From: Albert Chin (postfix-usersmlists.thewrittenword.com)
Date: Tue Jun 21 2005 - 16:56:34 CDT

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Tue, Jun 21, 2005 at 04:46:15PM -0500, Albert Chin wrote:
> On Tue, Jun 21, 2005 at 02:46:11PM -0500, Albert Chin wrote:
> > On Tue, Jun 21, 2005 at 03:18:13PM -0400, Wietse Venema wrote:
> > > If IRIX is too much trouble, consider putting
> > >
> > > #define CANT_USE_SEND_RECV_MSG
> > >
> > > in the appropriate sys_defs.h section.
> > >
> > > I would not recommend playing with _KERNEL, _POSIX and stuff
> > > like that.
> >
> > Same result. On the version of IRIX I have, IPV6 is available (but no
> > HAVE_GETIFADDRS or HAS_SIOCGLIF). With the above, I have the following
> > patch for IRIX:
> > --- src/util/sys_defs.h.orig 2005-06-21 12:47:04.000000000 -0500
> > +++ src/util/sys_defs.h 2005-06-21 14:22:56.000000000 -0500
> > -596,6 +609,13
> > #if defined(IRIX6)
> > #define HAS_POSIX_REGEXP
> > #define PIPES_CANT_FIONREAD
> > +#ifndef SCM_RIGHTS
> > +# define SCM_RIGHTS 0x01
> > +#endif
> > +#ifndef NO_IPV6
> > +# define HAS_IPV6
> > +#endif
> > +#define CANT_USE_SEND_RECV_MSG
> > #endif
> >
> > /*
> >
> > If I do not define HAS_IPV6, I get into problems with the custom
> > prototype for inet_ntop() in src/util/sys_defs.h.
> >
> > We have run 2.0.19 successfully on this version of IRIX but haven't
> > built anything since then except for the current 2.2.3.
>
> I tried the test you asked for earlier on this platform with the above
> patch:
> $ cd src/util
> $ make unix_recv_fd unix_send_fd
> $ ./unix_recv_fd unix:/tmp/test &
> $ ./unix_send_fd unix:/tmp/test unix_send_fd.c
> unknown: warning: unix_recv_fd: your system has no support for file descriptor passing
> [1] + done ./unix_recv_fd unix:/tmp/test
> unknown: path=unix_send_fd.c fd=4
> unknown: warning: unix_send_fd: your system has no support for file descriptor passing
> unknown: fatal: send file descriptor: Error 0

Ok, I looked at the source and this response is expected. Still trying
to isolate the "premature end-of-input ..." problem reported for this
platform earlier.

--
albert chin (chinathewrittenword.com)

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]