Indeed, that was my point. As an aside, clang has the —driver-mode option that lets you specify either gcc or g++, which is in turn equivalent to running clang.exe vs clang++.exe. Does gcc have this? If so we can require it to be the non ++ version, and then build the driver mode argument as necessary.
If not, we can compute the other executable name based on the one variable On Mon, Oct 23, 2017 at 4:52 PM Pavel Labath via Phabricator < revi...@reviews.llvm.org> wrote: > labath added a comment. > > In https://reviews.llvm.org/D39215#904536, @labath wrote: > > > In https://reviews.llvm.org/D39215#904510, @zturner wrote: > > > > > There is already a CMake variable called `LLDB_TEST_COMPILER`. This > `LLDB_TEST_CLANG` was introduced later, I guess unaware of the presence of > `LLDB_TEST_COMPILER`. Would it be possible to standardize on one > variable? This would also mean deleting the `TEST_C_COMPILER` and > TEST_CXX_COMPILER` variables. I don't see why we should ever have both. > Run the test suite a second time if you need to test with different C and > C++ compilers. > > > > > > Unlike msvc, where you just use CL to compile everything, on ***nix you > have different compiler drivers for C (gcc, clang) and C++ (g++, clang++). > The differences are minute, but we still need both (clang has > --driver-mode=g++ argument to emulate g++, but I don't know of any gcc > option like that). > > > > However, I do see a bit of space for simplification. I can remove > LLDB_TEST_CLANG altogether, and just have LLDB_TEST_???_COMPILER default to > clang. > > > Scratch that, dotest.py already has some magic to compute the c++ > compiler, given the c version, so I guess one variable is really enough. > > > https://reviews.llvm.org/D39215 > > > >
_______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
