On 07/26/2006 03:56 PM, root wrote:
Ralf, Gaby,
If you generate lisp from Aldor and parse the result you will have complete reflection information.
Well, you are always assuming that the target language is LISP. The aldor compiler is able to target other languages, too. So I would prefer a solution that is independent of the target system. A runtime library would be perfectly OK. The Trace package in libaldor provides a very limited reflection http://lists.nongnu.org/archive/html/axiom-developer/2006-07/msg00106.html If an extension of this package could be provided with the aldor compiler as a runtime support, that would be better than going the lisp way.
This can be done dynamically by an aldor domain since you can invoke the compiler and get the result. A 'refection' domain could have functions to access all of the information available.
Assume you distribute a standalone executable. That will normally not have the compiler living inside it. So your solution does not work in that case.
I believe reflections are not too complicated once we know how the internal datastructures for domains are. (Yes, yes, we must look inside the compiler sources, or does anybody already have a description of how domains are represented in memory?)
It would be a bit more complicated to figure out how functions could be called with the appropriate type. The latter is inherently type-unsafe and I am certainly not proposing it, but there are cases (like AldorUnit) where reflection and calling () -> () functions makes sense.
To make it clear, I am only in favour of reflections in order to be able to write nice programming support tools.
For ordinary users I cannot see an advantage of reflections. Does anybody else have reasonable use-cases? Ralf _______________________________________________ Axiom-developer mailing list Axiom-developer@nongnu.org http://lists.nongnu.org/mailman/listinfo/axiom-developer
