-
Committer:
stevenbird
-
Date:
2008-02-19 22:41:45 UTC
-
Revision ID:
svn-v3-trunk0:2b9c9e99-6f39-0410-b283-7f802c844ae2:trunk:514
More flexible control of test_case_parts via optional flag
stop_on_fail. This flag can be used at the outer and inner
levels. In the previous configuration, testing stopped
after the first test case failed, but within a given test
case, all test_case_parts were run. Staff requested the
converse behaviour in which all test cases were run, but
within a given test case, only test_case_parts up to the first
failing test_case_part were run. This permitted early
test_case_parts to serve as guards for later ones, simplifying
the logic required for test code (i.e. avoiding the need to
conjoin the complement of the antecedents of previous
test_case_parts to the current test_case_part). So, the default
configuration is:
test_case: stop_on_fail=False
test_case_part: stop_on_fail=True