search

mpark / variant

C++17 `std::variant` for C++11/14/17
https://mpark.github.io/variant
Boost Software License 1.0
664 stars 87 forks source link

Apply patch for icpc #78

Closed sbolding-LANL closed 3 years ago

sbolding-LANL commented 3 years ago

icpc in some way utilizes the preprocessor of the associated "developer tools" used by the compiler. This leads to, in some cases, a preprocessor claiming support for __tuple_element_packs, even though icpc (as of version 21.1) can't actually parse such code. Just use the MPARK_TUPLE_ELEMENT_PACK impl with __icc until icpc supports it.

Fixes #77

FYI, even with this patch, icpc 19.1.3 fails to compile one of the mpark tests (I couldn't seem to clone the llvm tests on my system, but probably because of our network). All tests but this one compiled and passed:

/Users/sbolding/codes/mcnp6/build/git-submodule-packages/mpark-variant/test/assign.fwd.cpp(53): error: static assertion failed with "variant<short, long> v; v = 42;"
    static_assert(std::is_assignable<mpark::variant<short, long>, int>{},

I imagine any Intel users are already working around this if they needed to SFINAE on something like this, so I am OK with this limitation until the new Intel clang front ends are widely available :), but without this patch we can't build at all on Catalina.

ax3l commented 3 years ago

@mpark any news on this one? :)

mpark commented 3 years ago

Read through a bit of the history and issue... it really doesn't seem like something I want to dig deep into. Going to merge this for now and maybe come back if there are other ICC issues that come up.