From dcc19503193c71596278a252064a8ce66331b3cd Mon Sep 17 00:00:00 2001
From: Daniel Veillard <veillard@redhat.com>
Date: Wed, 22 May 2013 22:56:45 +0200
Subject: [PATCH] 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 +++++-
 result/japancrlf.xml       |  4 ++++
 result/japancrlf.xml.rde   |  7 +++++++
 result/japancrlf.xml.rdr   |  7 +++++++
 result/japancrlf.xml.sax   | 11 +++++++++++
 result/japancrlf.xml.sax2  | 11 +++++++++++
 result/noent/japancrlf.xml |  4 ++++
 test/japancrlf.xml         |  6 ++++++
 8 files changed, 55 insertions(+), 1 deletion(-)
 create mode 100644 result/japancrlf.xml
 create mode 100644 result/japancrlf.xml.rde
 create mode 100644 result/japancrlf.xml.rdr
 create mode 100644 result/japancrlf.xml.sax
 create mode 100644 result/japancrlf.xml.sax2
 create mode 100644 result/noent/japancrlf.xml
 create mode 100644 test/japancrlf.xml

diff --git a/parser.c b/parser.c
index 4a442bb..4739add 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));
 }
 
 /**
diff --git a/result/japancrlf.xml b/result/japancrlf.xml
new file mode 100644
index 0000000..60b307c
--- /dev/null
+++ b/result/japancrlf.xml
@@ -0,0 +1,4 @@
+<?xml version="1.0"?>
+<入力メッセージ xmlns="http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826">
+	<c8c:Ele xmlns:c8c="http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826"/>
+</入力メッセージ>
diff --git a/result/japancrlf.xml.rde b/result/japancrlf.xml.rde
new file mode 100644
index 0000000..0bc5af7
--- /dev/null
+++ b/result/japancrlf.xml.rde
@@ -0,0 +1,7 @@
+0 1 入力メッセージ 0 0
+1 14 #text 0 1 
+	
+1 1 c8c:Ele 1 0
+1 14 #text 0 1 
+
+0 15 入力メッセージ 0 0
diff --git a/result/japancrlf.xml.rdr b/result/japancrlf.xml.rdr
new file mode 100644
index 0000000..0bc5af7
--- /dev/null
+++ b/result/japancrlf.xml.rdr
@@ -0,0 +1,7 @@
+0 1 入力メッセージ 0 0
+1 14 #text 0 1 
+	
+1 1 c8c:Ele 1 0
+1 14 #text 0 1 
+
+0 15 入力メッセージ 0 0
diff --git a/result/japancrlf.xml.sax b/result/japancrlf.xml.sax
new file mode 100644
index 0000000..c3bbdc6
--- /dev/null
+++ b/result/japancrlf.xml.sax
@@ -0,0 +1,11 @@
+SAX.setDocumentLocator()
+SAX.startDocument()
+SAX.startElement(入力メッセージ, xmlns='http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826')
+SAX.characters(
+	, 2)
+SAX.startElement(c8c:Ele, xmlns:c8c='http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826')
+SAX.endElement(c8c:Ele)
+SAX.characters(
+, 1)
+SAX.endElement(入力メッセージ)
+SAX.endDocument()
diff --git a/result/japancrlf.xml.sax2 b/result/japancrlf.xml.sax2
new file mode 100644
index 0000000..878eae4
--- /dev/null
+++ b/result/japancrlf.xml.sax2
@@ -0,0 +1,11 @@
+SAX.setDocumentLocator()
+SAX.startDocument()
+SAX.startElementNs(入力メッセージ, NULL, 'http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826', 1, xmlns='http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826', 0, 0)
+SAX.characters(
+	, 2)
+SAX.startElementNs(Ele, c8c, 'http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826', 1, xmlns:c8c='http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826', 0, 0)
+SAX.endElementNs(Ele, c8c, 'http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826')
+SAX.characters(
+, 1)
+SAX.endElementNs(入力メッセージ, NULL, 'http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826')
+SAX.endDocument()
diff --git a/result/noent/japancrlf.xml b/result/noent/japancrlf.xml
new file mode 100644
index 0000000..60b307c
--- /dev/null
+++ b/result/noent/japancrlf.xml
@@ -0,0 +1,4 @@
+<?xml version="1.0"?>
+<入力メッセージ xmlns="http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826">
+	<c8c:Ele xmlns:c8c="http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826"/>
+</入力メッセージ>
diff --git a/test/japancrlf.xml b/test/japancrlf.xml
new file mode 100644
index 0000000..480cb2d
--- /dev/null
+++ b/test/japancrlf.xml
@@ -0,0 +1,6 @@
+﻿<入力メッセージ
+	xmlns="http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826">
+	<c8c:Ele
+		xmlns:c8c="http://schemas.cordys.com/webapps/1.0/bpm/c8c8b82a-0ac0-3d19-01e2-bda74af9b826"
+	/>
+</入力メッセージ>
\ No newline at end of file
-- 
1.8.3.2