On Jan 9, 9:39 pm, Fredrik Lundh <[EMAIL PROTECTED]> wrote: > Paddy wrote: > > To see how they act against 'corner cases' and > > an exercise for me in trying to create corner cases. (I'm in to > > functional testing at the mo'). > > sounds more like "pulling requirements out of thin air". not sure that > helps the OP get a better understanding of Python, really. > > </F>
Not really out of thin air. The OP transforming alternating occurrences of _ so corners would include a null string; strings with 1, 2 and 3 occurrences of the character to be replaced and the char to be replaced separated or not by other chars.... Ideally I would create a randomized interesting string generator and try each prog against a lot of generated examples and computed 'correct' answers, but I am unsure of what should be the right answer in some cases so the earlier posting can be used to give extra information for the spec. to be fleshed out with. You may be right in that the OP might see the extended input and dismiss them because he knows that the input data is never going to be like that - or - he may find that his input validation might well allow some of the cases through and he needs to either improve his input validation or define what to do with more types of input. Your also right in that its mostly not Python specific but it helps me think of corner cases and the interpretation of specs which is very important in helping solve the right problem. - Paddy. -- http://mail.python.org/mailman/listinfo/python-list
