I don't think any XML parser reveals to the application what whitespace exists between attribute values, so there is no way the application (in this case the XSLT processor) can preserve this whitespace.
 
There's surely something seriously wrong with an XML application that fails if there's too much or too little whitespace between attributes.
 
Or have I misunderstood the problem?
 

Regards,

Michael Kay
http://www.saxonica.com/
http://twitter.com/michaelhkay




From: Ternola, Gerald [mailto:g-ternola@ti.com]
Sent: 17 February 2010 09:26
To: saxon-help@lists.sourceforge.net
Subject: [saxon] How to omit normalization in Saxon

Hi Michael!

            Thanks for helping us with the line Wrapping issue, We haven’t notice this problem until only today that we are doing a regression test with our system.

            Below code is the java codes that we had isolated that is causing a major problem to use. When we pass an xml

            Example

                        <element>

                                    <data

                                                id="12"

                                                lot="134"

                                    />

                                    <data

                                                id="13" lot="A"

                                    />

</element>

è Some of the elements are in multiple rows, but we need to maintain or preserve it’s structure when we do the schema validation.

 

The output was:

<element>

                  <data id="12" lot="134"/>

                  <data id="13" lot="A"/>

</element>

 

But we still need the original structure, Basically we need the XdmNode to be passed on the XML Pipeline “Calabash” for further processing.