From: Daniel Veillard <veillard@redhat.com>
Date: Wed, 22 May 2013 20:56:45 +0000
Subject: Fix a parsing bug on non-ascii element and CR/LF usage

https://bugzilla.gnome.org/show_bug.cgi?id=698550

Somehow the behaviour of the internal parser routine changed
slightly when encountering CR/LF, which led to a bug when
parsing document with non-ascii Names
---
 parser.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/parser.c b/parser.c
index b9df6d8..dd00399 100644
--- a/parser.c
+++ b/parser.c
@@ -3404,6 +3404,7 @@ xmlParseNCNameComplex(xmlParserCtxtPtr ctxt) {
     int len = 0, l;
     int c;
     int count = 0;
+    const xmlChar *end; /* needed because CUR_CHAR() can move cur on \r\n */
 
 #ifdef DEBUG
     nbParseNCNameComplex++;
@@ -3413,6 +3414,7 @@ xmlParseNCNameComplex(xmlParserCtxtPtr ctxt) {
      * Handler for more complex cases
      */
     GROW;
+    end = ctxt->input->cur;
     c = CUR_CHAR(l);
     if ((c == ' ') || (c == '>') || (c == '/') || /* accelerators */
 	(!xmlIsNameStartChar(ctxt, c) || (c == ':'))) {
@@ -3434,12 +3436,14 @@ xmlParseNCNameComplex(xmlParserCtxtPtr ctxt) {
 	}
 	len += l;
 	NEXTL(l);
+	end = ctxt->input->cur;
 	c = CUR_CHAR(l);
 	if (c == 0) {
 	    count = 0;
 	    GROW;
             if (ctxt->instate == XML_PARSER_EOF)
                 return(NULL);
+	    end = ctxt->input->cur;
 	    c = CUR_CHAR(l);
 	}
     }
@@ -3448,7 +3452,7 @@ xmlParseNCNameComplex(xmlParserCtxtPtr ctxt) {
         xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NCName");
         return(NULL);
     }
-    return(xmlDictLookup(ctxt->dict, ctxt->input->cur - len, len));
+    return(xmlDictLookup(ctxt->dict, end - len, len));
 }
 
 /**