Re: [RFC] Default constructors

Hi! > Well, as others have pointed out, HHVM behaves differently in this case, And? > so I guess what I'm getting at is: is this considered a feature *of the > language*, or just a known oddity of the *implementation*? If it's part I do not think the standard should require it, as whoever using this for anything, one way or another, is doing it very, very wrong, but that's how it works now. > other implementations can emulate it (or decide to deviate from it); if > not, then it seems appropriate to talk of "fixing" rather than > "changing" it. The choice of emotionally charged term is up to whoever uses it, my point was just that this was a deliberate intentional decision, not an unintentional mistake, and any change there has BC consequences. I think this point is well made by now, so unless there is something new to say, I'll refrain from discussing it further. -- Stas Malyshev smalyshev@gmail.com