GNU bug report logs - #3021
23.0.92; MEM FULL on Windows after some VC operations

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 3021 in the body.
You can then email your comments to 3021 AT debbugs.gnu.org in the normal way.

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

Message #5 received at submit <at> emacsbugs.donarmstrong.com (full text, mbox):

From: Reiner Steib <reinersteib+gmane <at> imap.cc>
To: emacs-pretest-bug <at> gnu.org
Subject: 23.0.92; MEM FULL on Windows after some VC operations
Date: Thu, 16 Apr 2009 21:29:27 +0200

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

Hi,

with the latest precompiled pretest version for Windows
(http://alpha.gnu.org/gnu/emacs/pretest/windows/emacs-23.0.92-bin-i386.zip),
I got "!MEM FULL!" after doing some VC / PCL-CVS commands (viewing the
revision log, diffs, ...) on rather big files.  The repository is on a
network share, i.e. no pserver, ssh, ... involved.

As requested by Eli, I attach the information reported by
Sysinternals' VMMap, both for the Emacs 23 process with "MEM FULL" and
an Emacs 22.3 process on the same box.

The laptop has 3 GB of memory.

In GNU Emacs 23.0.92.1 (i386-mingw-nt6.0.6001)
 of 2009-03-31 on SOFT-MJASON
Windowing system distributor `Microsoft Corp.', version 6.0.6001
configured using `configure --with-gcc (3.4)'

Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: nil
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: DEU
  value of $XMODIFIERS: nil
  locale-coding-system: cp1252
  default-enable-multibyte-characters: t

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

[emacs-23-0-92-mem-full.txt (text/plain, attachment)]

[emacs-22-3-okay.txt (text/plain, attachment)]

Message #10 received at 3021 <at> emacsbugs.donarmstrong.com (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Reiner Steib <Reiner.Steib <at> gmx.de>, 3021 <at> debbugs.gnu.org
Subject: Re: bug#3021: 23.0.92; MEM FULL on Windows after some VC operations
Date: Fri, 17 Apr 2009 00:35:16 +0300

> From: Reiner Steib <reinersteib+gmane <at> imap.cc>
> Date: Thu, 16 Apr 2009 21:29:27 +0200
> Cc: 
> 
> with the latest precompiled pretest version for Windows
> (http://alpha.gnu.org/gnu/emacs/pretest/windows/emacs-23.0.92-bin-i386.zip),
> I got "!MEM FULL!" after doing some VC / PCL-CVS commands (viewing the
> revision log, diffs, ...) on rather big files.  The repository is on a
> network share, i.e. no pserver, ssh, ... involved.
> 
> As requested by Eli, I attach the information reported by
> Sysinternals' VMMap, both for the Emacs 23 process with "MEM FULL" and
> an Emacs 22.3 process on the same box.

Thanks.

> The laptop has 3 GB of memory.
> [...]
> Process: emacs.exe
> PID: 7068
> 
> Type        Size        Committed  Total WS  Private WS  Shareable WS  Shared WS  Blocks  
> Total       400.344     363.360    291.112   270.448     20.664        17.636

Unless I'm missing something, looks like something might be wrong with
how we discover available memory on Windows: how come a process with a
400MB memory footprint says MEM FULL on a 3GB machine?

FWIW, I had a couple of days ago a similar incident: a 380MB Emacs
process running on a 1.5GB machine claimed to have passed 75% of
available memory.

Message #17 received at 3021 <at> emacsbugs.donarmstrong.com (full text, mbox):

From: Reiner Steib <reinersteib+gmane <at> imap.cc>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 3021 <at> debbugs.gnu.org
Subject: Re: bug#3021: 23.0.92; MEM FULL on Windows after some VC operations
Date: Mon, 20 Apr 2009 22:07:46 +0200

On Thu, Apr 16 2009, Eli Zaretskii wrote:

> Unless I'm missing something, looks like something might be wrong with
> how we discover available memory on Windows: how come a process with a
> 400MB memory footprint says MEM FULL on a 3GB machine?

Ideas anyone?

Can I help with more information?

> FWIW, I had a couple of days ago a similar incident: a 380MB Emacs
> process running on a 1.5GB machine claimed to have passed 75% of
> available memory.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Message #22 received at 3021 <at> emacsbugs.donarmstrong.com (full text, mbox):

From: Jason Rumney <jasonr <at> gnu.org>
To: Reiner.Steib <at> gmx.de
Cc: 3021 <at> debbugs.gnu.org
Subject: Re: bug#3021: 23.0.92; MEM FULL on Windows after some VC operations
Date: Tue, 21 Apr 2009 12:23:16 +0800

Reiner Steib wrote:
> On Thu, Apr 16 2009, Eli Zaretskii wrote:
>
>   
>> Unless I'm missing something, looks like something might be wrong with
>> how we discover available memory on Windows: how come a process with a
>> 400MB memory footprint says MEM FULL on a 3GB machine?
>>     
>
> Ideas anyone?
>
> Can I help with more information?
>   

Can you find the location where that message is emitted, and put a 
breakpoint there. When the bug next happens, have a look at the tests 
leading up to the decision to display the message and see if you can see 
anything amiss.

Message #29 received at 3021 <at> emacsbugs.donarmstrong.com (full text, mbox):

From: Reiner Steib <reinersteib+gmane <at> imap.cc>
To: 3021 <at> debbugs.gnu.org
Cc: Jason Rumney <jasonr <at> gnu.org>, Eli Zaretskii <eliz <at> gnu.org>
Subject: Re: bug#3021: 23.0.92; MEM FULL on Windows after some VC operations
Date: Sun, 05 Jul 2009 12:31:45 +0200

On Tue, Apr 21 2009, Jason Rumney wrote:

> Reiner Steib wrote:
>> On Thu, Apr 16 2009, Eli Zaretskii wrote:
>>> Unless I'm missing something, looks like something might be wrong with
>>> how we discover available memory on Windows: how come a process with a
>>> 400MB memory footprint says MEM FULL on a 3GB machine?
>>
>> Ideas anyone?
>>
>> Can I help with more information?
>
> Can you find the location where that message is emitted, and put a breakpoint
> there. When the bug next happens, have a look at the tests leading up to the
> decision to display the message and see if you can see anything amiss.

Sorry for the late reply, I have been on vacation for a few weeks and
busy at work otherwise.

On Thu, May 21 2009, Chong Yidong wrote:
> If you can reproduce this bug, could you try to investigate?  A
> backtrace would be useful.

I can (or better say could with 23.0.92 - haven't tried newer
pretests) reproduce the problem.  I'll try the most recent pretest
next week.

But at work (where the windows box lives) I don't have a debugger
installed and I don't now what I need to install/download to get a
working debugging environment for Emacs on Windows Vista.  So it will
take me quite a lot of time to figure out what is required and to set
it up.  In etc/DEBUG I can only see instructions for Microsoft Visual
C++, or am I missing something?

- What do I need to install to debug a MinGW/MSYS build (Jason's
  build) on MS Vista?

- To put a breakpoint, of course I need to install the Emacs sources.
  How do I have to combine the source (emacs-*.tar.gz) and the binary
  (windows/emacs-*-bin-i386.zip) correctly?

- Jason suggested: "find the location where that message is emitted,
  and put a breakpoint there"...

  Can someone assist me with this?  I'm not familiar with Emacs' C
  code. From rgrep on 'MEM.FULL\|memory_full' it's not obvious for me
  which would be the right function.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Message #34 received at 3021 <at> emacsbugs.donarmstrong.com (full text, mbox):

From: Jason Rumney <jasonr <at> gnu.org>
To: 3021 <at> debbugs.gnu.org
Subject: Re: bug#3021: 23.0.92; MEM FULL on Windows after some VC operations
Date: Mon, 06 Jul 2009 00:02:07 +0800

Reiner Steib wrote:

> - What do I need to install to debug a MinGW/MSYS build (Jason's
>   build) on MS Vista?
>   
gdb (from www.mingw.org)

> - To put a breakpoint, of course I need to install the Emacs sources.
>   How do I have to combine the source (emacs-*.tar.gz) and the binary
>   (windows/emacs-*-bin-i386.zip) correctly?
>   

Unzipping into the same directory should work - any files contained in 
both are identical, so no harm will come from replacing, or conversely 
from keeping the old ones.

> - Jason suggested: "find the location where that message is emitted,
>   and put a breakpoint there"...
>
>   Can someone assist me with this?  I'm not familiar with Emacs' C
>   code. From rgrep on 'MEM.FULL\|memory_full' it's not obvious for me
>   which would be the right function.
>   
So put a breakpoint on all locations that rgrep shows that message occuring.

To run gdb, follow the directions in etc/DEBUG and nt/INSTALL, notably:

Start Emacs from the src directory as: gdb ../bin/emacs.exe
This helps gdb find the source files, and also it will automatically 
include the extra functions in src/.gdbinit for inspecting lisp 
structures and other Emacs specific things.

To set a break point in gdb, type one of the following:

break functionname
break filename:linenum

to run emacs until a breakpoint is hit, type

run

this is the point to pass arguments to Emacs (not when starting gdb, or 
they will be interpreted as gdb arguments). So to run without init files

run -Q



You can also run gdb from inside Emacs, in which case setting 
breakpoints is easier - go to the line where you want the breakpoint in 
the source file's buffer, and press C-x SPC.  But I have found gdb to be 
unstable within Emacs under Windows, YMMV.

Message #39 received at 3021 <at> emacsbugs.donarmstrong.com (full text, mbox):

From: Reiner Steib <reinersteib+gmane <at> imap.cc>
To: 3021 <at> debbugs.gnu.org
Subject: Re: bug#3021: 23.0.92; MEM FULL on Windows after some VC operations
Date: Sat, 11 Jul 2009 13:52:11 +0200

On Sun, Jul 05 2009, Reiner Steib wrote:

> I'll try the most recent pretest next week.

During some quick tests with 23.0.95, I couldn't reproduce the
problem yet.  I'll try again with 23.0.96.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Message #44 received at 3021 <at> emacsbugs.donarmstrong.com (full text, mbox):

From: Reiner Steib <reinersteib+gmane <at> imap.cc>
To: Jason Rumney <jasonr <at> gnu.org>
Cc: 3021 <at> debbugs.gnu.org
Subject: Re: bug#3021: 23.0.92; MEM FULL on Windows after some VC operations
Date: Sat, 11 Jul 2009 13:57:39 +0200

On Sun, Jul 05 2009, Jason Rumney wrote:

> Reiner Steib wrote:
>> - What do I need to install to debug a MinGW/MSYS build (Jason's
>>   build) on MS Vista?
>>   
> gdb (from www.mingw.org)
>
>> - To put a breakpoint, of course I need to install the Emacs sources.
>>   How do I have to combine the source (emacs-*.tar.gz) and the binary
>>   (windows/emacs-*-bin-i386.zip) correctly?
>
> Unzipping into the same directory should work - any files contained in both
> are identical, so no harm will come from replacing, or conversely from keeping
> the old ones.

Thanks, that worked.  There was some warning/error from gdb (sorry, I
didn't write it down) when starting it (with a "continue y/n" prompt),
but I could run emacs.exe from gdb.

These instructions should be documented in some readme/debug files, I
think.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Message #49 received at 3021 <at> emacsbugs.donarmstrong.com (full text, mbox):

From: Reiner Steib <reinersteib+gmane <at> imap.cc>
To: 3021 <at> debbugs.gnu.org
Subject: Re: bug#3021: 23.0.92; MEM FULL on Windows after some VC operations
Date: Wed, 15 Jul 2009 20:49:58 +0200

[ The following message is a courtesy copy of an article that has
  been posted to news:gmane.emacs.bugs as well. ]

close 3021
quit

On Sat, Jul 11 2009, Reiner Steib wrote:
> On Sun, Jul 05 2009, Reiner Steib wrote:
>> I'll try the most recent pretest next week.
> During some quick tests with 23.0.95, I couldn't reproduce the
> problem yet.  I'll try again with 23.0.96.

I did a lot of VC operations with Emacs 23.0.95 today and there was no
MEM FULL message.

So I think we can close this bug. -> Done.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Previous Next

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