[Bug c++/79378] lambda init-capture adds const

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79378

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2017-02-06
     Ever confirmed|0                           |1

--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> ---
We also get the r2 case wrong for the example in [expr.prim.lamda] p20:

template <class T, class U> struct is_same { static constexpr bool value =
false; };
template <class T> struct is_same<T, T> { static constexpr bool value = true;
};
void f3() {
float x, &r = x;
[=] {
 // x and r are not captured (appearance in a decltype operand is not an
odr-use)
decltype(x) y1;
static_assert(is_same<decltype(y1), float>::value, "y1 has type float");
decltype((x)) y2 = y1;
static_assert(is_same<decltype(y2), float const&>::value, "y2 has type float
const& because this lambda is not mutable and x is an lvalue");
decltype(r) r1 = y1;
static_assert(is_same<decltype(r1), float&>::value, "r1 has type float&
(transformation not considered)");
decltype((r)) r2 = y2;
static_assert(is_same<decltype(r2), float const&>::value, "r2 has type float
const&");
};
}


lam2.cc: In lambda function:
lam2.cc:13:20: error: binding reference of type ‘float&’ to ‘const float’
discards qualifiers
 decltype((r)) r2 = y2;
                    ^~
lam2.cc:14:1: error: static assertion failed: r2 has type float const&
 static_assert(is_same<decltype(r2), float const&>::value, "r2 has type float
const&");
 ^~~~~~~~~~~~~