Re: TriG : call for technical review

On 07/09/13 20:04, Gregg Kellogg wrote:
> On Sep 6, 2013, at 9:50 PM, Gregg Kellogg <gregg@greggkellogg.net> wrote:
>
>> On Sep 6, 2013, at 6:20 PM, Gregg Kellogg <gregg@greggkellogg.net> wrote:
>>
>>> On Sep 3, 2013, at 2:33 AM, Andy Seaborne <andy@apache.org> wrote:
>>>
>>>> The TriG document should now be up-to-date with the WG decisions on datasets.  While it still requires editorial work, it is now in a state where technical review from the WG would be beneficial.
>>>>
>>>> Please could we have volunteers to review the document in preparation for Last Call.
>>>
>>> This is only a partial review, based on updating my own implementation.
>>>
>>> Production [5g] wrappedGraph does not match the empty graph. This is because wrappedGraph is defined as '{' triplesBlock '}' and triplesBlock is defined as triples ( '.' triplesBlock? )? As the triples production must have a subject, there's no way to state <g> {}. It could be changed to '{' triplesBlock? '}' to allow this, although there may be a better way.

Fixed - thanks for catching that.  That seems like the right way to fix it.

	Andy

>>>
>>> The following tests should be removed, as they were speculative: trig-bnodeplist-graph-01, trig-collection-graph-01, trig-collection-graph-02
>>
>> On further investigation, these three tests were in the tests2 directory, which included tests for the updated grammar. At this point the other tests in the tests2 directory should be merged with tests. Also, they really need comments, otherwise it's completely unclear what they're testing, and this shows up in the EARL report.
>
> I consolidated the tests into a single manifest, added comments for each test, and changed the three tests I mentioned into negative syntax tests.
>
> Gregg
>
>>> Otherwise, the grammar looks cleaner than before, and it runs as expected.
>>>
>>> Gregg
>>>
>>>
>>>> 	Thanks
>>>> 	Andy & Gavin
>>>>
>>>
>>
>

Received on Saturday, 7 September 2013 20:14:39 UTC