This should provide the times that would be yielded by a time series query in issue #34, but without computing the result.
Hmm, that implies that we must yield times for any event that could influence the result of the query, regardless of whether the value actually changes at that time from right-limit of its value.
This should provide the times that would be yielded by a time series query in issue #34, but without computing the result.
Hmm, that implies that we must yield times for any event that could influence the result of the query, regardless of whether the value actually changes at that time from right-limit of its value.