This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Make the new STR_WITH_LEN() affected compile under -Dusethreads.
[perl5.git] / handy.h
CommitLineData
a0d0e21e 1/* handy.h
a687059c 2 *
4bb101f2 3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999,
1d325971 4 * 2000, 2001, 2002, 2004, 2005 by Larry Wall and others
a687059c 5 *
6e21c824
LW
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
8d063cd8 8 *
8d063cd8
LW
9 */
10
85e6fe83 11#if !defined(__STDC__)
378cc40b
LW
12#ifdef NULL
13#undef NULL
14#endif
a687059c
LW
15#ifndef I286
16# define NULL 0
17#else
18# define NULL 0L
19#endif
85e6fe83
LW
20#endif
21
378cc40b 22#define Null(type) ((type)NULL)
954c1994
GS
23
24/*
ccfc67b7 25=head1 Handy Values
954c1994 26
3609ea0d 27=for apidoc AmU||Nullch
ccfc67b7 28Null character pointer.
2307c6d0 29
954c1994
GS
30=for apidoc AmU||Nullsv
31Null SV pointer.
32
33=cut
34*/
35
8d063cd8 36#define Nullch Null(char*)
760ac839 37#define Nullfp Null(PerlIO*)
79072805 38#define Nullsv Null(SV*)
8d063cd8 39
641d3f0b
PP
40#ifdef TRUE
41#undef TRUE
42#endif
43#ifdef FALSE
44#undef FALSE
45#endif
46#define TRUE (1)
47#define FALSE (0)
48
27d4fb96
PP
49
50/* XXX Configure ought to have a test for a boolean type, if I can
51 just figure out all the headers such a test needs.
52 Andy Dougherty August 1996
53*/
8e84507e 54/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
55 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
56 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 57 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
58 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
59 g++ can be identified by __GNUG__.
60 Andy Dougherty February 2000
5d94fbed 61*/
3609ea0d 62#ifdef __GNUG__ /* GNU g++ has bool built-in */
5d94fbed 63# ifndef HAS_BOOL
c1d22f6b 64# define HAS_BOOL 1
5d94fbed 65# endif
5d94fbed
AD
66#endif
67
641d3f0b
PP
68/* The NeXT dynamic loader headers will not build with the bool macro
69 So declare them now to clear confusion.
70*/
8f1f23e8 71#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
72# undef FALSE
73# undef TRUE
74 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
75# define ENUM_BOOL 1
76# ifndef HAS_BOOL
77# define HAS_BOOL 1
78# endif /* !HAS_BOOL */
8f1f23e8 79#endif /* NeXT || __NeXT__ */
641d3f0b 80
5d94fbed 81#ifndef HAS_BOOL
61bb5906 82# if defined(UTS) || defined(VMS)
5d94fbed
AD
83# define bool int
84# else
85# define bool char
86# endif
c1d22f6b 87# define HAS_BOOL 1
a687059c 88#endif
0d3e774c 89
46c6c7e2 90/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
91 * XXX Should really be a Configure probe, with HAS__FUNCTION__
92 * and FUNCTION__ as results.
93 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
94#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
95# define FUNCTION__ __func__
96#else
97# if (defined(_MSC_VER) && _MSC_VER < 1300) || /* Pre-MSVC 7.0 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 98 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
46c6c7e2
JH
99# define FUNCTION__ ""
100# else
101# define FUNCTION__ __FUNCTION__ /* Common extension. */
102# endif
103#endif
104
27d4fb96
PP
105/* XXX A note on the perl source internal type system. The
106 original intent was that I32 be *exactly* 32 bits.
107
108 Currently, we only guarantee that I32 is *at least* 32 bits.
109 Specifically, if int is 64 bits, then so is I32. (This is the case
110 for the Cray.) This has the advantage of meshing nicely with
111 standard library calls (where we pass an I32 and the library is
112 expecting an int), but the disadvantage that an I32 is not 32 bits.
113 Andy Dougherty August 1996
24fef2a7 114
dc45a647
MB
115 There is no guarantee that there is *any* integral type with
116 exactly 32 bits. It is perfectly legal for a system to have
117 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 118
dc45a647
MB
119 Similarly, there is no guarantee that I16 and U16 have exactly 16
120 bits.
693762b4 121
8e84507e
NIS
122 For dealing with issues that may arise from various 32/64-bit
123 systems, we will ask Configure to check out
8175356b 124
3609ea0d
JH
125 SHORTSIZE == sizeof(short)
126 INTSIZE == sizeof(int)
127 LONGSIZE == sizeof(long)
dc45a647 128 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 129 PTRSIZE == sizeof(void *)
dc45a647
MB
130 DOUBLESIZE == sizeof(double)
131 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 132
27d4fb96
PP
133*/
134
69512466
JH
135#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
136# include <inttypes.h>
dd0eed91
JH
137# ifdef INT32_MIN_BROKEN
138# undef INT32_MIN
139# define INT32_MIN (-2147483647-1)
140# endif
141# ifdef INT64_MIN_BROKEN
142# undef INT64_MIN
143# define INT64_MIN (-9223372036854775807LL-1)
144# endif
69512466
JH
145#endif
146
8175356b
JH
147typedef I8TYPE I8;
148typedef U8TYPE U8;
149typedef I16TYPE I16;
150typedef U16TYPE U16;
151typedef I32TYPE I32;
152typedef U32TYPE U32;
6b8eaf93
JH
153#ifdef PERL_CORE
154# ifdef HAS_QUAD
8175356b
JH
155typedef I64TYPE I64;
156typedef U64TYPE U64;
6b8eaf93
JH
157# endif
158#endif /* PERL_CORE */
8175356b 159
69512466
JH
160#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
161# ifndef UINT64_C /* usually from <inttypes.h> */
162# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
163# define INT64_C(c) CAT2(c,LL)
164# define UINT64_C(c) CAT2(c,ULL)
165# else
166# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
167# define INT64_C(c) CAT2(c,L)
168# define UINT64_C(c) CAT2(c,UL)
169# else
170# define INT64_C(c) ((I64TYPE)(c))
171# define UINT64_C(c) ((U64TYPE)(c))
172# endif
173# endif
e8c95190
JH
174# endif
175#endif
dc750f2e 176
cae3d67c 177/* HMB H.Merijn Brand - a placeholder for preparing Configure patches */
4f5da3e9 178#if defined(HAS_MALLOC_SIZE) && defined(HAS_SNPRINTF) && defined(HAS_VSNPRINTF)
cae3d67c 179/* Not (yet) used at top level, but mention them for metaconfig */
dc750f2e 180#endif
e8c95190 181
a22e52b9
JH
182/* Mention I8SIZE, U8SIZE, I16SIZE, U16SIZE, I32SIZE, U32SIZE,
183 I64SIZE, and U64SIZE here so that metaconfig pulls them in. */
184
d8668976 185#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 186
5ff3f7a4
GS
187/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
188 Please search CHAR_MAX in perl.h for further details. */
189#define U8_MAX UINT8_MAX
190#define U8_MIN UINT8_MIN
191
5ff3f7a4
GS
192#define I16_MAX INT16_MAX
193#define I16_MIN INT16_MIN
194#define U16_MAX UINT16_MAX
195#define U16_MIN UINT16_MIN
196
5ff3f7a4
GS
197#define I32_MAX INT32_MAX
198#define I32_MIN INT32_MIN
0e983133
GS
199#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
200# define U32_MAX UINT32_MAX
201#else
202# define U32_MAX 4294967295U
203#endif
5ff3f7a4
GS
204#define U32_MIN UINT32_MIN
205
206#else
207
5c9fa16e
KA
208/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
209 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
210#define U8_MAX PERL_UCHAR_MAX
211#define U8_MIN PERL_UCHAR_MIN
79072805 212
27d4fb96
PP
213#define I16_MAX PERL_SHORT_MAX
214#define I16_MIN PERL_SHORT_MIN
215#define U16_MAX PERL_USHORT_MAX
216#define U16_MIN PERL_USHORT_MIN
79072805 217
c4f23d77 218#if LONGSIZE > 4
27d4fb96
PP
219# define I32_MAX PERL_INT_MAX
220# define I32_MIN PERL_INT_MIN
221# define U32_MAX PERL_UINT_MAX
222# define U32_MIN PERL_UINT_MIN
79072805 223#else
27d4fb96
PP
224# define I32_MAX PERL_LONG_MAX
225# define I32_MIN PERL_LONG_MIN
226# define U32_MAX PERL_ULONG_MAX
227# define U32_MIN PERL_ULONG_MIN
79072805
LW
228#endif
229
5ff3f7a4
GS
230#endif
231
58a9a5d5 232/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
233#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
234#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
235#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
236
ff68c719 237#define Ctl(ch) ((ch) & 037)
8d063cd8 238
65177ae0 239/* concatenating with "" ensures that only literal strings are accepted as argument */
2efa8cc7
GA
240#define STR_WITH_LEN(s) (s ""), (sizeof(s)-1)
241
954c1994 242/*
ccfc67b7
JH
243=head1 Miscellaneous Functions
244
954c1994
GS
245=for apidoc Am|bool|strNE|char* s1|char* s2
246Test two strings to see if they are different. Returns true or
247false.
248
249=for apidoc Am|bool|strEQ|char* s1|char* s2
250Test two strings to see if they are equal. Returns true or false.
251
252=for apidoc Am|bool|strLT|char* s1|char* s2
253Test two strings to see if the first, C<s1>, is less than the second,
254C<s2>. Returns true or false.
255
256=for apidoc Am|bool|strLE|char* s1|char* s2
257Test two strings to see if the first, C<s1>, is less than or equal to the
258second, C<s2>. Returns true or false.
259
260=for apidoc Am|bool|strGT|char* s1|char* s2
261Test two strings to see if the first, C<s1>, is greater than the second,
262C<s2>. Returns true or false.
263
264=for apidoc Am|bool|strGE|char* s1|char* s2
265Test two strings to see if the first, C<s1>, is greater than or equal to
266the second, C<s2>. Returns true or false.
267
268=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
269Test two strings to see if they are different. The C<len> parameter
270indicates the number of bytes to compare. Returns true or false. (A
271wrapper for C<strncmp>).
272
273=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
274Test two strings to see if they are equal. The C<len> parameter indicates
275the number of bytes to compare. Returns true or false. (A wrapper for
276C<strncmp>).
277
278=cut
279*/
280
8d063cd8
LW
281#define strNE(s1,s2) (strcmp(s1,s2))
282#define strEQ(s1,s2) (!strcmp(s1,s2))
283#define strLT(s1,s2) (strcmp(s1,s2) < 0)
284#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
285#define strGT(s1,s2) (strcmp(s1,s2) > 0)
286#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
287#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
288#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 289
36477c24
PP
290#ifdef HAS_MEMCMP
291# define memNE(s1,s2,l) (memcmp(s1,s2,l))
292# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
293#else
294# define memNE(s1,s2,l) (bcmp(s1,s2,l))
295# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
296#endif
297
bbce6d69
PP
298/*
299 * Character classes.
300 *
301 * Unfortunately, the introduction of locales means that we
302 * can't trust isupper(), etc. to tell the truth. And when
303 * it comes to /\w+/ with tainting enabled, we *must* be able
304 * to trust our character classes.
305 *
306 * Therefore, the default tests in the text of Perl will be
307 * independent of locale. Any code that wants to depend on
308 * the current locale will use the tests that begin with "lc".
309 */
310
2304df62
AD
311#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
312# ifndef CTYPE256
313# define CTYPE256
314# endif
315#endif
316
954c1994 317/*
ccfc67b7
JH
318
319=head1 Character classes
320
954c1994 321=for apidoc Am|bool|isALNUM|char ch
4375e838 322Returns a boolean indicating whether the C C<char> is an ASCII alphanumeric
f1cbbd6e 323character (including underscore) or digit.
954c1994
GS
324
325=for apidoc Am|bool|isALPHA|char ch
4375e838 326Returns a boolean indicating whether the C C<char> is an ASCII alphabetic
954c1994
GS
327character.
328
329=for apidoc Am|bool|isSPACE|char ch
330Returns a boolean indicating whether the C C<char> is whitespace.
331
332=for apidoc Am|bool|isDIGIT|char ch
4375e838 333Returns a boolean indicating whether the C C<char> is an ASCII
954c1994
GS
334digit.
335
336=for apidoc Am|bool|isUPPER|char ch
337Returns a boolean indicating whether the C C<char> is an uppercase
338character.
339
340=for apidoc Am|bool|isLOWER|char ch
341Returns a boolean indicating whether the C C<char> is a lowercase
342character.
343
344=for apidoc Am|char|toUPPER|char ch
345Converts the specified character to uppercase.
346
347=for apidoc Am|char|toLOWER|char ch
348Converts the specified character to lowercase.
349
350=cut
351*/
352
bbce6d69
PP
353#define isALNUM(c) (isALPHA(c) || isDIGIT(c) || (c) == '_')
354#define isIDFIRST(c) (isALPHA(c) || (c) == '_')
355#define isALPHA(c) (isUPPER(c) || isLOWER(c))
356#define isSPACE(c) \
357 ((c) == ' ' || (c) == '\t' || (c) == '\n' || (c) =='\r' || (c) == '\f')
aaa51d5e
JF
358#define isPSXSPC(c) (isSPACE(c) || (c) == '\v')
359#define isBLANK(c) ((c) == ' ' || (c) == '\t')
bbce6d69 360#define isDIGIT(c) ((c) >= '0' && (c) <= '9')
9d116dd7
JH
361#ifdef EBCDIC
362 /* In EBCDIC we do not do locales: therefore() isupper() is fine. */
363# define isUPPER(c) isupper(c)
364# define isLOWER(c) islower(c)
b8c5462f
JH
365# define isALNUMC(c) isalnum(c)
366# define isASCII(c) isascii(c)
367# define isCNTRL(c) iscntrl(c)
368# define isGRAPH(c) isgraph(c)
9d116dd7 369# define isPRINT(c) isprint(c)
b8c5462f
JH
370# define isPUNCT(c) ispunct(c)
371# define isXDIGIT(c) isxdigit(c)
9d116dd7
JH
372# define toUPPER(c) toupper(c)
373# define toLOWER(c) tolower(c)
374#else
375# define isUPPER(c) ((c) >= 'A' && (c) <= 'Z')
376# define isLOWER(c) ((c) >= 'a' && (c) <= 'z')
b8c5462f
JH
377# define isALNUMC(c) (isALPHA(c) || isDIGIT(c))
378# define isASCII(c) ((c) <= 127)
7be5a6cf 379# define isCNTRL(c) ((c) < ' ' || (c) == 127)
b8c5462f 380# define isGRAPH(c) (isALNUM(c) || isPUNCT(c))
f79b3095 381# define isPRINT(c) (((c) > 32 && (c) < 127) || (c) == ' ')
b8c5462f 382# define isPUNCT(c) (((c) >= 33 && (c) <= 47) || ((c) >= 58 && (c) <= 64) || ((c) >= 91 && (c) <= 96) || ((c) >= 123 && (c) <= 126))
c302d089 383# define isXDIGIT(c) (isDIGIT(c) || ((c) >= 'a' && (c) <= 'f') || ((c) >= 'A' && (c) <= 'F'))
9d116dd7
JH
384# define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
385# define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
386#endif
bbce6d69
PP
387
388#ifdef USE_NEXT_CTYPE
389
390# define isALNUM_LC(c) \
37bd1396 391 (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
ff68c719
PP
392# define isIDFIRST_LC(c) \
393 (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
394# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
395# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
396# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
397# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
398# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
37bd1396 399# define isALNUMC_LC(c) NXIsAlNum((unsigned int)(c))
b8c5462f
JH
400# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
401# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
ff68c719 402# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 403# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
ff68c719
PP
404# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
405# define toLOWER_LC(c) NXToLower((unsigned int)(c))
bbce6d69
PP
406
407#else /* !USE_NEXT_CTYPE */
b8c5462f 408
bbce6d69
PP
409# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
410
b8c5462f 411# define isALNUM_LC(c) (isalnum((unsigned char)(c)) || (char)(c) == '_')
ff68c719 412# define isIDFIRST_LC(c) (isalpha((unsigned char)(c)) || (char)(c) == '_')
bbce6d69
PP
413# define isALPHA_LC(c) isalpha((unsigned char)(c))
414# define isSPACE_LC(c) isspace((unsigned char)(c))
415# define isDIGIT_LC(c) isdigit((unsigned char)(c))
416# define isUPPER_LC(c) isupper((unsigned char)(c))
417# define isLOWER_LC(c) islower((unsigned char)(c))
b8c5462f
JH
418# define isALNUMC_LC(c) isalnum((unsigned char)(c))
419# define isCNTRL_LC(c) iscntrl((unsigned char)(c))
420# define isGRAPH_LC(c) isgraph((unsigned char)(c))
bbce6d69 421# define isPRINT_LC(c) isprint((unsigned char)(c))
b8c5462f 422# define isPUNCT_LC(c) ispunct((unsigned char)(c))
bbce6d69
PP
423# define toUPPER_LC(c) toupper((unsigned char)(c))
424# define toLOWER_LC(c) tolower((unsigned char)(c))
425
426# else
427
3609ea0d 428# define isALNUM_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
bbce6d69
PP
429# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
430# define isALPHA_LC(c) (isascii(c) && isalpha(c))
431# define isSPACE_LC(c) (isascii(c) && isspace(c))
432# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
433# define isUPPER_LC(c) (isascii(c) && isupper(c))
434# define isLOWER_LC(c) (isascii(c) && islower(c))
b8c5462f
JH
435# define isALNUMC_LC(c) (isascii(c) && isalnum(c))
436# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
437# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
bbce6d69 438# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 439# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
bbce6d69
PP
440# define toUPPER_LC(c) toupper(c)
441# define toLOWER_LC(c) tolower(c)
442
443# endif
a0d0e21e 444#endif /* USE_NEXT_CTYPE */
55204971 445
aaa51d5e
JF
446#define isPSXSPC_LC(c) (isSPACE_LC(c) || (c) == '\v')
447#define isBLANK_LC(c) isBLANK(c) /* could be wrong */
448
a0ed51b3
LW
449#define isALNUM_uni(c) is_uni_alnum(c)
450#define isIDFIRST_uni(c) is_uni_idfirst(c)
451#define isALPHA_uni(c) is_uni_alpha(c)
452#define isSPACE_uni(c) is_uni_space(c)
453#define isDIGIT_uni(c) is_uni_digit(c)
454#define isUPPER_uni(c) is_uni_upper(c)
455#define isLOWER_uni(c) is_uni_lower(c)
b8c5462f
JH
456#define isALNUMC_uni(c) is_uni_alnumc(c)
457#define isASCII_uni(c) is_uni_ascii(c)
458#define isCNTRL_uni(c) is_uni_cntrl(c)
459#define isGRAPH_uni(c) is_uni_graph(c)
a0ed51b3 460#define isPRINT_uni(c) is_uni_print(c)
b8c5462f
JH
461#define isPUNCT_uni(c) is_uni_punct(c)
462#define isXDIGIT_uni(c) is_uni_xdigit(c)
a2a2844f
JH
463#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
464#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
465#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
b8d68ded 466#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
a0ed51b3 467
aaa51d5e
JF
468#define isPSXSPC_uni(c) (isSPACE_uni(c) ||(c) == '\f')
469#define isBLANK_uni(c) isBLANK(c) /* could be wrong */
470
9041c2e3
NIS
471#define isALNUM_LC_uvchr(c) (c < 256 ? isALNUM_LC(c) : is_uni_alnum_lc(c))
472#define isIDFIRST_LC_uvchr(c) (c < 256 ? isIDFIRST_LC(c) : is_uni_idfirst_lc(c))
473#define isALPHA_LC_uvchr(c) (c < 256 ? isALPHA_LC(c) : is_uni_alpha_lc(c))
474#define isSPACE_LC_uvchr(c) (c < 256 ? isSPACE_LC(c) : is_uni_space_lc(c))
475#define isDIGIT_LC_uvchr(c) (c < 256 ? isDIGIT_LC(c) : is_uni_digit_lc(c))
476#define isUPPER_LC_uvchr(c) (c < 256 ? isUPPER_LC(c) : is_uni_upper_lc(c))
477#define isLOWER_LC_uvchr(c) (c < 256 ? isLOWER_LC(c) : is_uni_lower_lc(c))
478#define isALNUMC_LC_uvchr(c) (c < 256 ? isALNUMC_LC(c) : is_uni_alnumc_lc(c))
479#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : is_uni_cntrl_lc(c))
480#define isGRAPH_LC_uvchr(c) (c < 256 ? isGRAPH_LC(c) : is_uni_graph_lc(c))
481#define isPRINT_LC_uvchr(c) (c < 256 ? isPRINT_LC(c) : is_uni_print_lc(c))
482#define isPUNCT_LC_uvchr(c) (c < 256 ? isPUNCT_LC(c) : is_uni_punct_lc(c))
a0ed51b3 483
aaa51d5e
JF
484#define isPSXSPC_LC_uni(c) (isSPACE_LC_uni(c) ||(c) == '\f')
485#define isBLANK_LC_uni(c) isBLANK(c) /* could be wrong */
486
a0ed51b3 487#define isALNUM_utf8(p) is_utf8_alnum(p)
82686b01
JH
488/* The ID_Start of Unicode is quite limiting: it assumes a L-class
489 * character (meaning that you cannot have, say, a CJK character).
490 * Instead, let's allow ID_Continue but not digits. */
491#define isIDFIRST_utf8(p) (is_utf8_idcont(p) && !is_utf8_digit(p))
a0ed51b3
LW
492#define isALPHA_utf8(p) is_utf8_alpha(p)
493#define isSPACE_utf8(p) is_utf8_space(p)
494#define isDIGIT_utf8(p) is_utf8_digit(p)
495#define isUPPER_utf8(p) is_utf8_upper(p)
496#define isLOWER_utf8(p) is_utf8_lower(p)
b8c5462f
JH
497#define isALNUMC_utf8(p) is_utf8_alnumc(p)
498#define isASCII_utf8(p) is_utf8_ascii(p)
499#define isCNTRL_utf8(p) is_utf8_cntrl(p)
500#define isGRAPH_utf8(p) is_utf8_graph(p)
a0ed51b3 501#define isPRINT_utf8(p) is_utf8_print(p)
b8c5462f
JH
502#define isPUNCT_utf8(p) is_utf8_punct(p)
503#define isXDIGIT_utf8(p) is_utf8_xdigit(p)
a2a2844f
JH
504#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
505#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
506#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
a0ed51b3 507
aaa51d5e
JF
508#define isPSXSPC_utf8(c) (isSPACE_utf8(c) ||(c) == '\f')
509#define isBLANK_utf8(c) isBLANK(c) /* could be wrong */
510
9041c2e3
NIS
511#define isALNUM_LC_utf8(p) isALNUM_LC_uvchr(utf8_to_uvchr(p, 0))
512#define isIDFIRST_LC_utf8(p) isIDFIRST_LC_uvchr(utf8_to_uvchr(p, 0))
513#define isALPHA_LC_utf8(p) isALPHA_LC_uvchr(utf8_to_uvchr(p, 0))
514#define isSPACE_LC_utf8(p) isSPACE_LC_uvchr(utf8_to_uvchr(p, 0))
515#define isDIGIT_LC_utf8(p) isDIGIT_LC_uvchr(utf8_to_uvchr(p, 0))
516#define isUPPER_LC_utf8(p) isUPPER_LC_uvchr(utf8_to_uvchr(p, 0))
517#define isLOWER_LC_utf8(p) isLOWER_LC_uvchr(utf8_to_uvchr(p, 0))
518#define isALNUMC_LC_utf8(p) isALNUMC_LC_uvchr(utf8_to_uvchr(p, 0))
519#define isCNTRL_LC_utf8(p) isCNTRL_LC_uvchr(utf8_to_uvchr(p, 0))
520#define isGRAPH_LC_utf8(p) isGRAPH_LC_uvchr(utf8_to_uvchr(p, 0))
521#define isPRINT_LC_utf8(p) isPRINT_LC_uvchr(utf8_to_uvchr(p, 0))
522#define isPUNCT_LC_utf8(p) isPUNCT_LC_uvchr(utf8_to_uvchr(p, 0))
a0ed51b3 523
aaa51d5e
JF
524#define isPSXSPC_LC_utf8(c) (isSPACE_LC_utf8(c) ||(c) == '\f')
525#define isBLANK_LC_utf8(c) isBLANK(c) /* could be wrong */
526
9d116dd7 527#ifdef EBCDIC
20b634c2
JH
528# ifdef PERL_IMPLICIT_CONTEXT
529# define toCTRL(c) Perl_ebcdic_control(aTHX_ c)
530# else
531# define toCTRL Perl_ebcdic_control
532# endif
9d116dd7
JH
533#else
534 /* This conversion works both ways, strangely enough. */
535# define toCTRL(c) (toUPPER(c) ^ 64)
536#endif
bbce6d69 537
dea28490
JJ
538/* Line numbers are unsigned, 32 bits. */
539typedef U32 line_t;
dea28490 540#define NOLINE ((line_t) 4294967295UL)
378cc40b 541
8c52afec 542
8e84507e 543/*
ccfc67b7
JH
544=head1 SV Manipulation Functions
545
954c1994
GS
546=for apidoc Am|SV*|NEWSV|int id|STRLEN len
547Creates a new SV. A non-zero C<len> parameter indicates the number of
548bytes of preallocated string space the SV should have. An extra byte for a
549tailing NUL is also reserved. (SvPOK is not set for the SV even if string
8e84507e 550space is allocated.) The reference count for the new SV is set to 1.
954c1994
GS
551C<id> is an integer id between 0 and 1299 (used to identify leaks).
552
ccfc67b7
JH
553=head1 Memory Management
554
a02a5408 555=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
556The XSUB-writer's interface to the C C<malloc> function.
557
c5008215
JC
558In 5.9.3, Newx() and friends replace the older New() API, and drops
559the first parameter, I<x>, a debug aid which allowed callers to identify
560themselves. This aid has been superceded by a new build option,
561PERL_MEM_LOG (see L<perlhack/PERL_MEM_LOG>). The older API is still
562there for use in XS modules supporting older perls.
563
a02a5408 564=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 565The XSUB-writer's interface to the C C<malloc> function, with
c5008215 566cast. See also C<Newx>.
954c1994 567
a02a5408 568=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 569The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 570memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 571
954c1994
GS
572=for apidoc Am|void|Renew|void* ptr|int nitems|type
573The XSUB-writer's interface to the C C<realloc> function.
574
575=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
576The XSUB-writer's interface to the C C<realloc> function, with
577cast.
578
49b8b560 579=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
580The XSUB-writer's interface to the C C<free> function.
581
582=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
583The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
584source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
585the type. Can do overlapping moves. See also C<Copy>.
586
e90e2364
NC
587=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
588Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
589optimise.
590
954c1994
GS
591=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
592The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
593source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
594the type. May fail on overlapping copies. See also C<Move>.
595
e90e2364
NC
596=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
597
598Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
599optimise.
600
954c1994
GS
601=for apidoc Am|void|Zero|void* dest|int nitems|type
602
603The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
604destination, C<nitems> is the number of items, and C<type> is the type.
605
e90e2364
NC
606=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
607
608Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
609optimise.
610
954c1994 611=for apidoc Am|void|StructCopy|type src|type dest|type
4375e838 612This is an architecture-independent macro to copy one structure to another.
954c1994 613
9965345d
JH
614=for apidoc Am|void|Poison|void* dest|int nitems|type
615
616Fill up memory with a pattern (byte 0xAB over and over again) that
617hopefully catches attempts to access uninitialized memory.
618
619=cut */
954c1994 620
ff06c60c
IZ
621#define NEWSV(x,len) newSV(len)
622
27d5b266 623#ifdef PERL_MALLOC_WRAP
fabdb6c0 624#define MEM_WRAP_CHECK(n,t) MEM_WRAP_CHECK_1(n,t,PL_memory_wrap)
27d5b266 625#define MEM_WRAP_CHECK_1(n,t,a) \
ed87b6ed 626 (void)((sizeof(t)>1?(n):1)>((MEM_SIZE)~0)/sizeof(t)?(Perl_croak_nocontext(a),0):0)
8b44ba4c 627#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 628
6fff79ce 629#define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > (MEM_SIZE)~0 - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (Perl_croak_nocontext(PL_memory_wrap),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1936d2a7 630
27d5b266
JH
631#else
632
410319be
NC
633#define MEM_WRAP_CHECK(n,t)
634#define MEM_WRAP_CHECK_1(n,t,a)
635#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
636#define MEM_WRAP_CHECK_(n,t)
637
6fff79ce 638#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 639
1936d2a7 640#endif
8b44ba4c 641
fe4f188c 642#ifdef PERL_MEM_LOG
46c6c7e2 643/*
9f653bb5 644 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
645 * go through functions, which are handy for debugging breakpoints, but
646 * which more importantly get the immediate calling environment (file and
e352bcff
JH
647 * line number, and C function name if available) passed in. This info can
648 * then be used for logging the calls, for which one gets a sample
649 * implementation if PERL_MEM_LOG_STDERR is defined.
3609ea0d 650 *
46c6c7e2
JH
651 * Known problems:
652 * - all memory allocs do not get logged, only those
653 * that go through Newx() and derivatives (while all
654 * Safefrees do get logged)
655 * - __FILE__ and __LINE__ do not work everywhere
656 * - __func__ or __FUNCTION__ even less so
657 * - I think more goes on after the perlio frees but
658 * the thing is that STDERR gets closed (as do all
659 * the file descriptors)
660 * - no deeper calling stack than the caller of the Newx()
661 * or the kind, but do I look like a C reflection/introspection
662 * utility to you?
663 * - the function prototypes for the logging functions
664 * probably should maybe be somewhere else than handy.h
665 * - one could consider inlining (macrofying) the logging
666 * for speed, but I am too lazy
667 * - one could imagine recording the allocations in a hash,
668 * (keyed by the allocation address?), and maintain that
669 * through reallocs and frees, but how to do that without
670 * any News() happening...?
671 */
672
673Malloc_t Perl_mem_log_alloc(const UV n, const UV typesize, const char *typename, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname);
674
675Malloc_t Perl_mem_log_realloc(const UV n, const UV typesize, const char *typename, Malloc_t oldalloc, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname);
676
677Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
678
fe4f188c
JH
679#endif
680
681#ifdef PERL_MEM_LOG
46c6c7e2
JH
682#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
683#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
684#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
685#endif
686
687#ifndef MEM_LOG_ALLOC
688#define MEM_LOG_ALLOC(n,t,a) (a)
689#endif
690#ifndef MEM_LOG_REALLOC
691#define MEM_LOG_REALLOC(n,t,v,a) (a)
692#endif
693#ifndef MEM_LOG_FREE
694#define MEM_LOG_FREE(a) (a)
695#endif
696
697#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) MEM_LOG_ALLOC(n,t,(t*)safemalloc((MEM_SIZE)((n)*sizeof(t))))))
698#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) MEM_LOG_ALLOC(n,t,(c*)safemalloc((MEM_SIZE)((n)*sizeof(t))))))
699#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) MEM_LOG_ALLOC(n,t,(t*)safemalloc((MEM_SIZE)((n)*sizeof(t)))))), \
ff68c719 700 memzero((char*)(v), (n)*sizeof(t))
a6f6820f
NC
701
702#ifndef PERL_CORE
a02a5408
JC
703/* pre 5.9.x compatibility */
704#define New(x,v,n,t) Newx(v,n,t)
705#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 706#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 707#endif
a02a5408 708
ff68c719 709#define Renew(v,n,t) \
fe4f188c 710 (v = (MEM_WRAP_CHECK_(n,t) MEM_LOG_REALLOC(n,t,v,(t*)saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 711#define Renewc(v,n,t,c) \
fe4f188c 712 (v = (MEM_WRAP_CHECK_(n,t) MEM_LOG_REALLOC(n,t,v,(c*)saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
713
714#ifdef PERL_POISON
715#define Safefree(d) \
06c0cc96 716 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 717#else
fe4f188c 718#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 719#endif
55497cff 720
8b44ba4c
NC
721#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
722#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
723#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 724
8b44ba4c
NC
725#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
726#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 727#ifdef HAS_MEMSET
8b44ba4c 728#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 729#else
8b44ba4c
NC
730/* Using bzero(), which returns void. */
731#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
732#endif
733
8b44ba4c 734#define Poison(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), 0xAB, (n) * sizeof(t)))
27d5b266 735
2304df62 736#ifdef USE_STRUCT_COPY
ff68c719 737#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
738#else
739#define StructCopy(s,d,t) Copy(s,d,1,t)
740#endif
2cc61e15 741
622913ab
IZ
742#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
743
2cc61e15
DD
744#ifdef NEED_VA_COPY
745# ifdef va_copy
746# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 747# else
a1866d1b
JH
748# if defined(__va_copy)
749# define Perl_va_copy(s, d) __va_copy(d, s)
750# else
751# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
752# endif
2cc61e15
DD
753# endif
754#endif
755
472d47bc
SB
756/* convenience debug macros */
757#ifdef USE_ITHREADS
758#define pTHX_FORMAT "Perl interpreter: 0x%p"
759#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
760#define pTHX_VALUE_ (void *)my_perl,
761#define pTHX_VALUE (void *)my_perl
762#define pTHX__VALUE_ ,(void *)my_perl,
763#define pTHX__VALUE ,(void *)my_perl
472d47bc 764#else
3609ea0d 765#define pTHX_FORMAT
472d47bc 766#define pTHX__FORMAT
3609ea0d 767#define pTHX_VALUE_
472d47bc 768#define pTHX_VALUE
3609ea0d 769#define pTHX__VALUE_
472d47bc
SB
770#define pTHX__VALUE
771#endif /* USE_ITHREADS */
3609ea0d 772