> is there any reason why you're using buffers in the render context rather
> than writing directly to the response output stream?

I need the Http request (which hold the user object among other things) and
the output buffer. As a result I had to create a container object
(RenderContext) to hold everything as it traversed the nodes and thus build
the resulting html.

> if you're not careful then re-using classes can cause unexpected problems
> at high concurrency.

By putting all the state into the user object, output(RenderContext )
doesn't remember from one output call to the next. If there is a potential
conflict which is rare, I just use 'synchronized'.


----- Original Message -----
From: "robert burrell donkin" <[EMAIL PROTECTED]>
To: "ECS Developers List" <[EMAIL PROTECTED]>
Sent: Saturday, November 17, 2001 9:07 AM
Subject: Re: Fw: additions to ConcreteElement


>
> On Thursday, November 15, 2001, at 09:45 PM, jnilson wrote:
>
> > "> i have concerns about performance "
>
> is there any reason why you're using buffers in the render context rather
> than writing directly to the response output stream?
>
> > This is one of the main reason I added output(RenderContext context).
The
> > rational was to "conserve" objects by allowing them to hang around in
> > memory
> > and change their look and feel to suit different needs. Without it I
> > found I
> > often had to instantiate a new class for every small difference.
>
> if you're not careful then re-using classes can cause unexpected problems
> at high concurrency.
>
> but - as i said before - this is something that i need to think about.
>
> - robert
>
>
> --
> 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]>

Reply via email to