Boost logo

Boost :

Subject: Re: [boost] Yap's formal review is starting now!
From: a.hagen-zanker_at_[hidden]
Date: 2018-02-16 10:41:03

> Peter Dimov:
> It now occurs to me that if my_expr is not <Kind, Tuple>, but <Kind, T...>, a transform would be able to match it without knowing the tuple type.

Zach, are you giving this serious consideration?

I think it sounds like an excellent suggestion; it would bring the interface closer to the conceptual model (i.e. an expression consists of a function and its arguments, which can also be expressions) . The discussion of std::tuple / hana::tuple / TupleLike suggests to me that something ended up in the interface that really should be an implementation detail.

I realise this would be a big refactor.

One other advantage, beyond conceptual clarity, would be that the Transform Matching will be much more concise.

Unsubscribe & other changes:

Boost list run by bdawes at, gregod at, cpdaniel at, john at