On 07/11/12 18:36, Rob Vesse wrote:
Ok good to know

If you can ping the list when the Turtle test suite is ready for public
consumption that'd be great

Will do.

Also, if you get a chance to use the test before they move or after, and think something isn't covered then shout.

I'd like to get a really quite tight set of tests. Turtle is sufficiently simple that I don't see why the test suite should not be very comprehensive. The triple generation is really pretty simple so it's about getting coverage in the syntax tests.

        Andy


Rob


On 11/7/12 10:26 AM, "Andy Seaborne" <[email protected]> wrote:

The tests are temporarily ...

http://svn.apache.org/repos/asf/jena/Experimental/riot-reader/testing/RIOT
/Lang/
  {Turtle,TurtleSubm}  + N-Triples syntax tests

but I have an action from RDF-WG to move them (Turtle, TurtleSubm,
N-Triples) to W3C space.  And add a README.

The submission tests are cleaned up ones - test submission 29 has been
removed ({|} can't in in IRIs even if they can be in RDF URI
References), and 28 corrected.

The Turtle/ are all new.

I'm due to merge the sets together as well - copy over the evaluation
tests from the submission - so there is one set for Turtle.

        Andy




On 07/11/12 18:06, Rob Vesse wrote:
Hey Andy

When you say you pass all tests I assume you are ignoring test-28.ttl
per

http://lists.w3.org/Archives/Public/public-rdf-comments/2012Jul/0049.html
?

Are the tests the ones currently at
http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/tests/ or is
there a
newer and more complete test suite floating around somewhere?

Rob



On 11/3/12 8:34 AM, "Andy Seaborne" <[email protected]> wrote:

I've got round to updating the Turtle parser (RIOT) for the changes for
the Last Call publication of the Turtle spec from the RDF-WG.

Editors working draft:
http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html#


The writers have not been updated.

The changes are all additional things except one item which is an
alteration.

Added:
1/ The local part of a prefixed names can have %xx

Now legal:

ns:foo%23bar

and this puts %-2-3 into the URI (it does not convert it to numeric
codepoint i.e. #)

2/ The local part of a prefixed names can have character escapes

'_' '~' '.' '-' '!' '$' '&' "'" '(' ')' '*' '+' ','
';' '=' '/' '?' '#' '@' '%'

Now legal:

ns:foo\#bar

3/ PREFIX and BASE  (any case) are synonyms for @prefix and @base
except
they don't take a DOT at the end of the declaration.

Change:

4/ XSD decimals now need a digital after the decimal point in their
short form.  Previously, the character sequence 123. used to be an XSD
decimal; it is now an XSD integer and a DOT.  As space-DOT is much,
much
more common, the WG hopes this will not be a major issue.  SPARQL 1.1
made the change and so far it has not been an issue there.


The RIOT turtle parser passes the current draft test suite [*], as does
Gregg Kellogg's mods for RDF.rb.  (The test suite is work-in-progress
in
RDF-WG and subject to change.)

        Andy

[*] Poacher and gamekeeper (I wrote the tests) but I promise I wrote
the
tests, *then* made the Turtle parser pass them.  Honest.





Reply via email to