Thread: Function Inlining

My school forbids function definitions and .c file inclusions in header files.
Is there a way to inline a function that is declared in a header file and defined in a .c file, while following the above mentioned constraints?

At school we use macs, which apparently, have some strange mix of gcc 4.2 and llvm 9.1.

Code:

 gcc --version
Configured with: --prefix=/Library/Developer/CommandLineTools/usr --with-gxx-include-dir=/usr/include/c++/4.2.1
Apple LLVM version 9.1.0 (clang-902.0.39.1)
Target: x86_64-apple-darwin17.5.0
Thread model: posix
InstalledDir: /Library/Developer/CommandLineTools/usr/bin

Thank you.

Without functions, all you can use are preprocessor macros. Why does your school forbid the use of functions though?

Originally Posted by Dren

My school forbids function definitions and .c file inclusions in header files.
Is there a way to inline a function that is declared in a header file and defined in a .c file, while following the above mentioned constraints?

The reason why you would declare a function in a header file is so that the header file could be included in multiple source files, and hence the function could be conveniently called in multiple translation units. For a function to be inline, it has to be defined in the same translation unit as it is declared, which implies that it has to be defined in the same translation unit as it is called, since it should be declared before being called.

Therefore, the answer to your question is yes: you could declare it as inline in the header, and then define it in each source file where it is used, but that is an absurd solution (and I think you'll run into redefinition errors unless you also declared it static so as to give it internal linkage).

Can you demonstrate that with inlining, your code will experience a speedup that is both significant and important to what you are trying to do? If so, approach your instructors to get the rule changed because you have demonstrated that it is a silly rule insofar as it does not consider function inlining. If not, then there's no real benefit inlining, so just write the code as prescribed and move on.

Originally Posted by Dren

I will *have* to declare it static as well, as that is another rule at my school.

What do you mean? In general, you should only declare a function static if it is meant to be implementation detail, since it would then have internal linkage, hence avoiding the problem of a name collision with another implementation detail function in a different translation unit. As such, how can that possibly be a rule at your school? Surely you don't mean that the rule requires you to declare all functions static, even those declared in header files so as to be included in multiple translation units?

Originally Posted by Dren

We can't use for loops, switch-case statements, or interlinked ternary operators either. Nor can we have more than 5 function definitions per source file...

Well, there are good reasons why you might not want to nest ternary operators (i.e., readability and it is quite plausible readers will get the precedence mixed up if you don't enforce grouping with parentheses), but the others are unusual unless it is for a beginner's exercise.

Originally Posted by Dren

My school forbids function definitions and .c file inclusions in header files.
Is there a way to inline a function that is declared in a header file and defined in a .c file, while following the above mentioned constraints?

At school we use macs, which apparently, have some strange mix of gcc 4.2 and llvm 9.1.


Thank you.

NOTE: I read the above as your school allows function prototypes in header files, is this correct?

Tim S.

Links Inline function - Wikipedia
and Inline Functions In C

Showed me something new the use of "static inline".

Tim S.

Originally Posted by jimblumberg

You should only have macros and function declarations (prototypes) in header files.

If it is indeed meant to be used in multiple translation units, an inline function with external linkage should be defined in a header file. If not, then perhaps it should have internal linkage instead and be defined in a source file. EDIT: come to think of it, if you want to define the multiply-used inline function externally, I guess that could still work, but I suspect that you'll then depend on the compiler doing whole program optimisation in order for inlining to actually happen even for functions that are good candidates for inlining.

Originally Posted by jimblumberg

I doubt that inline functions are allowed since function definitions are not allowed in header files

They didn't forbid it, and hence Dren's plan of defining inline functions with internal linkage within source files could still be within the rules.