Andrei Alexandrescu

I don't deny it's nicer. But there's work in reviewing, work in fixing known breakages of existing code, and more future work on fixing other breakages. It doesn't seem on...

Gave this a fresh look, I can't shed the feeling "why are we bothering with low impact matter when there's so many big rocks to move". Anyhow, what are the...

@MetaLang https://github.com/dlang/phobos/blob/master/std/typecons.d#L651 has a lot to do with this. Named tuples are subtypes of their unnamed counterparts. With this realization I think this PR is a go. Any breakage anyone...

@jmdavis can you please explain why `t[0]` is not equivalent to `t.name`? That would indeed be a problem. The visible comments in https://github.com/dlang/phobos/pull/6078 discuss the aftermath of the matter, I...

Got it, thanks. That's actually an easy fix - we need to use a non-names `Tuple` as backend for the respective named `Tuple`. Then no need to cast in CTFE....

>So @property ref inout(Tuple!Types) _Tuple_super() inout @trusted is the magic enemy he It's not the signature, it's the implementation. All we need is to get rid of that cast, and...

@wilzbach link doesn't work but at any rate we can make it work.

Eh, we should relax the style checker a bit: ``` std/traits.d(2803:19)[error]: Expected `{` instead of `;` std/traits.d(2803:19)[warn]: Empty declaration std/traits.d(2804:24)[error]: Expected `{` instead of `;` std/traits.d(2804:24)[warn]: Empty declaration std/traits.d(8357:1)[error]: Expected...

This is great feedback you guys, thanks.

@JohanEngelen I figure it's a mistake to eliminate qualifiers (with `Unqual`). Due to `immutable`, immutable and mutable structures do not have compatible layout - e.g. you can't seat one on...