Re:Proving that Android’s, Java’s and Python’s sorting algor...
by krick on Tuesday February 24, 2015 @12:24PM
(#9102035)
Does it "change behavior"? As far as I understand, the only change is that this test wouldn't crash anymore. Which certainly is "changing the behavior", but allocating more memory is as well. If so, I don't see any reason not to use formally verified version. Not that it is really important, but quite reasonable.
I wouldn't worry about an extra couple of bytes nearly as much as I would worry about changing the behaviour of a function used in an "astronomical number of program", so this looks like a pretty reasonable and conservative choice, at least for an immediate patch.