Defining VCL-like framework for Python

Alexander Staubo nospam-alex at mop.no
Sat May 22 09:51:59 EDT 1999


In article <Pine.WNT.4.04.9905211029100.289-100000 at rigoletto.ski.org>, 
da at ski.org says...
> On Fri, 21 May 1999, Alexander Staubo wrote:
> 
> > Unfortunately, as I pointed out earlier, cyclic references make 
> > parent/child relationship hard to implement cleanly. The best workaround 
> > I've found is to have classes implement a common "pre-destructor" that 
> > must clean up any external references to prepare it for destruction.
> 
> FYI: using environmental acquisition seems a natural for a parent-child
> relationship, and solves many cyclic reference problems.  I'd point you to
> a URL of Jim Fulton's code for acquisition, but Digital Creations just
> redid their website and I can't find it anymore.  I'm cc'ing
> webmaster at digicool, hoping that they'll let you know where those documents
> have gone.

I agree. I remember the DC document you mention. Personally I'm so 
steeped in the traditional containment systems my mind is still 
struggling to "accept" using acquisition in this context, but I see that 
it's vastly superior, at least for this project. I'm definitely going to 
have a stab at it -- at least it will cure my circular-ref ills.

-- 
Alexander Staubo             http://www.mop.no/~alex/
"Give me an underground laboratory, half a dozen atom smashers and a beautiful 
girl in a diaphanous veil waiting to be turned into a chimpanzee, and I care 
not who writes the nation's laws." --S. J. Perelman




More information about the Python-list mailing list