[5237] in SIPB bug reports
Re: Don't build on pickled-herring or other beta machines
daemon@ATHENA.MIT.EDU (Richard Basch)
Wed Mar 22 10:10:04 1995
Date: Wed, 22 Mar 1995 10:05:57 +0500
To: Sam Hartman <hartmans@MIT.EDU>
Cc: sipb-office@MIT.EDU, rs6kswat@MIT.EDU, bug-sipb@MIT.EDU
In-Reply-To: Sam Hartman's message of Wed, 22 Mar 1995 01:15:59 -0500,
<199503220615.BAA14606@tertius.mit.edu>
From: "Richard Basch" <basch@MIT.EDU>
Is this generally true?
I have been building on tardis, and my applications work on other
RISC/6000 machines, for the most part. I can believe that X
applications are susceptible to the failure you describe, but I have
seen no evidence that the general user will run into incompatibilities
with libc.a.
-Richard
Date: Wed, 22 Mar 1995 01:15:59 -0500
From: Sam Hartman <hartmans@MIT.EDU>
I just realized that people might accidentally build software
on machines running 3.2.5. This is a bad thing to do, because libc.a
changed between 3.2.3e and 3.2.5 in a non-backward-compatible way.
I.E. the old code works on both systems, but the new code doesn't work
on the old systems.
After most systems are 3.2.5, this won't be such a big deal,
but right now, any program you build for the RISC/6000 on a beta
machine will only run on a very limited set of systems. The likely
candidates for this problem are pickled-herring in the SIPB office and
washington in the test cluster.
--Sam
