Hi all!

Yes, I am still alive.  I've been doing a little bit of hacking on treecc 
lately, so thought I would check in and let you know what I am up to.

I've just checked into CVS a python binding for treecc, and I might even do a 
perl binding soon unless someone else beats me to it.

I'm doing some compiler research work at the moment on treecc.  It works very 
well at its original purpose: taking the pain out of abstract syntax tree 
creation and manipulation.  But it has always been less than helpful when it 
came to type systems and type inferencing (take a look at the type gathering 
code in cscc if you don't believe me :-) ).

I'm looking into ways that the aspect-oriented approach in treecc could be 
extended to handle type systems.  i.e. Instead of the programmer having to 
build a type system by hand, they could declare the type coercion and 
operator matching rules declaratively and let treecc figure out the algorithm 
details.

Not much to report as yet, as my experiments so far are bordering on term 
rewriting systems and/or logic programming, which I find to be a nightmare on 
larger language definitions.

I'm also researching ways that compilers can be quickly built using scripting 
languages (hence the python bindings).  While maybe not useful for a language 
the size of C#, there are some times when it would be nice to be able to 
quickly prototype a small-C compiler for PIC microcontrollers, Lego 
Mindstorms bricks, etc.  The small size of programs on such platforms means 
that compiler speed is not as important as getting a compiler going quickly.

So, hi again to everyone.  I'm really impressed at what you have all achieved 
on pnet and libjit in my absence.  Keep up the good work.

Cheers,

Rhys.
_______________________________________________
Developers mailing list
[email protected]
http://dotgnu.org/mailman/listinfo/developers

Reply via email to