Darling macOS compatibility layer adopts AI-assisted development policy and ARM64 support

Darling, the long-running macOS binary compatibility layer for Linux, has established an official policy governing AI-assisted contributions after more than a decade of development. The project now permits large language model generated code under strict conditions: contributors remain personally responsible for every line, must explain changes in their own words, and are forbidden from using disassembly, leaked materials, or Apple's restricted source code including corecrypto. The policy mandates implementation of documented contracts rather than application-specific hacks, preservation of macOS layering, and isolation of undocumented quirks.

Through AI-generated patches, the project has achieved notable progress including running the macOS version of League of Legends on Linux. Multiple community forks that embraced AI-generated code have emerged this year, and upstream Darling is now working to merge those contributions. The developers have also introduced initial ARM64 architecture support, expanding the project's reach beyond x86_64 systems.

AI policy and technical boundaries in Darling

• Clean-room implementation rules: Contributors must implement documented APIs rather than application-specific hacks, respect macOS subsystem layering, and isolate undocumented behavior with explicit labels; disassembly and Apple's restricted source including corecrypto are explicitly prohibited.
• Contributor accountability: Pull request authors bear responsibility for every line of AI-assisted code and must answer maintainer questions personally, not through automated responses, ensuring human verification of LLM output.
• ARM64 architecture expansion: Initial support for ARM64 processors has been added, extending Darling beyond its historical x86_64 focus and enabling compatibility with ARM-based Linux systems including Apple Silicon Macs running Linux and ARM servers.
• Upstream integration of fork work: Multiple community forks that adopted AI-generated code are being merged into the main Darling repository, consolidating dispersed development efforts under the new policy framework.
• Demonstrated capability: The macOS build of League of Legends now runs on Linux through AI-assisted patches, serving as a functional validation of the development approach for complex commercial software.

Why it matters

Darling's formalized AI policy offers a template for other resource-constrained open-source projects navigating LLM-assisted development while maintaining legal and ethical boundaries around reverse engineering. The ARM64 support broadens potential deployment to ARM servers and Linux-on-Mac hardware, while progress on commercial game compatibility signals potential competition with Wine for cross-platform binary execution.