How to use std::format without "macro redefinition" errors?

[s9w]

There's this info about cmake options, and this site about defines. What’s the relationship between the CMake option LOG4CXX_FORMAT_NAMESPACE=std and the define LOG4CXX_FORMAT_NS=std? Do I need to set one, or both?

Also the sentence "You would use the value 'std' if Log4cxx was built with LOG4CXX_FORMAT_NAMESPACE=fmt" is adding to my confusion.

[swebb2066]

Yes, it is a mess.

The CMake option LOG4CXX_FORMAT_NAMESPACE=std is used to set the default value for LOG4CXX_FORMAT_NS (i.e. when that preprocessor macro is not provided by the application developer). It is also used when building Google Benchmark checks. It allows developers to compare the performance of the fmt and standard libraries. Setting the LOG4CXX_FORMAT_NAMESPACE option when building Log4cxx does not prevent an application from using the fmt library to build a logging request message.

An application developer can use the standard library implementation of format (i.e. std::format) in an app's logging requests by adding LOG4CXX_FORMAT_NS=std to the cmake target_compile_definitions command. If that preprocessor macro is absent, the LOG4CXX_XXXXX_FMT logging macros will expand to the default value of LOG4CXX_FORMAT_NS. When LOG4CXX_FORMAT_NAMESPACE is not specified when building Log4cxx, fmt::format is used.

Let me know if that clarifies your understanding (or where it does not) so I can improve the website documentation.