[Python-ideas] Ideas towards GIL removal

[Greg Ewing]

Jim Jewett wrote:

> I have some vague memory (but couldn't find the references) that
> someone tried and it was too expensive.

Too expensive compared to what? The question isn't
whether it's more expensive than the current scheme,
but whether it helps when there's no GIL and you
have to lock the object to update the refcount.

--
Greg