MacOS is really frustrating with how it handles its libraries and compilers. It is also frustrating because it ships an unspecified version of LLVM, which generally isn't the latest stable release. You can, however, with a little tweaking, use the latest version of LLVM or GCC on your Mac, and reliably use it for your C and C++ tooling.
First, you need to install the latest version of LLVM. Most people nowadays are using Homebrew. If you don't have it, you can install LLVM from source, which takes a lot of time to compile. If you have brew, you can install LLVM with
brew install llvm
You have several options for installing LLVM, all of which will require brew to compile it from source. I usually install the latest stable version of LLDB and I build it against Homebrew's Python 2.
brew install llvm [email protected] --with-lldb
This took me roughly two hours to build on a 2015 rMBP 13" with an i5.
Once it's done, add Brew's LLVM to your path:
echo 'export PATH="/usr/local/opt/llvm/bin:$PATH"' >> ~/.zshrc # or ~/.bashrc depending on your shell
Most build tools (namely CMake and Make) respect certain environment variables, so you should set these once you have the LLVM suite in your path.
I usually go ahead and set up aliases so I can conveniently compile small programs from the command line with the latest version of clang.
For environment variables:
export CC=clang export CXX=clang++ export LD=ld.lld export AR=llvm-ar export RANLIB=llvm-ranlib
Note: if you don't have LLVM in your
$PATH, then you will need to add the
full path to your environment variables
These are the aliases I set:
alias cc=$CC alias c++=$CXX alias ld=$LD alias ar=$AR alias ranlib=$RANLIB
Working with XCode
If you use XCode or
xcodebuild, then you will realize that it does not use
the versions of LLVM/Clang that you set up in the environment. The LLVM
project actually provides a way to build an XCode toolchain that contains
everything you need to switch XCode to the latest versions of clang and the
other tools you need to compile your projects.
In order to do this, you need to manually build LLVM from source and build the toolchain.
You can follow this tutorial which shows you how to build LLVM from the Github mirrors.
When you get to section 3, change the cmake command slightly so you enable the XCode toolchain target:
mkdir -p build cd build cmake -G Ninja \ -DCMAKE_BUILD_TYPE=RelWithDebInfo \ -DLLVM_CREATE_XCODE_TOOLCHAIN=On
I also recommend using
Ninja rather than
Make to build LLVM, because it
will build significantly faster.
Now that you have the XCode toolchain, you can place it in the
directory in XCode (bump the version as necessary, this guide may not always be
up to date).
mv LLVM8.0.1.xctoolchain /Applications/Xcode.app/Contents/Developer/Toolchains
You need to instruct XCode to actually use the toolchain. You can do so in two ways: from your environment variables, and through the XCode app itself.
To set it through an environment variable:
Xcode.app, you can select
Xcode -> Toolchains -> org.llvm.8.0.1 in the
These steps should generally work with your standard C/C++ Make, CMake, and Xcode setups, though there can be a lot of quirks because of the way people install libraries and how Apple sets up their compiler suite.