Re: Tests: OFFSET and LIMIT

On 26 Jul 2007, at 13:21, Seaborne, Andy wrote:

> Steve Harris wrote:
>> On 25 Jul 2007, at 21:35, Lee Feigenbaum wrote:
>>> Seaborne, Andy wrote:
>>>> Some tests for LIMIT/OFFSET
>>>> All the test involve ORDER BY so that the results are consistent.
>>> Glitter passes all but slice-4 and slice-5. Upon further   
>>> inspection, these tests (maybe others?)
> There are no others.
>>> rely on sorting xsd:integer  typed literals and plain literals in  
>>> a consistent manner, which I  think we agreed is undefined by the  
>>> spec?
>> I fail slice-4 & 5 for the same reason.
> Fixed in CVS.

Excellent, I can now pass all of them. I actually don't without some  
sed action, but that's a known parser deficiency, the tests get my  
thumbs up.

> By the way - these are draft tests - you can fix them up.

I wasn't sure they were wrong. I have the impression that I've  
implemented < as per the spec, but one can never be sure.

- Steve

Received on Thursday, 26 July 2007 14:55:21 UTC