> Task :test CustomScriptTest > testScript(String) > net.sourceforge.kolmafia.CustomScriptTest.testScript(String) FAILED org.opentest4j.AssertionFailedError at CustomScriptTest.java:59 MaximizerTest > maximizeShouldNotRemoveEquipmentThatCanNoLongerBeEquipped() FAILED org.opentest4j.AssertionFailedError at MaximizerTest.java:102 MaximizerTest > maximizeGiveBestScoreWithEffectsAtNoncombatLimit() FAILED org.opentest4j.AssertionFailedError at MaximizerTest.java:74 529 tests completed, 3 failed, 7 skipped > Task :test FAILED
Ok, so in test, we want to start from a clean, predictable state. especially with the customScriptTest, which is an all-or-nothing text compare. Tests should either be able to count on a state at start or set one.My biased opinion is that adding defensive resets to unrelated tests is the wrong place to make these changes. Either fix the problem at the source, or fix it systematically in the test runner framework.