Re: __index returns truncated to one, why?

["Thiago L." <fakedme@...>]

On 10/04/2014 13:14, Sean Conner wrote:
> It was thus said that the Great Duncan Cross once stated:
> > On Thu, Apr 10, 2014 at 12:33 PM, steve donovan
> > <steve.j.donovan@gmail.com> wrote:
> > > On Thu, Apr 10, 2014 at 1:03 PM, Dirk Zoller <duz@sol-3.de> wrote:
> > > > Is there is a good technical reason for this truncation?
> > > It is in a very important part of the VM, and any extra work at this
> > > point is going to slow the usual case down. (By how much, I don't
> > > know.)
> > I just realised this could be another case for my old suggestion [1]
> > about '...' as a suffix operator which means, explicitly, "do not
> > truncate this list of values". My original suggestion was for function
> > call arguments, so:
> >
> >    func(a()..., b, c)
> >
> > ...would not truncate the return values of a(), as it normally would.
> > But another case could be for this case, so:
> >
> >   local a, b, c = t[1]...
> 	t = { 'one' , 'two' , 'three' }
> 	local a, b, c = t[1]...
>
>    a == 'one', that's fine.
>
>    What does b and c equal?
>
>    -spc
Whatever ({mt.__index(1)})[2] and ({mt.__index(1)})[3] would be...