XSL general entity treatment

[harebrain]

Hi all,

I have an XML document that I need to reformat using XSL, basically changing all the attribute="value" stuff to their own child elements. This works great EXCEPT & quot ; and & #34 ; get changed to " (34 is ASCII for &quot while & amp ; remains as-is.

Why? Aren't < > & &quot; and ' forbidden in XML aside from their special meanings? So why would an XSL transform screw with the entities? Better yet, why does it only screw with HALF of them?

I've tried

<xsl:value-of select=&quot;.&quot; disable-output-escaping=&quot;no&quot;/>

and

<xsl:value-of select=&quot;.&quot; disable-output-escaping=&quot;yes&quot;/>

to try to work around this, but it doesn't make any difference.

I've done this with different parsers, too, and the behavior is the same. What's up with this?

(I've seen this included in other questions here, but this peculiarity itself hasn't been addressed.)

--harebrain

 

[harebrain]

Well, because nobody else has an answer for this one, I'll share what I've found. According to

http://www.w3schools.com/xml/xml_cdata.asp

:

Only the characters &quot;<&quot; and &quot;&&quot; are strictly illegal in XML. Apostrophes, quotation marks and greater than signs are legal, but it is a good habit to replace them.

So if XML doesn't require use of an entity reference where a quote character is needed, there's no reason why XSLT should preserve those entity references in a result tree.

This also explains the contrary behavior regarding ampersands: the ampersand character is illegal in XML character data, so when data is returned from the result tree it would have to be converted back to an entity reference.

(To w3schools: Why is it a &quot;good habit&quot; to do something that is superfluous?)

(To the W3C: Thanks for the confusion! )

--harebrain