adude28 wrote:

> Here is what I have learned from this post.
> 
> 1. When you have an <!ENTITY and you are going to use a CDATA section
> then you must choose one type of quote mark internally and use the
> other to enclose the CDATA section.

Yes, this is mandated by the XML Spec. It's standard practice in
most languages that allow both types of quote, I think.

> 2. Inside your CDATA section you can not use "%" or "&" marks
> otherwise things will bomb out.

No, you can use % freely wherever you want EXCEPT in an ENTITY 
Declaration before the name of the entity, where it has a special
meaning. The & is a markup character like < and can be used inside 
a CDATA Section -- but it will get converted to &amp; when the 
content of the CDATA Section is processed.

I'm not clear why you think % and & are disallowed in CDATA Sections.
The text you gave as an example contained neither of these characters.
Is there something about your text you aren't telling us?

> I followed these two rules and things worked. Unfortunately, now I
> realize that this approach will probably not work because I can forsee
> times that I may need to use a "%" or "&" mark in my code.

You can use both of them whenever you wish. The % will be left as it
is. The & will be converted to &amp; of course.

Did you read the FAQ pages which explain alternative approaches?

> I am thinking of using the following format for my <!ENTITY reference
> 
> <!ENTITY MessageBodyContent SYSTEM
> 'http://www.url.com/MessageBodyContent.xml'>
> 
> that way I don't have to touch the content of my
> MessageBodyContent.xml and I then I can still post just one file to my
> processing page.

Any < or & characters in the file will still be converted to &lt; and 
&amp; when processed unless you kludge it with disable-output-escaping.

In any case, the "HTML" fragment you are encapsulating is not valid
HTML or XHTML, so any subsequent processor may gag on it (except a
browser).

///Peter