On Wed, Jan 12, 2011 at 12:27:40PM +0000, Frank Shearar wrote:
> On 2011/01/11 23:29, Igor Stasenko wrote:
> >>
> >> Added support for positional messages and matrix expressions, from cobalt. Used in OpenGL
> >>
> >
> > Personally i would implement these syntax extensions is separate
> > Parser/Class subclasses
> > and then put them in use locally by classes which using them, by
> > redefining  #parserClass/compilerClass.
> >
> > I don't think that everyone will be happy seeing these syntax
> > extensions enabled by default.
> > Let us keep base syntax in core classes. Compiler and Parser code
> > already clobbered enough up to the point to be completely thrown out
> > and replaced by implementation from scratch :)
> 
> I'd like to see the syntax for these expressions in the relevant places. 
> Positional messages I presume mean foo(1,2,3), but what does a matrix 
> expression look like?
> 
> I don't mind syntax extensions, as long as we don't go crazy with them. 
> Of course, like Igor points out, we can go as crazy as we like without 
> upsetting other parts of the system by using #parserClass/compilerClass 
> (like OMeta does, for instance, using a completely different language in 
> its methods).
> 
> frank
> 

this is no longer needed. Andreas removed the positional
methods from OpenGL a while ago (without me noticing), and I
moved this out of the inbox.

As for matrix expressions, I have no idea either

-- 
Matthew Fulmer (a.k.a. Tapple)