I've come to the conclusion our tests are too fragile. Many of the tests depend on test(s) prior to run so if test 3 of 50 fails 4 -through 50 don't get run. I recall reading about ways around this but can't seem to find them. Here's the scenario;
test one (no dependencies)
test two depends on one
test three depend on two
etc.
My thoughts (from what I remember reading) is the dependencies for each test other than one) already exist so that if a test fails subsequent test(s) can be completed.through the entire test run.
The error hander would move to the next test upon failure using Runner.Stop(true) in the 'OnLogError' event.
I'm sure this has been discussed before but I've been unable to find the posts.
