[XERCESC-1838] Unexpected validation result when comment is present

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.8.0, 3.0.0
• Fix Version/s: 3.0.1, 3.1.0
• Component/s: Validating Parser (XML Schema)
• Assignee: Boris Kolpackov
• Reporter: Alexey Miroshnichenko
• Created: 04/Nov/08 19:18
• Updated: 12/Feb/09 14:12
• Resolved: 12/Feb/09 14:12

Description

Schema:
<?xml version="1.0" encoding="UTF-8"?>
<xsd:schema targetNamespace="http://www.openuri.org/mySchema" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns="http://www.openuri.org/mySchema" elementFormDefault="qualified" version="1.0">
<xsd:element name="root">
<xsd:complexType>
<xsd:sequence>
<xsd:element name="test" type="MyType" maxOccurs="unbounded"/>
</xsd:sequence>
</xsd:complexType>
</xsd:element>
<xsd:complexType name="MyType">
<xsd:simpleContent>
<xsd:extension base="xsd:byte">
</xsd:extension>
</xsd:simpleContent>
</xsd:complexType>
</xsd:schema>

document:
<?xml version="1.0" encoding="UTF-8"?>
<p1:root xmlns:p1="http://www.openuri.org/mySchema"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://www.openuri.org/mySchema schema.xsd">
<p1:test><!-- text -->34</p1:test>
<p1:test>
<!-- text -->34</p1:test>
<p1:test>
34</p1:test>
<p1:test>34</p1:test>
<p1:test>
34
</p1:test>
</p1:root>

SAX2Print output:
<?xml version="1.0" encoding="LATIN1"?>
<p1:root xsi:schemaLocation="http://www.openuri.org/mySchema schema.xsd">
<p1:test>34</p1:test>
<p1:test> 34
Error at file Z:\eTools\bin\Debug/instance-1.xml, line 7, char 26
Message: Datatype error: Type:InvalidDatatypeValueException, Message:Value ' 34' does not match re
gular expression facet '[+\-]?[0-9]+'.
</p1:test>
<p1:test>34</p1:test>
<p1:test>34</p1:test>
<p1:test>34</p1:test>
</p1:root>

Note the difference between second and third 'test' element in the document: second has comment.

Suggested fix - in SchemaValidator::normalizeWhiteSpace function last lines:
was
if (fCurReader->isWhitespace(*(srcPtr-1)))
fTrailing = true;
else
fTrailing = false;

now
...
if (0 != (srcPtr) && fCurReader->isWhitespace((srcPtr-1)))
fTrailing = true;
else
fTrailing = false;
}

Idea: isWhitespace returns 'true' for empty string, but here 'empty string' does not mean whitepace.
Please, review.