logging
dynlib

warning compare with 0 in format_float.cpp