Bug ID: 37013
           Summary: clang -x cuda fails in some templates on >>>
           Product: clang
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: CUDA

Compiling C++11 template code with `>>>` fails in some cases as soon as `-x
cuda` is used.

Although `>>>` should be valid in C++11, they might be mismatched here with
CUDA kernel arguments.

  clang version 7.0.0 (
780143bbd516aa39d754e5cab50955321b71fba0) (
  nvcc 8.0.44

It would be great if closing multiple nested template arguments in such manner
works with the CUDA frontend as well.

Reproducer: Issue first seen with Boost.mp11 (1.66.0) with minimal example:

  #include <boost/mp11.hpp>
  int main() {}

Compiled with

  clang++ -std=c++11 -x cuda main.cpp  --cuda-path=$CUDA_ROOT

We could not boil it down to a non-mp11 code snippet yet, maybe it's related to
the heavy use of variadic template arguments in combination with `>>>` ?

You are receiving this mail because:
You are on the CC list for the bug.
llvm-bugs mailing list

Reply via email to