Hi!
On Jul 17, 2008, at 8:46 AM, Paul McCullagh wrote:
Pool of threads (done).
Change main select to be libevent (done in different tree).
Executue on read events. If we ever get a zero read on a non-block
call (everything is non-block) we yield after N events and wait
till we get sliced back in. This would be done on read/writes
(though just read first).
Thoughts?
OK, I see. Because libevent has already detected that data is
available, the thread does a "busy" read.
Is this faster than doing a blocking read?
It gives us an easy yield point. The idea is to always push data, etc.
We could long term look at solving the slow client/linger problem, but
that is not really in the current scope (also, I am not sure this
makes sense for a database).
Cheers,
-Brian
--
_______________________________________________________
Brian "Krow" Aker, brian at tangent.org
Seattle, Washington
http://krow.net/ <-- Me
http://tangent.org/ <-- Software
_______________________________________________________
You can't grep a dead tree.
_______________________________________________
Mailing list: https://launchpad.net/~drizzle-discuss
Post to : [email protected]
Unsubscribe : https://launchpad.net/~drizzle-discuss
More help : https://help.launchpad.net/ListHelp
