Re: Connection pool mgmt

[Stuart Peacock]

Thanks for your kind reply.

I completely understand maintaining similar semantics to other implementations.

What seems to be missing(and perhaps it's just me) is an alternative method of
pool management. To me, it matters little how long a connection "has" been used.
What matters to my pooling strategy is how long it "hasn't" been used. Perhaps
I'm just not up to speed on the overhead involved, but if my pool has grown to
10 connections, but each one is still getting used frequently then my pool is the right
size. If I have 10 and 6 of them are basicly sitting idle, them I'd prefer to drop them.

Is is desirable to have additional functionality in the mySql pooling which isn't present
in the others? I can live with it if it seems more reasonable to go the the lowest common denominator. But it will just mean I'll have to implement my pooling strategy outside of
the current driver.

Thanks for all your work!!!!!


sp