This isn't your fault, Branden, it's the fault of the current implementation. But I would really prefer to see a client-server contract that specified what *was* being searched for rather than what was *not* being searched for. (See [1] for a closely related point.) While the team discusses your question, could we maybe firm that contract up?
Best, Ray [1] https://jira.sakaiproject.org/browse/KERN-2828 On 6/27/12 4:50 AM, Branden Visser wrote: > Hi all, > > I'm cleaning up some search logic divergence as part of KERN-2908. > These two types of searches search for content by aggregating a list > of groups to which the user is associated (either directly or > indirectly). I'm wondering if some subtleties in group aggregation > should have been applied to both queries. Here is a summary: > > The /var/search/pool/auth-all.json search will include indirect groups > at an uncontrolled level of recursion, but exclude groups that are > deemed not to be "User Facing". This excludes: > * pseudo-groups > * groups that have sakai:excludeSearch=false > * contact groups > * collections > > All the /var/search/pool/me/* searches will include ALL indirect > groups at a controlled level of recursion (maximum of 6 levels of > indirection). Should any of the above exclusions have been applied to > these searches as well? At the very least, including those with > sakai:excludeSearch seems wrong to me. > _______________________________________________ oae-dev mailing list [email protected] http://collab.sakaiproject.org/mailman/listinfo/oae-dev
