Date: Thu, 10 Dec 2015 19:49:17 +0000 From: "Heyman, Jerrold" <Jerrold.Heyman@emc.com> To: "freebsd-questions@freebsd.org" <freebsd-questions@freebsd.org> Subject: <rpc/rpc.h> and rpc_createerr Message-ID: <9CDA60925D09954CA4BAD0284E2DFC43024EDE@MX204CL01.corp.emc.com>
next in thread | raw e-mail | index | archive | help
I've just installed FreeBSD 10.2 in order to determine the portability of m=
y companies code. Built gcc4.6 out of the ports/lang area, but see the sam=
e issue using /usr/bin/cc (clang 3.4.1).
in /usr/include/rpc/clnt.h the following snippet:
/*
* If a creation fails, the following allows the user to figure out why.
*/
struct rpc_createerr {
enum clnt_stat cf_stat;
struct rpc_err cf_error; /* userful when cf_stat =3D=3D RPC_PMAPFA=
ILURE */
};
__BEGIN_DECLS
extern struct rpc_createerr *__rpc_createeer(void);
__END_DECLS
#define rpc_createerr (*(__rpc_createeerr()))
Note that the #define becomes active once the file is included, and in my s=
ource code I have multiple
struct rpc_createerr *ce;
declarations. Both cc and gcc cite this as an error, though for different =
reasons.
gcc complains that a '(' is found where a '{' is expected.
The cc error message is 'error: declaration of anyonymous struct must be a =
definition'.
My other ports - Linux, AIX, Solaris, Mac OSX, do not have the #define in /=
usr/include/rpc/clnt.h.
The HP-UX does, but it is encapsulated within a #ifdef _REENTRANT / #endif =
block.
Is this an actual error, or is there something on FreeBSD that I need to do=
that is different than the other platforms?
Thanks in advance,
jerry
Jerry Heyman |
Principal Software Engineer | Software is the difference
EMC Data Domain | between hardware and reality
Jerrold.Heyman@emc.com / 919.597.7812 |
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9CDA60925D09954CA4BAD0284E2DFC43024EDE>