[Asterisk-Dev] Dialplan syntax changes.. Option: work on a radically different design..

[alex at pilosoft.com]

On Wed, 18 May 2005, Brian West wrote:

> > What is the dialplan?  It's really a procedural language, it seems.
> >
> > Do we already have well defined syntax(es) that would fit?
> >
> > Sure we do -- and most of them stem from the syntax of C in one way  
> > or another.
> >
> > So, can it make sense to make the Dialplan look more like C?
Finally someone who has enough sanity to recognize that line-based 
programming languages died away with ol' GWBASIC and FORTRAN-1945

I'm for it with both hands.

Now, let's discuss the BNF of the new language. Also, discuss whether we
need gotos, etc, etc - and whether possibly adoption of another existing
"domain specific language" would be preferable to rolling our own, writing
parsers and syntax/semantic analyzer, variable/array/etc support code 
- which needs to be debugged etc. Very boring stuff.

What should be language like? C? Perl? I venture to say, it should be 
closer to perl, as far as dynamic variables, loose typing and array/hash 
support.

Also, question this: If we are going to move toward perl-ish language for
dialpan, do we need really dialplan? Would integrating, say, perlagi with
asterisk in a way mod_perl (as in, no overhead) is integrated be able to
replace dialplan altogether?