This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
https://perl5.git.perl.org / perl5.git / blame_incremental
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (non-incremental) | history | HEAD
Test preamble: unify chdir 't' if -d 't';
[perl5.git] / handy.h
... / ...
CommitLineData
1/* handy.h
2 *
3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000,
4 * 2001, 2002, 2004, 2005, 2006, 2007, 2008, 2012 by Larry Wall and others
5 *
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.
8 *
9 */
10
11/* IMPORTANT NOTE: Everything whose name begins with an underscore is for
12 * internal core Perl use only. */
13
14#ifndef HANDY_H /* Guard against nested #inclusion */
15#define HANDY_H
16
17#if !defined(__STDC__)
18#ifdef NULL
19#undef NULL
20#endif
21# define NULL 0
22#endif
23
24#ifndef PERL_CORE
25# define Null(type) ((type)NULL)
26
27/*
28=head1 Handy Values
29
30=for apidoc AmU||Nullch
31Null character pointer. (No longer available when C<PERL_CORE> is
32defined.)
33
34=for apidoc AmU||Nullsv
35Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
36
37=cut
38*/
39
40# define Nullch Null(char*)
41# define Nullfp Null(PerlIO*)
42# define Nullsv Null(SV*)
43#endif
44
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
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
63#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
64# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
65#else
66# define MUTABLE_PTR(p) ((void *) (p))
67#endif
68
69#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
70#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
71#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
72#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
73#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
74#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
75
76#if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR)
77# include <stdbool.h>
78# ifndef HAS_BOOL
79# define HAS_BOOL 1
80# endif
81#endif
82
83/* bool is built-in for g++-2.6.3 and later, which might be used
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
86 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
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
90*/
91#ifdef __GNUG__ /* GNU g++ has bool built-in */
92# ifndef PERL_BOOL_AS_CHAR
93# ifndef HAS_BOOL
94# define HAS_BOOL 1
95# endif
96# endif
97#endif
98
99#ifndef HAS_BOOL
100# ifdef bool
101# undef bool
102# endif
103# define bool char
104# define HAS_BOOL 1
105#endif
106
107/* cast-to-bool. A simple (bool) cast may not do the right thing: if bool is
108 * defined as char for example, then the cast from int is
109 * implementation-defined (bool)!!(cbool) in a ternary triggers a bug in xlc on
110 * AIX */
111#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
112
113/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
114 * XXX Should really be a Configure probe, with HAS__FUNCTION__
115 * and FUNCTION__ as results.
116 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
117#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
118# define FUNCTION__ __func__
119#else
120# if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
121 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
122# define FUNCTION__ ""
123# else
124# define FUNCTION__ __FUNCTION__ /* Common extension. */
125# endif
126#endif
127
128/* XXX A note on the perl source internal type system. The
129 original intent was that I32 be *exactly* 32 bits.
130
131 Currently, we only guarantee that I32 is *at least* 32 bits.
132 Specifically, if int is 64 bits, then so is I32. (This is the case
133 for the Cray.) This has the advantage of meshing nicely with
134 standard library calls (where we pass an I32 and the library is
135 expecting an int), but the disadvantage that an I32 is not 32 bits.
136 Andy Dougherty August 1996
137
138 There is no guarantee that there is *any* integral type with
139 exactly 32 bits. It is perfectly legal for a system to have
140 sizeof(short) == sizeof(int) == sizeof(long) == 8.
141
142 Similarly, there is no guarantee that I16 and U16 have exactly 16
143 bits.
144
145 For dealing with issues that may arise from various 32/64-bit
146 systems, we will ask Configure to check out
147
148 SHORTSIZE == sizeof(short)
149 INTSIZE == sizeof(int)
150 LONGSIZE == sizeof(long)
151 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
152 PTRSIZE == sizeof(void *)
153 DOUBLESIZE == sizeof(double)
154 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
155
156*/
157
158#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
159# include <inttypes.h>
160# ifdef INT32_MIN_BROKEN
161# undef INT32_MIN
162# define INT32_MIN (-2147483647-1)
163# endif
164# ifdef INT64_MIN_BROKEN
165# undef INT64_MIN
166# define INT64_MIN (-9223372036854775807LL-1)
167# endif
168#endif
169
170typedef I8TYPE I8;
171typedef U8TYPE U8;
172typedef I16TYPE I16;
173typedef U16TYPE U16;
174typedef I32TYPE I32;
175typedef U32TYPE U32;
176#ifdef PERL_CORE
177# ifdef HAS_QUAD
178typedef I64TYPE I64;
179typedef U64TYPE U64;
180# endif
181#endif /* PERL_CORE */
182
183/* INT64_C/UINT64_C are C99 from <stdint.h> (so they will not be
184 * available in strict C89 mode), but they are nice, so let's define
185 * them if necessary. */
186#if defined(HAS_QUAD)
187# undef PeRl_INT64_C
188# undef PeRl_UINT64_C
189/* Prefer the native integer types (int and long) over long long
190 * (which is not C89) and Win32-specific __int64. */
191# if QUADKIND == QUAD_IS_INT && INTSIZE == 8
192# define PeRl_INT64_C(c) (c)
193# define PeRl_UINT64_C(c) CAT2(c,U)
194# endif
195# if QUADKIND == QUAD_IS_LONG && LONGSIZE == 8
196# define PeRl_INT64_C(c) CAT2(c,L)
197# define PeRl_UINT64_C(c) CAT2(c,UL)
198# endif
199# if QUADKIND == QUAD_IS_LONG_LONG && defined(HAS_LONG_LONG)
200# define PeRl_INT64_C(c) CAT2(c,LL)
201# define PeRl_UINT64_C(c) CAT2(c,ULL)
202# endif
203# if QUADKIND == QUAD_IS___INT64
204# define PeRl_INT64_C(c) CAT2(c,I64)
205# define PeRl_UINT64_C(c) CAT2(c,UI64)
206# endif
207# ifndef PeRl_INT64_C
208# define PeRl_INT64_C(c) ((I64TYPE)(c)) /* last resort */
209# define PeRl_UINT64_C(c) ((U64TYPE)(c))
210# endif
211/* In OS X the INT64_C/UINT64_C are defined with LL/ULL, which will
212 * not fly with C89-pedantic gcc, so let's undefine them first so that
213 * we can redefine them with our native integer preferring versions. */
214# if defined(PERL_DARWIN) && defined(PERL_GCC_PEDANTIC)
215# undef INT64_C
216# undef UINT64_C
217# endif
218# ifndef INT64_C
219# define INT64_C(c) PeRl_INT64_C(c)
220# endif
221# ifndef UINT64_C
222# define UINT64_C(c) PeRl_UINT64_C(c)
223# endif
224#endif
225
226#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
227
228/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
229 Please search CHAR_MAX in perl.h for further details. */
230#define U8_MAX UINT8_MAX
231#define U8_MIN UINT8_MIN
232
233#define I16_MAX INT16_MAX
234#define I16_MIN INT16_MIN
235#define U16_MAX UINT16_MAX
236#define U16_MIN UINT16_MIN
237
238#define I32_MAX INT32_MAX
239#define I32_MIN INT32_MIN
240#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
241# define U32_MAX UINT32_MAX
242#else
243# define U32_MAX 4294967295U
244#endif
245#define U32_MIN UINT32_MIN
246
247#else
248
249/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
250 Please search CHAR_MAX in perl.h for further details. */
251#define U8_MAX PERL_UCHAR_MAX
252#define U8_MIN PERL_UCHAR_MIN
253
254#define I16_MAX PERL_SHORT_MAX
255#define I16_MIN PERL_SHORT_MIN
256#define U16_MAX PERL_USHORT_MAX
257#define U16_MIN PERL_USHORT_MIN
258
259#if LONGSIZE > 4
260# define I32_MAX PERL_INT_MAX
261# define I32_MIN PERL_INT_MIN
262# define U32_MAX PERL_UINT_MAX
263# define U32_MIN PERL_UINT_MIN
264#else
265# define I32_MAX PERL_LONG_MAX
266# define I32_MIN PERL_LONG_MIN
267# define U32_MAX PERL_ULONG_MAX
268# define U32_MIN PERL_ULONG_MIN
269#endif
270
271#endif
272
273/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
274#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
275#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
276#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
277
278#define Ctl(ch) ((ch) & 037)
279
280/* This is a helper macro to avoid preprocessor issues, replaced by nothing
281 * unless under DEBUGGING, where it expands to an assert of its argument,
282 * followed by a comma (hence the comma operator). If we just used a straight
283 * assert(), we would get a comma with nothing before it when not DEBUGGING */
284#ifdef DEBUGGING
285# define __ASSERT_(statement) assert(statement),
286#else
287# define __ASSERT_(statement)
288#endif
289
290/*
291=head1 SV-Body Allocation
292
293=for apidoc Ama|SV*|newSVpvs|const char* s
294Like C<newSVpvn>, but takes a literal C<NUL>-terminated string instead of a
295string/length pair.
296
297=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
298Like C<newSVpvn_flags>, but takes a literal C<NUL>-terminated string instead of
299a string/length pair.
300
301=for apidoc Ama|SV*|newSVpvs_share|const char* s
302Like C<newSVpvn_share>, but takes a literal C<NUL>-terminated string instead of
303a string/length pair and omits the hash parameter.
304
305=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
306Like C<sv_catpvn_flags>, but takes a literal C<NUL>-terminated string instead
307of a string/length pair.
308
309=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
310Like C<sv_catpvn_nomg>, but takes a literal string instead of a
311string/length pair.
312
313=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
314Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
315
316=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
317Like C<sv_catpvn_mg>, but takes a literal string instead of a
318string/length pair.
319
320=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
321Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
322
323=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
324Like C<sv_setpvn_mg>, but takes a literal string instead of a
325string/length pair.
326
327=for apidoc Am|SV *|sv_setref_pvs|const char* s
328Like C<sv_setref_pvn>, but takes a literal string instead of a
329string/length pair.
330
331=head1 Memory Management
332
333=for apidoc Ama|char*|savepvs|const char* s
334Like C<savepvn>, but takes a literal C<NUL>-terminated string instead of a
335string/length pair.
336
337=for apidoc Ama|char*|savesharedpvs|const char* s
338A version of C<savepvs()> which allocates the duplicate string in memory
339which is shared between threads.
340
341=head1 GV Functions
342
343=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
344Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
345
346=head1 Hash Manipulation Functions
347
348=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
349Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
350
351=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
352Like C<hv_store>, but takes a literal string instead of a string/length pair
353and omits the hash parameter.
354
355=head1 Lexer interface
356
357=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
358
359Like L</lex_stuff_pvn>, but takes a literal string instead of a
360string/length pair.
361
362=cut
363*/
364
365/* concatenating with "" ensures that only literal strings are accepted as
366 * argument */
367#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
368
369/* note that STR_WITH_LEN() can't be used as argument to macros or functions
370 * that under some configurations might be macros, which means that it requires
371 * the full Perl_xxx(aTHX_ ...) form for any API calls where it's used.
372 */
373
374/* STR_WITH_LEN() shortcuts */
375#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
376#define newSVpvs_flags(str,flags) \
377 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
378#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
379#define sv_catpvs_flags(sv, str, flags) \
380 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
381#define sv_catpvs_nomg(sv, str) \
382 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
383#define sv_catpvs(sv, str) \
384 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
385#define sv_catpvs_mg(sv, str) \
386 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
387#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
388#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
389#define sv_setref_pvs(rv, classname, str) \
390 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
391#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
392#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
393#define gv_stashpvs(str, create) \
394 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
395#define gv_fetchpvs(namebeg, add, sv_type) \
396 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
397#define gv_fetchpvn(namebeg, len, add, sv_type) \
398 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
399#define sv_catxmlpvs(dsv, str, utf8) \
400 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
401#define hv_fetchs(hv,key,lval) \
402 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
403 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
404 : HV_FETCH_JUST_SV, NULL, 0))
405
406#define hv_stores(hv,key,val) \
407 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
408 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
409
410#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
411
412#define get_cvs(str, flags) \
413 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
414
415/*
416=head1 Miscellaneous Functions
417
418=for apidoc Am|bool|strNE|char* s1|char* s2
419Test two strings to see if they are different. Returns true or
420false.
421
422=for apidoc Am|bool|strEQ|char* s1|char* s2
423Test two strings to see if they are equal. Returns true or false.
424
425=for apidoc Am|bool|strLT|char* s1|char* s2
426Test two strings to see if the first, C<s1>, is less than the second,
427C<s2>. Returns true or false.
428
429=for apidoc Am|bool|strLE|char* s1|char* s2
430Test two strings to see if the first, C<s1>, is less than or equal to the
431second, C<s2>. Returns true or false.
432
433=for apidoc Am|bool|strGT|char* s1|char* s2
434Test two strings to see if the first, C<s1>, is greater than the second,
435C<s2>. Returns true or false.
436
437=for apidoc Am|bool|strGE|char* s1|char* s2
438Test two strings to see if the first, C<s1>, is greater than or equal to
439the second, C<s2>. Returns true or false.
440
441=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
442Test two strings to see if they are different. The C<len> parameter
443indicates the number of bytes to compare. Returns true or false. (A
444wrapper for C<strncmp>).
445
446=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
447Test two strings to see if they are equal. The C<len> parameter indicates
448the number of bytes to compare. Returns true or false. (A wrapper for
449C<strncmp>).
450
451=cut
452*/
453
454#define strNE(s1,s2) (strcmp(s1,s2))
455#define strEQ(s1,s2) (!strcmp(s1,s2))
456#define strLT(s1,s2) (strcmp(s1,s2) < 0)
457#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
458#define strGT(s1,s2) (strcmp(s1,s2) > 0)
459#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
460#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
461#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
462
463#ifdef HAS_MEMCMP
464# define memNE(s1,s2,l) (memcmp(s1,s2,l))
465# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
466#else
467# define memNE(s1,s2,l) (bcmp(s1,s2,l))
468# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
469#endif
470
471#define memEQs(s1, l, s2) \
472 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
473#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
474
475/*
476 * Character classes.
477 *
478 * Unfortunately, the introduction of locales means that we
479 * can't trust isupper(), etc. to tell the truth. And when
480 * it comes to /\w+/ with tainting enabled, we *must* be able
481 * to trust our character classes.
482 *
483 * Therefore, the default tests in the text of Perl will be
484 * independent of locale. Any code that wants to depend on
485 * the current locale will use the tests that begin with "lc".
486 */
487
488#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
489# ifndef CTYPE256
490# define CTYPE256
491# endif
492#endif
493
494/*
495
496=head1 Character classification
497This section is about functions (really macros) that classify characters
498into types, such as punctuation versus alphabetic, etc. Most of these are
499analogous to regular expression character classes. (See
500L<perlrecharclass/POSIX Character Classes>.) There are several variants for
501each class. (Not all macros have all variants; each item below lists the
502ones valid for it.) None are affected by C<use bytes>, and only the ones
503with C<LC> in the name are affected by the current locale.
504
505The base function, e.g., C<isALPHA()>, takes an octet (either a C<char> or a
506C<U8>) as input and returns a boolean as to whether or not the character
507represented by that octet is (or on non-ASCII platforms, corresponds to) an
508ASCII character in the named class based on platform, Unicode, and Perl rules.
509If the input is a number that doesn't fit in an octet, FALSE is returned.
510
511Variant C<isFOO_A> (e.g., C<isALPHA_A()>) is identical to the base function
512with no suffix C<"_A">.
513
514Variant C<isFOO_L1> imposes the Latin-1 (or EBCDIC equivlalent) character set
515onto the platform. That is, the code points that are ASCII are unaffected,
516since ASCII is a subset of Latin-1. But the non-ASCII code points are treated
517as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return
518true when called with the code point 0xDF, which is a word character in both
519ASCII and EBCDIC (though it represents different characters in each).
520
521Variant C<isFOO_uni> is like the C<isFOO_L1> variant, but accepts any UV code
522point as input. If the code point is larger than 255, Unicode rules are used
523to determine if it is in the character class. For example,
524C<isWORDCHAR_uni(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A
525WITH MACRON in Unicode, and is a word character.
526
527Variant C<isFOO_utf8> is like C<isFOO_uni>, but the input is a pointer to a
528(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
529classification of just the first (possibly multi-byte) character in the string
530is tested.
531
532Variant C<isFOO_LC> is like the C<isFOO_A> and C<isFOO_L1> variants, but the
533result is based on the current locale, which is what C<LC> in the name stands
534for. If Perl can determine that the current locale is a UTF-8 locale, it uses
535the published Unicode rules; otherwise, it uses the C library function that
536gives the named classification. For example, C<isDIGIT_LC()> when not in a
537UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always
538returned if the input won't fit into an octet. On some platforms where the C
539library function is known to be defective, Perl changes its result to follow
540the POSIX standard's rules.
541
542Variant C<isFOO_LC_uvchr> is like C<isFOO_LC>, but is defined on any UV. It
543returns the same as C<isFOO_LC> for input code points less than 256, and
544returns the hard-coded, not-affected-by-locale, Unicode results for larger ones.
545
546Variant C<isFOO_LC_utf8> is like C<isFOO_LC_uvchr>, but the input is a pointer to a
547(known to be well-formed) UTF-8 encoded string (C<U8*> or C<char*>). The
548classification of just the first (possibly multi-byte) character in the string
549is tested.
550
551=for apidoc Am|bool|isALPHA|char ch
552Returns a boolean indicating whether the specified character is an
553alphabetic character, analogous to C<m/[[:alpha:]]/>.
554See the L<top of this section|/Character classification> for an explanation of
555variants
556C<isALPHA_A>, C<isALPHA_L1>, C<isALPHA_uni>, C<isALPHA_utf8>, C<isALPHA_LC>,
557C<isALPHA_LC_uvchr>, and C<isALPHA_LC_utf8>.
558
559=for apidoc Am|bool|isALPHANUMERIC|char ch
560Returns a boolean indicating whether the specified character is a either an
561alphabetic character or decimal digit, analogous to C<m/[[:alnum:]]/>.
562See the L<top of this section|/Character classification> for an explanation of
563variants
564C<isALPHANUMERIC_A>, C<isALPHANUMERIC_L1>, C<isALPHANUMERIC_uni>,
565C<isALPHANUMERIC_utf8>, C<isALPHANUMERIC_LC>, C<isALPHANUMERIC_LC_uvchr>, and
566C<isALPHANUMERIC_LC_utf8>.
567
568=for apidoc Am|bool|isASCII|char ch
569Returns a boolean indicating whether the specified character is one of the 128
570characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
571On non-ASCII platforms, it returns TRUE iff this
572character corresponds to an ASCII character. Variants C<isASCII_A()> and
573C<isASCII_L1()> are identical to C<isASCII()>.
574See the L<top of this section|/Character classification> for an explanation of
575variants
576C<isASCII_uni>, C<isASCII_utf8>, C<isASCII_LC>, C<isASCII_LC_uvchr>, and
577C<isASCII_LC_utf8>. Note, however, that some platforms do not have the C
578library routine C<isascii()>. In these cases, the variants whose names contain
579C<LC> are the same as the corresponding ones without.
580
581Also note, that because all ASCII characters are UTF-8 invariant (meaning they
582have the exact same representation (always a single byte) whether encoded in
583UTF-8 or not), C<isASCII> will give the correct results when called with any
584byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> will
585work properly on any string encoded or not in UTF-8.
586
587=for apidoc Am|bool|isBLANK|char ch
588Returns a boolean indicating whether the specified character is a
589character considered to be a blank, analogous to C<m/[[:blank:]]/>.
590See the L<top of this section|/Character classification> for an explanation of
591variants
592C<isBLANK_A>, C<isBLANK_L1>, C<isBLANK_uni>, C<isBLANK_utf8>, C<isBLANK_LC>,
593C<isBLANK_LC_uvchr>, and C<isBLANK_LC_utf8>. Note, however, that some
594platforms do not have the C library routine C<isblank()>. In these cases, the
595variants whose names contain C<LC> are the same as the corresponding ones
596without.
597
598=for apidoc Am|bool|isCNTRL|char ch
599Returns a boolean indicating whether the specified character is a
600control character, analogous to C<m/[[:cntrl:]]/>.
601See the L<top of this section|/Character classification> for an explanation of
602variants
603C<isCNTRL_A>, C<isCNTRL_L1>, C<isCNTRL_uni>, C<isCNTRL_utf8>, C<isCNTRL_LC>,
604C<isCNTRL_LC_uvchr>, and C<isCNTRL_LC_utf8>
605On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant.
606
607=for apidoc Am|bool|isDIGIT|char ch
608Returns a boolean indicating whether the specified character is a
609digit, analogous to C<m/[[:digit:]]/>.
610Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
611See the L<top of this section|/Character classification> for an explanation of
612variants
613C<isDIGIT_uni>, C<isDIGIT_utf8>, C<isDIGIT_LC>, C<isDIGIT_LC_uvchr>, and
614C<isDIGIT_LC_utf8>.
615
616=for apidoc Am|bool|isGRAPH|char ch
617Returns a boolean indicating whether the specified character is a
618graphic character, analogous to C<m/[[:graph:]]/>.
619See the L<top of this section|/Character classification> for an explanation of
620variants
621C<isGRAPH_A>, C<isGRAPH_L1>, C<isGRAPH_uni>, C<isGRAPH_utf8>, C<isGRAPH_LC>,
622C<isGRAPH_LC_uvchr>, and C<isGRAPH_LC_utf8>.
623
624=for apidoc Am|bool|isLOWER|char ch
625Returns a boolean indicating whether the specified character is a
626lowercase character, analogous to C<m/[[:lower:]]/>.
627See the L<top of this section|/Character classification> for an explanation of
628variants
629C<isLOWER_A>, C<isLOWER_L1>, C<isLOWER_uni>, C<isLOWER_utf8>, C<isLOWER_LC>,
630C<isLOWER_LC_uvchr>, and C<isLOWER_LC_utf8>.
631
632=for apidoc Am|bool|isOCTAL|char ch
633Returns a boolean indicating whether the specified character is an
634octal digit, [0-7].
635The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to
636C<isOCTAL>.
637
638=for apidoc Am|bool|isPUNCT|char ch
639Returns a boolean indicating whether the specified character is a
640punctuation character, analogous to C<m/[[:punct:]]/>.
641Note that the definition of what is punctuation isn't as
642straightforward as one might desire. See L<perlrecharclass/POSIX Character
643Classes> for details.
644See the L<top of this section|/Character classification> for an explanation of
645variants
646C<isPUNCT_A>, C<isPUNCT_L1>, C<isPUNCT_uni>, C<isPUNCT_utf8>, C<isPUNCT_LC>,
647C<isPUNCT_LC_uvchr>, and C<isPUNCT_LC_utf8>.
648
649=for apidoc Am|bool|isSPACE|char ch
650Returns a boolean indicating whether the specified character is a
651whitespace character. This is analogous
652to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
653(experimentally), this also matches what C<m/[[:space:]]/> does.
654("Experimentally" means that this change may be backed out in 5.22 if
655field experience indicates that it was unwise.) Prior to 5.18, only the
656locale forms of this macro (the ones with C<LC> in their names) matched
657precisely what C<m/[[:space:]]/> does. In those releases, the only difference,
658in the non-locale variants, was that C<isSPACE()> did not match a vertical tab.
659(See L</isPSXSPC> for a macro that matches a vertical tab in all releases.)
660See the L<top of this section|/Character classification> for an explanation of
661variants
662C<isSPACE_A>, C<isSPACE_L1>, C<isSPACE_uni>, C<isSPACE_utf8>, C<isSPACE_LC>,
663C<isSPACE_LC_uvchr>, and C<isSPACE_LC_utf8>.
664
665=for apidoc Am|bool|isPSXSPC|char ch
666(short for Posix Space)
667Starting in 5.18, this is identical (experimentally) in all its forms to the
668corresponding C<isSPACE()> macros. ("Experimentally" means that this change
669may be backed out in 5.22 if field experience indicates that it
670was unwise.)
671The locale forms of this macro are identical to their corresponding
672C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the
673non-locale forms differ from their C<isSPACE()> forms only in that the
674C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do.
675Otherwise they are identical. Thus this macro is analogous to what
676C<m/[[:space:]]/> matches in a regular expression.
677See the L<top of this section|/Character classification> for an explanation of
678variants
679C<isPSXSPC_A>, C<isPSXSPC_L1>, C<isPSXSPC_uni>, C<isPSXSPC_utf8>, C<isPSXSPC_LC>,
680C<isPSXSPC_LC_uvchr>, and C<isPSXSPC_LC_utf8>.
681
682=for apidoc Am|bool|isUPPER|char ch
683Returns a boolean indicating whether the specified character is an
684uppercase character, analogous to C<m/[[:upper:]]/>.
685See the L<top of this section|/Character classification> for an explanation of
686variants
687C<isUPPER_A>, C<isUPPER_L1>, C<isUPPER_uni>, C<isUPPER_utf8>, C<isUPPER_LC>,
688C<isUPPER_LC_uvchr>, and C<isUPPER_LC_utf8>.
689
690=for apidoc Am|bool|isPRINT|char ch
691Returns a boolean indicating whether the specified character is a
692printable character, analogous to C<m/[[:print:]]/>.
693See the L<top of this section|/Character classification> for an explanation of
694variants
695C<isPRINT_A>, C<isPRINT_L1>, C<isPRINT_uni>, C<isPRINT_utf8>, C<isPRINT_LC>,
696C<isPRINT_LC_uvchr>, and C<isPRINT_LC_utf8>.
697
698=for apidoc Am|bool|isWORDCHAR|char ch
699Returns a boolean indicating whether the specified character is a character
700that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
701in a regular expression. A word character is an alphabetic character, a
702decimal digit, a connecting punctuation character (such as an underscore), or
703a "mark" character that attaches to one of those (like some sort of accent).
704C<isALNUM()> is a synonym provided for backward compatibility, even though a
705word character includes more than the standard C language meaning of
706alphanumeric.
707See the L<top of this section|/Character classification> for an explanation of
708variants
709C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uni>, and C<isWORDCHAR_utf8>.
710C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, and C<isWORDCHAR_LC_utf8> are also as
711described there, but additionally include the platform's native underscore.
712
713=for apidoc Am|bool|isXDIGIT|char ch
714Returns a boolean indicating whether the specified character is a hexadecimal
715digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
716and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
717See the L<top of this section|/Character classification> for an explanation of
718variants
719C<isXDIGIT_uni>, C<isXDIGIT_utf8>, C<isXDIGIT_LC>, C<isXDIGIT_LC_uvchr>, and
720C<isXDIGIT_LC_utf8>.
721
722=for apidoc Am|bool|isIDFIRST|char ch
723Returns a boolean indicating whether the specified character can be the first
724character of an identifier. This is very close to, but not quite the same as
725the official Unicode property C<XID_Start>. The difference is that this
726returns true only if the input character also matches L</isWORDCHAR>.
727See the L<top of this section|/Character classification> for an explanation of
728variants
729C<isIDFIRST_A>, C<isIDFIRST_L1>, C<isIDFIRST_uni>, C<isIDFIRST_utf8>,
730C<isIDFIRST_LC>, C<isIDFIRST_LC_uvchr>, and C<isIDFIRST_LC_utf8>.
731
732=for apidoc Am|bool|isIDCONT|char ch
733Returns a boolean indicating whether the specified character can be the
734second or succeeding character of an identifier. This is very close to, but
735not quite the same as the official Unicode property C<XID_Continue>. The
736difference is that this returns true only if the input character also matches
737L</isWORDCHAR>. See the L<top of this section|/Character classification> for
738an
739explanation of variants C<isIDCONT_A>, C<isIDCONT_L1>, C<isIDCONT_uni>,
740C<isIDCONT_utf8>, C<isIDCONT_LC>, C<isIDCONT_LC_uvchr>, and
741C<isIDCONT_LC_utf8>.
742
743=head1 Miscellaneous Functions
744
745=for apidoc Am|U8|READ_XDIGIT|char str*
746Returns the value of an ASCII-range hex digit and advances the string pointer.
747Behaviour is only well defined when isXDIGIT(*str) is true.
748
749=head1 Character case changing
750
751=for apidoc Am|U8|toUPPER|U8 ch
752Converts the specified character to uppercase. If the input is anything but an
753ASCII lowercase character, that input character itself is returned. Variant
754C<toUPPER_A> is equivalent.
755
756=for apidoc Am|UV|toUPPER_uni|UV cp|U8* s|STRLEN* lenp
757Converts the Unicode code point C<cp> to its uppercase version, and
758stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
759that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
760bytes since the uppercase version may be longer than the original character.
761
762The first code point of the uppercased version is returned
763(but note, as explained just above, that there may be more.)
764
765=for apidoc Am|UV|toUPPER_utf8|U8* p|U8* s|STRLEN* lenp
766Converts the UTF-8 encoded character at C<p> to its uppercase version, and
767stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
768that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
769bytes since the uppercase version may be longer than the original character.
770
771The first code point of the uppercased version is returned
772(but note, as explained just above, that there may be more.)
773
774The input character at C<p> is assumed to be well-formed.
775
776=for apidoc Am|U8|toFOLD|U8 ch
777Converts the specified character to foldcase. If the input is anything but an
778ASCII uppercase character, that input character itself is returned. Variant
779C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full
780Latin1 range, as the full generality of L</toFOLD_uni> is needed there.)
781
782=for apidoc Am|UV|toFOLD_uni|UV cp|U8* s|STRLEN* lenp
783Converts the Unicode code point C<cp> to its foldcase version, and
784stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
785that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
786bytes since the foldcase version may be longer than the original character.
787
788The first code point of the foldcased version is returned
789(but note, as explained just above, that there may be more.)
790
791=for apidoc Am|UV|toFOLD_utf8|U8* p|U8* s|STRLEN* lenp
792Converts the UTF-8 encoded character at C<p> to its foldcase version, and
793stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
794that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
795bytes since the foldcase version may be longer than the original character.
796
797The first code point of the foldcased version is returned
798(but note, as explained just above, that there may be more.)
799
800The input character at C<p> is assumed to be well-formed.
801
802=for apidoc Am|U8|toLOWER|U8 ch
803Converts the specified character to lowercase. If the input is anything but an
804ASCII uppercase character, that input character itself is returned. Variant
805C<toLOWER_A> is equivalent.
806
807=for apidoc Am|U8|toLOWER_L1|U8 ch
808Converts the specified Latin1 character to lowercase. The results are undefined if
809the input doesn't fit in a byte.
810
811=for apidoc Am|U8|toLOWER_LC|U8 ch
812Converts the specified character to lowercase using the current locale's rules,
813if possible; otherwise returns the input character itself.
814
815=for apidoc Am|UV|toLOWER_uni|UV cp|U8* s|STRLEN* lenp
816Converts the Unicode code point C<cp> to its lowercase version, and
817stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
818that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
819bytes since the lowercase version may be longer than the original character.
820
821The first code point of the lowercased version is returned
822(but note, as explained just above, that there may be more.)
823
824=for apidoc Am|UV|toLOWER_utf8|U8* p|U8* s|STRLEN* lenp
825Converts the UTF-8 encoded character at C<p> to its lowercase version, and
826stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
827that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
828bytes since the lowercase version may be longer than the original character.
829
830The first code point of the lowercased version is returned
831(but note, as explained just above, that there may be more.)
832
833The input character at C<p> is assumed to be well-formed.
834
835=for apidoc Am|U8|toLOWER_LC|U8 ch
836Converts the specified character to lowercase using the current locale's rules,
837if possible; otherwise returns the input character itself.
838
839=for apidoc Am|U8|toTITLE|U8 ch
840Converts the specified character to titlecase. If the input is anything but an
841ASCII lowercase character, that input character itself is returned. Variant
842C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 range,
843as the full generality of L</toTITLE_uni> is needed there. Titlecase is not a
844concept used in locale handling, so there is no functionality for that.)
845
846=for apidoc Am|UV|toTITLE_uni|UV cp|U8* s|STRLEN* lenp
847Converts the Unicode code point C<cp> to its titlecase version, and
848stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
849that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
850bytes since the titlecase version may be longer than the original character.
851
852The first code point of the titlecased version is returned
853(but note, as explained just above, that there may be more.)
854
855=for apidoc Am|UV|toTITLE_utf8|U8* p|U8* s|STRLEN* lenp
856Converts the UTF-8 encoded character at C<p> to its titlecase version, and
857stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note
858that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1>
859bytes since the titlecase version may be longer than the original character.
860
861The first code point of the titlecased version is returned
862(but note, as explained just above, that there may be more.)
863
864The input character at C<p> is assumed to be well-formed.
865
866=cut
867
868XXX Still undocumented isVERTWS_uni and _utf8; it's unclear what their names
869really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change.
870
871Note that these macros are repeated in Devel::PPPort, so should also be
872patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
873
874*/
875
876/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
877 * is known only in the perl core, and this macro can be called from outside
878 * that */
879#ifdef HAS_QUAD
880# define WIDEST_UTYPE U64TYPE
881#else
882# define WIDEST_UTYPE U32
883#endif
884
885/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
886 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
887 * bits of information are lost even on a 64-bit machine, but to get the
888 * compiler to optimize it out if possible. This is because Configure makes
889 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
890 * sizeof() guarantees that this evaluates to a constant true at compile time.
891 */
892#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
893
894#ifdef EBCDIC
895# ifndef _ALL_SOURCE
896 /* This returns the wrong results on at least z/OS unless this is
897 * defined. */
898# error _ALL_SOURCE should probably be defined
899# endif
900
901 /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
902 * likely not defined, and so we use the native function */
903# define isASCII(c) cBOOL(isascii(c))
904#else
905# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
906#endif
907
908#define isASCII_A(c) isASCII(c)
909#define isASCII_L1(c) isASCII(c)
910
911/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
912 * and the 8 possible permutations of those bits exactly comprise the 8 octal
913 * digits */
914#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
915
916/* ASCII range only */
917#ifdef H_PERL /* If have access to perl.h, lookup in its table */
918
919/* Character class numbers. For internal core Perl use only. The ones less
920 * than 32 are used in PL_charclass[] and the ones up through the one that
921 * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
922 * related files. PL_charclass ones use names used in l1_char_class_tab.h but
923 * their actual definitions are here. If that file has a name not used here,
924 * it won't compile.
925 *
926 * The first group of these is ordered in what I (khw) estimate to be the
927 * frequency of their use. This gives a slight edge to exiting a loop earlier
928 * (in reginclass() in regexec.c) */
929# define _CC_WORDCHAR 0 /* \w and [:word:] */
930# define _CC_DIGIT 1 /* \d and [:digit:] */
931# define _CC_ALPHA 2 /* [:alpha:] */
932# define _CC_LOWER 3 /* [:lower:] */
933# define _CC_UPPER 4 /* [:upper:] */
934# define _CC_PUNCT 5 /* [:punct:] */
935# define _CC_PRINT 6 /* [:print:] */
936# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
937# define _CC_GRAPH 8 /* [:graph:] */
938# define _CC_CASED 9 /* [:lower:] and [:upper:] under /i */
939
940#define _FIRST_NON_SWASH_CC 10
941/* The character classes above are implemented with swashes. The second group
942 * (just below) contains the ones implemented without. These are also sorted
943 * in rough order of the frequency of their use, except that \v should be last,
944 * as it isn't a real Posix character class, and some (small) inefficiencies in
945 * regular expression handling would be introduced by putting it in the middle
946 * of those that are. Also, cntrl and ascii come after the others as it may be
947 * useful to group these which have no members that match above Latin1, (or
948 * above ASCII in the latter case) */
949
950# define _CC_SPACE 10 /* \s */
951# define _CC_BLANK 11 /* [:blank:] */
952# define _CC_XDIGIT 12 /* [:xdigit:] */
953# define _CC_PSXSPC 13 /* [:space:] */
954# define _CC_CNTRL 14 /* [:cntrl:] */
955# define _CC_ASCII 15 /* [:ascii:] */
956# define _CC_VERTSPACE 16 /* \v */
957
958# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
959
960/* The members of the third group below do not need to be coordinated with data
961 * structures in regcomp.[ch] and regexec.c. */
962# define _CC_IDFIRST 17
963# define _CC_CHARNAME_CONT 18
964# define _CC_NONLATIN1_FOLD 19
965# define _CC_NONLATIN1_SIMPLE_FOLD 20
966# define _CC_QUOTEMETA 21
967# define _CC_NON_FINAL_FOLD 22
968# define _CC_IS_IN_SOME_FOLD 23
969# define _CC_MNEMONIC_CNTRL 24
970/* Unused: 25-31
971 * If more bits are needed, one could add a second word for non-64bit
972 * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd
973 * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it
974 * is used only for optimization (as of this writing), and differs in the
975 * Latin1 range from the ALPHA bit only in two relatively unimportant
976 * characters: the masculine and feminine ordinal indicators, so removing it
977 * would just cause /i regexes which match them to run less efficiently */
978
979#if defined(PERL_CORE) || defined(PERL_EXT)
980/* An enum version of the character class numbers, to help compilers
981 * optimize */
982typedef enum {
983 _CC_ENUM_ALPHA = _CC_ALPHA,
984 _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
985 _CC_ENUM_ASCII = _CC_ASCII,
986 _CC_ENUM_BLANK = _CC_BLANK,
987 _CC_ENUM_CASED = _CC_CASED,
988 _CC_ENUM_CNTRL = _CC_CNTRL,
989 _CC_ENUM_DIGIT = _CC_DIGIT,
990 _CC_ENUM_GRAPH = _CC_GRAPH,
991 _CC_ENUM_LOWER = _CC_LOWER,
992 _CC_ENUM_PRINT = _CC_PRINT,
993 _CC_ENUM_PSXSPC = _CC_PSXSPC,
994 _CC_ENUM_PUNCT = _CC_PUNCT,
995 _CC_ENUM_SPACE = _CC_SPACE,
996 _CC_ENUM_UPPER = _CC_UPPER,
997 _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
998 _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
999 _CC_ENUM_XDIGIT = _CC_XDIGIT
1000} _char_class_number;
1001#endif
1002
1003#define POSIX_SWASH_COUNT _FIRST_NON_SWASH_CC
1004#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
1005
1006#if defined(PERL_IN_UTF8_C) || defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
1007# if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
1008 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
1009 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9
1010 #error Need to adjust order of swash_property_names[]
1011# endif
1012
1013/* This is declared static in each of the few files that this is #defined for
1014 * to keep them from being publicly accessible. Hence there is a small amount
1015 * of wasted space */
1016
1017static const char* const swash_property_names[] = {
1018 "XPosixWord",
1019 "XPosixDigit",
1020 "XPosixAlpha",
1021 "XPosixLower",
1022 "XPosixUpper",
1023 "XPosixPunct",
1024 "XPosixPrint",
1025 "XPosixAlnum",
1026 "XPosixGraph",
1027 "Cased"
1028};
1029#endif
1030
1031# ifdef DOINIT
1032EXTCONST U32 PL_charclass[] = {
1033# include "l1_char_class_tab.h"
1034};
1035
1036# else /* ! DOINIT */
1037EXTCONST U32 PL_charclass[];
1038# endif
1039
1040 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
1041# define _CC_mask(classnum) (1U << (classnum))
1042
1043 /* For internal core Perl use only: the base macro for defining macros like
1044 * isALPHA */
1045# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
1046 && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
1047
1048 /* The mask for the _A versions of the macros; it just adds in the bit for
1049 * ASCII. */
1050# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
1051
1052 /* For internal core Perl use only: the base macro for defining macros like
1053 * isALPHA_A. The foo_A version makes sure that both the desired bit and
1054 * the ASCII bit are present */
1055# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
1056 && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
1057 == _CC_mask_A(classnum)))
1058
1059# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
1060# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
1061# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
1062# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
1063# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT) /* No non-ASCII digits */
1064# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
1065# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
1066# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
1067# define isPSXSPC_A(c) _generic_isCC_A(c, _CC_PSXSPC)
1068# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
1069# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
1070# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
1071# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
1072# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT) /* No non-ASCII xdigits */
1073# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
1074# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
1075# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
1076# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
1077
1078 /* continuation character for legal NAME in \N{NAME} */
1079# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
1080
1081# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
1082# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
1083# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
1084# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
1085# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
1086# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
1087# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
1088# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
1089# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
1090# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
1091
1092 /* Participates in a single-character fold with a character above 255 */
1093# define _HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_SIMPLE_FOLD)))
1094
1095 /* Like the above, but also can be part of a multi-char fold */
1096# 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)))
1097
1098# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
1099# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1100 _generic_isCC(c, _CC_NON_FINAL_FOLD)
1101# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1102 _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
1103# define _IS_MNEMONIC_CNTRL_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
1104 _generic_isCC(c, _CC_MNEMONIC_CNTRL)
1105#else /* else we don't have perl.h */
1106
1107 /* If we don't have perl.h, we are compiling a utility program. Below we
1108 * hard-code various macro definitions that wouldn't otherwise be available
1109 * to it. */
1110# ifdef EBCDIC
1111 /* Use the native functions. They likely will return false for all
1112 * non-ASCII values, but this makes sure */
1113# define isLOWER_A(c) (isASCII(c) && islower(c))
1114# define isPRINT_A(c) (isASCII(c) && isprint(c))
1115# define isUPPER_A(c) (isASCII(c) && isupper(c))
1116# else /* ASCII platform. These are coded based on first principals */
1117# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
1118# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
1119# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
1120# endif /* Below are common definitions for ASCII and non-ASCII */
1121# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
1122# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
1123# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
1124# define isCNTRL_A(c) (isASCII(c) && (! isPRINT_A(c)))
1125# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
1126# define isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
1127# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
1128# define isPUNCT_A(c) (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
1129# define isSPACE_A(c) ((c) == ' ' \
1130 || (c) == '\t' \
1131 || (c) == '\n' \
1132 || (c) == '\r' \
1133 || (c) == '\v' \
1134 || (c) == '\f')
1135# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
1136# define isXDIGIT_A(c) (isDIGIT_A(c) \
1137 || ((c) >= 'a' && (c) <= 'f') \
1138 || ((c) <= 'F' && (c) >= 'A'))
1139
1140 /* The _L1 macros may be unnecessary for the utilities; I (khw) added them
1141 * during debugging, and it seems best to keep them. */
1142# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
1143# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
1144# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
1145# define isBLANK_L1(c) (isBLANK_A(c) \
1146 || (FITS_IN_8_BITS(c) \
1147 && NATIVE_TO_LATIN1((U8) c) == 0xA0))
1148# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
1149# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
1150# define isLOWER_L1(c) (isLOWER_A(c) \
1151 || (FITS_IN_8_BITS(c) \
1152 && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
1153 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
1154 || NATIVE_TO_LATIN1((U8) c) == 0xAA \
1155 || NATIVE_TO_LATIN1((U8) c) == 0xBA \
1156 || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
1157# define isPRINT_L1(c) (isPRINT_A(c) \
1158 || (FITS_IN_8_BITS(c) \
1159 && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
1160# define isPSXSPC_L1(c) isSPACE_L1(c)
1161# define isPUNCT_L1(c) (isPUNCT_A(c) \
1162 || (FITS_IN_8_BITS(c) \
1163 && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
1164 || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
1165 || NATIVE_TO_LATIN1((U8) c) == 0xAB \
1166 || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
1167 || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
1168 || NATIVE_TO_LATIN1((U8) c) == 0xBB \
1169 || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
1170# define isSPACE_L1(c) (isSPACE_A(c) \
1171 || (FITS_IN_8_BITS(c) \
1172 && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
1173 || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
1174# define isUPPER_L1(c) (isUPPER_A(c) \
1175 || (FITS_IN_8_BITS(c) \
1176 && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
1177 && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
1178 && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
1179# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
1180# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
1181# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
1182 || isBLANK_L1(c) \
1183 || (c) == '-' \
1184 || (c) == '(' \
1185 || (c) == ')')
1186 /* The following are not fully accurate in the above-ASCII range. I (khw)
1187 * don't think it's necessary to be so for the purposes where this gets
1188 * compiled */
1189# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
1190# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
1191
1192 /* And these aren't accurate at all. They are useful only for above
1193 * Latin1, which utilities and bootstrapping don't deal with */
1194# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
1195# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
1196
1197 /* Many of the macros later in this file are defined in terms of these. By
1198 * implementing them with a function, which converts the class number into
1199 * a call to the desired macro, all of the later ones work. However, that
1200 * function won't be actually defined when building a utility program (no
1201 * perl.h), and so a compiler error will be generated if one is attempted
1202 * to be used. And the above-Latin1 code points require Unicode tables to
1203 * be present, something unlikely to be the case when bootstrapping */
1204# define _generic_isCC(c, classnum) \
1205 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
1206# define _generic_isCC_A(c, classnum) \
1207 (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
1208#endif /* End of no perl.h */
1209
1210#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
1211#define isALPHA(c) isALPHA_A(c)
1212#define isBLANK(c) isBLANK_A(c)
1213#define isCNTRL(c) isCNTRL_A(c)
1214#define isDIGIT(c) isDIGIT_A(c)
1215#define isGRAPH(c) isGRAPH_A(c)
1216#define isIDFIRST(c) isIDFIRST_A(c)
1217#define isLOWER(c) isLOWER_A(c)
1218#define isPRINT(c) isPRINT_A(c)
1219#define isPSXSPC(c) isPSXSPC_A(c)
1220#define isPUNCT(c) isPUNCT_A(c)
1221#define isSPACE(c) isSPACE_A(c)
1222#define isUPPER(c) isUPPER_A(c)
1223#define isWORDCHAR(c) isWORDCHAR_A(c)
1224#define isXDIGIT(c) isXDIGIT_A(c)
1225
1226/* ASCII casing. These could also be written as
1227 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
1228 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
1229 which uses table lookup and mask instead of subtraction. (This would
1230 work because the _MOD does not apply in the ASCII range) */
1231#define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c))
1232#define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c))
1233
1234/* In the ASCII range, these are equivalent to what they're here defined to be.
1235 * But by creating these definitions, other code doesn't have to be aware of
1236 * this detail */
1237#define toFOLD(c) toLOWER(c)
1238#define toTITLE(c) toUPPER(c)
1239
1240#define toLOWER_A(c) toLOWER(c)
1241#define toUPPER_A(c) toUPPER(c)
1242#define toFOLD_A(c) toFOLD(c)
1243#define toTITLE_A(c) toTITLE(c)
1244
1245/* Use table lookup for speed; returns the input itself if is out-of-range */
1246#define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \
1247 ? (c) \
1248 : PL_latin1_lc[ (U8) (c) ])
1249#define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */
1250
1251/* Modified uc. Is correct uc except for three non-ascii chars which are
1252 * all mapped to one of them, and these need special handling; returns the
1253 * input itself if is out-of-range */
1254#define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \
1255 ? (c) \
1256 : PL_mod_latin1_uc[ (U8) (c) ])
1257#define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale
1258
1259/* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */
1260
1261/* For internal core Perl use only: the base macro for defining macros like
1262 * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point
1263 * (0-255) to check. In a UTF-8 locale, the result is the same as calling
1264 * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like
1265 * _CC_UPPER, which gives the class number for doing this. For non-UTF-8
1266 * locales, the code to actually do the test this is passed in 'non_utf8'. If
1267 * 'c' is above 255, 0 is returned. For accessing the full range of possible
1268 * code points under locale rules, use the macros based on _generic_LC_uvchr
1269 * instead of this. */
1270#define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \
1271 (! FITS_IN_8_BITS(c) \
1272 ? 0 \
1273 : IN_UTF8_CTYPE_LOCALE \
1274 ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \
1275 : cBOOL(non_utf8))
1276
1277/* For internal core Perl use only: a helper macro for defining macros like
1278 * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to
1279 * actually do this test is passed in 'non_utf8_func', which is called on 'c',
1280 * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See
1281 * _generic_LC_base for more info */
1282#define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \
1283 _generic_LC_base(c,utf8_locale_classnum, \
1284 non_utf8_func( (_LC_CAST) (c)))
1285
1286/* For internal core Perl use only: like _generic_LC, but also returns TRUE if
1287 * 'c' is the platform's native underscore character */
1288#define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \
1289 _generic_LC_base(c, utf8_locale_classnum, \
1290 (non_utf8_func( (_LC_CAST) (c)) \
1291 || (char)(c) == '_'))
1292
1293/* These next three are also for internal core Perl use only: case-change
1294 * helper macros */
1295#define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \
1296 ? (c) \
1297 : (IN_UTF8_CTYPE_LOCALE) \
1298 ? PL_latin1_lc[ (U8) (c) ] \
1299 : function((cast)(c)))
1300
1301/* Note that the result can be larger than a byte in a UTF-8 locale. It
1302 * returns a single value, so can't adequately return the upper case of LATIN
1303 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1304 * values "SS"); instead it asserts against that under DEBUGGING, and
1305 * otherwise returns its input */
1306#define _generic_toUPPER_LC(c, function, cast) \
1307 (! FITS_IN_8_BITS(c) \
1308 ? (c) \
1309 : ((! IN_UTF8_CTYPE_LOCALE) \
1310 ? function((cast)(c)) \
1311 : ((((U8)(c)) == MICRO_SIGN) \
1312 ? GREEK_CAPITAL_LETTER_MU \
1313 : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \
1314 ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \
1315 : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \
1316 ? (__ASSERT_(0) (c)) \
1317 : PL_mod_latin1_uc[ (U8) (c) ])))))
1318
1319/* Note that the result can be larger than a byte in a UTF-8 locale. It
1320 * returns a single value, so can't adequately return the fold case of LATIN
1321 * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
1322 * values "ss"); instead it asserts against that under DEBUGGING, and
1323 * otherwise returns its input */
1324#define _generic_toFOLD_LC(c, function, cast) \
1325 ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \
1326 ? GREEK_SMALL_LETTER_MU \
1327 : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \
1328 || (c) != LATIN_SMALL_LETTER_SHARP_S) \
1329 _generic_toLOWER_LC(c, function, cast)))
1330
1331/* Use the libc versions for these if available. */
1332#if defined(HAS_ISASCII)
1333# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c)))
1334#else
1335# define isASCII_LC(c) isASCII(c)
1336#endif
1337
1338#if defined(HAS_ISBLANK)
1339# define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank)
1340#else /* Unlike isASCII, varies if in a UTF-8 locale */
1341# define isBLANK_LC(c) (IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c)
1342#endif
1343
1344#define _LC_CAST U8
1345
1346#ifdef WIN32
1347 /* The Windows functions don't bother to follow the POSIX standard, which
1348 * for example says that something can't both be a printable and a control.
1349 * But Windows treats the \t control as a printable, and does such things
1350 * as making superscripts into both digits and punctuation. This tames
1351 * these flaws by assuming that the definitions of both controls and space
1352 * are correct, and then making sure that other definitions don't have
1353 * weirdnesses, by making sure that isalnum() isn't also ispunct(), etc.
1354 * Not all possible weirdnesses are checked for, just the ones that were
1355 * detected on actual Microsoft code pages */
1356
1357# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1358# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1359
1360# define isALPHA_LC(c) (_generic_LC(c, _CC_ALPHA, isalpha) && isALPHANUMERIC_LC(c))
1361# define isALPHANUMERIC_LC(c) (_generic_LC(c, _CC_ALPHANUMERIC, isalnum) && ! isPUNCT_LC(c))
1362# define isDIGIT_LC(c) (_generic_LC(c, _CC_DIGIT, isdigit) && isALPHANUMERIC_LC(c))
1363# define isGRAPH_LC(c) (_generic_LC(c, _CC_GRAPH, isgraph) && isPRINT_LC(c))
1364# define isIDFIRST_LC(c) (((c) == '_') || (_generic_LC(c, _CC_IDFIRST, isalpha) && ! isPUNCT_LC(c)))
1365# define isLOWER_LC(c) (_generic_LC(c, _CC_LOWER, islower) && isALPHA_LC(c))
1366# define isPRINT_LC(c) (_generic_LC(c, _CC_PRINT, isprint) && ! isCNTRL_LC(c))
1367# define isPUNCT_LC(c) (_generic_LC(c, _CC_PUNCT, ispunct) && ! isCNTRL_LC(c))
1368# define isUPPER_LC(c) (_generic_LC(c, _CC_UPPER, isupper) && isALPHA_LC(c))
1369# define isWORDCHAR_LC(c) (((c) == '_') || isALPHANUMERIC_LC(c))
1370# define isXDIGIT_LC(c) (_generic_LC(c, _CC_XDIGIT, isxdigit) && isALPHANUMERIC_LC(c))
1371
1372# define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1373# define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1374# define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
1375
1376#elif defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
1377 /* For most other platforms */
1378
1379# define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha)
1380# define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum)
1381# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
1382# define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit)
1383# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph)
1384# define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha)
1385# define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower)
1386# define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint)
1387# define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct)
1388# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
1389# define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper)
1390# define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum)
1391# define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit)
1392
1393
1394# define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
1395# define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
1396# define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
1397
1398#else /* The final fallback position */
1399
1400# define isALPHA_LC(c) (isascii(c) && isalpha(c))
1401# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
1402# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
1403# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
1404# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
1405# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
1406# define isLOWER_LC(c) (isascii(c) && islower(c))
1407# define isPRINT_LC(c) (isascii(c) && isprint(c))
1408# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
1409# define isSPACE_LC(c) (isascii(c) && isspace(c))
1410# define isUPPER_LC(c) (isascii(c) && isupper(c))
1411# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
1412# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
1413
1414# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
1415# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
1416# define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c))
1417
1418#endif
1419
1420#define isIDCONT(c) isWORDCHAR(c)
1421#define isIDCONT_A(c) isWORDCHAR_A(c)
1422#define isIDCONT_L1(c) isWORDCHAR_L1(c)
1423#define isIDCONT_LC(c) isWORDCHAR_LC(c)
1424#define isPSXSPC_LC(c) isSPACE_LC(c)
1425
1426/* For internal core Perl use only: the base macros for defining macros like
1427 * isALPHA_uni. 'c' is the code point to check. 'classnum' is the POSIX class
1428 * number defined earlier in this file. _generic_uni() is used for POSIX
1429 * classes where there is a macro or function 'above_latin1' that takes the
1430 * single argument 'c' and returns the desired value. These exist for those
1431 * classes which have simple definitions, avoiding the overhead of a hash
1432 * lookup or inversion list binary search. _generic_swash_uni() can be used
1433 * for classes where that overhead is faster than a direct lookup.
1434 * _generic_uni() won't compile if 'c' isn't unsigned, as it won't match the
1435 * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so
1436 * have duplicate checks here, so could create versions of the macros that
1437 * don't, but experiments show that gcc optimizes them out anyway. */
1438
1439/* Note that all ignore 'use bytes' */
1440#define _generic_uni(classnum, above_latin1, c) ((c) < 256 \
1441 ? _generic_isCC(c, classnum) \
1442 : above_latin1(c))
1443#define _generic_swash_uni(classnum, c) ((c) < 256 \
1444 ? _generic_isCC(c, classnum) \
1445 : _is_uni_FOO(classnum, c))
1446#define isALPHA_uni(c) _generic_swash_uni(_CC_ALPHA, c)
1447#define isALPHANUMERIC_uni(c) _generic_swash_uni(_CC_ALPHANUMERIC, c)
1448#define isASCII_uni(c) isASCII(c)
1449#define isBLANK_uni(c) _generic_uni(_CC_BLANK, is_HORIZWS_cp_high, c)
1450#define isCNTRL_uni(c) isCNTRL_L1(c) /* All controls are in Latin1 */
1451#define isDIGIT_uni(c) _generic_swash_uni(_CC_DIGIT, c)
1452#define isGRAPH_uni(c) _generic_swash_uni(_CC_GRAPH, c)
1453#define isIDCONT_uni(c) _generic_uni(_CC_WORDCHAR, _is_uni_perl_idcont, c)
1454#define isIDFIRST_uni(c) _generic_uni(_CC_IDFIRST, _is_uni_perl_idstart, c)
1455#define isLOWER_uni(c) _generic_swash_uni(_CC_LOWER, c)
1456#define isPRINT_uni(c) _generic_swash_uni(_CC_PRINT, c)
1457
1458/* Posix and regular space are identical above Latin1 */
1459#define isPSXSPC_uni(c) _generic_uni(_CC_PSXSPC, is_XPERLSPACE_cp_high, c)
1460
1461#define isPUNCT_uni(c) _generic_swash_uni(_CC_PUNCT, c)
1462#define isSPACE_uni(c) _generic_uni(_CC_SPACE, is_XPERLSPACE_cp_high, c)
1463#define isUPPER_uni(c) _generic_swash_uni(_CC_UPPER, c)
1464#define isVERTWS_uni(c) _generic_uni(_CC_VERTSPACE, is_VERTWS_cp_high, c)
1465#define isWORDCHAR_uni(c) _generic_swash_uni(_CC_WORDCHAR, c)
1466#define isXDIGIT_uni(c) _generic_uni(_CC_XDIGIT, is_XDIGIT_cp_high, c)
1467
1468#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
1469#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
1470#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
1471#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
1472
1473/* For internal core Perl use only: the base macros for defining macros like
1474 * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code
1475 * point, not just 0-255. Like _generic_uni, there are two versions, one for
1476 * simple class definitions; the other for more complex. These are like
1477 * _generic_uni, so see it for more info. */
1478#define _generic_LC_uvchr(latin1, above_latin1, c) \
1479 (c < 256 ? latin1(c) : above_latin1(c))
1480#define _generic_LC_swash_uvchr(latin1, classnum, c) \
1481 (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
1482
1483#define isALPHA_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHA_LC, _CC_ALPHA, c)
1484#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_swash_uvchr(isALPHANUMERIC_LC, \
1485 _CC_ALPHANUMERIC, c)
1486#define isASCII_LC_uvchr(c) isASCII_LC(c)
1487#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, is_HORIZWS_cp_high, c)
1488#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
1489#define isDIGIT_LC_uvchr(c) _generic_LC_swash_uvchr(isDIGIT_LC, _CC_DIGIT, c)
1490#define isGRAPH_LC_uvchr(c) _generic_LC_swash_uvchr(isGRAPH_LC, _CC_GRAPH, c)
1491#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
1492 _is_uni_perl_idcont, c)
1493#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
1494 _is_uni_perl_idstart, c)
1495#define isLOWER_LC_uvchr(c) _generic_LC_swash_uvchr(isLOWER_LC, _CC_LOWER, c)
1496#define isPRINT_LC_uvchr(c) _generic_LC_swash_uvchr(isPRINT_LC, _CC_PRINT, c)
1497#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) /* space is identical to posix
1498 space under locale */
1499#define isPUNCT_LC_uvchr(c) _generic_LC_swash_uvchr(isPUNCT_LC, _CC_PUNCT, c)
1500#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
1501 is_XPERLSPACE_cp_high, c)
1502#define isUPPER_LC_uvchr(c) _generic_LC_swash_uvchr(isUPPER_LC, _CC_UPPER, c)
1503#define isWORDCHAR_LC_uvchr(c) _generic_LC_swash_uvchr(isWORDCHAR_LC, \
1504 _CC_WORDCHAR, c)
1505#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
1506
1507#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
1508
1509/* For internal core Perl use only: the base macros for defining macros like
1510 * isALPHA_utf8. These are like the earlier defined macros, but take an input
1511 * UTF-8 encoded string 'p'. If the input is in the Latin1 range, use
1512 * the Latin1 macro 'classnum' on 'p'. Otherwise use the value given by the
1513 * 'utf8' parameter. This relies on the fact that ASCII characters have the
1514 * same representation whether utf8 or not. Note that it assumes that the utf8
1515 * has been validated, and ignores 'use bytes' */
1516#define _generic_utf8(classnum, p, utf8) (UTF8_IS_INVARIANT(*(p)) \
1517 ? _generic_isCC(*(p), classnum) \
1518 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1519 ? _generic_isCC( \
1520 TWO_BYTE_UTF8_TO_NATIVE(*(p), \
1521 *((p)+1 )), \
1522 classnum) \
1523 : utf8)
1524/* Like the above, but calls 'above_latin1(p)' to get the utf8 value. 'above_latin1'
1525 * can be a macro */
1526#define _generic_func_utf8(classnum, above_latin1, p) \
1527 _generic_utf8(classnum, p, above_latin1(p))
1528/* Like the above, but passes classnum to _isFOO_utf8(), instead of having an
1529 * 'above_latin1' parameter */
1530#define _generic_swash_utf8(classnum, p) \
1531 _generic_utf8(classnum, p, _is_utf8_FOO(classnum, p))
1532
1533/* Like the above, but should be used only when it is known that there are no
1534 * characters in the upper-Latin1 range (128-255 on ASCII platforms) which the
1535 * class is TRUE for. Hence it can skip the tests for this range.
1536 * 'above_latin1' should include its arguments */
1537#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
1538 (UTF8_IS_INVARIANT(*(p)) \
1539 ? _generic_isCC(*(p), classnum) \
1540 : (UTF8_IS_ABOVE_LATIN1(*(p))) \
1541 ? above_latin1 \
1542 : 0)
1543
1544/* NOTE that some of these macros have very similar ones in regcharclass.h.
1545 * For example, there is (at the time of this writing) an 'is_SPACE_utf8()'
1546 * there, differing in name only by an underscore from the one here
1547 * 'isSPACE_utf8(). The difference is that the ones here are probably more
1548 * efficient and smaller, using an O(1) array lookup for Latin1-range code
1549 * points; the regcharclass.h ones are implemented as a series of
1550 * "if-else-if-else ..." */
1551
1552#define isALPHA_utf8(p) _generic_swash_utf8(_CC_ALPHA, p)
1553#define isALPHANUMERIC_utf8(p) _generic_swash_utf8(_CC_ALPHANUMERIC, p)
1554#define isASCII_utf8(p) isASCII(*p) /* Because ASCII is invariant under
1555 utf8, the non-utf8 macro works
1556 */
1557#define isBLANK_utf8(p) _generic_func_utf8(_CC_BLANK, is_HORIZWS_high, p)
1558
1559#ifdef EBCDIC
1560 /* Because all controls are UTF-8 invariants in EBCDIC, we can use this
1561 * more efficient macro instead of the more general one */
1562# define isCNTRL_utf8(p) isCNTRL_L1(p)
1563#else
1564# define isCNTRL_utf8(p) _generic_utf8(_CC_CNTRL, p, 0)
1565#endif
1566
1567#define isDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_DIGIT, p, \
1568 _is_utf8_FOO(_CC_DIGIT, p))
1569#define isGRAPH_utf8(p) _generic_swash_utf8(_CC_GRAPH, p)
1570#define isIDCONT_utf8(p) _generic_func_utf8(_CC_WORDCHAR, \
1571 _is_utf8_perl_idcont, p)
1572
1573/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
1574 * IDFIRST is an alnum. See
1575 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
1576 * ever wanted to know about. (In the ASCII range, there isn't a difference.)
1577 * This used to be not the XID version, but we decided to go with the more
1578 * modern Unicode definition */
1579#define isIDFIRST_utf8(p) _generic_func_utf8(_CC_IDFIRST, \
1580 _is_utf8_perl_idstart, p)
1581
1582#define isLOWER_utf8(p) _generic_swash_utf8(_CC_LOWER, p)
1583#define isPRINT_utf8(p) _generic_swash_utf8(_CC_PRINT, p)
1584
1585/* Posix and regular space are identical above Latin1 */
1586#define isPSXSPC_utf8(p) _generic_func_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
1587
1588#define isPUNCT_utf8(p) _generic_swash_utf8(_CC_PUNCT, p)
1589#define isSPACE_utf8(p) _generic_func_utf8(_CC_SPACE, is_XPERLSPACE_high, p)
1590#define isUPPER_utf8(p) _generic_swash_utf8(_CC_UPPER, p)
1591#define isVERTWS_utf8(p) _generic_func_utf8(_CC_VERTSPACE, is_VERTWS_high, p)
1592#define isWORDCHAR_utf8(p) _generic_swash_utf8(_CC_WORDCHAR, p)
1593#define isXDIGIT_utf8(p) _generic_utf8_no_upper_latin1(_CC_XDIGIT, p, \
1594 is_XDIGIT_high(p))
1595
1596#define toFOLD_utf8(p,s,l) to_utf8_fold(p,s,l)
1597#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
1598#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1599#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
1600
1601/* For internal core Perl use only: the base macros for defining macros like
1602 * isALPHA_LC_utf8. These are like _generic_utf8, but if the first code point
1603 * in 'p' is within the 0-255 range, it uses locale rules from the passed-in
1604 * 'macro' parameter */
1605#define _generic_LC_utf8(macro, p, utf8) \
1606 (UTF8_IS_INVARIANT(*(p)) \
1607 ? macro(*(p)) \
1608 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
1609 ? macro(TWO_BYTE_UTF8_TO_NATIVE(*(p), *((p)+1))) \
1610 : utf8)
1611
1612#define _generic_LC_swash_utf8(macro, classnum, p) \
1613 _generic_LC_utf8(macro, p, _is_utf8_FOO(classnum, p))
1614#define _generic_LC_func_utf8(macro, above_latin1, p) \
1615 _generic_LC_utf8(macro, p, above_latin1(p))
1616
1617#define isALPHANUMERIC_LC_utf8(p) _generic_LC_swash_utf8(isALPHANUMERIC_LC, \
1618 _CC_ALPHANUMERIC, p)
1619#define isALPHA_LC_utf8(p) _generic_LC_swash_utf8(isALPHA_LC, _CC_ALPHA, p)
1620#define isASCII_LC_utf8(p) isASCII_LC(*p)
1621#define isBLANK_LC_utf8(p) _generic_LC_func_utf8(isBLANK_LC, is_HORIZWS_high, p)
1622#define isCNTRL_LC_utf8(p) _generic_LC_utf8(isCNTRL_LC, p, 0)
1623#define isDIGIT_LC_utf8(p) _generic_LC_swash_utf8(isDIGIT_LC, _CC_DIGIT, p)
1624#define isGRAPH_LC_utf8(p) _generic_LC_swash_utf8(isGRAPH_LC, _CC_GRAPH, p)
1625#define isIDCONT_LC_utf8(p) _generic_LC_func_utf8(isIDCONT_LC, _is_utf8_perl_idcont, p)
1626#define isIDFIRST_LC_utf8(p) _generic_LC_func_utf8(isIDFIRST_LC, _is_utf8_perl_idstart, p)
1627#define isLOWER_LC_utf8(p) _generic_LC_swash_utf8(isLOWER_LC, _CC_LOWER, p)
1628#define isPRINT_LC_utf8(p) _generic_LC_swash_utf8(isPRINT_LC, _CC_PRINT, p)
1629#define isPSXSPC_LC_utf8(p) isSPACE_LC_utf8(p) /* space is identical to posix
1630 space under locale */
1631#define isPUNCT_LC_utf8(p) _generic_LC_swash_utf8(isPUNCT_LC, _CC_PUNCT, p)
1632#define isSPACE_LC_utf8(p) _generic_LC_func_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
1633#define isUPPER_LC_utf8(p) _generic_LC_swash_utf8(isUPPER_LC, _CC_UPPER, p)
1634#define isWORDCHAR_LC_utf8(p) _generic_LC_swash_utf8(isWORDCHAR_LC, \
1635 _CC_WORDCHAR, p)
1636#define isXDIGIT_LC_utf8(p) _generic_LC_func_utf8(isXDIGIT_LC, is_XDIGIT_high, p)
1637
1638/* Macros for backwards compatibility and for completeness when the ASCII and
1639 * Latin1 values are identical */
1640#define isALPHAU(c) isALPHA_L1(c)
1641#define isDIGIT_L1(c) isDIGIT_A(c)
1642#define isOCTAL(c) isOCTAL_A(c)
1643#define isOCTAL_L1(c) isOCTAL_A(c)
1644#define isXDIGIT_L1(c) isXDIGIT_A(c)
1645#define isALNUM(c) isWORDCHAR(c)
1646#define isALNUMU(c) isWORDCHAR_L1(c)
1647#define isALNUM_LC(c) isWORDCHAR_LC(c)
1648#define isALNUM_uni(c) isWORDCHAR_uni(c)
1649#define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c)
1650#define isALNUM_utf8(p) isWORDCHAR_utf8(p)
1651#define isALNUM_LC_utf8(p) isWORDCHAR_LC_utf8(p)
1652#define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */
1653#define isALNUMC_L1(c) isALPHANUMERIC_L1(c)
1654#define isALNUMC(c) isALPHANUMERIC(c)
1655#define isALNUMC_LC(c) isALPHANUMERIC_LC(c)
1656#define isALNUMC_uni(c) isALPHANUMERIC_uni(c)
1657#define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c)
1658#define isALNUMC_utf8(p) isALPHANUMERIC_utf8(p)
1659#define isALNUMC_LC_utf8(p) isALPHANUMERIC_LC_utf8(p)
1660
1661/* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII,
1662 * except that they don't necessarily mean the same characters, e.g. CTRL-D is
1663 * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC.
1664 * '?' is special-cased on EBCDIC to APC, which is the control there that is
1665 * the outlier from the block that contains the other controls, just like
1666 * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0
1667 * block. If it weren't special cased, it would yield a non-control.
1668 * The conversion works both ways, so CTRL('D') is 4, and CTRL(4) is D, etc. */
1669#ifndef EBCDIC
1670# define toCTRL(c) (toUPPER(c) ^ 64)
1671#else
1672# define toCTRL(c) ((isPRINT_A(c)) \
1673 ? UNLIKELY((c) == '?') \
1674 ? QUESTION_MARK_CTRL \
1675 : (NATIVE_TO_LATIN1(toUPPER(c)) ^ 64) \
1676 : UNLIKELY((c) == QUESTION_MARK_CTRL) \
1677 ? ((c) == '?') \
1678 : (LATIN1_TO_NATIVE((c) ^ 64)))
1679#endif
1680
1681/* Line numbers are unsigned, 32 bits. */
1682typedef U32 line_t;
1683#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
1684
1685/* Helpful alias for version prescan */
1686#define is_LAX_VERSION(a,b) \
1687 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1688
1689#define is_STRICT_VERSION(a,b) \
1690 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1691
1692#define BADVERSION(a,b,c) \
1693 if (b) { \
1694 *b = c; \
1695 } \
1696 return a;
1697
1698/* Converts a character known to represent a hexadecimal digit (0-9, A-F, or
1699 * a-f) to its numeric value. READ_XDIGIT's argument is a string pointer,
1700 * which is advanced. The input is validated only by an assert() in DEBUGGING
1701 * builds. In both ASCII and EBCDIC the last 4 bits of the digits are 0-9; and
1702 * the last 4 bits of A-F and a-f are 1-6, so adding 9 yields 10-15 */
1703#define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) (0xf & (isDIGIT(c) \
1704 ? (c) \
1705 : ((c) + 9))))
1706#define READ_XDIGIT(s) (__ASSERT_(isXDIGIT(*s)) (0xf & (isDIGIT(*(s)) \
1707 ? (*(s)++) \
1708 : (*(s)++ + 9))))
1709
1710/* Converts a character known to represent an octal digit (0-7) to its numeric
1711 * value. The input is validated only by an assert() in DEBUGGING builds. In
1712 * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */
1713#define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c)))
1714
1715/* Efficiently returns a boolean as to if two native characters are equivalent
1716 * case-insenstively. At least one of the characters must be one of [A-Za-z];
1717 * the ALPHA in the name is to remind you of that. This is asserted() in
1718 * DEBUGGING builds. Because [A-Za-z] are invariant under UTF-8, this macro
1719 * works (on valid input) for both non- and UTF-8-encoded bytes.
1720 *
1721 * When one of the inputs is a compile-time constant and gets folded by the
1722 * compiler, this reduces to an AND and a TEST. On both EBCDIC and ASCII
1723 * machines, 'A' and 'a' differ by a single bit; the same with the upper and
1724 * lower case of all other ASCII-range alphabetics. On ASCII platforms, they
1725 * are 32 apart; on EBCDIC, they are 64. At compile time, this uses an
1726 * exclusive 'or' to find that bit and then inverts it to form a mask, with
1727 * just a single 0, in the bit position where the upper- and lowercase differ.
1728 * */
1729#define isALPHA_FOLD_EQ(c1, c2) \
1730 (__ASSERT_(isALPHA_A(c1) || isALPHA_A(c2)) \
1731 ((c1) & ~('A' ^ 'a')) == ((c2) & ~('A' ^ 'a')))
1732#define isALPHA_FOLD_NE(c1, c2) (! isALPHA_FOLD_EQ((c1), (c2)))
1733
1734/*
1735=head1 Memory Management
1736
1737=for apidoc Am|void|Newx|void* ptr|int nitems|type
1738The XSUB-writer's interface to the C C<malloc> function.
1739
1740Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1741
1742In 5.9.3, Newx() and friends replace the older New() API, and drops
1743the first parameter, I<x>, a debug aid which allowed callers to identify
1744themselves. This aid has been superseded by a new build option,
1745PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
1746there for use in XS modules supporting older perls.
1747
1748=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
1749The XSUB-writer's interface to the C C<malloc> function, with
1750cast. See also C<Newx>.
1751
1752Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1753
1754=for apidoc Am|void|Newxz|void* ptr|int nitems|type
1755The XSUB-writer's interface to the C C<malloc> function. The allocated
1756memory is zeroed with C<memzero>. See also C<Newx>.
1757
1758Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1759
1760=for apidoc Am|void|Renew|void* ptr|int nitems|type
1761The XSUB-writer's interface to the C C<realloc> function.
1762
1763Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1764
1765=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1766The XSUB-writer's interface to the C C<realloc> function, with
1767cast.
1768
1769Memory obtained by this should B<ONLY> be freed with L<"Safefree">.
1770
1771=for apidoc Am|void|Safefree|void* ptr
1772The XSUB-writer's interface to the C C<free> function.
1773
1774This should B<ONLY> be used on memory obtained using L<"Newx"> and friends.
1775
1776=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1777The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
1778source, C<dest> is the destination, C<nitems> is the number of items, and
1779C<type> is the type. Can do overlapping moves. See also C<Copy>.
1780
1781=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1782Like C<Move> but returns dest. Useful
1783for encouraging compilers to tail-call
1784optimise.
1785
1786=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1787The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
1788source, C<dest> is the destination, C<nitems> is the number of items, and
1789C<type> is the type. May fail on overlapping copies. See also C<Move>.
1790
1791=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1792
1793Like C<Copy> but returns dest. Useful
1794for encouraging compilers to tail-call
1795optimise.
1796
1797=for apidoc Am|void|Zero|void* dest|int nitems|type
1798
1799The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1800destination, C<nitems> is the number of items, and C<type> is the type.
1801
1802=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1803
1804Like C<Zero> but returns dest. Useful
1805for encouraging compilers to tail-call
1806optimise.
1807
1808=for apidoc Am|void|StructCopy|type *src|type *dest|type
1809This is an architecture-independent macro to copy one structure to another.
1810
1811=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1812
1813Fill up memory with a byte pattern (a byte repeated over and over
1814again) that hopefully catches attempts to access uninitialized memory.
1815
1816=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1817
1818PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1819
1820=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
1821
1822PoisonWith(0xEF) for catching access to freed memory.
1823
1824=for apidoc Am|void|Poison|void* dest|int nitems|type
1825
1826PoisonWith(0xEF) for catching access to freed memory.
1827
1828=cut */
1829
1830/* Maintained for backwards-compatibility only. Use newSV() instead. */
1831#ifndef PERL_CORE
1832#define NEWSV(x,len) newSV(len)
1833#endif
1834
1835#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1836
1837/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1838 * overly eager compilers that will bleat about e.g.
1839 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
1840#ifdef PERL_MALLOC_WRAP
1841#define MEM_WRAP_CHECK(n,t) \
1842 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (croak_memory_wrap(),0))
1843#define MEM_WRAP_CHECK_1(n,t,a) \
1844 (void)(UNLIKELY(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t)) && (Perl_croak_nocontext("%s",(a)),0))
1845#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
1846
1847#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)))
1848#else
1849
1850#define MEM_WRAP_CHECK(n,t)
1851#define MEM_WRAP_CHECK_1(n,t,a)
1852#define MEM_WRAP_CHECK_2(n,t,a,b)
1853#define MEM_WRAP_CHECK_(n,t)
1854
1855#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1856
1857#endif
1858
1859#ifdef PERL_MEM_LOG
1860/*
1861 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
1862 * go through functions, which are handy for debugging breakpoints, but
1863 * which more importantly get the immediate calling environment (file and
1864 * line number, and C function name if available) passed in. This info can
1865 * then be used for logging the calls, for which one gets a sample
1866 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
1867 *
1868 * Known problems:
1869 * - not all memory allocs get logged, only those
1870 * that go through Newx() and derivatives (while all
1871 * Safefrees do get logged)
1872 * - __FILE__ and __LINE__ do not work everywhere
1873 * - __func__ or __FUNCTION__ even less so
1874 * - I think more goes on after the perlio frees but
1875 * the thing is that STDERR gets closed (as do all
1876 * the file descriptors)
1877 * - no deeper calling stack than the caller of the Newx()
1878 * or the kind, but do I look like a C reflection/introspection
1879 * utility to you?
1880 * - the function prototypes for the logging functions
1881 * probably should maybe be somewhere else than handy.h
1882 * - one could consider inlining (macrofying) the logging
1883 * for speed, but I am too lazy
1884 * - one could imagine recording the allocations in a hash,
1885 * (keyed by the allocation address?), and maintain that
1886 * through reallocs and frees, but how to do that without
1887 * any News() happening...?
1888 * - lots of -Ddefines to get useful/controllable output
1889 * - lots of ENV reads
1890 */
1891
1892PERL_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);
1893
1894PERL_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);
1895
1896PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
1897
1898# ifdef PERL_CORE
1899# ifndef PERL_MEM_LOG_NOIMPL
1900enum mem_log_type {
1901 MLT_ALLOC,
1902 MLT_REALLOC,
1903 MLT_FREE,
1904 MLT_NEW_SV,
1905 MLT_DEL_SV
1906};
1907# endif
1908# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
1909void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1910void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1911# endif
1912# endif
1913
1914#endif
1915
1916#ifdef PERL_MEM_LOG
1917#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1918#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
1919#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
1920#endif
1921
1922#ifndef MEM_LOG_ALLOC
1923#define MEM_LOG_ALLOC(n,t,a) (a)
1924#endif
1925#ifndef MEM_LOG_REALLOC
1926#define MEM_LOG_REALLOC(n,t,v,a) (a)
1927#endif
1928#ifndef MEM_LOG_FREE
1929#define MEM_LOG_FREE(a) (a)
1930#endif
1931
1932#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1933#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1934#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
1935
1936#ifndef PERL_CORE
1937/* pre 5.9.x compatibility */
1938#define New(x,v,n,t) Newx(v,n,t)
1939#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
1940#define Newz(x,v,n,t) Newxz(v,n,t)
1941#endif
1942
1943#define Renew(v,n,t) \
1944 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1945#define Renewc(v,n,t,c) \
1946 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
1947
1948#ifdef PERL_POISON
1949#define Safefree(d) \
1950 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
1951#else
1952#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
1953#endif
1954
1955#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1956#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1957#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
1958
1959#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1960#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1961#ifdef HAS_MEMSET
1962#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
1963#else
1964/* Using bzero(), which returns void. */
1965#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
1966#endif
1967
1968#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1969#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1970#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1971#define Poison(d,n,t) PoisonFree(d,n,t)
1972
1973#ifdef PERL_POISON
1974# define PERL_POISON_EXPR(x) x
1975#else
1976# define PERL_POISON_EXPR(x)
1977#endif
1978
1979#ifdef USE_STRUCT_COPY
1980#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
1981#else
1982#define StructCopy(s,d,t) Copy(s,d,1,t)
1983#endif
1984
1985/* C_ARRAY_LENGTH is the number of elements in the C array (so you
1986 * want your zero-based indices to be less than but not equal to).
1987 *
1988 * C_ARRAY_END is one past the last: half-open/half-closed range,
1989 * not last-inclusive range. */
1990#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
1991#define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a))
1992
1993#ifdef NEED_VA_COPY
1994# ifdef va_copy
1995# define Perl_va_copy(s, d) va_copy(d, s)
1996# else
1997# if defined(__va_copy)
1998# define Perl_va_copy(s, d) __va_copy(d, s)
1999# else
2000# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
2001# endif
2002# endif
2003#endif
2004
2005/* convenience debug macros */
2006#ifdef USE_ITHREADS
2007#define pTHX_FORMAT "Perl interpreter: 0x%p"
2008#define pTHX__FORMAT ", Perl interpreter: 0x%p"
2009#define pTHX_VALUE_ (void *)my_perl,
2010#define pTHX_VALUE (void *)my_perl
2011#define pTHX__VALUE_ ,(void *)my_perl,
2012#define pTHX__VALUE ,(void *)my_perl
2013#else
2014#define pTHX_FORMAT
2015#define pTHX__FORMAT
2016#define pTHX_VALUE_
2017#define pTHX_VALUE
2018#define pTHX__VALUE_
2019#define pTHX__VALUE
2020#endif /* USE_ITHREADS */
2021
2022/* Perl_deprecate was not part of the public API, and did not have a deprecate()
2023 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
2024 CPAN::Unpack show any users outside the core. */
2025#ifdef PERL_CORE
2026# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2027#endif
2028
2029/* Internal macros to deal with gids and uids */
2030#ifdef PERL_CORE
2031
2032# if Uid_t_size > IVSIZE
2033# define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid))
2034# define SvUID(sv) SvNV(sv)
2035# else
2036# if Uid_t_sign <= 0
2037# define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid))
2038# define SvUID(sv) SvIV(sv)
2039# else
2040# define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid))
2041# define SvUID(sv) SvUV(sv)
2042# endif
2043# endif /* Uid_t_size */
2044
2045# if Gid_t_size > IVSIZE
2046# define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid))
2047# define SvGID(sv) SvNV(sv)
2048# else
2049# if Gid_t_sign <= 0
2050# define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid))
2051# define SvGID(sv) SvIV(sv)
2052# else
2053# define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid))
2054# define SvGID(sv) SvUV(sv)
2055# endif
2056# endif /* Gid_t_size */
2057
2058#endif
2059
2060#endif /* HANDY_H */
2061
2062/*
2063 * Local variables:
2064 * c-indentation-style: bsd
2065 * c-basic-offset: 4
2066 * indent-tabs-mode: nil
2067 * End:
2068 *
2069 * ex: set ts=8 sts=4 sw=4 et:
2070 */
Unnamed repository; edit this file 'description' to name the repository.