llama.cpp b9505 adds server header for improved HTTP handling

The open-source community favorite llama.cpp has shipped version b9505, a focused maintenance release by ggml-org. The primary change is the addition of a header file to `tools/server/server-http.h`, which helps organize server-side HTTP handling code. This is a small but necessary step in keeping the server component modular and maintainable, especially as the project supports an ever-growing list of hardware backends.

The release is available across all major platforms and accelerators: macOS (Apple Silicon, Intel), Windows (CPU, CUDA 12/13, Vulkan, HIP), Linux (CPU, Vulkan, ROCm, OpenVINO, SYCL), and Android (arm64). Notably, it also includes an iOS XCFramework. While not a headline feature, b9505 ensures that llama.cpp remains robust for developers deploying local LLMs via its HTTP server, reinforcing its position as the go-to tool for running models like Llama and Mistral on consumer hardware.