It'd be good to add something like this, thanks. A few comments:
The time-related stat macros are already broken out into stat-time.h. Perhaps these size-related macros should be broken out into stat-size.h? That way, we could leave stat-macros.h alone. Many programs that need CHMOD_MODE_BITS (in the current stat-macros.h) don't need the heavy artillery used to configure the size-related macros. If we make this change, the module should be renamed to "stat-size" of course. The indenting for the '#' lines in stat-macros.h isn't consistent. The usual gnulib style is to not count "#ifndef STAT_MACROS_H" as a separate indenting level. > +# define ST_NBLOCKS(statbuf) \ > + (S_ISREG ((statbuf).st_mode) \ > + || S_ISDIR ((statbuf).st_mode) \ > + ? st_blocks ((statbuf).st_size) : 0) This indenting doesn't look right, as || and ? should be different levels. I'd combine the 2nd and 3rd lines into one line. There's another definition of ST_NBLOCKS that has the same problem. (I realize this problem is in coreutils but we might as well fix it....) > +# endif /* _CRAY */ > +# endif /* not AIX PS/2 */ > +# endif /* !hpux */ > +#endif /* HAVE_STRUCT_STAT_ST_BLOCKS */ I'd omit comments on the "else" and "endif" lines, such as these, as they make the code harder to read. They can be helpful for long #if blocks, but these are short. > + AC_CHECK_HEADERS([sys/param.h]) This should be AC_CHECK_HEADERS_ONCE. > Depends-on: > +sys_stat > +stdint We can break the dependency on the stdint module by using (size_t) -1 instead of SIZE_MAX. Why does stat-macros need to depend on sys_stat? Unless there's a specific need for the dependency, I'd remove it.
