Paul Fultz II

Well, if later I decide as an extension to add some basic algorithms for sequences, I wouldn't be able to call it `fold`.

I am not planning on it currently, however, I don't want to close the door to the possibility.

If its called `fold` then fold for a sequence would look like this: ``` cpp FIT_STATIC_LAMBDA_FUNCTION(fold) = [](auto&& sequence, auto f) { return fit::unpack(fit::fold(f))(std::forward(sequence)); }; ``` I think the double...

> The fold for a sequence should has the signature : function, sequence, not sequence, function. In C++ its always defined with the sequence first. Even Hana defines it that...

> I don't think it looks strange, actually I think it highlights the fact that fit::compress is really a fold. However, it will never be `Foldable` in the Hana sense.

> It is not auto unpacking, it is about the type of the function. If the function takes the sequence as argument, fold should not unpack, but if the function...

> Yes, but I think in the C++ community people would expect accumulate always to be numeric, and this is more generally useful. I think fold() and reverse_fold() tell everything...

> You never define Foldable in your library. As a user, why do I care that hana::Foldable exists and that it is different? Because a user may use their understanding...

Well, it appears that Boost.Config's test depend on Boost.Core, which depends on Boost.Config. I really would like to use Boost.Config, but the cycles need to be removed before it can...

There is really only two remedies: - Continue using the current configuration macros - Build another configuration library for boost that does not have cycles I don't like either of...