Paul Eggert wrote:
> Karl Fogel wrote:
>> My proposal would be to at least document it for `nthcdr'
> 
> It would be better to fix nthcdr so that it works on bignums

... and I did that by installing the attached patch.

Thanks for reporting the problem. It's helpful to see which functions should 
have higher priority for fixing these sorts of glitches. If you're bitten by 
similar glitches in other functions please let us know.