W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > January to March 2007

RE: AndyS's syntax tests now in CVS

From: Lee Feigenbaum <feigenbl@us.ibm.com>
Date: Sun, 11 Feb 2007 17:50:29 -0500
To: public-rdf-dawg@w3.org
Message-ID: <OFB0206079.79957EC2-ON8525727F.00787113-8525727F.007D78C7@us.ibm.com>

Andy Seaborne wrote on 02/09/2007 05:39:21 AM:

> Fixed by:
> 
> 1/ Added "@prefix dawgt: ..." to manifest.ttl
> 2/ Turned "bad-0X" into "bad-3X" in manifest.ttl

Thanks. I added the prefix declaration to the other 3 syntax manifests as 
well.

I also updated the syntax-sparql3/manifest.ttl to reflect some file name 
changes (some negative tests had been renamed to start with syn-bad- but 
the manifest was out of sync with that.)

> The test numbering should be that they can go with tests in
> syntax-sparql3 - that's why the numbering does not start at 1.

Ah, I see. Easier (for me, at least) now to just keep them where they are.

> Results to come.

My results:

The first time I ran it, I hadn't yet updated my code to reflect the fact 
that the same bnode label is prohibited in multiple BGPs. So I was 
surprised when I passed some of the negative tests. I saw that some of the 
negative tests had mismatched curly braces which made them negative for 
unintended reasons. I fixed the curlies in these tests:

syn-bad-35.rq
syn-bad-36.rq
syn-bad-37.rq

After doing this, I failed (i.e., successfully parsed) all the syn-bad 
tests in syntax-sparql4.

Once adapting my parser to reject bnode labels that are used in multiple 
BGPs, I pass all of the tests in syntax-sparql4.

Lee
Received on Sunday, 11 February 2007 22:50:38 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:35 GMT