corruption. With CFMX you only have to lock for avoidance of race
conditions, because the underlying java runtime deals the memory
issues.
Clustered environment doesn't matter, because CFLOCK is local to the
machine in question, not the cluster.
If application variables (including instance variables of any
application-scope CFCs) are set to a value and never change
(exceedingly rare), then yes, you can skip locking. However, you may
as well just set them into the request scope, since they're going to
be set every request. You can use CFLOCK in this situation to only
load the variables once and improve performace. If you're
instantiating large CFCs, the performance increase can be huge.
cheers,
barneyb
On Tue, 14 Sep 2004 14:40:25 -0400, Don Chunshen Li
<[EMAIL PROTECTED]> wrote:
> This seems to be a timeless question of possible confusion. This is my understanding of cf5.x and cf6.x, for a single cf server (not clustered) environment, application and server variables do not need to be locked when none of them needs to be modified in the course of applications execution, why? they are always the same for all users, so why lock them?
> while session variables is totally different. Any one disagree?
>
>
--
Barney Boisvert
[EMAIL PROTECTED]
360.319.6145
http://www.barneyb.com
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings] [Donations and Support]
