7015 - XML Parser not conforming to element declarations

Status: VERIFIED INVALID

(Core :: XML, defect, P3)

Description

[salwan.searty]

The XML parser in release M5 does not seem to enforce the grammar defined in
the element declarations. For example, the following element declarations
define the element DOCUMENT to contain at least one LEVEL1 element, and the
LEVEL1 element to contain atleast one LEVEL2 element:

<?xml version="1.0"?>
<!DOCTYPE DOCUMENT [
<!ELEMENT DOCUMENT (LEVEL1)+>
<!ELEMENT LEVEL1 (LEVEL2)+>
<!ELEMENT LEVEL2 (#PCDATA)>
]>

That means the following code should be legal:

<DOCUMENT>
	<LEVEL1>
		<LEVEL2>
		Just some text.
		</LEVEL2>
	</LEVEL1>
</DOCUMENT>

While the following code should NOT be legal: (because it doesn't conform to
the grammar defined above.)

<DOCUMENT>
	<LEVEL1>
	Just some text.
	</LEVEL1>
	<LEVEL2>
	Some more text.
	</LEVEL2>
</DOCUMENT>

However, even in the second code snippet, the Mozilla does not seem to generate
an error.

Comment 1

[Nisheeth Ranjan]

Thanks a lot for taking the time to file this bug report.

Expat is not a validating parser, so it only reports well-formedness errors.
Both the examples you mention are well-formed XML, so, you see no errors.

Mozilla will probably do document validation once the XML schema spec gets more
baked.  We are not planning on doing DTD based validation because DTDs will get
deprecated once XML schemas become a W3C Recommendation.

Comment 2

[James Ross]

*** Bug 182604 has been marked as a duplicate of this bug. ***