[XSL-LIST Mailing List Archive Home] [By Thread] [By Date]

Re: [xsl] Multiple CDATA tags


Subject: Re: [xsl] Multiple CDATA tags
From: David Carlisle <davidc@xxxxxxxxx>
Date: Tue, 15 Feb 2005 18:04:13 GMT

It is conformant behaviour (more or less anything from not using CDATA
at all, to putting a new CDATA section around each character is allowed)
The XSLT spec gives implementations a lot of freedom here. Just as they
have freedom wheher to use " or ' around attribute values, and other
"syntactic" differences in the markup.

Sometimes the system has to break the run into multiple CDATA sections,
for example if the content includes ]]> or any characters not in the
current encoding.

It looks like libxslt's method of ensuring that it catches any ]]> is to
split the CDATA as soon as it sees one ] which is presumably quicker and
sufficient, although not necessary in this case.

The output you get basically means the same thing as the output you
wanted, so it shouldn't really matter. generally speaking you shouldn't
ever really need cdata-section-elements=" as CDATA is really a mechanism
for simplifying hand authoring, saving you quote < and &; for machine
generated XML such as the output of XSLT, it isn't really doing very
much.

David

________________________________________________________________________
This e-mail has been scanned for all viruses by Star. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
http://www.star.net.uk
________________________________________________________________________


Current Thread
Keywords