Originally posted by mohit bahl:
But my intentions are only to validate that the function calls are returning what they were suppose to
How do you know what they are supposed to return.
and in case there is some code change done by some other developer then its impact can be known with the help of such test cases and can be caught at the time of an integrated build.
This is what the tools claiming to generate tests do. They provide regression tests assuming the current implementation is correct. I haven't found an implementation that I like though. They tend to either not generate good coverage (focusing on throwing bad input values at the function) or are so brittle that they break for normal low level implementation changes.