RE: Automatically copying an element's attributes and

[Michael Kay]

> I'm interested in people's views on "recoverable errors". In 
> the case of
> errors like this that are clearly under the control of the stylesheet
> author, should we be making them non-recoverable (fatal) 
> errors? Most of
> the new run-time error conditions in XSLT 2.0 are non-recoverable, but
> the old 1.0 ones haven't changed much. I think the original reasoning
> was that with client-side transformations in particular, it's 
> better to
> put anything on the screen rather than a syntax error message. Does
> anyone think that's a valid argument?

I think that recovery from errors in this class should be
imlementation-selectable, possibly with a specified default.  That way,
a processor could be used where you really want to recover and proceed,
as in a browser, and also where you want to use a different strategy.
The alternative would presumably be to throw an exception which could be
handled in some way that would allow the processor to continue if
desired.  This would give the same result, and would offer a way to give
the user a notification about the error.

Cheers,

Tom P

 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list