From b8a102d26f41e57b94359627a4df8f22af10c6fa Mon Sep 17 00:00:00 2001
From: nicm <nicm>
Date: Wed, 2 Mar 2016 15:36:02 +0000
Subject: Handle wcwidth() and mbtowc() failures in better style and drop
 characters where we can't find the width (wcwidth() fails) on input, the same
 as we drop invalid UTF-8. Suggested by schwarze@.

---
 input.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

(limited to 'input.c')

diff --git a/input.c b/input.c
index ae205024..18c8eb9a 100644
--- a/input.c
+++ b/input.c
@@ -1960,8 +1960,14 @@ input_utf8_close(struct input_ctx *ictx)
 {
 	struct utf8_data	*ud = &ictx->utf8data;
 
-	if (utf8_append(ud, ictx->ch) != UTF8_DONE)
-		fatalx("UTF-8 close invalid %#x", ictx->ch);
+	if (utf8_append(ud, ictx->ch) != UTF8_DONE) {
+		/*
+		 * An error here could be invalid UTF-8 or it could be a
+		 * nonprintable character for which we can't get the
+		 * width. Drop it.
+		 */
+		return (0);
+	}
 
 	log_debug("%s %hhu '%*s' (width %hhu)", __func__, ud->size,
 	    (int)ud->size, ud->data, ud->width);
-- 
cgit