Re: [RFD] Unified Methods

[Edgar Toernig <froese@...>]

Roberto Ierusalimschy wrote:
> 
> Do you have any estimates about the performance of this
> new implementation?

The user methods would have exactly the same performance as
they have now.

With the tag methods it's a little bit more complicated.  The
simplest (a "proof of concept") implementation would require
one hash table lookup every time a tag method may be needed.

But a simple implementation trick would give the same perform-
ance when no tag method is set and exactly one hash table
lookup when the tag method really needs to be called.

So, at worst it's the same categorie as the changes made with
the globals.

Ciao, ET.