On Fri, Jul 29, 2011 at 4:55 PM, Maurizio Giordano <[email protected]> wrote: > In my new language implementation the mset > is a fundamental data structure, I cannot rely only on standard > syntax objects like lists, strings, etc.
If by that, you mean you want msets to have a printed representation that can used in code (say, for quoted data) and also used with Racket's read and write functions, then I see your problem -- we don't really have good facilities for introducing new, phase-crossing datatypes. Racket relies on having a few core datatypes that all phases share (e.g. lists, strings, symbols, etc.), but anything implemented in Racket itself lives in only one phase. If, on the other hand, you are not relying directly on Racket's bindings for quote, quasiquote, read, or write, you may be able to work around this. For instance, you could use a read-syntax that turns mset literals into a prefab struct that your quote form interprets. You could also have separate read and write functions that interpret mset literals directly. Basically, the less Racket-like your language is, the more you have flexibility to work around our read/write behavior across phases. --Carl _________________________________________________ For list-related administrative tasks: http://lists.racket-lang.org/listinfo/users
