TCK_Contributors_Guide

Guide for contributing a new TCK test (at least an outline of a guide)

Some Basic Test Structure

• The typical test method begins by submitting a job via JobOperator#start.

  • There is no order of execution assumed among the tests, so to do pretty much anything requires a job to have run within the test method.

• In general, the validation logic (i.e. the check for success/failure) is either performed:

  • within the TestNG test method at various points within the sequence of calls to the job repository (i.e. to JobOperator).

  • OR within the job execution (i.e. within the batch artifacts themselves)

  • OR within some combination of both.

• The tests use the JobOperatorBridge utility provided by the TCK in order to encapsulate the waiting for "termination" (i.e. waiting for COMPLETED, FAILED, or STOPPED).

• Any test with any sort of wait introducing a timing dependency will include a specific, well-documented system property for each individual wait, in each individual test, to allow the executor to tune for his particular implementation and environment. A hopefully well-chosen default value will be hard-coded as well.

What files make up a single test?

1. First, the TestNG test method

   • Annotated with @Test (imported from org.testng.annotations)

   • Also annotated with @org.junit.Test

   • Using

Checklist

a *com.ibm.jbatch.tck.tests*

, @Test annotations - logging - bridge, - waiter - annotation - timeout (if necessary) * JSL * artifacts (impls of JSR 352) com.ibm.jbatch.tck.artifacts

• batch.xml

• update TestNG suite (if necessary)

• Maven

• bugzilla

• RI/spec