Re: Better custom metamethod/metafield support

[Philipp Janda <siffiejoe@...>]

Am 07.04.2014 07:51 schröbte steve donovan:
> On Mon, Apr 7, 2014 at 3:35 AM, Thiago L. <fakedme@gmail.com> wrote:
> > I keep arguing with sandbox writers that debug.getmetatable should be
> > allowed in sandboxes, even if limited to be unable to get the metatable of
> > specific objects (from a blacklist table or something probably).
>
> They would probably argue that if that functionality was needed, then
> you could provide a function to do it within the sandbox.  Since we
> don't use package.loadall, seems safe and convenient enough.

Since many objects share a common metatable, a per-object blacklist 
doesn't make sense. A per-metatable blacklist would be better. I'd 
probably implement that using a designated field in the metatables ... 
wait ... never mind ...

Philipp