Details
- Reviewers
beanz compnerd - Commits
- rGc451cf240b06: Add CLANG_DEFAULT_OBJCOPY to allow Clang to use llvm-objcopy for dwarf fission
rGcf1fee2d59c5: [CMake] Passthrough CMAKE_SYSROOT to external projects
rC317960: Add CLANG_DEFAULT_OBJCOPY to allow Clang to use llvm-objcopy for dwarf fission
rL317960: Add CLANG_DEFAULT_OBJCOPY to allow Clang to use llvm-objcopy for dwarf fission
rL317744: [CMake] Passthrough CMAKE_SYSROOT to external projects
Diff Detail
- Repository
- rL LLVM
Event Timeline
Sorry I've been bad at staying on top of your patches.
I think this shouldn't be an unconditional passthrough. I can imagine a situation where at least for the runtimes you might want different values set for different targets.
You've done most of the work in that area. What do you think?
We already use CMAKE_SYSROOT in build for our targets (here https://github.com/llvm-mirror/clang/blob/master/cmake/caches/Fuchsia-stage2.cmake#L45). The way this works is that the -DCMAKE_SYSROOT=... value we set in our cache file will be eventually passed to this function in ${ARG_CMAKE_ARGS} variable and since that one comes after the one I've added, CMake will pick that value as the final one. I admit it's not the prettiest solution though. I could change it to only set -DCMAKE_SYSROOT if it's defined and it's not already set in ${ARG_CMAKE_ARGS}, would you prefer that solution?
Sounds reasonable to me. The added complexity of not passing it doesn't seem worth it.