[XERCESJ-894] Too much malformed data is reported

• Type: Bug
• Status: Resolved
• Resolution: Fixed
• Affects Version/s: 2.6.1
• Component/s: SAX
• Environment: Operating System: All; Platform: All
• Bugzilla Id: 27081
• Assignee: Unassigned
• Reporter: elharo
• Created: 20/Feb/04 00:15
• Updated: 16/Feb/11 23:55
• Resolved: 13/Apr/04 01:43

Description

Consider the following malformed test case from the W3C XML conformance test
suite <xmltest/not-wf/sa/027.xml.html>:

<doc>
<!-- abc
</doc>

When processing this document with SAX, Xerces reports the content of the
malformed comment by invoking the characters method. That is, it reports <!--
abc as character data. There is no possible way anything later in the file could
change this into character data. Thus there is no way for the parser to think
that this is well-formed character data. Xerces should report a fatal error and
not report the <!-- abc as content.