Re: ... efficiency (was Re: class implementation)

[David Jones <drj@...>]

On Apr 07, 2006, at 00:27, Mark Hamburg wrote:

> on 4/6/06 8:17 AM, David Jones at drj@pobox.com wrote:
>
> > "for k,v in pairs{...}" is not quadratic, but it isn't fast.
>
> It's also incorrect if you care about nils or order. Ipairs is 
> incorrect if
> the arguments can contain nils.

Of course you're right.  Those damn nils again.
> The set of cases where nils matter when iterating over varargs is 
> probably
> small, but it is potentially an issue.

Agreed.  That makes me want apairs in the standard base library.

drj