OK. I will come up with a plan to extend the custom validators and not break backward compatibility. It seems like we could pass a keyword variable to be used instead of None should the value be empty for the formatter. Maybe a separate "empty" value for the parser (input to database) and formatter (output from database).
Looks like I'll be getting into those "diff" tools for proper patch submission after all... ;-) -- Joe B. On Monday, October 29, 2012 8:25:19 PM UTC-7, Massimo Di Pierro wrote: > > I am reverting this. sorry. > > On Monday, 29 October 2012 10:09:35 UTC-5, Massimo Di Pierro wrote: >> >> This is what was talking about. I have fixed all the validators but think >> may break third party custom validators.... this change is not backward >> compatible. >> >> >> --
