Re: [OK?] Re: SPARQL: Error handling

On Mon, 2006-01-30 at 00:02 -0500, Kendall Clark wrote:
> On Jan 29, 2006, at 11:45 PM, Dan Connolly wrote:
> >> Is there any reason to state the fault propagation rule redundantly?
> >
> > I'm having trouble seeing the redundancy; where does it already say
> > that Out Message must not be returned in the case of syntax errors?
> 
> The spec says that in the case of a syntax error, the MalformedQuery  
> fault should be returned.

That's not what I'm looking for. I'm looking for the text
that follows from this WG decision:

"for queries that are not SPARQL Query Strings, you should return
MalformedQuery and you must not return 2xx"
 -- http://www.w3.org/2005/08/16-dawg-minutes#item04


>  If this (or any other) fault is returned,  
> no Out Message can be returned, by the WSDL 2.0 fault propagation  
> rule our spec uses.
> 
> Thus, if you return MalformedQuery, you cannot return an Out Message.  
> That's what it says by saying that we're using Fault Replaces Message  
> fault propagation rule.
> 
> I guess it's more accurate to say that it implies this prohibition  
> rather than explicitly states it. But isn't that editorial rather  
> than substantive?
> 
> >> At any rate, in 1.108 I've updated this section in response to this
> >> discussion.
> 
> Is the relevant language insufficient as it stands?

Yes; that doesn't say that Out Message *MUST NOT* be returned
in the case of a syntax error.

> MalformedQuery
> 
> When a SPARQL query string is not a legal sequence of characters in  
> the language defined by the SPARQL grammar, this fault message should  
> be returned. See 2.1.4 query Fault Messages for SPARQL Protocol fault  
> propagation rules. In the case of HTTP bindings, an HTTP 2xx status  
> code must not be returned.
> 
> Cheers,
> Kendall

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Monday, 30 January 2006 13:50:52 UTC