Hi, Jiang:
I just got the answer from apache xalan mail-list. You can check it; 

Following is the apache xalan's response:

>Sender: Christoffer Dam Bruun [mailto:cdbr...@flyingpigs.dk] 
>Date: 2011/6/6 17:52
>Receiver: xalan-j-us...@xml.apache.org
>Subject: Re: a question of DOMSource

>If your xpath expressions starts with '/'  e.g. '/Records/Record' then you
need >a document  If you are 
>transforming a "sub" element you cannot start with '/' e.g. '/Record' but
must >use 'Record'

>Best regards
>Christoffer Bruun

>>On 04-06-2011 07:07, ext2 wrote: 
>>Hi:

>>When I using Xalan 2.7.1 to transform a DOM Source;

>>I find that:The Node combine with DOMSource must be a Document and It
cannot >>be just a Element; 
>>Does anyone know if this limit is correctly or wrong, and why?
>>Following is a simple program to illustrate it
>>Document doc = DocumentBuilder.parse(.)
>>//if I using document as input, the result is correct;
>>Transformer.transform(new DOMSource(doc), streamResult);
>>//but if I using element as input, the result is wrong;
>>Transformer.transform(new DOMSource(doc.getDocumentElement),
streamResult);
>>
>>Ps: xml  xslt I used:
>>Xml:
>><?xml version="1.0" encoding="UTF-8"?>
>><Record xmlns="http://www.tongtech.com/tiwen/record1";>
>><content1>a</content1>
>><content2>b</content2>
>><content3>c</content3>
>></Record>

>>Xslt:
>>It's too long, but it logical is very simple, just exchange field content2
>>& content 3; 
>>And I am sure it's correctly, because I have use a
StreamSource(xml-stream), >>and Document 's DOMSource as input, their result
both are correctly;
>>but if anyone interest with it, I can post it in reply;

 


> -----original -----
>Sender: Willem Jiang [mailto:willem.ji...@gmail.com]
>Date: 2011/6/7 16:51
>Receiver: users@camel.apache.org
>Subject: Re: A problem of XSLT
> 
> As you can't tell if there are some problems in your XSLT template,
> can you write a simple java code to call transformer yourself?
> 
> On 6/4/11 12:38 PM, ext2 wrote:
> > Hi:
> >
> > Recently I am using camel xslt to do xml transform;
> >
> > If the xslt processor's input is a xml string, the  transform result is
> > correct;
> >
> > But if the xslt processor's input is a org.w3c.xml.Node which is build
from
> > the exactly same xml string; then thing's will be very strange:
> >
> > If I using the latest sun's JAXP implementation , a exception (Namespace
for
> > prefix  has not been declared) always occurs; and if I switch to apache
> > xalan 2.7.1,the transform result is just a empty document;
> >
> > Actually, I think this problem is the XSLT engine's reason;
> > But after I explorer the source code of xalan, I find it seems xalan
always
> > ask for the input DOM Node should be a document container. I am not sure
if
> > this limit is a bug or a correct thing;
> > Do anyone know it?
> >
> > If the limit is correctly, Maybe the camel XSLT processor should wrapper
a
> > DOM with a DocumentFragment, then build a DOMSource of DocumentFragment
as
> > XSLT's input;
> > But if the limit is xslt engine bug, maybe we should waiting from xalan
or
> > sun to correct this;
> >
> >
> > Ps: route&  xml&  xslt;
> > Route:( same route from DOM input&  xml string input)
> > <from uri="direct:from"/>
> > <to uri="xslt:test/test.xsl?output=DOM"/>
> > <to uri="mock:result/>
> >
> >
> > Xml:
> > Input:
> > <?xml version="1.0" encoding="UTF-8"?>
> > <Record xmlns="http://test/record1";>
> > <content1>a</content1>
> > <content2>b</content2>
> > <content3>c</content3>
> > </Record>
> >
> > Successful result:
> > <?xml version="1.0" encoding="UTF-8"?>
> > <tag0:Record xmlns:tag0="http://www.tongtech.com/tiwen/record1";>
> > <tag0:content1>a</tag0:content1>
> > <tag0:content2>c</tag0:content2>
> > <tag0:content3>b</tag0:content3>
> > </tag0:Record>
> >
> > Xslt:
> > .....it's too long, and I am sure it's correct....;
> >
> >
> >
> 
> 
> --
> Willem
> ----------------------------------
> FuseSource
> Web: http://www.fusesource.com
> Blog:    http://willemjiang.blogspot.com (English)
>           http://jnn.javaeye.com (Chinese)
> Twitter: willemjiang
> Weibo: willemjiang


Reply via email to