Standard practice is to include two tests. One is for development purposes and human inspection. That is where the truly non-deterministic test should go. The other test is for automated use and injects a random number generator with a known seed into the class under test. The test results should basically be a frozen version of the development test.
On Sun, Jan 24, 2010 at 2:23 AM, Luc Maisonobe <luc.maison...@free.fr>wrote: > I am puzzled by tests that can randomly fail and belong to an automatic > test suite. > -- Ted Dunning, CTO DeepDyve
