[VYSPER-338] XMLParser.unescape throw exception

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.7
• Component/s: core protocol
• Environment: java1.6, windows7
• Assignee: Unassigned
• Reporter: Zhang JinYan
• Created: 24/Jan/13 06:52
• Updated: 24/Jan/13 07:19

Description

If message stanza contains text: "辽宁" （escape before send："辽宁"）
exception will be throw out:

Caused by: org.xml.sax.SAXParseException: For input string: "8FBD;&#x5B81"
at org.apache.vysper.xml.sax.impl.XMLParser.fatalError(XMLParser.java:499)
at org.apache.vysper.xml.sax.impl.XMLParser.parse(XMLParser.java:124)
at org.apache.vysper.xml.sax.impl.DefaultNonBlockingXMLReader.parse(DefaultNonBlockingXMLReader.java:185)
at org.apache.vysper.xml.decoder.XMPPDecoder.doDecode(XMPPDecoder.java:117)

Caused by：String org.apache.vysper.xml.sax.impl.XMLParser.unescape(String s)

private String unescape(String s) {
s = s.replace("&", "&").replace(">", ">").replace("<", "<").replace("'", "'").replace(""",
"\"");

StringBuffer sb = new StringBuffer();

Matcher matcher = UNESCAPE_UNICODE_PATTERN.matcher(s);
int end = 0;
while (matcher.find())

{ boolean isHex = matcher.group(1).equals("x"); String unicodeCode = matcher.group(2); int base = isHex ? 16 : 10; int i = Integer.valueOf(unicodeCode, base).intValue(); char[] c = Character.toChars(i); sb.append(s.substring(end, matcher.start())); end = matcher.end(); sb.append(c); }

sb.append(s.substring(end, s.length()));

return sb.toString();
}

Replace xml predefined entities before unescape change the context of escaped strings.
For example:
Input: "&#x8FBD;&#x5B81;"
After replace: "辽宁"

unescape use regex： Pattern.compile("\\&
#(x?)(.+);");
match：
group(1) = x
group(2) = 8FBD;&#x5B81

then Integer.valueOf(unicodeCode, base) will throw exception.
I fixed this bug, see the patch.