Re: openat includes

[Paul Eggert]

On 11/05/11 01:46, Bruno Haible wrote:
> lib/openat.c uses 'bool' and relies on openat.h for <stdbool.h>. This is a
> pitfall for future refactorings

I dunno, this sounds like overkill to me.

Any refactoring that would modify openat.h so much that
it would remove bool, would cause openat.h to no longer
specify openat_needs_fchdir -- which is a function that
openat.c must implement.  The chance that 'bool' would
be a pitfall without 'openat_needs_fchdir' being a larger
pitfall is essentially zero.

In general in my experience it's OK for foo.c to assume
interfaces provided by foo.h.  They're typically maintained
together and refactored together anyway.