Re: Avoiding boneheaded mistakes in XSLT?

On Thu, 30 Dec 2010 13:53:42 +0000
Michael Kay <mike@xxxxxxxxxxxx> wrote:


> The problem is that if someone writes
> 
> if (@version = "2.0") ...
> 
> and there is no @version attribute, it's very likely that the
> semantics of the vocabulary are that an absent version attribute is
> to be treated as if @version="1.0". So a warning that there is no
> version attribute would be inappropriate here.


You're making assumptions Mike?
  Simple fact, the author requested an attribute that
is missing from the instance.

If set to report it, IMHO that would be useful information
from Saxon for debugging?






-- 

regards 

-- 
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk