llama.cpp Release b11900 Improves Model Resolution Testing

llama.cpp release b11900 introduces test enhancements designed to accurately pin model resolution against real Hugging Face repository layouts and file naming conventions. It extends the resolution testing framework by integrating dozens of actual repository listings captured directly from the Hugging Face API. This update ensures robust compatibility and reliable automated testing when loading quantized models directly from Hugging Face repositories. Developers and local AI tools relying on llama.cpp benefit from more stable and predictable model loading behaviors. The update incorporates 63 verbatim repository listings and 171 convention-following listings to catch resolution bugs by verifying file headers. It also includes filesystem workarounds to safely skip overly long split paths that exceed Windows path length limits.