[Python-3000] Removing __del__

[Carl Friedrich Bolz]

Chermside, Michael wrote:
[snip]
> The other problem I discussed is illustrated by the following
> malicious code:
>
> evil_list = []
>
> class MyEvilClass(object):
>     def __close__(self):
>         evil_list.append(self)
>
>
>
> Do the proponents of __close__ propose a way of prohibiting
> this behavior? Or do we continue to include complicated
> logic the GC module to support it? I don't think anyone
> cares how this code behaves so long as it doesn't segfault.

I still think a rather nice solution would be to guarantee to call
__del__ (or __close__ or whatever) only once, as was discussed earlier:

http://mail.python.org/pipermail/python-dev/2005-August/055251.html

It solves all sorts of nasty problems with resurrection and cyclic GC
and it is the semantics you already get when using Jython and PyPy
(maybe IronPython too, I don't know how GC is handled in the CLR).

Now the implementation side of this is more messy, especially with
refcounting. You would need a way to store whether the object was
already finalized. I think you could steal one bit of the refcounting
field to store this information (and still have a very fast check for
whether the rest of the refcounting field is really zero, if the correct
bit is chosen).

Cheers,

Carl Fridrich Bolz