>Yeah, that may well be. I don't think we should have a policy of
>folding our arms and shouting "no" whenever someone asks us to clean
>up our namespace, but on the flip side one request (or even two) is
>probably not enough reason to do anything drastic, and this would be
>fairly drastic.
How about something less drastic? Could you at least eliminate "bool" from
interface structures that are intended to be compiled in multiple
environments? ("char" works fine, as does "pgbool"). Could you make c.h skip
the bool definition if it finds HAVE_BOOL defined? Then you could put in the
user guide where it talks about what header files and macros a server
extension needs if your program defines bool independently, define
HAVE_BOOL and if you want Postgres to define it, don't.
--
Bryan Henderson San Jose, California