If you want the output to be XML, then the special JavaScript characters
must be escaped according to XML conventions.

If you want the output to be HTML containing <script> elements, then use
method="html" on xsl:output, and the special rules for the <script> element
in HTML will be applied.

Michael Kay 

> -----Original Message-----
> From: Alex Scott [mailto:fastidious007@xxxxxxxxxxx] 
> Sent: 27 May 2004 10:15
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Subject: RE:  MSXML and unwanted entity output
> 
> So how can I get a bit of javascript that is full of &&'s and <'s to
> output 
> as &&'s and <'s
> 
> Do I have to use something like (Example fromn FAQ) for every 
> instance of
> these characters in the Javascript?:
> <xsl:value-of select="'&lt;BR/&gt;'" disable-output-escaping="yes"/>
> 
> 
> 
> On Thu, 27 May 2004 10:20:54 +0100, "Alex Scott"
> <fastidious007@xxxxxxxxxxx> said:
> > Ok, 
> > 
> > So are you saying that this can't be change in the XSL as whatever
> > characters I use it will not make any difference to the output,
> > and this is therefore a problem (Supposedly further down 
> the line)in the
> > XML parser?
> > ---
> > Alex.
> > 
> > On Thu, 27 May 2004 10:05:51 +0100, "Michael Kay" 
> <mhk@xxxxxxxxx> said:
> > > > 
> > > > <![CDATA[&&]]>
> > > > Will out put as an entity:
> > > > &amp;&amp;
> > > > 
> > > > 
> > > > Does this look like a problem in the output type, or is there 
> > > > a way I can
> > > > alter the XSLT so that I can get 
> > > > the HTML returned as the same stuff I put in i.e. 
> > > > && or <
> > > 
> > > This is a FAQ. 
> > > 
> > > An XSLT transformation parses your source XML, creates a tree
> > > representation
> > > of the source, transforms it, and then serializes the result tree.
> > > Lexical
> > > details of the source XML, for example whether attributes 
> use single
> > > quotes
> > > or double quotes, are lost in this process. The difference between
> > > <![CDATA[&&]]> and &amp;&amp; is one of these lexical details.
> > > 
> > > You would want the test "contains(., '&amp;')" to work 
> regardless how the
> > > ampersand was written in the source, surely?
> > > 
> > > Michael Kay