Fredrik Lundh wrote: > (and there was never 96 real errors in the code base, of course; > coverity is a remarkable tool, but there's plenty of room for > mistakes when used on a code base this large)
Yep - from the discussions on python-dev, a lot of the "error-fixing" was telling Coverity what things it could safely ignore ;) However, there were definitely things found that were real bugs, and subsequently fixed. It also provoked at least one discussion where the code was iffy, it was entirely understandable that it was flagged, and with some reworking it would become unambiguous. I'm not sure what the final result was though ... Tim Delaney -- http://mail.python.org/mailman/listinfo/python-list