Hi!
As a few of you may have noticed I've pushed up the rewrite for the SE
layer to now on Schemas. This doesn't do anything at all for the
average storage engine but if you are a network engine you can now
push schemas around like you do tables. All of the schema related code
is now owned by the SE layer (and there is a default plugin to handle
db.opt/etc).
I will be extending this a bit more, but from the look and feel of it
you should be able to understand how it now works.
One thing that I would like to do with the SE layer going forward is
make sure that operations that are IO based really hide behind the
layer.
Right now we have errno exposed for instance. That is great if you are
MyISAM, but for the other engines? A simple "yes" or "no" is all that
we need on the upper end. You can supply back messages to the user,
but you should keep the specifics to your engine, in the engine. A
better error logging system is needed but I don't see that coming in
the near future.
My big wish is to keep the operational details hidden to those who
don't need to know about them (and to log information that would be
useful for sysadmins).
Over time I am going to be modifying the system API to be a little
more straight forward as far as use is concerned. I'd also at some
point like to come up with an error system that will be compatible
with exceptions (something we don't have at the moment).
Cheers,
-Brian
_______________________________________________
Mailing list: https://launchpad.net/~drizzle-discuss
Post to : [email protected]
Unsubscribe : https://launchpad.net/~drizzle-discuss
More help : https://help.launchpad.net/ListHelp
