On Tue, Dec 14, 2004 at 09:32:51PM -0600, Michael Carman wrote: > > I think this is the most valuable part of the exercise - the bugs you find > > when you think 'its got 98% coverage, there cant possibly be any bugs > > left...oh, look' > > With a little luck, that's when it finally sinks that coverage != correctness; > it's just a tool to help achieve it. *That* is a valuable lesson.
That's just a corallary to "tests can't prove there are no bugs". But just in looking through the coverage for tiny points that I missed caused me to look at code that I've long ignored, like the failure diagnostic code for is_deeply(). Something I prefer not to look at. It caused me to tidy a few things up and yes, in covering the 4th possible outcome of an xor condition I found ANOTHER BUG. https://rt.cpan.org/NoAuth/Bug.html?id=8865 Minor, but still a bug. I even spotted a test that said this sort of outcome was ok. In the end, its a tool. Like all tools its not so much the tool but the hand that wields it. -- Michael G Schwern [EMAIL PROTECTED] http://www.pobox.com/~schwern/ "Congratulations, you're a thieving bastard... Now give me back my pants." "That's MISTER Pants!" -- Ian's Adventures In Morrowind http://www.machall.com/morrowind/page3.html
