XQuery error codes and rewriting

[David Carlisle]

> The rule about allowing rewrites that don't have the same error behavior is
> mainly there for vendors who translate XQuery into SQL.

What about ones who convert to other languages (er like xslt for
example;-)? The issues are similar: that it's a lot easier to go down,
generating code in the supporting language that does the right thing on
correct input than it is to trap any error condition that this other
language may raise, and map that to the specified XQuery error
condition.

> Another approach is the XPath 1.0 philosophy that errors don't happen

Yes that's one extreme:-) Although the problem here is not that errors
are specified as happening, but just that they have specified unique
identifiers. Getting them right is a lot of work, and if it's a person
rather than a machine on the receiving end of the error message it's
probably a complete waste of time as they'll skip over the code and read
the error message. I just thought it was worth musing in public as to
whether the balance is really right but wanted to do it here rather
than as an official comment on the spec, because it _is_ just musing in
public.


David

________________________________________________________________________
This e-mail has been scanned for all viruses by Star. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
http://www.star.net.uk
________________________________________________________________________