Re: [Xsltforms-support] applying a stylesheet in an instance to data in an instance

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

Rev. 550 includes a new experimental action named "setnode" with @inner 
or @outer attributes to perform a "loadXML" for the corresponding node.

Here is a test case:
<html xmlns="http://www.w3.org/1999/xhtml" 
xmlns:xf="http://www.w3.org/2002/xforms" 
xmlns:ev="http://www.w3.org/2001/xml-events">
   <head>
     <title>Inner &amp; Outer Set Node</title>
     <xf:model>
             <xf:instance>
                 <data xmlns="">
                     <item>a</item>
                 </data>
             </xf:instance>
         </xf:model>
   </head>
   <body>
     <h1>Inner &amp; Outer Set Node</h1>
     <xf:output value="item"/>
     <xf:trigger>
         <xf:label>Inner</xf:label>
         <xf:setnode ref="." ev:event="DOMActivate" 
inner="'&lt;item&gt;b&lt;/item&gt;'"/>
     </xf:trigger>
     <xf:trigger>
         <xf:label>Outer</xf:label>
         <xf:setnode ref="item" ev:event="DOMActivate" 
outer="'&lt;item&gt;c&lt;/item&gt;'"/>
     </xf:trigger>
   </body>
</html>

Le 16/07/2012 16:32, C. M. Sperberg-McQueen a écrit :
> For what it's worth, I think it would be useful to have a standard way
> to load the result of a transform into an instance.
>
> In case an example helps, the form I was working on this past week
> may serve as a simple use case.  It involves two XSLT processes:
> first, the user loads an XML document and a static stylesheet (the
> 'factory') runs over the user's document and produces a second
> stylesheet (the 'filter'), and then the user can configure the filter by
> changing various default behaviors.  Whey they are done, they can
> save the filter to disk or apply the filter to the user's document.
>
> Since the point of the filter is to produce untagged text for the user
> to feed into text analysis or statistical software that does not understand
> XML (it's sad to think that people still use such software, but I'm told
> they do), there is no particular need to load the XML structure of the
> second transformation result into an instance.  (Initially I used the
> 'text' output method in the transform, but all the browsers I tested
> appear to insist on producing either XML or HTML output from a
> transform, so I switched to the XML output method and wrapped
> the text result in an 'output' element.)
>
> But the result of the first transformation MUST be loaded into an
> instance for the form to serve its purpose; the point of the form is
> to illustrate, for a panel discussion at a conference next week,
> (a) that it's not enough just to 'strip the tags out' of an XML document
> and (b) that it's possible to provide an interface to allow an
> intelligent selection from the encoded document.  So the user
> must be able to edit the filter, which means the filter must be an
> XML instance.
>
> In this case, I handled the lack of 'loadXML' action by using a simple
> trick:  I submit the text result of the initial transformation to a script
> on the server, which does nothing but bounce it back to the client
> as text/xml.  The submission is defined this way:
>
>        <xf:submission id="load-userxslt"
> 		     ref="instance('filter')/text()"
> 		     method="post"
> 		     resource="../../../lib/reflect-as-xml.sh"
> 		     replace="instance"
> 		     instance="filter"
> 		     />
>
> The 'filter' instance initially contains the text-node output from the
> transformation; this submission sends that output (without the
> outermost wrapper element, which is not needed) to the server,
> where 'reflect-as-xml.sh' does nothing but send it back with an HTTP
> header.  Reduced to its essentials, reflect-as-xml.sh looks like this:
>
>    #!/bin/sh
>    echo "Content-Type: text/xml"
>    echo
>    cat
>
> It would be a nicer solution, however, if no server round-trip were
> required here.  (I thought about trying to figure out how to do that
> in Javascript with an extension function or extension action, but I
> decided to go with this quicker solution, for purposes of the demo.)
>
> In case it's of interest to readers of this list, the form in question
> is now visible at
>
>    http://blackmesatech.com/2012/07/dh/sdb.xml
>
> The form is intended as a toy to illustrate some general principles,
> and not as a tool for really serious work.  (So in particular, don't
> give it a very large document to work with.  It works fine on a
> TEI-encoded sonnet, and it takes several minutes [and counting] to
> handle one of Shakespeare's five-act plays.)
>
>