-
jplang authored
Instead of counting and retrieving results based on their timestamps, we now load all result ids then load the appropriate results by their ids. This also brings a 2x performance improvement as we search tokens in one of the 2 queries only. git-svn-id: https://svn.redmine.org/redmine/trunk@13739 e93f8b46-1217-0410-a6f0-8f06a7374b81
0b6f6362