Bruno Haible wrote: > Paul Eggert wrote: >> openat_needs_fchdir -- which is a function that >> openat.c must implement. > > If that was the only use of 'bool' in openat.c, I would not have done the > change. But it's also used in line 193. > >> In general in my experience it's OK for foo.c to assume >> interfaces provided by foo.h. > > Sure. But the interface of openat.c is <fcntl.h>. > > openat.h is just an auxiliary include file. You removed the > include <dirent.h> from it yesterday; you may want to remove > include <stdbool.h> from it tomorrow, when refactoring openat_needs_fchdir > in some way.
I see your point, but don't you think that if someone were to remove openat_needs_fchdir and the inclusion of <stdbool.h>, they would try to verify that openat.c still compiles?
