----------------- cut here ---->8
To: linux-kernel@vger.rutgers.edu
Subject: Linux 1.3.49 thru 53
X-Mailer: mh-e-5.0.2/Emacs-19.30.1/MH-6.8.3 on linux 1.3.53 i486
Date: Wed, 03 Jan 1996 18:00:33 -0500
From: "John A. Martin" <jmartin@cpcug.org>
Somewhere in the early '50s I upgraded my GNU Emacs to 19.30. Later I
found that dired-diff for example failed to show its output and M-x
man showed truncated man pages. After an exchange or two with GNU
Emacs Bug Reporting I received the following suggestion.
>>>>> "Richard" == Richard Stallman "Re: M-x man result chopped"
>>>>> Tue, 2 Jan 1996 23:12:47 -0500
Richard> I suspect that the problem is in process.c; that it is
Richard> losing whatever output is pending when the subprocess
Richard> terminates.
I have not had a chance to dig deeper, but I have checked that the
problem does _not_ occur for me with a 1.3.32 kernel but that the
failure to show the dired-diff output occurs with each of the kernels
from 1.3.49 through 1.3.53 that I have on hand. Also, xman sometimes
truncates a man page. The man command at a shell prompt does not, and
diff from a command line works normally,
The kernels mentioned here were all compiled for i486 elf with gcc
v2.7.0 and libc 5.1.3.
Does this anecdote point to anything in particular?
----------------- cut here ---->8
It seems that the problem occurs with M-x diff, dired-diff, and M-x
man with GNU emacs 19.29 and 19.30 whether compiled by me or obtained
precompiled in packages. It does not occur with linux-1.2.13 (perhaps
that is why RMS could not duplicate it on his GNU/Linux system).
Also, once in a while, one or another of the diff tricks will not fail
on my linux box. Would it be worthwhile to do a "binary search"
between 1.3.32 and 1.3.49? (Ugh..)
I's been a while since I've noticed xman losing output.
>>>>> "Theodore" == Theodore Ts'o "Re: 1.3.72 bad news"
>>>>> Thu, 14 Mar 1996 20:14:57 -0500
Theodore> I can reproduce this problem using emacs 19.28.
Theodore> Unfortunately, I can't duplicate it using any other
Theodore> program. typescript, xterm, all work fine. Even "cat
Theodore> /dev/ptypf" in one window, followed by "grep print *.c >
Theodore> /dev/ttypf" in another works fine.
Theodore> Furthermore, I've looked through the changes between
Theodore> 1.2.13 and 1.3.72 in pty.c and tty_io.c, and I don't see
Theodore> any changes that should make any difference. So, I'm
Theodore> puzzled.
Theodore> If anyone can come up with a (slightly smaller) test
Theodore> case than emacs, I'd really appreciate it. This is
Theodore> definitely a real bug, and it's likely to be in the tty
Theodore> subsystem. Now, if only I can figure out what it is...
Theodore> - Ted
jam