When there are lot of test failures, we consult results db for every failure. However, we can bypass the retry logic only if ALL the test failures are pre-existing. If even one failure is not pre-existing, we can't benefit from results database lookups. So, we can optimize the number of results database calls, by skipping consulting results-db for further tests the moment we know that one of the test failure is not pre-existing.

Number of test failures (and resulting number of results db calls) can be lot for some queues, e.g.: 170 on GTK-WK2-Tests-EWS queue in https://ews-build.webkit.org/#/builders/35/builds/34489/steps/23/logs/test-failures . Even on mac/ios queue it can be upto 60 failures (after https://commits.webkit.org/257442@main). So this change would help reduce unnecessary results database calls in those cases.