We are pleased to announce that CMake 4.0.2 is now available for download.
Please use the latest release from our download page:
https://cmake.org/download/
See also the release notes and release milestone
Thanks for your support!
Changes made since CMake 4.0.1:
Ben Boeckel (1):
- Help/cmake-cxxmodules: note GCC 15 support
Brad King (12):
- Help: Add 4.0 release note about file path case sensitivity on Windows Normalize PWD environment variable before using it
- CMakeDetermineCompilerABI: Comment implicit link directory exclusion
- macOS: Restore RPATH entry for libraries linked from /usr/local/lib
- ci: Tell CudaOnly.Architecture test what specific architecture to use
- ci: add tags to tie CUDA jobs to runners with hardware supporting them
- CUDA/Clang: Prefer NVCC’s default architecture for each CUDA Toolkit version
- Tests: Fix RunCMake.CTestCommandLine under ctest -T Coverage
- AUTOMOC: Avoid compiler warnings while computing predefines
- Help: Fix wording in tutorial workflow description
- macOS: Restore support for LLVM/Clang without explicit CMAKE_OSX_SYSROOT
- CMake 4.0.2
Craig Scott (2):
- Help: VS_DEBUGGER_WORKING_DIRECTORY precedence
- Xcode: Use DEBUGGER_WORKING_DIRECTORY as a fallback for scheme work dir
Gregor Jasny (1):
- macOS: Reliably apply workaround for system curl 8.{3,4,5} LibreSSL backend
Marc Chevrier (3):
- Ninja generator: restore correct encoding for policy CMP0181=OLD
- FindPython: fix the definition of SABI artifacts
- Help: block() command: enhance the
PROPAGATEexample
Matthew Woehlke (3):
- Help: Fix overlapping text in footnotes
- CPS: Fix mangling of relative requirement names
- find_package: Fix performance regression in 4.0.0 release
Raul Tambre (1):
- Clang: MSVC C++26 support
Robert Maynard (3):
- ci: Tell CudaOnly.CompileFlags test what specific architecture to use
- CUDA/Clang: Update architectures supported by CUDA 12.8
- CUDA: Add CUDA 12.9 new architectures and family syntax