We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Fillion-Robin & Schreiner - A modern build-backend for CPython C/C++/Fortran/Cython extensions
Learn about scikit-build-core, a modern build backend for CPython that simplifies building C/C++/Fortran/Cython projects using CMake. Covers key features & real-world impact.
-
scikit-build-core is a modern build backend for CPython extensions that simplifies building C, C++, Fortran, and Cython projects
-
Key improvements over traditional build systems:
- Decouples build backend from frontend per PEP 517/518
- Provides cross-platform support via CMake
- Enables automatic dependency management
- Supports editable mode for development
- Integrates with IDE tooling
-
Features powerful configuration options:
- Static configuration via pyproject.toml
- Dynamic metadata updates
- Plugin system for extensibility
- File generation capabilities
- Environment variable control
-
Handles build dependencies intelligently:
- Automatically manages CMake and Ninja requirements
- Can use system-installed tools when available
- Supports version-specific requirements
- Enables cross-compilation
-
Real-world adoption and impact:
- ~800,000 downloads/month for scikit-build-classic
- Used by major projects like NVIDIA
- CMake wheel: 6.5M downloads/month
- Ninja wheel: ~10M downloads/month
- Supports 250+ packages including scientific computing tools
-
Provides development-friendly features:
- Debugging support for native code
- Fast incremental builds via Ninja
- Integration with external packages
- Standardized wheel building process
- Support for multiple build configurations
-
Backwards compatible while modernizing:
- Maintains support for existing scikit-build-classic projects
- Provides migration path to new features
- Leverages modern Python packaging standards
- Eliminates need for setup.py