llama.cpp b9655 fixes grammar generator bug across all platforms

The llama.cpp project, with 117k stars on GitHub, released version b9655 focused on fixing a persistent grammar generator bug. The bug, described as an 'oldie but goodie,' was reintroduced during recent code changes and impacted grammar-constrained text generation. This feature is essential for applications requiring structured outputs, such as JSON generation or controlled text formatting. The fix ensures that grammar rules are correctly applied during inference, preventing errors like token mismatches or incomplete generations.

The release also highlights llama.cpp's extensive platform support. Builds are available for macOS (Apple Silicon and Intel), Linux (x64, arm64, s390x with Vulkan, ROCm, OpenVINO, SYCL), Windows (x64, arm64 with CUDA 12/13, Vulkan, SYCL, HIP), Android (arm64), and iOS (XCFramework). This breadth of support makes llama.cpp a versatile choice for deploying LLMs on edge devices, servers, and local machines. The bug fix ensures consistent behavior across all these environments, reinforcing the project's reliability for production use.