The Expressiveness of D

[%u]

Lutger Wrote:

> %u wrote:
> ...
> > 
> > template T(A...) { alias A T; }
> > 
> > void main() {
> >   auto a = true ? new B : new C;
> > // these don't work - why?
> > //  auto b = [new B, new C];
> > //  auto c = { return [1: new B,2: new C]; };
> 
> I think it should, it works for arithmetic types. See this bug report:
> 
> http://d.puremagic.com/issues/show_bug.cgi?id=3543

Thanks. I understand it's a known bug now. The explicit casts shown in this thread are what annoy me. In code review you have to add a comment "// totally unnecessary cast, but see dmd bug #3543". Confuses new users. Not anything serious and there's a workaround. Just annoys. Especially if code guide tells to avoid casts. Especially now that I found it by accident and always remember it.

>  
> >   T!(int,int) e = (1,2);
> 
> I am surprised this works since D doesn't have a tuple literal syntax.

I actually found these examples when doing a google search. Wanted to see how tuples in D work. I didn't realize there's a library solution. We should probably remove the old newsgroup posts of this errorenous "tuple". If we can hide it well enough, new users won't even notice DMD accepts this confusing code.

> 
> >   e = T!(3,4);
> > 
> > // ah - so (1,2) syntax on initialization, T!(1,2) when assigning!
> >   T!(int,int) d = T!(1,2);
> > 
> >   e = d;
> > 
> > // tuples aren't first class, why?
> > //  auto f = { return e; };
> > }
> 
> There has been discussion about this, as mentioned by bearophile. 
> 
> In summary, the reason for not having first class tuples is that current 
> language constructs provide sufficient means to express tuples as a library 
> type, see http://www.digitalmars.com/d/2.0/phobos/std_typecons.html#Tuple.
> 
> You can do most - but not all - of what builtin tuples could do with that 
> template.

Crazy.. yes, the library solution is enough to make the language orthogonal in this way. It just confuses me why dmd almost accepts what I did.