Subject: Re: Another stat13 problem
To: Brian C. Grayson <bgrayson@ece.utexas.edu>
From: Chris G. Demetriou <cgd@pa.dec.com>
List: current-users
Date: 11/07/1997 10:11:11
>   On a related note, we have ~5 NetBSD i386 machines, one of which is
> running -current as our guinea pig, and the others are a few
> months behind.  With the current __stat13 stuff, I can not create
> .o files on the -current box and then link successfully on an
> older box -- /usr/include/sys/stat.h adds an asm reference to
> __stat13, which of course doesn't exist in libc from a few months
> ago.

Why do you think it's reasonable to expect that code compiled on a
"new" system will work on an "old" system?  Certainly, the other way
around should work (for binaries; .o's are always sketchy, because the
headers used by your .o's might not be compatible with those used by
your libraries), but the situation you describe surely should not.

You don't try to run 1.2 binaries on 1.1 systems, or -current
(1.3_ALPHA) binaries on 1.2 systems, do you?  Why do you think you
should be able to use -current object on an old system?



cgd