Commit | Line | Data |
---|---|---|
a0d0e21e | 1 | /* handy.h |
a687059c | 2 | * |
1129b882 | 3 | * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000, |
da5d8dbb | 4 | * 2001, 2002, 2004, 2005, 2006, 2007, 2008, 2012 by Larry Wall and others |
a687059c | 5 | * |
6e21c824 LW |
6 | * You may distribute under the terms of either the GNU General Public |
7 | * License or the Artistic License, as specified in the README file. | |
8d063cd8 | 8 | * |
8d063cd8 LW |
9 | */ |
10 | ||
4650c663 KW |
11 | /* IMPORTANT NOTE: Everything whose name begins with an underscore is for |
12 | * internal core Perl use only. */ | |
13 | ||
6a5bc5ac KW |
14 | #ifndef PERL_HANDY_H_ /* Guard against nested #inclusion */ |
15 | #define PERL_HANDY_H_ | |
9d745869 | 16 | |
24792b8d NC |
17 | #ifndef PERL_CORE |
18 | # define Null(type) ((type)NULL) | |
954c1994 GS |
19 | |
20 | /* | |
3f620621 | 21 | =for apidoc_section $string |
78342678 | 22 | =for apidoc AmnU||Nullch |
72d33970 FC |
23 | Null character pointer. (No longer available when C<PERL_CORE> is |
24 | defined.) | |
2307c6d0 | 25 | |
3f620621 | 26 | =for apidoc_section $SV |
78342678 | 27 | =for apidoc AmnU||Nullsv |
72d33970 | 28 | Null SV pointer. (No longer available when C<PERL_CORE> is defined.) |
954c1994 GS |
29 | |
30 | =cut | |
51b56f5c KW |
31 | |
32 | Below are signatures of functions from config.h which can't easily be gleaned | |
33 | from it, and are very unlikely to change | |
34 | ||
3f620621 | 35 | =for apidoc_section $signals |
51b56f5c KW |
36 | =for apidoc Am|int|Sigsetjmp|jmp_buf env|int savesigs |
37 | =for apidoc Am|void|Siglongjmp|jmp_buf env|int val | |
38 | ||
3f620621 | 39 | =for apidoc_section $filesystem |
51b56f5c KW |
40 | =for apidoc Am|void *|FILE_ptr|FILE * f |
41 | =for apidoc Am|Size_t|FILE_cnt|FILE * f | |
42 | =for apidoc Am|void *|FILE_base|FILE * f | |
b290de04 | 43 | =for apidoc Am|Size_t|FILE_bufsiz|FILE *f |
51b56f5c | 44 | |
3f620621 | 45 | =for apidoc_section $string |
b290de04 KW |
46 | =for apidoc Amu|token|CAT2|token x|token y |
47 | =for apidoc Amu|string|STRINGIFY|token x | |
51b56f5c | 48 | |
3f620621 | 49 | =for apidoc_section $numeric |
b290de04 | 50 | =for apidoc Am|double|Drand01 |
51b56f5c KW |
51 | =for apidoc Am|void|seedDrand01|Rand_seed_t x |
52 | =for apidoc Am|char *|Gconvert|double x|Size_t n|bool t|char * b | |
53 | ||
54 | =cut | |
954c1994 GS |
55 | */ |
56 | ||
24792b8d NC |
57 | # define Nullch Null(char*) |
58 | # define Nullfp Null(PerlIO*) | |
59 | # define Nullsv Null(SV*) | |
60 | #endif | |
8d063cd8 | 61 | |
641d3f0b | 62 | #ifdef TRUE |
63 | #undef TRUE | |
64 | #endif | |
65 | #ifdef FALSE | |
66 | #undef FALSE | |
67 | #endif | |
68 | #define TRUE (1) | |
69 | #define FALSE (0) | |
70 | ||
df87895c | 71 | /* |
3f620621 | 72 | =for apidoc_section $SV |
df87895c KW |
73 | =for apidoc Am|void *|MUTABLE_PTR|void * p |
74 | =for apidoc_item |AV *|MUTABLE_AV|AV * p | |
75 | =for apidoc_item |CV *|MUTABLE_CV|CV * p | |
76 | =for apidoc_item |GV *|MUTABLE_GV|GV * p | |
77 | =for apidoc_item |HV *|MUTABLE_HV|HV * p | |
78 | =for apidoc_item |IO *|MUTABLE_IO|IO * p | |
79 | =for apidoc_item |SV *|MUTABLE_SV|SV * p | |
80 | ||
81 | The C<MUTABLE_I<*>>() macros cast pointers to the types shown, in such a way | |
82 | (compiler permitting) that casting away const-ness will give a warning; | |
83 | e.g.: | |
84 | ||
85 | const SV *sv = ...; | |
86 | AV *av1 = (AV*)sv; <== BAD: the const has been silently | |
87 | cast away | |
88 | AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn | |
89 | ||
54ba56f3 KW |
90 | C<MUTABLE_PTR> is the base macro used to derive new casts. The other |
91 | already-built-in ones return pointers to what their names indicate. | |
df87895c KW |
92 | |
93 | =cut | |
cf3f0ffb DM |
94 | */ |
95 | ||
041c1a23 | 96 | #if defined(PERL_USE_GCC_BRACE_GROUPS) |
6c2255e0 | 97 | # define MUTABLE_PTR(p) ({ void *p_ = (p); p_; }) |
b1bc3f34 NC |
98 | #else |
99 | # define MUTABLE_PTR(p) ((void *) (p)) | |
100 | #endif | |
101 | ||
a062e10d | 102 | #define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p)) |
ea726b52 | 103 | #define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p)) |
159b6efe | 104 | #define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p)) |
dbebbdb4 | 105 | #define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p)) |
a45c7426 | 106 | #define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p)) |
b1bc3f34 | 107 | #define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p)) |
27d4fb96 | 108 | |
f789f6a4 | 109 | #if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR) |
bd31be4b NC |
110 | # include <stdbool.h> |
111 | # ifndef HAS_BOOL | |
112 | # define HAS_BOOL 1 | |
113 | # endif | |
114 | #endif | |
115 | ||
8e84507e | 116 | /* bool is built-in for g++-2.6.3 and later, which might be used |
c1d22f6b GS |
117 | for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't |
118 | be sure _G_config.h will be included before this file. _G_config.h | |
8e84507e | 119 | also defines _G_HAVE_BOOL for both gcc and g++, but only g++ |
c1d22f6b GS |
120 | actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us. |
121 | g++ can be identified by __GNUG__. | |
122 | Andy Dougherty February 2000 | |
5d94fbed | 123 | */ |
3609ea0d | 124 | #ifdef __GNUG__ /* GNU g++ has bool built-in */ |
f789f6a4 | 125 | # ifndef PERL_BOOL_AS_CHAR |
5d94fbed | 126 | # ifndef HAS_BOOL |
c1d22f6b | 127 | # define HAS_BOOL 1 |
5d94fbed | 128 | # endif |
f789f6a4 | 129 | # endif |
5d94fbed AD |
130 | #endif |
131 | ||
132 | #ifndef HAS_BOOL | |
f789f6a4 FC |
133 | # ifdef bool |
134 | # undef bool | |
135 | # endif | |
70d5cb32 | 136 | # define bool char |
c1d22f6b | 137 | # define HAS_BOOL 1 |
a687059c | 138 | #endif |
0d3e774c | 139 | |
25ba28ce | 140 | /* |
3f620621 | 141 | =for apidoc_section $casting |
25ba28ce KW |
142 | =for apidoc Am|bool|cBOOL|bool expr |
143 | ||
144 | Cast-to-bool. A simple S<C<(bool) I<expr>>> cast may not do the right thing: | |
145 | if C<bool> is defined as C<char>, for example, then the cast from C<int> is | |
146 | implementation-defined. | |
147 | ||
148 | C<(bool)!!(cbool)> in a ternary triggers a bug in xlc on AIX | |
149 | ||
150 | =cut | |
151 | */ | |
18f5643b | 152 | #define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0) |
f2338a2e | 153 | |
46c6c7e2 | 154 | /* Try to figure out __func__ or __FUNCTION__ equivalent, if any. |
e352bcff JH |
155 | * XXX Should really be a Configure probe, with HAS__FUNCTION__ |
156 | * and FUNCTION__ as results. | |
157 | * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */ | |
46c6c7e2 JH |
158 | #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */ |
159 | # define FUNCTION__ __func__ | |
7adf2470 | 160 | #elif (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */ |
07798b17 | 161 | # define FUNCTION__ "" |
46c6c7e2 | 162 | #else |
07798b17 | 163 | # define FUNCTION__ __FUNCTION__ /* Common extension. */ |
46c6c7e2 JH |
164 | #endif |
165 | ||
27d4fb96 | 166 | /* XXX A note on the perl source internal type system. The |
167 | original intent was that I32 be *exactly* 32 bits. | |
168 | ||
169 | Currently, we only guarantee that I32 is *at least* 32 bits. | |
170 | Specifically, if int is 64 bits, then so is I32. (This is the case | |
171 | for the Cray.) This has the advantage of meshing nicely with | |
172 | standard library calls (where we pass an I32 and the library is | |
173 | expecting an int), but the disadvantage that an I32 is not 32 bits. | |
174 | Andy Dougherty August 1996 | |
24fef2a7 | 175 | |
dc45a647 MB |
176 | There is no guarantee that there is *any* integral type with |
177 | exactly 32 bits. It is perfectly legal for a system to have | |
178 | sizeof(short) == sizeof(int) == sizeof(long) == 8. | |
693762b4 | 179 | |
dc45a647 MB |
180 | Similarly, there is no guarantee that I16 and U16 have exactly 16 |
181 | bits. | |
693762b4 | 182 | |
8e84507e NIS |
183 | For dealing with issues that may arise from various 32/64-bit |
184 | systems, we will ask Configure to check out | |
8175356b | 185 | |
3609ea0d JH |
186 | SHORTSIZE == sizeof(short) |
187 | INTSIZE == sizeof(int) | |
188 | LONGSIZE == sizeof(long) | |
dc45a647 | 189 | LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG) |
3609ea0d | 190 | PTRSIZE == sizeof(void *) |
dc45a647 MB |
191 | DOUBLESIZE == sizeof(double) |
192 | LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE). | |
8175356b | 193 | |
27d4fb96 | 194 | */ |
195 | ||
69512466 JH |
196 | #ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */ |
197 | # include <inttypes.h> | |
dd0eed91 JH |
198 | # ifdef INT32_MIN_BROKEN |
199 | # undef INT32_MIN | |
200 | # define INT32_MIN (-2147483647-1) | |
201 | # endif | |
202 | # ifdef INT64_MIN_BROKEN | |
203 | # undef INT64_MIN | |
204 | # define INT64_MIN (-9223372036854775807LL-1) | |
205 | # endif | |
69512466 JH |
206 | #endif |
207 | ||
8175356b JH |
208 | typedef I8TYPE I8; |
209 | typedef U8TYPE U8; | |
210 | typedef I16TYPE I16; | |
211 | typedef U16TYPE U16; | |
212 | typedef I32TYPE I32; | |
213 | typedef U32TYPE U32; | |
16d89be8 | 214 | |
74b807c7 | 215 | #ifdef QUADKIND |
8175356b JH |
216 | typedef I64TYPE I64; |
217 | typedef U64TYPE U64; | |
16d89be8 | 218 | #endif |
8175356b | 219 | |
5ff3f7a4 GS |
220 | /* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type. |
221 | Please search CHAR_MAX in perl.h for further details. */ | |
26e4a9a9 KW |
222 | #ifdef UINT8_MAX |
223 | # define U8_MAX UINT8_MAX | |
0e983133 | 224 | #else |
26e4a9a9 | 225 | # define U8_MAX PERL_UCHAR_MAX |
0e983133 | 226 | #endif |
26e4a9a9 KW |
227 | #ifdef UINT8_MIN |
228 | # define U8_MIN UINT8_MIN | |
5ff3f7a4 | 229 | #else |
26e4a9a9 KW |
230 | # define U8_MIN PERL_UCHAR_MIN |
231 | #endif | |
5ff3f7a4 | 232 | |
26e4a9a9 KW |
233 | #ifdef INT16_MAX |
234 | # define I16_MAX INT16_MAX | |
79072805 | 235 | #else |
26e4a9a9 KW |
236 | # define I16_MAX PERL_SHORT_MAX |
237 | #endif | |
238 | #ifdef INT16_MIN | |
239 | # define I16_MIN INT16_MIN | |
240 | #else | |
241 | # define I16_MIN PERL_SHORT_MIN | |
242 | #endif | |
243 | #ifdef UINT16_MAX | |
244 | # define U16_MAX UINT16_MAX | |
245 | #else | |
246 | # define U16_MAX PERL_USHORT_MAX | |
247 | #endif | |
248 | #ifdef UINT16_MIN | |
249 | # define U16_MIN UINT16_MIN | |
250 | #else | |
251 | # define U16_MIN PERL_USHORT_MIN | |
79072805 LW |
252 | #endif |
253 | ||
26e4a9a9 KW |
254 | #ifdef INT32_MAX |
255 | # define I32_MAX INT32_MAX | |
256 | #elif LONGSIZE > 4 | |
257 | # define I32_MAX PERL_INT_MAX | |
258 | #else | |
259 | # define I32_MAX PERL_LONG_MAX | |
260 | #endif | |
261 | #ifdef INT32_MIN | |
262 | # define I32_MIN INT32_MIN | |
263 | #elif LONGSIZE > 4 | |
264 | # define I32_MIN PERL_INT_MIN | |
265 | #else | |
266 | # define I32_MIN PERL_LONG_MIN | |
267 | #endif | |
268 | #ifdef UINT32_MAX | |
269 | # ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */ | |
270 | # define U32_MAX UINT_MAX | |
271 | # else | |
272 | # define U32_MAX 4294967295U | |
273 | # endif | |
274 | #elif LONGSIZE > 4 | |
275 | # define U32_MAX PERL_UINT_MAX | |
276 | #else | |
277 | # define U32_MAX PERL_ULONG_MAX | |
278 | #endif | |
279 | #ifdef UINT32_MIN | |
280 | # define U32_MIN UINT32_MIN | |
281 | #elif LONGSIZE > 4 | |
282 | # define U32_MIN PERL_UINT_MIN | |
283 | #else | |
284 | # define U32_MIN PERL_ULONG_MIN | |
5ff3f7a4 GS |
285 | #endif |
286 | ||
23332c7d KW |
287 | /* |
288 | =for apidoc_section $integer | |
289 | =for apidoc Ay|| PERL_INT_FAST8_T | |
290 | =for apidoc_item PERL_INT_FAST16_T | |
291 | =for apidoc_item PERL_UINT_FAST8_T | |
292 | =for apidoc_item PERL_UINT_FAST16_T | |
293 | ||
294 | These are equivalent to the correspondingly-named C99 typedefs on platforms | |
295 | that have those; they evaluate to C<int> and C<unsigned int> on platforms that | |
296 | don't, so that you can portably take advantage of this C99 feature. | |
297 | ||
298 | =cut | |
299 | */ | |
247cee9f KW |
300 | # ifdef I_STDINT |
301 | typedef int_fast8_t PERL_INT_FAST8_T; | |
302 | typedef uint_fast8_t PERL_UINT_FAST8_T; | |
303 | typedef int_fast16_t PERL_INT_FAST16_T; | |
304 | typedef uint_fast16_t PERL_UINT_FAST16_T; | |
305 | # else | |
306 | typedef int PERL_INT_FAST8_T; | |
307 | typedef unsigned int PERL_UINT_FAST8_T; | |
308 | typedef int PERL_INT_FAST16_T; | |
309 | typedef unsigned int PERL_UINT_FAST16_T; | |
310 | # endif | |
247cee9f | 311 | |
464decb6 | 312 | /* log(2) (i.e., log base 10 of 2) is pretty close to 0.30103, just in case |
ab350cbd KW |
313 | * anyone is grepping for it. So BIT_DIGITS gives the number of decimal digits |
314 | * required to represent any possible unsigned number containing N bits. | |
315 | * TYPE_DIGITS gives the number of decimal digits required to represent any | |
316 | * possible unsigned number of type T. */ | |
464decb6 | 317 | #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log10(2) =~ 146/485 */ |
fc36a67e | 318 | #define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8) |
319 | #define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */ | |
320 | ||
88794300 | 321 | /* Unused by core; should be deprecated */ |
ff68c719 | 322 | #define Ctl(ch) ((ch) & 037) |
8d063cd8 | 323 | |
98fce2a4 KW |
324 | #if defined(PERL_CORE) || defined(PERL_EXT) |
325 | # ifndef MIN | |
326 | # define MIN(a,b) ((a) < (b) ? (a) : (b)) | |
327 | # endif | |
328 | # ifndef MAX | |
329 | # define MAX(a,b) ((a) > (b) ? (a) : (b)) | |
330 | # endif | |
331 | #endif | |
332 | ||
84ff4fa9 KW |
333 | /* Returns a boolean as to whether the input unsigned number is a power of 2 |
334 | * (2**0, 2**1, etc). In other words if it has just a single bit set. | |
335 | * If not, subtracting 1 would leave the uppermost bit set, so the & would | |
336 | * yield non-zero */ | |
337 | #if defined(PERL_CORE) || defined(PERL_EXT) | |
011b1419 | 338 | # define isPOWER_OF_2(n) ((n) && ((n) & ((n)-1)) == 0) |
84ff4fa9 KW |
339 | #endif |
340 | ||
d223e1ea | 341 | /* Returns a mask with the lowest n bits set */ |
fae1e72b | 342 | #define nBIT_MASK(n) ((UINTMAX_C(1) << (n)) - 1) |
d223e1ea | 343 | |
1381ccb1 KW |
344 | /* The largest unsigned number that will fit into n bits */ |
345 | #define nBIT_UMAX(n) nBIT_MASK(n) | |
346 | ||
8d9433eb | 347 | /* |
3f620621 | 348 | =for apidoc_section $directives |
d23778e6 | 349 | =for apidoc Am||__ASSERT_|bool expr |
8d9433eb KW |
350 | |
351 | This is a helper macro to avoid preprocessor issues, replaced by nothing | |
352 | unless under DEBUGGING, where it expands to an assert of its argument, | |
353 | followed by a comma (hence the comma operator). If we just used a straight | |
354 | assert(), we would get a comma with nothing before it when not DEBUGGING. | |
355 | ||
356 | =cut | |
357 | ||
358 | We also use empty definition under Coverity since the __ASSERT__ | |
359 | checks often check for things that Really Cannot Happen, and Coverity | |
360 | detects that and gets all excited. */ | |
3e94db23 | 361 | |
e7ae132e KW |
362 | #if defined(DEBUGGING) && !defined(__COVERITY__) \ |
363 | && ! defined(PERL_SMALL_MACRO_BUFFER) | |
0f092d08 KW |
364 | # define __ASSERT_(statement) assert(statement), |
365 | #else | |
366 | # define __ASSERT_(statement) | |
367 | #endif | |
368 | ||
3fe05580 | 369 | /* |
3f620621 | 370 | =for apidoc_section $SV |
3fe05580 | 371 | |
3bb9fd01 | 372 | =for apidoc Ama|SV*|newSVpvs|"literal string" |
1568d13a | 373 | Like C<newSVpvn>, but takes a literal string instead of a |
30a15352 | 374 | string/length pair. |
3fe05580 | 375 | |
3bb9fd01 | 376 | =for apidoc Ama|SV*|newSVpvs_flags|"literal string"|U32 flags |
1568d13a | 377 | Like C<newSVpvn_flags>, but takes a literal string instead of |
30a15352 | 378 | a string/length pair. |
84bafc02 | 379 | |
3bb9fd01 | 380 | =for apidoc Ama|SV*|newSVpvs_share|"literal string" |
1568d13a | 381 | Like C<newSVpvn_share>, but takes a literal string instead of |
30a15352 | 382 | a string/length pair and omits the hash parameter. |
3fe05580 | 383 | |
3bb9fd01 | 384 | =for apidoc Am|void|sv_catpvs_flags|SV* sv|"literal string"|I32 flags |
1568d13a | 385 | Like C<sv_catpvn_flags>, but takes a literal string instead |
30a15352 | 386 | of a string/length pair. |
9dcc53ea | 387 | |
3bb9fd01 | 388 | =for apidoc Am|void|sv_catpvs_nomg|SV* sv|"literal string" |
1568d13a | 389 | Like C<sv_catpvn_nomg>, but takes a literal string instead of |
0c395ea5 | 390 | a string/length pair. |
9dcc53ea | 391 | |
3bb9fd01 | 392 | =for apidoc Am|void|sv_catpvs|SV* sv|"literal string" |
1568d13a | 393 | Like C<sv_catpvn>, but takes a literal string instead of a |
0c395ea5 | 394 | string/length pair. |
3fe05580 | 395 | |
3bb9fd01 | 396 | =for apidoc Am|void|sv_catpvs_mg|SV* sv|"literal string" |
1568d13a | 397 | Like C<sv_catpvn_mg>, but takes a literal string instead of a |
9dcc53ea Z |
398 | string/length pair. |
399 | ||
3bb9fd01 | 400 | =for apidoc Am|void|sv_setpvs|SV* sv|"literal string" |
1568d13a | 401 | Like C<sv_setpvn>, but takes a literal string instead of a |
0c395ea5 | 402 | string/length pair. |
3fe05580 | 403 | |
3bb9fd01 | 404 | =for apidoc Am|void|sv_setpvs_mg|SV* sv|"literal string" |
1568d13a | 405 | Like C<sv_setpvn_mg>, but takes a literal string instead of a |
9dcc53ea Z |
406 | string/length pair. |
407 | ||
3bb9fd01 | 408 | =for apidoc Am|SV *|sv_setref_pvs|SV *const rv|const char *const classname|"literal string" |
1568d13a | 409 | Like C<sv_setref_pvn>, but takes a literal string instead of |
0c395ea5 | 410 | a string/length pair. |
9dcc53ea | 411 | |
3f620621 | 412 | =for apidoc_section $string |
3fe05580 | 413 | |
3bb9fd01 | 414 | =for apidoc Ama|char*|savepvs|"literal string" |
1568d13a | 415 | Like C<savepvn>, but takes a literal string instead of a |
30a15352 | 416 | string/length pair. |
3fe05580 | 417 | |
3bb9fd01 | 418 | =for apidoc Ama|char*|savesharedpvs|"literal string" |
9dcc53ea Z |
419 | A version of C<savepvs()> which allocates the duplicate string in memory |
420 | which is shared between threads. | |
421 | ||
3f620621 | 422 | =for apidoc_section $GV |
3fe05580 | 423 | |
3bb9fd01 | 424 | =for apidoc Am|HV*|gv_stashpvs|"name"|I32 create |
1568d13a | 425 | Like C<gv_stashpvn>, but takes a literal string instead of a |
0c395ea5 | 426 | string/length pair. |
3fe05580 | 427 | |
3f620621 | 428 | =for apidoc_section $HV |
3fe05580 | 429 | |
3bb9fd01 | 430 | =for apidoc Am|SV**|hv_fetchs|HV* tb|"key"|I32 lval |
1568d13a | 431 | Like C<hv_fetch>, but takes a literal string instead of a |
0c395ea5 | 432 | string/length pair. |
3fe05580 | 433 | |
3bb9fd01 | 434 | =for apidoc Am|SV**|hv_stores|HV* tb|"key"|SV* val |
1568d13a | 435 | Like C<hv_store>, but takes a literal string instead of a |
0c395ea5 | 436 | string/length pair |
3fe05580 MHM |
437 | and omits the hash parameter. |
438 | ||
3f620621 | 439 | =for apidoc_section $lexer |
510966aa | 440 | |
3bb9fd01 | 441 | =for apidoc Amx|void|lex_stuff_pvs|"pv"|U32 flags |
510966aa | 442 | |
1568d13a | 443 | Like L</lex_stuff_pvn>, but takes a literal string instead of |
0c395ea5 | 444 | a string/length pair. |
510966aa | 445 | |
3fe05580 MHM |
446 | =cut |
447 | */ | |
448 | ||
a34e53fc | 449 | /* |
3f620621 | 450 | =for apidoc_section $string |
2efa8cc7 | 451 | |
a34e53fc KW |
452 | =for apidoc Amu|pair|STR_WITH_LEN|"literal string" |
453 | ||
454 | Returns two comma separated tokens of the input literal string, and its length. | |
455 | This is convenience macro which helps out in some API calls. | |
456 | Note that it can't be used as an argument to macros or functions that under | |
457 | some configurations might be macros, which means that it requires the full | |
458 | Perl_xxx(aTHX_ ...) form for any API calls where it's used. | |
459 | ||
460 | =cut | |
461 | */ | |
462 | ||
a34e53fc | 463 | #define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1) |
ba3a79e7 GA |
464 | |
465 | /* STR_WITH_LEN() shortcuts */ | |
466 | #define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str)) | |
84bafc02 NC |
467 | #define newSVpvs_flags(str,flags) \ |
468 | Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags) | |
ba3a79e7 | 469 | #define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0) |
9dcc53ea Z |
470 | #define sv_catpvs_flags(sv, str, flags) \ |
471 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags) | |
472 | #define sv_catpvs_nomg(sv, str) \ | |
473 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0) | |
474 | #define sv_catpvs(sv, str) \ | |
475 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC) | |
476 | #define sv_catpvs_mg(sv, str) \ | |
477 | Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC) | |
3fe05580 | 478 | #define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str)) |
9dcc53ea Z |
479 | #define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str)) |
480 | #define sv_setref_pvs(rv, classname, str) \ | |
481 | Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str)) | |
ba3a79e7 | 482 | #define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str)) |
9dcc53ea Z |
483 | #define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str)) |
484 | #define gv_stashpvs(str, create) \ | |
485 | Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create) | |
3752a9fe KW |
486 | |
487 | #define gv_fetchpvs(namebeg, flags, sv_type) \ | |
488 | Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), flags, sv_type) | |
72659597 | 489 | #define gv_fetchpvn gv_fetchpvn_flags |
9dcc53ea Z |
490 | #define sv_catxmlpvs(dsv, str, utf8) \ |
491 | Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8) | |
4ac46235 | 492 | |
ba3a79e7 | 493 | |
510966aa Z |
494 | #define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags) |
495 | ||
b96d8cd9 NC |
496 | #define get_cvs(str, flags) \ |
497 | Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags)) | |
5c1737d1 | 498 | |
9b6e9510 | 499 | /* internal helpers */ |
4a1bbd3d | 500 | /* Transitional */ |
d67f622b N |
501 | #ifndef PERL_VERSION_MAJOR |
502 | # define PERL_VERSION_MAJOR PERL_REVISION | |
4a1bbd3d KW |
503 | #else |
504 | # undef PERL_REVISION /* We don't want code to be using these */ | |
505 | #endif | |
d67f622b N |
506 | #ifndef PERL_VERSION_MINOR |
507 | # define PERL_VERSION_MINOR PERL_VERSION | |
4a1bbd3d KW |
508 | #else |
509 | # undef PERL_VERSION | |
510 | #endif | |
d67f622b N |
511 | #ifndef PERL_VERSION_PATCH |
512 | # define PERL_VERSION_PATCH PERL_SUBVERSION | |
4a1bbd3d KW |
513 | #else |
514 | # undef PERL_SUBVERSION | |
515 | #endif | |
516 | ||
517 | #define PERL_JNP_TO_DECIMAL_(maJor,miNor,Patch) \ | |
518 | /* '10*' leaves room for things like alpha, beta, releases */ \ | |
519 | (10 * ((maJor) * 1000000) + ((miNor) * 1000) + (Patch)) | |
9b6e9510 | 520 | #define PERL_DECIMAL_VERSION_ \ |
d67f622b N |
521 | PERL_JNP_TO_DECIMAL_(PERL_VERSION_MAJOR, PERL_VERSION_MINOR, \ |
522 | PERL_VERSION_PATCH) | |
9b6e9510 KW |
523 | |
524 | /* | |
3f620621 | 525 | =for apidoc_section $versioning |
4a1bbd3d | 526 | =for apidoc AmR|bool|PERL_VERSION_EQ|const U8 major|const U8 minor|const U8 patch |
9de44d19 KW |
527 | =for apidoc_item PERL_VERSION_NE |
528 | =for apidoc_item PERL_VERSION_LT | |
529 | =for apidoc_item PERL_VERSION_LE | |
530 | =for apidoc_item PERL_VERSION_GT | |
531 | =for apidoc_item PERL_VERSION_GE | |
9b6e9510 | 532 | |
4a1bbd3d | 533 | Returns whether or not the perl currently being compiled has the specified |
9b6e9510 KW |
534 | relationship to the perl given by the parameters. For example, |
535 | ||
536 | #if PERL_VERSION_GT(5,24,2) | |
537 | code that will only be compiled on perls after v5.24.2 | |
538 | #else | |
539 | fallback code | |
540 | #endif | |
541 | ||
542 | Note that this is usable in making compile-time decisions | |
543 | ||
4a1bbd3d KW |
544 | You may use the special value '*' for the final number to mean ALL possible |
545 | values for it. Thus, | |
546 | ||
547 | #if PERL_VERSION_EQ(5,31,'*') | |
548 | ||
549 | means all perls in the 5.31 series. And | |
550 | ||
551 | #if PERL_VERSION_NE(5,24,'*') | |
552 | ||
553 | means all perls EXCEPT 5.24 ones. And | |
554 | ||
555 | #if PERL_VERSION_LE(5,9,'*') | |
556 | ||
557 | is effectively | |
558 | ||
559 | #if PERL_VERSION_LT(5,10,0) | |
560 | ||
561 | This means you don't have to think so much when converting from the existing | |
562 | deprecated C<PERL_VERSION> to using this macro: | |
563 | ||
564 | #if PERL_VERSION <= 9 | |
565 | ||
566 | becomes | |
567 | ||
568 | #if PERL_VERSION_LE(5,9,'*') | |
569 | ||
9b6e9510 KW |
570 | =cut |
571 | */ | |
572 | ||
4a1bbd3d KW |
573 | /* N.B. These don't work if the patch version is 42 or 92, as those are what |
574 | * '*' is in ASCII and EBCDIC respectively */ | |
575 | # define PERL_VERSION_EQ(j,n,p) \ | |
576 | (((p) == '*') \ | |
d67f622b N |
577 | ? ( (j) == PERL_VERSION_MAJOR \ |
578 | && (n) == PERL_VERSION_MINOR) \ | |
4a1bbd3d KW |
579 | : (PERL_DECIMAL_VERSION_ == PERL_JNP_TO_DECIMAL_(j,n,p))) |
580 | # define PERL_VERSION_NE(j,n,p) (! PERL_VERSION_EQ(j,n,p)) | |
581 | ||
582 | # define PERL_VERSION_LT(j,n,p) /* < '*' effectively means < 0 */ \ | |
583 | (PERL_DECIMAL_VERSION_ < PERL_JNP_TO_DECIMAL_( (j), \ | |
584 | (n), \ | |
585 | (((p) == '*') ? 0 : p))) | |
586 | # define PERL_VERSION_GE(j,n,p) (! PERL_VERSION_LT(j,n,p)) | |
587 | ||
588 | # define PERL_VERSION_LE(j,n,p) /* <= '*' effectively means < n+1 */ \ | |
589 | (PERL_DECIMAL_VERSION_ < PERL_JNP_TO_DECIMAL_( (j), \ | |
590 | (((p) == '*') ? ((n)+1) : (n)), \ | |
591 | (((p) == '*') ? 0 : p))) | |
592 | # define PERL_VERSION_GT(j,n,p) (! PERL_VERSION_LE(j,n,p)) | |
9b6e9510 | 593 | |
954c1994 | 594 | /* |
3f620621 | 595 | =for apidoc_section $string |
ccfc67b7 | 596 | |
954c1994 | 597 | =for apidoc Am|bool|strNE|char* s1|char* s2 |
dc6b0978 KW |
598 | Test two C<NUL>-terminated strings to see if they are different. Returns true |
599 | or false. | |
954c1994 GS |
600 | |
601 | =for apidoc Am|bool|strEQ|char* s1|char* s2 | |
dc6b0978 KW |
602 | Test two C<NUL>-terminated strings to see if they are equal. Returns true or |
603 | false. | |
954c1994 GS |
604 | |
605 | =for apidoc Am|bool|strLT|char* s1|char* s2 | |
dc6b0978 KW |
606 | Test two C<NUL>-terminated strings to see if the first, C<s1>, is less than the |
607 | second, C<s2>. Returns true or false. | |
954c1994 GS |
608 | |
609 | =for apidoc Am|bool|strLE|char* s1|char* s2 | |
dc6b0978 KW |
610 | Test two C<NUL>-terminated strings to see if the first, C<s1>, is less than or |
611 | equal to the second, C<s2>. Returns true or false. | |
954c1994 GS |
612 | |
613 | =for apidoc Am|bool|strGT|char* s1|char* s2 | |
dc6b0978 KW |
614 | Test two C<NUL>-terminated strings to see if the first, C<s1>, is greater than |
615 | the second, C<s2>. Returns true or false. | |
954c1994 GS |
616 | |
617 | =for apidoc Am|bool|strGE|char* s1|char* s2 | |
dc6b0978 KW |
618 | Test two C<NUL>-terminated strings to see if the first, C<s1>, is greater than |
619 | or equal to the second, C<s2>. Returns true or false. | |
954c1994 GS |
620 | |
621 | =for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len | |
dc6b0978 KW |
622 | Test two C<NUL>-terminated strings to see if they are different. The C<len> |
623 | parameter indicates the number of bytes to compare. Returns true or false. (A | |
954c1994 GS |
624 | wrapper for C<strncmp>). |
625 | ||
626 | =for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len | |
dc6b0978 KW |
627 | Test two C<NUL>-terminated strings to see if they are equal. The C<len> |
628 | parameter indicates the number of bytes to compare. Returns true or false. (A | |
629 | wrapper for C<strncmp>). | |
954c1994 | 630 | |
bd18bd40 KW |
631 | =for apidoc Am|bool|memEQ|char* s1|char* s2|STRLEN len |
632 | Test two buffers (which may contain embedded C<NUL> characters, to see if they | |
633 | are equal. The C<len> parameter indicates the number of bytes to compare. | |
634 | Returns zero if equal, or non-zero if non-equal. | |
635 | ||
3bb9fd01 | 636 | =for apidoc Am|bool|memEQs|char* s1|STRLEN l1|"s2" |
2d8eeddb KW |
637 | Like L</memEQ>, but the second string is a literal enclosed in double quotes, |
638 | C<l1> gives the number of bytes in C<s1>. | |
639 | Returns zero if equal, or non-zero if non-equal. | |
640 | ||
bd18bd40 KW |
641 | =for apidoc Am|bool|memNE|char* s1|char* s2|STRLEN len |
642 | Test two buffers (which may contain embedded C<NUL> characters, to see if they | |
643 | are not equal. The C<len> parameter indicates the number of bytes to compare. | |
644 | Returns zero if non-equal, or non-zero if equal. | |
645 | ||
3bb9fd01 | 646 | =for apidoc Am|bool|memNEs|char* s1|STRLEN l1|"s2" |
2d8eeddb KW |
647 | Like L</memNE>, but the second string is a literal enclosed in double quotes, |
648 | C<l1> gives the number of bytes in C<s1>. | |
649 | Returns zero if non-equal, or zero if non-equal. | |
650 | ||
4aada8b9 KW |
651 | =for apidoc Am|bool|memCHRs|"list"|char c |
652 | Returns the position of the first occurence of the byte C<c> in the literal | |
653 | string C<"list">, or NULL if C<c> doesn't appear in C<"list">. All bytes are | |
654 | treated as unsigned char. Thus this macro can be used to determine if C<c> is | |
655 | in a set of particular characters. Unlike L<strchr(3)>, it works even if C<c> | |
656 | is C<NUL> (and the set doesn't include C<NUL>). | |
657 | ||
954c1994 | 658 | =cut |
fc169e00 KW |
659 | |
660 | New macros should use the following conventions for their names (which are | |
661 | based on the underlying C library functions): | |
662 | ||
663 | (mem | str n? ) (EQ | NE | LT | GT | GE | (( BEGIN | END ) P? )) l? s? | |
664 | ||
665 | Each has two main parameters, string-like operands that are compared | |
666 | against each other, as specified by the macro name. Some macros may | |
667 | additionally have one or potentially even two length parameters. If a length | |
668 | parameter applies to both string parameters, it will be positioned third; | |
669 | otherwise any length parameter immediately follows the string parameter it | |
670 | applies to. | |
671 | ||
672 | If the prefix to the name is 'str', the string parameter is a pointer to a C | |
673 | language string. Such a string does not contain embedded NUL bytes; its | |
674 | length may be unknown, but can be calculated by C<strlen()>, since it is | |
675 | terminated by a NUL, which isn't included in its length. | |
676 | ||
a3815e44 | 677 | The optional 'n' following 'str' means that there is a third parameter, |
fc169e00 KW |
678 | giving the maximum number of bytes to look at in each string. Even if both |
679 | strings are longer than the length parameter, those extra bytes will be | |
680 | unexamined. | |
681 | ||
682 | The 's' suffix means that the 2nd byte string parameter is a literal C | |
683 | double-quoted string. Its length will automatically be calculated by the | |
684 | macro, so no length parameter will ever be needed for it. | |
685 | ||
686 | If the prefix is 'mem', the string parameters don't have to be C strings; | |
687 | they may contain embedded NUL bytes, do not necessarily have a terminating | |
688 | NUL, and their lengths can be known only through other means, which in | |
689 | practice are additional parameter(s) passed to the function. All 'mem' | |
690 | functions have at least one length parameter. Barring any 'l' or 's' suffix, | |
691 | there is a single length parameter, in position 3, which applies to both | |
692 | string parameters. The 's' suffix means, as described above, that the 2nd | |
693 | string is a literal double-quoted C string (hence its length is calculated by | |
694 | the macro, and the length parameter to the function applies just to the first | |
695 | string parameter, and hence is positioned just after it). An 'l' suffix | |
696 | means that the 2nd string parameter has its own length parameter, and the | |
697 | signature will look like memFOOl(s1, l1, s2, l2). | |
698 | ||
699 | BEGIN (and END) are for testing if the 2nd string is an initial (or final) | |
700 | substring of the 1st string. 'P' if present indicates that the substring | |
701 | must be a "proper" one in tha mathematical sense that the first one must be | |
702 | strictly larger than the 2nd. | |
703 | ||
954c1994 GS |
704 | */ |
705 | ||
62946e08 | 706 | |
75400963 KW |
707 | #define strNE(s1,s2) (strcmp(s1,s2) != 0) |
708 | #define strEQ(s1,s2) (strcmp(s1,s2) == 0) | |
8d063cd8 LW |
709 | #define strLT(s1,s2) (strcmp(s1,s2) < 0) |
710 | #define strLE(s1,s2) (strcmp(s1,s2) <= 0) | |
711 | #define strGT(s1,s2) (strcmp(s1,s2) > 0) | |
712 | #define strGE(s1,s2) (strcmp(s1,s2) >= 0) | |
62946e08 | 713 | |
75400963 KW |
714 | #define strnNE(s1,s2,l) (strncmp(s1,s2,l) != 0) |
715 | #define strnEQ(s1,s2,l) (strncmp(s1,s2,l) == 0) | |
378cc40b | 716 | |
9d3980bc KW |
717 | #define memEQ(s1,s2,l) (memcmp(((const void *) (s1)), ((const void *) (s2)), l) == 0) |
718 | #define memNE(s1,s2,l) (! memEQ(s1,s2,l)) | |
36477c24 | 719 | |
085b7534 | 720 | /* memEQ and memNE where second comparand is a string constant */ |
568a785a | 721 | #define memEQs(s1, l, s2) \ |
777fa2cb | 722 | (((sizeof(s2)-1) == (l)) && memEQ((s1), ("" s2 ""), (sizeof(s2)-1))) |
5f50c6c9 | 723 | #define memNEs(s1, l, s2) (! memEQs(s1, l, s2)) |
568a785a | 724 | |
fdbb9a7c KW |
725 | /* Keep these private until we decide it was a good idea */ |
726 | #if defined(PERL_CORE) || defined(PERL_EXT) || defined(PERL_EXT_POSIX) | |
727 | ||
728 | #define strBEGINs(s1,s2) (strncmp(s1,"" s2 "", sizeof(s2)-1) == 0) | |
729 | ||
bdb7e3f0 | 730 | #define memBEGINs(s1, l, s2) \ |
30a6480c | 731 | ( (Ptrdiff_t) (l) >= (Ptrdiff_t) sizeof(s2) - 1 \ |
bdb7e3f0 | 732 | && memEQ(s1, "" s2 "", sizeof(s2)-1)) |
de627158 | 733 | #define memBEGINPs(s1, l, s2) \ |
30a6480c | 734 | ( (Ptrdiff_t) (l) > (Ptrdiff_t) sizeof(s2) - 1 \ |
de627158 | 735 | && memEQ(s1, "" s2 "", sizeof(s2)-1)) |
bdb7e3f0 | 736 | #define memENDs(s1, l, s2) \ |
30a6480c | 737 | ( (Ptrdiff_t) (l) >= (Ptrdiff_t) sizeof(s2) - 1 \ |
bdb7e3f0 | 738 | && memEQ(s1 + (l) - (sizeof(s2) - 1), "" s2 "", sizeof(s2)-1)) |
b80f8424 | 739 | #define memENDPs(s1, l, s2) \ |
30a6480c | 740 | ( (Ptrdiff_t) (l) > (Ptrdiff_t) sizeof(s2) \ |
b80f8424 | 741 | && memEQ(s1 + (l) - (sizeof(s2) - 1), "" s2 "", sizeof(s2)-1)) |
fdbb9a7c | 742 | #endif /* End of making macros private */ |
bdb7e3f0 | 743 | |
062b6850 KW |
744 | #define memLT(s1,s2,l) (memcmp(s1,s2,l) < 0) |
745 | #define memLE(s1,s2,l) (memcmp(s1,s2,l) <= 0) | |
746 | #define memGT(s1,s2,l) (memcmp(s1,s2,l) > 0) | |
747 | #define memGE(s1,s2,l) (memcmp(s1,s2,l) >= 0) | |
748 | ||
4aada8b9 KW |
749 | #define memCHRs(s1,c) ((const char *) memchr("" s1 "" , c, sizeof(s1)-1)) |
750 | ||
bbce6d69 | 751 | /* |
752 | * Character classes. | |
753 | * | |
754 | * Unfortunately, the introduction of locales means that we | |
755 | * can't trust isupper(), etc. to tell the truth. And when | |
756 | * it comes to /\w+/ with tainting enabled, we *must* be able | |
757 | * to trust our character classes. | |
758 | * | |
759 | * Therefore, the default tests in the text of Perl will be | |
760 | * independent of locale. Any code that wants to depend on | |
761 | * the current locale will use the tests that begin with "lc". | |
762 | */ | |
763 | ||
2304df62 AD |
764 | #ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */ |
765 | # ifndef CTYPE256 | |
766 | # define CTYPE256 | |
767 | # endif | |
768 | #endif | |
769 | ||
954c1994 | 770 | /* |
ccfc67b7 | 771 | |
dcccc8ff | 772 | =head1 Character classification |
243effed KW |
773 | This section is about functions (really macros) that classify characters |
774 | into types, such as punctuation versus alphabetic, etc. Most of these are | |
775 | analogous to regular expression character classes. (See | |
776 | L<perlrecharclass/POSIX Character Classes>.) There are several variants for | |
777 | each class. (Not all macros have all variants; each item below lists the | |
778 | ones valid for it.) None are affected by C<use bytes>, and only the ones | |
779 | with C<LC> in the name are affected by the current locale. | |
780 | ||
d713f9d9 KW |
781 | The base function, e.g., C<isALPHA()>, takes any signed or unsigned value, |
782 | treating it as a code point, and returns a boolean as to whether or not the | |
783 | character represented by it is (or on non-ASCII platforms, corresponds to) an | |
6aff1f14 KW |
784 | ASCII character in the named class based on platform, Unicode, and Perl rules. |
785 | If the input is a number that doesn't fit in an octet, FALSE is returned. | |
243effed | 786 | |
c98722a4 | 787 | Variant C<isI<FOO>_A> (e.g., C<isALPHA_A()>) is identical to the base function |
550da823 KW |
788 | with no suffix C<"_A">. This variant is used to emphasize by its name that |
789 | only ASCII-range characters can return TRUE. | |
4b9734bf | 790 | |
d60679e1 | 791 | Variant C<isI<FOO>_L1> imposes the Latin-1 (or EBCDIC equivalent) character set |
4b9734bf KW |
792 | onto the platform. That is, the code points that are ASCII are unaffected, |
793 | since ASCII is a subset of Latin-1. But the non-ASCII code points are treated | |
794 | as if they are Latin-1 characters. For example, C<isWORDCHAR_L1()> will return | |
795 | true when called with the code point 0xDF, which is a word character in both | |
4650c663 | 796 | ASCII and EBCDIC (though it represents different characters in each). |
d713f9d9 KW |
797 | If the input is a number that doesn't fit in an octet, FALSE is returned. |
798 | (Perl's documentation uses a colloquial definition of Latin-1, to include all | |
799 | code points below 256.) | |
243effed | 800 | |
d713f9d9 KW |
801 | Variant C<isI<FOO>_uvchr> is exactly like the C<isI<FOO>_L1> variant, for |
802 | inputs below 256, but if the code point is larger than 255, Unicode rules are | |
803 | used to determine if it is in the character class. For example, | |
d0da05db | 804 | C<isWORDCHAR_uvchr(0x100)> returns TRUE, since 0x100 is LATIN CAPITAL LETTER A |
6aff1f14 | 805 | WITH MACRON in Unicode, and is a word character. |
243effed | 806 | |
059703b0 KW |
807 | Variants C<isI<FOO>_utf8> and C<isI<FOO>_utf8_safe> are like C<isI<FOO>_uvchr>, |
808 | but are used for UTF-8 encoded strings. The two forms are different names for | |
809 | the same thing. Each call to one of these classifies the first character of | |
810 | the string starting at C<p>. The second parameter, C<e>, points to anywhere in | |
811 | the string beyond the first character, up to one byte past the end of the | |
812 | entire string. Although both variants are identical, the suffix C<_safe> in | |
813 | one name emphasizes that it will not attempt to read beyond S<C<e - 1>>, | |
814 | provided that the constraint S<C<s E<lt> e>> is true (this is asserted for in | |
815 | C<-DDEBUGGING> builds). If the UTF-8 for the input character is malformed in | |
816 | some way, the program may croak, or the function may return FALSE, at the | |
817 | discretion of the implementation, and subject to change in future releases. | |
243effed | 818 | |
d713f9d9 KW |
819 | Variant C<isI<FOO>_LC> is like the C<isI<FOO>_A> and C<isI<FOO>_L1> variants, |
820 | but the result is based on the current locale, which is what C<LC> in the name | |
821 | stands for. If Perl can determine that the current locale is a UTF-8 locale, | |
822 | it uses the published Unicode rules; otherwise, it uses the C library function | |
823 | that gives the named classification. For example, C<isDIGIT_LC()> when not in | |
824 | a UTF-8 locale returns the result of calling C<isdigit()>. FALSE is always | |
1a83413c KW |
825 | returned if the input won't fit into an octet. On some platforms where the C |
826 | library function is known to be defective, Perl changes its result to follow | |
827 | the POSIX standard's rules. | |
243effed | 828 | |
d713f9d9 KW |
829 | Variant C<isI<FOO>_LC_uvchr> acts exactly like C<isI<FOO>_LC> for inputs less |
830 | than 256, but for larger ones it returns the Unicode classification of the code | |
831 | point. | |
243effed | 832 | |
059703b0 KW |
833 | Variants C<isI<FOO>_LC_utf8> and C<isI<FOO>_LC_utf8_safe> are like |
834 | C<isI<FOO>_LC_uvchr>, but are used for UTF-8 encoded strings. The two forms | |
835 | are different names for the same thing. Each call to one of these classifies | |
836 | the first character of the string starting at C<p>. The second parameter, | |
837 | C<e>, points to anywhere in the string beyond the first character, up to one | |
838 | byte past the end of the entire string. Although both variants are identical, | |
839 | the suffix C<_safe> in one name emphasizes that it will not attempt to read | |
840 | beyond S<C<e - 1>>, provided that the constraint S<C<s E<lt> e>> is true (this | |
841 | is asserted for in C<-DDEBUGGING> builds). If the UTF-8 for the input | |
842 | character is malformed in some way, the program may croak, or the function may | |
843 | return FALSE, at the discretion of the implementation, and subject to change in | |
844 | future releases. | |
ccfc67b7 | 845 | |
6fdd32c3 KW |
846 | =for apidoc Am|bool|isALPHA|UV ch |
847 | =for apidoc_item ||isALPHA_A|UV ch | |
848 | =for apidoc_item ||isALPHA_L1|UV ch | |
849 | =for apidoc_item ||isALPHA_uvchr|UV ch | |
fa470d81 KW |
850 | =for apidoc_item ||isALPHA_utf8_safe|U8 * s|U8 * end |
851 | =for apidoc_item ||isALPHA_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
852 | =for apidoc_item ||isALPHA_LC|UV ch |
853 | =for apidoc_item ||isALPHA_LC_uvchr|UV ch | |
fa470d81 | 854 | =for apidoc_item ||isALPHA_LC_utf8_safe|U8 * s| U8 *end |
d713f9d9 KW |
855 | Returns a boolean indicating whether the specified input is one of C<[A-Za-z]>, |
856 | analogous to C<m/[[:alpha:]]/>. | |
dcccc8ff | 857 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 | 858 | the variants. |
8a58bdcf | 859 | |
f16858ed KW |
860 | =cut |
861 | ||
f1460a66 | 862 | Here and below, we add the prototypes of these macros for downstream programs |
f16858ed KW |
863 | that would be interested in them, such as Devel::PPPort |
864 | ||
6fdd32c3 KW |
865 | =for apidoc Am|bool|isALPHANUMERIC|UV ch |
866 | =for apidoc_item ||isALPHANUMERIC_A|UV ch | |
867 | =for apidoc_item ||isALPHANUMERIC_L1|UV ch | |
868 | =for apidoc_item ||isALPHANUMERIC_uvchr|UV ch | |
fa470d81 KW |
869 | =for apidoc_item ||isALPHANUMERIC_utf8_safe|U8 * s|U8 * end |
870 | =for apidoc_item ||isALPHANUMERIC_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
871 | =for apidoc_item ||isALPHANUMERIC_LC|UV ch |
872 | =for apidoc_item ||isALPHANUMERIC_LC_uvchr|UV ch | |
fa470d81 | 873 | =for apidoc_item ||isALPHANUMERIC_LC_utf8_safe|U8 * s| U8 *end |
6fdd32c3 KW |
874 | =for apidoc_item ||isALNUMC|UV ch |
875 | =for apidoc_item ||isALNUMC_A|UV ch | |
876 | =for apidoc_item ||isALNUMC_L1|UV ch | |
877 | =for apidoc_item ||isALNUMC_LC|UV ch | |
878 | =for apidoc_item ||isALNUMC_LC_uvchr|UV ch | |
d713f9d9 KW |
879 | Returns a boolean indicating whether the specified character is one of |
880 | C<[A-Za-z0-9]>, analogous to C<m/[[:alnum:]]/>. | |
dcccc8ff | 881 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 | 882 | the variants. |
15861f94 | 883 | |
255b632a KW |
884 | A (discouraged from use) synonym is C<isALNUMC> (where the C<C> suffix means |
885 | this corresponds to the C language alphanumeric definition). Also | |
886 | there are the variants | |
887 | C<isALNUMC_A>, C<isALNUMC_L1> | |
888 | C<isALNUMC_LC>, and C<isALNUMC_LC_uvchr>. | |
889 | ||
6fdd32c3 KW |
890 | =for apidoc Am|bool|isASCII|UV ch |
891 | =for apidoc_item ||isASCII_A|UV ch | |
892 | =for apidoc_item ||isASCII_L1|UV ch | |
893 | =for apidoc_item ||isASCII_uvchr|UV ch | |
fa470d81 KW |
894 | =for apidoc_item ||isASCII_utf8_safe|U8 * s|U8 * end |
895 | =for apidoc_item ||isASCII_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
896 | =for apidoc_item ||isASCII_LC|UV ch |
897 | =for apidoc_item ||isASCII_LC_uvchr|UV ch | |
fa470d81 | 898 | =for apidoc_item ||isASCII_LC_utf8_safe|U8 * s| U8 *end |
8a58bdcf | 899 | Returns a boolean indicating whether the specified character is one of the 128 |
243effed | 900 | characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>. |
e5ad6aba | 901 | On non-ASCII platforms, it returns TRUE iff this |
8a58bdcf KW |
902 | character corresponds to an ASCII character. Variants C<isASCII_A()> and |
903 | C<isASCII_L1()> are identical to C<isASCII()>. | |
dcccc8ff | 904 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 | 905 | the variants. |
059703b0 KW |
906 | Note, however, that some platforms do not have the C library routine |
907 | C<isascii()>. In these cases, the variants whose names contain C<LC> are the | |
908 | same as the corresponding ones without. | |
243effed | 909 | |
d98532ea KW |
910 | Also note, that because all ASCII characters are UTF-8 invariant (meaning they |
911 | have the exact same representation (always a single byte) whether encoded in | |
912 | UTF-8 or not), C<isASCII> will give the correct results when called with any | |
059703b0 KW |
913 | byte in any string encoded or not in UTF-8. And similarly C<isASCII_utf8> and |
914 | C<isASCII_utf8_safe> will work properly on any string encoded or not in UTF-8. | |
d98532ea | 915 | |
6fdd32c3 KW |
916 | =for apidoc Am|bool|isBLANK|UV ch |
917 | =for apidoc_item ||isBLANK_A|UV ch | |
918 | =for apidoc_item ||isBLANK_L1|UV ch | |
919 | =for apidoc_item ||isBLANK_uvchr|UV ch | |
fa470d81 KW |
920 | =for apidoc_item ||isBLANK_utf8_safe|U8 * s|U8 * end |
921 | =for apidoc_item ||isBLANK_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
922 | =for apidoc_item ||isBLANK_LC|UV ch |
923 | =for apidoc_item ||isBLANK_LC_uvchr|UV ch | |
fa470d81 | 924 | =for apidoc_item ||isBLANK_LC_utf8_safe|U8 * s| U8 *end |
243effed | 925 | Returns a boolean indicating whether the specified character is a |
6aff1f14 | 926 | character considered to be a blank, analogous to C<m/[[:blank:]]/>. |
dcccc8ff | 927 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
928 | the variants. |
929 | Note, | |
da8c1a98 KW |
930 | however, that some platforms do not have the C library routine |
931 | C<isblank()>. In these cases, the variants whose names contain C<LC> are | |
932 | the same as the corresponding ones without. | |
243effed | 933 | |
6fdd32c3 KW |
934 | =for apidoc Am|bool|isCNTRL|UV ch |
935 | =for apidoc_item ||isCNTRL_A|UV ch | |
936 | =for apidoc_item ||isCNTRL_L1|UV ch | |
937 | =for apidoc_item ||isCNTRL_uvchr|UV ch | |
fa470d81 KW |
938 | =for apidoc_item ||isCNTRL_utf8_safe|U8 * s|U8 * end |
939 | =for apidoc_item ||isCNTRL_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
940 | =for apidoc_item ||isCNTRL_LC|UV ch |
941 | =for apidoc_item ||isCNTRL_LC_uvchr|UV ch | |
fa470d81 | 942 | =for apidoc_item ||isCNTRL_LC_utf8_safe|U8 * s| U8 *end |
f16858ed | 943 | |
243effed | 944 | Returns a boolean indicating whether the specified character is a |
6aff1f14 | 945 | control character, analogous to C<m/[[:cntrl:]]/>. |
dcccc8ff | 946 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
947 | the variants. |
948 | On EBCDIC platforms, you almost always want to use the C<isCNTRL_L1> variant. | |
949 | ||
6fdd32c3 KW |
950 | =for apidoc Am|bool|isDIGIT|UV ch |
951 | =for apidoc_item ||isDIGIT_A|UV ch | |
952 | =for apidoc_item ||isDIGIT_L1|UV ch | |
953 | =for apidoc_item ||isDIGIT_uvchr|UV ch | |
fa470d81 KW |
954 | =for apidoc_item ||isDIGIT_utf8_safe|U8 * s|U8 * end |
955 | =for apidoc_item ||isDIGIT_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
956 | =for apidoc_item ||isDIGIT_LC|UV ch |
957 | =for apidoc_item ||isDIGIT_LC_uvchr|UV ch | |
fa470d81 KW |
958 | =for apidoc_item ||isDIGIT_LC_utf8_safe|U8 * s| U8 *end |
959 | ||
2787a470 | 960 | Returns a boolean indicating whether the specified character is a |
6aff1f14 | 961 | digit, analogous to C<m/[[:digit:]]/>. |
8a58bdcf | 962 | Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>. |
dcccc8ff | 963 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
964 | the variants. |
965 | ||
6fdd32c3 KW |
966 | =for apidoc Am|bool|isGRAPH|UV ch |
967 | =for apidoc_item ||isGRAPH_A|UV ch | |
968 | =for apidoc_item ||isGRAPH_L1|UV ch | |
969 | =for apidoc_item ||isGRAPH_uvchr|UV ch | |
fa470d81 KW |
970 | =for apidoc_item ||isGRAPH_utf8_safe|U8 * s|U8 * end |
971 | =for apidoc_item ||isGRAPH_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
972 | =for apidoc_item ||isGRAPH_LC|UV ch |
973 | =for apidoc_item ||isGRAPH_LC_uvchr|UV ch | |
fa470d81 | 974 | =for apidoc_item ||isGRAPH_LC_utf8_safe|U8 * s| U8 *end |
243effed | 975 | Returns a boolean indicating whether the specified character is a |
6aff1f14 | 976 | graphic character, analogous to C<m/[[:graph:]]/>. |
dcccc8ff | 977 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
978 | the variants. |
979 | ||
6fdd32c3 KW |
980 | =for apidoc Am|bool|isLOWER|UV ch |
981 | =for apidoc_item ||isLOWER_A|UV ch | |
982 | =for apidoc_item ||isLOWER_L1|UV ch | |
983 | =for apidoc_item ||isLOWER_uvchr|UV ch | |
fa470d81 KW |
984 | =for apidoc_item ||isLOWER_utf8_safe|U8 * s|U8 * end |
985 | =for apidoc_item ||isLOWER_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
986 | =for apidoc_item ||isLOWER_LC|UV ch |
987 | =for apidoc_item ||isLOWER_LC_uvchr|UV ch | |
fa470d81 | 988 | =for apidoc_item ||isLOWER_LC_utf8_safe|U8 * s| U8 *end |
2787a470 | 989 | Returns a boolean indicating whether the specified character is a |
6aff1f14 | 990 | lowercase character, analogous to C<m/[[:lower:]]/>. |
dcccc8ff | 991 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
992 | the variants |
993 | ||
6fdd32c3 KW |
994 | =for apidoc Am|bool|isOCTAL|UV ch |
995 | =for apidoc_item ||isOCTAL_A|UV ch | |
996 | =for apidoc_item ||isOCTAL_L1|UV ch | |
2787a470 | 997 | Returns a boolean indicating whether the specified character is an |
6aff1f14 | 998 | octal digit, [0-7]. |
243effed KW |
999 | The only two variants are C<isOCTAL_A> and C<isOCTAL_L1>; each is identical to |
1000 | C<isOCTAL>. | |
1001 | ||
6fdd32c3 KW |
1002 | =for apidoc Am|bool|isPUNCT|UV ch |
1003 | =for apidoc_item ||isPUNCT_A|UV ch | |
1004 | =for apidoc_item ||isPUNCT_L1|UV ch | |
1005 | =for apidoc_item ||isPUNCT_uvchr|UV ch | |
fa470d81 KW |
1006 | =for apidoc_item ||isPUNCT_utf8_safe|U8 * s|U8 * end |
1007 | =for apidoc_item ||isPUNCT_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
1008 | =for apidoc_item ||isPUNCT_LC|UV ch |
1009 | =for apidoc_item ||isPUNCT_LC_uvchr|UV ch | |
fa470d81 | 1010 | =for apidoc_item ||isPUNCT_LC_utf8_safe|U8 * s| U8 *end |
243effed | 1011 | Returns a boolean indicating whether the specified character is a |
6aff1f14 KW |
1012 | punctuation character, analogous to C<m/[[:punct:]]/>. |
1013 | Note that the definition of what is punctuation isn't as | |
243effed KW |
1014 | straightforward as one might desire. See L<perlrecharclass/POSIX Character |
1015 | Classes> for details. | |
dcccc8ff | 1016 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
1017 | the variants. |
1018 | ||
6fdd32c3 KW |
1019 | =for apidoc Am|bool|isSPACE|UV ch |
1020 | =for apidoc_item ||isSPACE_A|UV ch | |
1021 | =for apidoc_item ||isSPACE_L1|UV ch | |
1022 | =for apidoc_item ||isSPACE_uvchr|UV ch | |
fa470d81 KW |
1023 | =for apidoc_item ||isSPACE_utf8_safe|U8 * s|U8 * end |
1024 | =for apidoc_item ||isSPACE_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
1025 | =for apidoc_item ||isSPACE_LC|UV ch |
1026 | =for apidoc_item ||isSPACE_LC_uvchr|UV ch | |
fa470d81 | 1027 | =for apidoc_item ||isSPACE_LC_utf8_safe|U8 * s| U8 *end |
2787a470 | 1028 | Returns a boolean indicating whether the specified character is a |
6aff1f14 | 1029 | whitespace character. This is analogous |
398d098a | 1030 | to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18 |
779cf272 | 1031 | this also matches what C<m/[[:space:]]/> does. Prior to 5.18, only the |
398d098a KW |
1032 | locale forms of this macro (the ones with C<LC> in their names) matched |
1033 | precisely what C<m/[[:space:]]/> does. In those releases, the only difference, | |
1034 | in the non-locale variants, was that C<isSPACE()> did not match a vertical tab. | |
1035 | (See L</isPSXSPC> for a macro that matches a vertical tab in all releases.) | |
dcccc8ff | 1036 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
1037 | the variants. |
1038 | ||
6fdd32c3 KW |
1039 | =for apidoc Am|bool|isPSXSPC|UV ch |
1040 | =for apidoc_item ||isPSXSPC_A|UV ch | |
1041 | =for apidoc_item ||isPSXSPC_L1|UV ch | |
1042 | =for apidoc_item ||isPSXSPC_uvchr|UV ch | |
fa470d81 KW |
1043 | =for apidoc_item ||isPSXSPC_utf8_safe|U8 * s|U8 * end |
1044 | =for apidoc_item ||isPSXSPC_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
1045 | =for apidoc_item ||isPSXSPC_LC|UV ch |
1046 | =for apidoc_item ||isPSXSPC_LC_uvchr|UV ch | |
fa470d81 | 1047 | =for apidoc_item ||isPSXSPC_LC_utf8_safe|U8 * s| U8 *end |
398d098a | 1048 | (short for Posix Space) |
779cf272 KW |
1049 | Starting in 5.18, this is identical in all its forms to the |
1050 | corresponding C<isSPACE()> macros. | |
398d098a KW |
1051 | The locale forms of this macro are identical to their corresponding |
1052 | C<isSPACE()> forms in all Perl releases. In releases prior to 5.18, the | |
1053 | non-locale forms differ from their C<isSPACE()> forms only in that the | |
1054 | C<isSPACE()> forms don't match a Vertical Tab, and the C<isPSXSPC()> forms do. | |
1055 | Otherwise they are identical. Thus this macro is analogous to what | |
1056 | C<m/[[:space:]]/> matches in a regular expression. | |
dcccc8ff | 1057 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
1058 | the variants. |
1059 | ||
6fdd32c3 KW |
1060 | =for apidoc Am|bool|isUPPER|UV ch |
1061 | =for apidoc_item ||isUPPER_A|UV ch | |
1062 | =for apidoc_item ||isUPPER_L1|UV ch | |
1063 | =for apidoc_item ||isUPPER_uvchr|UV ch | |
fa470d81 KW |
1064 | =for apidoc_item ||isUPPER_utf8_safe|U8 * s|U8 * end |
1065 | =for apidoc_item ||isUPPER_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
1066 | =for apidoc_item ||isUPPER_LC|UV ch |
1067 | =for apidoc_item ||isUPPER_LC_uvchr|UV ch | |
fa470d81 | 1068 | =for apidoc_item ||isUPPER_LC_utf8_safe|U8 * s| U8 *end |
2787a470 | 1069 | Returns a boolean indicating whether the specified character is an |
6aff1f14 | 1070 | uppercase character, analogous to C<m/[[:upper:]]/>. |
dcccc8ff | 1071 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
1072 | the variants. |
1073 | ||
6fdd32c3 KW |
1074 | =for apidoc Am|bool|isPRINT|UV ch |
1075 | =for apidoc_item ||isPRINT_A|UV ch | |
1076 | =for apidoc_item ||isPRINT_L1|UV ch | |
1077 | =for apidoc_item ||isPRINT_uvchr|UV ch | |
fa470d81 KW |
1078 | =for apidoc_item ||isPRINT_utf8_safe|U8 * s|U8 * end |
1079 | =for apidoc_item ||isPRINT_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
1080 | =for apidoc_item ||isPRINT_LC|UV ch |
1081 | =for apidoc_item ||isPRINT_LC_uvchr|UV ch | |
fa470d81 | 1082 | =for apidoc_item ||isPRINT_LC_utf8_safe|U8 * s| U8 *end |
8eea39dd | 1083 | Returns a boolean indicating whether the specified character is a |
6aff1f14 | 1084 | printable character, analogous to C<m/[[:print:]]/>. |
dcccc8ff | 1085 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
1086 | the variants. |
1087 | ||
6fdd32c3 KW |
1088 | =for apidoc Am|bool|isWORDCHAR|UV ch |
1089 | =for apidoc_item ||isWORDCHAR_A|UV ch | |
1090 | =for apidoc_item ||isWORDCHAR_L1|UV ch | |
1091 | =for apidoc_item ||isWORDCHAR_uvchr|UV ch | |
fa470d81 KW |
1092 | =for apidoc_item ||isWORDCHAR_utf8_safe|U8 * s|U8 * end |
1093 | =for apidoc_item ||isWORDCHAR_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
1094 | =for apidoc_item ||isWORDCHAR_LC|UV ch |
1095 | =for apidoc_item ||isWORDCHAR_LC_uvchr|UV ch | |
fa470d81 | 1096 | =for apidoc_item ||isWORDCHAR_LC_utf8_safe|U8 * s| U8 *end |
6fdd32c3 KW |
1097 | =for apidoc_item ||isALNUM|UV ch |
1098 | =for apidoc_item ||isALNUM_A|UV ch | |
1099 | =for apidoc_item ||isALNUM_LC|UV ch | |
1100 | =for apidoc_item ||isALNUM_LC_uvchr|UV ch | |
243effed KW |
1101 | Returns a boolean indicating whether the specified character is a character |
1102 | that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match | |
1103 | in a regular expression. A word character is an alphabetic character, a | |
1104 | decimal digit, a connecting punctuation character (such as an underscore), or | |
1105 | a "mark" character that attaches to one of those (like some sort of accent). | |
1106 | C<isALNUM()> is a synonym provided for backward compatibility, even though a | |
1107 | word character includes more than the standard C language meaning of | |
1108 | alphanumeric. | |
dcccc8ff | 1109 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
1110 | the variants. |
1111 | C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uvchr>, | |
1112 | C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, C<isWORDCHAR_LC_utf8>, and | |
1113 | C<isWORDCHAR_LC_utf8_safe> are also as described there, but additionally | |
1114 | include the platform's native underscore. | |
1115 | ||
6fdd32c3 KW |
1116 | =for apidoc Am|bool|isXDIGIT|UV ch |
1117 | =for apidoc_item ||isXDIGIT_A|UV ch | |
1118 | =for apidoc_item ||isXDIGIT_L1|UV ch | |
1119 | =for apidoc_item ||isXDIGIT_uvchr|UV ch | |
fa470d81 KW |
1120 | =for apidoc_item ||isXDIGIT_utf8_safe|U8 * s|U8 * end |
1121 | =for apidoc_item ||isXDIGIT_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
1122 | =for apidoc_item ||isXDIGIT_LC|UV ch |
1123 | =for apidoc_item ||isXDIGIT_LC_uvchr|UV ch | |
fa470d81 | 1124 | =for apidoc_item ||isXDIGIT_LC_utf8_safe|U8 * s| U8 *end |
8a58bdcf | 1125 | Returns a boolean indicating whether the specified character is a hexadecimal |
243effed KW |
1126 | digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()> |
1127 | and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>. | |
dcccc8ff | 1128 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
1129 | the variants. |
1130 | ||
6fdd32c3 KW |
1131 | =for apidoc Am|bool|isIDFIRST|UV ch |
1132 | =for apidoc_item ||isIDFIRST_A|UV ch | |
1133 | =for apidoc_item ||isIDFIRST_L1|UV ch | |
1134 | =for apidoc_item ||isIDFIRST_uvchr|UV ch | |
fa470d81 KW |
1135 | =for apidoc_item ||isIDFIRST_utf8_safe|U8 * s|U8 * end |
1136 | =for apidoc_item ||isIDFIRST_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
1137 | =for apidoc_item ||isIDFIRST_LC|UV ch |
1138 | =for apidoc_item ||isIDFIRST_LC_uvchr|UV ch | |
fa470d81 | 1139 | =for apidoc_item ||isIDFIRST_LC_utf8_safe|U8 * s| U8 *end |
3c3ecf18 KW |
1140 | Returns a boolean indicating whether the specified character can be the first |
1141 | character of an identifier. This is very close to, but not quite the same as | |
1142 | the official Unicode property C<XID_Start>. The difference is that this | |
1143 | returns true only if the input character also matches L</isWORDCHAR>. | |
dcccc8ff | 1144 | See the L<top of this section|/Character classification> for an explanation of |
fa470d81 KW |
1145 | the variants. |
1146 | ||
6fdd32c3 KW |
1147 | =for apidoc Am|bool|isIDCONT|UV ch |
1148 | =for apidoc_item ||isIDCONT_A|UV ch | |
1149 | =for apidoc_item ||isIDCONT_L1|UV ch | |
1150 | =for apidoc_item ||isIDCONT_uvchr|UV ch | |
fa470d81 KW |
1151 | =for apidoc_item ||isIDCONT_utf8_safe|U8 * s|U8 * end |
1152 | =for apidoc_item ||isIDCONT_utf8|U8 * s|U8 * end | |
6fdd32c3 KW |
1153 | =for apidoc_item ||isIDCONT_LC|UV ch |
1154 | =for apidoc_item ||isIDCONT_LC_uvchr|UV ch | |
fa470d81 | 1155 | =for apidoc_item ||isIDCONT_LC_utf8_safe|U8 * s| U8 *end |
3c3ecf18 KW |
1156 | Returns a boolean indicating whether the specified character can be the |
1157 | second or succeeding character of an identifier. This is very close to, but | |
1158 | not quite the same as the official Unicode property C<XID_Continue>. The | |
1159 | difference is that this returns true only if the input character also matches | |
dcccc8ff | 1160 | L</isWORDCHAR>. See the L<top of this section|/Character classification> for |
fa470d81 | 1161 | an explanation of the variants. |
f16858ed | 1162 | |
3f620621 | 1163 | =for apidoc_section $numeric |
8eea39dd | 1164 | |
95a59cab | 1165 | =for apidoc Am|U8|READ_XDIGIT|char str* |
243effed | 1166 | Returns the value of an ASCII-range hex digit and advances the string pointer. |
95a59cab YO |
1167 | Behaviour is only well defined when isXDIGIT(*str) is true. |
1168 | ||
e7c1e6c1 | 1169 | =head1 Character case changing |
21da7284 KW |
1170 | Perl uses "full" Unicode case mappings. This means that converting a single |
1171 | character to another case may result in a sequence of more than one character. | |
1172 | For example, the uppercase of C<E<223>> (LATIN SMALL LETTER SHARP S) is the two | |
1173 | character sequence C<SS>. This presents some complications The lowercase of | |
1174 | all characters in the range 0..255 is a single character, and thus | |
1175 | C<L</toLOWER_L1>> is furnished. But, C<toUPPER_L1> can't exist, as it couldn't | |
1176 | return a valid result for all legal inputs. Instead C<L</toUPPER_uvchr>> has | |
1177 | an API that does allow every possible legal result to be returned.) Likewise | |
1178 | no other function that is crippled by not being able to give the correct | |
1179 | results for the full range of possible inputs has been implemented here. | |
e7c1e6c1 | 1180 | |
d713f9d9 | 1181 | =for apidoc Am|U8|toUPPER|int ch |
1f607577 KW |
1182 | Converts the specified character to uppercase. If the input is anything but an |
1183 | ASCII lowercase character, that input character itself is returned. Variant | |
c753c8d3 | 1184 | C<toUPPER_A> is equivalent. |
954c1994 | 1185 | |
d0da05db KW |
1186 | =for apidoc Am|UV|toUPPER_uvchr|UV cp|U8* s|STRLEN* lenp |
1187 | Converts the code point C<cp> to its uppercase version, and | |
1188 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code | |
1189 | point is interpreted as native if less than 256; otherwise as Unicode. Note | |
1f607577 KW |
1190 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> |
1191 | bytes since the uppercase version may be longer than the original character. | |
1192 | ||
1193 | The first code point of the uppercased version is returned | |
21da7284 KW |
1194 | (but note, as explained at L<the top of this section|/Character case |
1195 | changing>, that there may be more.) | |
1f607577 | 1196 | |
059703b0 | 1197 | =for apidoc Am|UV|toUPPER_utf8|U8* p|U8* e|U8* s|STRLEN* lenp |
912c8141 | 1198 | =for apidoc_item toUPPER_utf8_safe |
a239b1e2 KW |
1199 | Converts the first UTF-8 encoded character in the sequence starting at C<p> and |
1200 | extending no further than S<C<e - 1>> to its uppercase version, and | |
1f607577 KW |
1201 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note |
1202 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> | |
1203 | bytes since the uppercase version may be longer than the original character. | |
1204 | ||
1205 | The first code point of the uppercased version is returned | |
21da7284 KW |
1206 | (but note, as explained at L<the top of this section|/Character case |
1207 | changing>, that there may be more). | |
1f607577 | 1208 | |
059703b0 KW |
1209 | It will not attempt to read beyond S<C<e - 1>>, provided that the constraint |
1210 | S<C<s E<lt> e>> is true (this is asserted for in C<-DDEBUGGING> builds). If | |
1211 | the UTF-8 for the input character is malformed in some way, the program may | |
1212 | croak, or the function may return the REPLACEMENT CHARACTER, at the discretion | |
1213 | of the implementation, and subject to change in future releases. | |
a239b1e2 | 1214 | |
912c8141 | 1215 | C<toUPPER_utf8_safe> is now just a different spelling of plain C<toUPPER_utf8> |
1f607577 | 1216 | |
25200305 KW |
1217 | =for apidoc Am|U8|toFOLD|U8 ch |
1218 | Converts the specified character to foldcase. If the input is anything but an | |
1219 | ASCII uppercase character, that input character itself is returned. Variant | |
1220 | C<toFOLD_A> is equivalent. (There is no equivalent C<to_FOLD_L1> for the full | |
d0da05db | 1221 | Latin1 range, as the full generality of L</toFOLD_uvchr> is needed there.) |
25200305 | 1222 | |
d0da05db KW |
1223 | =for apidoc Am|UV|toFOLD_uvchr|UV cp|U8* s|STRLEN* lenp |
1224 | Converts the code point C<cp> to its foldcase version, and | |
1225 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code | |
1226 | point is interpreted as native if less than 256; otherwise as Unicode. Note | |
1f607577 KW |
1227 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> |
1228 | bytes since the foldcase version may be longer than the original character. | |
1229 | ||
1230 | The first code point of the foldcased version is returned | |
21da7284 KW |
1231 | (but note, as explained at L<the top of this section|/Character case |
1232 | changing>, that there may be more). | |
1f607577 | 1233 | |
059703b0 | 1234 | =for apidoc Am|UV|toFOLD_utf8|U8* p|U8* e|U8* s|STRLEN* lenp |
912c8141 | 1235 | =for apidoc_item toFOLD_utf8_safe |
a239b1e2 KW |
1236 | Converts the first UTF-8 encoded character in the sequence starting at C<p> and |
1237 | extending no further than S<C<e - 1>> to its foldcase version, and | |
1f607577 KW |
1238 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note |
1239 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> | |
1240 | bytes since the foldcase version may be longer than the original character. | |
1241 | ||
1242 | The first code point of the foldcased version is returned | |
21da7284 KW |
1243 | (but note, as explained at L<the top of this section|/Character case |
1244 | changing>, that there may be more). | |
1f607577 | 1245 | |
059703b0 | 1246 | It will not attempt |
a239b1e2 KW |
1247 | to read beyond S<C<e - 1>>, provided that the constraint S<C<s E<lt> e>> is |
1248 | true (this is asserted for in C<-DDEBUGGING> builds). If the UTF-8 for the | |
1249 | input character is malformed in some way, the program may croak, or the | |
1250 | function may return the REPLACEMENT CHARACTER, at the discretion of the | |
1251 | implementation, and subject to change in future releases. | |
1252 | ||
912c8141 | 1253 | C<toFOLD_utf8_safe> is now just a different spelling of plain C<toFOLD_utf8> |
1f607577 | 1254 | |
3cb048e5 KW |
1255 | =for apidoc Am|UV|toLOWER|UV cp |
1256 | =for apidoc_item |UV|toLOWER_A|UV cp | |
1257 | =for apidoc_item |UV|toLOWER_L1|UV cp | |
1258 | =for apidoc_item |UV|toLOWER_LATIN1|UV cp | |
1259 | =for apidoc_item |UV|toLOWER_LC|UV cp | |
1260 | =for apidoc_item |UV|toLOWER_uvchr|UV cp|U8* s|STRLEN* lenp | |
1261 | =for apidoc_item |UV|toLOWER_utf8|U8* p|U8* e|U8* s|STRLEN* lenp | |
1262 | =for apidoc_item |UV|toLOWER_utf8_safe|U8* p|U8* e|U8* s|STRLEN* lenp | |
1263 | ||
1264 | These all return the lowercase of a character. The differences are what domain | |
1265 | they operate on, and whether the input is specified as a code point (those | |
1266 | forms with a C<cp> parameter) or as a UTF-8 string (the others). In the latter | |
1267 | case, the code point to use is the first one in the buffer of UTF-8 encoded | |
1268 | code points, delineated by the arguments S<C<p .. e - 1>>. | |
1269 | ||
1270 | C<toLOWER> and C<toLOWER_A> are synonyms of each other. They return the | |
1271 | lowercase of any uppercase ASCII-range code point. All other inputs are | |
1272 | returned unchanged. Since these are macros, the input type may be any integral | |
1273 | one, and the output will occupy the same number of bits as the input. | |
1274 | ||
1275 | C<toLOWER_L1> and C<toLOWER_LATIN1> are synonyms of each other. They behave | |
1276 | identically as C<toLOWER> for ASCII-range input. But additionally will return | |
1277 | the lowercase of any uppercase code point in the entire 0..255 range, assuming | |
1278 | a Latin-1 encoding (or the EBCDIC equivalent on such platforms). | |
1279 | ||
1280 | C<toLOWER_LC> returns the lowercase of the input code point according to the | |
1281 | rules of the current POSIX locale. Input code points outside the range 0..255 | |
1282 | are returned unchanged. | |
1283 | ||
1284 | C<toLOWER_uvchr> returns the lowercase of any Unicode code point. The return | |
1285 | value is identical to that of C<toLOWER_L1> for input code points in the 0..255 | |
1286 | range. The lowercase of the vast majority of Unicode code points is the same | |
1287 | as the code point itself. For these, and for code points above the legal | |
1288 | Unicode maximum, this returns the input code point unchanged. It additionally | |
1289 | stores the UTF-8 of the result into the buffer beginning at C<s>, and its | |
1290 | length in bytes into C<*lenp>. The caller must have made C<s> large enough to | |
1291 | contain at least C<UTF8_MAXBYTES_CASE+1> bytes to avoid possible overflow. | |
1292 | ||
1293 | NOTE: the lowercase of a code point may be more than one code point. The | |
1294 | return value of this function is only the first of these. The entire lowercase | |
1295 | is returned in C<s>. To determine if the result is more than a single code | |
1296 | point, you can do something like this: | |
1297 | ||
1298 | uc = toLOWER_uvchr(cp, s, &len); | |
1299 | if (len > UTF8SKIP(s)) { is multiple code points } | |
1300 | else { is a single code point } | |
1301 | ||
1302 | C<toLOWER_utf8> and C<toLOWER_utf8_safe> are synonyms of each other. The only | |
1303 | difference between these and C<toLOWER_uvchr> is that the source for these is | |
1304 | encoded in UTF-8, instead of being a code point. It is passed as a buffer | |
1305 | starting at C<p>, with C<e> pointing to one byte beyond its end. The C<p> | |
1306 | buffer may certainly contain more than one code point; but only the first one | |
1307 | (up through S<C<e - 1>>) is examined. If the UTF-8 for the input character is | |
1308 | malformed in some way, the program may croak, or the function may return the | |
1309 | REPLACEMENT CHARACTER, at the discretion of the implementation, and subject to | |
1310 | change in future releases. | |
1f607577 | 1311 | |
25200305 KW |
1312 | =for apidoc Am|U8|toTITLE|U8 ch |
1313 | Converts the specified character to titlecase. If the input is anything but an | |
1314 | ASCII lowercase character, that input character itself is returned. Variant | |
b7d90381 | 1315 | C<toTITLE_A> is equivalent. (There is no C<toTITLE_L1> for the full Latin1 |
d0da05db | 1316 | range, as the full generality of L</toTITLE_uvchr> is needed there. Titlecase is |
b7d90381 | 1317 | not a concept used in locale handling, so there is no functionality for that.) |
25200305 | 1318 | |
d0da05db KW |
1319 | =for apidoc Am|UV|toTITLE_uvchr|UV cp|U8* s|STRLEN* lenp |
1320 | Converts the code point C<cp> to its titlecase version, and | |
1321 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. The code | |
1322 | point is interpreted as native if less than 256; otherwise as Unicode. Note | |
1f607577 KW |
1323 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> |
1324 | bytes since the titlecase version may be longer than the original character. | |
1325 | ||
1326 | The first code point of the titlecased version is returned | |
21da7284 KW |
1327 | (but note, as explained at L<the top of this section|/Character case |
1328 | changing>, that there may be more). | |
1f607577 | 1329 | |
059703b0 | 1330 | =for apidoc Am|UV|toTITLE_utf8|U8* p|U8* e|U8* s|STRLEN* lenp |
912c8141 KW |
1331 | =for apidoc_item toTITLE_utf8_safe |
1332 | Convert the first UTF-8 encoded character in the sequence starting at C<p> and | |
a239b1e2 | 1333 | extending no further than S<C<e - 1>> to its titlecase version, and |
1f607577 KW |
1334 | stores that in UTF-8 in C<s>, and its length in bytes in C<lenp>. Note |
1335 | that the buffer pointed to by C<s> needs to be at least C<UTF8_MAXBYTES_CASE+1> | |
1336 | bytes since the titlecase version may be longer than the original character. | |
1337 | ||
1338 | The first code point of the titlecased version is returned | |
21da7284 KW |
1339 | (but note, as explained at L<the top of this section|/Character case |
1340 | changing>, that there may be more). | |
1f607577 | 1341 | |
059703b0 | 1342 | It will not attempt |
a239b1e2 KW |
1343 | to read beyond S<C<e - 1>>, provided that the constraint S<C<s E<lt> e>> is |
1344 | true (this is asserted for in C<-DDEBUGGING> builds). If the UTF-8 for the | |
1345 | input character is malformed in some way, the program may croak, or the | |
1346 | function may return the REPLACEMENT CHARACTER, at the discretion of the | |
1347 | implementation, and subject to change in future releases. | |
1348 | ||
912c8141 | 1349 | C<toTITLE_utf8_safe> is now just a different spelling of plain C<toTITLE_utf8> |
1f607577 | 1350 | |
954c1994 | 1351 | =cut |
353c9b6f | 1352 | |
d0da05db | 1353 | XXX Still undocumented isVERTWS_uvchr and _utf8; it's unclear what their names |
1e222e4f KW |
1354 | really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change, |
1355 | and aren't general purpose as they don't work on U+DF, and assert against that. | |
243effed | 1356 | |
8a58bdcf | 1357 | Note that these macros are repeated in Devel::PPPort, so should also be |
62fa66b6 KW |
1358 | patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc |
1359 | ||
954c1994 GS |
1360 | */ |
1361 | ||
8f5283f4 KW |
1362 | /* |
1363 | void below because that's the best fit, and works for Devel::PPPort | |
3f620621 | 1364 | =for apidoc_section $integer |
545bca17 | 1365 | =for apidoc AyT||WIDEST_UTYPE |
8f5283f4 KW |
1366 | |
1367 | Yields the widest unsigned integer type on the platform, currently either | |
326c768d | 1368 | C<U32> or C<U64>. This can be used in declarations such as |
8f5283f4 KW |
1369 | |
1370 | WIDEST_UTYPE my_uv; | |
1371 | ||
1372 | or casts | |
1373 | ||
1374 | my_uv = (WIDEST_UTYPE) val; | |
1375 | ||
1376 | =cut | |
1377 | ||
1378 | */ | |
de9e2639 KW |
1379 | #ifdef QUADKIND |
1380 | # define WIDEST_UTYPE U64 | |
7c062697 KW |
1381 | #else |
1382 | # define WIDEST_UTYPE U32 | |
1383 | #endif | |
1384 | ||
3912bc88 KW |
1385 | /* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in |
1386 | * the lower 8. It is designed to be hopefully bomb-proof, making sure that no | |
1387 | * bits of information are lost even on a 64-bit machine, but to get the | |
1388 | * compiler to optimize it out if possible. This is because Configure makes | |
1389 | * sure that the machine has an 8-bit byte, so if c is stored in a byte, the | |
1390 | * sizeof() guarantees that this evaluates to a constant true at compile time. | |
7e75d1a1 JH |
1391 | * |
1392 | * For Coverity, be always true, because otherwise Coverity thinks | |
1393 | * it finds several expressions that are always true, independent | |
1394 | * of operands. Well, they are, but that is kind of the point. | |
220c71bf | 1395 | */ |
7e75d1a1 | 1396 | #ifndef __COVERITY__ |
6c5b02ac KW |
1397 | /* The '| 0' part ensures a compiler error if c is not integer (like e.g., a |
1398 | * pointer) */ | |
1399 | #define FITS_IN_8_BITS(c) ( (sizeof(c) == 1) \ | |
ace3ad0f | 1400 | || !(((WIDEST_UTYPE)((c) | 0)) & ~0xFF)) |
7e75d1a1 JH |
1401 | #else |
1402 | #define FITS_IN_8_BITS(c) (1) | |
1403 | #endif | |
cf301eb7 | 1404 | |
45f4bb73 | 1405 | /* Returns true if l <= c <= (l + n), where 'l' and 'n' are non-negative |
833b0f46 | 1406 | * Written this way so that after optimization, only one conditional test is |
76d3ad4c KW |
1407 | * needed. (The NV casts stop any warnings about comparison always being true |
1408 | * if called with an unsigned. The cast preserves the sign, which is all we | |
1409 | * care about.) */ | |
1410 | #define withinCOUNT(c, l, n) (__ASSERT_((NV) (l) >= 0) \ | |
1411 | __ASSERT_((NV) (n) >= 0) \ | |
1412 | (((WIDEST_UTYPE) (((c)) - ((l) | 0))) <= (((WIDEST_UTYPE) ((n) | 0))))) | |
833b0f46 | 1413 | |
94250c4f KW |
1414 | /* Returns true if c is in the range l..u, where 'l' is non-negative |
1415 | * Written this way so that after optimization, only one conditional test is | |
4758c20d | 1416 | * needed. */ |
1eaefa6e | 1417 | #define inRANGE(c, l, u) (__ASSERT_((u) >= (l)) \ |
4758c20d | 1418 | ( (sizeof(c) == sizeof(U8)) ? withinCOUNT(((U8) (c)), (l), ((u) - (l))) \ |
4758c20d KW |
1419 | : (sizeof(c) == sizeof(U32)) ? withinCOUNT(((U32) (c)), (l), ((u) - (l))) \ |
1420 | : (__ASSERT_(sizeof(c) == sizeof(WIDEST_UTYPE)) \ | |
45f4bb73 | 1421 | withinCOUNT(((WIDEST_UTYPE) (c)), (l), ((u) - (l)))))) |
305fe86e | 1422 | |
41f43cc2 | 1423 | #ifdef EBCDIC |
b6340bd0 | 1424 | # ifndef _ALL_SOURCE |
0852beac KW |
1425 | /* The native libc isascii() et.al. functions return the wrong results |
1426 | * on at least z/OS unless this is defined. */ | |
b6340bd0 KW |
1427 | # error _ALL_SOURCE should probably be defined |
1428 | # endif | |
41f43cc2 | 1429 | #else |
0852beac KW |
1430 | /* There is a simple definition of ASCII for ASCII platforms. But the |
1431 | * EBCDIC one isn't so simple, so is defined using table look-up like the | |
9c903d59 | 1432 | * other macros below. |
3f3c579d KW |
1433 | * |
1434 | * The cast here is used instead of '(c) >= 0', because some compilers emit | |
1435 | * a warning that that test is always true when the parameter is an | |
1436 | * unsigned type. khw supposes that it could be written as | |
1437 | * && ((c) == '\0' || (c) > 0) | |
1438 | * to avoid the message, but the cast will likely avoid extra branches even | |
1439 | * with stupid compilers. | |
1440 | * | |
1441 | * The '| 0' part ensures a compiler error if c is not integer (like e.g., | |
1442 | * a pointer) */ | |
9c903d59 | 1443 | # define isASCII(c) ((WIDEST_UTYPE)((c) | 0) < 128) |
41f43cc2 KW |
1444 | #endif |
1445 | ||
38694112 KW |
1446 | /* Take the eight possible bit patterns of the lower 3 bits and you get the |
1447 | * lower 3 bits of the 8 octal digits, in both ASCII and EBCDIC, so those bits | |
1448 | * can be ignored. If the rest match '0', we have an octal */ | |
1449 | #define isOCTAL_A(c) (((WIDEST_UTYPE)((c) | 0) & ~7) == '0') | |
c2da0b36 | 1450 | |
9fb1bf9d | 1451 | #ifdef H_PERL /* If have access to perl.h, lookup in its table */ |
f4cdb42c | 1452 | |
a500dc72 KW |
1453 | /* Character class numbers. For internal core Perl use only. The ones less |
1454 | * than 32 are used in PL_charclass[] and the ones up through the one that | |
1455 | * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and | |
1456 | * related files. PL_charclass ones use names used in l1_char_class_tab.h but | |
1457 | * their actual definitions are here. If that file has a name not used here, | |
1458 | * it won't compile. | |
1709d539 KW |
1459 | * |
1460 | * The first group of these is ordered in what I (khw) estimate to be the | |
31c7f561 | 1461 | * frequency of their use. This gives a slight edge to exiting a loop earlier |
58a3ba2c KW |
1462 | * (in reginclass() in regexec.c). Except \v should be last, as it isn't a |
1463 | * real Posix character class, and some (small) inefficiencies in regular | |
1464 | * expression handling would be introduced by putting it in the middle of those | |
1465 | * that are. Also, cntrl and ascii come after the others as it may be useful | |
1466 | * to group these which have no members that match above Latin1, (or above | |
1467 | * ASCII in the latter case) */ | |
1468 | ||
1709d539 KW |
1469 | # define _CC_WORDCHAR 0 /* \w and [:word:] */ |
1470 | # define _CC_DIGIT 1 /* \d and [:digit:] */ | |
1471 | # define _CC_ALPHA 2 /* [:alpha:] */ | |
1472 | # define _CC_LOWER 3 /* [:lower:] */ | |
1473 | # define _CC_UPPER 4 /* [:upper:] */ | |
1474 | # define _CC_PUNCT 5 /* [:punct:] */ | |
1475 | # define _CC_PRINT 6 /* [:print:] */ | |
15861f94 | 1476 | # define _CC_ALPHANUMERIC 7 /* [:alnum:] */ |
1709d539 | 1477 | # define _CC_GRAPH 8 /* [:graph:] */ |
359b005e | 1478 | # define _CC_CASED 9 /* [:lower:] or [:upper:] under /i */ |
779cf272 | 1479 | # define _CC_SPACE 10 /* \s, [:space:] */ |
b0d691b2 KW |
1480 | # define _CC_BLANK 11 /* [:blank:] */ |
1481 | # define _CC_XDIGIT 12 /* [:xdigit:] */ | |
779cf272 KW |
1482 | # define _CC_CNTRL 13 /* [:cntrl:] */ |
1483 | # define _CC_ASCII 14 /* [:ascii:] */ | |
1484 | # define _CC_VERTSPACE 15 /* \v */ | |
1709d539 | 1485 | |
a0947d7b KW |
1486 | # define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE |
1487 | ||
1709d539 | 1488 | /* The members of the third group below do not need to be coordinated with data |
3ffc8c70 | 1489 | * structures in regcomp.[ch] and regexec.c. */ |
779cf272 KW |
1490 | # define _CC_IDFIRST 16 |
1491 | # define _CC_CHARNAME_CONT 17 | |
1492 | # define _CC_NONLATIN1_FOLD 18 | |
1493 | # define _CC_NONLATIN1_SIMPLE_FOLD 19 | |
1494 | # define _CC_QUOTEMETA 20 | |
1495 | # define _CC_NON_FINAL_FOLD 21 | |
1496 | # define _CC_IS_IN_SOME_FOLD 22 | |
2ae9030c KW |
1497 | # define _CC_BINDIGIT 23 |
1498 | # define _CC_OCTDIGIT 24 | |
1499 | # define _CC_MNEMONIC_CNTRL 25 | |
073c22b3 KW |
1500 | |
1501 | /* This next group is only used on EBCDIC platforms, so theoretically could be | |
1502 | * shared with something entirely different that's only on ASCII platforms */ | |
abb8abf6 | 1503 | # define _CC_UTF8_START_BYTE_IS_FOR_AT_LEAST_SURROGATE 31 |
0654f0ab | 1504 | /* Unused: 26-30 |
f4cdb42c KW |
1505 | * If more bits are needed, one could add a second word for non-64bit |
1506 | * QUAD_IS_INT systems, using some #ifdefs to distinguish between having a 2nd | |
37ede926 KW |
1507 | * word or not. The IS_IN_SOME_FOLD bit is the most easily expendable, as it |
1508 | * is used only for optimization (as of this writing), and differs in the | |
1509 | * Latin1 range from the ALPHA bit only in two relatively unimportant | |
a500dc72 | 1510 | * characters: the masculine and feminine ordinal indicators, so removing it |
073c22b3 KW |
1511 | * would just cause /i regexes which match them to run less efficiently. |
1512 | * Similarly the EBCDIC-only bits are used just for speed, and could be | |
1513 | * replaced by other means */ | |
96ac0975 | 1514 | |
3a371f2f KW |
1515 | #if defined(PERL_CORE) || defined(PERL_EXT) |
1516 | /* An enum version of the character class numbers, to help compilers | |
1517 | * optimize */ | |
1518 | typedef enum { | |
3a371f2f | 1519 | _CC_ENUM_ALPHA = _CC_ALPHA, |
e8d596e0 KW |
1520 | _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC, |
1521 | _CC_ENUM_ASCII = _CC_ASCII, | |
1522 | _CC_ENUM_BLANK = _CC_BLANK, | |
b0d691b2 | 1523 | _CC_ENUM_CASED = _CC_CASED, |
e8d596e0 | 1524 | _CC_ENUM_CNTRL = _CC_CNTRL, |
3a371f2f KW |
1525 | _CC_ENUM_DIGIT = _CC_DIGIT, |
1526 | _CC_ENUM_GRAPH = _CC_GRAPH, | |
1527 | _CC_ENUM_LOWER = _CC_LOWER, | |
1528 | _CC_ENUM_PRINT = _CC_PRINT, | |
1529 | _CC_ENUM_PUNCT = _CC_PUNCT, | |
e8d596e0 | 1530 | _CC_ENUM_SPACE = _CC_SPACE, |
3a371f2f | 1531 | _CC_ENUM_UPPER = _CC_UPPER, |
e8d596e0 | 1532 | _CC_ENUM_VERTSPACE = _CC_VERTSPACE, |
3a371f2f | 1533 | _CC_ENUM_WORDCHAR = _CC_WORDCHAR, |
e8d596e0 | 1534 | _CC_ENUM_XDIGIT = _CC_XDIGIT |
3a371f2f KW |
1535 | } _char_class_number; |
1536 | #endif | |
1537 | ||
86f72d56 | 1538 | #define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1) |
63c61c3f | 1539 | |
6635f04f | 1540 | START_EXTERN_C |
96ac0975 NC |
1541 | # ifdef DOINIT |
1542 | EXTCONST U32 PL_charclass[] = { | |
1543 | # include "l1_char_class_tab.h" | |
1544 | }; | |
1545 | ||
1546 | # else /* ! DOINIT */ | |
1547 | EXTCONST U32 PL_charclass[]; | |
1548 | # endif | |
6635f04f | 1549 | END_EXTERN_C |
96ac0975 | 1550 | |
265c1f46 | 1551 | /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */ |
430b7c70 | 1552 | # define _CC_mask(classnum) (1U << (classnum)) |
4650c663 KW |
1553 | |
1554 | /* For internal core Perl use only: the base macro for defining macros like | |
1555 | * isALPHA */ | |
ff7ecfc3 | 1556 | # define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \ |
f4cd282c | 1557 | && (PL_charclass[(U8) (c)] & _CC_mask(classnum))) |
4eeeb416 | 1558 | |
f4cdb42c KW |
1559 | /* The mask for the _A versions of the macros; it just adds in the bit for |
1560 | * ASCII. */ | |
1561 | # define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII)) | |
1562 | ||
4650c663 KW |
1563 | /* For internal core Perl use only: the base macro for defining macros like |
1564 | * isALPHA_A. The foo_A version makes sure that both the desired bit and | |
1565 | * the ASCII bit are present */ | |
b7d90381 KW |
1566 | # define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \ |
1567 | && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \ | |
1568 | == _CC_mask_A(classnum))) | |
f4cdb42c | 1569 | |
26c1d9d8 KW |
1570 | /* On ASCII platforms certain classes form a single range. It's faster to |
1571 | * special case these. isDIGIT is a single range on all platforms */ | |
b877c1ff KW |
1572 | # ifdef EBCDIC |
1573 | # define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA) | |
1574 | # define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH) | |
1575 | # define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER) | |
1576 | # define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT) | |
1577 | # define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER) | |
1578 | # else | |
26c1d9d8 | 1579 | /* By folding the upper and lowercase, we can use a single range */ |
b877c1ff | 1580 | # define isALPHA_A(c) inRANGE((~('A' ^ 'a') & (c)), 'A', 'Z') |
26c1d9d8 | 1581 | # define isGRAPH_A(c) inRANGE(c, ' ' + 1, 0x7e) |
b877c1ff KW |
1582 | # define isLOWER_A(c) inRANGE(c, 'a', 'z') |
1583 | # define isPRINT_A(c) inRANGE(c, ' ', 0x7e) | |
1584 | # define isUPPER_A(c) inRANGE(c, 'A', 'Z') | |
1585 | # endif | |
15861f94 | 1586 | # define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC) |
f4cdb42c KW |
1587 | # define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK) |
1588 | # define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL) | |
b877c1ff | 1589 | # define isDIGIT_A(c) inRANGE(c, '0', '9') |
f4cdb42c KW |
1590 | # define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT) |
1591 | # define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE) | |
f4cdb42c | 1592 | # define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR) |
b7d90381 KW |
1593 | # define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT) /* No non-ASCII xdigits |
1594 | */ | |
d95f8b6a | 1595 | # define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST) |
3ded5eb0 KW |
1596 | # define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA) |
1597 | # define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC) | |
1598 | # define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK) | |
1599 | ||
1600 | /* continuation character for legal NAME in \N{NAME} */ | |
1601 | # define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT) | |
1602 | ||
1603 | # define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL) | |
1604 | # define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH) | |
1605 | # define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER) | |
1606 | # define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT) | |
b7d90381 | 1607 | # define isPSXSPC_L1(c) isSPACE_L1(c) |
3ded5eb0 KW |
1608 | # define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT) |
1609 | # define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE) | |
1610 | # define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER) | |
1611 | # define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR) | |
1612 | # define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST) | |
f4cdb42c | 1613 | |
0852beac KW |
1614 | # ifdef EBCDIC |
1615 | # define isASCII(c) _generic_isCC(c, _CC_ASCII) | |
1616 | # endif | |
1617 | ||
f12c0118 | 1618 | /* Participates in a single-character fold with a character above 255 */ |
c62fdeb7 KW |
1619 | # if defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C) |
1620 | # define HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE(c) \ | |
1621 | (( ! cBOOL(FITS_IN_8_BITS(c))) \ | |
1622 | || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_SIMPLE_FOLD))) | |
1623 | ||
1624 | # define IS_NON_FINAL_FOLD(c) _generic_isCC(c, _CC_NON_FINAL_FOLD) | |
1625 | # define IS_IN_SOME_FOLD_L1(c) _generic_isCC(c, _CC_IS_IN_SOME_FOLD) | |
1626 | # endif | |
f12c0118 KW |
1627 | |
1628 | /* Like the above, but also can be part of a multi-char fold */ | |
c62fdeb7 KW |
1629 | # define HAS_NONLATIN1_FOLD_CLOSURE(c) \ |
1630 | ( (! cBOOL(FITS_IN_8_BITS(c))) \ | |
1631 | || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_FOLD))) | |
430b7c70 | 1632 | |
4eeeb416 | 1633 | # define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA) |
5e6ebb12 KW |
1634 | |
1635 | /* is c a control character for which we have a mnemonic? */ | |
1636 | # if defined(PERL_CORE) || defined(PERL_EXT) | |
1637 | # define isMNEMONIC_CNTRL(c) _generic_isCC(c, _CC_MNEMONIC_CNTRL) | |
1638 | # endif | |
687c8d01 | 1639 | #else /* else we don't have perl.h H_PERL */ |
3ded5eb0 KW |
1640 | |
1641 | /* If we don't have perl.h, we are compiling a utility program. Below we | |
1642 | * hard-code various macro definitions that wouldn't otherwise be available | |
fc273927 | 1643 | * to it. Most are coded based on first principles. These are written to |
74665a89 | 1644 | * avoid EBCDIC vs. ASCII #ifdef's as much as possible. */ |
182c4ace | 1645 | # define isDIGIT_A(c) inRANGE(c, '0', '9') |
0852beac | 1646 | # define isBLANK_A(c) ((c) == ' ' || (c) == '\t') |
74665a89 KW |
1647 | # define isSPACE_A(c) (isBLANK_A(c) \ |
1648 | || (c) == '\n' \ | |
1649 | || (c) == '\r' \ | |
1650 | || (c) == '\v' \ | |
0852beac | 1651 | || (c) == '\f') |
74665a89 KW |
1652 | /* On EBCDIC, there are gaps between 'i' and 'j'; 'r' and 's'. Same for |
1653 | * uppercase. The tests for those aren't necessary on ASCII, but hurt only | |
1654 | * performance (if optimization isn't on), and allow the same code to be | |
1655 | * used for both platform types */ | |
182c4ace KW |
1656 | # define isLOWER_A(c) inRANGE((c), 'a', 'i') \ |
1657 | || inRANGE((c), 'j', 'r') \ | |
1658 | || inRANGE((c), 's', 'z') | |
1659 | # define isUPPER_A(c) inRANGE((c), 'A', 'I') \ | |
1660 | || inRANGE((c), 'J', 'R') \ | |
1661 | || inRANGE((c), 'S', 'Z') | |
a4d7a999 KW |
1662 | # define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c)) |
1663 | # define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c)) | |
3ded5eb0 | 1664 | # define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_') |
0852beac | 1665 | # define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_') |
182c4ace KW |
1666 | # define isXDIGIT_A(c) ( isDIGIT_A(c) \ |
1667 | || inRANGE((c), 'a', 'f') \ | |
1668 | || inRANGE((c), 'A', 'F') | |
74665a89 KW |
1669 | # define isPUNCT_A(c) ((c) == '-' || (c) == '!' || (c) == '"' \ |
1670 | || (c) == '#' || (c) == '$' || (c) == '%' \ | |
1671 | || (c) == '&' || (c) == '\'' || (c) == '(' \ | |
1672 | || (c) == ')' || (c) == '*' || (c) == '+' \ | |
1673 | || (c) == ',' || (c) == '.' || (c) == '/' \ | |
1674 | || (c) == ':' || (c) == ';' || (c) == '<' \ | |
1675 | || (c) == '=' || (c) == '>' || (c) == '?' \ | |
1676 | || (c) == '@' || (c) == '[' || (c) == '\\' \ | |
1677 | || (c) == ']' || (c) == '^' || (c) == '_' \ | |
1678 | || (c) == '`' || (c) == '{' || (c) == '|' \ | |
1679 | || (c) == '}' || (c) == '~') | |
1680 | # define isGRAPH_A(c) (isALPHANUMERIC_A(c) || isPUNCT_A(c)) | |
1681 | # define isPRINT_A(c) (isGRAPH_A(c) || (c) == ' ') | |
3ded5eb0 | 1682 | |
0852beac | 1683 | # ifdef EBCDIC |
74665a89 KW |
1684 | /* The below is accurate for the 3 EBCDIC code pages traditionally |
1685 | * supported by perl. The only difference between them in the controls | |
1686 | * is the position of \n, and that is represented symbolically below */ | |
1687 | # define isCNTRL_A(c) ((c) == '\0' || (c) == '\a' || (c) == '\b' \ | |
1688 | || (c) == '\f' || (c) == '\n' || (c) == '\r' \ | |
1689 | || (c) == '\t' || (c) == '\v' \ | |
182c4ace | 1690 | || inRANGE((c), 1, 3) /* SOH, STX, ETX */ \ |
8ec0a736 | 1691 | || (c) == 7F /* U+7F DEL */ \ |
182c4ace KW |
1692 | || inRANGE((c), 0x0E, 0x13) /* SO SI DLE \ |
1693 | DC[1-3] */ \ | |
74665a89 KW |
1694 | || (c) == 0x18 /* U+18 CAN */ \ |
1695 | || (c) == 0x19 /* U+19 EOM */ \ | |
182c4ace | 1696 | || inRANGE((c), 0x1C, 0x1F) /* [FGRU]S */ \ |
74665a89 KW |
1697 | || (c) == 0x26 /* U+17 ETB */ \ |
1698 | || (c) == 0x27 /* U+1B ESC */ \ | |
1699 | || (c) == 0x2D /* U+05 ENQ */ \ | |
1700 | || (c) == 0x2E /* U+06 ACK */ \ | |
1701 | || (c) == 0x32 /* U+16 SYN */ \ | |
1702 | || (c) == 0x37 /* U+04 EOT */ \ | |
1703 | || (c) == 0x3C /* U+14 DC4 */ \ | |
1704 | || (c) == 0x3D /* U+15 NAK */ \ | |
1705 | || (c) == 0x3F)/* U+1A SUB */ | |
0852beac | 1706 | # define isASCII(c) (isCNTRL_A(c) || isPRINT_A(c)) |
74665a89 KW |
1707 | # else /* isASCII is already defined for ASCII platforms, so can use that to |
1708 | define isCNTRL */ | |
1709 | # define isCNTRL_A(c) (isASCII(c) && ! isPRINT_A(c)) | |
0852beac KW |
1710 | # endif |
1711 | ||
3ffc8c70 | 1712 | /* The _L1 macros may be unnecessary for the utilities; I (khw) added them |
caa94d35 KW |
1713 | * during debugging, and it seems best to keep them. We may be called |
1714 | * without NATIVE_TO_LATIN1 being defined. On ASCII platforms, it doesn't | |
1715 | * do anything anyway, so make it not a problem */ | |
1716 | # if ! defined(EBCDIC) && ! defined(NATIVE_TO_LATIN1) | |
1717 | # define NATIVE_TO_LATIN1(ch) (ch) | |
1718 | # endif | |
3ded5eb0 KW |
1719 | # define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c)) |
1720 | # define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c)) | |
1721 | # define isBLANK_L1(c) (isBLANK_A(c) \ | |
1722 | || (FITS_IN_8_BITS(c) \ | |
1723 | && NATIVE_TO_LATIN1((U8) c) == 0xA0)) | |
1724 | # define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c))) | |
1725 | # define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c))) | |
1726 | # define isLOWER_L1(c) (isLOWER_A(c) \ | |
1727 | || (FITS_IN_8_BITS(c) \ | |
ae683a5f | 1728 | && (( NATIVE_TO_LATIN1((U8) c) >= 0xDF \ |
3ded5eb0 KW |
1729 | && NATIVE_TO_LATIN1((U8) c) != 0xF7) \ |
1730 | || NATIVE_TO_LATIN1((U8) c) == 0xAA \ | |
1731 | || NATIVE_TO_LATIN1((U8) c) == 0xBA \ | |
1732 | || NATIVE_TO_LATIN1((U8) c) == 0xB5))) | |
1733 | # define isPRINT_L1(c) (isPRINT_A(c) \ | |
1734 | || (FITS_IN_8_BITS(c) \ | |
1735 | && NATIVE_TO_LATIN1((U8) c) >= 0xA0)) | |
3ded5eb0 KW |
1736 | # define isPUNCT_L1(c) (isPUNCT_A(c) \ |
1737 | || (FITS_IN_8_BITS(c) \ | |
ae683a5f | 1738 | && ( NATIVE_TO_LATIN1((U8) c) == 0xA1 \ |
3ded5eb0 KW |
1739 | || NATIVE_TO_LATIN1((U8) c) == 0xA7 \ |
1740 | || NATIVE_TO_LATIN1((U8) c) == 0xAB \ | |
1741 | || NATIVE_TO_LATIN1((U8) c) == 0xB6 \ | |
1742 | || NATIVE_TO_LATIN1((U8) c) == 0xB7 \ | |
1743 | || NATIVE_TO_LATIN1((U8) c) == 0xBB \ | |
1744 | || NATIVE_TO_LATIN1((U8) c) == 0xBF))) | |
1745 | # define isSPACE_L1(c) (isSPACE_A(c) \ | |
1746 | || (FITS_IN_8_BITS(c) \ | |
ae683a5f | 1747 | && ( NATIVE_TO_LATIN1((U8) c) == 0x85 \ |
3ded5eb0 KW |
1748 | || NATIVE_TO_LATIN1((U8) c) == 0xA0))) |
1749 | # define isUPPER_L1(c) (isUPPER_A(c) \ | |
1750 | || (FITS_IN_8_BITS(c) \ | |
182c4ace KW |
1751 | && ( IN_RANGE(NATIVE_TO_LATIN1((U8) c), \ |
1752 | 0xC0, 0xDE) \ | |
3ded5eb0 KW |
1753 | && NATIVE_TO_LATIN1((U8) c) != 0xD7))) |
1754 | # define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c)) | |
1755 | # define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_') | |
1756 | # define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \ | |
1757 | || isBLANK_L1(c) \ | |
1758 | || (c) == '-' \ | |
1759 | || (c) == '(' \ | |
1760 | || (c) == ')') | |
1761 | /* The following are not fully accurate in the above-ASCII range. I (khw) | |
1762 | * don't think it's necessary to be so for the purposes where this gets | |
1763 | * compiled */ | |
1764 | # define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c)) | |
1765 | # define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c) | |
1766 | ||
1767 | /* And these aren't accurate at all. They are useful only for above | |
1768 | * Latin1, which utilities and bootstrapping don't deal with */ | |
1769 | # define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0 | |
6838b41e | 1770 | # define _HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0 |
3ded5eb0 KW |
1771 | # define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0 |
1772 | ||
1773 | /* Many of the macros later in this file are defined in terms of these. By | |
1774 | * implementing them with a function, which converts the class number into | |
1775 | * a call to the desired macro, all of the later ones work. However, that | |
1776 | * function won't be actually defined when building a utility program (no | |
1777 | * perl.h), and so a compiler error will be generated if one is attempted | |
1778 | * to be used. And the above-Latin1 code points require Unicode tables to | |
1779 | * be present, something unlikely to be the case when bootstrapping */ | |
1780 | # define _generic_isCC(c, classnum) \ | |
1781 | (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE)) | |
1782 | # define _generic_isCC_A(c, classnum) \ | |
1783 | (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE)) | |
687c8d01 | 1784 | #endif /* End of no perl.h H_PERL */ |
8a58bdcf | 1785 | |
e66b99e9 KW |
1786 | #define isALPHANUMERIC(c) isALPHANUMERIC_A(c) |
1787 | #define isALPHA(c) isALPHA_A(c) | |
0852beac KW |
1788 | #define isASCII_A(c) isASCII(c) |
1789 | #define isASCII_L1(c) isASCII(c) | |
e66b99e9 KW |
1790 | #define isBLANK(c) isBLANK_A(c) |
1791 | #define isCNTRL(c) isCNTRL_A(c) | |
1792 | #define isDIGIT(c) isDIGIT_A(c) | |
1793 | #define isGRAPH(c) isGRAPH_A(c) | |
1794 | #define isIDFIRST(c) isIDFIRST_A(c) | |
1795 | #define isLOWER(c) isLOWER_A(c) | |
1796 | #define isPRINT(c) isPRINT_A(c) | |
779cf272 | 1797 | #define isPSXSPC_A(c) isSPACE_A(c) |
e66b99e9 | 1798 | #define isPSXSPC(c) isPSXSPC_A(c) |
779cf272 | 1799 | #define isPSXSPC_L1(c) isSPACE_L1(c) |
e66b99e9 KW |
1800 | #define isPUNCT(c) isPUNCT_A(c) |
1801 | #define isSPACE(c) isSPACE_A(c) | |
1802 | #define isUPPER(c) isUPPER_A(c) | |
1803 | #define isWORDCHAR(c) isWORDCHAR_A(c) | |
1804 | #define isXDIGIT(c) isXDIGIT_A(c) | |
1805 | ||
1806 | /* ASCII casing. These could also be written as | |
1807 | #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c)) | |
1808 | #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c)) | |
1809 | which uses table lookup and mask instead of subtraction. (This would | |
c5e9991e KW |
1810 | work because the _MOD does not apply in the ASCII range). |
1811 | ||
1812 | These actually are UTF-8 invariant casing, not just ASCII, as any non-ASCII | |
1813 | UTF-8 invariants are neither upper nor lower. (Only on EBCDIC platforms are | |
1814 | there non-ASCII invariants, and all of them are controls.) */ | |
68067e4e DM |
1815 | #define toLOWER(c) (isUPPER(c) ? (U8)((c) + ('a' - 'A')) : (c)) |
1816 | #define toUPPER(c) (isLOWER(c) ? (U8)((c) - ('a' - 'A')) : (c)) | |
bbce6d69 | 1817 | |
25200305 KW |
1818 | /* In the ASCII range, these are equivalent to what they're here defined to be. |
1819 | * But by creating these definitions, other code doesn't have to be aware of | |
c5e9991e KW |
1820 | * this detail. Actually this works for all UTF-8 invariants, not just the |
1821 | * ASCII range. (EBCDIC platforms can have non-ASCII invariants.) */ | |
25200305 | 1822 | #define toFOLD(c) toLOWER(c) |
25200305 KW |
1823 | #define toTITLE(c) toUPPER(c) |
1824 | ||
c753c8d3 KW |
1825 | #define toLOWER_A(c) toLOWER(c) |
1826 | #define toUPPER_A(c) toUPPER(c) | |
25200305 KW |
1827 | #define toFOLD_A(c) toFOLD(c) |
1828 | #define toTITLE_A(c) toTITLE(c) | |
1a0901db | 1829 | |
4650c663 | 1830 | /* Use table lookup for speed; returns the input itself if is out-of-range */ |
b2bf251f | 1831 | #define toLOWER_LATIN1(c) ((! FITS_IN_8_BITS(c)) \ |
8e7c6e7d | 1832 | ? (c) \ |
f4cd282c | 1833 | : PL_latin1_lc[ (U8) (c) ]) |
c753c8d3 KW |
1834 | #define toLOWER_L1(c) toLOWER_LATIN1(c) /* Synonym for consistency */ |
1835 | ||
1a0901db | 1836 | /* Modified uc. Is correct uc except for three non-ascii chars which are |
4650c663 KW |
1837 | * all mapped to one of them, and these need special handling; returns the |
1838 | * input itself if is out-of-range */ | |
b2bf251f | 1839 | #define toUPPER_LATIN1_MOD(c) ((! FITS_IN_8_BITS(c)) \ |
8e7c6e7d | 1840 | ? (c) \ |
f4cd282c | 1841 | : PL_mod_latin1_uc[ (U8) (c) ]) |
31f05a37 | 1842 | #define IN_UTF8_CTYPE_LOCALE PL_in_utf8_CTYPE_locale |
84061b6a | 1843 | |
beab9ebe KW |
1844 | /* Use foo_LC_uvchr() instead of these for beyond the Latin1 range */ |
1845 | ||
1846 | /* For internal core Perl use only: the base macro for defining macros like | |
1847 | * isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point | |
31f05a37 KW |
1848 | * (0-255) to check. In a UTF-8 locale, the result is the same as calling |
1849 | * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like | |
1850 | * _CC_UPPER, which gives the class number for doing this. For non-UTF-8 | |
1851 | * locales, the code to actually do the test this is passed in 'non_utf8'. If | |
1852 | * 'c' is above 255, 0 is returned. For accessing the full range of possible | |
1853 | * code points under locale rules, use the macros based on _generic_LC_uvchr | |
1854 | * instead of this. */ | |
beab9ebe KW |
1855 | #define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \ |
1856 | (! FITS_IN_8_BITS(c) \ | |
1857 | ? 0 \ | |
31f05a37 KW |
1858 | : IN_UTF8_CTYPE_LOCALE \ |
1859 | ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \ | |
beab9ebe KW |
1860 | : cBOOL(non_utf8)) |
1861 | ||
1862 | /* For internal core Perl use only: a helper macro for defining macros like | |
1863 | * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to | |
1864 | * actually do this test is passed in 'non_utf8_func', which is called on 'c', | |
1865 | * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See | |
1866 | * _generic_LC_base for more info */ | |
1867 | #define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \ | |
1868 | _generic_LC_base(c,utf8_locale_classnum, \ | |
1869 | non_utf8_func( (_LC_CAST) (c))) | |
1870 | ||
1871 | /* For internal core Perl use only: like _generic_LC, but also returns TRUE if | |
1872 | * 'c' is the platform's native underscore character */ | |
1873 | #define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \ | |
1874 | _generic_LC_base(c, utf8_locale_classnum, \ | |
1875 | (non_utf8_func( (_LC_CAST) (c)) \ | |
1876 | || (char)(c) == '_')) | |
1877 | ||
1878 | /* These next three are also for internal core Perl use only: case-change | |
247985d4 KW |
1879 | * helper macros. The reason for using the PL_latin arrays is in case the |
1880 | * system function is defective; it ensures uniform results that conform to the | |
b257a28c | 1881 | * Unicod standard. It does not handle the anomalies in UTF-8 Turkic locales */ |
beab9ebe KW |
1882 | #define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \ |
1883 | ? (c) \ | |
31f05a37 KW |
1884 | : (IN_UTF8_CTYPE_LOCALE) \ |
1885 | ? PL_latin1_lc[ (U8) (c) ] \ | |
5a10328c | 1886 | : (cast)function((cast)(c))) |
beab9ebe | 1887 | |
31f05a37 KW |
1888 | /* Note that the result can be larger than a byte in a UTF-8 locale. It |
1889 | * returns a single value, so can't adequately return the upper case of LATIN | |
1890 | * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two | |
1891 | * values "SS"); instead it asserts against that under DEBUGGING, and | |
b257a28c KW |
1892 | * otherwise returns its input. It does not handle the anomalies in UTF-8 |
1893 | * Turkic locales. */ | |
beab9ebe KW |
1894 | #define _generic_toUPPER_LC(c, function, cast) \ |
1895 | (! FITS_IN_8_BITS(c) \ | |
1896 | ? (c) \ | |
31f05a37 | 1897 | : ((! IN_UTF8_CTYPE_LOCALE) \ |
b7d90381 | 1898 | ? (cast)function((cast)(c)) \ |
31f05a37 KW |
1899 | : ((((U8)(c)) == MICRO_SIGN) \ |
1900 | ? GREEK_CAPITAL_LETTER_MU \ | |
1901 | : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \ | |
1902 | ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \ | |
1903 | : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \ | |
1904 | ? (__ASSERT_(0) (c)) \ | |
1905 | : PL_mod_latin1_uc[ (U8) (c) ]))))) | |
1906 | ||
1907 | /* Note that the result can be larger than a byte in a UTF-8 locale. It | |
1908 | * returns a single value, so can't adequately return the fold case of LATIN | |
1909 | * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two | |
1910 | * values "ss"); instead it asserts against that under DEBUGGING, and | |
b257a28c KW |
1911 | * otherwise returns its input. It does not handle the anomalies in UTF-8 |
1912 | * Turkic locales */ | |
beab9ebe | 1913 | #define _generic_toFOLD_LC(c, function, cast) \ |
4d7db1b9 | 1914 | ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \ |
31f05a37 | 1915 | ? GREEK_SMALL_LETTER_MU \ |
4d7db1b9 KW |
1916 | : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \ |
1917 | || (c) != LATIN_SMALL_LETTER_SHARP_S) \ | |
1918 | _generic_toLOWER_LC(c, function, cast))) | |
beab9ebe | 1919 | |
84061b6a | 1920 | /* Use the libc versions for these if available. */ |
f05550c0 | 1921 | #if defined(HAS_ISASCII) |
84061b6a KW |
1922 | # define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c))) |
1923 | #else | |
1924 | # define isASCII_LC(c) isASCII(c) | |
1925 | #endif | |
1926 | ||
f05550c0 | 1927 | #if defined(HAS_ISBLANK) |
84061b6a | 1928 | # define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank) |
95f34b6f | 1929 | #else /* Unlike isASCII, varies if in a UTF-8 locale */ |
278b2b56 | 1930 | # define isBLANK_LC(c) ((IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c)) |
84061b6a KW |
1931 | #endif |
1932 | ||
f05550c0 | 1933 | #define _LC_CAST U8 |
bbce6d69 | 1934 | |
f05550c0 | 1935 | #ifdef WIN32 |
375f5f06 KW |
1936 | /* The Windows functions don't bother to follow the POSIX standard, which |
1937 | * for example says that something can't both be a printable and a control. | |
1938 | * But Windows treats the \t control as a printable, and does such things | |
1939 | * as making superscripts into both digits and punctuation. This tames | |
1940 | * these flaws by assuming that the definitions of both controls and space | |
1941 | * are correct, and then making sure that other definitions don't have | |
1942 | * weirdnesses, by making sure that isalnum() isn't also ispunct(), etc. | |
1943 | * Not all possible weirdnesses are checked for, just the ones that were | |
1944 | * detected on actual Microsoft code pages */ | |
1945 | ||
b7d90381 KW |
1946 | # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl) |
1947 | # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace) | |
1948 | ||
1949 | # define isALPHA_LC(c) (_generic_LC(c, _CC_ALPHA, isalpha) \ | |
1950 | && isALPHANUMERIC_LC(c)) | |
1951 | # define isALPHANUMERIC_LC(c) (_generic_LC(c, _CC_ALPHANUMERIC, isalnum) && \ | |
1952 | ! isPUNCT_LC(c)) | |
1953 | # define isDIGIT_LC(c) (_generic_LC(c, _CC_DIGIT, isdigit) && \ | |
1954 | isALPHANUMERIC_LC(c)) | |
1955 | # define isGRAPH_LC(c) (_generic_LC(c, _CC_GRAPH, isgraph) && isPRINT_LC(c)) | |
1956 | # define isIDFIRST_LC(c) (((c) == '_') \ | |
1957 | || (_generic_LC(c, _CC_IDFIRST, isalpha) && ! isPUNCT_LC(c))) | |
1958 | # define isLOWER_LC(c) (_generic_LC(c, _CC_LOWER, islower) && isALPHA_LC(c)) | |
1959 | # define isPRINT_LC(c) (_generic_LC(c, _CC_PRINT, isprint) && ! isCNTRL_LC(c)) | |
1960 | # define isPUNCT_LC(c) (_generic_LC(c, _CC_PUNCT, ispunct) && ! isCNTRL_LC(c)) | |
1961 | # define isUPPER_LC(c) (_generic_LC(c, _CC_UPPER, isupper) && isALPHA_LC(c)) | |
f05550c0 | 1962 | # define isWORDCHAR_LC(c) (((c) == '_') || isALPHANUMERIC_LC(c)) |
b7d90381 KW |
1963 | # define isXDIGIT_LC(c) (_generic_LC(c, _CC_XDIGIT, isxdigit) \ |
1964 | && isALPHANUMERIC_LC(c)) | |
f05550c0 BF |
1965 | |
1966 | # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8) | |
1967 | # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8) | |
1968 | # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8) | |
1969 | ||
1970 | #elif defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII)) | |
4650c663 | 1971 | /* For most other platforms */ |
beab9ebe | 1972 | |
f05550c0 BF |
1973 | # define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha) |
1974 | # define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum) | |
1975 | # define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl) | |
1976 | # define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit) | |
1977 | # define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph) | |
1978 | # define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha) | |
1979 | # define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower) | |
1980 | # define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint) | |
1981 | # define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct) | |
1982 | # define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace) | |
1983 | # define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper) | |
1984 | # define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum) | |
1985 | # define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit) | |
1986 | ||
1987 | ||
1988 | # define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8) | |
1989 | # define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8) | |
1990 | # define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8) | |
1991 | ||
1992 | #else /* The final fallback position */ | |
1993 | ||
b7d90381 KW |
1994 | # define isALPHA_LC(c) (isascii(c) && isalpha(c)) |
1995 | # define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c)) | |
1996 | # define isCNTRL_LC(c) (isascii(c) && iscntrl(c)) | |
1997 | # define isDIGIT_LC(c) (isascii(c) && isdigit(c)) | |
1998 | # define isGRAPH_LC(c) (isascii(c) && isgraph(c)) | |
f05550c0 | 1999 | # define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_')) |
b7d90381 KW |
2000 | # define isLOWER_LC(c) (isascii(c) && islower(c)) |
2001 | # define isPRINT_LC(c) (isascii(c) && isprint(c)) | |
2002 | # define isPUNCT_LC(c) (isascii(c) && ispunct(c)) | |
2003 | # define isSPACE_LC(c) (isascii(c) && isspace(c)) | |
2004 | # define isUPPER_LC(c) (isascii(c) && isupper(c)) | |
f05550c0 | 2005 | # define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_')) |
b7d90381 | 2006 | # define isXDIGIT_LC(c) (isascii(c) && isxdigit(c)) |
f05550c0 BF |
2007 | |
2008 | # define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c)) | |
2009 | # define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c)) | |
2010 | # define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c)) | |
bbce6d69 | 2011 | |
f05550c0 | 2012 | #endif |
55204971 | 2013 | |
eba68aa0 KW |
2014 | #define isIDCONT(c) isWORDCHAR(c) |
2015 | #define isIDCONT_A(c) isWORDCHAR_A(c) | |
2016 | #define isIDCONT_L1(c) isWORDCHAR_L1(c) | |
2017 | #define isIDCONT_LC(c) isWORDCHAR_LC(c) | |
13380643 | 2018 | #define isPSXSPC_LC(c) isSPACE_LC(c) |
aaa51d5e | 2019 | |
4650c663 | 2020 | /* For internal core Perl use only: the base macros for defining macros like |
d0da05db KW |
2021 | * isALPHA_uvchr. 'c' is the code point to check. 'classnum' is the POSIX class |
2022 | * number defined earlier in this file. _generic_uvchr() is used for POSIX | |
4650c663 KW |
2023 | * classes where there is a macro or function 'above_latin1' that takes the |
2024 | * single argument 'c' and returns the desired value. These exist for those | |
2366ba44 KW |
2025 | * classes which have simple definitions, avoiding the overhead of an inversion |
2026 | * list binary search. _generic_invlist_uvchr() can be used | |
4650c663 | 2027 | * for classes where that overhead is faster than a direct lookup. |
d0da05db | 2028 | * _generic_uvchr() won't compile if 'c' isn't unsigned, as it won't match the |
4650c663 KW |
2029 | * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so |
2030 | * have duplicate checks here, so could create versions of the macros that | |
2031 | * don't, but experiments show that gcc optimizes them out anyway. */ | |
66c17564 KW |
2032 | |
2033 | /* Note that all ignore 'use bytes' */ | |
1e222e4f KW |
2034 | #define _generic_uvchr(classnum, above_latin1, c) ((c) < 256 \ |
2035 | ? _generic_isCC(c, classnum) \ | |
cd500f2f | 2036 | : above_latin1(c)) |
2366ba44 | 2037 | #define _generic_invlist_uvchr(classnum, c) ((c) < 256 \ |
1e222e4f | 2038 | ? _generic_isCC(c, classnum) \ |
922e8cb4 | 2039 | : _is_uni_FOO(classnum, c)) |
2366ba44 KW |
2040 | #define isALPHA_uvchr(c) _generic_invlist_uvchr(_CC_ALPHA, c) |
2041 | #define isALPHANUMERIC_uvchr(c) _generic_invlist_uvchr(_CC_ALPHANUMERIC, c) | |
d0da05db KW |
2042 | #define isASCII_uvchr(c) isASCII(c) |
2043 | #define isBLANK_uvchr(c) _generic_uvchr(_CC_BLANK, is_HORIZWS_cp_high, c) | |
2044 | #define isCNTRL_uvchr(c) isCNTRL_L1(c) /* All controls are in Latin1 */ | |
2366ba44 KW |
2045 | #define isDIGIT_uvchr(c) _generic_invlist_uvchr(_CC_DIGIT, c) |
2046 | #define isGRAPH_uvchr(c) _generic_invlist_uvchr(_CC_GRAPH, c) | |
1e222e4f KW |
2047 | #define isIDCONT_uvchr(c) \ |
2048 | _generic_uvchr(_CC_WORDCHAR, _is_uni_perl_idcont, c) | |
2049 | #define isIDFIRST_uvchr(c) \ | |
2050 | _generic_uvchr(_CC_IDFIRST, _is_uni_perl_idstart, c) | |
2366ba44 KW |
2051 | #define isLOWER_uvchr(c) _generic_invlist_uvchr(_CC_LOWER, c) |
2052 | #define isPRINT_uvchr(c) _generic_invlist_uvchr(_CC_PRINT, c) | |
d0da05db | 2053 | |
2366ba44 | 2054 | #define isPUNCT_uvchr(c) _generic_invlist_uvchr(_CC_PUNCT, c) |
d0da05db KW |
2055 | #define isSPACE_uvchr(c) _generic_uvchr(_CC_SPACE, is_XPERLSPACE_cp_high, c) |
2056 | #define isPSXSPC_uvchr(c) isSPACE_uvchr(c) | |
2057 | ||
2366ba44 | 2058 | #define isUPPER_uvchr(c) _generic_invlist_uvchr(_CC_UPPER, c) |
d0da05db | 2059 | #define isVERTWS_uvchr(c) _generic_uvchr(_CC_VERTSPACE, is_VERTWS_cp_high, c) |
2366ba44 | 2060 | #define isWORDCHAR_uvchr(c) _generic_invlist_uvchr(_CC_WORDCHAR, c) |
d0da05db KW |
2061 | #define isXDIGIT_uvchr(c) _generic_uvchr(_CC_XDIGIT, is_XDIGIT_cp_high, c) |
2062 | ||
2063 | #define toFOLD_uvchr(c,s,l) to_uni_fold(c,s,l) | |
2064 | #define toLOWER_uvchr(c,s,l) to_uni_lower(c,s,l) | |
2065 | #define toTITLE_uvchr(c,s,l) to_uni_title(c,s,l) | |
2066 | #define toUPPER_uvchr(c,s,l) to_uni_upper(c,s,l) | |
2067 | ||
2068 | /* For backwards compatibility, even though '_uni' should mean official Unicode | |
2069 | * code points, in Perl it means native for those below 256 */ | |
2070 | #define isALPHA_uni(c) isALPHA_uvchr(c) | |
2071 | #define isALPHANUMERIC_uni(c) isALPHANUMERIC_uvchr(c) | |
2072 | #define isASCII_uni(c) isASCII_uvchr(c) | |
2073 | #define isBLANK_uni(c) isBLANK_uvchr(c) | |
2074 | #define isCNTRL_uni(c) isCNTRL_uvchr(c) | |
2075 | #define isDIGIT_uni(c) isDIGIT_uvchr(c) | |
2076 | #define isGRAPH_uni(c) isGRAPH_uvchr(c) | |
2077 | #define isIDCONT_uni(c) isIDCONT_uvchr(c) | |
2078 | #define isIDFIRST_uni(c) isIDFIRST_uvchr(c) | |
2079 | #define isLOWER_uni(c) isLOWER_uvchr(c) | |
2080 | #define isPRINT_uni(c) isPRINT_uvchr(c) | |
2081 | #define isPUNCT_uni(c) isPUNCT_uvchr(c) | |
2082 | #define isSPACE_uni(c) isSPACE_uvchr(c) | |
2083 | #define isPSXSPC_uni(c) isPSXSPC_uvchr(c) | |
2084 | #define isUPPER_uni(c) isUPPER_uvchr(c) | |
2085 | #define isVERTWS_uni(c) isVERTWS_uvchr(c) | |
2086 | #define isWORDCHAR_uni(c) isWORDCHAR_uvchr(c) | |
2087 | #define isXDIGIT_uni(c) isXDIGIT_uvchr(c) | |
2088 | #define toFOLD_uni(c,s,l) toFOLD_uvchr(c,s,l) | |
2089 | #define toLOWER_uni(c,s,l) toLOWER_uvchr(c,s,l) | |
2090 | #define toTITLE_uni(c,s,l) toTITLE_uvchr(c,s,l) | |
2091 | #define toUPPER_uni(c,s,l) toUPPER_uvchr(c,s,l) | |
a0ed51b3 | 2092 | |
4650c663 KW |
2093 | /* For internal core Perl use only: the base macros for defining macros like |
2094 | * isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code | |
d0da05db | 2095 | * point, not just 0-255. Like _generic_uvchr, there are two versions, one for |
4650c663 | 2096 | * simple class definitions; the other for more complex. These are like |
d0da05db | 2097 | * _generic_uvchr, so see it for more info. */ |
cd500f2f KW |
2098 | #define _generic_LC_uvchr(latin1, above_latin1, c) \ |
2099 | (c < 256 ? latin1(c) : above_latin1(c)) | |
2366ba44 | 2100 | #define _generic_LC_invlist_uvchr(latin1, classnum, c) \ |
cd500f2f KW |
2101 | (c < 256 ? latin1(c) : _is_uni_FOO(classnum, c)) |
2102 | ||
2366ba44 KW |
2103 | #define isALPHA_LC_uvchr(c) _generic_LC_invlist_uvchr(isALPHA_LC, _CC_ALPHA, c) |
2104 | #define isALPHANUMERIC_LC_uvchr(c) _generic_LC_invlist_uvchr(isALPHANUMERIC_LC, \ | |
cd500f2f | 2105 | _CC_ALPHANUMERIC, c) |
b7d90381 KW |
2106 | #define isASCII_LC_uvchr(c) isASCII_LC(c) |
2107 | #define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, \ | |
2108 | is_HORIZWS_cp_high, c) | |
feeab5a9 | 2109 | #define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0) |
2366ba44 KW |
2110 | #define isDIGIT_LC_uvchr(c) _generic_LC_invlist_uvchr(isDIGIT_LC, _CC_DIGIT, c) |
2111 | #define isGRAPH_LC_uvchr(c) _generic_LC_invlist_uvchr(isGRAPH_LC, _CC_GRAPH, c) | |
b7d90381 | 2112 | #define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \ |
eba68aa0 | 2113 | _is_uni_perl_idcont, c) |
b7d90381 | 2114 | #define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \ |
cd500f2f | 2115 | _is_uni_perl_idstart, c) |
2366ba44 KW |
2116 | #define isLOWER_LC_uvchr(c) _generic_LC_invlist_uvchr(isLOWER_LC, _CC_LOWER, c) |
2117 | #define isPRINT_LC_uvchr(c) _generic_LC_invlist_uvchr(isPRINT_LC, _CC_PRINT, c) | |
b7d90381 | 2118 | #define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c) |
2366ba44 | 2119 | #define isPUNCT_LC_uvchr(c) _generic_LC_invlist_uvchr(isPUNCT_LC, _CC_PUNCT, c) |
b7d90381 | 2120 | #define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \ |
509fb054 | 2121 | is_XPERLSPACE_cp_high, c) |
2366ba44 KW |
2122 | #define isUPPER_LC_uvchr(c) _generic_LC_invlist_uvchr(isUPPER_LC, _CC_UPPER, c) |
2123 | #define isWORDCHAR_LC_uvchr(c) _generic_LC_invlist_uvchr(isWORDCHAR_LC, \ | |
cd500f2f | 2124 | _CC_WORDCHAR, c) |
b7d90381 KW |
2125 | #define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, \ |
2126 | is_XDIGIT_cp_high, c) | |
e712593e | 2127 | |
b7d90381 | 2128 | #define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c)) |
aaa51d5e | 2129 | |
da8c1a98 KW |
2130 | /* The "_safe" macros make sure that we don't attempt to read beyond 'e', but |
2131 | * they don't otherwise go out of their way to look for malformed UTF-8. If | |
2132 | * they can return accurate results without knowing if the input is otherwise | |
2133 | * malformed, they do so. For example isASCII is accurate in spite of any | |
2134 | * non-length malformations because it looks only at a single byte. Likewise | |
2135 | * isDIGIT looks just at the first byte for code points 0-255, as all UTF-8 | |
2136 | * variant ones return FALSE. But, if the input has to be well-formed in order | |
2137 | * for the results to be accurate, the macros will test and if malformed will | |
2138 | * call a routine to die | |
2139 | * | |
2140 | * Except for toke.c, the macros do assume that e > p, asserting that on | |
2141 | * DEBUGGING builds. Much code that calls these depends on this being true, | |
2142 | * for other reasons. toke.c is treated specially as using the regular | |
2143 | * assertion breaks it in many ways. All strings that these operate on there | |
2144 | * are supposed to have an extra NUL character at the end, so that *e = \0. A | |
2145 | * bunch of code in toke.c assumes that this is true, so the assertion allows | |
2146 | * for that */ | |
2147 | #ifdef PERL_IN_TOKE_C | |
2148 | # define _utf8_safe_assert(p,e) ((e) > (p) || ((e) == (p) && *(p) == '\0')) | |
2149 | #else | |
2150 | # define _utf8_safe_assert(p,e) ((e) > (p)) | |
2151 | #endif | |
2152 | ||
2153 | #define _generic_utf8_safe(classnum, p, e, above_latin1) \ | |
c81b3562 KW |
2154 | ((! _utf8_safe_assert(p, e)) \ |
2155 | ? (_force_out_malformed_utf8_message((U8 *) (p), (U8 *) (e), 0, 1), 0)\ | |
2156 | : (UTF8_IS_INVARIANT(*(p))) \ | |
da8c1a98 KW |
2157 | ? _generic_isCC(*(p), classnum) \ |
2158 | : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \ | |
2159 | ? ((LIKELY((e) - (p) > 1 && UTF8_IS_CONTINUATION(*((p)+1)))) \ | |
2160 | ? _generic_isCC(EIGHT_BIT_UTF8_TO_NATIVE(*(p), *((p)+1 )), \ | |
2161 | classnum) \ | |
2162 | : (_force_out_malformed_utf8_message( \ | |
2163 | (U8 *) (p), (U8 *) (e), 0, 1), 0)) \ | |
2164 | : above_latin1)) | |
b7d90381 KW |
2165 | /* Like the above, but calls 'above_latin1(p)' to get the utf8 value. |
2166 | * 'above_latin1' can be a macro */ | |
da8c1a98 KW |
2167 | #define _generic_func_utf8_safe(classnum, above_latin1, p, e) \ |
2168 | _generic_utf8_safe(classnum, p, e, above_latin1(p, e)) | |
2366ba44 | 2169 | #define _generic_non_invlist_utf8_safe(classnum, above_latin1, p, e) \ |
da8c1a98 KW |
2170 | _generic_utf8_safe(classnum, p, e, \ |
2171 | (UNLIKELY((e) - (p) < UTF8SKIP(p)) \ | |
2172 | ? (_force_out_malformed_utf8_message( \ | |
2173 | (U8 *) (p), (U8 *) (e), 0, 1), 0) \ | |
2174 | : above_latin1(p))) | |
2366ba44 KW |
2175 | /* Like the above, but passes classnum to _isFOO_utf8(), instead of having an |
2176 | * 'above_latin1' parameter */ | |
2177 | #define _generic_invlist_utf8_safe(classnum, p, e) \ | |
2178 | _generic_utf8_safe(classnum, p, e, _is_utf8_FOO(classnum, p, e)) | |
922e8cb4 | 2179 | |
cc8ab7c0 | 2180 | /* Like the above, but should be used only when it is known that there are no |
ff7ecfc3 KW |
2181 | * characters in the upper-Latin1 range (128-255 on ASCII platforms) which the |
2182 | * class is TRUE for. Hence it can skip the tests for this range. | |
2183 | * 'above_latin1' should include its arguments */ | |
da8c1a98 KW |
2184 | #define _generic_utf8_safe_no_upper_latin1(classnum, p, e, above_latin1) \ |
2185 | (__ASSERT_(_utf8_safe_assert(p, e)) \ | |
2186 | (UTF8_IS_INVARIANT(*(p))) \ | |
2187 | ? _generic_isCC(*(p), classnum) \ | |
2188 | : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \ | |
2189 | ? 0 /* Note that doesn't check validity for latin1 */ \ | |
2190 | : above_latin1) | |
2191 | ||
84238efa | 2192 | |
059703b0 KW |
2193 | #define isALPHA_utf8(p, e) isALPHA_utf8_safe(p, e) |
2194 | #define isALPHANUMERIC_utf8(p, e) isALPHANUMERIC_utf8_safe(p, e) | |
2195 | #define isASCII_utf8(p, e) isASCII_utf8_safe(p, e) | |
2196 | #define isBLANK_utf8(p, e) isBLANK_utf8_safe(p, e) | |
2197 | #define isCNTRL_utf8(p, e) isCNTRL_utf8_safe(p, e) | |
2198 | #define isDIGIT_utf8(p, e) isDIGIT_utf8_safe(p, e) | |
2199 | #define isGRAPH_utf8(p, e) isGRAPH_utf8_safe(p, e) | |
2200 | #define isIDCONT_utf8(p, e) isIDCONT_utf8_safe(p, e) | |
2201 | #define isIDFIRST_utf8(p, e) isIDFIRST_utf8_safe(p, e) | |
2202 | #define isLOWER_utf8(p, e) isLOWER_utf8_safe(p, e) | |
2203 | #define isPRINT_utf8(p, e) isPRINT_utf8_safe(p, e) | |
2204 | #define isPSXSPC_utf8(p, e) isPSXSPC_utf8_safe(p, e) | |
2205 | #define isPUNCT_utf8(p, e) isPUNCT_utf8_safe(p, e) | |
2206 | #define isSPACE_utf8(p, e) isSPACE_utf8_safe(p, e) | |
2207 | #define isUPPER_utf8(p, e) isUPPER_utf8_safe(p, e) | |
2208 | #define isVERTWS_utf8(p, e) isVERTWS_utf8_safe(p, e) | |
2209 | #define isWORDCHAR_utf8(p, e) isWORDCHAR_utf8_safe(p, e) | |
2210 | #define isXDIGIT_utf8(p, e) isXDIGIT_utf8_safe(p, e) | |
e8fa43e2 | 2211 | |
2366ba44 | 2212 | #define isALPHA_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_ALPHA, p, e) |
da8c1a98 | 2213 | #define isALPHANUMERIC_utf8_safe(p, e) \ |
2366ba44 | 2214 | _generic_invlist_utf8_safe(_CC_ALPHANUMERIC, p, e) |
da8c1a98 KW |
2215 | #define isASCII_utf8_safe(p, e) \ |
2216 | /* Because ASCII is invariant under utf8, the non-utf8 macro \ | |
2217 | * works */ \ | |
2218 | (__ASSERT_(_utf8_safe_assert(p, e)) isASCII(*(p))) | |
2219 | #define isBLANK_utf8_safe(p, e) \ | |
2366ba44 | 2220 | _generic_non_invlist_utf8_safe(_CC_BLANK, is_HORIZWS_high, p, e) |
da8c1a98 | 2221 | |
e8fa43e2 KW |
2222 | #ifdef EBCDIC |
2223 | /* Because all controls are UTF-8 invariants in EBCDIC, we can use this | |
2224 | * more efficient macro instead of the more general one */ | |
da8c1a98 | 2225 | # define isCNTRL_utf8_safe(p, e) \ |
56d02b8c | 2226 | (__ASSERT_(_utf8_safe_assert(p, e)) isCNTRL_L1(*(p))) |
e8fa43e2 | 2227 | #else |
da8c1a98 | 2228 | # define isCNTRL_utf8_safe(p, e) _generic_utf8_safe(_CC_CNTRL, p, e, 0) |
e8fa43e2 KW |
2229 | #endif |
2230 | ||
da8c1a98 KW |
2231 | #define isDIGIT_utf8_safe(p, e) \ |
2232 | _generic_utf8_safe_no_upper_latin1(_CC_DIGIT, p, e, \ | |
2366ba44 KW |
2233 | _is_utf8_FOO(_CC_DIGIT, p, e)) |
2234 | #define isGRAPH_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_GRAPH, p, e) | |
da8c1a98 | 2235 | #define isIDCONT_utf8_safe(p, e) _generic_func_utf8_safe(_CC_WORDCHAR, \ |
dd1a3ba7 | 2236 | _is_utf8_perl_idcont, p, e) |
e5dcd934 | 2237 | |
c11ff943 KW |
2238 | /* To prevent S_scan_word in toke.c from hanging, we have to make sure that |
2239 | * IDFIRST is an alnum. See | |
8034715d | 2240 | * https://github.com/Perl/perl5/issues/10275 for more detail than you |
f91dcd13 KW |
2241 | * ever wanted to know about. (In the ASCII range, there isn't a difference.) |
2242 | * This used to be not the XID version, but we decided to go with the more | |
2243 | * modern Unicode definition */ | |
da8c1a98 KW |
2244 | #define isIDFIRST_utf8_safe(p, e) \ |
2245 | _generic_func_utf8_safe(_CC_IDFIRST, \ | |
dd1a3ba7 | 2246 | _is_utf8_perl_idstart, (U8 *) (p), (U8 *) (e)) |
da8c1a98 | 2247 | |
2366ba44 KW |
2248 | #define isLOWER_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_LOWER, p, e) |
2249 | #define isPRINT_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_PRINT, p, e) | |
da8c1a98 | 2250 | #define isPSXSPC_utf8_safe(p, e) isSPACE_utf8_safe(p, e) |
2366ba44 | 2251 | #define isPUNCT_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_PUNCT, p, e) |
da8c1a98 | 2252 | #define isSPACE_utf8_safe(p, e) \ |
2366ba44 KW |
2253 | _generic_non_invlist_utf8_safe(_CC_SPACE, is_XPERLSPACE_high, p, e) |
2254 | #define isUPPER_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_UPPER, p, e) | |
da8c1a98 | 2255 | #define isVERTWS_utf8_safe(p, e) \ |
2366ba44 | 2256 | _generic_non_invlist_utf8_safe(_CC_VERTSPACE, is_VERTWS_high, p, e) |
da8c1a98 | 2257 | #define isWORDCHAR_utf8_safe(p, e) \ |
2366ba44 | 2258 | _generic_invlist_utf8_safe(_CC_WORDCHAR, p, e) |
da8c1a98 KW |
2259 | #define isXDIGIT_utf8_safe(p, e) \ |
2260 | _generic_utf8_safe_no_upper_latin1(_CC_XDIGIT, p, e, \ | |
2261 | (UNLIKELY((e) - (p) < UTF8SKIP(p)) \ | |
2262 | ? (_force_out_malformed_utf8_message( \ | |
2263 | (U8 *) (p), (U8 *) (e), 0, 1), 0) \ | |
2264 | : is_XDIGIT_high(p))) | |
a0ed51b3 | 2265 | |
059703b0 KW |
2266 | #define toFOLD_utf8(p,e,s,l) toFOLD_utf8_safe(p,e,s,l) |
2267 | #define toLOWER_utf8(p,e,s,l) toLOWER_utf8_safe(p,e,s,l) | |
2268 | #define toTITLE_utf8(p,e,s,l) toTITLE_utf8_safe(p,e,s,l) | |
2269 | #define toUPPER_utf8(p,e,s,l) toUPPER_utf8_safe(p,e,s,l) | |
2e8adce6 | 2270 | |
567b353c | 2271 | /* For internal core use only, subject to change */ |
059703b0 KW |
2272 | #define _toFOLD_utf8_flags(p,e,s,l,f) _to_utf8_fold_flags (p,e,s,l,f) |
2273 | #define _toLOWER_utf8_flags(p,e,s,l,f) _to_utf8_lower_flags(p,e,s,l,f) | |
2274 | #define _toTITLE_utf8_flags(p,e,s,l,f) _to_utf8_title_flags(p,e,s,l,f) | |
2275 | #define _toUPPER_utf8_flags(p,e,s,l,f) _to_utf8_upper_flags(p,e,s,l,f) | |
a1a5ec35 KW |
2276 | |
2277 | #define toFOLD_utf8_safe(p,e,s,l) _toFOLD_utf8_flags(p,e,s,l, FOLD_FLAGS_FULL) | |
2278 | #define toLOWER_utf8_safe(p,e,s,l) _toLOWER_utf8_flags(p,e,s,l, 0) | |
2279 | #define toTITLE_utf8_safe(p,e,s,l) _toTITLE_utf8_flags(p,e,s,l, 0) | |
2280 | #define toUPPER_utf8_safe(p,e,s,l) _toUPPER_utf8_flags(p,e,s,l, 0) | |
567b353c | 2281 | |
059703b0 KW |
2282 | #define isALPHA_LC_utf8(p, e) isALPHA_LC_utf8_safe(p, e) |
2283 | #define isALPHANUMERIC_LC_utf8(p, e) isALPHANUMERIC_LC_utf8_safe(p, e) | |
2284 | #define isASCII_LC_utf8(p, e) isASCII_LC_utf8_safe(p, e) | |
2285 | #define isBLANK_LC_utf8(p, e) isBLANK_LC_utf8_safe(p, e) | |
2286 | #define isCNTRL_LC_utf8(p, e) isCNTRL_LC_utf8_safe(p, e) | |
2287 | #define isDIGIT_LC_utf8(p, e) isDIGIT_LC_utf8_safe(p, e) | |
2288 | #define isGRAPH_LC_utf8(p, e) isGRAPH_LC_utf8_safe(p, e) | |
2289 | #define isIDCONT_LC_utf8(p, e) isIDCONT_LC_utf8_safe(p, e) | |
2290 | #define isIDFIRST_LC_utf8(p, e) isIDFIRST_LC_utf8_safe(p, e) | |
2291 | #define isLOWER_LC_utf8(p, e) isLOWER_LC_utf8_safe(p, e) | |
2292 | #define isPRINT_LC_utf8(p, e) isPRINT_LC_utf8_safe(p, e) | |
2293 | #define isPSXSPC_LC_utf8(p, e) isPSXSPC_LC_utf8_safe(p, e) | |
2294 | #define isPUNCT_LC_utf8(p, e) isPUNCT_LC_utf8_safe(p, e) | |
2295 | #define isSPACE_LC_utf8(p, e) isSPACE_LC_utf8_safe(p, e) | |
2296 | #define isUPPER_LC_utf8(p, e) isUPPER_LC_utf8_safe(p, e) | |
2297 | #define isWORDCHAR_LC_utf8(p, e) isWORDCHAR_LC_utf8_safe(p, e) | |
2298 | #define isXDIGIT_LC_utf8(p, e) isXDIGIT_LC_utf8_safe(p, e) | |
34aeb2e9 | 2299 | |
da8c1a98 KW |
2300 | /* For internal core Perl use only: the base macros for defining macros like |
2301 | * isALPHA_LC_utf8_safe. These are like _generic_utf8, but if the first code | |
2302 | * point in 'p' is within the 0-255 range, it uses locale rules from the | |
2303 | * passed-in 'macro' parameter */ | |
2304 | #define _generic_LC_utf8_safe(macro, p, e, above_latin1) \ | |
2305 | (__ASSERT_(_utf8_safe_assert(p, e)) \ | |
2306 | (UTF8_IS_INVARIANT(*(p))) \ | |
2307 | ? macro(*(p)) \ | |
2308 | : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \ | |
2309 | ? ((LIKELY((e) - (p) > 1 && UTF8_IS_CONTINUATION(*((p)+1)))) \ | |
2310 | ? macro(EIGHT_BIT_UTF8_TO_NATIVE(*(p), *((p)+1))) \ | |
2311 | : (_force_out_malformed_utf8_message( \ | |
2312 | (U8 *) (p), (U8 *) (e), 0, 1), 0)) \ | |
2313 | : above_latin1)) | |
2314 | ||
2366ba44 KW |
2315 | #define _generic_LC_invlist_utf8_safe(macro, classnum, p, e) \ |
2316 | _generic_LC_utf8_safe(macro, p, e, \ | |
2317 | _is_utf8_FOO(classnum, p, e)) | |
da8c1a98 KW |
2318 | |
2319 | #define _generic_LC_func_utf8_safe(macro, above_latin1, p, e) \ | |
2320 | _generic_LC_utf8_safe(macro, p, e, above_latin1(p, e)) | |
2321 | ||
2366ba44 | 2322 | #define _generic_LC_non_invlist_utf8_safe(classnum, above_latin1, p, e) \ |
da8c1a98 KW |
2323 | _generic_LC_utf8_safe(classnum, p, e, \ |
2324 | (UNLIKELY((e) - (p) < UTF8SKIP(p)) \ | |
2325 | ? (_force_out_malformed_utf8_message( \ | |
2326 | (U8 *) (p), (U8 *) (e), 0, 1), 0) \ | |
2327 | : above_latin1(p))) | |
2328 | ||
2329 | #define isALPHANUMERIC_LC_utf8_safe(p, e) \ | |
2366ba44 | 2330 | _generic_LC_invlist_utf8_safe(isALPHANUMERIC_LC, \ |
da8c1a98 KW |
2331 | _CC_ALPHANUMERIC, p, e) |
2332 | #define isALPHA_LC_utf8_safe(p, e) \ | |
2366ba44 | 2333 | _generic_LC_invlist_utf8_safe(isALPHA_LC, _CC_ALPHA, p, e) |
da8c1a98 KW |
2334 | #define isASCII_LC_utf8_safe(p, e) \ |
2335 | (__ASSERT_(_utf8_safe_assert(p, e)) isASCII_LC(*(p))) | |
2336 | #define isBLANK_LC_utf8_safe(p, e) \ | |
2366ba44 | 2337 | _generic_LC_non_invlist_utf8_safe(isBLANK_LC, is_HORIZWS_high, p, e) |
da8c1a98 KW |
2338 | #define isCNTRL_LC_utf8_safe(p, e) \ |
2339 | _generic_LC_utf8_safe(isCNTRL_LC, p, e, 0) | |
2340 | #define isDIGIT_LC_utf8_safe(p, e) \ | |
2366ba44 | 2341 | _generic_LC_invlist_utf8_safe(isDIGIT_LC, _CC_DIGIT, p, e) |
da8c1a98 | 2342 | #define isGRAPH_LC_utf8_safe(p, e) \ |
2366ba44 | 2343 | _generic_LC_invlist_utf8_safe(isGRAPH_LC, _CC_GRAPH, p, e) |
da8c1a98 KW |
2344 | #define isIDCONT_LC_utf8_safe(p, e) \ |
2345 | _generic_LC_func_utf8_safe(isIDCONT_LC, \ | |
dd1a3ba7 | 2346 | _is_utf8_perl_idcont, p, e) |
da8c1a98 KW |
2347 | #define isIDFIRST_LC_utf8_safe(p, e) \ |
2348 | _generic_LC_func_utf8_safe(isIDFIRST_LC, \ | |
dd1a3ba7 | 2349 | _is_utf8_perl_idstart, p, e) |
da8c1a98 | 2350 | #define isLOWER_LC_utf8_safe(p, e) \ |
2366ba44 | 2351 | _generic_LC_invlist_utf8_safe(isLOWER_LC, _CC_LOWER, p, e) |
da8c1a98 | 2352 | #define isPRINT_LC_utf8_safe(p, e) \ |
2366ba44 | 2353 | _generic_LC_invlist_utf8_safe(isPRINT_LC, _CC_PRINT, p, e) |
da8c1a98 KW |
2354 | #define isPSXSPC_LC_utf8_safe(p, e) isSPACE_LC_utf8_safe(p, e) |
2355 | #define isPUNCT_LC_utf8_safe(p, e) \ | |
2366ba44 | 2356 | _generic_LC_invlist_utf8_safe(isPUNCT_LC, _CC_PUNCT, p, e) |
da8c1a98 | 2357 | #define isSPACE_LC_utf8_safe(p, e) \ |
2366ba44 | 2358 | _generic_LC_non_invlist_utf8_safe(isSPACE_LC, is_XPERLSPACE_high, p, e) |
da8c1a98 | 2359 | #define isUPPER_LC_utf8_safe(p, e) \ |
2366ba44 | 2360 | _generic_LC_invlist_utf8_safe(isUPPER_LC, _CC_UPPER, p, e) |
da8c1a98 | 2361 | #define isWORDCHAR_LC_utf8_safe(p, e) \ |
2366ba44 | 2362 | _generic_LC_invlist_utf8_safe(isWORDCHAR_LC, _CC_WORDCHAR, p, e) |
da8c1a98 | 2363 | #define isXDIGIT_LC_utf8_safe(p, e) \ |
2366ba44 | 2364 | _generic_LC_non_invlist_utf8_safe(isXDIGIT_LC, is_XDIGIT_high, p, e) |
aaa51d5e | 2365 | |
fbc19f27 KW |
2366 | /* Macros for backwards compatibility and for completeness when the ASCII and |
2367 | * Latin1 values are identical */ | |
b7d90381 KW |
2368 | #define isALPHAU(c) isALPHA_L1(c) |
2369 | #define isDIGIT_L1(c) isDIGIT_A(c) | |
2370 | #define isOCTAL(c) isOCTAL_A(c) | |
2371 | #define isOCTAL_L1(c) isOCTAL_A(c) | |
2372 | #define isXDIGIT_L1(c) isXDIGIT_A(c) | |
2373 | #define isALNUM(c) isWORDCHAR(c) | |
a377c856 | 2374 | #define isALNUM_A(c) isALNUM(c) |
b7d90381 KW |
2375 | #define isALNUMU(c) isWORDCHAR_L1(c) |
2376 | #define isALNUM_LC(c) isWORDCHAR_LC(c) | |
2377 | #define isALNUM_uni(c) isWORDCHAR_uni(c) | |
2e28f0b9 | 2378 | #define isALNUM_LC_uvchr(c) isWORDCHAR_LC_uvchr(c) |
059703b0 | 2379 | #define isALNUM_utf8(p,e) isWORDCHAR_utf8(p,e) |
4c1d9526 | 2380 | #define isALNUM_utf8_safe(p,e) isWORDCHAR_utf8_safe(p,e) |
059703b0 | 2381 | #define isALNUM_LC_utf8(p,e)isWORDCHAR_LC_utf8(p,e) |
4c1d9526 | 2382 | #define isALNUM_LC_utf8_safe(p,e)isWORDCHAR_LC_utf8_safe(p,e) |
b7d90381 KW |
2383 | #define isALNUMC_A(c) isALPHANUMERIC_A(c) /* Mnemonic: "C's alnum" */ |
2384 | #define isALNUMC_L1(c) isALPHANUMERIC_L1(c) | |
2385 | #define isALNUMC(c) isALPHANUMERIC(c) | |
2386 | #define isALNUMC_LC(c) isALPHANUMERIC_LC(c) | |
2387 | #define isALNUMC_uni(c) isALPHANUMERIC_uni(c) | |
15861f94 | 2388 | #define isALNUMC_LC_uvchr(c) isALPHANUMERIC_LC_uvchr(c) |
059703b0 | 2389 | #define isALNUMC_utf8(p,e) isALPHANUMERIC_utf8(p,e) |
4c1d9526 KW |
2390 | #define isALNUMC_utf8_safe(p,e) isALPHANUMERIC_utf8_safe(p,e) |
2391 | #define isALNUMC_LC_utf8_safe(p,e) isALPHANUMERIC_LC_utf8_safe(p,e) | |
fbc19f27 | 2392 | |
2bd1cbf6 KW |
2393 | /* On EBCDIC platforms, CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII, |
2394 | * except that they don't necessarily mean the same characters, e.g. CTRL-D is | |
2395 | * 4 on both systems, but that is EOT on ASCII; ST on EBCDIC. | |
2396 | * '?' is special-cased on EBCDIC to APC, which is the control there that is | |
2397 | * the outlier from the block that contains the other controls, just like | |
2398 | * toCTRL('?') on ASCII yields DEL, the control that is the outlier from the C0 | |
2399 | * block. If it weren't special cased, it would yield a non-control. | |
88794300 KW |
2400 | * The conversion works both ways, so toCTRL('D') is 4, and toCTRL(4) is D, |
2401 | * etc. */ | |
2bd1cbf6 | 2402 | #ifndef EBCDIC |
75763b3a | 2403 | # define toCTRL(c) (__ASSERT_(FITS_IN_8_BITS(c)) toUPPER(((U8)(c))) ^ 64) |
2bd1cbf6 | 2404 | #else |
75763b3a KW |
2405 | # define toCTRL(c) (__ASSERT_(FITS_IN_8_BITS(c)) \ |
2406 | ((isPRINT_A(c)) \ | |
2407 | ? (UNLIKELY((c) == '?') \ | |
2408 | ? QUESTION_MARK_CTRL \ | |
2409 | : (NATIVE_TO_LATIN1(toUPPER((U8) (c))) ^ 64)) \ | |
2410 | : (UNLIKELY((c) == QUESTION_MARK_CTRL) \ | |
2411 | ? '?' \ | |
2412 | : (LATIN1_TO_NATIVE(((U8) (c)) ^ 64))))) | |
2bd1cbf6 | 2413 | #endif |
bbce6d69 | 2414 | |
dea28490 JJ |
2415 | /* Line numbers are unsigned, 32 bits. */ |
2416 | typedef U32 line_t; | |
e5dcd934 | 2417 | #define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */ |
378cc40b | 2418 | |
91152fc1 DG |
2419 | /* Helpful alias for version prescan */ |
2420 | #define is_LAX_VERSION(a,b) \ | |
2421 | (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL)) | |
2422 | ||
2423 | #define is_STRICT_VERSION(a,b) \ | |
2424 | (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL)) | |
2425 | ||
2426 | #define BADVERSION(a,b,c) \ | |
2427 | if (b) { \ | |
2428 | *b = c; \ | |
2429 | } \ | |
2430 | return a; | |
8c52afec | 2431 | |
1ce77b7d KW |
2432 | /* Converts a character KNOWN to represent a hexadecimal digit (0-9, A-F, or |
2433 | * a-f) to its numeric value without using any branches. The input is | |
2434 | * validated only by an assert() in DEBUGGING builds. | |
2435 | * | |
2436 | * It works by right shifting and isolating the bit that is 0 for the digits, | |
2437 | * and 1 for at least the alphas A-F, a-f. The bit is shifted to the ones | |
2438 | * position, and then to the eights position. Both are added together to form | |
2439 | * 0 if the input is '0'-'9' and to form 9 if alpha. This is added to the | |
2440 | * final four bits of the input to form the correct value. */ | |
2441 | #define XDIGIT_VALUE(c) (__ASSERT_(isXDIGIT(c)) \ | |
2442 | ((NATIVE_TO_LATIN1(c) >> 6) & 1) /* 1 if alpha; 0 if not */ \ | |
2443 | + ((NATIVE_TO_LATIN1(c) >> 3) & 8) /* 8 if alpha; 0 if not */ \ | |
2444 | + ((c) & 0xF)) /* 0-9 if input valid hex digit */ | |
2445 | ||
2446 | /* The argument is a string pointer, which is advanced. */ | |
2447 | #define READ_XDIGIT(s) ((s)++, XDIGIT_VALUE(*((s) - 1))) | |
95a59cab | 2448 | |
cb27eebd KW |
2449 | /* Converts a character known to represent an octal digit (0-7) to its numeric |
2450 | * value. The input is validated only by an assert() in DEBUGGING builds. In | |
2451 | * both ASCII and EBCDIC the last 3 bits of the octal digits range from 0-7. */ | |
2452 | #define OCTAL_VALUE(c) (__ASSERT_(isOCTAL(c)) (7 & (c))) | |
2453 | ||
305b8651 | 2454 | /* Efficiently returns a boolean as to if two native characters are equivalent |
f1460a66 | 2455 | * case-insensitively. At least one of the characters must be one of [A-Za-z]; |
305b8651 KW |
2456 | * the ALPHA in the name is to remind you of that. This is asserted() in |
2457 | * DEBUGGING builds. Because [A-Za-z] are invariant under UTF-8, this macro | |
2458 | * works (on valid input) for both non- and UTF-8-encoded bytes. | |
2459 | * | |
2460 | * When one of the inputs is a compile-time constant and gets folded by the | |
2461 | * compiler, this reduces to an AND and a TEST. On both EBCDIC and ASCII | |
2462 | * machines, 'A' and 'a' differ by a single bit; the same with the upper and | |
2463 | * lower case of all other ASCII-range alphabetics. On ASCII platforms, they | |
96ca48da KW |
2464 | * are 32 apart; on EBCDIC, they are 64. At compile time, this uses an |
2465 | * exclusive 'or' to find that bit and then inverts it to form a mask, with | |
2466 | * just a single 0, in the bit position where the upper- and lowercase differ. | |
2467 | * */ | |
305b8651 KW |
2468 | #define isALPHA_FOLD_EQ(c1, c2) \ |
2469 | (__ASSERT_(isALPHA_A(c1) || isALPHA_A(c2)) \ | |
2470 | ((c1) & ~('A' ^ 'a')) == ((c2) & ~('A' ^ 'a'))) | |
2471 | #define isALPHA_FOLD_NE(c1, c2) (! isALPHA_FOLD_EQ((c1), (c2))) | |
2472 | ||
8e84507e | 2473 | /* |
3f620621 | 2474 | =for apidoc_section $memory |
ccfc67b7 | 2475 | |
a02a5408 | 2476 | =for apidoc Am|void|Newx|void* ptr|int nitems|type |
954c1994 GS |
2477 | The XSUB-writer's interface to the C C<malloc> function. |
2478 | ||
596f7718 | 2479 | Memory obtained by this should B<ONLY> be freed with L</"Safefree">. |
0d7b2759 | 2480 | |
c5008215 JC |
2481 | In 5.9.3, Newx() and friends replace the older New() API, and drops |
2482 | the first parameter, I<x>, a debug aid which allowed callers to identify | |
37b8b4c9 | 2483 | themselves. This aid has been superseded by a new build option, |
d10b4965 | 2484 | PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still |
c5008215 JC |
2485 | there for use in XS modules supporting older perls. |
2486 | ||
a02a5408 | 2487 | =for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast |
954c1994 | 2488 | The XSUB-writer's interface to the C C<malloc> function, with |
fbe13c60 | 2489 | cast. See also C<L</Newx>>. |
954c1994 | 2490 | |
596f7718 | 2491 | Memory obtained by this should B<ONLY> be freed with L</"Safefree">. |
0d7b2759 | 2492 | |
a02a5408 | 2493 | =for apidoc Am|void|Newxz|void* ptr|int nitems|type |
954c1994 | 2494 | The XSUB-writer's interface to the C C<malloc> function. The allocated |
fbe13c60 | 2495 | memory is zeroed with C<memzero>. See also C<L</Newx>>. |
a02a5408 | 2496 | |
596f7718 | 2497 | Memory obtained by this should B<ONLY> be freed with L</"Safefree">. |
0d7b2759 | 2498 | |
954c1994 GS |
2499 | =for apidoc Am|void|Renew|void* ptr|int nitems|type |
2500 | The XSUB-writer's interface to the C C<realloc> function. | |
2501 | ||
596f7718 | 2502 | Memory obtained by this should B<ONLY> be freed with L</"Safefree">. |
0d7b2759 | 2503 | |
954c1994 GS |
2504 | =for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast |
2505 | The XSUB-writer's interface to the C C<realloc> function, with | |
2506 | cast. | |
2507 | ||
596f7718 | 2508 | Memory obtained by this should B<ONLY> be freed with L</"Safefree">. |
0d7b2759 | 2509 | |
49b8b560 | 2510 | =for apidoc Am|void|Safefree|void* ptr |
954c1994 GS |
2511 | The XSUB-writer's interface to the C C<free> function. |
2512 | ||
596f7718 | 2513 | This should B<ONLY> be used on memory obtained using L</"Newx"> and friends. |
0d7b2759 | 2514 | |
3f620621 | 2515 | =for apidoc_section $string |
954c1994 GS |
2516 | =for apidoc Am|void|Move|void* src|void* dest|int nitems|type |
2517 | The XSUB-writer's interface to the C C<memmove> function. The C<src> is the | |
926bb54c | 2518 | source, C<dest> is the destination, C<nitems> is the number of items, and |
fbe13c60 | 2519 | C<type> is the type. Can do overlapping moves. See also C<L</Copy>>. |
954c1994 | 2520 | |
e90e2364 | 2521 | =for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type |
796b6530 | 2522 | Like C<Move> but returns C<dest>. Useful |
72d33970 | 2523 | for encouraging compilers to tail-call |
e90e2364 NC |
2524 | optimise. |
2525 | ||
954c1994 GS |
2526 | =for apidoc Am|void|Copy|void* src|void* dest|int nitems|type |
2527 | The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the | |
926bb54c | 2528 | source, C<dest> is the destination, C<nitems> is the number of items, and |
fbe13c60 | 2529 | C<type> is the type. May fail on overlapping copies. See also C<L</Move>>. |
954c1994 | 2530 | |
e90e2364 NC |
2531 | =for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type |
2532 | ||
796b6530 | 2533 | Like C<Copy> but returns C<dest>. Useful |
72d33970 | 2534 | for encouraging compilers to tail-call |
e90e2364 NC |
2535 | optimise. |
2536 | ||
954c1994 GS |
2537 | =for apidoc Am|void|Zero|void* dest|int nitems|type |
2538 | ||
2539 | The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the | |
2540 | destination, C<nitems> is the number of items, and C<type> is the type. | |
2541 | ||
e90e2364 NC |
2542 | =for apidoc Am|void *|ZeroD|void* dest|int nitems|type |
2543 | ||
72d33970 FC |
2544 | Like C<Zero> but returns dest. Useful |
2545 | for encouraging compilers to tail-call | |
e90e2364 NC |
2546 | optimise. |
2547 | ||
3f620621 | 2548 | =for apidoc_section $utility |
e538104b | 2549 | =for apidoc Amu|void|StructCopy|type *src|type *dest|type |
4375e838 | 2550 | This is an architecture-independent macro to copy one structure to another. |
954c1994 | 2551 | |
7e337ee0 JH |
2552 | =for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte |
2553 | ||
2554 | Fill up memory with a byte pattern (a byte repeated over and over | |
2555 | again) that hopefully catches attempts to access uninitialized memory. | |
2556 | ||
2557 | =for apidoc Am|void|PoisonNew|void* dest|int nitems|type | |
2558 | ||
2559 | PoisonWith(0xAB) for catching access to allocated but uninitialized memory. | |
2560 | ||
1c12ffb4 | 2561 | =for apidoc Am|void|PoisonFree|void* dest|int nitems|type |
7e337ee0 JH |
2562 | |
2563 | PoisonWith(0xEF) for catching access to freed memory. | |
2564 | ||
9965345d JH |
2565 | =for apidoc Am|void|Poison|void* dest|int nitems|type |
2566 | ||
7e337ee0 | 2567 | PoisonWith(0xEF) for catching access to freed memory. |
9965345d JH |
2568 | |
2569 | =cut */ | |
954c1994 | 2570 | |
561b68a9 SH |
2571 | /* Maintained for backwards-compatibility only. Use newSV() instead. */ |
2572 | #ifndef PERL_CORE | |
ff06c60c | 2573 | #define NEWSV(x,len) newSV(len) |
561b68a9 | 2574 | #endif |
ff06c60c | 2575 | |
b7112dce | 2576 | #define MEM_SIZE_MAX ((MEM_SIZE)-1) |
19a94d75 | 2577 | |
a500027b | 2578 | #define _PERL_STRLEN_ROUNDUP_UNCHECKED(n) (((n) - 1 + PERL_STRLEN_ROUNDUP_QUANTUM) & ~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM - 1)) |
e6bdf523 | 2579 | |
27d5b266 | 2580 | #ifdef PERL_MALLOC_WRAP |
e6bdf523 DM |
2581 | |
2582 | /* This expression will be constant-folded at compile time. It checks | |
2583 | * whether or not the type of the count n is so small (e.g. U8 or U16, or | |
2584 | * U32 on 64-bit systems) that there's no way a wrap-around could occur. | |
2585 | * As well as avoiding the need for a run-time check in some cases, it's | |
2586 | * designed to avoid compiler warnings like: | |
2587 | * comparison is always false due to limited range of data type | |
73e8ff00 DM |
2588 | * It's mathematically equivalent to |
2589 | * max(n) * sizeof(t) > MEM_SIZE_MAX | |
e6bdf523 DM |
2590 | */ |
2591 | ||
2592 | # define _MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) \ | |
445198b9 LM |
2593 | ( sizeof(MEM_SIZE) < sizeof(n) \ |
2594 | || sizeof(t) > ((MEM_SIZE)1 << 8*(sizeof(MEM_SIZE) - sizeof(n)))) | |
e6bdf523 | 2595 | |
88f9f128 | 2596 | /* This is written in a slightly odd way to avoid various spurious |
d98e5cde DM |
2597 | * compiler warnings. We *want* to write the expression as |
2598 | * _MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) && (n > C) | |
2599 | * (for some compile-time constant C), but even when the LHS | |
2600 | * constant-folds to false at compile-time, g++ insists on emitting | |
2601 | * warnings about the RHS (e.g. "comparison is always false"), so instead | |
2602 | * we write it as | |
e6bdf523 | 2603 | * |
d98e5cde | 2604 | * (cond ? n : X) > C |
88f9f128 | 2605 | * |
d98e5cde DM |
2606 | * where X is a constant with X > C always false. Choosing a value for X |
2607 | * is tricky. If 0, some compilers will complain about 0 > C always being | |
2608 | * false; if 1, Coverity complains when n happens to be the constant value | |
2609 | * '1', that cond ? 1 : 1 has the same value on both branches; so use C | |
2610 | * for X and hope that nothing else whines. | |
e6bdf523 DM |
2611 | */ |
2612 | ||
2613 | # define _MEM_WRAP_WILL_WRAP(n,t) \ | |
88f9f128 DM |
2614 | ((_MEM_WRAP_NEEDS_RUNTIME_CHECK(n,t) ? (MEM_SIZE)(n) : \ |
2615 | MEM_SIZE_MAX/sizeof(t)) > MEM_SIZE_MAX/sizeof(t)) | |
e6bdf523 DM |
2616 | |
2617 | # define MEM_WRAP_CHECK(n,t) \ | |
2618 | (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \ | |
2619 | && (croak_memory_wrap(),0)) | |
2620 | ||
2621 | # define MEM_WRAP_CHECK_1(n,t,a) \ | |
2622 | (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \ | |
2623 | && (Perl_croak_nocontext("%s",(a)),0)) | |
2624 | ||
814eedc8 DD |
2625 | /* "a" arg must be a string literal */ |
2626 | # define MEM_WRAP_CHECK_s(n,t,a) \ | |
2627 | (void)(UNLIKELY(_MEM_WRAP_WILL_WRAP(n,t)) \ | |
2628 | && (Perl_croak_nocontext("" a ""),0)) | |
2629 | ||
8b44ba4c | 2630 | #define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t), |
27d5b266 | 2631 | |
a500027b | 2632 | #define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (croak_memory_wrap(),0) : 0), _PERL_STRLEN_ROUNDUP_UNCHECKED(n)) |
27d5b266 JH |
2633 | #else |
2634 | ||
410319be NC |
2635 | #define MEM_WRAP_CHECK(n,t) |
2636 | #define MEM_WRAP_CHECK_1(n,t,a) | |
814eedc8 | 2637 | #define MEM_WRAP_CHECK_s(n,t,a) |
8b44ba4c NC |
2638 | #define MEM_WRAP_CHECK_(n,t) |
2639 | ||
a500027b | 2640 | #define PERL_STRLEN_ROUNDUP(n) _PERL_STRLEN_ROUNDUP_UNCHECKED(n) |
27d5b266 | 2641 | |
1936d2a7 | 2642 | #endif |
8b44ba4c | 2643 | |
fe4f188c | 2644 | #ifdef PERL_MEM_LOG |
46c6c7e2 | 2645 | /* |
9f653bb5 | 2646 | * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s |
46c6c7e2 JH |
2647 | * go through functions, which are handy for debugging breakpoints, but |
2648 | * which more importantly get the immediate calling environment (file and | |
e352bcff JH |
2649 | * line number, and C function name if available) passed in. This info can |
2650 | * then be used for logging the calls, for which one gets a sample | |
73d1d973 | 2651 | * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined. |
3609ea0d | 2652 | * |
46c6c7e2 | 2653 | * Known problems: |
94e892a6 | 2654 | * - not all memory allocs get logged, only those |
46c6c7e2 | 2655 | * that go through Newx() and derivatives (while all |
94e892a6 | 2656 | * Safefrees do get logged) |
46c6c7e2 JH |
2657 | * - __FILE__ and __LINE__ do not work everywhere |
2658 | * - __func__ or __FUNCTION__ even less so | |
2659 | * - I think more goes on after the perlio frees but | |
2660 | * the thing is that STDERR gets closed (as do all | |
2661 | * the file descriptors) | |
2662 | * - no deeper calling stack than the caller of the Newx() | |
2663 | * or the kind, but do I look like a C reflection/introspection | |
2664 | * utility to you? | |
2665 | * - the function prototypes for the logging functions | |
2666 | * probably should maybe be somewhere else than handy.h | |
2667 | * - one could consider inlining (macrofying) the logging | |
2668 | * for speed, but I am too lazy | |
2669 | * - one could imagine recording the allocations in a hash, | |
2670 | * (keyed by the allocation address?), and maintain that | |
2671 | * through reallocs and frees, but how to do that without | |
2672 | * any News() happening...? | |
73d1d973 | 2673 | * - lots of -Ddefines to get useful/controllable output |
b953482e | 2674 | * - lots of ENV reads |
46c6c7e2 JH |
2675 | */ |
2676 | ||
0b0ab801 | 2677 | # ifdef PERL_CORE |
73d1d973 | 2678 | # ifndef PERL_MEM_LOG_NOIMPL |
0b0ab801 MHM |
2679 | enum mem_log_type { |
2680 | MLT_ALLOC, | |
2681 | MLT_REALLOC, | |
d7a2c63c MHM |
2682 | MLT_FREE, |
2683 | MLT_NEW_SV, | |
2684 | MLT_DEL_SV | |
0b0ab801 MHM |
2685 | }; |
2686 | # endif | |
12754f92 | 2687 | # if defined(PERL_IN_SV_C) /* those are only used in sv.c */ |
d7a2c63c MHM |
2688 | void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); |
2689 | void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname); | |
12754f92 | 2690 | # endif |
0b0ab801 MHM |
2691 | # endif |
2692 | ||
fe4f188c JH |
2693 | #endif |
2694 | ||
2695 | #ifdef PERL_MEM_LOG | |
d1401ee9 MHM |
2696 | #define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__) |
2697 | #define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__) | |
46c6c7e2 | 2698 | #define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__) |
fe4f188c JH |
2699 | #endif |
2700 | ||
2701 | #ifndef MEM_LOG_ALLOC | |
2702 | #define MEM_LOG_ALLOC(n,t,a) (a) | |
2703 | #endif | |
2704 | #ifndef MEM_LOG_REALLOC | |
2705 | #define MEM_LOG_REALLOC(n,t,v,a) (a) | |
2706 | #endif | |
2707 | #ifndef MEM_LOG_FREE | |
2708 | #define MEM_LOG_FREE(a) (a) | |
2709 | #endif | |
2710 | ||
d1401ee9 MHM |
2711 | #define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) |
2712 | #define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t)))))) | |
2713 | #define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t))))) | |
a6f6820f NC |
2714 | |
2715 | #ifndef PERL_CORE | |
a02a5408 JC |
2716 | /* pre 5.9.x compatibility */ |
2717 | #define New(x,v,n,t) Newx(v,n,t) | |
2718 | #define Newc(x,v,n,t,c) Newxc(v,n,t,c) | |
4541904d | 2719 | #define Newz(x,v,n,t) Newxz(v,n,t) |
a6f6820f | 2720 | #endif |
a02a5408 | 2721 | |
ff68c719 | 2722 | #define Renew(v,n,t) \ |
d1401ee9 | 2723 | (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) |
ff68c719 | 2724 | #define Renewc(v,n,t,c) \ |
d1401ee9 | 2725 | (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t)))))) |
94010e71 NC |
2726 | |
2727 | #ifdef PERL_POISON | |
2728 | #define Safefree(d) \ | |
06c0cc96 | 2729 | ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0) |
94010e71 | 2730 | #else |
fe4f188c | 2731 | #define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d))) |
94010e71 | 2732 | #endif |
55497cff | 2733 | |
dbb57106 YO |
2734 | /* assert that a valid ptr has been supplied - use this instead of assert(ptr) * |
2735 | * as it handles cases like constant string arguments without throwing warnings * | |
2736 | * the cast is required, as is the inequality check, to avoid warnings */ | |
45908e4d | 2737 | #define perl_assert_ptr(p) assert( ((void*)(p)) != 0 ) |
55497cff | 2738 | |
45908e4d YO |
2739 | |
2740 | #define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) | |
2741 | #define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) | |
2742 | #define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), (void)memzero((char*)(d), (n) * sizeof(t))) | |
2743 | ||
bdd1531d | 2744 | /* Like above, but returns a pointer to 'd' */ |
45908e4d YO |
2745 | #define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), memmove((char*)(d),(const char*)(s), (n) * sizeof(t))) |
2746 | #define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), perl_assert_ptr(s), memcpy((char*)(d),(const char*)(s), (n) * sizeof(t))) | |
45908e4d | 2747 | #define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) perl_assert_ptr(d), memzero((char*)(d), (n) * sizeof(t))) |
e90e2364 | 2748 | |
7e337ee0 JH |
2749 | #define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t))) |
2750 | #define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB) | |
2751 | #define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF) | |
2752 | #define Poison(d,n,t) PoisonFree(d,n,t) | |
27d5b266 | 2753 | |
caa674f3 DD |
2754 | #ifdef PERL_POISON |
2755 | # define PERL_POISON_EXPR(x) x | |
2756 | #else | |
2757 | # define PERL_POISON_EXPR(x) | |
2758 | #endif | |
2759 | ||
ff68c719 | 2760 | #define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s))) |
2cc61e15 | 2761 | |
1b7e2294 | 2762 | /* |
3f620621 | 2763 | =for apidoc_section $utility |
1b7e2294 KW |
2764 | |
2765 | =for apidoc Am|STRLEN|C_ARRAY_LENGTH|void *a | |
2766 | ||
2767 | Returns the number of elements in the input C array (so you want your | |
2768 | zero-based indices to be less than but not equal to). | |
2769 | ||
2770 | =for apidoc Am|void *|C_ARRAY_END|void *a | |
2771 | ||
2772 | Returns a pointer to one element past the final element of the input C array. | |
2773 | ||
2774 | =cut | |
2775 | ||
2776 | C_ARRAY_END is one past the last: half-open/half-closed range, not | |
2777 | last-inclusive range. | |
2778 | */ | |
622913ab | 2779 | #define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0])) |
c3caa5c3 | 2780 | #define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a)) |
622913ab | 2781 | |
2cc61e15 DD |
2782 | #ifdef NEED_VA_COPY |
2783 | # ifdef va_copy | |
2784 | # define Perl_va_copy(s, d) va_copy(d, s) | |
07798b17 AC |
2785 | # elif defined(__va_copy) |
2786 | # define Perl_va_copy(s, d) __va_copy(d, s) | |
2cc61e15 | 2787 | # else |
07798b17 | 2788 | # define Perl_va_copy(s, d) Copy(s, d, 1, va_list) |
2cc61e15 DD |
2789 | # endif |
2790 | #endif | |
2791 | ||
472d47bc SB |
2792 | /* convenience debug macros */ |
2793 | #ifdef USE_ITHREADS | |
2794 | #define pTHX_FORMAT "Perl interpreter: 0x%p" | |
2795 | #define pTHX__FORMAT ", Perl interpreter: 0x%p" | |
f54cb97a AL |
2796 | #define pTHX_VALUE_ (void *)my_perl, |
2797 | #define pTHX_VALUE (void *)my_perl | |
2798 | #define pTHX__VALUE_ ,(void *)my_perl, | |
2799 | #define pTHX__VALUE ,(void *)my_perl | |
472d47bc | 2800 | #else |
3609ea0d | 2801 | #define pTHX_FORMAT |
472d47bc | 2802 | #define pTHX__FORMAT |
3609ea0d | 2803 | #define pTHX_VALUE_ |
472d47bc | 2804 | #define pTHX_VALUE |
3609ea0d | 2805 | #define pTHX__VALUE_ |
472d47bc SB |
2806 | #define pTHX__VALUE |
2807 | #endif /* USE_ITHREADS */ | |
3609ea0d | 2808 | |
2acdbac1 NC |
2809 | /* Perl_deprecate was not part of the public API, and did not have a deprecate() |
2810 | shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor | |
2811 | CPAN::Unpack show any users outside the core. */ | |
2812 | #ifdef PERL_CORE | |
dc6e8de0 A |
2813 | # define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \ |
2814 | "Use of " s " is deprecated") | |
c9680906 A |
2815 | # define deprecate_disappears_in(when,message) \ |
2816 | Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \ | |
2817 | message ", and will disappear in Perl " when) | |
ac641426 A |
2818 | # define deprecate_fatal_in(when,message) \ |
2819 | Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \ | |
2820 | message ". Its use will be fatal in Perl " when) | |
2acdbac1 NC |
2821 | #endif |
2822 | ||
dfff4baf BF |
2823 | /* Internal macros to deal with gids and uids */ |
2824 | #ifdef PERL_CORE | |
2825 | ||
2826 | # if Uid_t_size > IVSIZE | |
2827 | # define sv_setuid(sv, uid) sv_setnv((sv), (NV)(uid)) | |
2828 | # define SvUID(sv) SvNV(sv) | |
07798b17 AC |
2829 | # elif Uid_t_sign <= 0 |
2830 | # define sv_setuid(sv, uid) sv_setiv((sv), (IV)(uid)) | |
2831 | # define SvUID(sv) SvIV(sv) | |
dfff4baf | 2832 | # else |
07798b17 AC |
2833 | # define sv_setuid(sv, uid) sv_setuv((sv), (UV)(uid)) |
2834 | # define SvUID(sv) SvUV(sv) | |
dfff4baf BF |
2835 | # endif /* Uid_t_size */ |
2836 | ||
2837 | # if Gid_t_size > IVSIZE | |
2838 | # define sv_setgid(sv, gid) sv_setnv((sv), (NV)(gid)) | |
2839 | # define SvGID(sv) SvNV(sv) | |
07798b17 AC |
2840 | # elif Gid_t_sign <= 0 |
2841 | # define sv_setgid(sv, gid) sv_setiv((sv), (IV)(gid)) | |
2842 | # define SvGID(sv) SvIV(sv) | |
dfff4baf | 2843 | # else |
07798b17 AC |
2844 | # define sv_setgid(sv, gid) sv_setuv((sv), (UV)(gid)) |
2845 | # define SvGID(sv) SvUV(sv) | |
dfff4baf BF |
2846 | # endif /* Gid_t_size */ |
2847 | ||
2848 | #endif | |
2849 | ||
6a5bc5ac | 2850 | #endif /* PERL_HANDY_H_ */ |
9d745869 | 2851 | |
e9a8c099 | 2852 | /* |
14d04a33 | 2853 | * ex: set ts=8 sts=4 sw=4 et: |
e9a8c099 | 2854 | */ |