On Wed, Jan 9, 2013 at 12:05 PM, Timon Gehr <[email protected]> wrote:
> Yes, it is in D. Nothing is released yet. It needs to be polished a little
> so that there are no known embarrassing shortcomings anymore.
What? That's FOSS, release early, release often!
> (DMD chokes on both.)
>
> Furthermore, I need to implement exceptions, modules, and some parts of
> compile time reflection + tons of really small features. (Where all
> ambiguities and contradictions are detected according to well-defined rules
> instead of resolved or choked on randomly as DMD likes to do.)
Makes me salivate.
> Also, it obviously needs a repl. :-)
Obviously.
And direct programmatic access to the lexer and the parser. I'm coding
a macro system for D right now, as an over-layer above DMD (like rdmd)
and having to create the AST by myself to transform them according to
the user-defined macros is a pain.
> CTFE is basically done (as a portable byte code interpreter, but other
> strategies, such as JIT, could be easily plugged).
Great!
> This is a snippet of my regression test suite:
>
> auto dynRangePrimes(){
> DynRange!int impl(int start)=>
>
> dynRange(cons(start,delay(()=>filter!(a=>a%start)(impl(start+1)))));
> return impl(2);
> }
>
> static assert(array(take(dynRangePrimes(), 20)) ==
> [2,3,5,7,11,13,17,19,23,29,31,37,41,43,47,53,59,61,67,71]);
Lazy recursive range, right? As the usual Haskell definition for a
prime generator. Nice!
> I also need to decide on a licence.
Unless you've reason not to, I'd advise using the same as Phobos: Boost.
> I assume that the alpha will be out in late spring. (I am busy until early
> spring.)
You can count me in to test it (I gather to prefer to code this as you
see fit, right now). You could also present it at the D conf.
In any cases, congratulations for what seems to be a elegantly done D
implementation.
