https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117174
Bug ID: 117174
Summary: Compiler seems to incorrectly cache SFINAE condition
evaluation results
Product: gcc
Version: 14.2.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: ivan.solovev at qt dot io
Target Milestone: ---
https://godbolt.org/z/197vedfYz
```
namespace A {
template <typename T>
bool func(T *) { return true; }
} // namespace A
namespace B {
namespace C {
using A::func;
template <typename T, typename = void>
constexpr bool inline hasFunc = false;
template <typename T>
constexpr bool inline hasFunc<
T, std::void_t<decltype(func(std::declval<T>()))>
> = true;
} // namespace C
template <typename T, typename = void>
constexpr bool inline hasCustomFunc = false;
template <typename T>
constexpr bool inline hasCustomFunc<
T, std::void_t<decltype(func(std::declval<T>()))>
> = true;
} // namespace B
struct S {};
bool func(S *) { return false; }
static_assert(B::hasCustomFunc<S*>);
static_assert(!B::hasCustomFunc<int*>); // FAILS here!
static_assert(B::C::hasFunc<int*>);
```
This code uses the same condition
std::void_t<decltype(func(std::declval<T>()))>
to figure out if the function overload is available for type T.
The difference is that `B::C::hasFunc` also sees the template from namespace A,
while `B::hasCustomFunc` should not see it.
It looks like GCC evaluates the condition only once and reuses it in the second
check, which is incorrect. Updating the second condition to look like
std::void_t<decltype(func(std::declval<T>()) == true)>
fixes the issue.
