> I added some comments in the PR regarding the possibility of forcing the tracking on the visiting function to redirect the cost to gc passes instead of when calling next()...

Yep, I agree that this approach is better.

> ...but there is another option: not untracking tuples that have refcount == 1.

I do like the idea of not having to fix this in ~10 different places, but I'm not sure this is enough. The GC visit may happen when the refcount of the result tuple is > 1, but it may drop back to 1 before __next__ called again

z = zip([42, []])
item = next(zip)
gc.collect()  # Refcount of item == 2, untracked by GC.
del item  # Refcount of item == 1, ready to be recycled.
next(zip)  # This should be tracked, but won't be.