Restore code to handle wcwidth failure so that unknown codepoints still

do the most likely right thing. GitHub issue 3427, patch based on an
diff from Jesse Luehrs in GitHub issue 3003.

1 files changed, 10 insertions, 2 deletions

diff --git a/utf8.c b/utf8.c
index 05ab9cfe..042ddf89 100644
--- a/utf8.c
+++ b/utf8.c
@@ -229,13 +229,21 @@ utf8_width(struct utf8_data *ud, int *width)
 	case 0:
 		return (UTF8_ERROR);
 	}
+	log_debug("UTF-8 %.*s is %08X", (int)ud->size, ud->data, (u_int)wc);
 #ifdef HAVE_UTF8PROC
 	*width = utf8proc_wcwidth(wc);
+	log_debug("utf8proc_wcwidth(%08X) returned %d", (u_int)wc, *width);
 #else
 	*width = wcwidth(wc);
+	log_debug("wcwidth(%08X) returned %d", (u_int)wc, *width);
+	if (*width < 0) {
+		/*
+		 * C1 control characters are nonprintable, so they are always
+		 * zero width.
+		 */
+		*width = (wc >= 0x80 && wc <= 0x9f) ? 0 : 1;
+	}
 #endif
-	log_debug("UTF-8 %.*s %#x, wcwidth() %d", (int)ud->size, ud->data,
-	    (u_int)wc, *width);
 	if (*width >= 0 && *width <= 0xff)
 		return (UTF8_DONE);
 	return (UTF8_ERROR);