This is an interesting improvement:
http://d.puremagic.com/issues/show_bug.cgi?id=9963
But I think we have also to think about priorities. I think
people interested in D performance today use GDC or LDC, and they
avoid DMD. And maybe this is not going to change.
On the other hand if D misses some of its parts then all D users
are affected. So isn't it more important for Walter to use his
time (this is just an example) to define the exact semantics of
"scope" keyword and to implement it, instead of optimizing small
operations done by the DMD compiler?
When is "scope" going to be worked on?
(And even regarding just bugs, I think Issue 3789 is one hundred
times more important and urgent than Issue 9963.)
Bye,
bearophile
