[BusyBox] Re: which shell?

[Larry Doolittle]

> > That's both true and false.  There are indeed dramatic differences
> > between them, but some of that is fluff.  I think we (the maintainers)
> > would benefit if some or all of the shells shared a common library of
> > variable handling and tty handling.  Just like they share Vladimir's
> > command line editor.  Also, one of my to-do items for hush is to drop
> > Aaron's math handling into hush.  The more of this kind of work that
> > gets done, the more similar the shells will become.
> > 
> > You never know where this giant Ouija board will take us.  I think
> > it's too early to tell if some or all of these shells might end up
> > close enough to have their source code merged with a compile-time
> > selectable feature/size tradeoff.
> 
> As I hope busybox reflects for use in the builtin systems it would be strange, 
> that some environments would enter into it (already in a binary kind). 

I don't understand what you are trying to say here.

> And as external functions increase size such approach very much is not
> pleasant to me.  I would prefer, that mathematical processing, for
> example, was connected as:
> #include "aritch.c" 
> in each shell.

Hmmm.  This is certainly one way for code to be shared between "callers".
It's kind of out of fashion right now.  Before I support it, I would have
to run some tests to show that it really did reduce the size of the final
executable compared to a normal, modern, library of subroutines approach.

The mechanism of sharing code between the shells is less important to me
than the sharing itself.  The former is an implementation detail.  The
latter is the means to focusing limited attention on fewer lines of code,
which should yield better debugged code.

      - Larry