rtest [testcase[.test] ...]
This is the most simple case. You have some commands that produce output and it must match the sample you give.
TEST_DEF TestName Short Description <<END_TEST one or more shell commands .... ___EXPECT___ expected output from the commands above END_TEST
In this case, the sample output comes from a separate file (testname.baseline). You would use this form when the output is really large, or binary
TEST_DEF TestName Another Description <<END_TEST ___EXPECT_FILE__ testname.baseline END_TEST
see the test cases in the source code for more examples