xref: /openbsd/usr.bin/rs/utf8.c (revision 2450f62a)
1 /*
2  * Copyright (c) 2015 Ingo Schwarze <schwarze@openbsd.org>
3  *
4  * Permission to use, copy, modify, and distribute this software for any
5  * purpose with or without fee is hereby granted, provided that the above
6  * copyright notice and this permission notice appear in all copies.
7  *
8  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
15  */
16 
17 #include <err.h>
18 #include <stdlib.h>
19 #include <string.h>
20 #include <wchar.h>
21 
22 int
mbsavis(char ** outp,const char * mbs)23 mbsavis(char** outp, const char *mbs)
24 {
25 	const char *src;  /* Iterate mbs. */
26 	char	 *dst;  /* Iterate *outp. */
27 	wchar_t	  wc;
28 	int	  total_width;  /* Display width of the whole string. */
29 	int	  width;  /* Display width of a single Unicode char. */
30 	int	  len;  /* Length in bytes of UTF-8 encoded string. */
31 
32 	len = strlen(mbs);
33 	if ((*outp = malloc(len + 1)) == NULL)
34 		err(1, NULL);
35 
36 	if (MB_CUR_MAX == 1) {
37 		memcpy(*outp, mbs, len + 1);
38 		return len;
39 	}
40 
41 	src = mbs;
42 	dst = *outp;
43 	total_width = 0;
44 	while (*src != '\0') {
45 		if ((len = mbtowc(&wc, src, MB_CUR_MAX)) == -1) {
46 			total_width++;
47 			*dst++ = '?';
48 			src++;
49 		} else if ((width = wcwidth(wc)) == -1) {
50 			total_width++;
51 			*dst++ = '?';
52 			src += len;
53 		} else {
54 			total_width += width;
55 			while (len-- > 0)
56 				*dst++ = *src++;
57 		}
58 	}
59 	*dst = '\0';
60 	return total_width;
61 }
62