Richard Meyer							Feb. 15th 1994




			Test-Suite for Wild-LIFE
			========================


If you want to check that the Wild-LIFE interpreter has been correctly compiled
and conforms to expected behaviour, then run the script: 'check_all'.  This will
generate the log file 'check.log' containing a list of all test files and
which ones failed.

The test-suite consists of more than 300 LIFE programs of all sizes and
complexities.  The "check" script runs the programs and compares both the
output and error messages to the 'reference' data. This suite is run each time
a change is made to the system to ensure that nothing has been broken or that
all differences in behaviour are explainable and correct.

Certain files may have unavoidably different output depending on your system, in
most cases this is benign. We have made sure that no absolute path names are
compared or CPU times. Notable differences are:

  - Floating point calculations: NaN, Infinity etc... may not be supported
    or may displayed differently (NAN, INF etc...).

  - Word size: on Alpha (and other 64-bit CPUs) some of the tests which check
    the encoding of the type hierarchy produce codes of different lengths. This
    test-suite is configured for a 32-bit system.

As a result, do not be surprised if the following tests result in
different output on an Alpha:

  andor5.lf
  babcode.lf
  caret.lf
  not.lf
  t.lf
  t8104.lf
  z_gctoken1.lf

The largest failing of the test suite is that it cannot check graphically
interactive programs, typically the X interface.

Directories & files:

LF		: source of test programs
IN		: terminal input to test programs

OUT		: stdout of last test run
REFOUT		: reference output
REFDIFF		: difference between output and reference output

ERR		: stderr of last run
REFERR		: reference error output
ERRDIFF		: difference between error output and reference error output

FILES		: auxiliary files used by certain tests.

check		: script to run one or several tests
approve		: script to approve the output of a test
showdiff	: script to display the differences graphically
check_all	: script to run all the tests
check.log	: log of all tests (created by "check_all")
README		: don't read this file if you don't have an occurs-check ;o)
