Pure higher order functions

[Jonathan M Davis]

On 2011-07-06 16:19, bearophile wrote:
> Jonathan M Davis:
> > I believe that the problem here is essentially what's being argued over
> > in the dmd-beta list right now.
> 
> In D a HOF is implemented with a struct or a class. So, what's a pure
> struct/class, generally?

I don't think that it makes any sense to talk about a pure struct or class. 
Functions are pure, not types. However, given that applying a modifier to a 
type seems to generally end up applying it to all of the type's functions, 
then marking a type as pure should mark all of its functions as pure. That's 
not what I was talking about though. I was talking about how weakly pure was 
essentially stripped out of the language with recent compiler changes. Whether 
marking a type as pure does anything is another issue entirely.

- Jonathan M Davis