1 /* mbutil.c -- readline multibyte character utility functions */
3 /* Copyright (C) 2001-2017 Free Software Foundation, Inc.
5 This file is part of the GNU Readline Library (Readline), a library
6 for reading lines of text with interactive input and history editing.
8 Readline is free software: you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation, either version 3 of the License, or
11 (at your option) any later version.
13 Readline is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with Readline. If not, see <http://www.gnu.org/licenses/>.
22 #define READLINE_LIBRARY
24 #if defined (HAVE_CONFIG_H)
28 #include <sys/types.h>
32 #if defined (HAVE_UNISTD_H)
33 # include <unistd.h> /* for _POSIX_VERSION */
34 #endif /* HAVE_UNISTD_H */
36 #if defined (HAVE_STDLIB_H)
39 # include "ansi_stdlib.h"
40 #endif /* HAVE_STDLIB_H */
45 /* System-specific feature definitions and include files. */
49 #if defined (TIOCSTAT_IN_SYS_IOCTL)
50 # include <sys/ioctl.h>
51 #endif /* TIOCSTAT_IN_SYS_IOCTL */
53 /* Some standard library routines. */
56 #include "rlprivate.h"
59 /* Declared here so it can be shared between the readline and history
61 #if defined (HANDLE_MULTIBYTE)
62 int rl_byte_oriented = 0;
64 int rl_byte_oriented = 1;
68 int _rl_utf8locale = 0;
70 /* **************************************************************** */
72 /* Multibyte Character Utility Functions */
74 /* **************************************************************** */
76 #if defined(HANDLE_MULTIBYTE)
78 /* **************************************************************** */
80 /* UTF-8 specific Character Utility Functions */
82 /* **************************************************************** */
84 /* Return the length in bytes of the possibly-multibyte character beginning
85 at S. Encoding is UTF-8. */
87 _rl_utf8_mblen (const char *s, size_t n)
92 return (0); /* no shift states */
96 c = (unsigned char)*s;
101 c1 = (unsigned char)s[1];
104 if (n >= 2 && (s[1] ^ 0x80) < 0x40)
110 && (s[1] ^ 0x80) < 0x40 && (s[2] ^ 0x80) < 0x40
111 && (c >= 0xe1 || c1 >= 0xa0)
112 && (c != 0xed || c1 < 0xa0))
118 && (s[1] ^ 0x80) < 0x40 && (s[2] ^ 0x80) < 0x40
119 && (s[3] ^ 0x80) < 0x40
120 && (c >= 0xf1 || c1 >= 0x90)
121 && (c < 0xf4 || (c == 0xf4 && c1 < 0x90)))
125 /* invalid or incomplete multibyte character */
130 _rl_find_next_mbchar_internal (char *string, int seed, int count, int find_non_zero)
139 memset(&ps, 0, sizeof (mbstate_t));
145 point = seed + _rl_adjust_point (string, seed, &ps);
146 /* if _rl_adjust_point returns -1, the character or string is invalid.
148 if (point == seed - 1) /* invalid */
151 /* if this is true, means that seed was not pointing to a byte indicating
152 the beginning of a multibyte character. Correct the point and consume
159 len = strlen (string + point);
162 if (_rl_utf8locale && UTF8_SINGLEBYTE(string[point]))
165 wc = (wchar_t) string[point];
166 memset(&ps, 0, sizeof(mbstate_t));
169 tmp = mbrtowc (&wc, string+point, len, &ps);
170 if (MB_INVALIDCH ((size_t)tmp))
172 /* invalid bytes. assume a byte represents a character */
176 memset(&ps, 0, sizeof(mbstate_t));
178 else if (MB_NULLWCH (tmp))
179 break; /* found wide '\0' */
186 if (WCWIDTH (wc) == 0)
198 tmp = mbrtowc (&wc, string + point, strlen (string + point), &ps);
199 while (MB_NULLWCH (tmp) == 0 && MB_INVALIDCH (tmp) == 0 && WCWIDTH (wc) == 0)
202 tmp = mbrtowc (&wc, string + point, strlen (string + point), &ps);
210 _rl_find_prev_mbchar_internal (char *string, int seed, int find_non_zero)
213 int prev, non_zero_prev, point, length;
217 memset(&ps, 0, sizeof(mbstate_t));
218 length = strlen(string);
222 else if (length < seed)
225 prev = non_zero_prev = point = 0;
228 if (_rl_utf8locale && UTF8_SINGLEBYTE(string[point]))
231 wc = (wchar_t) string[point];
232 memset(&ps, 0, sizeof(mbstate_t));
235 tmp = mbrtowc (&wc, string + point, length - point, &ps);
236 if (MB_INVALIDCH ((size_t)tmp))
238 /* in this case, bytes are invalid or too short to compose
239 multibyte char, so assume that the first byte represents
240 a single character anyway. */
242 /* clear the state of the byte sequence, because
243 in this case effect of mbstate is undefined */
244 memset(&ps, 0, sizeof (mbstate_t));
246 /* Since we're assuming that this byte represents a single
247 non-zero-width character, don't forget about it. */
250 else if (MB_NULLWCH (tmp))
251 break; /* Found '\0' char. Can this happen? */
256 if (WCWIDTH (wc) != 0)
269 /* return the number of bytes parsed from the multibyte sequence starting
270 at src, if a non-L'\0' wide character was recognized. It returns 0,
271 if a L'\0' wide character was recognized. It returns (size_t)(-1),
272 if an invalid multibyte sequence was encountered. It returns (size_t)(-2)
273 if it couldn't parse a complete multibyte character. */
275 _rl_get_char_len (char *src, mbstate_t *ps)
280 /* Look at no more than MB_CUR_MAX characters */
281 l = (size_t)strlen (src);
282 if (_rl_utf8locale && l > 0 && UTF8_SINGLEBYTE(*src))
283 tmp = (*src != 0) ? 1 : 0;
286 mb_cur_max = MB_CUR_MAX;
287 tmp = mbrlen((const char *)src, (l < mb_cur_max) ? l : mb_cur_max, ps);
289 if (tmp == (size_t)(-2))
291 /* too short to compose multibyte char */
293 memset (ps, 0, sizeof(mbstate_t));
296 else if (tmp == (size_t)(-1))
298 /* invalid to compose multibyte char */
299 /* initialize the conversion state */
301 memset (ps, 0, sizeof(mbstate_t));
304 else if (tmp == (size_t)0)
310 /* compare the specified two characters. If the characters matched,
311 return 1. Otherwise return 0. */
313 _rl_compare_chars (char *buf1, int pos1, mbstate_t *ps1, char *buf2, int pos2, mbstate_t *ps2)
317 if ((w1 = _rl_get_char_len (&buf1[pos1], ps1)) <= 0 ||
318 (w2 = _rl_get_char_len (&buf2[pos2], ps2)) <= 0 ||
320 (buf1[pos1] != buf2[pos2]))
323 for (i = 1; i < w1; i++)
324 if (buf1[pos1+i] != buf2[pos2+i])
330 /* adjust pointed byte and find mbstate of the point of string.
331 adjusted point will be point <= adjusted_point, and returns
332 differences of the byte(adjusted_point - point).
333 if point is invalid (point < 0 || more than string length),
336 _rl_adjust_point (char *string, int point, mbstate_t *ps)
343 length = strlen(string);
351 if (_rl_utf8locale && UTF8_SINGLEBYTE(string[pos]))
354 tmp = mbrlen (string + pos, length - pos, ps);
355 if (MB_INVALIDCH ((size_t)tmp))
357 /* in this case, bytes are invalid or too short to compose
358 multibyte char, so assume that the first byte represents
359 a single character anyway. */
361 /* clear the state of the byte sequence, because
362 in this case effect of mbstate is undefined */
364 memset (ps, 0, sizeof (mbstate_t));
366 else if (MB_NULLWCH (tmp))
372 return (pos - point);
376 _rl_is_mbchar_matched (char *string, int seed, int end, char *mbchar, int length)
380 if ((end - seed) < length)
383 for (i = 0; i < length; i++)
384 if (string[seed + i] != mbchar[i])
390 _rl_char_value (char *buf, int ind)
397 if (MB_LEN_MAX == 1 || rl_byte_oriented)
398 return ((wchar_t) buf[ind]);
399 if (_rl_utf8locale && UTF8_SINGLEBYTE(buf[ind]))
400 return ((wchar_t) buf[ind]);
403 return ((wchar_t) buf[ind]);
404 if (l < ind) /* Sanity check */
405 l = strlen (buf+ind);
406 memset (&ps, 0, sizeof (mbstate_t));
407 tmp = mbrtowc (&wc, buf + ind, l - ind, &ps);
408 if (MB_INVALIDCH (tmp) || MB_NULLWCH (tmp))
409 return ((wchar_t) buf[ind]);
412 #endif /* HANDLE_MULTIBYTE */
414 /* Find next `count' characters started byte point of the specified seed.
415 If flags is MB_FIND_NONZERO, we look for non-zero-width multibyte
417 #undef _rl_find_next_mbchar
419 _rl_find_next_mbchar (char *string, int seed, int count, int flags)
421 #if defined (HANDLE_MULTIBYTE)
422 return _rl_find_next_mbchar_internal (string, seed, count, flags);
424 return (seed + count);
428 /* Find previous character started byte point of the specified seed.
429 Returned point will be point <= seed. If flags is MB_FIND_NONZERO,
430 we look for non-zero-width multibyte characters. */
431 #undef _rl_find_prev_mbchar
433 _rl_find_prev_mbchar (char *string, int seed, int flags)
435 #if defined (HANDLE_MULTIBYTE)
436 return _rl_find_prev_mbchar_internal (string, seed, flags);
438 return ((seed == 0) ? seed : seed - 1);