TBH, my protocol for figuring out what a valid URL is has been entirely inductive (in the non-mathematical sense) so far -- basically I come up with an approximation and then look for specimens of URLs that fail. This bug does remind me, though, that I really ought to create a simple configuration file so people can work around problems on their own until I patch them. (the regexp you want is in urlscan.py right now :) )
Daniel
signature.asc
Description: Digital signature
