Allow tests to define run_once instead of execute.
This allows us to move a whole lot of boilerplate out of each test into
a common place in test.py
Also allow for a postprocess phase for tests - to provide a cleaner alternative
to run_once
Signed-off-by: Martin J. Bligh <mbligh@google.com>
git-svn-id: http://test.kernel.org/svn/autotest/trunk@1846 592f7852-d20e-0410-864c-8624ca9c26a4
diff --git a/client/tests/tsc/tsc.py b/client/tests/tsc/tsc.py
index 6e7ca64..2dbdcf5 100755
--- a/client/tests/tsc/tsc.py
+++ b/client/tests/tsc/tsc.py
@@ -12,6 +12,5 @@
utils.system('make')
- def execute(self, iterations = 1, args = ''):
- for i in range(iterations):
- utils.system(self.srcdir + '/checktsc ' + args)
+ def run_once(self, args = ''):
+ utils.system(self.srcdir + '/checktsc ' + args)