Hello Michael,

I've been listening to you very carefully. I understand what you mean when you say "©' and '&_#x000A9;' are identical as these are two different ways of writing the same copyright character".

Questions:
"What is the reason for wanting to do this - why do I need the unicode version in my output?
Answer: Actually I am working on a pagination platform, which uses xml as an input. Before importing my actual xml (which infact contains unicode values for different standard symbols in it like &_#x000A9; [without underscore]) to it I am adding few attributes and elements to it (using XSLT) to make things easy for composition purposes.

Once the xml is imported into this typesetting engine, all the unicode entities gets converted into their corresponding symbols.

Anyways, its not an issue at composition stage. But once the composition is finished, we are exporting the xml from the composition engine itself (as this may contain many textual correction it), with symbols (entities) it.
I am using XSLT to re-transform exported xml back to the required format, which includes:

1. Removal of additional elements and attributes.
2. Converting unicode symbols (for © ) back to their unicode values (&_#x000A9;-without underscore).

I don't find any problem with first one but stucked with the unicode ones

In fact I've tested to replace the literal © to another character e.g. a Big "X" instead , but it doesn't seems to be working, and which is worrying me most. It seems I'm making mistake in defining my style sheet (seems to be very tired today and thanks for your patience as I am almost novice in XSLT, who's trying his of its own). Here's my style sheet:

<?xml version="1.0" encoding="ISO-8859-1"?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0" xmlns:aid="http://ns.adobe.com/AdobeInDesign/4.0/">


 <xsl:template name="replace-string">
    <xsl:param name="text"/>
    <xsl:param name="from"/>
    <xsl:param name="to"/>

    <xsl:choose>
      <xsl:when test="contains($text, $from)">

    <xsl:variable name="before" select="substring-before($text, $from)"/>
    <xsl:variable name="after" select="substring-after($text, $from)"/>
    <xsl:variable name="prefix" select="concat($before, $to)"/>

    <xsl:value-of select="$before"/>
    <xsl:value-of select="$to"/>
        <xsl:call-template name="replace-string">
      <xsl:with-param name="text" select="$after"/>
      <xsl:with-param name="from" select="$from"/>
      <xsl:with-param name="to" select="$to"/>
    </xsl:call-template>
      </xsl:when>
      <xsl:otherwise>
        <xsl:value-of select="$text"/>
      </xsl:otherwise>
    </xsl:choose>
 </xsl:template>


<xsl:template match="text()">
<xsl:call-template name="replace-string">
        <xsl:with-param name="text" select="."/>
        <xsl:with-param name="from">©</xsl:with-param>
        <xsl:with-param name="to" select="P"/>
</xsl:call-template>
</xsl:template>

<xsl:template match="*|@*">
      <xsl:choose>
      <xsl:when test="name()='XXX'">
      </xsl:when>
      <xsl:when test="name()='YYY'">
      </xsl:when>

<xsl:when test="name()='char'">
      </xsl:when>
     <xsl:otherwise>
    <xsl:copy>
        <xsl:apply-templates select="text()|*|@*"/>
        </xsl:copy>
     </xsl:otherwise>
     </xsl:choose>
</xsl:template>
</xsl:stylesheet>

Hope I've made myself clear now.

Pankaj

<xsl:with-param name="to" select="P"/>

Try

<xsl:with-param name="to" select="'P'"/>

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

I am glad you agree that the two symbols are identical, but what you don't seem to realise is that replacing one with the other makes no difference in the slightest to what is output by the XML serialization after you have performed your transformation.

The way to affect the output from the XML serialization is to use the <xsl:output> declaration.

As for why your example doesn't work - thats cause its got bugs in it.

You're first example had single quotes around the "." - which meant you where searching the string '.' for the copyright symbol, instead of searching the current context node.

The second example didn't have quotes around the "P" in the "to" parameter - so unless you have an element in your XML input called <P> that wouldn't even run.

/- Sam Judson : Wrox Technical Editor -/

Thanks Sam and Michael.

I will look into it tomorrow, seems to be tired today as I've started making blunders now. I am sure there is a bug in my code.

Thanks for your patience and guidance.

Pankaj

Well it seems to me that I need to switch over to PERL script specifically for replacing symbols into unicode entities. I've tested it with Perl and getting perfect results.

Alas I wanted it with XSLT itself, but seems to be of no use as © (including other symbols) are converting into © only.

Thanks for help by the way.

Pankaj

Did you try the solution proposed using <xsl:output>?

XSLT itself works on a more abstract level that isn't concerned with how characters are serialized, on the basis that it doesn't actually matter. But the serializer does give you some control, driven by xsl:output.

Using Perl is fine, but be aware of the dangers. It isn't XML-aware, so you might find yourself translating characters in contexts where numeric character references aren't recognized or allowed (for example in comments, in CDATA sections, or in element names), and it might deliver the output file in an encoding that doesn't match the declared encoding in the XML declaration.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Thanks Michael. I did tried your <xsl:output> suggestion for unicode conversion and in fact I am testing it right now. I was not able to test it properly few days back as I was out of office for some urgent work. I too want to get it done thru XSLT only and would not prefer to unnecessarily execute a perl or any other program to convert symbols to unicode.

Let me test it thoroughly, will revert back in case I see any trouble in it.

Thanks.
Pankaj

Hi,

I've tried <xsl:output encoding="us-ascii"/> to translate all symbols, but not getting the desired output. Fo example converting © gives me the value #169;. On the other hand their are some places in text where the I am getting strange output like #169;#8233;.

Below is the snippet my stylesheet:
<?xml version="1.0" encoding="ISO-8859-1"?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0" xmlns:aid="http://ns.adobe.com/AdobeInDesign/4.0/">



<xsl:output encoding="us-ascii"/>

 <xsl:template name="replace-string">
    <xsl:param name="text"/>
    <xsl:param name="from"/>
    <xsl:param name="to"/>

    <xsl:choose>
      <xsl:when test="contains($text, $from)">
    <xsl:variable name="before" select="substring-before($text, $from)"/>
    <xsl:variable name="after" select="substring-after($text, $from)"/>
    <xsl:variable name="prefix" select="concat($before, $to)"/>

    <xsl:value-of select="$before"/>
    <xsl:value-of select="$to"/>
     <xsl:call-template name="replace-string">
      <xsl:with-param name="text" select="$after"/>
      <xsl:with-param name="from" select="$from"/>
      <xsl:with-param name="to" select="$to"/>

    </xsl:call-template>
      </xsl:when>
      <xsl:otherwise>
        <xsl:value-of select="$text"/>
      </xsl:otherwise>
    </xsl:choose>
</xsl:template>



<xsl:template match="text()|*|@*">
  <xsl:choose>
      <xsl:when test="name()='aid:pstyle'">
      </xsl:when>
      <xsl:when test="name()='aid:cstyle'">
      </xsl:when>

    <xsl:when test="name()='char'">
      </xsl:when>
     <xsl:otherwise>
    <xsl:copy>
        <xsl:apply-templates select="text()|*|@*"/>
    </xsl:copy>
     </xsl:otherwise>
 </xsl:choose>
</xsl:template>


<xsl:template match="text()">
    <xsl:call-template name="replace-string">
        <xsl:with-param name="text" select="'©'"/>
        <xsl:with-param name="from">©</xsl:with-param>
        <xsl:with-param name="to" select="'#x000A9;'"/>
    </xsl:call-template>
    <xsl:copy>
        <xsl:apply-templates select="text()|*|@*"/>
    </xsl:copy>

</xsl:template>

</xsl:stylesheet>

Any suggestions please where I possibly wrong.

I am working on TestXSLT using "Sablotron" XSLT proc. I have also tried using LibXSLT but of no use. By the way what's major difference in "Salotron" and LibXSLT? which one is better?
 Sorry for asking so much.


Pankaj

Please note I am using &_#x000A9 without Underscore and same is te case with resulting output i.e., &_#169.

hex A9 is the same as decimal 169, so this output is correct.

If there's a character 8223 in the output then that's because this character is present in the input. 8223 is hex 201F, which is a typographical ("99") closing quotation mark.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference