From fa832504c217fa68b9376c1969a7db95f6e8d24e Mon Sep 17 00:00:00 2001
From: nyamatongwe <unknown>
Date: Wed, 21 Mar 2012 23:39:03 +1100
Subject: Ensure segment discovery always makes progress even for invalid
 UTF-8.

---
 src/Document.cxx | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

(limited to 'src')

diff --git a/src/Document.cxx b/src/Document.cxx
index fd71a1363..fae97856a 100644
--- a/src/Document.cxx
+++ b/src/Document.cxx
@@ -721,7 +721,12 @@ int Document::SafeSegment(const char *text, int length, int lengthSegment) {
 		lastEncodingAllowedBreak = j;
 
 		if (dbcsCodePage == SC_CP_UTF8) {
-			j += (ch < 0x80) ? 1 : BytesFromLead(ch);
+			if (ch < 0x80) {
+				j++;
+			} else {
+				int bytes = BytesFromLead(ch);
+				j += bytes ? bytes : 1;
+			}
 		} else if (dbcsCodePage) {
 			j += IsDBCSLeadByte(ch) ? 2 : 1;
 		} else {
-- 
cgit v1.2.3