xsltforms-support

[Xsltforms-support] Hints on using the transform function

[Philip F. <Phi...@ma...>]

I've been experimenting with XSLTForms' extension transform function but I can't seem to get it to work or, maybe, I'm not using it correctly. I had hoped I could use it to apply a transform to one instance and deposit the result into another by using an xf:insert:

<xf:insert origin="transform(instance('source'), '/resources/schemas/schematron/atom-feed.xsl')" position="after" nodeset="." context="instance('svrlReport')"/>

I've tested the transform in the browser and it does work, I also know that the plain insert works but when I invoke the 'insert' in my form with the transform function as described above - nothing happens.

How should I be invoking the transform function?


Regards

Philip

Re: [Xsltforms-support] Hints on using the transform function

[Alain C. <ala...@ag...>]

Hello Philip,
> I've been experimenting with XSLTForms' extension transform function 
> but I can't seem to get it to work or, maybe, I'm not using it 
> correctly. I had hoped I could use it to apply a transform to one 
> instance and deposit the result into another by using an xf:insert:
> <xf:insert origin="transform(instance('source'), 
> '/resources/schemas/schematron/atom-feed.xsl')" position="after" 
> nodeset="." context="instance('svrlReport')"/>
The current transform() function in XSLTForms returns a string.  A 
parse() function could be added to get an XML Document instead.

Regards,

-Alain

Re: [Xsltforms-support] Hints on using the transform function

[Philip F. <Phi...@ma...>]

Alain,

Thank you. Yes, I can see that the function works when I use xf:setvalue to send the result to a text node. Would an alternative implementation of the transform function return a node and then you could use the existing serialize function to generate the textual representation from the result of transform?


Regards

Philip

From: Alain Couthures [mailto:ala...@ag...]
Sent: Tuesday, May 10, 2011 8:00 AM
To: Philip Fennell
Cc: xsl...@li...
Subject: Re: [Xsltforms-support] Hints on using the transform function

Hello Philip,
I've been experimenting with XSLTForms' extension transform function but I can't seem to get it to work or, maybe, I'm not using it correctly. I had hoped I could use it to apply a transform to one instance and deposit the result into another by using an xf:insert:

<xf:insert origin="transform(instance('source'), '/resources/schemas/schematron/atom-feed.xsl')" position="after" nodeset="." context="instance('svrlReport')"/>
The current transform() function in XSLTForms returns a string.  A parse() function could be added to get an XML Document instead.

Regards,

-Alain

Re: [Xsltforms-support] Hints on using the transform function

[Alain C. <ala...@ag...>]

Philip,
> Would an alternative implementation of the transform function return a 
> node and then you could use the existing serialize function to 
> generate the textual representation from the result of transform?

Wouldn't the result always be a document root or a document element?

Considering that I don't know if people are already using the current 
transform() function, wouldn't it be better to name this function 
transform-doc(), for example?

What if the result of the transformation is a text?

Regards,

-Alain

Re: [Xsltforms-support] Hints on using the transform function

[Philip F. <Phi...@ma...>]

Alain,

> What if the result of the transformation is a text?

Currently, if the transform returns just text then Firefox wraps the result of the transform in an element but, of course, this gets serialized too.

May be there should be a standard 'wrapper' element for plain text results in the same way that XProc will wrap the result of a p:data instruction defined in an input source port (used to retrieve non-XML content) in a c:data element.


Regards

Philip

From: Alain Couthures [mailto:ala...@ag...]
Sent: Tuesday, May 10, 2011 4:13 PM
To: Philip Fennell
Cc: xsl...@li...
Subject: Re: [Xsltforms-support] Hints on using the transform function

Philip,
Would an alternative implementation of the transform function return a node and then you could use the existing serialize function to generate the textual representation from the result of transform?

Wouldn't the result always be a document root or a document element?

Considering that I don't know if people are already using the current transform() function, wouldn't it be better to name this function transform-doc(), for example?

What if the result of the transformation is a text?

Regards,

-Alain