[Python-Dev] GC Proposal

["Martin v. Löwis"]

> The effect is similar for the "batch allocation" case, but opposite
> for the "long-running program" case.

I don't understand. Where is the difference?

> My proposal can be made equivalent to Martin's proposal by removing
> all of its pending traces when an untraced object is deleted.  We
> could even change this at runtime, by adding a counter for pending
> objects.

What is a "pending trace"?

> Come to think of it, I think Martin's proposal needs to be implemented
> as mine.  He wants the middle generation to be 10% larger than the
> oldest generation

Not exactly: 10% of the oldest generation, not 10% larger. So if the
oldest generation has 1,000,000 objects, I want to collect when the
survivors from the middle generation reach 100,000, not when they reach
1,100,000.

> but to find out the size you need to either iterate
> it (reintroducing the original problem), or keep some counters.  With
> counters, his middle generation size is my "pending traces".

Yes, I have two counters: one for the number of objects in the oldest
generation (established at the last collection, and unchanged
afterwards), and one for the number of survivors from the middle
collection (increased every time objects move to the oldest
generation).

So it seems there are minor difference (such as whether a counter
for the total number of traceable objects is maintained, which you
seem to be suggesting), but otherwise, I still think the proposals
are essentially the same.

Regards,
Martin