GNU bug report logs - #17749
24.3.91; Crash on mwheel-scroll of org file

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 17749 in the body.
You can then email your comments to 17749 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: "Vladimir Alexiev" <vladimir.alexiev <at> ontotext.com>
To: <bug-gnu-emacs <at> gnu.org>,
	<emacs-orgmode <at> gnu.org>
Subject: 24.3.91; Crash on mwheel-scroll of org file
Date: Tue, 10 Jun 2014 19:02:30 +0300

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

Hi! 

I think I found a new crash, at least I can't find it with google.

Scenario:
1. Open emacs-crash-test.org
2. Make the window smaller (e.g. 20 lines tall)
2. Roll the mouse wheel down until you see the end of file, then up. 
3. CRASH

I cannot reproduce it with "emacs -Q" yet.
I cannot always reproduce it with emacs-crash-test.org. I made this by replacing all letter with "a" in an original file that I can't share.
Please let me know whether this is a new crash and you are willing to debug it.
Then I can dig deeper to find which exact situation, or my code, or ELPA package causes this.

I saved emacs_backtrace.txt, then did
$ cat emacs_backtrace.txt | addr2line -C -f -i -p -e c:/emacs/bin/emacs > emacs_addr.txt
Both files are attached.
I also attach the info from report-emacs-bug.

Do I still need to do gdb "bt full" or "xbacktrace", or are the attachments enough?
Sorry, I can't recompile or debug emacs.

$ emacs &
[1] 5344

../../emacs-24.3.91/src/xdisp.c:6066: Emacs fatal error: assertion failed: (BUFFERP (it->object
) && IT_CHARPOS (*it) == it->bidi_it.charpos && IT_BYTEPOS (*it) == it->bidi_it.bytepos) || (ST
RINGP (it->object) && IT_STRING_CHARPOS (*it) == it->bidi_it.charpos && IT_STRING_BYTEPOS (*it)
 == it->bidi_it.bytepos) || (CONSP (it->object) && it->method == GET_FROM_STRETCH)

$ gdb -5 5344
Attaching to process 5344
[New Thread 5344.0x1148]
[New Thread 5344.0x214c]
[New Thread 5344.0x1574]
Reading symbols from /cygdrive/c/emacs/bin/emacs...done.

warning: `/cygdrive/c/Windows/SYSTEM32/ntdll.dll': Shared library architecture i386:x86-64 is n
ot compatible with target architecture i386.
(and several more warnings about Windows system libraries)

(gdb) continue
Continuing.
[Thread 5344.0x1574 exited with code 0]
[Inferior 1 (process 5344) exited with code 01400]
(gdb)
The program is not being run.

[emacs-crash-test.org (application/octet-stream, attachment)]

[emacs_addr.txt (text/plain, attachment)]

[emacs_backtrace.txt (text/plain, attachment)]

[emacs-report-info.txt (text/plain, attachment)]

Message #8 received at 17749 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: vladimir.alexiev <at> ontotext.com
Cc: 17749 <at> debbugs.gnu.org, emacs-orgmode <at> gnu.org
Subject: Re: bug#17749: 24.3.91; Crash on mwheel-scroll of org file
Date: Tue, 10 Jun 2014 20:54:25 +0300

> From: "Vladimir Alexiev" <vladimir.alexiev <at> ontotext.com>
> Date: Tue, 10 Jun 2014 19:02:30 +0300
> 
> 1. Open emacs-crash-test.org
> 2. Make the window smaller (e.g. 20 lines tall)
> 2. Roll the mouse wheel down until you see the end of file, then up. 
> 3. CRASH
> 
> I cannot reproduce it with "emacs -Q" yet.

I cannot reproduce this with the Org file you've sent, if I perform
the steps above starting from "emacs -Q".

> I cannot always reproduce it with emacs-crash-test.org. I made this by replacing all letter with "a" in an original file that I can't share.
> Please let me know whether this is a new crash and you are willing to debug it.
> Then I can dig deeper to find which exact situation, or my code, or ELPA package causes this.

Please do, I need a reproducible recipe starting from "emacs -Q" in
order to debug this.

> ../../emacs-24.3.91/src/xdisp.c:6066: Emacs fatal error: assertion failed: (BUFFERP (it->object) && IT_CHARPOS (*it) == it->bidi_it.charpos && IT_BYTEPOS (*it) == it->bidi_it.bytepos) || (STRINGP (it->object) && IT_STRING_CHARPOS (*it) == it->bidi_it.charpos && IT_STRING_BYTEPOS (*it) == it->bidi_it.bytepos) || (CONSP (it->object) && it->method == GET_FROM_STRETCH)

This is an assertion violation, not a crash.  Some assumption in the
code turned out to be false, but it is not possible to know which one,
and whether it's just a wrong assumption or a real problem.

So a reproducible recipe is required, I hope you can find it.

Thanks.

Message #11 received at 17749 <at> debbugs.gnu.org (full text, mbox):

From: "Vladimir Alexiev" <vladimir.alexiev <at> ontotext.com>
To: "'Eli Zaretskii'" <eliz <at> gnu.org>,
	<gerd <at> gnu.org>
Cc: 17749 <at> debbugs.gnu.org, emacs-orgmode <at> gnu.org
Subject: RE: bug#17749: 24.3.91; Crash on mwheel-scroll of org file
Date: Tue, 5 Aug 2014 11:10:16 +0300

> > 1. Open emacs-crash-test.org
> > 2. Make the window smaller (e.g. 20 lines tall)
> > 2. Roll the mouse wheel down until you see the end of file, then up.
> > 3. CRASH

The specific org file doesn't matter.
The window size doesn't matter.
This crash happens to me occasionally (2-3 times a week), always on an org file, always on mwheel-up (which invokes mwheel-scroll).

Eli, can you tell something from the backtrace here?
http://lists.gnu.org/archive/html/emacs-orgmode/2014-06/txtkIBvnmHi_F.txt

The critical lines are:

die at C:\msys\home\Dani\emacs\build\src/../../emacs-24.3.91/src/alloc.c:6826
pop_it at C:\msys\home\Dani\emacs\build\src/../../emacs-24.3.91/src/xdisp.c:6060
set_iterator_to_next at C:\msys\home\Dani\emacs\build\src/../../emacs-24.3.91/src/xdisp.c:7523
move_it_in_display_line_to at C:\msys\home\Dani\emacs\build\src/../../emacs-24.3.91/src/xdisp.c:8838

> > ../../emacs-24.3.91/src/xdisp.c:6066: Emacs fatal error: assertion failed: (BUFFERP (it->object) && IT_CHARPOS (*it) == it-
> >bidi_it.charpos && IT_BYTEPOS (*it) == it->bidi_it.bytepos) || (STRINGP (it->object) && IT_STRING_CHARPOS (*it) == it-
> >bidi_it.charpos && IT_STRING_BYTEPOS (*it) == it->bidi_it.bytepos) || (CONSP (it->object) && it->method ==
> GET_FROM_STRETCH)

The code can be seen here:
https://github.com/mirrors/emacs/blob/master/src/xdisp.c#L6127 

> This is an assertion violation, not a crash.

The failing assertion causes a crash, so I'm not sure what's the difference. 
Is there an emacs build that doesn't crash on eassert() failure?

> Some assumption in the
> code turned out to be false, but it is not possible to know which one,
> and whether it's just a wrong assumption or a real problem.

It seems the assertion does basic consistency checks of the IT structure.

What is it->bidi_p? I don't use BIDI text: this happens on ascii org files.


I guess I'll upgrade Emacs and Org and hope for the best...

Message #14 received at 17749 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: vladimir.alexiev <at> ontotext.com
Cc: 17749 <at> debbugs.gnu.org, emacs-orgmode <at> gnu.org
Subject: Re: bug#17749: 24.3.91; Crash on mwheel-scroll of org file
Date: Tue, 05 Aug 2014 17:21:12 +0300

> From: "Vladimir Alexiev" <vladimir.alexiev <at> ontotext.com>
> Cc: <17749 <at> debbugs.gnu.org>,
> 	<emacs-orgmode <at> gnu.org>
> Date: Tue, 5 Aug 2014 11:10:16 +0300
> 
> > > 1. Open emacs-crash-test.org
> > > 2. Make the window smaller (e.g. 20 lines tall)
> > > 2. Roll the mouse wheel down until you see the end of file, then up.
> > > 3. CRASH
> 
> The specific org file doesn't matter.
> The window size doesn't matter.
> This crash happens to me occasionally (2-3 times a week), always on an org file, always on mwheel-up (which invokes mwheel-scroll).
> 
> Eli, can you tell something from the backtrace here?
> http://lists.gnu.org/archive/html/emacs-orgmode/2014-06/txtkIBvnmHi_F.txt

No, sorry.  I really need to reproduce this on my machine and run
Emacs under a debugger to see what happens and why.

Please try to come up with a recipe starting from "emacs -Q".  It is
OK to include customizations and loading of optional packages, but
please try to find the minimal set of those, as too many of them
complicate debugging.

> > This is an assertion violation, not a crash.
> 
> The failing assertion causes a crash, so I'm not sure what's the difference. 
> Is there an emacs build that doesn't crash on eassert() failure?

If Emacs is configured without --enable-checking (which is what
happens in released versions), all the eassert lines compile to
nothing.

> > Some assumption in the
> > code turned out to be false, but it is not possible to know which one,
> > and whether it's just a wrong assumption or a real problem.
> 
> It seems the assertion does basic consistency checks of the IT structure.

Yes.  But without details, I cannot know what caused it or how to fix
it.

> What is it->bidi_p? I don't use BIDI text: this happens on ascii org files.

The display engine always works in bidirectional mode, this has been
so since Emacs 24.1.

Message #17 received at 17749 <at> debbugs.gnu.org (full text, mbox):

From: Lars Ingebrigtsen <larsi <at> gnus.org>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 17749 <at> debbugs.gnu.org, emacs-orgmode <at> gnu.org, vladimir.alexiev <at> ontotext.com
Subject: Re: bug#17749: 24.3.91; Crash on mwheel-scroll of org file
Date: Sat, 26 Dec 2015 16:26:11 +0100

Eli Zaretskii <eliz <at> gnu.org> writes:

> No, sorry.  I really need to reproduce this on my machine and run
> Emacs under a debugger to see what happens and why.
>
> Please try to come up with a recipe starting from "emacs -Q".  It is
> OK to include customizations and loading of optional packages, but
> please try to find the minimal set of those, as too many of them
> complicate debugging.

More information was requested, but no response was given within a few
months, so I'm closing this bug report.  If the problem still exists,
please reopen this bug report.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Previous Next

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