Use try_for_each in into_py_dict

[bschoenmaeckers]

This applies the suggestion in #4493 (comment) to see if this will lead to any improved performance (or not 🤷‍♂️).

[davidhewitt]

Thanks for following up with this! I think we actually lack benchmarks for this case (or at least I didn't spot any), so might need to add some? 🤔

[bschoenmaeckers]

The dict_new benchmark uses into_py_dict but does not show any difference.

[bschoenmaeckers]

I suspect it only has an advantage on the free-threaded build because of the critical-section wrapping the whole loop.

[davidhewitt]

Ah, yes.

I think it makes sense to proceed to merge this; as I understand it, for complex iterators this is better for the optimizer and it's definitely better for our own free-threaded iterators!