Don't use the parameter x in assert(x) if NDEBUG is defined to avoid any problems where (x) may contain symbols that are only defined if !NDEBUG Signed-off-by: Jon Ringle <jon at ringle.org> --- usr/include/assert.h | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/usr/include/assert.h b/usr/include/assert.h index d89e082..e59ed94 100644 --- a/usr/include/assert.h +++ b/usr/include/assert.h @@ -9,7 +9,7 @@ #ifdef NDEBUG -#define assert(x) ((void)(x)) +#define assert(x) ((void)(0)) #else -- 1.6.5.rc2
several autoconf tests use ferror() Stream errors are not being tracked by klibc implementation, so provide a dummy ferror() that just returns 0. Signed-off-by: Jon Ringle <jon at ringle.org> --- usr/include/stdio.h | 7 +++++++ 1 files changed, 7 insertions(+), 0 deletions(-) diff --git a/usr/include/stdio.h b/usr/include/stdio.h index 1993f75..7d2c86d 100644 --- a/usr/include/stdio.h +++ b/usr/include/stdio.h @@ -115,6 +115,13 @@ static __inline__ int fflush(FILE * __f) return 0; } +/* stream errors are not kept track of by klibc implementation */ +static __inline__ int ferror(FILE * __f) +{ + (void)__f; + return 0; +} + __extern int sscanf(const char *, const char *, ...); __extern int vsscanf(const char *, const char *, va_list); -- 1.6.5.rc2
Apparently Analagous Threads
- [PATCH 1/2] Fix must_inline macro in klibc/compiler.h for gcc-4.3
- [RFC] Usage of NDEBUG as a guard for non-assert debug code
- Question: Should we consider valid a variable defined #ifndef NDEBUG scope and used in assert?
- Question: Should we consider valid a variable defined #ifndef NDEBUG scope and used in assert?
- [RFC] Usage of NDEBUG as a guard for non-assert debug code