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

On Tue, 2013-03-26 at 01:06 -0400, Eric Prud'hommeaux wrote:
> On Feb 24, 2013, at 10:13 PM, David Robillard <d@drobilla.net> wrote:
> 
> > I finally took some time to implement most of the new Turtle spec.
> > 
> > To be able to run the test suite in my existing test framework, I needed
> > to fix up the test output slightly.  Attached is a patch to:
> > 
> … [eliding bnode label issues to be addressed in a separate sub-thread]
> > * Re-order some triples in the tests-ttl output, so output triple order
> >   matches input triple order (for streaming parser/serialisers)
> > 
> > These change make testing streaming implementations nicer, but shouldn't
> > break anything since the RDF is the same.
> 
> Per this comment and comment 22
> <http://www.w3.org/2011/rdf-wg/wiki/Turtle_Candidate_Recommendation_Comments#c22>, 
> turtle-subm-26.nt has been reordered, see <https://dvcs.w3.org/hg/rdf/rev/4c8bfe7ae424>.
> 
> 
> If you are satisfied with this resolution of ordering in Turtle tests,
> please reply with [RESOLVED] in the subject.

The new turtle-subm-26.nt matches the output of serdi exactly.

However turtle-subm-06.nt and turtle-subm-08.nt still do not (the
required changes were also in my earlier patches).  For inline blank
objects, to be able to do streaming+abbreviating the statement with that
as an object *before* it emits the following statements using it as a
subject, i.e. the statement order exactly as they are encountered in the
input.

Otherwise, if you attempt to re-serialize those triples as abbreviated
Turtle, you don't have the first statement at the time you need it, i.e.
it is impossible to reconstruct the Turtle input from the output in
constant space.  This can be very important when files get very large.

Not a big deal for me, but the order of these would have to be fixed if
my implementation (or any streaming+abbreviating implementation) is to
pass the test suite exactly without modification, which would be nice.

Thanks,

-dr