Inline functions (Re: [PATCH/RFC v7 1/2] Add infrastructure for translating Git with gettext)

Ævar Arnfjörð Bjarmason wrote:

As an aside, I'd appreciate a document pointer, what are the
implications here exactly?

Oh, good point!  I clean forgot to provide that.

"static inline" functions are just like macros, except each argument
is evaluated exactly once and the arguments are typechecked.

In the older dialect, what I was suggesting would be written as
follows:

 #ifdef NO_GETTEXT
 #define git_setup_gettext() do { } while(0)
 #else
 extern void git_setup_gettext(void);
 #endif

The feature has been part of GCC and various compilers for a long time
now.  It was finally officially adopted when the C99 standard came
out.  It "degrades gracefully" in that the Makefile will define inline
to nothing on those very old compilers known not to support it, making
these into static functions (which is kind of ugly but with most
optimizers it tends to produce the same result).

GCC documentation:

  http://gcc.gnu.org/onlinedocs/gcc/Inline.html

C99 documentation (see the "most current draft" of the revised C99
and navigate with PDF bookmarks to Language → Declarations →
Function specifiers → Semantics):

  http://www.open-std.org/jtc1/sc22/wg14/www/projects#9899

More important than all that is how the construct gets used and why.
For this, see SECTION 2 item 2 ("Ifdefs are ugly") of
linux-2.6/Documentation/SubmittingPatches[1]: and take a look at some
of the surrounding code (maybe include/linux/kernel.h would be
analogous) for examples.

Hope that helps,
Jonathan

[1] http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=Documentation/SubmittingPatches