- From: Matthew Perry <matthew.perry@oracle.com>
- Date: Fri, 03 Feb 2012 16:04:32 -0500
- To: public-rdf-dawg@w3.org
Hi Greg, I see your point. If you go by the equivalent DROP + INSERT, there's no error to suppress, and :g1 goes away on a quad store. I would be fine with getting rid of these tests. - Matt On 2/3/2012 3:31 PM, Gregory Williams wrote: > On Feb 3, 2012, at 3:10 PM, Matthew Perry wrote: > >> Hi Greg, >> >> Both those tests use SILENT. I was testing to make sure any errors are suppressed. > Matt, > > Sorry -- I missed the fact that SILENT is used. But my point was that the results will be different depending on whether the system uses a graphstore or a quadstore. A graphstore will end up with the results as the test case currently expects (with the graph :g1 being unchanged), while a quadstore will end up with :g1 being empty, because the (empty) graph :g4 is copied/moved to :g1. Does that sound right? > > .greg > >
Received on Friday, 3 February 2012 21:04:58 UTC