Hrvoje> Martin, we'll *have* to resolve current INLINE yuckage somehow. I, Hrvoje> for one, would expect INLINE to expand either to `inline' or nothing. Yup. Especially given that: - CONST resolving to either `const' or `' suggests INLINE should resolve to `inline' or `'. - Stallmacs 20.2 has this (primitive) code that at least conforms to the standard intuition: /* If using GNU, then support inline function declarations. */ #ifdef __GNUC__ #define INLINE __inline__ #else #define INLINE #endif In 21.1 we should (will?): - nuke CONST by renaming to `const', in accord with AC_C_CONST. - start using `inline' liberally in .c files for small or heavily called static functions. - rename INLINE to something more meaningful. Hmmmmm..... Can't think of a good name .... maybe EXTERN_INLINE or HEADER_INLINE or INLINE_FUNCTION_USED_WITHIN_MULTIPLE_COMPILATION_UNITS ?? Martin