| 1 | /* mbutil.c -- readline multibyte character utility functions */ |
| 2 | |
| 3 | /* Copyright (C) 2001-2005 Free Software Foundation, Inc. |
| 4 | |
| 5 | This file is part of the GNU Readline Library, a library for |
| 6 | reading lines of text with interactive input and history editing. |
| 7 | |
| 8 | The GNU Readline Library is free software; you can redistribute it |
| 9 | and/or modify it under the terms of the GNU General Public License |
| 10 | as published by the Free Software Foundation; either version 2, or |
| 11 | (at your option) any later version. |
| 12 | |
| 13 | The GNU Readline Library is distributed in the hope that it will be |
| 14 | useful, but WITHOUT ANY WARRANTY; without even the implied warranty |
| 15 | of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 16 | GNU General Public License for more details. |
| 17 | |
| 18 | The GNU General Public License is often shipped with GNU software, and |
| 19 | is generally kept in a file called COPYING or LICENSE. If you do not |
| 20 | have a copy of the license, write to the Free Software Foundation, |
| 21 | 59 Temple Place, Suite 330, Boston, MA 02111 USA. */ |
| 22 | #define READLINE_LIBRARY |
| 23 | |
| 24 | #if defined (HAVE_CONFIG_H) |
| 25 | # include <config.h> |
| 26 | #endif |
| 27 | |
| 28 | #include <sys/types.h> |
| 29 | #include <fcntl.h> |
| 30 | #include "posixjmp.h" |
| 31 | |
| 32 | #if defined (HAVE_UNISTD_H) |
| 33 | # include <unistd.h> /* for _POSIX_VERSION */ |
| 34 | #endif /* HAVE_UNISTD_H */ |
| 35 | |
| 36 | #if defined (HAVE_STDLIB_H) |
| 37 | # include <stdlib.h> |
| 38 | #else |
| 39 | # include "ansi_stdlib.h" |
| 40 | #endif /* HAVE_STDLIB_H */ |
| 41 | |
| 42 | #include <stdio.h> |
| 43 | #include <ctype.h> |
| 44 | |
| 45 | /* System-specific feature definitions and include files. */ |
| 46 | #include "rldefs.h" |
| 47 | #include "rlmbutil.h" |
| 48 | |
| 49 | #if defined (TIOCSTAT_IN_SYS_IOCTL) |
| 50 | # include <sys/ioctl.h> |
| 51 | #endif /* TIOCSTAT_IN_SYS_IOCTL */ |
| 52 | |
| 53 | /* Some standard library routines. */ |
| 54 | #include "readline.h" |
| 55 | |
| 56 | #include "rlprivate.h" |
| 57 | #include "xmalloc.h" |
| 58 | |
| 59 | /* Declared here so it can be shared between the readline and history |
| 60 | libraries. */ |
| 61 | #if defined (HANDLE_MULTIBYTE) |
| 62 | int rl_byte_oriented = 0; |
| 63 | #else |
| 64 | int rl_byte_oriented = 1; |
| 65 | #endif |
| 66 | |
| 67 | /* **************************************************************** */ |
| 68 | /* */ |
| 69 | /* Multibyte Character Utility Functions */ |
| 70 | /* */ |
| 71 | /* **************************************************************** */ |
| 72 | |
| 73 | #if defined(HANDLE_MULTIBYTE) |
| 74 | |
| 75 | static int |
| 76 | _rl_find_next_mbchar_internal (string, seed, count, find_non_zero) |
| 77 | char *string; |
| 78 | int seed, count, find_non_zero; |
| 79 | { |
| 80 | size_t tmp; |
| 81 | mbstate_t ps; |
| 82 | int point; |
| 83 | wchar_t wc; |
| 84 | |
| 85 | tmp = 0; |
| 86 | |
| 87 | memset(&ps, 0, sizeof (mbstate_t)); |
| 88 | if (seed < 0) |
| 89 | seed = 0; |
| 90 | if (count <= 0) |
| 91 | return seed; |
| 92 | |
| 93 | point = seed + _rl_adjust_point (string, seed, &ps); |
| 94 | /* if this is true, means that seed was not pointed character |
| 95 | started byte. So correct the point and consume count */ |
| 96 | if (seed < point) |
| 97 | count--; |
| 98 | |
| 99 | while (count > 0) |
| 100 | { |
| 101 | tmp = mbrtowc (&wc, string+point, strlen(string + point), &ps); |
| 102 | if (MB_INVALIDCH ((size_t)tmp)) |
| 103 | { |
| 104 | /* invalid bytes. asume a byte represents a character */ |
| 105 | point++; |
| 106 | count--; |
| 107 | /* reset states. */ |
| 108 | memset(&ps, 0, sizeof(mbstate_t)); |
| 109 | } |
| 110 | else if (MB_NULLWCH (tmp)) |
| 111 | break; /* found wide '\0' */ |
| 112 | else |
| 113 | { |
| 114 | /* valid bytes */ |
| 115 | point += tmp; |
| 116 | if (find_non_zero) |
| 117 | { |
| 118 | if (wcwidth (wc) == 0) |
| 119 | continue; |
| 120 | else |
| 121 | count--; |
| 122 | } |
| 123 | else |
| 124 | count--; |
| 125 | } |
| 126 | } |
| 127 | |
| 128 | if (find_non_zero) |
| 129 | { |
| 130 | tmp = mbrtowc (&wc, string + point, strlen (string + point), &ps); |
| 131 | while (tmp > 0 && wcwidth (wc) == 0) |
| 132 | { |
| 133 | point += tmp; |
| 134 | tmp = mbrtowc (&wc, string + point, strlen (string + point), &ps); |
| 135 | if (MB_NULLWCH (tmp) || MB_INVALIDCH (tmp)) |
| 136 | break; |
| 137 | } |
| 138 | } |
| 139 | |
| 140 | return point; |
| 141 | } |
| 142 | |
| 143 | static int |
| 144 | _rl_find_prev_mbchar_internal (string, seed, find_non_zero) |
| 145 | char *string; |
| 146 | int seed, find_non_zero; |
| 147 | { |
| 148 | mbstate_t ps; |
| 149 | int prev, non_zero_prev, point, length; |
| 150 | size_t tmp; |
| 151 | wchar_t wc; |
| 152 | |
| 153 | memset(&ps, 0, sizeof(mbstate_t)); |
| 154 | length = strlen(string); |
| 155 | |
| 156 | if (seed < 0) |
| 157 | return 0; |
| 158 | else if (length < seed) |
| 159 | return length; |
| 160 | |
| 161 | prev = non_zero_prev = point = 0; |
| 162 | while (point < seed) |
| 163 | { |
| 164 | tmp = mbrtowc (&wc, string + point, length - point, &ps); |
| 165 | if (MB_INVALIDCH ((size_t)tmp)) |
| 166 | { |
| 167 | /* in this case, bytes are invalid or shorted to compose |
| 168 | multibyte char, so assume that the first byte represents |
| 169 | a single character anyway. */ |
| 170 | tmp = 1; |
| 171 | /* clear the state of the byte sequence, because |
| 172 | in this case effect of mbstate is undefined */ |
| 173 | memset(&ps, 0, sizeof (mbstate_t)); |
| 174 | |
| 175 | /* Since we're assuming that this byte represents a single |
| 176 | non-zero-width character, don't forget about it. */ |
| 177 | prev = point; |
| 178 | } |
| 179 | else if (MB_NULLWCH (tmp)) |
| 180 | break; /* Found '\0' char. Can this happen? */ |
| 181 | else |
| 182 | { |
| 183 | if (find_non_zero) |
| 184 | { |
| 185 | if (wcwidth (wc) != 0) |
| 186 | prev = point; |
| 187 | } |
| 188 | else |
| 189 | prev = point; |
| 190 | } |
| 191 | |
| 192 | point += tmp; |
| 193 | } |
| 194 | |
| 195 | return prev; |
| 196 | } |
| 197 | |
| 198 | /* return the number of bytes parsed from the multibyte sequence starting |
| 199 | at src, if a non-L'\0' wide character was recognized. It returns 0, |
| 200 | if a L'\0' wide character was recognized. It returns (size_t)(-1), |
| 201 | if an invalid multibyte sequence was encountered. It returns (size_t)(-2) |
| 202 | if it couldn't parse a complete multibyte character. */ |
| 203 | int |
| 204 | _rl_get_char_len (src, ps) |
| 205 | char *src; |
| 206 | mbstate_t *ps; |
| 207 | { |
| 208 | size_t tmp; |
| 209 | |
| 210 | tmp = mbrlen((const char *)src, (size_t)strlen (src), ps); |
| 211 | if (tmp == (size_t)(-2)) |
| 212 | { |
| 213 | /* shorted to compose multibyte char */ |
| 214 | if (ps) |
| 215 | memset (ps, 0, sizeof(mbstate_t)); |
| 216 | return -2; |
| 217 | } |
| 218 | else if (tmp == (size_t)(-1)) |
| 219 | { |
| 220 | /* invalid to compose multibyte char */ |
| 221 | /* initialize the conversion state */ |
| 222 | if (ps) |
| 223 | memset (ps, 0, sizeof(mbstate_t)); |
| 224 | return -1; |
| 225 | } |
| 226 | else if (tmp == (size_t)0) |
| 227 | return 0; |
| 228 | else |
| 229 | return (int)tmp; |
| 230 | } |
| 231 | |
| 232 | /* compare the specified two characters. If the characters matched, |
| 233 | return 1. Otherwise return 0. */ |
| 234 | int |
| 235 | _rl_compare_chars (buf1, pos1, ps1, buf2, pos2, ps2) |
| 236 | char *buf1; |
| 237 | int pos1; |
| 238 | mbstate_t *ps1; |
| 239 | char *buf2; |
| 240 | int pos2; |
| 241 | mbstate_t *ps2; |
| 242 | { |
| 243 | int i, w1, w2; |
| 244 | |
| 245 | if ((w1 = _rl_get_char_len (&buf1[pos1], ps1)) <= 0 || |
| 246 | (w2 = _rl_get_char_len (&buf2[pos2], ps2)) <= 0 || |
| 247 | (w1 != w2) || |
| 248 | (buf1[pos1] != buf2[pos2])) |
| 249 | return 0; |
| 250 | |
| 251 | for (i = 1; i < w1; i++) |
| 252 | if (buf1[pos1+i] != buf2[pos2+i]) |
| 253 | return 0; |
| 254 | |
| 255 | return 1; |
| 256 | } |
| 257 | |
| 258 | /* adjust pointed byte and find mbstate of the point of string. |
| 259 | adjusted point will be point <= adjusted_point, and returns |
| 260 | differences of the byte(adjusted_point - point). |
| 261 | if point is invalied (point < 0 || more than string length), |
| 262 | it returns -1 */ |
| 263 | int |
| 264 | _rl_adjust_point(string, point, ps) |
| 265 | char *string; |
| 266 | int point; |
| 267 | mbstate_t *ps; |
| 268 | { |
| 269 | size_t tmp = 0; |
| 270 | int length; |
| 271 | int pos = 0; |
| 272 | |
| 273 | length = strlen(string); |
| 274 | if (point < 0) |
| 275 | return -1; |
| 276 | if (length < point) |
| 277 | return -1; |
| 278 | |
| 279 | while (pos < point) |
| 280 | { |
| 281 | tmp = mbrlen (string + pos, length - pos, ps); |
| 282 | if (MB_INVALIDCH ((size_t)tmp)) |
| 283 | { |
| 284 | /* in this case, bytes are invalid or shorted to compose |
| 285 | multibyte char, so assume that the first byte represents |
| 286 | a single character anyway. */ |
| 287 | pos++; |
| 288 | /* clear the state of the byte sequence, because |
| 289 | in this case effect of mbstate is undefined */ |
| 290 | if (ps) |
| 291 | memset (ps, 0, sizeof (mbstate_t)); |
| 292 | } |
| 293 | else if (MB_NULLWCH (tmp)) |
| 294 | pos++; |
| 295 | else |
| 296 | pos += tmp; |
| 297 | } |
| 298 | |
| 299 | return (pos - point); |
| 300 | } |
| 301 | |
| 302 | int |
| 303 | _rl_is_mbchar_matched (string, seed, end, mbchar, length) |
| 304 | char *string; |
| 305 | int seed, end; |
| 306 | char *mbchar; |
| 307 | int length; |
| 308 | { |
| 309 | int i; |
| 310 | |
| 311 | if ((end - seed) < length) |
| 312 | return 0; |
| 313 | |
| 314 | for (i = 0; i < length; i++) |
| 315 | if (string[seed + i] != mbchar[i]) |
| 316 | return 0; |
| 317 | return 1; |
| 318 | } |
| 319 | |
| 320 | wchar_t |
| 321 | _rl_char_value (buf, ind) |
| 322 | char *buf; |
| 323 | int ind; |
| 324 | { |
| 325 | size_t tmp; |
| 326 | wchar_t wc; |
| 327 | mbstate_t ps; |
| 328 | int l; |
| 329 | |
| 330 | if (MB_LEN_MAX == 1 || rl_byte_oriented) |
| 331 | return ((wchar_t) buf[ind]); |
| 332 | l = strlen (buf); |
| 333 | if (ind >= l - 1) |
| 334 | return ((wchar_t) buf[ind]); |
| 335 | memset (&ps, 0, sizeof (mbstate_t)); |
| 336 | tmp = mbrtowc (&wc, buf + ind, l - ind, &ps); |
| 337 | if (MB_INVALIDCH (tmp) || MB_NULLWCH (tmp)) |
| 338 | return ((wchar_t) buf[ind]); |
| 339 | return wc; |
| 340 | } |
| 341 | #endif /* HANDLE_MULTIBYTE */ |
| 342 | |
| 343 | /* Find next `count' characters started byte point of the specified seed. |
| 344 | If flags is MB_FIND_NONZERO, we look for non-zero-width multibyte |
| 345 | characters. */ |
| 346 | #undef _rl_find_next_mbchar |
| 347 | int |
| 348 | _rl_find_next_mbchar (string, seed, count, flags) |
| 349 | char *string; |
| 350 | int seed, count, flags; |
| 351 | { |
| 352 | #if defined (HANDLE_MULTIBYTE) |
| 353 | return _rl_find_next_mbchar_internal (string, seed, count, flags); |
| 354 | #else |
| 355 | return (seed + count); |
| 356 | #endif |
| 357 | } |
| 358 | |
| 359 | /* Find previous character started byte point of the specified seed. |
| 360 | Returned point will be point <= seed. If flags is MB_FIND_NONZERO, |
| 361 | we look for non-zero-width multibyte characters. */ |
| 362 | #undef _rl_find_prev_mbchar |
| 363 | int |
| 364 | _rl_find_prev_mbchar (string, seed, flags) |
| 365 | char *string; |
| 366 | int seed, flags; |
| 367 | { |
| 368 | #if defined (HANDLE_MULTIBYTE) |
| 369 | return _rl_find_prev_mbchar_internal (string, seed, flags); |
| 370 | #else |
| 371 | return ((seed == 0) ? seed : seed - 1); |
| 372 | #endif |
| 373 | } |