[2377] in testers
ld.so.1 suddenly started losing
daemon@ATHENA.MIT.EDU (Dr. Craig A. Counterman)
Tue Aug 17 13:13:08 1993
To: testers@Athena.MIT.EDU, bug-sparc@Athena.MIT.EDU
Date: Tue, 17 Aug 93 13:12:30 EDT
From: "Dr. Craig A. Counterman" <ccount@Athena.MIT.EDU>
There I was merrily trying to get programs ported and working on the
Sun, and suddenly things started to fail:
m4-167-5% perl
ld.so.1: perl: can't find file libc.so.1.7
Killed
m4-167-5% ci
ld.so.1: ci: can't find file libc.so.1.7
Killed
m4-167-5% co
ld.so.1: co: can't find file libc.so.1.7
Killed
I logged out and immediately back in, and it still lost. The
workstation seems OK now, someone logged in a few mins after I logged
out.
Near as I remember, what I had done near that time:
compiling in one window, doing co's and ci's in another. I messed up
the log message on a ci, so I aborted it with ^C, and it dumped core.
(THIS IS ANOTHER BUG, but relatively small). I couldn't ci the file,
since it was marked as in use. I first copied the version from
OldFiles. Actually, first I tried mv, which complained because it
could't unlink the OldFiles file, then I used cp. Then I realized
that the lock was a hidden file in the RCS directory, deleted that,
and tried to ci again. That's when ld.so started complaining.
I've tried to reproduce that sequence, but things seem to be OK.
Oh, another note, that workstation has a file /misc owned by some
random user, dated the end of July. I noticed because I use filename
completion and typed /mi(complete-char) and lost instead of getting
/mit/ as I usually do (hey, it saves one keystroke).
Thanks and good luck getting these machines ready for prime time.
Craig
