llama.cpp Release b11467 Fixes C++ Header Include for imatrix
llama.cpp released build b11467, introducing a code patch that explicitly includes the standard <clocale> header for std::setlocale in the imatrix tool. Explicit header inclusions ensure consistent C++ compilation across diverse platform toolchains and compilers. This prevents unexpected build failures for developers and users building llama.cpp from source. The patch stems from PR #30079 co-authored by github user dazzywi to fix missing declarations required by std::setlocale. Additionally, release notes indicate that pre-built macOS Apple Silicon binaries with KleidiAI optimizations remain disabled.
## BACKGROUND
llama.cpp is a high-performance open-source C/C++ framework for running Large Language Model (LLM) inference locally across various hardware backends. The imatrix (importance matrix) tool inside llama.cpp calculates token activation statistics to maintain model accuracy during GGUF quantization. In C++, functions dealing with localization like std::setlocale require including the <clocale> header, without which strict standard compilers may reject the code.