ftp.delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp-workers/2002/05/12/01:23:08

X-Authentication-Warning: delorie.com: mailnull set sender to djgpp-workers-bounces using -f
Date: Sun, 12 May 2002 08:20:40 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Laurynas Biveinis <lauras AT softhome DOT net>
cc: djgpp-workers AT delorie DOT com
Subject: Re: emacs under w2k
In-Reply-To: <1515097249.20020511203147@softhome.net>
Message-ID: <Pine.SUN.3.91.1020512080610.329D-100000@is>
MIME-Version: 1.0
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com

On Sat, 11 May 2002, Laurynas Biveinis wrote:

> I've rebuilt emacs 20.5 with 2.03 refresh on my w2k machine. It
> mostly works OK, however I'm experiencing unexpected exits time
> from time.

A known problem, I think.  See below.

> I don't see any reasonble pattern there (sometimes it
> exits from 'M-x tetris', sometimes when trying to set global font
> locking etcetera).

If this is the problem I'm aware of, it's related to keyboard 
keypresses.  Try to pay attention to what keys did you press immediately 
before it exits (I won't tell you which ones I think cause that, so that 
if the descriptions I had in the past from other users won't mask some 
new problem, or maybe a more complete description of the problem ;-).

> I've searched djgpp-workers archives to no avail about this one. Is
> it known behaviour? Is there any way I could help to track it down?

As far as known to me, this happens on all descendants of the NT family.  
The problem was first reported (on c.o.m.d.) for NT4.  Richard Dawe 
reported a similar problem a few days ago (in a private email) on W2K.

Unfortunately, my access to NT and W2K is very sporadic, so debugging 
this is not easy for me, especially since these seem to be not crashes, 
but aborts: for some reason, NT simply kills Emacs without any notice.  
What users described was very similar to what happens when you run on NT
a program that catches fatal signals such as SIGINT or SIGFPE: when the 
signal happens, you are kicked back to the shell prompt, as if the 
program silently exited.

Emacs doesn't do anything fancy with the keyboard, it just invokes BIOS 
(Int 16h, functions 10h, 11h, and 12h).  So I don't have any clear idea 
where to start looking for a possible problem.

Help in debugging this is greatly appreciated.  Let me know if you need 
any guidance in your wandering through the maze of Emacs code.

One thing to try is to build Emacs 21.2 on the same machine.  It probably 
won't change anything, but Emacs 20.x had a nasty memory-allocation bug, 
so I'd like to know whether it has any relevance.  (I think it doesn't, 
since it has no relation to keyboard handling.)  At least you will be 
running the latest codebase, which will make it easier to exchange 
patches.

TIA

- Raw text -


  webmaster     delorie software   privacy  
  Copyright 2019   by DJ Delorie     Updated Jul 2019