diff options
Diffstat (limited to 'utf8.c')
| -rw-r--r-- | utf8.c | 27 |
1 files changed, 26 insertions, 1 deletions
@@ -22,7 +22,6 @@ #include <errno.h> #include <stdlib.h> #include <string.h> -#include <vis.h> #include <wchar.h> #include "tmux.h" @@ -111,9 +110,27 @@ utf8_width(wchar_t wc) { int width; +#ifdef HAVE_UTF8PROC + width = utf8proc_wcwidth(wc); +#else width = wcwidth(wc); +#endif if (width < 0 || width > 0xff) { log_debug("Unicode %04lx, wcwidth() %d", (long)wc, width); + +#ifndef __OpenBSD__ + /* + * Many platforms (particularly and inevitably OS X) have no + * width for relatively common characters (wcwidth() returns + * -1); assume width 1 in this case. This will be wrong for + * genuinely nonprintable characters, but they should be + * rare. We may pass through stuff that ideally we would block, + * but this is no worse than sending the same to the terminal + * without tmux. + */ + if (width < 0) + return (1); +#endif return (-1); } return (width); @@ -123,7 +140,11 @@ utf8_width(wchar_t wc) enum utf8_state utf8_combine(const struct utf8_data *ud, wchar_t *wc) { +#ifdef HAVE_UTF8PROC + switch (utf8proc_mbtowc(wc, ud->data, ud->size)) { +#else switch (mbtowc(wc, ud->data, ud->size)) { +#endif case -1: log_debug("UTF-8 %.*s, mbtowc() %d", (int)ud->size, ud->data, errno); @@ -143,7 +164,11 @@ utf8_split(wchar_t wc, struct utf8_data *ud) char s[MB_LEN_MAX]; int slen; +#ifdef HAVE_UTF8PROC + slen = utf8proc_wctomb(s, wc); +#else slen = wctomb(s, wc); +#endif if (slen <= 0 || slen > (int)sizeof ud->data) return (UTF8_ERROR); |