RE: Why no metamethods for logical operators?

["Firestone, Rameses" <rameses.firestone@...>]

> -----Original Message-----
> From: lua-bounces@bazar2.conectiva.com.br [mailto:lua-bounces@bazar2.conectiva.com.br]On Behalf Of > > RLake@oxfam.org.pe
> Sent: Wednesday, March 24, 2004 8:53 AM
> To: Lua list
> Subject: RE: Why no metamethods for logical operators?
> 
> 
> 
> > I believe C++ must also decide at runtime. 

> C++ can tell at compile time because the short-circuit behaviour 
> only applies to primitive types, not classes. (You cannot subclass 
> a primitive type in C++).
 
Ok yes I agree, I misunderstood the original answer.

> In my opinion, the fact that you can override short-circuiting 
> operators in C++ is just plain wierd. But the fact that you 
> can override , is pretty wierd too. C++ gives you lots of 
> latitude to write unreadable programs. 

This is true. Like many other tools it can be abused. But it is 
also powerful in it's own way. Also there is almost no other choice 
of language if you wish to write fast reasonably portable object 
oriented code.

What I like about Lua is it is mainly targeted as an extension 
language. It doesn't try to claim it's the best language for all 
applications. It seems like some people like to push Java in this  
this manner and in my experience it often falls flat on It's face. 

> In any event, the use of "+" for "or" and "*" for "and" makes 
> a lot of sense; few datatypes would require different semantics. 

My main problem is I need + for + and * for *. So this wouldn't
work for me. I can always use a function. The only problem is the 
syntax doesn't look so clear.

Ram