On Wed, 07 Jan 2009 21:13:12 -0500, Norman Ramsey wrote: >> Is it true that the only types on C-- are the bits8, bits16, etc., up >> to > > some small limit? (although they may be given new names by a type > > definition) > > > > That it recognises nothing like C's structs and unions and so forth, > > and that that is entirely the job of the front end, which will likely > > be generating code containing a lot of numerical constants for field > > offsets and the like? > > Yes, that's exactly right. > > There is also a boolean type, but there are no variables of boolean > type, only expressions. > > In hindsight we have learned that it is convenient to write parts of the > run-time system in C--, and we have an embryonic proposal for a dialect > called "Systems C--" that will provide some rudimentary facilities, > using the same syntactic mechanisms used to lay out initialized data. > But I doubt this dialect will be implemented unless a volunteer steps > forward, or unless we get funding for new work on run-time systems.
I mentioned C-- to my son as an interesting language, emphasizing that it seemed to have the spirit of assembler, without the obnoxious syntax, and without register restrictions. He remarked that it would be interesting to see how it would fare as a target for a type-checking macro processor. Maybe such a thing would be useful? I imagine he's imagining something along the lines of Scheme's "hygienic" macros melded with the so-called "Typed Scheme". But applied to low-level stuff, because he's really interested in writing high-performance video games. -- hendrik _______________________________________________ Cminusminus mailing list [email protected] https://cminusminus.org/mailman/listinfo/cminusminus
