Wishlist/Roadmap for 2.0?

[Tristam MacDonald]

Graham St Jack Wrote:

> Tristam MacDonald wrote:
> > Is anyone currently running, or has thought about running, a tentative roadmap for D 2.0? Or a polled wishlist of some sort to give the Developers string feedback on what sort of issues the general D community would like to see?
> > 
> > I don't have any complaints about the way D is developed, and Walter is doing an incredible job seeing as he is doing it alone. But I do sometimes wonder if the development priorities as they seem to be are very beneficial, for instance, I have a feeling that CTFE is less useful to the general developer than say meaningful constructors/destructors/opAssign on structs.
> > 
> > Again, I am not complaining, merely suggesting a way to get balanced feedback on what is most important to the community in general...
> 
> First I have to say that (especially with the addition of const stuff) D 
> is already fantastic, and we have to be careful not to compromise it 
> with too much feature bloat.
> 
> 
> That said, the things I would like to see added for 2.x are:
> 
> Multiple return values (tuples).
> 
> Adding pthread-like conditions into the language, using the same mutex 
> as the synchronized keyword. It needs to be done in a way that allows 
> any number of conditions per object, rather than Java's 0 or 1.
> 
> A way to keep track of what exceptions are thrown, similar to the way 
> Java does it (perhaps optional via a compiler switch?).
> 

While we are on the topic of exceptions, better debugger support would be very welcome. I know it is not directly part of the language, but just being able to get line numbers and file names out of GDB would be a huge help. The compiler is already much more helpful on errors than common C++ compilers, but runtime errors are a little hard to diagnose when you are digging through x86 assembly trying to correlate with the D source code...