Why does the C preprocessor interpret the word “linux” as the constant “1”?

Re:Why does the C preprocessor interpret the word “linux” as... by twoodfin on Wednesday October 09, 2013 @10:45AM (#6521526)

I'm increasingly turning to -E to debug these hard-to-understand problems. I'd love a -E2 or what-have-you that would be even more verbose, and list all preprocess transformations applied to a line and the source lines where they were defined.

Unfortunately, the standard preprocessor output doesn't provide a convenient way to do this without a comment format. I guess you could use a do-nothing #pragma.

Eventually, I'm sure somebody will spin up some extension to clang to show you the complete evolution of a block of code and what other code contributed to that evolution.

[ Parent | on HN ]

Re:Why does the C preprocessor interpret the word “linux” as... by simias on Wednesday October 09, 2013 @10:52AM (#6521566)

Emacs has a "c-macro-expand" function which is quite hackish but extremely useful, IMO: it attempts to macro expand the region.

For instance running it on the following code in the middle of a C source file:

    #define BOGO_MAX(a, b) a > b ? a : b

    int test()
    {
      return BOGO_MAX(3 + 4, 5);
    }

Creates a new buffer containing:

    int test()
    {
      return 3 + 4 > 5 ? 3 + 4: 5;
    }

[ Parent | on HN ]

Re:Why does the C preprocessor interpret the word “linux” as... by tedunangst on Wednesday October 09, 2013 @11:46AM (#6521982)

How well does that play with ifdef? The problem is rarely what does this macro expand to, but which macro is going to be expanded.

[ Parent | on HN ]

Re:Why does the C preprocessor interpret the word “linux” as... by simias on Wednesday October 09, 2013 @11:49AM (#6521999)

Well, I guess it depends on what your code looks like.

That being said if I'm not sure if some code is being compiled I just add an "#error foo" and rebuild. Or even simpler I just type in some garbage to trigger a compilation error.

[ Parent | on HN ]