Re: Proposal for table length operator

[Enrico Colombini <erix@...>]

On 14/12/2010 9.34, Axel Kittenberger wrote:
> As far I understood this proposal table.remove(t) has O(n) costs.
> Since
>
> t[500] = 1
> t[750] = 1
> t[1000] = 1
> table.remove(t)
> now it has to scan all values to find the largest, 750.

If I understood Keith's proposal correctly, #t would just be 999; there 
would be no guarantee of it pointing to a non-nil value (in fact, that's 
done exactly to allow nil values to be counted as valid array entries).

--
  Enrico