Re: os.setenv()???

[Rici Lake <lua@...>]

On 8-Jan-06, at 5:04 PM, Ryanne Thomas Dolan wrote:

>  Is it easier to check that a given function is implemented in the 
> current environment when the function is a high-level abstraction?  Is 
> checking for std.putenv (a thin interface) harder than checking for 
> os.setenv (your abstraction)?

With my abstraction, the Lua program only has to check for the 
existence of a single interface. Without the abstraction, it has to 
check for N possibilities, where N is not bounded. So it seems simpler 
to me, yes.

>   I don't see that it is, but maybe a thin interface would introduce 
> more checks than necessary, defeating the performance boost gained 
> from a simpler translation.  I dunno.

Where is the performance boost? Changing a name between the Lua and the 
C interface does not cost anything.

You might be thinking that my proposed __index metamethod on a global 
env table is somehow slower than an explicit call. I think you'll find 
that it is actually faster, although the difference is not particularly 
significant.