With 7.1 llvm-config looks for libLLVM-7.1.dylib, but the shortVersion
was still 7 so only 7 and 7.1.0 where aliased before.
$ llvm-config --link-shared
llvm-config: error: libLLVM-7.1.dylib is missing
Got this trick from stack overflow to avoid needing compiler-rt to link
the test exe before building compiler-rt. a static lib isn't linked at
all, and so breaks the cycle.
Adds pkgsCross.wasm32 and pkgsCross.wasm64. Use it to build Nixpkgs
with a WebAssembly toolchain.
stdenv/cross: use static overlay on isWasm
isWasm doesn’t make sense dynamically linked.
Backport the `--unwindlib` flag, so it don't try to use libgcc on Linux
for hysterical reasons. The alternative of passing `-lunwind` always
would mess up C.
this adds libc++ to the LLVM cross, giving us access to the full
Nixpkgs set. This requires 4 stages of wrapped compilers:
- Clang with no libraries
- Clang with just compiler-rt
- Clang with Libc, and compiler-rt
- Clang with Libc++, Libc, and compiler-rt
This iteration was long, about five weeks (2fcb11a2), I think.
Darwin: it's missing a few thousand binaries and there's a make-netbsd
regression, but I suppose these aren't merge blockers.
When building with -DCOMPILER_RT_DEFAULT_TARGET_ONLY=ON the compiler-rt
CMake configuration attempts to use CMAKE_*_COMPILER_TARGET variables,
which are usually only defined in cross-compilation mode.
You can build (partially) with LLVM toolchain using the useLLVM flag.
This works like so:
nix-build -A hello --arg crossSystem '{ system =
"aarch64-unknown-linux-musl"; useLLVM = true }'
also don’t separate debug info in lldClang
It doesn’t work currently with that setup hook. Missing build-id?
LLVM should be target independent because it will work with all
machine types. This is different from GCC where it needs to know what
target to build ahead of time.
* llvm: build w/polly in-tree, optionally
Don't enable by default yet, defer rebuilds.
* top-level: llvm-polly, clang-polly
* bit hacky re:overrides
* need to explicitly set cmake flags for clang to link in polly
LLVM should be target independent because it will work with all
machine types. This is different from GCC where it needs to know what
target to build ahead of time.
