Tim Peters

- No thought whatsoever, or timing, went into the current `BITLIM` value. I see that, on my disk, I last left it as 3 :wink:. So as to provoke the...

I'm not going to argue about this. As I said, I didn't find the code hard to understand to begin with, so this is just random thrashing to me. I...

Neil, I think the one _basic_ thing this is missing is an asymptotically faster str->int function. Conversions in both directions were the real topic of the original issue report, although...

I agree these should be as similar as possible. I don't really care which style is picked. I have no interest in being as fast as possible here. "Major improvement"...

EDIT: oops! This originally forgot to add `mem[w] = result` to cache the result. No effect on correctness, just speed. The problem with using `@functools.cache` is that then the cache...

> Is the goal here to make memory, not time, the limiting factor with hopes that there would be no need for a limit on `str(int)` conversion to protect against...

The int->str and str->int code are very similar in concept, and should be next to each other in the code (not with unrelated division between them). While the concepts are...

> Where's a good place to show it? (If I understand correctly, this issue is not the place for further alternatives.) Right, this pull request is about the specific code...

I don't intend to argue with Neil :wink:, but it's my original issue report, and I wholly intended it to be open to discuss any and all relevant algorithms. This...

Thanks! It's at least cute that a well-intentioned small boost in precision to absorb rounding errors unintentionally convinced later code to do an entirely wrong thing :wink:. Can't make time...