On 07/08/2015 11:32 PM, Martin Nowak wrote:
On 07/08/2015 07:33 PM, Timon Gehr wrote:
The compiler complexity introduced by this change is trivial.
Except trivial doesn't mean it's cheap. Would probably take 1-2 days to
fully spec/implement/optimize/test/document it.
In that time I could finish a thread cache for the GC, or GC-less smart
pointers, or finish dub watch, or reduce binary sizes, or finish OSX
shared libraries, or join 10 more forum discussions.
This is completely self-contained.
But is it useful enough to spend the effort and extend the language?
I just noted that increased compiler complexity is absolutely not a
reason not to implement this.
