On Friday 04 May 2012 03:37:29 Bernhard Reutner-Fischer wrote: > On 4 May 2012 01:22, Mike Frysinger <[email protected]> wrote: > > On Thursday 03 May 2012 04:58:24 Bernhard Reutner-Fischer wrote: > >> Mike, it would have been much simpler to just leave the header impl-free > > > > yes, and it would cause unused bloat in the C lib. the func is dirt > > simple, so for the rare case where someone actually calls this, they get > > a few memory lookup/compares. for everyone else, there is no overhead > > with an unused function in the C library. > > > >> --- a/include/stdlib.h > >> +++ b/include/stdlib.h > >> > >> -__nonnull ((1)) __wur > >> -static inline int rpmatch (__const char *__response) > >> +/* __THROW */ > >> +static __inline __nonnull ((1)) __wur int rpmatch (__const char > >> *__response) > > > > i don't see what the problem is here. the code compiles fine for me with > > C and C++ compilers. > > it did not parse for me, libcheck (from check.sf.net) was > misconfigured due to parse errors in the above.
seems -ansi does not like "static inline". could easily change it to __always_inline. -mike
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ uClibc mailing list [email protected] http://lists.busybox.net/mailman/listinfo/uclibc
