State of Play

[Walter Bright]

Tomas Lindquist Olsen wrote:
> D1 does have some missing features that are in D2, and could be
> backported to D1 without breaking any code.
> This isn't going to happen for the sake of stability. But if I want to
> use some of the new features, I have to get all the cruft that made me
> look into D in the first place as well. A major reason I started with
> D was because of simple syntax, GC and lack of the const hell.
> 
> D2 is no longer a simple language, you have to know all kinds of shit
> to be able to use it correctly.
> 
> All my projects at the moment are in C++. And TBH I don't see that
> changing any time soon. The stuff I did in D no longer works, and I
> don't have time to debug the entire thing to figure out how/where the
> compiler changed.

I've worked with C/C++ for decades, and it's a miracle if code that 
compiled 10 years ago compiles today without changes. Code never works 
without changes when porting to a different compiler. Code targeted at 
VC tends to target specific versions of VC. "Portable" libraries like 
Boost, STL, and Hans Boehm GC are full of #ifdef's.

The only real difference from D is that:
1. The evolution of C++ is about 10 times slower than D.
2. We're all so used to working around these problems in C++, we don't 
notice it.

I understand your frustrations with the changes, just not why that means 
using C++ <g>.

> And yes, the Phobos vs. Tango (which in turn keeps breaking) situation
> of course isn't making things better.