Non-copy-constructibility tests fail

[eyalroz]

See also issue #16.

The following tests are failing:

bool test1 = ! std::is_copy_constructible
    <strf::constrained_fpe<is_64, facet<0, enable_only_move>>>
    ::value;

bool test3 = ! std::is_copy_constructible
    <strf::facets_pack<facet<0, enable_only_move>>>
    ::value;

bool test5 =  ! std::is_copy_constructible
    < strf::facets_pack
         < strf::constrained_fpe
             < is_64, facet<0, enable_only_move> >>>
    :: value;

bool test7 = ! std::is_move_constructible
    < strf::facets_pack<facet<0, disable_copy_and_move>> >
    ::value;

It's not so clear to me why that's happening, though. Perhaps @robhz786 can help?

[robhz786]

If you declare a copy constructor, then std::is_copy_constructible<your_class>::value is true unconditionally, even if the copy constructor fails to compile. And I don't think it is possible to use any sfinae trick here. The only solution seem to defaulting the constructors ( either by omitting the declaration or by using =default; ). If this is not possible, I think the issue is unsolvable.

https://akrzemi1.wordpress.com/2015/03/02/a-conditional-copy-constructor/

[eyalroz]

@robhz786 : But are we talking about declarations I added?

Oh, I remember there were someplaces I had to replace = default; with manually writing a default ctor, to make it STRF_HD. Maybe that's the cause here?

[robhz786]

@eyalroz Yes, that's the cause.
I created a pull request now. Take a look: #18
Question: What happens if we declare the copy and move constructors with = default; but without STRF_HD ?

[eyalroz]

All tests now pass.