Re: Reducing the size of Lua.

[skaller <skaller@...>]

On Tue, 2005-02-01 at 11:16, Greg McCreath wrote:

> 
> What options do we have to reduce the size of the Lua.  We do not need
> many of the newish language features.  I notice a steady progression
> of complexity since the earlier releases (Coroutines and such).  

Yeah, this concerns me a bit too. However ..

> Can we remove them?  

I think it is possible to factor the Lua code even more.
My own approach to minimalism has been:

(a) start with something simple

(b) add new features -- even problematic ones

(c) figure out a way to extend the basic core
    to make most of the features 'pluggable'

I'm doing this in Felix, where the idea is to support
a core language plus Domain Specific Sub Languages (DSSLs)
which can be 'plugged in'. At present, they can't be plugged
in, but without some hardwired examples it is hard to
figure out 'how would I make this feature pluggable'?

Lua is now adding a package manager, which is a step.
It is not enough, but a start. Some parts of hardwired
Lua core could be factored out. For example the 
bytecode compiler is not always required, so it
should be just another loadable package.

Still I find it hard to see how you got 215K for the lib,
on my x86 Linux box, the C++ version of the interpreter:

[skaller@pelican] ~/links/flx>ls -lasp bin/flx_lua
  20 -rwxrwxr-x    1 skaller  skaller     18512 Feb  1 13:48 bin/flx_lua

seems a bit smaller. This is statically linked but doesn't include
C/C++ run times... this is 10 times smaller than what you see..??

-- 
John Skaller, mailto:skaller@users.sf.net
voice: 061-2-9660-0850, 
snail: PO BOX 401 Glebe NSW 2037 Australia
Checkout the Felix programming language http://felix.sf.net