Re: [CDBI] Make name_lc go fast

Aaron Trevena wrote:
> I kind of assumed that was all it was doing anyway - so I'd be glad to
> see it optimised away as much as possible, but I still don't see it as
> a significant bottleneck.

Its called every single time a column is used as a string.  Which means
retrieve, insert (many times via find_column), and the accessor (via _attrs).
 The accessor is the killer.  Anything that speeds up the accessor to access
already retrieved data is a good thing.