Hacker News new | ask | show | jobs
by pjmlp 387 days ago
I think the biggest issue has been trying to always chase the next big thing that eventually could bring mindshare to D, while not finishing the previous attempts, so there are quite a few half baked features by now.

Even Andrei Alexandrescu eventually refocused on C++, and is contributing to some of the C++26 reflection papers.
1 comments
fuzztester 387 days ago
>while not finishing the previous attempts

I agree, and that applies to many software projects, and not just programming languages only.

>so there are quite a few half baked features by now

what are some of those half baked features?

link
pjmlp 387 days ago
The new allocators, some corner cases of the destroy and destructors, not everything on Phobos is @nogc friendly, BetterC still chockes on many common C extensions, DIP 1000, the whole set of @live semantics.

Now there is a new GC being redesigned, and there are discussions about a possible Phobos V3.

link
fuhsnn 386 days ago
> still chockes on many common C extensions

I play with ImportC occasionally, a lot of those can actually be opt out by undef'ing __GNUC__ on the preprocessor invocation, idk why they don't do that. Oh, now it chokes on C23 features as well because system cpp defines __STDC_VERSION__=202311L now. Edit: that was solved: dlang/dmd/pull/21372

link
tored 386 days ago
You mean ImportC and not BetterC, right?
link
pjmlp 386 days ago
Right, got that one wrong.

Although, on the context of BetterC, there is the debate about having more regular features available in that mode as well.

link
fuzztester 382 days ago
thanks.

interesting, didn't know about the new GC, or possible Phobos V3

link