Ethan suspicion is that the problem is actually a general problem with "or" that somehow gets exposed in this options case.
http://test.dwagn.org/wagn/beyond_type+*right+*options does not restrict the autofill that comes up when you try to add pointer items to http://test.dwagn.org/wagn/foo+beyond_type
Issue was that there were multiple wql keys that got translated into "cond" keys, and they were all conflicting with each other. Fixed. (did not write auto test -- boo for me)
Do you want to note that somewhere before we close this?
--John Abbe.....Tue May 31 21:54:23 -0700 2011
I don't think so. the coverage issue is really about the combinatorial problems in wql -- certain things work independently but not together. I fixed one particular combination issue, and it's not the kind of thing that's likely to reoccur. But I'm confident there are other combinatorial issues out there, particularly surrounding relationships that involve lots of joins.
Anyway, the point is that we need a systemic approach to feel good about our wql coverage, so I wouldn't hold up this ticket for that.
--Ethan McCutchen.....Wed Jun 01 09:33:53 -0700 2011
--John Abbe.....Wed Jun 01 14:53:04 -0700 2011