GNU bug report logs - #10192
Emacs 24.0.92 windows build, fatal error with M-< or M->

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 10192 in the body.
You can then email your comments to 10192 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Da Zhang <zhangda82 <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: Emacs 24.0.92 windows build, fatal error with M-< or M->
Date: Fri, 02 Dec 2011 00:20:21 -0500

[Message part 1 (text/plain, inline)]

Dear Yidong,

I am a dedicated Emacs user who is excited for the new Emacs 24. I have tried the 3 pretests of Emacs 24, and I saw improvements of its stability from each pretest. For example, with my current extensive .emacs configuration file, the first 2 pretests were not able to start normally. However, all the three pretests have the following problems:
1. after the .emacs file is loaded into the pretest Emacs, when I press M-< or M->, the program abort with a fatal error: "A fatal error has occurred! Would you like to attach a debugger? Select Yes to debug, No to abort Emacs..."
2. certain part of eieio is not available, and the emacs prompts with "Symbol's function definition is void: eieio--defgeneric-init-form"

I am not sure if other windows users have experienced the same problem. I attached my .emacs file. Thanks for your attention.

Da Zhang

[Message part 2 (text/html, inline)]

[.emacs (text/plain, attachment)]

Message #8 received at 10192 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Da Zhang <zhangda82 <at> gmail.com>
Cc: 10192 <at> debbugs.gnu.org
Subject: Re: bug#10192: Emacs 24.0.92 windows build,
	fatal error with M-< or M->
Date: Fri, 02 Dec 2011 09:53:38 +0200

> Date: Fri, 02 Dec 2011 00:20:21 -0500
> From: Da Zhang <zhangda82 <at> gmail.com>
> 
> 1. after the .emacs file is loaded into the pretest Emacs, when I press M-< or M->, the program abort with a fatal error: "A fatal error has occurred! Would you like to attach a debugger? Select Yes to debug, No to abort Emacs..."

Does this happen with "emacs -Q"?  If not, there's something in your
customizations that triggers these aborts.  Your .emacs is 3000-line
long, and there are obviously many optional packages that you load
there, so it's not easy to guess which customizations are responsible.

We need a reproducible recipe for causing these aborts, to be able to
find the reason.  Could you please bisect your .emacs and find the
minimal set of customizations that cause these aborts?  Then post only
those parts here.

Thanks.

Message #11 received at 10192 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> IRO.UMontreal.CA>
To: Da Zhang <zhangda82 <at> gmail.com>
Cc: 10192 <at> debbugs.gnu.org
Subject: Re: bug#10192: Emacs 24.0.92 windows build,
	fatal error with M-< or M->
Date: Fri, 02 Dec 2011 09:08:38 -0500

> 2. certain part of eieio is not available, and the emacs prompts with
> "Symbol's function definition is void: eieio--defgeneric-init-form"

eieio--defgeneric-init-form is defined in eieio.el, so maybe you have
some other eieio.el(c) somewhere in your load-path, or maybe you use
a package that does something incorrect like (eval-when-compile
(require 'eieio))?


        Stefan

Message #16 received at 10192-done <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Da Zhang <zhangda82 <at> gmail.com>
Cc: 10192-done <at> debbugs.gnu.org
Subject: Re: bug#10192: Emacs 24.0.92 windows build,
	fatal error with M-< or M->
Date: Sat, 03 Dec 2011 12:05:44 +0200

> Date: Sat, 03 Dec 2011 03:33:19 -0500
> From: Da Zhang <zhangda82 <at> gmail.com>
> 
> The system quit with fatal error after  running this code:
> ;; for smooth scrolling and disabling the automatical recentering of emacs when moving the cursor
> (setq-default scroll-margin 0
>        scroll-conservatively 0
>        scroll-up-aggressively 0.01
>        scroll-down-aggressively 0.01)

This should be fixed now (revision 106588 on the trunk).

If you can rebuild Emacs, the patch is below.  If you cannot rebuild,
set scroll-margin to 1, it should avoid the crashes.  (A better
solution is below, read on.)

By the way, if these customizations were needed to prevent Emacs from
recentering the display, you can now do it like this:

  (setq scroll-conservatively 101)

That is, set this single variable to a value larger than 100.  Then
Emacs will never recenter (and the crashes you had will also
disappear, even if you don't rebuild).  In previous versions of Emacs,
all kinds of weird tricks, such as the strange settings of
scroll-up-aggressively you show above, were necessary to prevent
recentering.  But in Emacs 24, no recentering will ever happen if
scroll-conservatively is set to a large number.

Here's the patch to prevent aborts with your original customizations:

=== modified file 'src/xdisp.c'
--- src/xdisp.c	2011-12-03 07:59:23 +0000
+++ src/xdisp.c	2011-12-03 09:55:27 +0000
@@ -15617,7 +15617,7 @@ redisplay_window (Lisp_Object window, in
 	      pt_offset = float_amount * WINDOW_BOX_TEXT_HEIGHT (w);
 	      if (pt_offset == 0 && float_amount > 0)
 		pt_offset = 1;
-	      if (pt_offset)
+	      if (pt_offset && margin > 0)
 		margin -= 1;
 	    }
 	  /* Compute how much to move the window start backward from

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.