llama.cpp Release b10798 Makes Build Info Output Stream Configurable
Release b10798 of llama.cpp updates the build info function llama_print_build_info to accept a custom output stream parameter instead of hardcoding stderr. As a result, the version command in llama-app now directs plain version output to stdout. Directing version details to stdout rather than stderr aligns llama.cpp with standard Unix command-line conventions, improving scripting and output parsing. It also allows developers embedding llama.cpp to redirect build diagnostic logs to custom output destinations. The newly added FILE* parameter in llama_print_build_info defaults to stderr, ensuring complete backward compatibility for existing library callers. The pull request (#28322) was submitted by Adrien Gallouët from Hugging Face.
## BACKGROUND
llama.cpp is a widely used open-source C/C++ framework designed for running Large Language Models locally across diverse CPU and GPU architectures. In standard C and Unix environments, stdout is intended for normal program results, while stderr is reserved for error messages and operational diagnostics.