[Guy Harris <gharris@xxxxxxxxx>]

Guy Harris wrote:

> -ansi -pedantic might, but it might also catch extensions we *want* to 
> use and that are properly protected by #ifdefs (e.g., declaring some 
> functions as being printf-like, so that the compiler can check the 
> format string against the argument list).  I'll try using the 
> --with-extra-gcc-checks option to turn on -pedantic (and some other 
> warnings), to see how well -ansi -pedantic works.

-ansi -pedantic produced a bunch of warnings that I cleaned up (a bunch 
of complaints about casting away constness - unfortunately, some can't 
be eliminated, either due to function signatures in external libraries 
that should have const but don't, or due to there not being a way I know 
of in C to say "once I've allocated this object and filled it in, and 
assigned the pointer to this variable or structure member, nobody's 
allowed to modify it, but when we're done with it I can free it", so you 
have to cast away the constness when freeing it - and some other cast 
issues).

However, -pedantic without -ansi produced the same warnings; presumably 
that's because, at least on the version of OS X I'm using, no 
declaration or definition that's omitted with __STRICT_ANSI__ is 
required, or something such as that.