John McNally <[EMAIL PROTECTED]> writes:
> I think there is some confusion on refresh() at least there is for me.
> The method that was called when an object is returned to the pool is
> dispose(). I would generally write
>
> refresh()
> {
> dispose();
> }
>
> or vice versa. and put all the logic in one method.
This is rather confusing, as dispose() is not part of the
ApplicationTool interface -- it's something from PoolService, I
believe. If I'm using PullService, I shouldn't have to care about the
details of the underlying object pooling implementation. Everything
that I need to care about should be part of the ApplicationTool
interface. Perhaps we just need more docs? Still, it is non-obvious.
> On Thu, 2002-06-06 at 10:32, Daniel Rall wrote:
> > "Peter Lynch" <[EMAIL PROTECTED]> writes:
>>
>> > ----- Original Message -----
>> > From: "James Taylor" <[EMAIL PROTECTED]>
>> > To: <[EMAIL PROTECTED]>
>> > Sent: Tuesday, June 04, 2002 5:02 AM
>> > Subject: Re: Turbine3 - RequestTools refreshing
>> >
>> >
>> >> On Mon, 2002-06-03 at 23:41, Evan Koffler wrote:
>> >> > I love the new pipeline. It's good to see it coming along.
>> >>
>> >> Yay! I love the pipeline too, and am glad that it works for you.
>> >>
>> >> > I have a web flow that uses a request tool to capture inform/error
>> >> > messages. In testing, it continued to grow. Basically, because the
>> >> > request tools weren't being refreshed. I took a stab at implementing
>> >> > the correction. All 10 lines of code. Below is the patch to existing
>> >> > code and attached is a new java source file. I tried my best to
>> >> > comply.
>> >>
>> >> As far as releasing tools, that is already done by DefaultTargetValve,
>> >> which on line 199 calls Module::requestFinished, which in turn calls
>> >> TurbinePull::releaseTools.
>> >>
>> >> As for refreshing, should this happen when the tools are released or
>> >> when they are requested from the pool. Right now request tools always
>> >> get initted when they are pulled from the pool, perhaps with the
>> >> assumption that init makes refresh unnecessary.
>> >>
>> >
>> > Never could understand why a pull tool would be returned to the pool, without
>> > having it's internal object references cleaned up. Otherwise you get a bunch of
>> > objects in the pool with stale references to other objects, but since it is in
>> > the pool, what is the point.
>> >
>> > Seems a tool should be 'cleaned' or refreshed before going into the pool. Then
>> > on the way out initing still makes sense by itself.
>>
>> So refresh()'ing on the way back into the pool would be the more
>> correct behavior. Peter, if Fulcrum is not already doing this, would
>> you be interested in submitting a patch
>> <http://jakarta.apache.org/site/source.html> to make it so?
>>
>> - Dan
>>
>> --
>> To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
>> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
>>
>
>
>
> --
> To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
>
>
> SPAM: ---- Start SpamAssassin results
> SPAM: -4.4 hits, 5 required;
> SPAM: * -4.4 -- 'In-Reply-To' line found
> SPAM: * -0.0 -- AWL: Auto-whitelist adjustment
> SPAM:
> SPAM: ---- End of SpamAssassin results
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
