Hi Joseph,

Unfortunately I haven't been able to get specifically to this issue but there is a build based on the latest Saxon 8.5.1 bits that I haven't been able to thouroughly test well enough to release publicly.  But why don't we start by getting you access to these bits, see where that leaves things, and if the same issue is taking place then I will be able to work with you directly to track down the source of the problem and fix it either tomorrow evening or on Saturday.

You can access the bits here via the following links:

http://www.x2x2x.org/downloads/saxon.net/Saxon.NET.0.1.8.51.bin.debug.zip

http://www.x2x2x.org/downloads/saxon.net/Saxon.NET.0.1.8.51.bin.release.zip

The debug file contains the pdb's which won't help you fix the problem, but will certainly give you a better feel for whats going on behind the scenes which, if you are able to track down, will certainly be helpful as this will cut out the repro side of things, something I'm sure you know can be a pain in the butt when it wants to be. :)

Hopefully the above will fix things, but if not we can definitely get things worked out in the next two days.

Cheers :)

<M:D/>
On 10/20/05, Joseph Levine <jlevine@cropsci.ncsu.edu> wrote:

I wondered if you had a chance to look at this issue yet?

I've been running into this problem with a number of different transforms: they run fine using transform.exe, or through Saxon with JAXP, but return incorrect results applied in code in a .net class.  In most cases, the issue seems to be that the processor is treating <xsl:value-of statements in the manner of xslt 1.0 – i.e., selecting the first match in the sequence, and dumping the rest.

 

I'd be glad to start trying to find out what's causing this in the Saxon.net source, but before I begin looking in that direction, I'd like to be certain I'm not just invoking the transformer incorrectly.

 

Thanks,

Joe Levine

 


From: saxondotnet-help-admin@lists.sourceforge.net [mailto:saxondotnet-help-admin@lists.sourceforge.net] On Behalf Of M. David Peterson
Sent: Tuesday, October 11, 2005 6:24 PM
To: saxondotnet-help@lists.sourceforge.net
Subject: Re: [Saxondotnet-help] differences between saxon, saxon.net?

 

Hi Joe,

I am right smack dab in the middle of trying to meet a deadline but as soon as I am finished I will look into this and get back to you later this evening.

Cheers :)

On 10/11/05, Joseph Levine <jlevine@cropsci.ncsu.edu> wrote:

I'm having trouble figuring out why the following transform works as I expected it to in Saxon invoked by JAXP, and from the Saxon.net command line, but produces different results when invoked from an Asp.net application:

 

Here's the xslt in question – the url is http://www2.turffiles.ncsu.edu/ContentManagement/CMXml/TestTransform.xsl

 

   <xsl:template match="PropertyGroupSet">

               

                <xsl:variable name="fieldNames">                                             

                                <xsl:value-of select="Property/Name" separator=","></xsl:value-of>                                            

                </xsl:variable>

                               

                <xsl:value-of select="$fieldNames"></xsl:value-of>

                <xsl:choose>

                        <xsl:when test="Property/IsSingleValueProperty=0">    

                                <br></br>

                                <fieldset>

                                        <xsl:call-template name="multipleValues"/>

                                        <legend>                                             

                                                <xsl:value-of select="@Name" />

                                                <img class="groupOptions" src="http://www.turffiles.ncsu.edu/images/fwd.gif " onclick="ShowPossibilities(event,{@PropertyID},true,'{$fieldNames}')" />                                             

                                        </legend>

                                        <xsl:call-template name="formItem"/>

                                </fieldset>                                     

                        </xsl:when>

                        <xsl:otherwise>

                                <xsl:call-template name="formItem"/>

                                <img class="Options" src="http://www.turffiles.ncsu.edu/images/fwd.gif" onclick="ShowPossibilities(event,{Property/PropertyID},false,'{$fieldNames}')" />     

                        </xsl:otherwise>

                </xsl:choose>             

        </xsl:template>

 

The source document is at http://www2.turffiles.ncsu.edu/ContentManagement/CMXml/TestProps.xml

 

The relevant section is the following PropertyGroupSet node with 2 name descendants:

 

<PropertyGroupSet Name="Author" PropertyID="17">

    <Property>

      <PropertyID>5</PropertyID>

      <Name>LastName</Name>

      <FriendlyName>Last Name</FriendlyName>

      <Type>String</Type>

      <IsUnique>false</IsUnique>

      <IsSingleValueProperty>0</IsSingleValueProperty>

      <PropertyGroupSet>Author</PropertyGroupSet>

    </Property>

    <Property>

      <PropertyID>6</PropertyID>

      <Name>FirstName</Name>

      <FriendlyName>First Name</FriendlyName>

      <Type>String</Type>

      <IsUnique>false</IsUnique>

      <IsSingleValueProperty>0</IsSingleValueProperty>

      <PropertyGroupSet>Author</PropertyGroupSet>

    </Property>

  </PropertyGroupSet>

 

I expect that the $fieldNames variable will select both Name nodes in this group; instead, it only selects the first of these.  More oddly still, the xpath expression count(Property/Name) correctly returns the value 2 here.  Property/Name[2] returns 'FirstName', as I expect it to.  Is there an error in my xslt that I'm not seeing?  Or is this an issue in Saxon.Net?  As I said, this works fine from the command line for both Saxon and Saxon.net, and invoked by jaxp in java.

 

On a somewhat related issue: I also had trouble with a couple of other xpath functions: node-name() and distinct-values() both seemed to choke the saxon.net processor, at least when I tried them.

 

Thanks,

Joe Levine

       

        




--
<M:D/>

M. David Peterson
http://www.xsltblog.com/




--
<M:D/>

M. David Peterson
http://www.xsltblog.com/