Thanks David,

Yes I am aware of the cdata-section-elements XSLT attribute however it does
not really solve my problem - about 5% of the elements in the source
document are <value/> elements not all of which contain CDATA. If we added
new CDATA sections to the output this would probably not satisfy the
customer either.

Kind regards,

Michael

On Wed, Sep 9, 2009 at 12:05 PM, David Carlisle <davidc@n...> wrote:

>
>
> > wondering whether anyone could point me to the relevant place which
> > specifies that this is expected behaviour of the parser/xslt
> > processor.
>
> The XSLT engine has no information about whether CDATA was used so when
> it linearises the tree it just does whatever it does, it has to quote <
> somehow, but it is free to use cdtat sections or numeric or entity
> references. This is exactly the same as using ' or " around attribute
> values, the XSLT engine can not preserve those either: it will most
> likely always use ".
>
> What you can do is tell the XSLT engine to always use CDATA sections for
> those elements add cdata-section-elements="value" to the xslt
> stylesheet.
>
> David
>
>
>
> ________________________________________________________________________
> The Numerical Algorithms Group Ltd is a company registered in England
> and Wales with company number 1249803. The registered office is:
> Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.
>
> This e-mail has been scanned for all viruses by Star. The service is
> powered by MessageLabs.
> ________________________________________________________________________
>