Re: [proposal] Type metatables per environment

[Daurnimator <quae@...>]

On 22 December 2011 12:29, Miles Bader <miles@gnu.org> wrote:

Patrick Rapin <toupie300@gmail.com> writes:
> My proposal is the following: provided it is possible, why not limit
> the metatable scope to the environment of the current function?
> Even better, I could imagine that the type metatables could be placed
> in the metatable of _ENV and manipulated from there.

Soooo what are the details?  How does the compiler implement this?
How would it impact performance?

Do you mean that, for instance, OP_GETTABLE would need a 4th argument
to pass in _ENV? ... and _every_ indexing operation (maybe the most
common operation in Lua bytecode) would then need to get and pass in
_ENV?

A not too odd assumption to make is that the current environment (ie, _ENV) is always available for a function/VM instruction.