[EMAIL PROTECTED] wrote:
Yip Ng <[EMAIL PROTECTED]> writes:
Some observations:
If org.apache.derbyTesting.functionTests.harness.RunTest is used to run
views.sql, then it have 2 newline diffs.
If org.apache.derbyTesting.functionTests.tests.lang.LangScripts is used to
run, then the ordering of one of the outputs differs.
I filed a jira for this: DERBY-2048
Hmm, I think this demonstrates the weakness of canon-based testing,
regardless of the harness being used to run them. In this particlular
case, using JUnit actually makes it more difficult since it doesn't
show you a context diff...
Not sure JUnit makes it any more difficult, the diff the old harness
used had its problems, and didn't produce a context diff. If I see a
failure I use the following, where $DM is the location of the master files:
diff -c $DM views.out
Dan.
