RE: Assertion in XercesDocumentBridge::createBridgeNode (xalan 1. 1)

[Dave Connet <co...@entelos.com>]

Our original code first builds a DOM from our internal data format, and then
the DOM is either transformed or dumped - so I just created the bridge to
use what was already built. (Before the bridge, I wrote the data to a temp
file and then passed that file to xalan)

In time I want to do things differently - but due to time constraints,
that's not likely to happen too quickly... Since the data we're dealing with
can be very large (20+M when written to an xml file), I really don't want
the whole thing in memory at one time!

Dave Connet


-----Original Message-----
From: David_N_Bertoni@lotus.com [mailto:David_N_Bertoni@lotus.com]
Sent: Wednesday, February 21, 2001 4:02 PM
To: xalan-dev@xml.apache.org
Subject: Re: Assertion in XercesDocumentBridge::createBridgeNode (xalan
1.1)



Yes, there should probably be a break, but this is why I hate this bridge.
The problem is that if someone tries to map the Xerces node, it's a very
bad thing to not be able to map it.

Just a question -- why are you using the Xerces DOM and the bridge?  The
performance is much better with the new Xalan source tree.  See the samples
or process.cpp in testXSLT for more information.

Dave



 

                    Dave Connet

                    <connet@entel        To:
"'xalan-dev@xml.apache.org'" <xa...@xml.apache.org>           
                    os.com>              cc:     (bcc: David N
Bertoni/CAM/Lotus)                                  
                                         Subject:     Assertion in
XercesDocumentBridge::createBridgeNode (xalan   
                    02/21/2001           1.1)

                    05:04 PM

                    Please

                    respond to

                    xalan-dev

 

 




While transforming, I'm getting an assertion in the default switch of
XercesDocumentBridge::createBridgeNode (line 894). The node type is
XML_DECL_NODE (which is not handled, neither is DOCUMENT_NODE).

It seems that XML_DECL_NODE should probably just have an empty case there.
892a893,895
>       case DOM_Node::XML_DECL_NODE:
>               break;
>

But what about DOCUMENT_NODE?

Dave Connet
Entelos, Inc.