Re: cvs.transvirtual.com unreachable and other news

Date view Thread view Subject view Author view

From: Tim Wilkinson (tim@transvirtual.com)
Date: Mon Oct 19 1998 - 14:13:18 EDT


Yes, all uses of kaffe.util.Ptr must have a finalizer to handle what it points to.

Tim

Archie Cobbs wrote:

> Tim Wilkinson writes:
> > kaffe.util.Ptr is used to reference bits of memory allocated by native
> > libraries or the underlying system which don't conform to the standard Object
> > model and must be explicitly managed - there are plenty of examples with the X
> > lib, the zlib and others.
> >
> > A 'safe' way to do this might be to encapulate these pointers in another final
> > class with a finalzier which will always be called then the pointer should be
> > freed - but this extra level of indirection is a bit of a hassel.
>
> I guess it depends on what the use of the memory is.
>
> If an instance of kaffe.util.Ptr is for a "special" chunk of memory
> (where "special" means not only does the memory pointed to need
> to be free'd, but some other thing (like the closing of a file)
> needs to happen as well), then it makes sense to use kaffe.util.Ptr
> and just make sure that the object containing the reference has
> a finalizer that does the right thing.
>
> On the other hand, if the only reason kaffe.util.Ptr is being used
> is because some native code allocated a buffer, then byte[] would
> seem more appropriate and less memory-leak prone... but I guess
> it's up to the developer to make that decision about the extra
> overhead incurred.
>
> In any case, can we safely state that all objects that use
> kaffe.util.Ptr must have an explicit finalizer?
>
> -Archie
>
> ___________________________________________________________________________
> Archie Cobbs * Whistle Communications, Inc. * http://www.whistle.com

--
  Tim Wilkinson                         Tel:     +1 510 704 1660
  Transvirtual Technologies, Inc.,      Fax:     +1 510 704 1893
  Berkeley, CA, USA.                    Email:   tim@transvirtual.com


Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Sat Sep 23 2000 - 19:57:00 EDT