JMeter will create one instance of a JavaSamplerClient implementation for each user/thread in the test. Additional instances may be created for internal use by JMeter (for example, to find out what parameters are supported by the client).
When the test is started, setupTest() will be called on each thread's JavaSamplerClient instance to initialize the client. Then runTest() will be called for each iteration of the test. Finally, teardownTest() will be called to allow the client to do any necessary clean-up.
The JMeter JavaSampler GUI allows a list of parameters to be defined for the test. These are passed to the various test methods through the {@link JavaSamplerContext}. A list of default parameters can be defined through the getDefaultParameters() method. These parameters and any default values associated with them will be shown in the GUI. Users can add other parameters as well.
When possible, Java tests should extend {@link AbstractJavaSamplerClient AbstractJavaSamplerClient} rather than implementing JavaSamplerClientdirectly. This should protect your tests from future changes to the interface. While it may be necessary to make changes to the JavaSamplerClient interface from time to time (therefore requiring changes to any implementations of this interface), we intend to make this abstract class provide reasonable default implementations of any new methods so that subclasses do not necessarily need to be updated for new versions. Implementing JavaSamplerClient directly will continue to be supported for cases where extending this class is not possible (for example, when the client class is already a subclass of some other class).
See {@link org.apache.jmeter.protocol.java.test.SleepTest} for anexample of how to implement this interface. @author Brad Kiewel @author Jeremy Arnold @version $Revision: 1.7 $
| |