1 /* Multibyte character data type.
2 Copyright (C) 2001, 2005-2006 Free Software Foundation, Inc.
3
4 This program is free software; you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation; either version 2, or (at your option)
7 any later version.
8
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
13
14 You should have received a copy of the GNU General Public License
15 along with this program; if not, write to the Free Software Foundation,
16 Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
17
18 /* Written by Bruno Haible <bruno@clisp.org>. */
19
20 /* A multibyte character is a short subsequence of a char* string,
21 representing a single wide character.
22
23 We use multibyte characters instead of wide characters because of
24 the following goals:
25 1) correct multibyte handling, i.e. operate according to the LC_CTYPE
26 locale,
27 2) ease of maintenance, i.e. the maintainer needs not know all details
28 of the ISO C 99 standard,
29 3) don't fail grossly if the input is not in the encoding set by the
30 locale, because often different encodings are in use in the same
31 countries (ISO-8859-1/UTF-8, EUC-JP/Shift_JIS, ...),
32 4) fast in the case of ASCII characters,
33 5) portability, i.e. don't make unportable assumptions about wchar_t.
34
35 Multibyte characters are only accessed through the mb* macros.
36
37 mb_ptr (mbc)
38 return a pointer to the beginning of the multibyte sequence.
39
40 mb_len (mbc)
41 returns the number of bytes occupied by the multibyte sequence.
42 Always > 0.
43
44 mb_iseq (mbc, sc)
45 returns true if mbc is the standard ASCII character sc.
46
47 mb_isnul (mbc)
48 returns true if mbc is the nul character.
49
50 mb_cmp (mbc1, mbc2)
51 returns a positive, zero, or negative value depending on whether mbc1
52 sorts after, same or before mbc2.
53
54 mb_casecmp (mbc1, mbc2)
55 returns a positive, zero, or negative value depending on whether mbc1
56 sorts after, same or before mbc2, modulo upper/lowercase conversion.
57
58 mb_equal (mbc1, mbc2)
59 returns true if mbc1 and mbc2 are equal.
60
61 mb_caseequal (mbc1, mbc2)
62 returns true if mbc1 and mbc2 are equal modulo upper/lowercase conversion.
63
64 mb_isalnum (mbc)
65 returns true if mbc is alphanumeric.
66
67 mb_isalpha (mbc)
68 returns true if mbc is alphabetic.
69
70 mb_isascii(mbc)
71 returns true if mbc is plain ASCII.
72
73 mb_isblank (mbc)
74 returns true if mbc is a blank.
75
76 mb_iscntrl (mbc)
77 returns true if mbc is a control character.
78
79 mb_isdigit (mbc)
80 returns true if mbc is a decimal digit.
81
82 mb_isgraph (mbc)
83 returns true if mbc is a graphic character.
84
85 mb_islower (mbc)
86 returns true if mbc is lowercase.
87
88 mb_isprint (mbc)
89 returns true if mbc is a printable character.
90
91 mb_ispunct (mbc)
92 returns true if mbc is a punctuation character.
93
94 mb_isspace (mbc)
95 returns true if mbc is a space character.
96
97 mb_isupper (mbc)
98 returns true if mbc is uppercase.
99
100 mb_isxdigit (mbc)
101 returns true if mbc is a hexadecimal digit.
102
103 mb_width (mbc)
104 returns the number of columns on the output device occupied by mbc.
105 Always >= 0.
106
107 mb_putc (mbc, stream)
108 outputs mbc on stream, a byte oriented FILE stream opened for output.
109
110 mb_setascii (&mbc, sc)
111 assigns the standard ASCII character sc to mbc.
112
113 mb_copy (&destmbc, &srcmbc)
114 copies srcmbc to destmbc.
115
116 Here are the function prototypes of the macros.
117
118 extern const char * mb_ptr (const mbchar_t mbc);
119 extern size_t mb_len (const mbchar_t mbc);
120 extern bool mb_iseq (const mbchar_t mbc, char sc);
121 extern bool mb_isnul (const mbchar_t mbc);
122 extern int mb_cmp (const mbchar_t mbc1, const mbchar_t mbc2);
123 extern int mb_casecmp (const mbchar_t mbc1, const mbchar_t mbc2);
124 extern bool mb_equal (const mbchar_t mbc1, const mbchar_t mbc2);
125 extern bool mb_caseequal (const mbchar_t mbc1, const mbchar_t mbc2);
126 extern bool mb_isalnum (const mbchar_t mbc);
127 extern bool mb_isalpha (const mbchar_t mbc);
128 extern bool mb_isascii (const mbchar_t mbc);
129 extern bool mb_isblank (const mbchar_t mbc);
130 extern bool mb_iscntrl (const mbchar_t mbc);
131 extern bool mb_isdigit (const mbchar_t mbc);
132 extern bool mb_isgraph (const mbchar_t mbc);
133 extern bool mb_islower (const mbchar_t mbc);
134 extern bool mb_isprint (const mbchar_t mbc);
135 extern bool mb_ispunct (const mbchar_t mbc);
136 extern bool mb_isspace (const mbchar_t mbc);
137 extern bool mb_isupper (const mbchar_t mbc);
138 extern bool mb_isxdigit (const mbchar_t mbc);
139 extern int mb_width (const mbchar_t mbc);
140 extern void mb_putc (const mbchar_t mbc, FILE *stream);
141 extern void mb_setascii (mbchar_t *new, char sc);
142 extern void mb_copy (mbchar_t *new, const mbchar_t *old);
143 */
144
145 #ifndef _MBCHAR_H
146 #define _MBCHAR_H 1
147
148 #include <stdbool.h>
149 #include <string.h>
150
151 /* Tru64 with Desktop Toolkit C has a bug: <stdio.h> must be included before
152 <wchar.h>.
153 BSD/OS 4.1 has a bug: <stdio.h> and <time.h> must be included before
154 <wchar.h>. */
155 #include <stdio.h>
156 #include <time.h>
157 #include <wchar.h>
158
159 /* BeOS 5 has the functions but no <wctype.h>. */
160 #if HAVE_WCTYPE_H
161 # include <wctype.h>
162 #endif
163 /* FreeBSD 4.4 to 4.11 has <wctype.h> but lacks the functions.
164 Assume all 12 functions are implemented the same way, or not at all. */
165 #if !defined iswalnum && !HAVE_ISWCNTRL
166 static inline int
iswalnum(wint_t wc)167 iswalnum (wint_t wc)
168 {
169 return (wc >= 0 && wc < 128
170 ? (wc >= '0' && wc <= '9') || ((wc & ~0x20) >= 'A' && (wc & ~0x20) <= 'Z')
171 : 0);
172 }
173 # define iswalnum iswalnum
174 #endif
175 #if !defined iswalpha && !HAVE_ISWCNTRL
176 static inline int
iswalpha(wint_t wc)177 iswalpha (wint_t wc)
178 {
179 return (wc >= 0 && wc < 128
180 ? (wc & ~0x20) >= 'A' && (wc & ~0x20) <= 'Z'
181 : 0);
182 }
183 # define iswalpha iswalpha
184 #endif
185 #if !defined iswblank && !HAVE_ISWCNTRL
186 static inline int
iswblank(wint_t wc)187 iswblank (wint_t wc)
188 {
189 return (wc >= 0 && wc < 128
190 ? wc == ' ' || wc == '\t'
191 : 0);
192 }
193 # define iswblank iswblank
194 #endif
195 #if !defined iswcntrl && !HAVE_ISWCNTRL
196 static inline int
iswcntrl(wint_t wc)197 iswcntrl (wint_t wc)
198 {
199 return (wc >= 0 && wc < 128
200 ? (wc & ~0x1f) == 0 || wc == 0x7f
201 : 0);
202 }
203 # define iswcntrl iswcntrl
204 #endif
205 #if !defined iswdigit && !HAVE_ISWCNTRL
206 static inline int
iswdigit(wint_t wc)207 iswdigit (wint_t wc)
208 {
209 return (wc >= '0' && wc <= '9');
210 }
211 # define iswdigit iswdigit
212 #endif
213 #if !defined iswgraph && !HAVE_ISWCNTRL
214 static inline int
iswgraph(wint_t wc)215 iswgraph (wint_t wc)
216 {
217 return (wc >= 0 && wc < 128
218 ? wc >= '!' && wc <= '~'
219 : 1);
220 }
221 # define iswgraph iswgraph
222 #endif
223 #if !defined iswlower && !HAVE_ISWCNTRL
224 static inline int
iswlower(wint_t wc)225 iswlower (wint_t wc)
226 {
227 return (wc >= 0 && wc < 128
228 ? wc >= 'a' && wc <= 'z'
229 : 0);
230 }
231 # define iswlower iswlower
232 #endif
233 #if !defined iswprint && !HAVE_ISWCNTRL
234 static inline int
iswprint(wint_t wc)235 iswprint (wint_t wc)
236 {
237 return (wc >= 0 && wc < 128
238 ? wc >= ' ' && wc <= '~'
239 : 1);
240 }
241 # define iswprint iswprint
242 #endif
243 #if !defined iswpunct && !HAVE_ISWCNTRL
244 static inline int
iswpunct(wint_t wc)245 iswpunct (wint_t wc)
246 {
247 return (wc >= 0 && wc < 128
248 ? wc >= '!' && wc <= '~'
249 && !((wc >= '0' && wc <= '9')
250 || ((wc & ~0x20) >= 'A' && (wc & ~0x20) <= 'Z'))
251 : 1);
252 }
253 # define iswpunct iswpunct
254 #endif
255 #if !defined iswspace && !HAVE_ISWCNTRL
256 static inline int
iswspace(wint_t wc)257 iswspace (wint_t wc)
258 {
259 return (wc >= 0 && wc < 128
260 ? wc == ' ' || wc == '\t'
261 || wc == '\n' || wc == '\v' || wc == '\f' || wc == '\r'
262 : 0);
263 }
264 # define iswspace iswspace
265 #endif
266 #if !defined iswupper && !HAVE_ISWCNTRL
267 static inline int
iswupper(wint_t wc)268 iswupper (wint_t wc)
269 {
270 return (wc >= 0 && wc < 128
271 ? wc >= 'A' && wc <= 'Z'
272 : 0);
273 }
274 # define iswupper iswupper
275 #endif
276 #if !defined iswxdigit && !HAVE_ISWCNTRL
277 static inline int
iswxdigit(wint_t wc)278 iswxdigit (wint_t wc)
279 {
280 return (wc >= '0' && wc <= '9') || ((wc & ~0x20) >= 'A' && (wc & ~0x20) <= 'F');
281 }
282 # define iswxdigit iswxdigit
283 #endif
284
285 #include "wcwidth.h"
286
287 #define MBCHAR_BUF_SIZE 24
288
289 struct mbchar
290 {
291 const char *ptr; /* pointer to current character */
292 size_t bytes; /* number of bytes of current character, > 0 */
293 bool wc_valid; /* true if wc is a valid wide character */
294 wchar_t wc; /* if wc_valid: the current character */
295 char buf[MBCHAR_BUF_SIZE]; /* room for the bytes, used for file input only */
296 };
297
298 /* EOF (not a real character) is represented with bytes = 0 and
299 wc_valid = false. */
300
301 typedef struct mbchar mbchar_t;
302
303 /* Access the current character. */
304 #define mb_ptr(mbc) ((mbc).ptr)
305 #define mb_len(mbc) ((mbc).bytes)
306
307 /* Comparison of characters. */
308 #define mb_iseq(mbc, sc) ((mbc).wc_valid && (mbc).wc == (sc))
309 #define mb_isnul(mbc) ((mbc).wc_valid && (mbc).wc == 0)
310 #define mb_cmp(mbc1, mbc2) \
311 ((mbc1).wc_valid \
312 ? ((mbc2).wc_valid \
313 ? (int) (mbc1).wc - (int) (mbc2).wc \
314 : -1) \
315 : ((mbc2).wc_valid \
316 ? 1 \
317 : (mbc1).bytes == (mbc2).bytes \
318 ? memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) \
319 : (mbc1).bytes < (mbc2).bytes \
320 ? (memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) > 0 ? 1 : -1) \
321 : (memcmp ((mbc1).ptr, (mbc2).ptr, (mbc2).bytes) >= 0 ? 1 : -1)))
322 #define mb_casecmp(mbc1, mbc2) \
323 ((mbc1).wc_valid \
324 ? ((mbc2).wc_valid \
325 ? (int) towlower ((mbc1).wc) - (int) towlower ((mbc2).wc) \
326 : -1) \
327 : ((mbc2).wc_valid \
328 ? 1 \
329 : (mbc1).bytes == (mbc2).bytes \
330 ? memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) \
331 : (mbc1).bytes < (mbc2).bytes \
332 ? (memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) > 0 ? 1 : -1) \
333 : (memcmp ((mbc1).ptr, (mbc2).ptr, (mbc2).bytes) >= 0 ? 1 : -1)))
334 #define mb_equal(mbc1, mbc2) \
335 ((mbc1).wc_valid && (mbc2).wc_valid \
336 ? (mbc1).wc == (mbc2).wc \
337 : (mbc1).bytes == (mbc2).bytes \
338 && memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) == 0)
339 #define mb_caseequal(mbc1, mbc2) \
340 ((mbc1).wc_valid && (mbc2).wc_valid \
341 ? towlower ((mbc1).wc) == towlower ((mbc2).wc) \
342 : (mbc1).bytes == (mbc2).bytes \
343 && memcmp ((mbc1).ptr, (mbc2).ptr, (mbc1).bytes) == 0)
344
345 /* <ctype.h>, <wctype.h> classification. */
346 #define mb_isascii(mbc) \
347 ((mbc).wc_valid && (mbc).wc >= 0 && (mbc).wc <= 127)
348 #define mb_isalnum(mbc) ((mbc).wc_valid && iswalnum ((mbc).wc))
349 #define mb_isalpha(mbc) ((mbc).wc_valid && iswalpha ((mbc).wc))
350 #define mb_isblank(mbc) ((mbc).wc_valid && iswblank ((mbc).wc))
351 #define mb_iscntrl(mbc) ((mbc).wc_valid && iswcntrl ((mbc).wc))
352 #define mb_isdigit(mbc) ((mbc).wc_valid && iswdigit ((mbc).wc))
353 #define mb_isgraph(mbc) ((mbc).wc_valid && iswgraph ((mbc).wc))
354 #define mb_islower(mbc) ((mbc).wc_valid && iswlower ((mbc).wc))
355 #define mb_isprint(mbc) ((mbc).wc_valid && iswprint ((mbc).wc))
356 #define mb_ispunct(mbc) ((mbc).wc_valid && iswpunct ((mbc).wc))
357 #define mb_isspace(mbc) ((mbc).wc_valid && iswspace ((mbc).wc))
358 #define mb_isupper(mbc) ((mbc).wc_valid && iswupper ((mbc).wc))
359 #define mb_isxdigit(mbc) ((mbc).wc_valid && iswxdigit ((mbc).wc))
360
361 /* Extra <wchar.h> function. */
362
363 /* Unprintable characters appear as a small box of width 1. */
364 #define MB_UNPRINTABLE_WIDTH 1
365
366 static inline int
mb_width_aux(wint_t wc)367 mb_width_aux (wint_t wc)
368 {
369 int w = wcwidth (wc);
370 /* For unprintable characters, arbitrarily return 0 for control characters
371 and MB_UNPRINTABLE_WIDTH otherwise. */
372 return (w >= 0 ? w : iswcntrl (wc) ? 0 : MB_UNPRINTABLE_WIDTH);
373 }
374
375 #define mb_width(mbc) \
376 ((mbc).wc_valid ? mb_width_aux ((mbc).wc) : MB_UNPRINTABLE_WIDTH)
377
378 /* Output. */
379 #define mb_putc(mbc, stream) fwrite ((mbc).ptr, 1, (mbc).bytes, (stream))
380
381 /* Assignment. */
382 #define mb_setascii(mbc, sc) \
383 ((mbc)->ptr = (mbc)->buf, (mbc)->bytes = 1, (mbc)->wc_valid = 1, \
384 (mbc)->wc = (mbc)->buf[0] = (sc))
385
386 /* Copying a character. */
387 static inline void
mb_copy(mbchar_t * new_mbc,const mbchar_t * old_mbc)388 mb_copy (mbchar_t *new_mbc, const mbchar_t *old_mbc)
389 {
390 if (old_mbc->ptr == &old_mbc->buf[0])
391 {
392 memcpy (&new_mbc->buf[0], &old_mbc->buf[0], old_mbc->bytes);
393 new_mbc->ptr = &new_mbc->buf[0];
394 }
395 else
396 new_mbc->ptr = old_mbc->ptr;
397 new_mbc->bytes = old_mbc->bytes;
398 if ((new_mbc->wc_valid = old_mbc->wc_valid))
399 new_mbc->wc = old_mbc->wc;
400 }
401
402
403 /* is_basic(c) tests whether the single-byte character c is in the
404 ISO C "basic character set".
405 This is a convenience function, and is in this file only to share code
406 between mbiter_multi.h and mbfile_multi.h. */
407 #if (' ' == 32) && ('!' == 33) && ('"' == 34) && ('#' == 35) \
408 && ('%' == 37) && ('&' == 38) && ('\'' == 39) && ('(' == 40) \
409 && (')' == 41) && ('*' == 42) && ('+' == 43) && (',' == 44) \
410 && ('-' == 45) && ('.' == 46) && ('/' == 47) && ('0' == 48) \
411 && ('1' == 49) && ('2' == 50) && ('3' == 51) && ('4' == 52) \
412 && ('5' == 53) && ('6' == 54) && ('7' == 55) && ('8' == 56) \
413 && ('9' == 57) && (':' == 58) && (';' == 59) && ('<' == 60) \
414 && ('=' == 61) && ('>' == 62) && ('?' == 63) && ('A' == 65) \
415 && ('B' == 66) && ('C' == 67) && ('D' == 68) && ('E' == 69) \
416 && ('F' == 70) && ('G' == 71) && ('H' == 72) && ('I' == 73) \
417 && ('J' == 74) && ('K' == 75) && ('L' == 76) && ('M' == 77) \
418 && ('N' == 78) && ('O' == 79) && ('P' == 80) && ('Q' == 81) \
419 && ('R' == 82) && ('S' == 83) && ('T' == 84) && ('U' == 85) \
420 && ('V' == 86) && ('W' == 87) && ('X' == 88) && ('Y' == 89) \
421 && ('Z' == 90) && ('[' == 91) && ('\\' == 92) && (']' == 93) \
422 && ('^' == 94) && ('_' == 95) && ('a' == 97) && ('b' == 98) \
423 && ('c' == 99) && ('d' == 100) && ('e' == 101) && ('f' == 102) \
424 && ('g' == 103) && ('h' == 104) && ('i' == 105) && ('j' == 106) \
425 && ('k' == 107) && ('l' == 108) && ('m' == 109) && ('n' == 110) \
426 && ('o' == 111) && ('p' == 112) && ('q' == 113) && ('r' == 114) \
427 && ('s' == 115) && ('t' == 116) && ('u' == 117) && ('v' == 118) \
428 && ('w' == 119) && ('x' == 120) && ('y' == 121) && ('z' == 122) \
429 && ('{' == 123) && ('|' == 124) && ('}' == 125) && ('~' == 126)
430 /* The character set is ISO-646, not EBCDIC. */
431 # define IS_BASIC_ASCII 1
432
433 extern const unsigned int is_basic_table[];
434
435 static inline bool
is_basic(char c)436 is_basic (char c)
437 {
438 return (is_basic_table [(unsigned char) c >> 5] >> ((unsigned char) c & 31))
439 & 1;
440 }
441
442 #else
443
444 static inline bool
is_basic(char c)445 is_basic (char c)
446 {
447 switch (c)
448 {
449 case '\t': case '\v': case '\f':
450 case ' ': case '!': case '"': case '#': case '%':
451 case '&': case '\'': case '(': case ')': case '*':
452 case '+': case ',': case '-': case '.': case '/':
453 case '0': case '1': case '2': case '3': case '4':
454 case '5': case '6': case '7': case '8': case '9':
455 case ':': case ';': case '<': case '=': case '>':
456 case '?':
457 case 'A': case 'B': case 'C': case 'D': case 'E':
458 case 'F': case 'G': case 'H': case 'I': case 'J':
459 case 'K': case 'L': case 'M': case 'N': case 'O':
460 case 'P': case 'Q': case 'R': case 'S': case 'T':
461 case 'U': case 'V': case 'W': case 'X': case 'Y':
462 case 'Z':
463 case '[': case '\\': case ']': case '^': case '_':
464 case 'a': case 'b': case 'c': case 'd': case 'e':
465 case 'f': case 'g': case 'h': case 'i': case 'j':
466 case 'k': case 'l': case 'm': case 'n': case 'o':
467 case 'p': case 'q': case 'r': case 's': case 't':
468 case 'u': case 'v': case 'w': case 'x': case 'y':
469 case 'z': case '{': case '|': case '}': case '~':
470 return 1;
471 default:
472 return 0;
473 }
474 }
475
476 #endif
477
478 #endif /* _MBCHAR_H */
479