Re: SSQLSv2 design discussion

[Warren Young]

Joel Fielder wrote:
> Sorry, I wasn't actually particularly clear with the whole transform
> thing - it would need to be possible either at field level *or* object
> level.  LocaliseDate would therefore need std::string
> operator()(mysqlpp::Date& date) to be applied to each date field as it's
> bunged into the SSQLS, whereas an object transformer would have SQLStock
> operator()(mysqlpp::Row& row).

Perhaps you are overengineering it.  What's wrong with just subclassing 
mysqlpp::Date and telling ssqlsxlat to use your subclass?  Every 
assignment to it can run the text through your extra logic.