C ++头文件中的#ifdef __linux和#ifdef __unix之间的区别

Question

I've seen some boilerplate header file code for dealing with different OS's, like the following:

#ifdef __WIN32
    //windows
#elif __linux
    //linux
#elif __unix
    //all unix not captured above
#endif

If __unix catches all unices not captured by __linux , why not just use __unix ? I'm assuming it's some sort of safeguard, like "if __linux isn't enough, use this".

However, there has to be some reason both are used, instead of just having a catch-all with __unix (once again, this assumption may be wrong).

What's the deal?

Answer 1

As have been stated in the comments Unix and Linux are two different operating systems and although a large amount of software is compatible they have a separate licensing structure. This means that a number of tools that are available in Linux are not available in Unix and vice versa. An example of this is Unix's ZFS file system which cannot be distributed under the same license as Linux and as a result support is not included in stock Linux kernels (although it is possible to compile a custom kernel with support). In terms of the particular boiler plate it is likely that it is just covering all basis and could be used to take advantage of particular tools that are available on each operating system.