Difference between revisions of "RDM Responder Testing"

Revision as of 22:34, 1 January 2011

OLA comes with a suite of tests for RDM responders which can be used to check how well an RDM device implements the E1.20 specification. Writing RDM Responder Tests explains how to add to / improve the tests.

Setup the Test Rig

The following controller devices are supported:

• RDM-TRI
• DMXter4 RDM / MiniDMXter

Connect the device under test to the controller device and start olad. Patch the output port on the controller device to a universe (UNIVERSE_NUMBER). Then run ola_rdm_discover, you should see the responder's UID appear:

 $ ola_rdm_discover -u UNIVERSE_NUMBER
 00a1:00010003
 7a70:ffffff00

Running the Tests

The tests are written in Python and run using ola_rdm_test.py. Below is the output from a typical test run:

 ./ola_rdm_test.py --universe 1  --pid_file ../../python/pids.config  00a1:00010003
 Starting tests, universe 3, UID 00a1:00010003
 SetManufacturerLabel: Passed
 SetSoftwareVersionLabel: Passed
 GetManufacturerLabel: Passed
 GetSoftwareVersionLabelWithData: Failed
 ...
 ------------- Warnings --------------
 ------------ By Category ------------
   Product Information:  7 /  7   100%
       RDM Information:  1 /  1   100%
    Core Functionality:  2 /  2   100%
      Error Conditions: 10 / 16   62%
          DMX512 Setup:  3 /  3   100%
 -------------------------------------
 29 / 30 tests run, 23 passed, 6 failed, 0 broken

Useful Options

ola_rdm_test.py has some options which can assist in debugging failures. For a full list of options run with -h

-d, --debug

Show all debugging output, including actual & expected responses.

-l, --log

Log the output of the tests to a file. The UID and timestamp is appended to the filename

-t Test1,Test2 , --tests=Test1,Test2

Only run a subset of the Tests. Only the tests listed (and their dependencies) will be run.

Information on Tests

Some tests have dependencies, which are other tests that need to be completed before the test can be run. Dependencies can be used to check for supported parameters and other conditions that may affect responder behavior.

There are 4 result states for a test:

Passed

The responder replied with the expected result

Failed

The responder failed to reply, or replied with an un-expected result

Not Run

This test wasn't run because the responder doesn't support the required functionality

Broken

An internal error occurred, this indicates a programming error or an error with the test rig.