Details
-
Bug
-
Status: Resolved
-
Resolution: Incomplete
-
2.2.0
-
None
-
None
-
Operating System: Windows NT/2K
Platform: PC
-
14876
Description
I've been reviewing the DOMEcho example provided in JAXP (inside the
java_xml_pack-summer-02_01 download from sun) and it is not clear to me
why we have to set the schemaSource on the DocumentBuilderFactory. I feel that
if the schemaSource is set to null (or not set at all), the DocumentBuilder(s)
that are created from that factory should take into account the schemaLocation
from inside of the xml document. By ignoring the schemaLocation declarations
within an xml document, Xerces will not be able to handle localized namespace
declarations and the use of multiple independent schemas by one document.
It is also not convenient to have to go into an xml document to extract the
schemaLocation so that we can set the schemaSource before we try
parsing/validating the document. This means we need to parse twice whenever we
want to validate.
I know that the Xerces2-J parser does provide a Parser class that can do this.
I can see it in the Dom.Count example. Why can't we integrate that same
behavior under the JAXP APIs?
Please feel free to contact me for any clarifications.
Thanks.