Hi Steve, and good work finding the place to change.
I think the change will need to be more complicated. One of the possible use cases right now is the ability to have very short hold shelf expire times -- think of a 4-hour window to come get some high-demand object from the desk.
We can't really base that on anything circ related, such as fine interval, since the circ doesn't exist yet, so I think the best option is probably a YAOUS that indicates whether the shelf lifetime is date-granular or precise. Then the library can decide. We already pull two other org settings for use in this query, so a third shouldn't materially impact performance.
There's also the opportunity to provide a tiny optimization by using a literal in the query ('today'::timestamptz or 'now'::timestamptz for date-granular or precise, respectively) rather than functions.
Hi Steve, and good work finding the place to change.
I think the change will need to be more complicated. One of the possible use cases right now is the ability to have very short hold shelf expire times -- think of a 4-hour window to come get some high-demand object from the desk.
We can't really base that on anything circ related, such as fine interval, since the circ doesn't exist yet, so I think the best option is probably a YAOUS that indicates whether the shelf lifetime is date-granular or precise. Then the library can decide. We already pull two other org settings for use in this query, so a third shouldn't materially impact performance.
There's also the opportunity to provide a tiny optimization by using a literal in the query ('today' ::timestamptz or 'now'::timestamptz for date-granular or precise, respectively) rather than functions.
Thoughts?