On Thu, Dec 18, 2008 at 04:05:09PM -0800, Brian Aker wrote: >> drizzle could avoid the horrible slowness of information_schema, and > > In the current roadmap is to rework I_S so the are not materialized (and > use the same execution path as the engine).
With the realisation that looking at datadict is really just looking at tables, and realising that this is a scanning/index method just like normal tables, I'm 99.99993241% convinced it should just be the same interface (engine provides handlerton for producing handlers to scan its metadata). although implementing this does mean finishing frm work or something to have magic discovery things... -- Stewart Smith _______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp
