0.11 (4 Apr 2003):

  * Two new test methods, test_true() and test_false(), are intended 
    to replace test_bool(), which is still supported but now deprecated.

  * Make the default setup() and shutdown() methods empty instead
    of raising NotImplementedError.

  * Revised run_sancho_tests so that directories on the command line
    are always recursively searched for test scripts. The -r option
    is no longer optional.  Added -c option to to get a code_coverage
    summary for each test script.

  * The path of each test script is put on the front of sys.path while
    the test script runs so that modules within the test directory can
    be imported.

  * Fix buglet when a module lacks __name__.


0.1 (14 Jun 2002):

  * Rewrote the reporting mechanism.  The name of the scenario and the name
    of the test case method is now always printed before any failures in
    that method.  The results of tests are represented as an object rather
    than as a tuple.

  * Catch exceptions raised in setup() and shutdown() methods and report
    them.
  
  * Drop test method docstring requirements.  The test counts proved to be
    nothing but trouble.  The test framework knows when a case fails
    anyhow.

  * Drop support for test_cases attribute.  All methods starting with
    check_ are considered test cases.

  * If an exception is raised inside a test_* method, don't print
    redundant excption information.

  * Add simple example.

  * An option to drop into pdb, the Python debugger, on an exception
    has been added. run_sancho_tests also has a new -i option for
    enabling this option.  (Implemented by David Binger)

0.0.0 (29 Jan 2002):

  * Initial release.
