TUnit

From TinyOS Wiki
Revision as of 04:25, 23 November 2009 by Sissou (talk | contribs) (Case Studies)
Jump to: navigation, search
Tunit.png

TUnit is a unit testing framework geared for TinyOS and sensor networks. The philosophy surrounding the design of TUnit stems from years of accepted industry standard testing in other languages. TUnit is used daily in TinyOS Test Driven Development, decreasing the time spent manually testing and debugging software by literally hundreds of man-hours.

Open Source Testing

The TUnit unit testing framework is available open source to be used both publicly and privately.

Unit tests can be contributed by the public to the public automated unit testing system in tinyos-2.x-contrib/tests. These tests are automatically checked out and executed on several dedicated mote hardware platforms. Results and lava lamp status indicators are available live at http://www.lavalampmotemasters.com.

Privately, any group can run TUnit standalone or as an automated process in their own team development environment.

Tutorials

Case Studies

  • TestRxFifo
    • Isolation test testing the receive functionality of a radio stack from the rest of the system, allowing us to test edge cases observed by receiving various types of data in the software implemented RX FIFO.
  • CsmaBehavior
    • Isolation test of the CSMA layer of the CC2500 radio stack.
  • TestMac
    • Multi-Node Testing and Characterization: Multiple statistics logging
  • TestRxThroughputWithNoCca
    • Multi-Node Test: Upon receiving the first message from the Driving Node, the Supporting Node receiver starts a timer and accumulates received messages. When the timer fires, it reports performance and verifies enough packets got through. The performance is characterized with statistics logging.
  • TestSwAcks
    • Multi-Node Test: An idle Supporting Node simply sends back an acknowledgment to the Driving Node.
  • TestLplDefaultContinuousDelivery
    • Multi-Node Test: A Test Suite where 2 to 19 Supporting Nodes deliver LPL messages to a single listener, the Driving Node. At the end of a period of time, a dynamic threshold is calculated at +/- 10% and assertions are made against each of the nodes in the network to determine if the channel is being shared fairly amongst the multiple transmitters.

Progress

External Links