From: Victor Wagner <vitus@wagner.pp.ru>
Date: Tue, 17 Oct 2006 18:07:06 +0000 (+0000)
Subject: Fix bug #1 - avoid broken UTF-8 at line breaks
X-Git-Tag: REL_0_95~21
X-Git-Url: http://wagner.pp.ru/gitweb/?a=commitdiff_plain;h=c83cd8895214b3d0a6481123a7f3f126bb9f0a06;p=oss%2Fcatdoc.git

Fix bug #1 - avoid broken UTF-8 at line breaks
---

diff --git a/src/writer.c b/src/writer.c
index 5448d6c..5f95ba2 100644
--- a/src/writer.c
+++ b/src/writer.c
@@ -54,12 +54,18 @@ void out_char(const char *chunk) {
 	} else if (bufpos>wrap_margin) {
 		char *q=outputbuffer,*p=outputbuffer+wrap_margin;
 		
-		while (p>outputbuffer&&!isspace(*p)) p--;
+		while (p>outputbuffer&&*p!=' '&& *p!='\t') p--;
 		if (p==outputbuffer) {
 			/*worst case - nowhere to wrap. Will use brute force */
-			fwrite(outputbuffer,wrap_margin,1,stdout);
+			int i = wrap_margin;
+			if (from_unicode == to_utf8) {
+				/* go back to start of nearest utf-8 character */
+				while(i>0 && (outputbuffer[i] & 0xC0) == 0x80) i--;
+			}			
+			fwrite(outputbuffer,i,1,stdout);
+
 			fputc('\n',stdout);
-			p=outputbuffer+wrap_margin;
+			p=outputbuffer+i;
 		} else {
 			*p=0;p++;
 			fputs(outputbuffer,stdout);