Re: [xsl] nbsp fails transformation

[Abel Braaksma <abel.online@xxxxxxxxx>]

> If someone sends you a document that isn't well-formed XML, the best 
> strategy is to get the people who produced it to mend their ways. 

True. However, having   in an XML file and finding out that all of 
a sudden XML is not XML anymore must be among the most frequent 
unpleasant surprises fresh XML programmers have to deal with. I believe 
it was among one of my first questions to this list as well. And my 
first reaction was: that cannot be, everybody knows  , how can it 
_not_ be XML?

The thing is, XML is a very generic and expandable language, and 
entities is one thing that can be expanded upon (above the five that are 
always allowed: < > &, &apos and "). This is done by 
declaring entities in DTD declarations like Patrick suggested, or can be 
done by using an external DTD file and link to it.

If your input comes from XHTML or HTML, this happens often. The fix is 
to use the original doctype declaration and make sure that the DTD's it 
refers to are available. That way other entities like —, ¨ 
© are also recognized in the majority of cases.

You can find the declaration of all these entities here: 
http://www.w3.org/TR/xhtml1/dtds.html#a_dtd_Latin-1_characters, it also 
shows a typical declaration for use in XML. Download the file at 
http://www.w3.org/TR/xhtml1/DTD/xhtml-lat1.ent, use it locally to refer 
to it and you can work with almost all XHTML/HTML input, as long as the 
rest is well-formed.

Kind regards,
Abel Braaksma



------------------------------------------------------------------------
From: 	Michael Kay <mike@xxxxxxxxxxxx>
Sent: 	Wednesday, August 10, 2011 10:19:17 AM
To: 	xsl-list
Cc: 	
Subject: 	Re: [xsl] nbsp fails transformation



> > Now since i can't even transform those files i can't throw those
> > entities out.
> >
> > How do i handle this  !?
> If someone sends you a document that isn't well-formed XML, the best 
> strategy is to get the people who produced it to mend their ways. Once 
> you start accepting bad XML (or non-XML, as I prefer to call it), all 
> the benefits of using XML for interchange quickly become lost, and you 
> might as well revert to using some proprietary interchange format.
>
> Michael Kay
> Saxonica