test: add dynamic test discovery and validity checks.

Modified benchmark_test.go to dynamically discover all .test files in the
tests directory instead of using hardcoded paths.
Added TestSamplesValidity integration test that validates each test file
against its corresponding .expected file.
Added runSampleWithOutput helper function to capture interpreter output.
Added new test cases with expected outputs for validation.

tests/church_5^5.test

@@ -0,0 +1,8 @@
+0 := \f.\x.x
+inc n := \f x.(f (n f x))
+exp n m := (m n)
+print n := (n F X)
+
+N := (inc (inc (inc (inc (inc 0)))))
+
+(print (exp N N))