Red Hat's issues in considering the D language

Stefan Koch via Digitalmars-d digitalmars-d at puremagic.com
Wed Dec 21 19:31:11 PST 2016


On Thursday, 22 December 2016 at 03:18:42 UTC, Jerry wrote:
>
> Not using AliasSeq if that's what you mean. I don't know if the 
> "tupleof" for a struct would be considered the same as "T..." 
> but basically what I was doing:
>
> foreach(ref field; myLargeStruct.tupleof)
> {
> }

Yes that is a compiler-tuple as well.
which means that the foreach is not loop at all.
Rather it's body gets duplicated myLargeStruct.tupleof.length 
times.
Leading to giant numbers statements, at those conditions
  N^{2,3,N} algorithms in the optimizer do not scale gracefully.


More information about the Digitalmars-d mailing list