This PR definitely isn't ready for prime time, but I figured I might as well exploit Cunningham's Law here.
The idea is that when the compiler supports __cpp_impl_trivially_relocatable (P1144's proposed feature-test macro), then we should define -DBOOST_CONTAINER_ATTRIBUTE_TRIVIALLY_RELOCATABLE_IF(x)=[[trivially_relocatable(x)]]. (And/or set it to a vendor-specific attribute if-and-when anyone supports a vendor-specific attribute; no mainstream compiler does, yet.)
The big problem with this patch as it stands right now is that I have no idea where these macros are configured. We want to do the same thing that BOOST_CONTAINER_ATTRIBUTE_NODISCARD does; but that's not configured anywhere in this repo, either.
I see that Boost.Container is just inheriting all its type-traits from Boost.Move, so probably dtl::is_trivially_relocatable belongs in Boost.Move, not here?
This PR definitely isn't ready for prime time, but I figured I might as well exploit Cunningham's Law here.
The idea is that when the compiler supports
__cpp_impl_trivially_relocatable(P1144's proposed feature-test macro), then we should define-DBOOST_CONTAINER_ATTRIBUTE_TRIVIALLY_RELOCATABLE_IF(x)=[[trivially_relocatable(x)]]. (And/or set it to a vendor-specific attribute if-and-when anyone supports a vendor-specific attribute; no mainstream compiler does, yet.)The big problem with this patch as it stands right now is that I have no idea where these macros are configured. We want to do the same thing that
BOOST_CONTAINER_ATTRIBUTE_NODISCARDdoes; but that's not configured anywhere in this repo, either.I see that Boost.Container is just inheriting all its type-traits from Boost.Move, so probably
dtl::is_trivially_relocatablebelongs in Boost.Move, not here?Fixes #258.