Our regression test suite had a failure rate of about 15/300 tests.
This morning's run was 50/300 tests - it was obvious that the latest
checkins were the problem. Hopefully the rest of the team will also
see this. The confounding problem is that one of the failures looks just like a
known intermittent failure - but it was happening at a rate of and now happened more like 15/300 or 20/300. Also some of them are