GNU bug report logs - #15247
24.3.50; Emacs Cygwin crashes when replying to a mail

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 15247 in the body.
You can then email your comments to 15247 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: Katsumi Yamaoka <yamaoka <at> jpl.org>
To: bug-gnu-emacs <at> gnu.org
Subject: 24.3.50; Emacs Cygwin crashes when replying to a mail
Date: Mon, 02 Sep 2013 15:02:16 +0900

Hi,

Recently Emacs on Cygwin crashes pretty often when I try to reply
to a mail in a new frame[1] using Gnus.  I feel the frequency gets
high if a mail cites Japanese and Chinese mixture text (so, it's
due to decoding of characters?).  But it hardly crashes if I have
a new empty message frame in advance besides the summary frame.
It happened too also when I used the LUCID toolkit last week.
I don't know what's wrong, sorry, but a gdb backtrace is attached.

Thanks.

[1] I use:
    http://www.jpl.org/ftp/pub/elisp/message-multiple-frames.el.gz
    It pops up a message frame after composing it.

In GNU Emacs 24.3.50.1 (i686-pc-cygwin, GTK+ Version 3.8.2)
 of 2013-09-02 on localhost
Bzr revision: 114096 dmantipov <at> yandex.ru-20130902033906-qqu9pis0ksze2lmg
Windowing system distributor `The Cygwin/X Project', version 11.0.11402000
Configured using:
Configured using:
 `configure --verbose --with-x-toolkit=gtk3 --without-imagemagick
 --without-dbus --without-gconf --without-gsettings'

Important settings:
  value of $LC_CTYPE: ja_JP.UTF-8
  value of $LC_MESSAGES: C
  value of $LANG: C
  locale-coding-system: utf-8-unix
  default enable-multibyte-characters: t

A backtrace follows:

GNU gdb (GDB) 7.6.50.20130728-cvs (cygwin-special)
[...]
Fatal error 6: Aborted
Program received signal SIGABRT, Aborted.
terminate_due_to_signal (sig=2265648, backtrace_limit=164) at emacs.c:372
372       exit (1);
(gdb) bt
#0  terminate_due_to_signal (sig=2265648, backtrace_limit=164) at emacs.c:372
#1  0x3818cf46 in ?? ()
#2  0x00229230 in ?? ()
#3  0x7693543c in WriteFile () from /cygdrive/c/windows/system32/kernel32.dll
#4  0x610dc5a0 in sig_send(_pinfo*, siginfo_t&, _cygtls*)@12 (p=0x60fd0000,
    p <at> entry=0x0, si=..., tls=tls <at> entry=0x0)
    at /usr/src/debug/cygwin-1.7.24-1/winsup/cygwin/sigproc.cc:624
#5  0x610d9c4d in _pinfo::kill(siginfo_t&)@8 (this=0x0, si=...)
    at /usr/src/debug/cygwin-1.7.24-1/winsup/cygwin/signal.cc:248
#6  0x00229274 in ?? ()
Backtrace stopped: previous frame inner to this frame (corrupt stack?)
(gdb)

Message #10 received at 15247-done <at> debbugs.gnu.org (full text, mbox):

From: Katsumi Yamaoka <yamaoka <at> jpl.org>
To: 15247-done <at> debbugs.gnu.org
Subject: Re: bug#15247: 24.3.50; Emacs Cygwin crashes when replying to a mail
Date: Mon, 02 Sep 2013 19:02:37 +0900

Katsumi Yamaoka wrote:
> Recently Emacs on Cygwin crashes pretty often when I try to reply
> to a mail in a new frame[1] using Gnus.  I feel the frequency gets
> high if a mail cites Japanese and Chinese mixture text (so, it's
> due to decoding of characters?).  But it hardly crashes if I have
> a new empty message frame in advance besides the summary frame.

I realized that this is caused by `make-frame-invisible' that runs
when there is only one Emacs frame.  At least for Cygwin, you should
be able to reproduce it in this way:

emacs -Q --eval '(make-frame-invisible (selected-frame) t)'

I also guess that it was harmless until about a week ago.

Why such an odd thing was run is because I was using this advice
to make `raise-frame' work on Cygwin.

(defadvice raise-frame (before make-it-work (&optional frame) activate)
  "Make it work on Cygwin."
  (make-frame-invisible frame t))

As I wrote in <http://thread.gmane.org/gmane.emacs.devel/150533>,
`raise-frame' doesn't raise iconified, invisible, or hidden frame.
Now I changed it into:

(defadvice raise-frame (before make-it-work (&optional frame) activate)
  "Make it work on Cygwin."
  (or (eq frame (selected-frame))
      (make-frame-invisible frame)))

Even if Emacs should not crash for any Lisp code, it's a trivial
one among many(?) issues concerning Cygwin.  So, I'm closing this
bug.  Sorry for the noise.

Message #13 received at 15247 <at> debbugs.gnu.org (full text, mbox):

From: martin rudalics <rudalics <at> gmx.at>
To: 15247 <at> debbugs.gnu.org, yamaoka <at> jpl.org
Subject: Re: bug#15247: 24.3.50; Emacs Cygwin crashes when replying to a mail
Date: Mon, 02 Sep 2013 14:29:51 +0200

> I realized that this is caused by `make-frame-invisible' that runs
> when there is only one Emacs frame.  At least for Cygwin, you should
> be able to reproduce it in this way:
>
> emacs -Q --eval '(make-frame-invisible (selected-frame) t)'
>
> I also guess that it was harmless until about a week ago.

Should be harmless again with revision 114106.

> Why such an odd thing was run is because I was using this advice
> to make `raise-frame' work on Cygwin.
>
> (defadvice raise-frame (before make-it-work (&optional frame) activate)
>   "Make it work on Cygwin."
>   (make-frame-invisible frame t))
>
> As I wrote in <http://thread.gmane.org/gmane.emacs.devel/150533>,
> `raise-frame' doesn't raise iconified, invisible, or hidden frame.
> Now I changed it into:
>
> (defadvice raise-frame (before make-it-work (&optional frame) activate)
>   "Make it work on Cygwin."
>   (or (eq frame (selected-frame))
>       (make-frame-invisible frame)))

I don't understand fully: Above you say that "`raise-frame' doesn't
raise iconified, invisible, or hidden frame" and in the before-advice
or `raise-frame' you make the frame invisible?	

> Even if Emacs should not crash for any Lisp code,

It should not, indeed.

> it's a trivial
> one among many(?) issues concerning Cygwin.  So, I'm closing this
> bug.  Sorry for the noise.

martin

Message #16 received at 15247 <at> debbugs.gnu.org (full text, mbox):

From: Katsumi Yamaoka <yamaoka <at> jpl.org>
To: martin rudalics <rudalics <at> gmx.at>
Cc: 15247 <at> debbugs.gnu.org
Subject: Re: bug#15247: 24.3.50; Emacs Cygwin crashes when replying to a mail
Date: Tue, 03 Sep 2013 08:27:23 +0900

martin rudalics wrote:
>> emacs -Q --eval '(make-frame-invisible (selected-frame) t)'
>> I also guess that it was harmless until about a week ago.

> Should be harmless again with revision 114106.

Thanks!

[...]
>> As I wrote in <http://thread.gmane.org/gmane.emacs.devel/150533>,
>> `raise-frame' doesn't raise iconified, invisible, or hidden frame.
>> Now I changed it into:
>>
>> (defadvice raise-frame (before make-it-work (&optional frame) activate)
>>   "Make it work on Cygwin."
>>   (or (eq frame (selected-frame))
>>       (make-frame-invisible frame)))

> I don't understand fully: Above you say that "`raise-frame' doesn't
> raise iconified, invisible, or hidden frame" and in the before-advice
> or `raise-frame' you make the frame invisible?

Yes, it does.  So far once vanishing a frame, that is alive but
is iconified or hidden behind other frames, from a Windows desktop
is the only means to make `raise-frame' raise the frame to the top.
Another way is:

(defadvice raise-frame (around make-it-work (&optional frame) activate)
;;                      ^^^^^^
  "Make it work on Cygwin."
  (unless (eq frame (selected-frame))
    (make-frame-invisible frame)
    (make-frame-visible frame)))

I found it through trial and trial and trial ... and error. ;)

Previous Next

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