#{

All tests are expressed as pairs of verbatim strings with tag "test":
@[test(id)...], where "id" is optional and use only for error reporting.

The first of each pair is considered as codes to be tested, and the second is
the expected results to which the results of the first to be checked against.

The expected results can be expressed in three forms (are checked in this order):
1. Value that should be produced by the test codes;
2. Raw printed output;
3. String pattern that should match the printed output;

The first is checked only if the test codes do not print anything,
and the expected result is not an empty string.

#}


# A positive test that produces a tuple (1,2):
@[test(first)]
	a = 1
	b = ( a, 2 )
@[test(first)]
@[test(first)]
	( 1, 2 )
@[test(first)]


# A positive test that prints "hello world!":
@[test(second)]
	io.writeln( "hello world!" )
@[test(second)]
@[test(second)]
	hello world!
@[test(second)]


# A negative test that should raise a parsing error:
@[test(third)]
	a : array<string>;  # array type cannot contain string
@[test(third)]
@[test(third)]
	Invalid %s type %s form
@[test(third)]