Re: Explicit casts in ctype.h suppress compiler warnings

[Rich Felker <dalias@libc.org>]

On Fri, Apr 17, 2015 at 01:59:14PM +0300, Alexander Monakov wrote:
> For the following erroneous source code:
> 
> #include <ctype.h>
> int f(char *c)
> {
>   return isdigit(c) || isspace(c);
> }
> 
> GCC warns only for passing a pointer to isspace; isdigit is implemented as a
> macro that casts its argument to unsigned, and the warning is suppresed
> because the origin of the cast is in a system header.  Since isspace is
> implemented with a static inline helper function, there is a warning.  With
> glibc headers, no warning is issued in either case for a similar reason.
> 
> I think it would be nice if musl's ctype.h could aid the compiler in
> diagnosing erroneous use, like it happens today for only for isspace() of all
> macros declared there.  The cost of restructuring the header to achieve that
> does not seem too high.  Thoughts?

Do you have an idea in mind for how we could achieve that? I suspect
the macros are still better optimizable than the inline function
approach, so I'd lean towards doing a macro that avoids evaluating c
and just checks its type, which would involve using ?: I think.

Rich