[SYCL][NFC] Remove some work-arounds for C++11

[v-klochkov]

C++14 is required after #3053
Thus some of C++11 workarounds in the code became unnecessary.

Signed-off-by: Vyacheslav N Klochkov [email protected]

[bader]

What about similar work-arounds in other files (e.g. accessor.hpp, buffer.hpp, etc.)?

[AGindinson]

@v-klochkov, do you have plans on type traits' refactoring in scope of this shift, or is this something I could volunteer for instead of finalizing #2778?

[v-klochkov]

@v-klochkov, do you have plans on type traits' refactoring in scope of this shift, or is this something I could volunteer for instead of finalizing #2778?

What about similar work-arounds in other files (e.g. accessor.hpp, buffer.hpp, etc.)?

@bader, This fix removes only some workarounds, not all of them. In particular it removes a bit more than was promised here: #2528 (review)

[v-klochkov]

@v-klochkov, do you have plans on type traits' refactoring in scope of this shift, or is this something I could volunteer for instead of finalizing #2778?

@AGindinson , Sure, you can volunteer to do more code cleanup. With C++14 sycl::detail::enable_if_t can be replaced with std::enable_if_t, etc..
C++14 though still does not let you write something like std::is_same_v<T1,T2>

[bader]

@v-klochkov, do you have plans on type traits' refactoring in scope of this shift, or is this something I could volunteer for instead of finalizing #2778?

What about similar work-arounds in other files (e.g. accessor.hpp, buffer.hpp, etc.)?

@bader, This fix removes only some workarounds, not all of them. In particular it removes a bit more than was promised here: #2528 (review)

That was my question. Why do we remove only some workarounds and not all of them?

[v-klochkov]

That was my question. Why do we remove only some workarounds and not all of them?

There is no rush here. Having everything in one change would require a huge patch, which then would suffer from conflicts, etc.
For example, If someone would want to replace detail::enable_if_t with std::enable_if_t, it would be 44+ files.

In this patch I changed only places with "#if __cplusplus >= 201402L"
Other places that you mentioned above are different as they have (">", not ">=") : "#if __cplusplus > 201402L"