Policy mix-ins sounds like a nice approach. They have the benefit of
compile-time rather than run-time overhead, and easily retain backward
compatibility.

-- ES

On 6/27/06, Marten Svanfeldt <[EMAIL PROTECTED]> wrote:
> There shouldn't be any overhead (except a small compiletime dito) of
> doing a change such as the one res suggested, if it is done (for
> example) by using policy mix-ins for the different behaviours.
>
> -Marten
>
> Quoting Jorrit Tyberghein <[EMAIL PROTECTED]>:
> > Be aware that csBlockAllocator should be very efficient. If you try to
> > make it too general you risk adding overhead (for example virtual
> > function call overhead).
> >
> > On 6/27/06, res <[EMAIL PROTECTED]> wrote:
> >> On 27.06.2006 07:29, Eric Sunshine wrote:
> >> > Rather than changing the policy of an allocator (csBlockAllocator)
> >> > which was designed specifically for automated construction/
> >> > destruction of vended objects, perhaps it would make more sense
> >> > simply to use a pure heap allocator/manager, such as csMemoryPool or
> >> > CS::Memory::Heap?
> >>
> >> Perhaps csBlockAllocator<> does too much and should be split up to
> >> accomodate the different use cases: at the base, there is an allocator
> >> for fixed-size bits of data (but no object construction/destruction),
> >> and layered on top of that different classes: e.g. one that does return
> >> a constructed object and reports undestructed objects (for fast
> >> allocation that ought to be cleaned up), one that just auto-destroys all
> >> objects (for use as a simple garbage collector).

Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Crystal-main mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/crystal-main
Unsubscribe: mailto:[EMAIL PROTECTED]

Reply via email to