Demjava regression tests.

Quick Start:

To run regression tests on a version of demjava set up your demeter
variables (DEMJAVA_HOME, etc) to point to the version of demjava
that you'd like to test and type 'RunRegression' in this directory.

Usage:

The RunRegression script takes one argument, the path to the directory
that contains the test suites it should run.  It runs all the suites
in its target directory, prints a summary of their results and exits.

A test suite is a directory whose name begins with the prefix 'suite-'
which contains any number of tests.  A suite is executed by running
each of the tests it contains.  A suite is successful iff all of its
tests execute sucessfully.

A test is a directory whose name begins with the prefix 'j-' which
contains a make file with the targets 'make clean' and 'make test'.  A
test is executed by running 'make clean' followed by 'make test' in
its directory.  A test is successful if the word 'SUCCESS' and none
of the following list of words appear in the output of 'make test':

Parse error
Syntax error
Compilation Failed
Invalid
NullPointerException
Memory fault
FAILURE

The argument to RunRegression can be a directory that doesn't contain
any suites but does contain some tests.  In that case, the tests are
run.

Each test's 'make test' output is saved in a file named 'output' in
its directory. (if the test is named j-bypassing1, its output will
be left in the file j-bypassing1/output).

Adding New Regression Tests:

Make a directory with a name that begins with the prefix 'j-'.  Put a
makefile in it with the targets 'clean' and 'test'.  Make the 'test'
target print 'SUCCESS' if the test succeds and 'FAILURE' if it
doesn't.  Finally, move the test directory into a suite directory if
you feel like it.

NEW---
Now 'RunRegression' does what it used to previously, and in the end
shows all the failures(only). 
It also creates a file "results.failure"

In order to see only the new failures, create a file "known.failure"
and  run 'CheckFailure' .

"known.failure" is a simple file containing the path of each test-case
known to fail.
A simple way to create the log "known.failure" is to run the script 
'MakeKnownFail', which creates the file from "results.failure". If
required, you can later edit the "known.failure".