This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
handy.h: White-space, comments, pod nit only
[perl5.git] / handy.h
CommitLineData
a0d0e21e 1/* handy.h
a687059c 2 *
1129b882 3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000,
da5d8dbb 4 * 2001, 2002, 2004, 2005, 2006, 2007, 2008, 2012 by Larry Wall and others
a687059c 5 *
6e21c824
LW
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
8d063cd8 8 *
8d063cd8
LW
9 */
10
4650c663
KW
11/* IMPORTANT NOTE: Everything whose name begins with an underscore is for
12 * internal core Perl use only. */
13
9d745869
KW
14#ifndef HANDY_H /* Guard against nested #inclusion */
15#define HANDY_H
16
85e6fe83 17#if !defined(__STDC__)
378cc40b
LW
18#ifdef NULL
19#undef NULL
20#endif
a687059c 21# define NULL 0
85e6fe83
LW
22#endif
23
24792b8d
NC
24#ifndef PERL_CORE
25# define Null(type) ((type)NULL)
954c1994
GS
26
27/*
ccfc67b7 28=head1 Handy Values
954c1994 29
3609ea0d 30=for apidoc AmU||Nullch
72d33970
FC
31Null character pointer. (No longer available when C<PERL_CORE> is
32defined.)
2307c6d0 33
954c1994 34=for apidoc AmU||Nullsv
72d33970 35Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
36
37=cut
38*/
39
24792b8d
NC
40# define Nullch Null(char*)
41# define Nullfp Null(PerlIO*)
42# define Nullsv Null(SV*)
43#endif
8d063cd8 44
641d3f0b
PP
45#ifdef TRUE
46#undef TRUE
47#endif
48#ifdef FALSE
49#undef FALSE
50#endif
51#define TRUE (1)
52#define FALSE (0)
53
cf3f0ffb
DM
54/* The MUTABLE_*() macros cast pointers to the types shown, in such a way
55 * (compiler permitting) that casting away const-ness will give a warning;
56 * e.g.:
57 *
58 * const SV *sv = ...;
59 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
60 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
61 */
62
b1bc3f34 63#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 64# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
65#else
66# define MUTABLE_PTR(p) ((void *) (p))
67#endif
68
a062e10d 69#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 70#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 71#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 72#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 73#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 74#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 75
f789f6a4 76#if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR)
bd31be4b
NC
77# include <stdbool.h>
78# ifndef HAS_BOOL
79# define HAS_BOOL 1
80# endif
81#endif
82
8e84507e 83/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
84 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
85 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 86 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
87 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
88 g++ can be identified by __GNUG__.
89 Andy Dougherty February 2000
5d94fbed 90*/
3609ea0d 91#ifdef __GNUG__ /* GNU g++ has bool built-in */
f789f6a4 92# ifndef PERL_BOOL_AS_CHAR
5d94fbed 93# ifndef HAS_BOOL
c1d22f6b 94# define HAS_BOOL 1
5d94fbed 95# endif
f789f6a4 96# endif
5d94fbed
AD
97#endif
98
641d3f0b
PP
99/* The NeXT dynamic loader headers will not build with the bool macro
100 So declare them now to clear confusion.
101*/
8f1f23e8 102#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
103# undef FALSE
104# undef TRUE
105 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
106# define ENUM_BOOL 1
107# ifndef HAS_BOOL
108# define HAS_BOOL 1
109# endif /* !HAS_BOOL */
8f1f23e8 110#endif /* NeXT || __NeXT__ */
641d3f0b 111
5d94fbed 112#ifndef HAS_BOOL
f789f6a4
FC
113# ifdef bool
114# undef bool
115# endif
70d5cb32 116# define bool char
c1d22f6b 117# define HAS_BOOL 1
a687059c 118#endif
0d3e774c 119
236415e2
KW
120/* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
121 * defined as char for example, then the cast from int is
122 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
123 * AIX */
18f5643b 124#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 125
46c6c7e2 126/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
127 * XXX Should really be a Configure probe, with HAS__FUNCTION__
128 * and FUNCTION__ as results.
129 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
130#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
131# define FUNCTION__ __func__
132#else
9ea40801 133# if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 134 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
46c6c7e2
JH
135# define FUNCTION__ ""
136# else
137# define FUNCTION__ __FUNCTION__ /* Common extension. */
138# endif
139#endif
140
27d4fb96
PP
141/* XXX A note on the perl source internal type system. The
142 original intent was that I32 be *exactly* 32 bits.
143
144 Currently, we only guarantee that I32 is *at least* 32 bits.
145 Specifically, if int is 64 bits, then so is I32. (This is the case
146 for the Cray.) This has the advantage of meshing nicely with
147 standard library calls (where we pass an I32 and the library is
148 expecting an int), but the disadvantage that an I32 is not 32 bits.
149 Andy Dougherty August 1996
24fef2a7 150
dc45a647
MB
151 There is no guarantee that there is *any* integral type with
152 exactly 32 bits. It is perfectly legal for a system to have
153 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 154
dc45a647
MB
155 Similarly, there is no guarantee that I16 and U16 have exactly 16
156 bits.
693762b4 157
8e84507e
NIS
158 For dealing with issues that may arise from various 32/64-bit
159 systems, we will ask Configure to check out
8175356b 160
3609ea0d
JH
161 SHORTSIZE == sizeof(short)
162 INTSIZE == sizeof(int)
163 LONGSIZE == sizeof(long)
dc45a647 164 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 165 PTRSIZE == sizeof(void *)
dc45a647
MB
166 DOUBLESIZE == sizeof(double)
167 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 168
27d4fb96
PP
169*/
170
69512466
JH
171#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
172# include <inttypes.h>
dd0eed91
JH
173# ifdef INT32_MIN_BROKEN
174# undef INT32_MIN
175# define INT32_MIN (-2147483647-1)
176# endif
177# ifdef INT64_MIN_BROKEN
178# undef INT64_MIN
179# define INT64_MIN (-9223372036854775807LL-1)
180# endif
69512466
JH
181#endif
182
8175356b
JH
183typedef I8TYPE I8;
184typedef U8TYPE U8;
185typedef I16TYPE I16;
186typedef U16TYPE U16;
187typedef I32TYPE I32;
188typedef U32TYPE U32;
6b8eaf93
JH
189#ifdef PERL_CORE
190# ifdef HAS_QUAD
8175356b
JH
191typedef I64TYPE I64;
192typedef U64TYPE U64;
6b8eaf93
JH
193# endif
194#endif /* PERL_CORE */
8175356b 195
69512466 196#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
80524f33
JL
197# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
198# define PeRl_INT64_C(c) CAT2(c,LL)
199# define PeRl_UINT64_C(c) CAT2(c,ULL)
200# else
52865553
SH
201# if QUADKIND == QUAD_IS___INT64
202# define PeRl_INT64_C(c) CAT2(c,I64)
203# define PeRl_UINT64_C(c) CAT2(c,UI64)
69512466 204# else
52865553
SH
205# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
206# define PeRl_INT64_C(c) CAT2(c,L)
207# define PeRl_UINT64_C(c) CAT2(c,UL)
69512466 208# else
80524f33
JL
209# define PeRl_INT64_C(c) ((I64TYPE)(c))
210# define PeRl_UINT64_C(c) ((U64TYPE)(c))
69512466
JH
211# endif
212# endif
e8c95190 213# endif
80524f33
JL
214# ifndef UINT64_C
215# define UINT64_C(c) PeRl_UINT64_C(c)
216# endif
217# ifndef INT64_C
218# define INT64_C(c) PeRl_INT64_C(c)
219# endif
e8c95190 220#endif
dc750f2e 221
d8668976 222#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 223
5ff3f7a4
GS
224/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
225 Please search CHAR_MAX in perl.h for further details. */
226#define U8_MAX UINT8_MAX
227#define U8_MIN UINT8_MIN
228
5ff3f7a4
GS
229#define I16_MAX INT16_MAX
230#define I16_MIN INT16_MIN
231#define U16_MAX UINT16_MAX
232#define U16_MIN UINT16_MIN
233
5ff3f7a4
GS
234#define I32_MAX INT32_MAX
235#define I32_MIN INT32_MIN
0e983133
GS
236#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
237# define U32_MAX UINT32_MAX
238#else
239# define U32_MAX 4294967295U
240#endif
5ff3f7a4
GS
241#define U32_MIN UINT32_MIN
242
243#else
244
5c9fa16e
KA
245/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
246 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
247#define U8_MAX PERL_UCHAR_MAX
248#define U8_MIN PERL_UCHAR_MIN
79072805 249
27d4fb96
PP
250#define I16_MAX PERL_SHORT_MAX
251#define I16_MIN PERL_SHORT_MIN
252#define U16_MAX PERL_USHORT_MAX
253#define U16_MIN PERL_USHORT_MIN
79072805 254
c4f23d77 255#if LONGSIZE > 4
27d4fb96
PP
256# define I32_MAX PERL_INT_MAX
257# define I32_MIN PERL_INT_MIN
258# define U32_MAX PERL_UINT_MAX
259# define U32_MIN PERL_UINT_MIN
79072805 260#else
27d4fb96
PP
261# define I32_MAX PERL_LONG_MAX
262# define I32_MIN PERL_LONG_MIN
263# define U32_MAX PERL_ULONG_MAX
264# define U32_MIN PERL_ULONG_MIN
79072805
LW
265#endif
266
5ff3f7a4
GS
267#endif
268
58a9a5d5 269/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
270#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
271#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
272#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
273
ff68c719 274#define Ctl(ch) ((ch) & 037)
8d063cd8 275
4650c663
KW
276/* This is a helper macro to avoid preprocessor issues, replaced by nothing
277 * unless under DEBUGGING, where it expands to an assert of its argument,
278 * followed by a comma (hence the comma operator). If we just used a straight
279 * assert(), we would get a comma with nothing before it when not DEBUGGING */
0f092d08
KW
280#ifdef DEBUGGING
281# define __ASSERT_(statement) assert(statement),
282#else
283# define __ASSERT_(statement)
284#endif
285
3fe05580
MHM
286/*
287=head1 SV-Body Allocation
288
289=for apidoc Ama|SV*|newSVpvs|const char* s
290Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
291
84bafc02
NC
292=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
293Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
294pair.
295
3fe05580
MHM
296=for apidoc Ama|SV*|newSVpvs_share|const char* s
297Like C<newSVpvn_share>, but takes a literal string instead of a string/length
298pair and omits the hash parameter.
299
9dcc53ea
Z
300=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
301Like C<sv_catpvn_flags>, but takes a literal string instead of a
302string/length pair.
303
304=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
305Like C<sv_catpvn_nomg>, but takes a literal string instead of a
306string/length pair.
307
7b4e7469 308=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
309Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
310
9dcc53ea
Z
311=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
312Like C<sv_catpvn_mg>, but takes a literal string instead of a
313string/length pair.
314
7b4e7469 315=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
316Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
317
9dcc53ea
Z
318=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
319Like C<sv_setpvn_mg>, but takes a literal string instead of a
320string/length pair.
321
322=for apidoc Am|SV *|sv_setref_pvs|const char* s
323Like C<sv_setref_pvn>, but takes a literal string instead of a
324string/length pair.
325
3fe05580
MHM
326=head1 Memory Management
327
328=for apidoc Ama|char*|savepvs|const char* s
329Like C<savepvn>, but takes a literal string instead of a string/length pair.
330
9dcc53ea
Z
331=for apidoc Ama|char*|savesharedpvs|const char* s
332A version of C<savepvs()> which allocates the duplicate string in memory
333which is shared between threads.
334
3fe05580
MHM
335=head1 GV Functions
336
337=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
338Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
339
340=head1 Hash Manipulation Functions
341
342=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
343Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
344
345=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
346Like C<hv_store>, but takes a literal string instead of a string/length pair
347and omits the hash parameter.
348
510966aa
Z
349=head1 Lexer interface
350
351=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
352
353Like L</lex_stuff_pvn>, but takes a literal string instead of a
354string/length pair.
355
3fe05580
MHM
356=cut
357*/
358
926bb54c
KW
359/* concatenating with "" ensures that only literal strings are accepted as
360 * argument */
26bfd924 361#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 362
926bb54c
KW
363/* note that STR_WITH_LEN() can't be used as argument to macros or functions
364 * that under some configurations might be macros, which means that it requires
365 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
ba3a79e7
GA
366 */
367
368/* STR_WITH_LEN() shortcuts */
369#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
370#define newSVpvs_flags(str,flags) \
371 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 372#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
373#define sv_catpvs_flags(sv, str, flags) \
374 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
375#define sv_catpvs_nomg(sv, str) \
376 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
377#define sv_catpvs(sv, str) \
378 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
379#define sv_catpvs_mg(sv, str) \
380 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 381#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
382#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
383#define sv_setref_pvs(rv, classname, str) \
384 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 385#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
386#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
387#define gv_stashpvs(str, create) \
388 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
389#define gv_fetchpvs(namebeg, add, sv_type) \
390 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
391#define gv_fetchpvn(namebeg, len, add, sv_type) \
392 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
393#define sv_catxmlpvs(dsv, str, utf8) \
394 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
395#define hv_fetchs(hv,key,lval) \
396 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
397 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
398 : HV_FETCH_JUST_SV, NULL, 0))
399
400#define hv_stores(hv,key,val) \
401 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
402 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 403
510966aa
Z
404#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
405
b96d8cd9
NC
406#define get_cvs(str, flags) \
407 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 408
954c1994 409/*
ccfc67b7
JH
410=head1 Miscellaneous Functions
411
954c1994
GS
412=for apidoc Am|bool|strNE|char* s1|char* s2
413Test two strings to see if they are different. Returns true or
414false.
415
416=for apidoc Am|bool|strEQ|char* s1|char* s2
417Test two strings to see if they are equal. Returns true or false.
418
419=for apidoc Am|bool|strLT|char* s1|char* s2
420Test two strings to see if the first, C<s1>, is less than the second,
421C<s2>. Returns true or false.
422
423=for apidoc Am|bool|strLE|char* s1|char* s2
424Test two strings to see if the first, C<s1>, is less than or equal to the
425second, C<s2>. Returns true or false.
426
427=for apidoc Am|bool|strGT|char* s1|char* s2
428Test two strings to see if the first, C<s1>, is greater than the second,
429C<s2>. Returns true or false.
430
431=for apidoc Am|bool|strGE|char* s1|char* s2
432Test two strings to see if the first, C<s1>, is greater than or equal to
433the second, C<s2>. Returns true or false.
434
435=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
436Test two strings to see if they are different. The C<len> parameter
72d33970 437indicates the number of bytes to compare. Returns true or false. (A
954c1994
GS
438wrapper for C<strncmp>).
439
440=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
441Test two strings to see if they are equal. The C<len> parameter indicates
72d33970 442the number of bytes to compare. Returns true or false. (A wrapper for
954c1994
GS
443C<strncmp>).
444
445=cut
446*/
447
8d063cd8
LW
448#define strNE(s1,s2) (strcmp(s1,s2))
449#define strEQ(s1,s2) (!strcmp(s1,s2))
450#define strLT(s1,s2) (strcmp(s1,s2) < 0)
451#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
452#define strGT(s1,s2) (strcmp(s1,s2) > 0)
453#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
454#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
455#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 456
36477c24
PP
457#ifdef HAS_MEMCMP
458# define memNE(s1,s2,l) (memcmp(s1,s2,l))
459# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
460#else
461# define memNE(s1,s2,l) (bcmp(s1,s2,l))
462# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
463#endif
464
568a785a 465#define memEQs(s1, l, s2) \
8e6497b6 466 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
467#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
468
bbce6d69
PP
469/*
470 * Character classes.
471 *
472 * Unfortunately, the introduction of locales means that we
473 * can't trust isupper(), etc. to tell the truth. And when
474 * it comes to /\w+/ with tainting enabled, we *must* be able
475 * to trust our character classes.
476 *
477 * Therefore, the default tests in the text of Perl will be
478 * independent of locale. Any code that wants to depend on
479 * the current locale will use the tests that begin with "lc".
480 */
481
2304df62
AD
482#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
483# ifndef CTYPE256
484# define CTYPE256
485# endif
486#endif
487
954c1994 488/*
ccfc67b7
JH
489
490=head1 Character classes
243effed
KW
491This section is about functions (really macros) that classify characters
492into types, such as punctuation versus alphabetic, etc. Most of these are
493analogous to regular expression character classes. (See
494L<perlrecharclass/POSIX Character Classes>.) There are several variants for
495each class. (Not all macros have all variants; each item below lists the
496ones valid for it.) None are affected by C<use bytes>, and only the ones
497with C<LC> in the name are affected by the current locale.
498
499The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
500C<U8>) as input and returns a boolean as to whether or not the character
6aff1f14
KW
501represented by that octet is (or on non-ASCII platforms, corresponds to) an
502ASCII character in the named class based on platform, Unicode, and Perl rules.
503If the input is a number that doesn't fit in an octet, FALSE is returned.
243effed 504
6aff1f14
KW
505Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
506with no suffix C<"_A">.
4b9734bf
KW
507
508Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
509onto the platform. That is, the code points that are ASCII are unaffected,
510since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
511as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
512true when called with the code point 0xDF, which is a word character in both
4650c663 513ASCII and EBCDIC (though it represents different characters in each).
243effed
KW
514
515Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
516point as input. If the code point is larger than 255, Unicode rules are used
517to determine if it is in the character class. For example,
6aff1f14
KW
518C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
519WITH MACRON in Unicode, and is a word character.
243effed
KW
520
521Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
522(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
523classification of just the first (possibly multi-byte) character in the string
524is tested.
243effed
KW
525
526Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but uses
527the C library function that gives the named classification instead of
528hard-coded rules. For example, C<isDIGIT_LC()> returns the result of calling
529C<isdigit()>. This means that the result is based on the current locale, which
530is what C<LC> in the name stands for. FALSE is always returned if the input
531won't fit into an octet.
532
533Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
534returns the same as C<isFOO_LC> for input code points less than 256, and
535returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
536
537Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
538(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
6aff1f14
KW
539classification of just the first (possibly multi-byte) character in the string
540is tested.
ccfc67b7 541
954c1994 542=for apidoc Am|bool|isALPHA|char ch
2787a470 543Returns a boolean indicating whether the specified character is an
6aff1f14 544alphabetic character, analogous to C<m/[[:alpha:]]/>.
8a58bdcf 545See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 546C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
243effed 547C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
8a58bdcf 548
15861f94
KW
549=for apidoc Am|bool|isALPHANUMERIC|char ch
550Returns a boolean indicating whether the specified character is a either an
6aff1f14 551alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
15861f94
KW
552See the L<top of this section|/Character classes> for an explanation of variants
553C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
e5ad6aba 554C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
15861f94
KW
555C<isALPHANUMERIC_LC_utf8>.
556
8a58bdcf
KW
557=for apidoc Am|bool|isASCII|char ch
558Returns a boolean indicating whether the specified character is one of the 128
243effed 559characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
e5ad6aba 560On non-ASCII platforms, it returns TRUE iff this
8a58bdcf
KW
561character corresponds to an ASCII character. Variants C<isASCII_A()> and
562C<isASCII_L1()> are identical to C<isASCII()>.
243effed
KW
563See the L<top of this section|/Character classes> for an explanation of variants
564C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
565C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
566library routine C<isascii()>. In these cases, the variants whose names contain
567C<LC> are the same as the corresponding ones without.
568
d98532ea
KW
569Also note, that because all ASCII characters are UTF-8 invariant (meaning they
570have the exact same representation (always a single byte) whether encoded in
571UTF-8 or not), C<isASCII> will give the correct results when called with any
572byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
573work properly on any string encoded or not in UTF-8.
574
243effed
KW
575=for apidoc Am|bool|isBLANK|char ch
576Returns a boolean indicating whether the specified character is a
6aff1f14 577character considered to be a blank, analogous to C<m/[[:blank:]]/>.
243effed 578See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 579C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
243effed
KW
580C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
581platforms do not have the C library routine C<isblank()>. In these cases, the
582variants whose names contain C<LC> are the same as the corresponding ones
583without.
584
585=for apidoc Am|bool|isCNTRL|char ch
586Returns a boolean indicating whether the specified character is a
6aff1f14 587control character, analogous to C<m/[[:cntrl:]]/>.
243effed 588See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 589C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
6aff1f14
KW
590C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
591On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
954c1994 592
954c1994 593=for apidoc Am|bool|isDIGIT|char ch
2787a470 594Returns a boolean indicating whether the specified character is a
6aff1f14 595digit, analogous to C<m/[[:digit:]]/>.
8a58bdcf 596Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
243effed 597See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 598C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
243effed
KW
599C<isDIGIT_LC_utf8>.
600
601=for apidoc Am|bool|isGRAPH|char ch
602Returns a boolean indicating whether the specified character is a
6aff1f14 603graphic character, analogous to C<m/[[:graph:]]/>.
243effed 604See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 605C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
243effed 606C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
954c1994 607
0c82b6df 608=for apidoc Am|bool|isLOWER|char ch
2787a470 609Returns a boolean indicating whether the specified character is a
6aff1f14 610lowercase character, analogous to C<m/[[:lower:]]/>.
8a58bdcf 611See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 612C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
243effed 613C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
0c82b6df 614
c99e91e9 615=for apidoc Am|bool|isOCTAL|char ch
2787a470 616Returns a boolean indicating whether the specified character is an
6aff1f14 617octal digit, [0-7].
243effed
KW
618The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
619C<isOCTAL>.
620
621=for apidoc Am|bool|isPUNCT|char ch
622Returns a boolean indicating whether the specified character is a
6aff1f14
KW
623punctuation character, analogous to C<m/[[:punct:]]/>.
624Note that the definition of what is punctuation isn't as
243effed
KW
625straightforward as one might desire. See L<perlrecharclass/POSIX Character
626Classes> for details.
627See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 628C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
243effed 629C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
c99e91e9 630
0c82b6df 631=for apidoc Am|bool|isSPACE|char ch
2787a470 632Returns a boolean indicating whether the specified character is a
6aff1f14 633whitespace character. This is analogous
398d098a
KW
634to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
635(experimentally), this also matches what C<m/[[:space:]]/> does.
636("Experimentally" means that this change may be backed out in 5.20 or 5.22 if
637field experience indicates that it was unwise.) Prior to 5.18, only the
638locale forms of this macro (the ones with C<LC> in their names) matched
639precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
640in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
641(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
8a58bdcf 642See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 643C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
243effed 644C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
0c82b6df 645
398d098a
KW
646=for apidoc Am|bool|isPSXSPC|char ch
647(short for Posix Space)
648Starting in 5.18, this is identical (experimentally) in all its forms to the
649corresponding C<isSPACE()> macros. ("Experimentally" means that this change
650may be backed out in 5.20 or 5.22 if field experience indicates that it
651was unwise.)
652The locale forms of this macro are identical to their corresponding
653C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
654non-locale forms differ from their C<isSPACE()> forms only in that the
655C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
656Otherwise they are identical. Thus this macro is analogous to what
657C<m/[[:space:]]/> matches in a regular expression.
658See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 659C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
398d098a
KW
660C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
661
954c1994 662=for apidoc Am|bool|isUPPER|char ch
2787a470 663Returns a boolean indicating whether the specified character is an
6aff1f14 664uppercase character, analogous to C<m/[[:upper:]]/>.
8a58bdcf 665See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 666C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
243effed 667C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
954c1994 668
243effed 669=for apidoc Am|bool|isPRINT|char ch
8eea39dd 670Returns a boolean indicating whether the specified character is a
6aff1f14 671printable character, analogous to C<m/[[:print:]]/>.
8a58bdcf 672See the L<top of this section|/Character classes> for an explanation of variants
e5ad6aba 673C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
243effed
KW
674C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
675
676=for apidoc Am|bool|isWORDCHAR|char ch
677Returns a boolean indicating whether the specified character is a character
678that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
679in a regular expression. A word character is an alphabetic character, a
680decimal digit, a connecting punctuation character (such as an underscore), or
681a "mark" character that attaches to one of those (like some sort of accent).
682C<isALNUM()> is a synonym provided for backward compatibility, even though a
683word character includes more than the standard C language meaning of
684alphanumeric.
685See the L<top of this section|/Character classes> for an explanation of variants
686C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, C<isWORDCHAR_utf8>,
687C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8>.
8a58bdcf
KW
688
689=for apidoc Am|bool|isXDIGIT|char ch
690Returns a boolean indicating whether the specified character is a hexadecimal
243effed
KW
691digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
692and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
693See the L<top of this section|/Character classes> for an explanation of variants
694C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
695C<isXDIGIT_LC_utf8>.
696
3c3ecf18
KW
697=for apidoc Am|bool|isIDFIRST|char ch
698Returns a boolean indicating whether the specified character can be the first
699character of an identifier. This is very close to, but not quite the same as
700the official Unicode property C<XID_Start>. The difference is that this
701returns true only if the input character also matches L</isWORDCHAR>.
702See the L<top of this section|/Character classes> for an explanation of variants
703C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
704C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
705
706=for apidoc Am|bool|isIDCONT|char ch
707Returns a boolean indicating whether the specified character can be the
708second or succeeding character of an identifier. This is very close to, but
709not quite the same as the official Unicode property C<XID_Continue>. The
710difference is that this returns true only if the input character also matches
711L</isWORDCHAR>. See the L<top of this section|/Character classes> for an
712explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
713C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
714C<isIDCONT_LC_utf8>.
715
243effed 716=head1 Miscellaneous Functions
8eea39dd 717
95a59cab 718=for apidoc Am|U8|READ_XDIGIT|char str*
243effed 719Returns the value of an ASCII-range hex digit and advances the string pointer.
95a59cab
YO
720Behaviour is only well defined when isXDIGIT(*str) is true.
721
e7c1e6c1
KW
722=head1 Character case changing
723
1f607577
KW
724=for apidoc Am|U8|toUPPER|U8 ch
725Converts the specified character to uppercase. If the input is anything but an
726ASCII lowercase character, that input character itself is returned. Variant
c753c8d3 727C<toUPPER_A> is equivalent.
954c1994 728
1f607577
KW
729=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
730Converts the Unicode code point C<cp> to its uppercase version, and
731stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
732that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
733bytes since the uppercase version may be longer than the original character.
734
735The first code point of the uppercased version is returned
736(but note, as explained just above, that there may be more.)
737
738=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
739Converts the UTF-8 encoded character at C<p> to its uppercase version, and
740stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
741that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
742bytes since the uppercase version may be longer than the original character.
743
744The first code point of the uppercased version is returned
745(but note, as explained just above, that there may be more.)
746
747The input character at C<p> is assumed to be well-formed.
748
25200305
KW
749=for apidoc Am|U8|toFOLD|U8 ch
750Converts the specified character to foldcase. If the input is anything but an
751ASCII uppercase character, that input character itself is returned. Variant
752C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
753Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
754
1f607577
KW
755=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
756Converts the Unicode code point C<cp> to its foldcase version, and
757stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
758that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
759bytes since the foldcase version may be longer than the original character.
760
761The first code point of the foldcased version is returned
762(but note, as explained just above, that there may be more.)
763
764=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
765Converts the UTF-8 encoded character at C<p> to its foldcase version, and
766stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
767that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
768bytes since the foldcase version may be longer than the original character.
769
770The first code point of the foldcased version is returned
771(but note, as explained just above, that there may be more.)
772
773The input character at C<p> is assumed to be well-formed.
774
775=for apidoc Am|U8|toLOWER|U8 ch
776Converts the specified character to lowercase. If the input is anything but an
777ASCII uppercase character, that input character itself is returned. Variant
c753c8d3 778C<toLOWER_A> is equivalent.
954c1994 779
1f607577
KW
780=for apidoc Am|U8|toLOWER_L1|U8 ch
781Converts the specified Latin1 character to lowercase. The results are undefined if
782the input doesn't fit in a byte.
783
784=for apidoc Am|U8|toLOWER_LC|U8 ch
785Converts the specified character to lowercase using the current locale's rules,
786if possible; otherwise returns the input character itself.
787
788=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
789Converts the Unicode code point C<cp> to its lowercase version, and
790stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
791that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
792bytes since the lowercase version may be longer than the original character.
793
794The first code point of the lowercased version is returned
795(but note, as explained just above, that there may be more.)
796
797=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
798Converts the UTF-8 encoded character at C<p> to its lowercase version, and
799stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
800that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
801bytes since the lowercase version may be longer than the original character.
802
803The first code point of the lowercased version is returned
804(but note, as explained just above, that there may be more.)
805
806The input character at C<p> is assumed to be well-formed.
807
808=for apidoc Am|U8|toLOWER_LC|U8 ch
809Converts the specified character to lowercase using the current locale's rules,
810if possible; otherwise returns the input character itself.
811
25200305
KW
812=for apidoc Am|U8|toTITLE|U8 ch
813Converts the specified character to titlecase. If the input is anything but an
814ASCII lowercase character, that input character itself is returned. Variant
815C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
816as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
817concept used in locale handling, so there is no functionality for that.)
818
1f607577
KW
819=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
820Converts the Unicode code point C<cp> to its titlecase version, and
821stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
822that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
823bytes since the titlecase version may be longer than the original character.
824
825The first code point of the titlecased version is returned
826(but note, as explained just above, that there may be more.)
827
828=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
829Converts the UTF-8 encoded character at C<p> to its titlecase version, and
830stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
831that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
832bytes since the titlecase version may be longer than the original character.
833
834The first code point of the titlecased version is returned
835(but note, as explained just above, that there may be more.)
836
837The input character at C<p> is assumed to be well-formed.
838
954c1994 839=cut
353c9b6f 840
1f607577 841XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
25200305 842really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
243effed 843
8a58bdcf 844Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
845patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
846
954c1994
GS
847*/
848
7c062697
KW
849/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
850 * is known only in the perl core, and this macro can be called from outside
851 * that */
852#ifdef HAS_QUAD
853# define WIDEST_UTYPE U64TYPE
854#else
855# define WIDEST_UTYPE U32
856#endif
857
3912bc88
KW
858/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
859 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
860 * bits of information are lost even on a 64-bit machine, but to get the
861 * compiler to optimize it out if possible. This is because Configure makes
862 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
863 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
864 */
865#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 866
41f43cc2 867#ifdef EBCDIC
b6340bd0
KW
868# ifndef _ALL_SOURCE
869 /* This returns the wrong results on at least z/OS unless this is
870 * defined. */
871# error _ALL_SOURCE should probably be defined
872# endif
873
874 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
875 * likely not defined, and so we use the native function */
876# define isASCII(c) isascii(c)
41f43cc2
KW
877#else
878# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
879#endif
880
ec79cfbb 881#define isASCII_A(c) isASCII(c)
3509bb53 882#define isASCII_L1(c) isASCII(c)
6515cb76 883
c2da0b36
KW
884/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
885 * and the 8 possible permutations of those bits exactly comprise the 8 octal
886 * digits */
3d33b3e3 887#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36 888
ec79cfbb 889/* ASCII range only */
9fb1bf9d 890#ifdef H_PERL /* If have access to perl.h, lookup in its table */
f4cdb42c 891
a500dc72
KW
892/* Character class numbers. For internal core Perl use only. The ones less
893 * than 32 are used in PL_charclass[] and the ones up through the one that
894 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
895 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
896 * their actual definitions are here. If that file has a name not used here,
897 * it won't compile.
1709d539
KW
898 *
899 * The first group of these is ordered in what I (khw) estimate to be the
31c7f561
KW
900 * frequency of their use. This gives a slight edge to exiting a loop earlier
901 * (in reginclass() in regexec.c) */
1709d539
KW
902# define _CC_WORDCHAR 0 /* \w and [:word:] */
903# define _CC_DIGIT 1 /* \d and [:digit:] */
904# define _CC_ALPHA 2 /* [:alpha:] */
905# define _CC_LOWER 3 /* [:lower:] */
906# define _CC_UPPER 4 /* [:upper:] */
907# define _CC_PUNCT 5 /* [:punct:] */
908# define _CC_PRINT 6 /* [:print:] */
15861f94 909# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
1709d539 910# define _CC_GRAPH 8 /* [:graph:] */
b0d691b2 911# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
1709d539 912
b0d691b2 913#define _FIRST_NON_SWASH_CC 10
1709d539
KW
914/* The character classes above are implemented with swashes. The second group
915 * (just below) contains the ones implemented without. These are also sorted
916 * in rough order of the frequency of their use, except that \v should be last,
917 * as it isn't a real Posix character class, and some (small) inefficiencies in
918 * regular expression handling would be introduced by putting it in the middle
919 * of those that are. Also, cntrl and ascii come after the others as it may be
920 * useful to group these which have no members that match above Latin1, (or
921 * above ASCII in the latter case) */
922
b0d691b2
KW
923# define _CC_SPACE 10 /* \s */
924# define _CC_BLANK 11 /* [:blank:] */
925# define _CC_XDIGIT 12 /* [:xdigit:] */
926# define _CC_PSXSPC 13 /* [:space:] */
927# define _CC_CNTRL 14 /* [:cntrl:] */
928# define _CC_ASCII 15 /* [:ascii:] */
929# define _CC_VERTSPACE 16 /* \v */
1709d539 930
a0947d7b
KW
931# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
932
1709d539 933/* The members of the third group below do not need to be coordinated with data
3ded5eb0
KW
934 * structures in regcomp.[ch] and regexec.c. But they should be added to
935 * bootstrap_ctype() */
b0d691b2
KW
936# define _CC_IDFIRST 17
937# define _CC_CHARNAME_CONT 18
938# define _CC_NONLATIN1_FOLD 19
939# define _CC_QUOTEMETA 20
940# define _CC_NON_FINAL_FOLD 21
941# define _CC_IS_IN_SOME_FOLD 22
4d68ffa0
KW
942# define _CC_BACKSLASH_FOO_LBRACE_IS_META 31 /* temp, see mk_PL_charclass.pl */
943/* Unused: 23-30
f4cdb42c
KW
944 * If more bits are needed, one could add a second word for non-64bit
945 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
37ede926
KW
946 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
947 * is used only for optimization (as of this writing), and differs in the
948 * Latin1 range from the ALPHA bit only in two relatively unimportant
a500dc72
KW
949 * characters: the masculine and feminine ordinal indicators, so removing it
950 * would just cause /i regexes which match them to run less efficiently */
96ac0975 951
3a371f2f
KW
952#if defined(PERL_CORE) || defined(PERL_EXT)
953/* An enum version of the character class numbers, to help compilers
954 * optimize */
955typedef enum {
3a371f2f 956 _CC_ENUM_ALPHA = _CC_ALPHA,
e8d596e0
KW
957 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
958 _CC_ENUM_ASCII = _CC_ASCII,
959 _CC_ENUM_BLANK = _CC_BLANK,
b0d691b2 960 _CC_ENUM_CASED = _CC_CASED,
e8d596e0 961 _CC_ENUM_CNTRL = _CC_CNTRL,
3a371f2f
KW
962 _CC_ENUM_DIGIT = _CC_DIGIT,
963 _CC_ENUM_GRAPH = _CC_GRAPH,
964 _CC_ENUM_LOWER = _CC_LOWER,
965 _CC_ENUM_PRINT = _CC_PRINT,
e8d596e0 966 _CC_ENUM_PSXSPC = _CC_PSXSPC,
3a371f2f 967 _CC_ENUM_PUNCT = _CC_PUNCT,
e8d596e0 968 _CC_ENUM_SPACE = _CC_SPACE,
3a371f2f 969 _CC_ENUM_UPPER = _CC_UPPER,
e8d596e0 970 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
3a371f2f 971 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
e8d596e0 972 _CC_ENUM_XDIGIT = _CC_XDIGIT
3a371f2f
KW
973} _char_class_number;
974#endif
975
63c61c3f 976#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
86f72d56 977#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
63c61c3f 978
3018b823 979#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
922e8cb4 980# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
b0d691b2
KW
981 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
982 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
922e8cb4
KW
983 #error Need to adjust order of swash_property_names[]
984# endif
985
b44daeeb
KW
986/* This is declared static in each of the few files that this is #defined for
987 * to keep them from being publicly accessible. Hence there is a small amount
988 * of wasted space */
989
922e8cb4
KW
990static const char* const swash_property_names[] = {
991 "XPosixWord",
992 "XPosixDigit",
993 "XPosixAlpha",
994 "XPosixLower",
995 "XPosixUpper",
996 "XPosixPunct",
997 "XPosixPrint",
998 "XPosixAlnum",
b0d691b2
KW
999 "XPosixGraph",
1000 "Cased"
922e8cb4
KW
1001};
1002#endif
1003
96ac0975
NC
1004# ifdef DOINIT
1005EXTCONST U32 PL_charclass[] = {
1006# include "l1_char_class_tab.h"
1007};
1008
1009# else /* ! DOINIT */
1010EXTCONST U32 PL_charclass[];
1011# endif
3ded5eb0
KW
1012#endif /* Has perl.h */
1013
1014#if defined(H_PERL) && ! defined(BOOTSTRAP_CHARSET)
96ac0975 1015
265c1f46 1016 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
430b7c70 1017# define _CC_mask(classnum) (1U << (classnum))
4650c663
KW
1018
1019 /* For internal core Perl use only: the base macro for defining macros like
1020 * isALPHA */
430b7c70 1021# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
f4cd282c 1022 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
4eeeb416 1023
f4cdb42c
KW
1024 /* The mask for the _A versions of the macros; it just adds in the bit for
1025 * ASCII. */
1026# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1027
4650c663
KW
1028 /* For internal core Perl use only: the base macro for defining macros like
1029 * isALPHA_A. The foo_A version makes sure that both the desired bit and
1030 * the ASCII bit are present */
3ded5eb0 1031# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
f4cd282c 1032 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
f4cdb42c
KW
1033 == _CC_mask_A(classnum)))
1034
f4cdb42c 1035# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
15861f94 1036# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
f4cdb42c
KW
1037# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1038# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1039# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT)
1040# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1041# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1042# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1043# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1044# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1045# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1046# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1047# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1048# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
d95f8b6a 1049# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
3ded5eb0
KW
1050# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1051# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1052# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1053
1054 /* continuation character for legal NAME in \N{NAME} */
1055# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1056
1057# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1058# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1059# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1060# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1061# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1062# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1063# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1064# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1065# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1066# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
f4cdb42c 1067
fb9bfbf7
KW
1068 /* Either participates in a fold with a character above 255, or is a
1069 * multi-char fold */
f4cd282c 1070# 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 1071
4eeeb416 1072# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
26faadbd
KW
1073# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1074 _generic_isCC(c, _CC_NON_FINAL_FOLD)
37ede926
KW
1075# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1076 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
3ded5eb0
KW
1077#else /* Either don't have perl.h or don't want to use char_class_tab.h */
1078
1079 /* If we don't have perl.h, we are compiling a utility program. Below we
1080 * hard-code various macro definitions that wouldn't otherwise be available
1081 * to it. We can also get here if we are configured to bootstrap up Perl
1082 * on a non-ASCII platform that doesn't have a working Perl (currently only
1083 * EBCDIC). For these we currently use the native definitions to get
1084 * things going. (It should also be possible to use the translation
1085 * function NATIVE_TO_LATIN1(), but that is an extra layer of dependence on
1086 * Perl, so it is currently avoided for the macros where it's possible to
1087 * do so.) */
9fb1bf9d 1088# ifdef EBCDIC
3ded5eb0 1089 /* Use the native functions. They likely will return false for all
4f48cecf 1090 * non-ASCII values, but this makes sure */
b6340bd0
KW
1091# define isALPHA_A(c) (isASCII(c) && isalpha(c))
1092# define isALPHANUMERIC_A(c) (isASCII(c) && isalnum(c))
1093# define isCNTRL_A(c) (isASCII(c) && iscntrl(c))
1094# define isDIGIT_A(c) (isASCII(c) && isdigit(c))
1095# define isGRAPH_A(c) (isASCII(c) && isgraph(c))
1096# define isLOWER_A(c) (isASCII(c) && islower(c))
1097# define isPRINT_A(c) (isASCII(c) && isprint(c))
1098# define isPUNCT_A(c) (isASCII(c) && ispunct(c))
1099# define isSPACE_A(c) (isASCII(c) && isspace(c))
1100# define isUPPER_A(c) (isASCII(c) && isupper(c))
1101# define isXDIGIT_A(c) (isASCII(c) && isxdigit(c))
3ded5eb0 1102# else /* ASCII platform. These are coded based on first principals */
9fb1bf9d 1103# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
15861f94 1104# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
3ded5eb0 1105# define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
2181da29 1106# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
3ded5eb0 1107# define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
9fb1bf9d
KW
1108# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1109# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
3ded5eb0
KW
1110# define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1111# define isSPACE_A(c) ((c) == ' ' \
1112 || (c) == '\t' \
1113 || (c) == '\n' \
1114 || (c) == '\r' \
1115 || (c) =='\v' \
926bb54c 1116 || (c) == '\f')
2181da29 1117# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
3ded5eb0
KW
1118# define isXDIGIT_A(c) (isDIGIT_A(c) \
1119 || ((c) >= 'a' && (c) <= 'f') \
1120 || ((c) <= 'F' && (c) >= 'A'))
1121# endif /* Below are common definitions for ASCII and non-ASCII */
1122# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1123# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1124# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1125
1126 /* The _L1 macros may be unnecessary for both the utilities and for
1127 * bootstrapping; I (khw) added them during debugging of bootstrapping, and
1128 * it seems best to keep them. */
1129# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1130# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1131# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1132# define isBLANK_L1(c) (isBLANK_A(c) \
1133 || (FITS_IN_8_BITS(c) \
1134 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1135# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1136# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1137# define isLOWER_L1(c) (isLOWER_A(c) \
1138 || (FITS_IN_8_BITS(c) \
1139 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1140 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1141 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1142 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1143 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1144# define isPRINT_L1(c) (isPRINT_A(c) \
1145 || (FITS_IN_8_BITS(c) \
1146 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1147# define isPSXSPC_L1(c) isSPACE_L1(c)
1148# define isPUNCT_L1(c) (isPUNCT_A(c) \
1149 || (FITS_IN_8_BITS(c) \
1150 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1151 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1152 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1153 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1154 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1155 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1156 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1157# define isSPACE_L1(c) (isSPACE_A(c) \
1158 || (FITS_IN_8_BITS(c) \
1159 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1160 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1161# define isUPPER_L1(c) (isUPPER_A(c) \
1162 || (FITS_IN_8_BITS(c) \
1163 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1164 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1165 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1166# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1167# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1168# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1169 || isBLANK_L1(c) \
1170 || (c) == '-' \
1171 || (c) == '(' \
1172 || (c) == ')')
1173 /* The following are not fully accurate in the above-ASCII range. I (khw)
1174 * don't think it's necessary to be so for the purposes where this gets
1175 * compiled */
1176# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1177# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1178
1179 /* And these aren't accurate at all. They are useful only for above
1180 * Latin1, which utilities and bootstrapping don't deal with */
1181# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1182# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1183
1184 /* Many of the macros later in this file are defined in terms of these. By
1185 * implementing them with a function, which converts the class number into
1186 * a call to the desired macro, all of the later ones work. However, that
1187 * function won't be actually defined when building a utility program (no
1188 * perl.h), and so a compiler error will be generated if one is attempted
1189 * to be used. And the above-Latin1 code points require Unicode tables to
1190 * be present, something unlikely to be the case when bootstrapping */
1191# define _generic_isCC(c, classnum) \
1192 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1193# define _generic_isCC_A(c, classnum) \
1194 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
1195#endif /* End of no perl.h or have BOOTSTRAP_CHARSET */
8a58bdcf 1196
e66b99e9
KW
1197#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1198#define isALPHA(c) isALPHA_A(c)
1199#define isBLANK(c) isBLANK_A(c)
1200#define isCNTRL(c) isCNTRL_A(c)
1201#define isDIGIT(c) isDIGIT_A(c)
1202#define isGRAPH(c) isGRAPH_A(c)
1203#define isIDFIRST(c) isIDFIRST_A(c)
1204#define isLOWER(c) isLOWER_A(c)
1205#define isPRINT(c) isPRINT_A(c)
1206#define isPSXSPC(c) isPSXSPC_A(c)
1207#define isPUNCT(c) isPUNCT_A(c)
1208#define isSPACE(c) isSPACE_A(c)
1209#define isUPPER(c) isUPPER_A(c)
1210#define isWORDCHAR(c) isWORDCHAR_A(c)
1211#define isXDIGIT(c) isXDIGIT_A(c)
1212
1213/* ASCII casing. These could also be written as
1214 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1215 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1216 which uses table lookup and mask instead of subtraction. (This would
1217 work because the _MOD does not apply in the ASCII range) */
68067e4e
DM
1218#define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1219#define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
bbce6d69 1220
25200305
KW
1221/* In the ASCII range, these are equivalent to what they're here defined to be.
1222 * But by creating these definitions, other code doesn't have to be aware of
1223 * this detail */
1224#define toFOLD(c) toLOWER(c)
1225#define toFOLD_LC(c) toLOWER_LC(c)
1226#define toTITLE(c) toUPPER(c)
1227
c753c8d3
KW
1228#define toLOWER_A(c) toLOWER(c)
1229#define toUPPER_A(c) toUPPER(c)
25200305
KW
1230#define toFOLD_A(c) toFOLD(c)
1231#define toTITLE_A(c) toTITLE(c)
1a0901db 1232
4650c663 1233/* Use table lookup for speed; returns the input itself if is out-of-range */
b2bf251f 1234#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1235 ? (c) \
f4cd282c 1236 : PL_latin1_lc[ (U8) (c) ])
c753c8d3
KW
1237#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1238
1a0901db 1239/* Modified uc. Is correct uc except for three non-ascii chars which are
4650c663
KW
1240 * all mapped to one of them, and these need special handling; returns the
1241 * input itself if is out-of-range */
b2bf251f 1242#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
8e7c6e7d 1243 ? (c) \
f4cd282c 1244 : PL_mod_latin1_uc[ (U8) (c) ])
4650c663 1245#ifdef USE_NEXT_CTYPE /* NeXT computers */
bbce6d69 1246
15861f94 1247# define isALPHANUMERIC_LC(c) NXIsAlNum((unsigned int)(c))
ff68c719 1248# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 1249# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 1250# define isBLANK_LC(c) isBLANK((unsigned int)(c))
b8c5462f 1251# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
509fb054 1252# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
b8c5462f 1253# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
509fb054
KW
1254# define isIDFIRST_LC(c) (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
1255# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
ff68c719 1256# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 1257# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
509fb054
KW
1258# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
1259# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
1260# define isWORDCHAR_LC(c) (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
e712593e 1261# define isXDIGIT_LC(c) NXIsXDigit((unsigned int)(c))
ff68c719 1262# define toLOWER_LC(c) NXToLower((unsigned int)(c))
509fb054 1263# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
bbce6d69
PP
1264
1265#else /* !USE_NEXT_CTYPE */
b8c5462f 1266
bbce6d69 1267# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
4650c663 1268 /* For most other platforms */
76fde96a 1269# define isALPHA_LC(c) (FITS_IN_8_BITS(c) && isalpha((unsigned char)(c)))
15861f94
KW
1270# define isALPHANUMERIC_LC(c) (FITS_IN_8_BITS(c) \
1271 && isalnum((unsigned char)(c)))
07315176 1272# ifdef HAS_ISASCII
76fde96a 1273# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii((unsigned char)(c)))
07315176 1274# else
76fde96a 1275# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isASCII((unsigned char)(c)))
07315176 1276# endif
67addccf 1277# ifdef HAS_ISBLANK
76fde96a 1278# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isblank((unsigned char)(c)))
67addccf 1279# else
76fde96a 1280# define isBLANK_LC(c) (FITS_IN_8_BITS(c) && isBLANK((unsigned char)(c)))
67addccf 1281# endif
76fde96a 1282# define isCNTRL_LC(c) (FITS_IN_8_BITS(c) && iscntrl((unsigned char)(c)))
509fb054 1283# define isDIGIT_LC(c) (FITS_IN_8_BITS(c) && isdigit((unsigned char)(c)))
76fde96a 1284# define isGRAPH_LC(c) (FITS_IN_8_BITS(c) && isgraph((unsigned char)(c)))
509fb054
KW
1285# define isIDFIRST_LC(c) (FITS_IN_8_BITS(c) \
1286 && (isalpha((unsigned char)(c)) || (char)(c) == '_'))
1287# define isLOWER_LC(c) (FITS_IN_8_BITS(c) && islower((unsigned char)(c)))
76fde96a
KW
1288# define isPRINT_LC(c) (FITS_IN_8_BITS(c) && isprint((unsigned char)(c)))
1289# define isPUNCT_LC(c) (FITS_IN_8_BITS(c) && ispunct((unsigned char)(c)))
509fb054
KW
1290# define isSPACE_LC(c) (FITS_IN_8_BITS(c) && isspace((unsigned char)(c)))
1291# define isUPPER_LC(c) (FITS_IN_8_BITS(c) && isupper((unsigned char)(c)))
1292# define isWORDCHAR_LC(c) (FITS_IN_8_BITS(c) \
1293 && (isalnum((unsigned char)(c)) || (char)(c) == '_'))
e712593e 1294# define isXDIGIT_LC(c) (FITS_IN_8_BITS(c) && isxdigit((unsigned char)(c)))
548d896a
DM
1295# define toLOWER_LC(c) (FITS_IN_8_BITS(c) ? (UV)tolower((unsigned char)(c)) : (c))
1296# define toUPPER_LC(c) (FITS_IN_8_BITS(c) ? (UV)toupper((unsigned char)(c)) : (c))
bbce6d69 1297
4650c663 1298# else /* The final fallback position */
bbce6d69 1299
bbce6d69 1300# define isALPHA_LC(c) (isascii(c) && isalpha(c))
15861f94 1301# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
07315176 1302# define isASCII_LC(c) isascii(c)
67addccf
KW
1303# ifdef HAS_ISBLANK
1304# define isBLANK_LC(c) (isascii(c) && isblank(c))
1305# else
3e66ab74 1306# define isBLANK_LC(c) isBLANK_A(c)
67addccf 1307# endif
b8c5462f 1308# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
509fb054 1309# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
b8c5462f 1310# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
509fb054
KW
1311# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1312# define isLOWER_LC(c) (isascii(c) && islower(c))
bbce6d69 1313# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 1314# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
509fb054
KW
1315# define isSPACE_LC(c) (isascii(c) && isspace(c))
1316# define isUPPER_LC(c) (isascii(c) && isupper(c))
1317# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
e712593e 1318# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
4650c663 1319
76fde96a 1320# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
509fb054 1321# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
bbce6d69
PP
1322
1323# endif
a0d0e21e 1324#endif /* USE_NEXT_CTYPE */
55204971 1325
eba68aa0
KW
1326#define isIDCONT(c) isWORDCHAR(c)
1327#define isIDCONT_A(c) isWORDCHAR_A(c)
1328#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1329#define isIDCONT_LC(c) isWORDCHAR_LC(c)
13380643 1330#define isPSXSPC_LC(c) isSPACE_LC(c)
aaa51d5e 1331
4650c663
KW
1332/* For internal core Perl use only: the base macros for defining macros like
1333 * isALPHA_uni. 'c' is the code point to check. 'classnum' is the POSIX class
1334 * number defined earlier in this file. _generic_uni() is used for POSIX
1335 * classes where there is a macro or function 'above_latin1' that takes the
1336 * single argument 'c' and returns the desired value. These exist for those
1337 * classes which have simple definitions, avoiding the overhead of a hash
1338 * lookup or inversion list binary search. _generic_swash_uni() can be used
1339 * for classes where that overhead is faster than a direct lookup.
1340 * _generic_uni() won't compile if 'c' isn't unsigned, as it won't match the
1341 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so
1342 * have duplicate checks here, so could create versions of the macros that
1343 * don't, but experiments show that gcc optimizes them out anyway. */
66c17564
KW
1344
1345/* Note that all ignore 'use bytes' */
cd500f2f
KW
1346#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1347 ? _generic_isCC(c, classnum) \
1348 : above_latin1(c))
1349#define _generic_swash_uni(classnum, c) ((c) < 256 \
1350 ? _generic_isCC(c, classnum) \
922e8cb4 1351 : _is_uni_FOO(classnum, c))
cd500f2f
KW
1352#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1353#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
509fb054
KW
1354#define isASCII_uni(c) isASCII(c)
1355#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1356#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
cd500f2f
KW
1357#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1358#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
eba68aa0 1359#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
509fb054 1360#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
cd500f2f
KW
1361#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1362#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
aea0c08c 1363
509fb054
KW
1364/* Posix and regular space are identical above Latin1 */
1365#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
b18192e9 1366
cd500f2f 1367#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
509fb054 1368#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
cd500f2f 1369#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
509fb054 1370#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
cd500f2f 1371#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
509fb054 1372#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
b18192e9 1373
b8d68ded 1374#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
509fb054
KW
1375#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1376#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1377#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
a0ed51b3 1378
4650c663
KW
1379/* For internal core Perl use only: the base macros for defining macros like
1380 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code
1381 * point, not just 0-255. Like _generic_uni, there are two versions, one for
1382 * simple class definitions; the other for more complex. These are like
1383 * _generic_uni, so see it for more info. */
cd500f2f
KW
1384#define _generic_LC_uvchr(latin1, above_latin1, c) \
1385 (c < 256 ? latin1(c) : above_latin1(c))
1386#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1387 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1388
1389#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1390#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1391 _CC_ALPHANUMERIC, c)
509fb054 1392#define isASCII_LC_uvchr(c) isASCII_LC(c)
cd500f2f 1393#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
feeab5a9 1394#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
cd500f2f
KW
1395#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1396#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
eba68aa0
KW
1397#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1398 _is_uni_perl_idcont, c)
cd500f2f
KW
1399#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1400 _is_uni_perl_idstart, c)
1401#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1402#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
509fb054
KW
1403#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1404 space under locale */
cd500f2f
KW
1405#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1406#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
509fb054 1407 is_XPERLSPACE_cp_high, c)
cd500f2f
KW
1408#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1409#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1410 _CC_WORDCHAR, c)
1411#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
e712593e 1412
36b00f93 1413#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
aaa51d5e 1414
4650c663
KW
1415/* For internal core Perl use only: the base macros for defining macros like
1416 * isALPHA_utf8. These are like the earlier defined macros, but take an input
1417 * UTF-8 encoded string 'p'. If the input is in the Latin1 range, use
1418 * the Latin1 macro 'classnum' on 'p'. Otherwise use the value given by the
1419 * 'utf8' parameter. This relies on the fact that ASCII characters have the
1420 * same representation whether utf8 or not. Note that it assumes that the utf8
1421 * has been validated, and ignores 'use bytes' */
cd500f2f 1422#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
aea0c08c 1423 ? _generic_isCC(*(p), classnum) \
7d7a6efc 1424 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
aea0c08c 1425 ? _generic_isCC( \
94bb8c36 1426 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
aea0c08c 1427 *((p)+1 )), \
94bb8c36 1428 classnum) \
252810d0 1429 : utf8)
cd500f2f
KW
1430/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1431 * can be a macro */
1432#define _generic_func_utf8(classnum, above_latin1, p) \
1433 _generic_utf8(classnum, p, above_latin1(p))
4650c663 1434/* Like the above, but passes classnum to _isFOO_utf8(), instead of having an
cd500f2f 1435 * 'above_latin1' parameter */
922e8cb4 1436#define _generic_swash_utf8(classnum, p) \
cd500f2f 1437 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
922e8cb4 1438
cc8ab7c0
KW
1439/* Like the above, but should be used only when it is known that there are no
1440 * characters in the range 128-255 which the class is TRUE for. Hence it can
cd500f2f 1441 * skip the tests for this range. 'above_latin1' should include its arguments */
3ded5eb0 1442#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
cc8ab7c0
KW
1443 (UTF8_IS_INVARIANT(*(p)) \
1444 ? _generic_isCC(*(p), classnum) \
1445 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
3ded5eb0 1446 ? above_latin1 \
cc8ab7c0
KW
1447 : 0)
1448
84238efa
KW
1449/* NOTE that some of these macros have very similar ones in regcharclass.h.
1450 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1451 * there, differing in name only by an underscore from the one here
1452 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1453 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1454 * points; the regcharclass.h ones are implemented as a series of
1455 * "if-else-if-else ..." */
1456
922e8cb4
KW
1457#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1458#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
509fb054
KW
1459#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1460 utf8, the non-utf8 macro works
1461 */
cd500f2f
KW
1462#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1463#define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1464#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1465 _is_utf8_FOO(_CC_DIGIT, p))
922e8cb4 1466#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
eba68aa0 1467#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
2f4622f0 1468 _is_utf8_perl_idcont, p)
e5dcd934 1469
c11ff943
KW
1470/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1471 * IDFIRST is an alnum. See
c8362b00 1472 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
f91dcd13
KW
1473 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1474 * This used to be not the XID version, but we decided to go with the more
1475 * modern Unicode definition */
cd500f2f 1476#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
509fb054 1477 _is_utf8_perl_idstart, p)
e5dcd934 1478
922e8cb4
KW
1479#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1480#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
aea0c08c 1481
509fb054 1482/* Posix and regular space are identical above Latin1 */
cd500f2f 1483#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
aea0c08c 1484
922e8cb4 1485#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
cd500f2f 1486#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
922e8cb4 1487#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
cd500f2f 1488#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
922e8cb4 1489#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
cd500f2f
KW
1490#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1491 is_XDIGIT_high(p))
a0ed51b3 1492
7d2b61dd 1493#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
509fb054
KW
1494#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1495#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1496#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
2e8adce6 1497
4650c663
KW
1498/* For internal core Perl use only: the base macros for defining macros like
1499 * isALPHA_LC_utf8. These are like _generic_utf8, but if the first code point
1500 * in 'p' is within the 0-255 range, it uses locale rules from the passed-in
1501 * 'macro' parameter */
94bb8c36
KW
1502#define _generic_LC_utf8(macro, p, utf8) \
1503 (UTF8_IS_INVARIANT(*(p)) \
1504 ? macro(*(p)) \
1505 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1506 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
9201a572
KW
1507 : utf8)
1508
cd500f2f
KW
1509#define _generic_LC_swash_utf8(macro, classnum, p) \
1510 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1511#define _generic_LC_func_utf8(macro, above_latin1, p) \
4650c663 1512 _generic_LC_utf8(macro, p, above_latin1(p))
9201a572 1513
cd500f2f
KW
1514#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1515 _CC_ALPHANUMERIC, p)
1516#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
509fb054 1517#define isASCII_LC_utf8(p) isASCII_LC(*p)
cd500f2f
KW
1518#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1519#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1520#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1521#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
eba68aa0 1522#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
cd500f2f
KW
1523#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1524#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1525#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
509fb054
KW
1526#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1527 space under locale */
cd500f2f
KW
1528#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1529#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1530#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1531#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1532 _CC_WORDCHAR, p)
1533#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
aaa51d5e 1534
fbc19f27
KW
1535/* Macros for backwards compatibility and for completeness when the ASCII and
1536 * Latin1 values are identical */
fbc19f27
KW
1537#define isALPHAU(c) isALPHA_L1(c)
1538#define isDIGIT_L1(c) isDIGIT_A(c)
1539#define isOCTAL(c) isOCTAL_A(c)
1540#define isOCTAL_L1(c) isOCTAL_A(c)
1541#define isXDIGIT_L1(c) isXDIGIT_A(c)
2e28f0b9
KW
1542#define isALNUM(c) isWORDCHAR(c)
1543#define isALNUMU(c) isWORDCHAR_L1(c)
1544#define isALNUM_LC(c) isWORDCHAR_LC(c)
1545#define isALNUM_uni(c) isWORDCHAR_uni(c)
1546#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1547#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1548#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
15861f94
KW
1549#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1550#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1551#define isALNUMC(c) isALPHANUMERIC(c)
1552#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1553#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1554#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1555#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1556#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
fbc19f27 1557
6ae9f32a 1558/* This conversion works both ways, strangely enough. On EBCDIC platforms,
e5dcd934
KW
1559 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, except that they don't
1560 * necessarily mean the same characters, e.g. CTRL-D is 4 on both systems, but
1561 * that is EOT on ASCII; ST on EBCDIC */
8e7c6e7d 1562# define toCTRL(c) (toUPPER(NATIVE_TO_LATIN1(c)) ^ 64)
bbce6d69 1563
dea28490
JJ
1564/* Line numbers are unsigned, 32 bits. */
1565typedef U32 line_t;
e5dcd934 1566#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
378cc40b 1567
91152fc1
DG
1568/* Helpful alias for version prescan */
1569#define is_LAX_VERSION(a,b) \
1570 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1571
1572#define is_STRICT_VERSION(a,b) \
1573 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1574
1575#define BADVERSION(a,b,c) \
1576 if (b) { \
1577 *b = c; \
1578 } \
1579 return a;
8c52afec 1580
cb27eebd
KW
1581/* Converts a character known to represent a hexadecimal digit (0-9, A-F, or
1582 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer,
1583 * which is advanced. The input is validated only by an assert() in DEBUGGING
1584 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and
1585 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */
1586#define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \
1587 ? (c) \
1588 : ((c) + 9))))
902008b8
KW
1589#define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \
1590 ? (*(s)++) \
1591 : (*(s)++ + 9))))
95a59cab 1592
cb27eebd
KW
1593/* Converts a character known to represent an octal digit (0-7) to its numeric
1594 * value. The input is validated only by an assert() in DEBUGGING builds. In
1595 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */
1596#define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c)))
1597
8e84507e 1598/*
ccfc67b7
JH
1599=head1 Memory Management
1600
a02a5408 1601=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1602The XSUB-writer's interface to the C C<malloc> function.
1603
0d7b2759
MH
1604Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1605
c5008215
JC
1606In 5.9.3, Newx() and friends replace the older New() API, and drops
1607the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1608themselves. This aid has been superseded by a new build option,
d10b4965 1609PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1610there for use in XS modules supporting older perls.
1611
a02a5408 1612=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1613The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1614cast. See also C<Newx>.
954c1994 1615
0d7b2759
MH
1616Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1617
a02a5408 1618=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1619The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1620memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1621
0d7b2759
MH
1622Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1623
954c1994
GS
1624=for apidoc Am|void|Renew|void* ptr|int nitems|type
1625The XSUB-writer's interface to the C C<realloc> function.
1626
0d7b2759
MH
1627Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1628
954c1994
GS
1629=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1630The XSUB-writer's interface to the C C<realloc> function, with
1631cast.
1632
0d7b2759
MH
1633Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1634
49b8b560 1635=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1636The XSUB-writer's interface to the C C<free> function.
1637
0d7b2759
MH
1638This should B<ONLY> be used on memory obtained using L<"Newx"> and friends.
1639
954c1994
GS
1640=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1641The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
926bb54c
KW
1642source, C<dest> is the destination, C<nitems> is the number of items, and
1643C<type> is the type. Can do overlapping moves. See also C<Copy>.
954c1994 1644
e90e2364 1645=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
72d33970
FC
1646Like C<Move> but returns dest. Useful
1647for encouraging compilers to tail-call
e90e2364
NC
1648optimise.
1649
954c1994
GS
1650=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1651The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
926bb54c
KW
1652source, C<dest> is the destination, C<nitems> is the number of items, and
1653C<type> is the type. May fail on overlapping copies. See also C<Move>.
954c1994 1654
e90e2364
NC
1655=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1656
72d33970
FC
1657Like C<Copy> but returns dest. Useful
1658for encouraging compilers to tail-call
e90e2364
NC
1659optimise.
1660
954c1994
GS
1661=for apidoc Am|void|Zero|void* dest|int nitems|type
1662
1663The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1664destination, C<nitems> is the number of items, and C<type> is the type.
1665
e90e2364
NC
1666=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1667
72d33970
FC
1668Like C<Zero> but returns dest. Useful
1669for encouraging compilers to tail-call
e90e2364
NC
1670optimise.
1671
da5d8dbb 1672=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1673This is an architecture-independent macro to copy one structure to another.
954c1994 1674
7e337ee0
JH
1675=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1676
1677Fill up memory with a byte pattern (a byte repeated over and over
1678again) that hopefully catches attempts to access uninitialized memory.
1679
1680=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1681
1682PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1683
1c12ffb4 1684=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1685
1686PoisonWith(0xEF) for catching access to freed memory.
1687
9965345d
JH
1688=for apidoc Am|void|Poison|void* dest|int nitems|type
1689
7e337ee0 1690PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1691
1692=cut */
954c1994 1693
561b68a9
SH
1694/* Maintained for backwards-compatibility only. Use newSV() instead. */
1695#ifndef PERL_CORE
ff06c60c 1696#define NEWSV(x,len) newSV(len)
561b68a9 1697#endif
ff06c60c 1698
19a94d75
JH
1699#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1700
2720f793
JH
1701/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1702 * overly eager compilers that will bleat about e.g.
1703 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1704#ifdef PERL_MALLOC_WRAP
f019c49e 1705#define MEM_WRAP_CHECK(n,t) \
d1decf2b 1706 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (croak_memory_wrap(),0))
27d5b266 1707#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1708 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1709#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1710
d1decf2b 1711#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
1712#else
1713
410319be
NC
1714#define MEM_WRAP_CHECK(n,t)
1715#define MEM_WRAP_CHECK_1(n,t,a)
1716#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1717#define MEM_WRAP_CHECK_(n,t)
1718
6fff79ce 1719#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1720
1936d2a7 1721#endif
8b44ba4c 1722
fe4f188c 1723#ifdef PERL_MEM_LOG
46c6c7e2 1724/*
9f653bb5 1725 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1726 * go through functions, which are handy for debugging breakpoints, but
1727 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1728 * line number, and C function name if available) passed in. This info can
1729 * then be used for logging the calls, for which one gets a sample
73d1d973 1730 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1731 *
46c6c7e2 1732 * Known problems:
94e892a6 1733 * - not all memory allocs get logged, only those
46c6c7e2 1734 * that go through Newx() and derivatives (while all
94e892a6 1735 * Safefrees do get logged)
46c6c7e2
JH
1736 * - __FILE__ and __LINE__ do not work everywhere
1737 * - __func__ or __FUNCTION__ even less so
1738 * - I think more goes on after the perlio frees but
1739 * the thing is that STDERR gets closed (as do all
1740 * the file descriptors)
1741 * - no deeper calling stack than the caller of the Newx()
1742 * or the kind, but do I look like a C reflection/introspection
1743 * utility to you?
1744 * - the function prototypes for the logging functions
1745 * probably should maybe be somewhere else than handy.h
1746 * - one could consider inlining (macrofying) the logging
1747 * for speed, but I am too lazy
1748 * - one could imagine recording the allocations in a hash,
1749 * (keyed by the allocation address?), and maintain that
1750 * through reallocs and frees, but how to do that without
1751 * any News() happening...?
73d1d973 1752 * - lots of -Ddefines to get useful/controllable output
b953482e 1753 * - lots of ENV reads
46c6c7e2
JH
1754 */
1755
12754f92 1756PERL_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 1757
12754f92 1758PERL_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 1759
12754f92 1760PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1761
0b0ab801 1762# ifdef PERL_CORE
73d1d973 1763# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1764enum mem_log_type {
1765 MLT_ALLOC,
1766 MLT_REALLOC,
d7a2c63c
MHM
1767 MLT_FREE,
1768 MLT_NEW_SV,
1769 MLT_DEL_SV
0b0ab801
MHM
1770};
1771# endif
12754f92 1772# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1773void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1774void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1775# endif
0b0ab801
MHM
1776# endif
1777
fe4f188c
JH
1778#endif
1779
1780#ifdef PERL_MEM_LOG
d1401ee9
MHM
1781#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1782#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1783#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1784#endif
1785
1786#ifndef MEM_LOG_ALLOC
1787#define MEM_LOG_ALLOC(n,t,a) (a)
1788#endif
1789#ifndef MEM_LOG_REALLOC
1790#define MEM_LOG_REALLOC(n,t,v,a) (a)
1791#endif
1792#ifndef MEM_LOG_FREE
1793#define MEM_LOG_FREE(a) (a)
1794#endif
1795
d1401ee9
MHM
1796#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1797#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1798#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1799
1800#ifndef PERL_CORE
a02a5408
JC
1801/* pre 5.9.x compatibility */
1802#define New(x,v,n,t) Newx(v,n,t)
1803#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1804#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1805#endif
a02a5408 1806
ff68c719 1807#define Renew(v,n,t) \
d1401ee9 1808 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1809#define Renewc(v,n,t,c) \
d1401ee9 1810 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1811
1812#ifdef PERL_POISON
1813#define Safefree(d) \
06c0cc96 1814 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1815#else
fe4f188c 1816#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1817#endif
55497cff 1818
8b44ba4c
NC
1819#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1820#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1821#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1822
8b44ba4c
NC
1823#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1824#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1825#ifdef HAS_MEMSET
8b44ba4c 1826#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1827#else
8b44ba4c
NC
1828/* Using bzero(), which returns void. */
1829#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1830#endif
1831
7e337ee0
JH
1832#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1833#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1834#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1835#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1836
caa674f3
DD
1837#ifdef PERL_POISON
1838# define PERL_POISON_EXPR(x) x
1839#else
1840# define PERL_POISON_EXPR(x)
1841#endif
1842
2304df62 1843#ifdef USE_STRUCT_COPY
ff68c719 1844#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1845#else
1846#define StructCopy(s,d,t) Copy(s,d,1,t)
1847#endif
2cc61e15 1848
622913ab 1849#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1850#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1851
2cc61e15
DD
1852#ifdef NEED_VA_COPY
1853# ifdef va_copy
1854# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1855# else
a1866d1b
JH
1856# if defined(__va_copy)
1857# define Perl_va_copy(s, d) __va_copy(d, s)
1858# else
1859# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1860# endif
2cc61e15
DD
1861# endif
1862#endif
1863
472d47bc
SB
1864/* convenience debug macros */
1865#ifdef USE_ITHREADS
1866#define pTHX_FORMAT "Perl interpreter: 0x%p"
1867#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1868#define pTHX_VALUE_ (void *)my_perl,
1869#define pTHX_VALUE (void *)my_perl
1870#define pTHX__VALUE_ ,(void *)my_perl,
1871#define pTHX__VALUE ,(void *)my_perl
472d47bc 1872#else
3609ea0d 1873#define pTHX_FORMAT
472d47bc 1874#define pTHX__FORMAT
3609ea0d 1875#define pTHX_VALUE_
472d47bc 1876#define pTHX_VALUE
3609ea0d 1877#define pTHX__VALUE_
472d47bc
SB
1878#define pTHX__VALUE
1879#endif /* USE_ITHREADS */
3609ea0d 1880
2acdbac1
NC
1881/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1882 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1883 CPAN::Unpack show any users outside the core. */
1884#ifdef PERL_CORE
d1d15184 1885# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1886#endif
1887
dfff4baf
BF
1888/* Internal macros to deal with gids and uids */
1889#ifdef PERL_CORE
1890
1891# if Uid_t_size > IVSIZE
1892# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
1893# define SvUID(sv) SvNV(sv)
1894# else
1895# if Uid_t_sign <= 0
1896# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
1897# define SvUID(sv) SvIV(sv)
1898# else
1899# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
1900# define SvUID(sv) SvUV(sv)
1901# endif
1902# endif /* Uid_t_size */
1903
1904# if Gid_t_size > IVSIZE
1905# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
1906# define SvGID(sv) SvNV(sv)
1907# else
1908# if Gid_t_sign <= 0
1909# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
1910# define SvGID(sv) SvIV(sv)
1911# else
1912# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
1913# define SvGID(sv) SvUV(sv)
1914# endif
1915# endif /* Gid_t_size */
1916
1917#endif
1918
9d745869
KW
1919#endif /* HANDY_H */
1920
e9a8c099
MHM
1921/*
1922 * Local variables:
1923 * c-indentation-style: bsd
1924 * c-basic-offset: 4
14d04a33 1925 * indent-tabs-mode: nil
e9a8c099
MHM
1926 * End:
1927 *
14d04a33 1928 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1929 */