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1497#define _generic_LC_func_utf8(macro, above_latin1, p) \
1498 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1499
cd500f2f
KW
1500#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1501 _CC_ALPHANUMERIC, p)
1502#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1503#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1504#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1505#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1506#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1507#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1508#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1509#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1510#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1511#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1512#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1513 space under locale */
cd500f2f
KW
1514#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1515#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1516#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1517#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1518 _CC_WORDCHAR, p)
1519#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1520
fbc19f27
KW
1521/* Macros for backwards compatibility and for completeness when the ASCII and
1522 * Latin1 values are identical */
fbc19f27
KW
1523#define isALPHAU(c) isALPHA_L1(c)
1524#define isDIGIT_L1(c) isDIGIT_A(c)
1525#define isOCTAL(c) isOCTAL_A(c)
1526#define isOCTAL_L1(c) isOCTAL_A(c)
1527#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1528#define isALNUM(c) isWORDCHAR(c)
1529#define isALNUMU(c) isWORDCHAR_L1(c)
1530#define isALNUM_LC(c) isWORDCHAR_LC(c)
1531#define isALNUM_uni(c) isWORDCHAR_uni(c)
1532#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1533#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1534#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1535#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1536#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1537#define isALNUMC(c) isALPHANUMERIC(c)
1538#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1539#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1540#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1541#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1542#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1543
6ae9f32a 1544/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1545 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1546 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1547 * that is EOT on ASCII; ST on EBCDIC */
8e7c6e7d 1548# define toCTRL(c) (toUPPER(NATIVE_TO_LATIN1(c)) ^ 64)
bbce6d69 1549
dea28490
JJ
1550/* Line numbers are unsigned, 32 bits. */
1551typedef U32 line_t;
e5dcd934 1552#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1553
91152fc1
DG
1554/* Helpful alias for version prescan */
1555#define is_LAX_VERSION(a,b) \
1556 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1557
1558#define is_STRICT_VERSION(a,b) \
1559 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1560
1561#define BADVERSION(a,b,c) \
1562 if (b) { \
1563 *b = c; \
1564 } \
1565 return a;
8c52afec 1566
14f6cdfd
KW
1567/* Converts a hex digit in a string to its numeric value, advancing the
1568 * pointer. The input must be known to be 0-9, A-F, or a-f. In both ASCII and
1569 * EBCDIC the last 4 bits of the digits are 0-9; and the last 4 bits of A-F and
1570 * a-f are 1-6, so adding 9 yields 10-15 */
902008b8
KW
1571#define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \
1572 ? (*(s)++) \
1573 : (*(s)++ + 9))))
95a59cab 1574
8e84507e 1575/*
ccfc67b7
JH
1576=head1 Memory Management
1577
a02a5408 1578=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1579The XSUB-writer's interface to the C C<malloc> function.
1580
0d7b2759
MH
1581Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1582
c5008215
JC
1583In 5.9.3, Newx() and friends replace the older New() API, and drops
1584the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1585themselves. This aid has been superseded by a new build option,
d10b4965 1586PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1587there for use in XS modules supporting older perls.
1588
a02a5408 1589=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1590The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1591cast. See also C<Newx>.
954c1994 1592
0d7b2759
MH
1593Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1594
a02a5408 1595=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1596The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1597memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1598
0d7b2759
MH
1599Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1600
954c1994
GS
1601=for apidoc Am|void|Renew|void* ptr|int nitems|type
1602The XSUB-writer's interface to the C C<realloc> function.
1603
0d7b2759
MH
1604Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1605
954c1994
GS
1606=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1607The XSUB-writer's interface to the C C<realloc> function, with
1608cast.
1609
0d7b2759
MH
1610Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1611
49b8b560 1612=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1613The XSUB-writer's interface to the C C<free> function.
1614
0d7b2759
MH
1615This should B<ONLY> be used on memory obtained using L<"Newx"> and friends.
1616
954c1994
GS
1617=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1618The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1619source, C<dest> is the destination, C<nitems> is the number of items, and
1620C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1621
e90e2364 1622=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
72d33970
FC
1623Like C<Move> but returns dest. Useful
1624for encouraging compilers to tail-call
e90e2364
NC
1625optimise.
1626
954c1994
GS
1627=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1628The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1629source, C<dest> is the destination, C<nitems> is the number of items, and
1630C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1631
e90e2364
NC
1632=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1633
72d33970
FC
1634Like C<Copy> but returns dest. Useful
1635for encouraging compilers to tail-call
e90e2364
NC
1636optimise.
1637
954c1994
GS
1638=for apidoc Am|void|Zero|void* dest|int nitems|type
1639
1640The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1641destination, C<nitems> is the number of items, and C<type> is the type.
1642
e90e2364
NC
1643=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1644
72d33970
FC
1645Like C<Zero> but returns dest. Useful
1646for encouraging compilers to tail-call
e90e2364
NC
1647optimise.
1648
da5d8dbb 1649=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1650This is an architecture-independent macro to copy one structure to another.
954c1994 1651
7e337ee0
JH
1652=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1653
1654Fill up memory with a byte pattern (a byte repeated over and over
1655again) that hopefully catches attempts to access uninitialized memory.
1656
1657=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1658
1659PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1660
1c12ffb4 1661=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1662
1663PoisonWith(0xEF) for catching access to freed memory.
1664
9965345d
JH
1665=for apidoc Am|void|Poison|void* dest|int nitems|type
1666
7e337ee0 1667PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1668
1669=cut */
954c1994 1670
561b68a9
SH
1671/* Maintained for backwards-compatibility only. Use newSV() instead. */
1672#ifndef PERL_CORE
ff06c60c 1673#define NEWSV(x,len) newSV(len)
561b68a9 1674#endif
ff06c60c 1675
19a94d75
JH
1676#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1677
2720f793
JH
1678/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1679 * overly eager compilers that will bleat about e.g.
1680 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1681#ifdef PERL_MALLOC_WRAP
f019c49e 1682#define MEM_WRAP_CHECK(n,t) \
d1decf2b 1683 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
27d5b266 1684#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1685 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1686#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1687
d1decf2b 1688#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
1689#else
1690
410319be
NC
1691#define MEM_WRAP_CHECK(n,t)
1692#define MEM_WRAP_CHECK_1(n,t,a)
1693#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1694#define MEM_WRAP_CHECK_(n,t)
1695
6fff79ce 1696#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1697
1936d2a7 1698#endif
8b44ba4c 1699
fe4f188c 1700#ifdef PERL_MEM_LOG
46c6c7e2 1701/*
9f653bb5 1702 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1703 * go through functions, which are handy for debugging breakpoints, but
1704 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1705 * line number, and C function name if available) passed in. This info can
1706 * then be used for logging the calls, for which one gets a sample
73d1d973 1707 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1708 *
46c6c7e2 1709 * Known problems:
94e892a6 1710 * - not all memory allocs get logged, only those
46c6c7e2 1711 * that go through Newx() and derivatives (while all
94e892a6 1712 * Safefrees do get logged)
46c6c7e2
JH
1713 * - __FILE__ and __LINE__ do not work everywhere
1714 * - __func__ or __FUNCTION__ even less so
1715 * - I think more goes on after the perlio frees but
1716 * the thing is that STDERR gets closed (as do all
1717 * the file descriptors)
1718 * - no deeper calling stack than the caller of the Newx()
1719 * or the kind, but do I look like a C reflection/introspection
1720 * utility to you?
1721 * - the function prototypes for the logging functions
1722 * probably should maybe be somewhere else than handy.h
1723 * - one could consider inlining (macrofying) the logging
1724 * for speed, but I am too lazy
1725 * - one could imagine recording the allocations in a hash,
1726 * (keyed by the allocation address?), and maintain that
1727 * through reallocs and frees, but how to do that without
1728 * any News() happening...?
73d1d973 1729 * - lots of -Ddefines to get useful/controllable output
b953482e 1730 * - lots of ENV reads
46c6c7e2
JH
1731 */
1732
12754f92 1733PERL_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 1734
12754f92 1735PERL_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 1736
12754f92 1737PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1738
0b0ab801 1739# ifdef PERL_CORE
73d1d973 1740# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1741enum mem_log_type {
1742 MLT_ALLOC,
1743 MLT_REALLOC,
d7a2c63c
MHM
1744 MLT_FREE,
1745 MLT_NEW_SV,
1746 MLT_DEL_SV
0b0ab801
MHM
1747};
1748# endif
12754f92 1749# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1750void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1751void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1752# endif
0b0ab801
MHM
1753# endif
1754
fe4f188c
JH
1755#endif
1756
1757#ifdef PERL_MEM_LOG
d1401ee9
MHM
1758#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1759#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1760#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1761#endif
1762
1763#ifndef MEM_LOG_ALLOC
1764#define MEM_LOG_ALLOC(n,t,a) (a)
1765#endif
1766#ifndef MEM_LOG_REALLOC
1767#define MEM_LOG_REALLOC(n,t,v,a) (a)
1768#endif
1769#ifndef MEM_LOG_FREE
1770#define MEM_LOG_FREE(a) (a)
1771#endif
1772
d1401ee9
MHM
1773#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1774#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1775#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1776
1777#ifndef PERL_CORE
a02a5408
JC
1778/* pre 5.9.x compatibility */
1779#define New(x,v,n,t) Newx(v,n,t)
1780#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1781#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1782#endif
a02a5408 1783
ff68c719 1784#define Renew(v,n,t) \
d1401ee9 1785 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1786#define Renewc(v,n,t,c) \
d1401ee9 1787 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1788
1789#ifdef PERL_POISON
1790#define Safefree(d) \
06c0cc96 1791 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1792#else
fe4f188c 1793#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1794#endif
55497cff 1795
8b44ba4c
NC
1796#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1797#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1798#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1799
8b44ba4c
NC
1800#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1801#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1802#ifdef HAS_MEMSET
8b44ba4c 1803#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1804#else
8b44ba4c
NC
1805/* Using bzero(), which returns void. */
1806#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1807#endif
1808
7e337ee0
JH
1809#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1810#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1811#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1812#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1813
caa674f3
DD
1814#ifdef PERL_POISON
1815# define PERL_POISON_EXPR(x) x
1816#else
1817# define PERL_POISON_EXPR(x)
1818#endif
1819
2304df62 1820#ifdef USE_STRUCT_COPY
ff68c719 1821#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1822#else
1823#define StructCopy(s,d,t) Copy(s,d,1,t)
1824#endif
2cc61e15 1825
622913ab 1826#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1827#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1828
2cc61e15
DD
1829#ifdef NEED_VA_COPY
1830# ifdef va_copy
1831# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1832# else
a1866d1b
JH
1833# if defined(__va_copy)
1834# define Perl_va_copy(s, d) __va_copy(d, s)
1835# else
1836# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1837# endif
2cc61e15
DD
1838# endif
1839#endif
1840
472d47bc
SB
1841/* convenience debug macros */
1842#ifdef USE_ITHREADS
1843#define pTHX_FORMAT "Perl interpreter: 0x%p"
1844#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1845#define pTHX_VALUE_ (void *)my_perl,
1846#define pTHX_VALUE (void *)my_perl
1847#define pTHX__VALUE_ ,(void *)my_perl,
1848#define pTHX__VALUE ,(void *)my_perl
472d47bc 1849#else
3609ea0d 1850#define pTHX_FORMAT
472d47bc 1851#define pTHX__FORMAT
3609ea0d 1852#define pTHX_VALUE_
472d47bc 1853#define pTHX_VALUE
3609ea0d 1854#define pTHX__VALUE_
472d47bc
SB
1855#define pTHX__VALUE
1856#endif /* USE_ITHREADS */
3609ea0d 1857
2acdbac1
NC
1858/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1859 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1860 CPAN::Unpack show any users outside the core. */
1861#ifdef PERL_CORE
d1d15184 1862# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1863#endif
1864
dfff4baf
BF
1865/* Internal macros to deal with gids and uids */
1866#ifdef PERL_CORE
1867
1868# if Uid_t_size > IVSIZE
1869# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
1870# define SvUID(sv) SvNV(sv)
1871# else
1872# if Uid_t_sign <= 0
1873# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
1874# define SvUID(sv) SvIV(sv)
1875# else
1876# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
1877# define SvUID(sv) SvUV(sv)
1878# endif
1879# endif /* Uid_t_size */
1880
1881# if Gid_t_size > IVSIZE
1882# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
1883# define SvGID(sv) SvNV(sv)
1884# else
1885# if Gid_t_sign <= 0
1886# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
1887# define SvGID(sv) SvIV(sv)
1888# else
1889# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
1890# define SvGID(sv) SvUV(sv)
1891# endif
1892# endif /* Gid_t_size */
1893
1894#endif
1895
9d745869
KW
1896#endif /* HANDY_H */
1897
e9a8c099
MHM
1898/*
1899 * Local variables:
1900 * c-indentation-style: bsd
1901 * c-basic-offset: 4
14d04a33 1902 * indent-tabs-mode: nil
e9a8c099
MHM
1903 * End:
1904 *
14d04a33 1905 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1906 */