[SoapRMI] preserving prefixes in document fragments

[Andy Harrison]

This is a multi-part message in MIME format.
--------------060001010307040103000408
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

Hi,

I am doing some work with WordML 2003. I wish to alter the contents of a
file, modifying only certain elements whilst leaving everything else.
I'm loading a document as a template and then inserting images and text
for a report.

I pass over most of the xml... headers, font info, namespaces etc and
flush that to a BufferedWriter, then I parse the chunks I am interested
in using builder.parseFragment(XppPullParser xpp).

What I want to know is, is it possible to get the
builder.serializeToXxxxx functions to not invent their own prefixes and
use the prefixes that are already loaded??

I've attached an example where I don't change any of the data but the
serialized output wants to declare its own prefixes. All of the prefixes
are declared at the top of the file, which I skipped over with
xpp.next(), so the new xmlns="" declarations are superfluous.

Not that I mind a few extra declarations, but it causes MS Word not to
show the picture!

thanks,

andy

--------------060001010307040103000408
Content-Type: text/plain;
 name="word test namespaces.txt"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="word test namespaces.txt"

ORIGINAL FRAGMENT FROM WORD 2003 ML

myFragment = builder.parseFragment(xpp);

<w:pict>
	<v:shape adj="0,,0" coordsize="" id="_x0_0_180_240.0094" o:spid="_x0000_s1026" o:spt="100" path="" style="stuff">
		<v:stroke joinstyle="round"/>
		<w:binData w:name="wordml://02000001.jpg">base64stuff/9j==</w:binData>
		<v:imagedata o:title="graphics1" src="wordml://02000001.jpg"/>
		<v:formulas/>
		<v:path o:connecttype="segments"/>
	</v:shape>
</w:pict>


A QUICK DUMP OF myFragment

...the original prefixes are there eg <v:shape o:spid="blah">

/body/sect/p/r/pict {prefix='w',namespaceName='http://schemas.microsoft.com/office/word/2003/wordml'}
/body/sect/p/r/pict/shape {prefix='v',namespaceName='urn:schemas-microsoft-com:vml'}
-> id: {prefix='null',namespaceName=''}
-> spid: {prefix='o',namespaceName='urn:schemas-microsoft-com:office:office'}
-> style: {prefix='null',namespaceName=''}
-> coordsize: {prefix='null',namespaceName=''}
-> spt: {prefix='o',namespaceName='urn:schemas-microsoft-com:office:office'}
-> adj: {prefix='null',namespaceName=''}
-> path: {prefix='null',namespaceName=''}
/body/sect/p/r/pict/shape/stroke {prefix='v',namespaceName='urn:schemas-microsoft-com:vml'}
-> joinstyle: {prefix='null',namespaceName=''}
/body/sect/p/r/pict/shape/binData {prefix='w',namespaceName='http://schemas.microsoft.com/office/word/2003/wordml'}
-> name: {prefix='w',namespaceName='http://schemas.microsoft.com/office/word/2003/wordml'}
/body/sect/p/r/pict/shape/imagedata {prefix='v',namespaceName='urn:schemas-microsoft-com:vml'}
-> src: {prefix='null',namespaceName=''}
-> title: {prefix='o',namespaceName='urn:schemas-microsoft-com:office:office'}
/body/sect/p/r/pict/shape/formulas {prefix='v',namespaceName='urn:schemas-microsoft-com:vml'}
/body/sect/p/r/pict/shape/path {prefix='v',namespaceName='urn:schemas-microsoft-com:vml'}



READ AND THEN RE-SERIALIZE...

builder.serializeToWriter(myFragment,outputWriter);

...the original prefixes are gone the serializer creates new ones and MS Word doesn't like it

eg. <v:shape n1:spid="blah" xmlns:n1="urn:schemas-microsoft-com:office:office">

<w:pict>
	<v:shape adj="0,,0" coordsize="" id="_x0_0_180_240.0094" n1:spid="_x0000_s1026" n1:spt="100" path="stuff" style="" xmlns:n1="urn:schemas-microsoft-com:office:office" xmlns:v="urn:schemas-microsoft-com:vml">
		<v:stroke joinstyle="round"/>
		<w:binData w:name="wordml://02000001.jpg">base64stuff/9j==</w:binData>
		<v:imagedata n1:title="graphics1" src="wordml://02000001.jpg"/>
		<v:formulas/>
		<v:path n1:connecttype="segments"/>
	</v:shape>
</w:pict>

--------------060001010307040103000408--