Re: [ANN] Lua 5.3.0 (alpha) now available

[Sean Conner <sean@...>]

It was thus said that the Great Coda Highland once stated:
> On Fri, Aug 1, 2014 at 11:04 AM, Coroutines <coroutines@gmail.com> wrote:
> > On Fri, Aug 1, 2014 at 11:00 AM, Roberto Ierusalimschy
> > <roberto@inf.puc-rio.br> wrote:
> >>> It was mentioned that "table.clear" (or similar) was likely. Was it decided
> >>> that it would not be useful?
> >>
> >> I don't think we even agreed on what it would do.
> >
> > As far as I see it, tables and closures are the only real mutable base
> > types in Lua -- we generally don't need to nil'ify all the upvalues of
> > a closure/function, but it would be nice to quickly dispense the array
> > and hash portion of a table so the same object can be returned, and
> > filled thereafter -- with the same associated metatable.  Just
> > something faster than looping through the table in Lua to unreference
> > everything?  I'm not sure what extra arguments it might take to
> > selectively clear portions of the table.  Maybe the array or hash
> > portion, or a starting and ending numeric index?  I might be very
> > crazy... :p  Stop me if I have weird ideas.
> >
> 
> Oh STAHP! ;)
> 
> Seriously, though, I can't see any real uses for it beyond just a
> complete and total table wipe. Anything else is either adequately
> served by nil assignment, or is an implementation detail that isn't
> exposed anyway.

	table.clear(x)

vs

	do
	  local y = getmetatable(x)
	  x = {}
	  setmetatable(x,y)
	end

  -spc (Hmmm ... now that I think about it, slap that code into a function
	and there you go ... )