Re: reserved character in XML element value

I think this discussion has made a little sense out of something in one of our products - I had to HTML-encode things like apostrophes - &apos did not always work - had to use &27, as I recall.

Now I just saw a site that differentiated HTML-encoding vs XML-encoding - the former has only 4 reserved characters, the latter has 5. The 5th one is &apos - not part of HTML, therefore, not consistently rendered, IME.

Nice!!

On 6/5/2012 4:05 AM, Larry Ducie wrote:

Hi David,


Firstly, the parser should not parse the content of a CDATA block. That doesn't mean it wont though. You are at the mercy of the implementation and they are not all standard. Anybody who has had to talk to many SOAP servers learns this lesson quickly.
See http://www.w3schools.com/xml/xml_cdata.asp for some details confirming your belief.

Secondly, when escaping entities yourself there are a few characters to watch out for - not just the ampersand (&):

Certain characters are reserved because they form the structure of the markup language. they are<>'".

To escape them they are replaced by representations:

< goes to&lt;

goes to&gt;

' goes to&apos;
" goes to&quot;

Because the escaped representations begin with an ampersand (&) that character is now also reserved! So we have to escape it too:

& goes to&amp;

If you want to escape the characters yourself NEVER escape the ampersand last because escaping all the others introduces new ampersands into the markup and you want to leave those ones alone.

Finally, why work this out yourself? Use an open source tool to do this. Henrik has built a good one, and you can download mine at http://sourceforge.net/projects/xmli/files/ They're free and much used.


Cheers

Larry Ducie