On 02/23/2011 08:11 AM, Daniel P. Berrange wrote:
> Huh, dispatch.c doesn't include "internal.h". Nor
does dispatch.h or
> remote.h. This only compiled because you included "memory.h", which
> eventually included "internal.h".
I'm not really seeing that as a problem. This is true for anything from
internal.h, not merely stdbool usage.
> If we go with this approach, shouldn't we also ensure that all .c files
> include "internal.h"?
I'm not sure why we need to treat internal.h specially. We rely on
transitively including things everywhere.
If you're okay with transitive rather than explicit inclusion, then I
can live with that policy. It's easy enough to test, after all (did it
compile?). So:
ACK.
--
Eric Blake eblake(a)redhat.com +1-801-349-2682
Libvirt virtualization library
http://libvirt.org