Re: Turtle tests blank ID patches, and EARL report for Serd

On Mon, 2013-02-25 at 16:06 -0800, Gregg Kellogg wrote:
> Thanks David, I encountered a couple of minor issues in incorporating your output in to the rollup report:
> 
> Your FOAF document is RDF/XML, but the content-type returned is text/html, so the tool fails to parse it because it selects the RDFa/. Easiest may be to just include the relevant resources in the EARL report directly as Turtle.

The document is conneg'd, but the MIME type for the XML form was wrong.
It has been fixed, you should get application/rdf+xml.

Note that if your client requests text/turtle or application/x-turtle
you will get a Turtle document.

> The tests you asserted which are not defined in either maniffest are ignored in the results.

Oops, this is because the URI for the tests-ttl tests was wrong (the
tests are run directly from the manifests and only include those).  A
(hopefully) fixed report is attached.

> I've updated the report with your results [3].

Great, thanks.

> P.S., I'll leave it to others in the WG to respond to your test suggestions. I know that the coverage tests are intended to cover 100% of the features, and to the degree that they don't, additional tests would be appreciated.

I will try to find some time to extend the test suite with a proper
manifest and such.  Even if it is not formally adopted I suppose it
would still be useful.

>  I have no problem with changing blank node labels and re-ordering.

This would be very appreciated to minimize divergence on my end (and for
any other streaming implementations that come along).  Without those
changes the test suite is much more difficult to rigorously verify.
When testing, trust nothing :)

Thanks,

-dr