Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!think!rlk
From: rlk@think.COM (Robert Krawitz)
Newsgroups: comp.emacs
Subject: Re: Possible bug in GNU Rmail
Message-ID: <7326@think.UUCP>
Date: Fri, 7-Aug-87 12:09:37 EDT
Article-I.D.: think.7326
Posted: Fri Aug  7 12:09:37 1987
Date-Received: Sun, 9-Aug-87 03:23:55 EDT
Sender: news@think.UUCP
Reply-To: rlk@THINK.COM
Organization: Thinking Machines Corporation, Cambridge, MA
Lines: 48

In article <2195@vdsvax.steinmetz.UUCP> duffd@vdsvax.steinmetz.UUCP (Duff David A) writes:
]When I say "M-x rmail" to read new mail, I sometimes get a message
]that says "File has changed since last visited or saved.  Read from
]disk? (yes or no) ".  

I assume you're using NFS?  This sort of behavior apparently happens
rarely under NFS.  It is more common under Brunhoff RFS, although a
fix to that was posted a while back.  Your description below makes it
sound like something else is to blame.

		       I usually say yes, because I am a little worried
]that I may overwrite and lose some incoming mail if I don't.  No
]matter what I say, though, I usually get another error right away that
]says "Invalid search bound (wrong side of point)" and rmail gets
]screwed up.  

This sounds like the size of the RMAIL buffer changed for some unknown
reason.  I think that what's happening here is that RMAIL wants to
search forward up until a certain point, but point is already past
that (something went to the end of the buffer).  I agree that it's a
less than ideal error message, but it only happens when something
serious has gone wrong.

By the way, you don't have your mailer stick incoming mail directly on
the end of your RMAIL file, do you?  That would account for all these
symptoms.  If you are doing this, then stop RIGHT NOW!!!  This is a
sure way to get clobbered.

	      Instead of displaying only the current message, I see the
]"raw" RMAIL file in my RMAIL buffer with all sorts of control
]characters and header information and the normal rmail commands don't
]work.  

Right.  Rmail is very confused at this point.  Its list of message
pointers is somewhat screwed up.  The error message isn't the best,
but it does alert you that something serious is wrong.

]Of course, this only happens when I invoke the rmail function after
]the RMAIL buffer already exists.  I can get around it by just killing
]the buffer and re-executing the rmail command, but I get a little
]nervous about the possibility of losing incoming mail.
]
]Why would the RMAIL file get changed?  Perhaps GNU somehow gets messed
]up and loses track of the correct last-save-date for the RMAIL buffer.
]I can't really think of anything else that could be happening.  My
]RMAIL file is protected.  Does anyone know what is going on here?

Robert^Z