llama.cpp Releases Version b10276 with Minor Security Update
llama.cpp has released version b10276, which introduces a minor security improvement by switching from 'npm install' to 'npm ci' in its build process. Using 'npm ci' instead of 'npm install' in automated environments ensures reproducible builds and prevents unexpected dependency updates, reducing potential supply chain security risks. The update is implemented via pull request #26601. Additionally, this release disables the KleidiAI-enabled build for macOS Apple Silicon (arm64) and disables openEuler builds.
## BACKGROUND
llama.cpp is a popular open-source C/C++ implementation for running LLMs locally with high performance. In Node.js development, 'npm install' can update the package-lock.json file and install newer compatible dependency versions, whereas 'npm ci' strictly installs the exact versions specified in the lockfile, making it safer for automated builds. KleidiAI is an open-source library by Arm that optimizes AI workload performance on Arm CPUs.