help debugging the Windows hangs

607 views
Skip to first unread message

Peter Bienstman

unread,
Jul 9, 2012, 2:48:42 PM7/9/12
to mnemosyne-...@googlegroups.com
Hi,

As you know, there have been quite some reports of Mnemosyne sometimes
hanging on Windows. I had hoped that the new PyQt version would fix
this, but unfortunately, that turned out not be the case. The hangs
never happen on my system, so I need help to debug this.

To summarise what I understand the problem to be (but feel free to
correct me), Mnemosyne sometimes hangs when trying to grade a card,
and this happens regardsless of whether the cards contain images or
not. It happens both on XP 32 bit and on W7 64 bit.

I prepared a new test version here:

http://users.ugent.be/~pbienst/pub/mnemosyne-20120709-setup.exe

I changed some small packaging/configuration options. I'm not sure if
they will have a big impact, but you never know :-)

If you still have hangs, it would be interesting to see if how the
frequency of hangs changes if you try e.g. 'autosave after 1 reps' vs
'autosave after 20 reps'.

If you have a virus scanner like Microsoft security essentials
running, please try disabling it and see if it makes a difference.
Ditto for Windows desktop search.

If you have a separate version of Python installed on your machine,
please remove it.

Do a search for default.db on your file system, and check if
default.db has write permissions, together with all the folder levels
above it.

Do you have funny characters in your windows account name?

Is your database on the C drive?

Do you also have the problem if you never used the card browser after
starting the program?

If you're technically inclined, run Sysinternal process monitor while
running Mnemosyne, to see if anything suspect happens with database
access (filter to show only events in the 'file system' class) (
http://technet.microsoft.com/en-us/sysinternals/bb896645 )

Make sure you have enough free disk space and try running a defragmenter.

Any other pattern you observe or thought you might have to solve this,
is welcome!

Peter

Chris

unread,
Jul 9, 2012, 4:43:51 PM7/9/12
to mnemosyne-...@googlegroups.com
COPYING my comment from other, less applicable thread, regarding frequency of crashes.  I had a weak suspicion about the autosave before I read your post, too:

It's a strange frequency. For one day, it would crash, I would restart, it would crash in two cards, and it did that repeatedly, and I gave up frustrated.  The next day I was able to get through about 100 cards before it crashed, so I thought I'd be patient and just try and get through as many cards as I could each day.  Today it let me get through about 15 before the freeze.  One thing which seems consistent, though, is that it seem to crash quicker and quicker if I try and restart it repeatedly?  If I give it a day, it seems to be in a brighter mood for a while. This makes no sense to me based on the little I know about computers, so I suppose it could be a coincidence of some sort.  The one other thing I was wondering about is if it has something to do with the autosave, because I put the autosave at "every 1 card" in order that I wouldn't lose information each time it crashed, but there have been a few times where I'll still be a few, even maybe 10 cards behind where I was when I open the program again.  No idea if this helps?

   Thanks for your quick help; I'll be of any help I can - I would be glad to run any experiments you'd like me to try on my machine, if that could in any way help,  Chris S.

Chris

unread,
Jul 9, 2012, 4:52:50 PM7/9/12
to mnemosyne-...@googlegroups.com
REGARDING your questions here:

I am running Windows 7 Starter, which I believe is 32-bit, but I may be mistaken.
No anti-virus running.
Don't really know what python is, so probably don't have another version of it running, hopefully!
Also not sure how to find default.db, doesn't come up on a file search of My Computer.
No funny characters in my username ("Chris")
Everything is on C drive.
Have had the problem without using the card browser, repeatedly.
I'll try the defragment this afternoon, and read up on the process monitor to see if I could figure it out without too much risk!  :)

   Good luck, Chris S.

Peter Bienstman

unread,
Jul 10, 2012, 1:06:30 AM7/10/12
to mnemosyne-...@googlegroups.com
Thanks for the info. More importantly, did you try the new version?

Peter

Chris <the.whis...@gmail.com> wrote:
--
You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group.
To view this discussion on the web visit https://groups.google.com/d/msg/mnemosyne-proj-users/-/nMK7uOamiloJ.
To post to this group, send email to mnemosyne-...@googlegroups.com.
To unsubscribe from this group, send email to mnemosyne-proj-u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/mnemosyne-proj-users?hl=en.

Chris

unread,
Jul 10, 2012, 1:48:35 AM7/10/12
to mnemosyne-...@googlegroups.com
Yes, just did.  Let me finish the 16 remaining scheduled cards for the day, then froze after about 5 unlearned cards.  The program will hang for me both on the regularly scheduled cards, and in new, unlearned cards, which I neglected to mention before (I mention it because I noticed another poster said they have it on learning new cards).  I did the defragging, didn't seem to have an effect (my computer was only at 1% fragmentation to start).  The sysinternal trick seems a bit too advanced for me. 

I was thinking, I could load the program on another computer that I only rarely use, mount the database in that copy of th program, and see if it occurs there; which may let us know if it's something to do with the database, or the computer?  This may or may not be of help, I really don't know.  That other computer is a Windows 7 64-bit desktop, this one is a Windows 7 Starter netbook.  Let me know if it is worthwhile.  I may just do it temporarily so I can keep up on my card repetitions, and then start using my netbook again once the problem is fixed.  I prefer not to have to use the desktop comp for review though. 

   Thanks again, Chris Shanks.
Chris <the.whis...@gmail.com> wrote:
To post to this group, send email to mnemosyne-proj-users@googlegroups.com.
To unsubscribe from this group, send email to mnemosyne-proj-users+unsub...@googlegroups.com.

Peter Bienstman

unread,
Jul 10, 2012, 2:48:25 AM7/10/12
to mnemosyne-...@googlegroups.com
On Monday, July 09, 2012 10:48:35 PM Chris wrote:
> Yes, just did. Let me finish the 16 remaining scheduled cards for the day,
> then froze after about 5 unlearned cards.

Too bad...

We are talking about this version, right?

http://users.ugent.be/~pbienst/pub/mnemosyne-20120709-setup.exe

> I was thinking, I could load the program on another computer that I only
> rarely use, mount the database in that copy of th program, and see if it
> occurs there; which may let us know if it's something to do with the
> database, or the computer?

That could indeed be interesting...

Peter

Peter Bienstman

unread,
Jul 10, 2012, 3:22:33 PM7/10/12
to mnemosyne-...@googlegroups.com
Yet another new version to try out:

http://users.ugent.be/~pbienst/pub/mnemosyne-2.0.1-test1-setup.exe

This one also always has a 50/50 split for question and answer window, to eliminate all possible causes of hang.

Peter

Scott Youngman

unread,
Jul 11, 2012, 1:21:19 PM7/11/12
to mnemosyne-...@googlegroups.com
My observations:
1. Hangs with either known or new cards.
- Sometimes I can go through all known cards without a hang (usually ~30-40 cards), then it will hang repeatedly in the new cards.
- Sometimes it will hang on the second card of the known cards.
2. Almost always the hang occurs when grading a card, but occasionally it can hang after grading when attempting to advance to the next card.
3. Hangs even without images on the cards.
4. Hangs even if the card browser has not been used in the session.

Variables changed without fixing the problem:
1. Autosave after 1 rep or 20 reps.
2. Real-time monitoring by Microsoft Security Essentials on or off.
3. Excluded mnemosyne.exe as a process in Microsoft Security Essentials.

Other factors:
1. No other version of Python.
2. No Win desktop search.
3. No special characters in username.
4. Default.db is on my personal data partition, and has write permissions.
5. Ran Process Monitor, but I'm not sure what to look for, and may not have set it correctly.
6. Plenty of disk space is available.
7. Low fragmentation level.

Scott Youngman

unread,
Jul 11, 2012, 1:33:22 PM7/11/12
to mnemosyne-...@googlegroups.com
Further observations:
1. In Process Monitor, everything with Mnemosyne is marked Successful.
2. default.db-journal gets written about the time Mnemosyne hangs.

Peter Bienstman

unread,
Jul 11, 2012, 2:16:06 PM7/11/12
to mnemosyne-...@googlegroups.com
On Wednesday 11 Jul 2012 10:21:19 Scott Youngman wrote:

> 2. Almost always the hang occurs when grading a card, but occasionally it
> can hang after grading when attempting to advance to the next card.

What difference do you mean between these two? Normally, after grading, you 
should advance automatically to the next card.

> Variables changed without fixing the problem:
> 1. Autosave after 1 rep or 20 reps.

Does it hang more/sooner with autosave at 1? That should confirm the hypothesis 
that it's an issue in the database layer.

> Further observations:
> 1. In Process Monitor, everything with Mnemosyne is marked Successful.

What you should be looking for is other processes trying to access default.db 
or default.db-journal. There is a filter option in the process monitor to only 
show the file access operations, so that you don't get overwhelmed by the 
registry access operations.

> 2. default.db-journal gets written about the time Mnemosyne hangs.

Interesting!

BTW, in 2.0.1-test1, default.db-journal should never get deleted, even after 
exiting the program. Can you confirm that this is indeed the case, after you 
start 2.0.1-test1 for a second time?

Thanks for all the info!

Peter

schwarz.c...@gmail.com

unread,
Jul 11, 2012, 2:23:19 PM7/11/12
to mnemosyne-...@googlegroups.com
Hi Peter, I still get the same ol' hang on this too.
http://users.ugent.be/~pbienst/pub/mnemosyne-2.0.1-test1-setup.exe

There's no notable change from autosave on 1 to autosave on 25. It crashed within 5 cards each time.

I do run MSE. Turning it on/off made no difference. I don't run Windows Desktop Search.

I don't have an extra version of python.

My account name is "Chris"

My database is in the default location on the C drive.

I reproduce the problem every time without starting the card browser.

I have enough disk space. I don't have time to run a defragmenter for this, sorry.

Windows 7, x64.

In process explorer, when I go to the next card, the series of writes does end with default.db-journal, but this is true whether or not it just crashed. The activity for a crash case and an okay case looks the same.

Although after trying it a few times, sometimes it ends in a call to the webkit dll:
11:18:29.5116801 AM mnemosyne.exe 14052 Thread Create SUCCESS Thread ID: 4268
11:19:06.5175272 AM mnemosyne.exe 14052 Thread Exit SUCCESS Thread ID: 11428, User Time: 0.0000000, Kernel Time: 0.0000000
11:19:22.9357077 AM mnemosyne.exe 14052 ReadFile C:\Program Files (x86)\Mnemosyne\QtWebKit4.dll SUCCESS Offset: 2,565,120, Length: 32,768, I/O Flags: Non-cached, Paging I/O, Synchronous Paging I/O, Priority: Normal
11:19:29.5332648 AM mnemosyne.exe 14052 Thread Exit SUCCESS Thread ID: 9788, User Time: 0.0000000, Kernel Time: 0.0000000

But most of the time it ends in the db-journal, just like it does in the working case.

Sorry none of this seems terribly helpful but I hope it is anyway!

schwarz.c...@gmail.com

unread,
Jul 11, 2012, 2:28:58 PM7/11/12
to mnemosyne-...@googlegroups.com
The db-journal file does remain after a proper exit.

Scott Youngman

unread,
Jul 11, 2012, 5:11:58 PM7/11/12
to mnemosyne-...@googlegroups.com
Sorry, my description was wrong on point 2. (see your question quoted below). What I should have said is:

2. Almost always the hang occurs when attempting to show the answer; the Q is displayed, but the A never appears.
Occasionally it can hang after grading a card; the A is displayed, but it does not advance to the next card.

Scott Youngman

unread,
Jul 11, 2012, 6:00:28 PM7/11/12
to mnemosyne-...@googlegroups.com
In answer to your question quoted below...

With 2.0.1-test1, I can't even show the answer to the first card, whether autosave = 1 or 20.

So I have to go back to build 20120709. With that build and in my testing today, yes, I was able to get further (more cards before a hang) when Autosave = 20 than = 1. For Autosave = 1, I could go between 2 and ~10 cards before hanging. With Autosave = 20, I went 45 cards before hanging, then after restarting I went ~60 cards without a hang until I ran out of cards for the day.

On Wednesday, July 11, 2012 12:16:06 PM UTC-6, Peter Bienstman wrote:

Scott Youngman

unread,
Jul 11, 2012, 6:06:11 PM7/11/12
to mnemosyne-...@googlegroups.com
In reply to your comment quoted below...

In 20120709, default.db-journal remains after a proper shutdown. (As stated above, I can't use 2.0.1-test1 because it won't even progress to the A of the first card.)

On Wednesday, July 11, 2012 12:16:06 PM UTC-6, Peter Bienstman wrote:

Scott Youngman

unread,
Jul 11, 2012, 6:25:39 PM7/11/12
to mnemosyne-...@googlegroups.com
Not so fast...
I just reopened 20120709. Autosave = 20, and it hung after 2 cards :^(

On Wednesday, July 11, 2012 12:16:06 PM UTC-6, Peter Bienstman wrote:

Chris

unread,
Jul 12, 2012, 12:58:33 AM7/12/12
to mnemosyne-...@googlegroups.com


On Monday, July 9, 2012 11:48:25 PM UTC-7, Peter Bienstman wrote:
On Monday, July 09, 2012 10:48:35 PM Chris wrote:
> Yes, just did.  Let me finish the 16 remaining scheduled cards for the day,
> then froze after about 5 unlearned cards.

Too bad...

We are talking about this version, right?

http://users.ugent.be/~pbienst/pub/mnemosyne-20120709-setup.exe


 Yes.
 
> I was thinking, I could load the program on another computer that I only
> rarely use, mount the database in that copy of th program, and see if it
> occurs there; which may let us know if it's something to do with the
> database, or the computer?
 
That could indeed be interesting...


I will try to get through my scheduled cards for tonight, and then perhaps try this tomorrow.  I have been having very little time with my practicum, work, and school, so hopefully I can fit some time in there.

Should I update to the new 2.0.1 you posted, or wait, as the same problems are reported for it?

  - Chris Shanks

ezehfr...@gmail.com

unread,
Jul 12, 2012, 10:55:59 PM7/12/12
to mnemosyne-...@googlegroups.com
On Thursday, July 12, 2012 12:58:33 AM UTC-4, Chris wrote:
> On Monday, July 9, 2012 11:48:25 PM UTC-7, Peter Bienstman wrote:<blockquote class="gmail_quote" style="margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex">On Monday, July 09, 2012 10:48:35 PM Chris wrote:
>
> &gt; Yes, just did.  Let me finish the 16 remaining scheduled cards for the day,
>
> &gt; then froze after about 5 unlearned cards.
>
>
>
> Too bad...
>
>
>
> We are talking about this version, right?
>
>
>
> <a href="http://users.ugent.be/%7Epbienst/pub/mnemosyne-20120709-setup.exe" target="_blank">http://users.ugent.be/~<WBR>pbienst/pub/mnemosyne-<WBR>20120709-setup.exe</a>
>
>
>
> </blockquote>
>
>  Yes.
>  </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">&gt; I was thinking, I could load the program on another computer that I only
>
> &gt; rarely use, mount the database in that copy of th program, and see if it
>
> &gt; occurs there; which may let us know if it&#39;s something to do with the
>
> &gt; database, or the computer?
>
>
>  </blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">That could indeed be interesting...
>
>
>
> </blockquote>
>
> I will try to get through my scheduled cards for tonight, and then perhaps try this tomorrow.  I have been having very little time with my practicum, work, and school, so hopefully I can fit some time in there.
>
> Should I update to the new 2.0.1 you posted, or wait, as the same problems are reported for it?
>
>   - Chris Shanks
> </div>

Having the same problem. It freezes after 2 cards.
Windows 7. I even tried running the software in safe mode. same result.

Peter Bienstman

unread,
Jul 13, 2012, 2:39:09 AM7/13/12
to mnemosyne-...@googlegroups.com
On Wednesday, July 11, 2012 09:58:33 PM Chris wrote:

> I will try to get through my scheduled cards for tonight, and then perhaps
> try this tomorrow. I have been having very little time with my practicum,
> work, and school, so hopefully I can fit some time in there.

No problem, take your time.

> Should I update to the new 2.0.1 you posted, or wait, as the same problems
> are reported for it?

At the moment, I'd suggest to wait, as I'll try to make another new installer
tonight.

Thanks!

Peter

Peter Bienstman

unread,
Jul 13, 2012, 3:15:03 AM7/13/12
to mnemosyne-...@googlegroups.com
Actually, another thing people can definitely try even with the current
installer is making sure that is much as possible, Mnemosyne is the only
program running.

So, don't have Window Explorer open when you launch Mnemosyne, or another
windows explorer alternative like Free Commander, Total commander, Directory
Opus ...

Other culprits can include windows indexing service / Windows search service,
Eldos callback filesystem .

Also, disable things like third party internet explorer / shell extensions,
such as Win7Shell, Tortoise SVN, ...

Peter

Oisín

unread,
Jul 13, 2012, 8:34:55 AM7/13/12
to mnemosyne-...@googlegroups.com


On 13 July 2012 02:55, <ezehfr...@gmail.com> wrote:
Having the same problem. It freezes after 2 cards.
Windows 7. I even tried running the software in safe mode. same result.

Sorry if this is something people have checked already, but have you tried with a fresh deck?
A few days ago I installed the latest test snapshot on my Windows 7 (64-bit) laptop and downloaded an XML deck from the HSK Flashcards site, and tested 100 or so cards and reviews, but could not replicate the crash/hang.

Peter Bienstman

unread,
Jul 13, 2012, 2:36:44 PM7/13/12
to mnemosyne-...@googlegroups.com
My current hypothesis for the hangs is that there is some sort of bad interaction between libraries mnemosyne uses (database library, installer library, ...) and/or a third party program running on people's computer (Freecommander, ...)

To figure out if removing the installer library from the mix helps, I here provide instructions for an alternate way of installing Mnemosyne. Rather than having a separate one-click install, the idea is to install Mnemosyne as a package in the system's Python installation.

Run all of the following installers:


Then, run C:\Python2.7\Scripts\mnemosyne and choose 'pythonw' to execute that script.

As extra benefit, Mnemosyne might start a little bit faster, and Windows power users can now also call libmnemosyne's functions from their python scripts, just as they can do under Linux.
Message has been deleted

Chris

unread,
Jul 13, 2012, 6:45:04 PM7/13/12
to mnemosyne-...@googlegroups.com
I got the following warning after installing the final mnemosyne (after all the other installs) above: (attached image)
Might not have installed correctly.png

Chris

unread,
Jul 13, 2012, 6:53:12 PM7/13/12
to mnemosyne-...@googlegroups.com
Okay, after all that the program opened.  I will try and go through cards to see if it freezes.  A couple questions, as this is getting pretty confusing/complicated from my biology and medicine (i.e., not computer science) perspective! :) Am I going to have to open it that same way with pythonw each time I use it now?  And secondly, the mnemosyne that is on my desktop, is that now an old version of the program that I should not use / remove, or has it been updated by this process?  And finally, when I run this pythonw script, is the database for the program and the image folder still in the old location that it was before, or has it made a copy somewhere else? 

  Sorry for all the questions, Chris Shanks.

Chris

unread,
Jul 13, 2012, 10:58:45 PM7/13/12
to mnemosyne-...@googlegroups.com
Awesome, made it through 237 scheduled cards and about 40 new, not memorized cards.  We'll see if this works again tomorrow.  Only strange thing I've noticed about this version so far is when looking at the SHOW STATISTICS > SCHEDULE graph, the values seem right (between 97 and 237 cards scheduled) for the last week, except for Day -1, which is showing 730 (which is/was totally wrong).  I will continue using this version of the program and report back any freezing issues, which I luckily have not yet discovered.

   Wishing continued luck, Chris S.

Chris

unread,
Jul 13, 2012, 11:36:10 PM7/13/12
to mnemosyne-...@googlegroups.com
Well... I spoke just too soon.   Just did three more unlearned cards ... froze.  AHHH! 

  - Chris

Peter Bienstman

unread,
Jul 14, 2012, 1:35:40 PM7/14/12
to mnemosyne-...@googlegroups.com
Quoting Chris <the.whis...@gmail.com>:

> Okay, after all that the program opened.

I also have no idea about this error message, but apparently, it was harmless.

> I will try and go through cards
> to see if it freezes. A couple questions, as this is getting pretty
> confusing/complicated from my biology and medicine (i.e., not computer
> science) perspective! :) Am I going to have to open it that same way with
> pythonw each time I use it now?

Tip: change 'mnemosyne' to 'mnemosyne.py' and you'll be able to skip
the pythont selecting step. You can also add a shortcut to this file
to your desktop for easier access.

> And secondly, the mnemosyne that is on my
> desktop, is that now an old version of the program that I should not use /
> remove, or has it been updated by this process?

You essentially now have to completely separate Mnemosyne installs,
where one of them does not update/influence the other one. Both
install still use the same data directory, though.

> Sorry for all the questions, Chris Shanks.

No problem, sorry for all the hassle in helping figuring out what's
going wrong :-)

Peter

Peter Bienstman

unread,
Jul 14, 2012, 1:36:55 PM7/14/12
to mnemosyne-...@googlegroups.com
Quoting Chris <the.whis...@gmail.com>:

> Only strange
> thing I've noticed about this version so far is when looking at the SHOW
> STATISTICS > SCHEDULE graph, the values seem right (between 97 and 237
> cards scheduled) for the last week, except for Day -1, which is showing 730
> (which is/was totally wrong).

Strange, feel free to send me your database privately, so that I can
have a look at it later. Could be caused by the crashes, though.

Peter


Peter Bienstman

unread,
Jul 14, 2012, 1:42:13 PM7/14/12
to mnemosyne-...@googlegroups.com
Quoting Chris <the.whis...@gmail.com>:

> Well... I spoke just too soon. Just did three more unlearned cards ...
> froze. AHHH!

Too bad, but at least it seems the hangs seem less frequent...

Well, it'd be interesting to see how this evolves over the next few
days, as well as getting more reports from other people too.

Anyway, this is what I'll plan to do: first take a week's holiday
break (I will be replying to email, though), and then I'll try to
update/replace the database library we use. It will take some time,
though, as it will require some architectural changes.

Basically, after updating PyQt and eliminating the installer library,
this is the last third party library I can change. If that doesn't
work, I don't see any other way out. The database library update
promises to bring some speed improvements, so in case it's interesting
to do.

Cheers,

Peter


Chris

unread,
Jul 14, 2012, 5:12:34 PM7/14/12
to mnemosyne-...@googlegroups.com
Since then, it has hung repeatedly, quickly after opening (i.e., just a few cards). 
 - Chris

Chris

unread,
Jul 14, 2012, 6:09:27 PM7/14/12
to mnemosyne-...@googlegroups.com
As a last-ditch, final hope effort, I just installed Mnemosyne on my desktop (Windows 7 64-bit). Froze on the second card.
I will continue to wait patiently.  Many thanks for all your efforts so far, Peter.  Have a good holiday, and look forward to hearing from you when you get back.  I will be Zen in my patient repetitions, or at least attempt to be.  I had so hoped my desktop would have been able to pull it off!

  - Thanks again for your time, Chris Shanks.

George Wade

unread,
Jul 14, 2012, 6:21:56 PM7/14/12
to mnemosyne-...@googlegroups.com
As Windows has so many combinations of OS and hardware of all kinds, would it be helpful — more than disruptive, to suggest that sometimes: a Linux distro on a usb-stick or even DVD-RW could be tried for better compatibility. The expense would be the USB-stick or DVD + a few hours of invaluable time. In case it works ow t the payback could be considerable.

George
Retreating into safe depths
of the BC countryside...

Chris

unread,
Jul 14, 2012, 6:28:54 PM7/14/12
to mnemosyne-...@googlegroups.com
I just tried a third install (this time the python 2.0.1 test).  Each of the installs froze on the same card, even after totally restarting and mounting the same database.  I don't know if that's a hint of any sort.  I'll wait till later and try again.  That seemed to help before.

 - Chris

PS Re: George, I don't think I'm confident enough to try and run linux.

Peter Bienstman

unread,
Jul 15, 2012, 2:18:17 AM7/15/12
to mnemosyne-...@googlegroups.com
OK, at the moment we know that the one-click installer has these issues.

What I'm mostly interested in at the moment is if the 'moderately
good' results you got with the multiclick installer were a one-off.

Cheers,

Peter
> --
> You received this message because you are subscribed to the Google
> Groups "mnemosyne-proj-users" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/mnemosyne-proj-users/-/Z_At1A21YXcJ.
> To post to this group, send email to mnemosyne-...@googlegroups.com.
> To unsubscribe from this group, send email to
> mnemosyne-proj-u...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/mnemosyne-proj-users?hl=en.
>
>



Chris

unread,
Jul 16, 2012, 12:44:02 AM7/16/12
to mnemosyne-...@googlegroups.com
I just took about 12 hours learning how Linux works, wiping my operating system on one of my computers, installing Linux, figuring out how to manually install a whole bunch of things in Linux, figured out how to use the command prompt in Linux, stayed up all night doing it....  It took me hours of learning and much frustration, and I wiped my computer for it .... AND IT FROZE.  IT FROZE IN LINUX.

   - Chris Shanks


On Saturday, July 14, 2012 11:18:17 PM UTC-7, Peter Bienstman wrote:
OK, at the moment we know that the one-click installer has these issues.

What I'm mostly interested in at the moment is if the 'moderately  
good' results you got with the multiclick installer were a one-off.

Cheers,

Peter

Quoting Chris <the.whis...@gmail.com>:

> I just tried a third install (this time the python 2.0.1 test).  Each of
> the installs froze on the same card, even after totally restarting and
> mounting the same database.  I don't know if that's a hint of any sort.
> I'll wait till later and try again.  That seemed to help before.
>
>  - Chris
>
> PS Re: George, I don't think I'm confident enough to try and run linux.
>
> --
> You received this message because you are subscribed to the Google  
> Groups "mnemosyne-proj-users" group.
> To view this discussion on the web visit  
> https://groups.google.com/d/msg/mnemosyne-proj-users/-/Z_At1A21YXcJ.
> To post to this group, send email to mnemosyne-proj-users@googlegroups.com.
> To unsubscribe from this group, send email to  
> mnemosyne-proj-users+unsub...@googlegroups.com.

Chris

unread,
Jul 16, 2012, 12:54:00 AM7/16/12
to mnemosyne-...@googlegroups.com
Repeatedly.

Chris

unread,
Jul 16, 2012, 1:05:53 AM7/16/12
to mnemosyne-...@googlegroups.com
I am happy, however, that I finally get this XKCD comic: http://xkcd.com/149/   .... I tried putting the final install command into the terminal in Linux repeatedly, couldn't figure out why it wouldn't work, after about 20 tries, I thought maybe put SUDO in front of it, like I'd read somewhere...  Worked perfectly.

On Sunday, July 15, 2012 9:54:00 PM UTC-7, Chris wrote:
Repeatedly.

Peter Bienstman

unread,
Jul 16, 2012, 1:13:03 AM7/16/12
to mnemosyne-...@googlegroups.com
Did it freeze in the same way? Does it always freeze on the same card?
Feel free to privately mail me your database if that's the case.

Does it also freeze when you create a fresh dummy database with a few
dummy cards?
>> mnemosyne-...@googlegroups.com.
>> > To unsubscribe from this group, send email to
>> > mnemosyne-proj-u...@googlegroups.com.
>> > For more options, visit this group at
>> > http://groups.google.com/group/mnemosyne-proj-users?hl=en.
>> >
>> >
>>
>>
>>
>>
>
> --
> You received this message because you are subscribed to the Google
> Groups "mnemosyne-proj-users" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/mnemosyne-proj-users/-/ILkIzpNprBgJ.
> To post to this group, send email to mnemosyne-...@googlegroups.com.
> To unsubscribe from this group, send email to
> mnemosyne-proj-u...@googlegroups.com.

Chris

unread,
Jul 16, 2012, 1:34:58 AM7/16/12
to mnemosyne-...@googlegroups.com
Doesn't always freeze on the same card.  Sometimes it will freeze on the same card for a couple of tries, but this is when it's really acting up and always freezing on the second card... but I think this must be a coincidence, because then it will be able to get past it if I leave it for a while.  I was wondering, when I leave it alone for a while, is there a period of time which will shuffle the database?  Perhaps this is when it can get through more cards.  I really don't know.  I will create fake database and get back to you.  I guess it must be something to do with my large database though, given I've tried two Windows OS's and one LINUX!?  I'll get back to you in a few minutes.

   - Chris S.
>> mnemosyne-proj-users@googlegroups.com.
>> > To unsubscribe from this group, send email to
>> > mnemosyne-proj-users+unsub...@googlegroups.com.
>> > For more options, visit this group at
>> > http://groups.google.com/group/mnemosyne-proj-users?hl=en.
>> >
>> >
>>
>>
>>
>>
>
> --
> You received this message because you are subscribed to the Google  
> Groups "mnemosyne-proj-users" group.
> To view this discussion on the web visit  
> https://groups.google.com/d/msg/mnemosyne-proj-users/-/ILkIzpNprBgJ.
> To post to this group, send email to mnemosyne-proj-users@googlegroups.com.
> To unsubscribe from this group, send email to  
> mnemosyne-proj-users+unsub...@googlegroups.com.

Chris

unread,
Jul 16, 2012, 1:49:45 AM7/16/12
to mnemosyne-...@googlegroups.com
Okay, I made up a 6-card database (in the Linux copy).  I cycled through it many times using learn ahead of schedule.  Didn't ever freeze.  So I tried a very old database that I made as a backup years ago (imported XML from Mnemosyne 1), it had about 500 cards.  Went through a couple hundred, really quickly (not actually paying attention to card info), failing some and passing some randomly.  Didn't freeze.  Then I tried another very old one (couple years ago, 5000 + cards imported from XML).  Got through a few hundred from it, but it had over 5000 scheduled, so I couldn't get to newly learned.  But it never froze.  So it would seem to be something to do with some fault in the one database that I am currently using?  Perhaps this is for everybody who is having problems?  The database I have been having the problem with on all the computers is about 7000 cards (the only special features on any of them are images).  If it would help you diagnose the problem, I will gladly send it to you.  How do I privately send it?

   - Chris.

Chris

unread,
Jul 16, 2012, 2:06:05 AM7/16/12
to mnemosyne-...@googlegroups.com
Went back to my important database, froze after about 10 cards.  Freezes the exact same way it does in windows.  Can't say exactly when it is though, because it is clearly sometimes when a card is revealed, or graded, but other times I've noticed it will be showing a card, and I will click anywhere on the window, or on the file menu, and it will realize it is frozen (i.e., it will give the "not responding" message in windows, will go grey in Linux ubuntu).

Scott Youngman

unread,
Jul 16, 2012, 10:54:59 PM7/16/12
to mnemosyne-...@googlegroups.com
I installed the multiple parts (not the single installation package) and ran it on my regular default.db. Autosave after 20 repetitions.
The number of cards shown before hanging in each of ten successive runs (restart after each hang) is: 13, 3, 13, 3, 13, 15, 20, 20, 53, 14 cards.
That is essentially the same as with any of the recent single-installer versions. (Win7-64.)

Peter Bienstman

unread,
Jul 17, 2012, 1:16:45 AM7/17/12
to mnemosyne-...@googlegroups.com
Well, the good news is that at least there seems to be progress in
zooming in on the bug :-)

You can mail me the database at Peter.Bienstman (at) UGent.be

(Note that previously when people sent me their database for these
Windows hangs, I could not reproduce it, though, so perhaps this is a
different issue.)

Anyway, thanks for all your heroic efforts in helping to track this down!

Peter
>>>> >> mnemosyne-...@googlegroups.com.
>>>> >> > To unsubscribe from this group, send email to
>>>> >> > mnemosyne-proj-u...@googlegroups.com.
>>>> >> > For more options, visit this group at
>>>> >> > http://groups.google.com/group/mnemosyne-proj-users?hl=en.
>>>> >> >
>>>> >> >
>>>> >>
>>>> >>
>>>> >>
>>>> >>
>>>> >
>>>> > --
>>>> > You received this message because you are subscribed to the Google
>>>> > Groups "mnemosyne-proj-users" group.
>>>> > To view this discussion on the web visit
>>>> > https://groups.google.com/d/msg/mnemosyne-proj-users/-/ILkIzpNprBgJ.
>>>> > To post to this group, send email to
>>>> mnemosyne-...@googlegroups.com.
>>>> > To unsubscribe from this group, send email to
>>>> > mnemosyne-proj-u...@googlegroups.com.
>>>> > For more options, visit this group at
>>>> > http://groups.google.com/group/mnemosyne-proj-users?hl=en.
>>>> >
>>>> >
>>>>
>>>>
>>>>
>>>>
>
> --
> You received this message because you are subscribed to the Google
> Groups "mnemosyne-proj-users" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/mnemosyne-proj-users/-/B0WKQw-CvNUJ.
> To post to this group, send email to mnemosyne-...@googlegroups.com.
> To unsubscribe from this group, send email to
> mnemosyne-proj-u...@googlegroups.com.

Chris

unread,
Jul 17, 2012, 3:22:49 AM7/17/12
to mnemosyne-...@googlegroups.com
Scott... that's interesting.  My distribution of freezes is much wider, but clusters... so if it starts freezing, it will be something like 2, 2, 4, 2, 3, 2, and then if I give it a day, sometimes I can get through >200 before it freezes... then it starts doing the tight cluster again.  I figured out if I screw around with my computer's time/date settings, I can get through many cards.  However, I am worried that this might somehow irreparably corrupt my database, because I flip it back in time?  Maybe Peter would know if this is a dangerous method to use?  I'll try my cards tonight, and get back to you tomorrow... maybe I'll send the database, if I have time tomorrow.  Other than that, maybe I'll just have to wait for an export feature which will allow me to import back into Ver. 1.  But I do love those cloze deletion cards... I've added a few hundred of those already.  I might give up soon; I'm starting to feel like I'm getting more annoying than helpful!  I've just invested so much in this great program! :)

   - Till tomorrow, Chris.

Chris

unread,
Jul 17, 2012, 3:58:54 AM7/17/12
to mnemosyne-...@googlegroups.com
Well, as usual, spoke too soon... with no monkeying around with time/date settings, today I got 46, 21, 103, 10, 3, 1, 2, 1...give up.  Surprising the freezing didn't keep increasing in frequency after 21.

querido

unread,
Jul 17, 2012, 5:39:24 AM7/17/12
to mnemosyne-...@googlegroups.com
Pardon me, I'm not an expert.
 
There are troubleshooting tools that can step through one statement at a time, right? And there are methods for stepping through larger blocks, like printing to a log or pop-up boxes that say "...start function x...", "...end function x...", "...start function y...", "...end function y...", etc., right? But you can't do this because the bug doesn't occur on your machine.
 
If those with the bug are not programmers you could provide a debugging version with these additional statements, then have them send you the log.
 
Here's an idea: One of those with the bug could give you "remote desktop" access, and you could debug it on their machine. Anyone?

Scott Youngman

unread,
Jul 17, 2012, 8:14:03 PM7/17/12
to mnemosyne-...@googlegroups.com
One of the difficult things about this bug is that nothing turns out to be a repeatable pattern. There have been so many times I thought I saw a regularity, but it failed within a day or two.

Peter Bienstman

unread,
Jul 18, 2012, 1:15:50 AM7/18/12
to mnemosyne-...@googlegroups.com
Quoting querido <tworoads...@gmail.com>:

> Pardon me, I'm not an expert.

No problem, thanks for your feedback!

> There are troubleshooting tools that can step through one statement at a
> time, right? And there are methods for stepping through larger blocks, like
> printing to a log or pop-up boxes that say "...start function x...",
> "...end function x...", "...start function y...", "...end function y...",
> etc., right? But you can't do this because the bug doesn't occur on your
> machine.

Indeed. The problem is that I suspect that the issue is not in
Mnemosyne's own Python code, but rather in the C code of one of the
libraries Mnemosyne uses. So, using a debugger to step through the
Python code would probably not be very helpful.

Still, it could be worthwhile doing this to validate this assumption.
Running Mnemosyne in a debugger is not possible with the one-click
installer, but is possible with the multi-click installer. I could
provide some instructions later on on how to do this if anybody is
interested.

Cheers,

Peter


Peter Bienstman

unread,
Jul 18, 2012, 1:25:28 AM7/18/12
to mnemosyne-...@googlegroups.com
Quoting Chris <the.whis...@gmail.com>:

> Scott... that's interesting. My distribution of freezes is much wider, but
> clusters... so if it starts freezing, it will be something like 2, 2, 4, 2,
> 3, 2, and then if I give it a day, sometimes I can get through >200 before
> it freezes... then it starts doing the tight cluster again. I figured out
> if I screw around with my computer's time/date settings, I can get through
> many cards. However, I am worried that this might somehow irreparably
> corrupt my database, because I flip it back in time? Maybe Peter would
> know if this is a dangerous method to use?

I really don't recommend doing this, as it will mess up your
scheduling information. You might even have added a second, different
problem in that way :-)

> I'll try my cards tonight, and
> get back to you tomorrow... maybe I'll send the database, if I have time
> tomorrow.

Please do, a database that hangs under Linux is a first for me, and
crucial to help me figure out what's going on, and whether we're
looking at one or at multiple confounding bugs.

Also, contrary to other people, new dummy databases seem to work in
your case, so if you have a new database which also hangs, that would
rule out the database layer as cause of the problems.

> Other than that, maybe I'll just have to wait for an export
> feature which will allow me to import back into Ver. 1.

That's currently not something I plan, as that would result in loss of
information, since there are many features which 1.x simply not
supports.

I'd rather spend my time fixing the 2.0 issue :-)

> I'm starting to feel like I'm getting more annoying
> than helpful!

Please don't, you're certainly not annoying, but crucial in helping
out to figure what's going on.

Believe me, it's also very frustrating for me, not being able to
reproduce these hangs, and not having any idea how widespread this
problem is.

Thanks for your continued feedback!

Cheers,

Peter

Chris

unread,
Jul 18, 2012, 3:29:53 AM7/18/12
to mnemosyne-...@googlegroups.com
I thought you wouldn't be creating a way to directly export back to one, but if you created an exporter, it would have some method like XML with card learning statistics, right?  Then I could manually turn around and load it back into 1, and it would even be able to keep things like the close deletion / double-sided cards, though it would lose some of its abilities, like the recognition of sister cards. 

Regarding recreating the issue with a new database.  I am suspicious, given that everybody else's seemed to freeze, that mine didn't. I did it so quick, though, and I didn't let a day pass for it to attempt to reschedule itself.  Perhaps I'll make another dummy one and try and run it for a few days.

  - Chris.

PS.  That linux bit blew me away too.  I thought if I took such desperate, whole-hearted measures, it would be sure to work fine.  You should have seen my stunned reaction when it froze after all.

A...@live.co.uk

unread,
Jul 18, 2012, 6:58:52 AM7/18/12
to mnemosyne-...@googlegroups.com
All my cards are gone.... mnemosyne is not starting up... it says error in config and uncaught eception... i feel like crying... please sir... include the export option... I want to export my cards.... it's probably too late... i see no cards.. even on other pc.... please help me...

silverbear

unread,
Jul 18, 2012, 8:04:29 AM7/18/12
to mnemosyne-...@googlegroups.com
Coming late to this party as I've been away, but as far as finding out how widespread this problem is:
I switch m/c's regularly, syncing between them, and have had no hangs at all. My m/c's are Vista 32, Vista 64 and XP 32. My processors & available memory are high in respect of the operating system being used. The windows indexing option is on/off depending on the m/c.
My database was upgraded from a v1 database at about the time you started releasing v2 test versions; it has 7K+ cards in it of the Vocab type, many have sound but no pictures. These are in a separate db, not the default one. Some cards were loaded via the front end, some as a bulk import.

The only problem I have (minor) is that the sound is a bit delayed on the first couple of cards when starting up, but this doesn't freeze the program.
'Hope this helps.......



On Monday, 9 July 2012 19:48:42 UTC+1, Peter Bienstman wrote:
Hi,

As you know, there have been quite some reports of Mnemosyne sometimes  
hanging on Windows. I had hoped that the new PyQt version would fix  
this, but unfortunately, that turned out not be the case. The hangs  
never happen on my system, so I need help to debug this.

To summarise what I understand the problem to be (but feel free to  
correct me), Mnemosyne sometimes hangs when trying to grade a card,  
and this happens regardsless of whether the cards contain images or  
not. It happens both on XP 32 bit and on W7 64 bit.

I prepared a new test version here:

http://users.ugent.be/~pbienst/pub/mnemosyne-20120709-setup.exe

I changed some small packaging/configuration options. I'm not sure if  
they will have a big impact, but you never know :-)

If you still have hangs, it would be interesting to see if how the  
frequency of hangs changes if you try e.g. 'autosave after 1 reps' vs  
'autosave after 20 reps'.

If you have a virus scanner like Microsoft security essentials  
running, please try disabling it and see if it makes a difference.  
Ditto for Windows desktop search.

If you have a separate version of Python installed on your machine,  
please remove it.

Do a search for default.db on your file system, and check if  
default.db has write permissions, together with all the folder levels  
above it.

Do you have funny characters in your windows account name?

Is your database on the C drive?

Do you also have the problem if you never used the card browser after  
starting the program?

If you're technically inclined, run Sysinternal process monitor while  
running Mnemosyne, to see if anything suspect happens with database  
access (filter to show only events in the 'file system' class) (  
http://technet.microsoft.com/en-us/sysinternals/bb896645 )

Make sure you have enough free disk space and try running a defragmenter.

Any other pattern you observe or thought you might have to solve this,  
is welcome!

Peter

Peter Bienstman

unread,
Jul 18, 2012, 1:45:45 PM7/18/12
to mnemosyne-...@googlegroups.com
Don't panic, just tell me exactly what the error message is.

Which version are you using?

Peter

All my cards are gone.... mnemosyne is not starting up... it says error in config and uncaught eception... i feel like crying... please sir... include the export option... I want to export my cards.... it's probably too late... i see no cards.. even on other pc.... please help me... 

-- 
You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group.
To post to this group, send email to mnemosyne-...@googlegroups.com.
To unsubscribe from this group, send email to mnemosyne-proj-u...@googlegroups.com.

Peter Bienstman

unread,
Jul 18, 2012, 2:01:17 PM7/18/12
to mnemosyne-...@googlegroups.com
Quoting Chris <the.whis...@gmail.com>:

> I thought you wouldn't be creating a way to directly export back to one,
> but if you created an exporter, it would have some method like XML with
> card learning statistics, right? Then I could manually turn around and
> load it back into 1, and it would even be able to keep things like the
> close deletion / double-sided cards, though it would lose some of its
> abilities, like the recognition of sister cards.

The structure of the 2.0 xml is very different to support all the new
features, and cannot be read by 1.x, sorry...

> Regarding recreating the issue with a new database. I am suspicious, given
> that everybody else's seemed to freeze, that mine didn't. I did it so
> quick, though, and I didn't let a day pass for it to attempt to reschedule
> itself. Perhaps I'll make another dummy one and try and run it for a few
> days.

OK.

> PS. That linux bit blew me away too. I thought if I took such desperate,
> whole-hearted measures, it would be sure to work fine. You should have
> seen my stunned reaction when it froze after all.

Please send the file privately, it's the most important thing you can
do at the moment :-)

Peter



xor...@gmail.com

unread,
Jul 18, 2012, 7:06:08 PM7/18/12
to mnemosyne-...@googlegroups.com
On Monday, July 9, 2012 11:48:42 AM UTC-7, Peter Bienstman wrote:
> Hi,
>
> As you know, there have been quite some reports of Mnemosyne sometimes
> hanging on Windows. I had hoped that the new PyQt version would fix
> this, but unfortunately, that turned out not be the case. The hangs
> never happen on my system, so I need help to debug this.
>
> To summarise what I understand the problem to be (but feel free to
> correct me), Mnemosyne sometimes hangs when trying to grade a card,
> and this happens regardsless of whether the cards contain images or
> not. It happens both on XP 32 bit and on W7 64 bit.

Another report for your collection:

Running Windows 7 64-bit with the original 2.0 release of Mnemosyne. Crashes primarily when revealing an answer (ie, I've made a guess, and am wanting to see the correct answer.

No Python installed, but I am running on a development machine, so I've got sundry other software installed. Not sure what might be relevant.

Running on a DB created from cards imported from several 1.x .mem files. Many of my cards have LaTeX in them.

Typical crash comes after 20-40 cards, then repeats more frequently, sometimes much more so. Sometimes deleting db.db-journal prevents the increase in frequency, but that may just be random.

I've also noted that a few of my cards seem to lock up repeatedly over several attempts. Not sure if there's something up with those cards, or not.

Happy to send a copy of my DB; I'm viewing this discussion via the web portal, however, and your (Peter) email address isn't anywhere I can locate it. Ping me, and I'll send it.

Will try the more recent one, and post if that improves things.

I'm muddling through at the moment with frequent auto-saves..

Scott Youngman

unread,
Jul 18, 2012, 9:09:16 PM7/18/12
to mnemosyne-...@googlegroups.com
"You should have seen my stunned reaction when it froze after all."

Chris, I can imagine it vividly right now... even feel the shock myself.

Chris

unread,
Jul 19, 2012, 3:28:25 AM7/19/12
to mnemosyne-...@googlegroups.com
Scott Youngman said it perfectly here:


On Tuesday, July 17, 2012 5:14:03 PM UTC-7, Scott Youngman wrote:
One of the difficult things about this bug is that nothing turns out to be a repeatable pattern. There have been so many times I thought I saw a regularity, but it failed within a day or two.

I had thought when my freezing started repeating quite quickly, I had to give up for a day or so.  I discovered over the past few days, if I just keep restarting and restarting, I would eventually again get past a big chunk of cards. 

I noticed (though I will be pensive about it), that sometimes if it is freezing repeatedly quickly, if I go to SETTINGS > CONFIGURE MNEMOSYNE > REVIEW MEMORIZED CARDS IN RANDOM ORDER > OK, sometimes that automatically shuffles the deck, allowing it to get past a sticky card.  However, for some reason, sometimes it does immediately shuffle the deck and show a different card on clicking "OK," at other times, it doesn't.  I don't know if this is meaningful or a combination of statistical chance + my confirmation bias (as each of my other hypotheses seems to have been).

  - Chris S.

P.S. Peter, I have sent you my database.

Peter Bienstman

unread,
Jul 19, 2012, 1:01:51 PM7/19/12
to mnemosyne-...@googlegroups.com
Quoting Chris <the.whis...@gmail.com>:
> P.S. Peter, I have sent you my database.

Got it, thanks!

Unfortunately, the power supply of my linux PC died today, and it
could take some time to get it fixed...

Oh, well...

Peter


Peter Bienstman

unread,
Jul 19, 2012, 1:19:27 PM7/19/12
to mnemosyne-...@googlegroups.com
BTW, on my Windows laptop, I pressed enter on your database for about 250 times, and I didn't not see a hang...
 
Peter

Chris

unread,
Jul 20, 2012, 2:51:28 AM7/20/12
to mnemosyne-...@googlegroups.com
This is so bizarre. So you haven't been able to reproduce it from anybody's databases you've looked at? To put it bluntly, what the hell could be causing this? I've used three separate computers, each with a separate OS, with that database, Win 7 64-bit desktop, Win 7 starter netbook (with two separate installs), and Ubuntu Linux 10 laptop, each freezing. I have moved the database back and forth repeatedly, and when the program is freezing, it will freeze quickly in each version. Oh well.

Murray James Morrison

unread,
Jul 20, 2012, 5:55:10 AM7/20/12
to mnemosyne-...@googlegroups.com
Hi Chris,

I also have a Windows 7 64-bit desktop running the latest version of Mnemosyne 2 (and before that, every Beta in between). I have 8000 cards in my database, but have never had any freezing problems.

If you need an extra pair of eyes, I could try your database on my machine. Not sure if the problem is with the database, but if it is, it should freeze for me too.

Do you have an Android/Java phone? Have you tried reviewing using Mnemogogo?

--murrayjames


On Fri, Jul 20, 2012 at 2:51 PM, Chris <the.whis...@gmail.com> wrote:
This is so bizarre.  So you haven't been able to reproduce it from anybody's databases you've looked at?  To put it bluntly, what the hell could be causing this?  I've used three separate computers, each with a separate OS, with that database, Win 7 64-bit desktop, Win 7 starter netbook (with two separate installs), and Ubuntu Linux 10 laptop, each freezing.  I have moved the database back and forth repeatedly, and when the program is freezing, it will freeze quickly in each version.  Oh well.
--
You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group.
To post to this group, send email to mnemosyne-...@googlegroups.com.
To unsubscribe from this group, send email to mnemosyne-proj-u...@googlegroups.com.




--
=================
Murray James Morrison
Saxophonist, Composer, Music Educator

Tel.: +86-18608001531 (China)

Peter Bienstman

unread,
Jul 20, 2012, 1:32:59 PM7/20/12
to mnemosyne-...@googlegroups.com
My linux box is repaired now, and what's even more bizarre is that
your database works there without problems too...

To be sure, it's freezing both with keyboard input (pressing enter to
do the default action) and mouse input?

Peter

Oisín

unread,
Jul 20, 2012, 11:35:38 PM7/20/12
to mnemosyne-...@googlegroups.com
On 20 July 2012 10:55, Murray James Morrison <murra...@murrayjames.net> wrote:
Hi Chris,

I also have a Windows 7 64-bit desktop running the latest version of Mnemosyne 2 (and before that, every Beta in between). I have 8000 cards in my database, but have never had any freezing problems.

If you need an extra pair of eyes, I could try your database on my machine. Not sure if the problem is with the database, but if it is, it should freeze for me too.


I couldn't make a hang happen on my Win 7 64-bit laptop, either. Someone posted that they made a new deck with a smaller number of test cards, and that hung too?
That deck seems like a good place to start, then. I'd also happily test if someone wants to share a test deck which has been causing hangs for them?

It's very curious... perhaps Qt or the Latex code are having an interaction effect with something else running on the system, maybe calling an external program which is masked by something in the users' paths?


Scott Youngman

unread,
Jul 20, 2012, 11:55:33 PM7/20/12
to mnemosyne-...@googlegroups.com
I and at least one other poster experienced the hang when running in Safe Mode (Win7-64 for me). That's about as far as one can go to remove potentially conflicting programs or processes in the background.

Oisín

unread,
Jul 21, 2012, 12:57:36 AM7/21/12
to mnemosyne-...@googlegroups.com
On 21 July 2012 04:55, Scott Youngman <syo...@gmail.com> wrote:
I and at least one other poster experienced the hang when running in Safe Mode (Win7-64 for me). That's about as far as one can go to remove potentially conflicting programs or processes in the background.

No, I'd say a more likely cause would be programs or libraries in the system path which are being called instead of the ones Mnemosyne intends to call... maybe. Safe mode might not prevent that.
Actual running processes interfering with Mnemosyne, like a virus killer (or virus) are possible but very unlikely,

Anyway like I said, if you can create a deck with a very small number of test cards which hangs for you, upload it somewhere and more of us can try to trigger the same behaviour and solve this problem.


On Friday, July 20, 2012 9:35:38 PM UTC-6, Oisín Mac Fhearaí wrote:
... perhaps Qt or the Latex code are having an interaction effect with something else running on the system, maybe calling an external program which is masked by something in the users' paths?

--
You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group.
To post to this group, send email to mnemosyne-...@googlegroups.com.
To unsubscribe from this group, send email to mnemosyne-proj-u...@googlegroups.com.

Chris

unread,
Jul 21, 2012, 6:30:06 AM7/21/12
to mnemosyne-...@googlegroups.com, murra...@murrayjames.net
Murray James, I'll send you my database, but if it works on both of Peter's, it probably will on yours.  I don't have either of the Memogogo-compatable phones.  I will try to get the database to you tomorrow.

  - Chris Shanks


On Friday, July 20, 2012 2:55:10 AM UTC-7, Murray James Morrison wrote:
Hi Chris,

I also have a Windows 7 64-bit desktop running the latest version of Mnemosyne 2 (and before that, every Beta in between). I have 8000 cards in my database, but have never had any freezing problems.

If you need an extra pair of eyes, I could try your database on my machine. Not sure if the problem is with the database, but if it is, it should freeze for me too.

Do you have an Android/Java phone? Have you tried reviewing using Mnemogogo?

--murrayjames

On Fri, Jul 20, 2012 at 2:51 PM, Chris <the.whis...@gmail.com> wrote:
This is so bizarre.  So you haven't been able to reproduce it from anybody's databases you've looked at?  To put it bluntly, what the hell could be causing this?  I've used three separate computers, each with a separate OS, with that database, Win 7 64-bit desktop, Win 7 starter netbook (with two separate installs), and Ubuntu Linux 10 laptop, each freezing.  I have moved the database back and forth repeatedly, and when the program is freezing, it will freeze quickly in each version.  Oh well.

--
You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group.
To post to this group, send email to mnemosyne-proj-users@googlegroups.com.
To unsubscribe from this group, send email to mnemosyne-proj-users+unsub...@googlegroups.com.

Chris

unread,
Jul 21, 2012, 6:33:18 AM7/21/12
to mnemosyne-...@googlegroups.com
Peter: I've tried both methods repeatedly, and had it freeze on both.  I tried switching it on "sticky" cards, too (when one card keeps freezing for a while), but it didn't help.  I've basically tried every simple approach to fix it, with the level of computer skills I have (and the limits of my imagination/creativity).  I'm actually willing to buy a new computer for the sole purpose of getting this program working.  I would just have an unfortunate adult temper-tantrum if that didn't work. :) 

   - Chris Shanks


On Friday, July 20, 2012 10:32:59 AM UTC-7, Peter Bienstman wrote:
My linux box is repaired now, and what's even more bizarre is that  
your database works there without problems too...

To be sure, it's freezing both with keyboard input (pressing enter to  
do the default action) and mouse input?

Peter

Quoting Chris <the.whis...@gmail.com>:

> This is so bizarre.  So you haven't been able to reproduce it from  
> anybody's databases you've looked at?  To put it bluntly, what the  
> hell could be causing this?  I've used three separate computers,  
> each with a separate OS, with that database, Win 7 64-bit desktop,  
> Win 7 starter netbook (with two separate installs), and Ubuntu Linux  
> 10 laptop, each freezing.  I have moved the database back and forth  
> repeatedly, and when the program is freezing, it will freeze quickly  
> in each version.  Oh well.
>
> --
> You received this message because you are subscribed to the Google  
> Groups "mnemosyne-proj-users" group.
> To post to this group, send email to mnemosyne-proj-users@googlegroups.com.
> To unsubscribe from this group, send email to  
> mnemosyne-proj-users+unsub...@googlegroups.com.

Chris

unread,
Jul 21, 2012, 6:39:00 AM7/21/12
to mnemosyne-...@googlegroups.com
Oisín:  I could just send you my whole database.  I would suppose it could be some other program interfering with it, but I don't have much stuff on my computers.  And the Linux I ran it on was a fresh install just for it.  I would think there's a low chance of some strange program interfering independently on all three of my computers, and on neither of Peter's.  But I have no idea at this point, especially with a fresh Ubuntu.

I wasn't able to get a new deck to hang, but others reported they could.  I was going to try and run a test deck for a longer and see if the problems start with it.  I'll try and send you the database in the next few days.  I have just been spending so much time trying to get through my scheduled cards with the freezes that I haven't had much patience to try and keep running a fake deck at the same time.

  - Thanks for all your suggestions, Chris S.



On Friday, July 20, 2012 8:35:38 PM UTC-7, Oisín Mac Fhearaí wrote:
On 20 July 2012 10:55, Murray James Morrison <murra...@murrayjames.net> wrote:
Hi Chris,

I also have a Windows 7 64-bit desktop running the latest version of Mnemosyne 2 (and before that, every Beta in between). I have 8000 cards in my database, but have never had any freezing problems.

If you need an extra pair of eyes, I could try your database on my machine. Not sure if the problem is with the database, but if it is, it should freeze for me too.


I couldn't make a hang happen on my Win 7 64-bit laptop, either. Someone posted that they made a new deck with a smaller number of test cards, and that hung too?
That deck seems like a good place to start, then. I'd also happily test if someone wants to share a test deck which has been causing hangs for them?

It's very curious... perhaps Qt or the Latex code are having an interaction effect with something else running on the system, maybe calling an external program which is masked by something in the users' paths?


Do you have an Android/Java phone? Have you tried reviewing using Mnemogogo?

--murrayjames

On Fri, Jul 20, 2012 at 2:51 PM, Chris <the.whis...@gmail.com> wrote:
This is so bizarre.  So you haven't been able to reproduce it from anybody's databases you've looked at?  To put it bluntly, what the hell could be causing this?  I've used three separate computers, each with a separate OS, with that database, Win 7 64-bit desktop, Win 7 starter netbook (with two separate installs), and Ubuntu Linux 10 laptop, each freezing.  I have moved the database back and forth repeatedly, and when the program is freezing, it will freeze quickly in each version.  Oh well.

--
You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group.
To post to this group, send email to mnemosyne-proj-users@googlegroups.com.
To unsubscribe from this group, send email to mnemosyne-proj-users+unsub...@googlegroups.com.
--
=================
Murray James Morrison
Saxophonist, Composer, Music Educator

Tel.: +86-18608001531 (China)

--
You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group.
To post to this group, send email to mnemosyne-proj-users@googlegroups.com.
To unsubscribe from this group, send email to mnemosyne-proj-users+unsub...@googlegroups.com.

George Wade

unread,
Jul 21, 2012, 7:34:14 AM7/21/12
to mnemosyne-...@googlegroups.com
I have heard that some well built and slightly expensive computers give less trouble arising from hardware quirks:  that makes them cheaper long term.  See if you can get a guarantee that you can change brands within a month if you first purchase does not work out.  Future Shop in North America seems to promote that marketing style.  A Mac with WINE should be very stable...;~)  if you can put up with being teased.

George


On 21 Jul 2012, at 03:33, Chris wrote:

Peter: I've tried both methods repeatedly, and had it freeze on both.  I tried switching it on "sticky" cards, too (when one card keeps freezing for a while), but it didn't help.  I've basically tried every simple approach to fix it, with the level of computer skills I have (and the limits of my imagination/creativity).  I'm actually willing to buy a new computer for the sole purpose of getting this program working.  I would just have an unfortunate adult temper-tantrum if that didn't work. :) 

   - Chris Shanks

On Friday, July 20, 2012 10:32:59 AM UTC-7, Peter Bienstman wrote:
My linux box is repaired now, and what's even more bizarre is that  
your database works there without problems too...

To be sure, it's freezing both with keyboard input (pressing enter to  
do the default action) and mouse input?

Peter
.....
--

Oisín

unread,
Jul 21, 2012, 12:16:26 PM7/21/12
to mnemosyne-...@googlegroups.com
On 21 July 2012 11:39, Chris <the.whis...@gmail.com> wrote:
Oisín:  I could just send you my whole database.  I would suppose it could be some other program interfering with it, but I don't have much stuff on my computers.  And the Linux I ran it on was a fresh install just for it.  I would think there's a low chance of some strange program interfering independently on all three of my computers, and on neither of Peter's.  But I have no idea at this point, especially with a fresh Ubuntu.

Sure, I'll try to make the hang happen if you send me a copy of your affected database.
If there's a particular card or subset of cards causing it _sometimes_, how about a testing strategy like this:
1. Copy the entire deck, call it "test"
2. Remove the bottom half of cards in the test deck
3. Copy "test" to "testB"
4. Open testB in Mnemosyne and run through that deck, giving random grades until either running out of cards or hanging
5. If no hang, quit Mnemosyne
6. Either go to step 3 and repeat the test until you're confident that either the hang happens for these cards or does not happen, or
7. Go back to step 1, performing a new search on either the top half of the original deck (if you couldn't find the hang this time), or the bottom half of the test deck, and keep narrowing it down until you have a deck with only one or a few cards in it which reliably hangs.

Oisín

To post to this group, send email to mnemosyne-...@googlegroups.com.
To unsubscribe from this group, send email to mnemosyne-proj-u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msg/mnemosyne-proj-users/-/_wN4XeWJnAEJ.

Scott Youngman

unread,
Jul 21, 2012, 11:29:45 PM7/21/12
to mnemosyne-...@googlegroups.com
I don't intend to discourage anyone from trying your method, Oisín. But I think the quest for "only one or a few cards ... which reliably hangs" has been tested in other ways, and there is no reliable or predictable pattern. Here's an example. I set autosave after 20 repetitions. In eight successive runs of the program, it hung on this numbered card: 3, 13, 3, 13, 15, 20, 20, 53. Because my progress was not autosaved until it got to card 21, each of the eight runs was performed on the exact same sequence of the same cards numbered 1-20. You can see that although there is some repetition of which cards hung, it is also not regular or predictable. For example, it hung on card 3 twice, but passed it successfully six times (25% failure on card 3). It hung on card 13 twice, but passed it successfully four times (2/6 or 33% failure on card 13). Lack of a repeated, predictable pattern is also evident in the results Chris has posted. So even if one could go through an entire deck as you describe and identify one or more cards which "reliably hang," running the test again would very likely identify a different set of "reliably hanging" cards. Or to say that another way, just because a hang happens on a certain card, it may not to hang on that card in a different run. I'm just speaking from frequent experience with this bug for the last five months, beginning with beta 10: https://groups.google.com/forum/?fromgroups#!topic/mnemosyne-proj-users/TcNvXsGSEnM

I think there are two other indications that the problem is not in the cards themselves. First, Peter has examined decks from some users who have experienced the problem, and he hasn't reported anything wrong in their structure. Second, decks which hang in one person's computer don't hang in another user's computer.


Peter Bienstman

unread,
Jul 22, 2012, 1:34:25 AM7/22/12
to mnemosyne-...@googlegroups.com
Quoting Chris <the.whis...@gmail.com>:

> I'm actually willing to buy a new computer for the sole purpose of
> getting this program working.

Don't go there :-)

Peter


Chris

unread,
Jul 22, 2012, 2:47:32 AM7/22/12
to mnemosyne-...@googlegroups.com
Peter!  I froze another database!  It was a bit of a marathon, going through about 600 cards.  The second time, about 75.  I know it might be an annoying key-mashing prospect, but maybe try going through one or two of the databases were sent by victims to 1000+ cards? 

  - Chris Shanks

Chris

unread,
Jul 22, 2012, 3:04:48 AM7/22/12
to mnemosyne-...@googlegroups.com
My regular database is being really obnoxious tonight, freezing repeatedly on the last cards I have in the queue to learn. But I just got this strange error. No idea if it has anything to do with anything, but I'm taking stabs in the dark now.  See the attached picture.

Come to think of it, way back when I first installed 2.0, I got some sort of Runtime error the first couple times I tried, and the program wouldn't run.  I just reinstalled and it worked.  I believe I saw someone else report something similar to this somewhere else on the boards.  Once again, probably has nothing to do with it.

 - Chris Shanks
Mnemosyne error.png

querido

unread,
Jul 22, 2012, 9:35:20 AM7/22/12
to mnemosyne-...@googlegroups.com
Haha. But what about my "remote desktop" idea?
 

On Sunday, July 22, 2012 1:34:25 AM UTC-4, Peter Bienstman wrote:
Quoting Chris
 
> I'm actually willing to buy a new computer for the sole purpose of  

Oisín

unread,
Jul 22, 2012, 9:48:05 AM7/22/12
to mnemosyne-...@googlegroups.com
On 22 July 2012 04:29, Scott Youngman <syo...@gmail.com> wrote:
I think there are two other indications that the problem is not in the cards themselves. First, Peter has examined decks from some users who have experienced the problem, and he hasn't reported anything wrong in their structure. Second, decks which hang in one person's computer don't hang in another user's computer.

Ah, okay. Well that's a shame, as it makes it so much harder to debug. Are you saying then that even if you create a blank database with 10 dummy cards, Mnemosyne will eventually hang even on that deck? Or only decks with some picture/audio cards, etc?

Have you tried using a tool like Process Hacker (replacement for Windows task manager) to examine the hung Mnemosyne process?
If you get it to hang, then double-click the Mnemosyne process in Process Hacker and go to the "threads" tab, you'll see a list of the different threads and can look at the call stack of each one.

I'm not sure if you reported whether the hang was a CPU-hogging infinite loop, or a deadlock or some other bad blocking situation, but this could help to find out. You might find the main thread is waiting on a call to Qt or sqlite which never completes?

The call traces aren't too long, so I'm sure it would be helpful to post them here.

Oisín

Chris

unread,
Jul 22, 2012, 5:48:10 PM7/22/12
to mnemosyne-...@googlegroups.com
I installed process hacker and watched it freeze.  I have no idea what any of this means, or how to report it here.  Nothing really seemed to change in the threads tab, except CPU usage went up to ~50%.  I discovered if I double clicked on any of the threads, this would further bring up some sort of stack of those threads, but they were constantly changing and still working on something, so I don't know how to report them.  Attached are a bunch of pictures during a freeze.  No idea if any of them are useful.  Like I say, keep in mind in the pictures those secondary "STACK - THREAD XXXX" windows were constantly changing, so the 4 print screens I took of each of the TID entries in the window behind were not at the exact same moment in time, so they may not have all been functioning at the same time.  As Will Ferrel once said when describing The Matrix, "Ergo, concordantly, vis-a-vis. You know what? I have no idea what the hell I'm saying. "

Chris

unread,
Jul 22, 2012, 5:49:29 PM7/22/12
to mnemosyne-...@googlegroups.com
Process Hacker 1.png
Process Hacker 2.png
Process Hacker 3.png
Process Hacker 4.png
Process Hacker 5.png

Oisín

unread,
Jul 22, 2012, 11:59:20 PM7/22/12
to mnemosyne-...@googlegroups.com
On 22 July 2012 22:49, Chris <the.whis...@gmail.com> wrote:


On Sunday, July 22, 2012 2:48:10 PM UTC-7, Chris wrote:
I installed process hacker and watched it freeze.  I have no idea what any of this means, or how to report it here.  Nothing really seemed to change in the threads tab, except CPU usage went up to ~50%.  I discovered if I double clicked on any of the threads, this would further bring up some sort of stack of those threads, but they were constantly changing and still working on something, so I don't know how to report them.  Attached are a bunch of pictures during a freeze.  No idea if any of them are useful.

Great! The main mnemosyne thread stacktrace is most interesting - it suggests that the sqlite database driver may be running out of memory (Sqlite returns NOMEM if it can't dynamically allocate memory when processing a statement)?

I'm not sure why sqlite would run out of memory even on small test decks, but hopefully Peter will be able to track this down. It'd be nice if there was a log of the SQL statements Mnemosyne tries to execute - this could be a bug in the database driver, but there might be a workaround.

Oisín


Like I say, keep in mind in the pictures those secondary "STACK - THREAD XXXX" windows were constantly changing, so the 4 print screens I took of each of the TID entries in the window behind were not at the exact same moment in time, so they may not have all been functioning at the same time.  As Will Ferrel once said when describing The Matrix, "Ergo, concordantly, vis-a-vis. You know what? I have no idea what the hell I'm saying. "



On Sunday, July 22, 2012 6:48:05 AM UTC-7, Oisín Mac Fhearaí wrote:


On 22 July 2012 04:29, Scott Youngman <syo...@gmail.com> wrote:
I think there are two other indications that the problem is not in the cards themselves. First, Peter has examined decks from some users who have experienced the problem, and he hasn't reported anything wrong in their structure. Second, decks which hang in one person's computer don't hang in another user's computer.

Ah, okay. Well that's a shame, as it makes it so much harder to debug. Are you saying then that even if you create a blank database with 10 dummy cards, Mnemosyne will eventually hang even on that deck? Or only decks with some picture/audio cards, etc?

Have you tried using a tool like Process Hacker (replacement for Windows task manager) to examine the hung Mnemosyne process?
If you get it to hang, then double-click the Mnemosyne process in Process Hacker and go to the "threads" tab, you'll see a list of the different threads and can look at the call stack of each one.

I'm not sure if you reported whether the hang was a CPU-hogging infinite loop, or a deadlock or some other bad blocking situation, but this could help to find out. You might find the main thread is waiting on a call to Qt or sqlite which never completes?

The call traces aren't too long, so I'm sure it would be helpful to post them here.

Oisín

--
You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group.
To post to this group, send email to mnemosyne-...@googlegroups.com.
To unsubscribe from this group, send email to mnemosyne-proj-u...@googlegroups.com.

Peter Bienstman

unread,
Jul 23, 2012, 3:26:52 AM7/23/12
to mnemosyne-...@googlegroups.com
On Sunday, July 22, 2012 12:04:48 AM Chris wrote:
> My regular database is being really obnoxious tonight, freezing repeatedly
> on the last cards I have in the queue to learn. But I just got this strange
> error. No idea if it has anything to do with anything, but I'm taking stabs
> in the dark now. See the attached picture.
>
> Come to think of it, way back when I first installed 2.0, I got some sort
> of Runtime error the first couple times I tried, and the program wouldn't
> run. I just reinstalled and it worked. I believe I saw someone else
> report something similar to this somewhere else on the boards. Once again,
> probably has nothing to do with it.

This is the kind of bug I like: a bug inside the Mnemosyne code itself, which
therefore gives a nice stack trace which makes it easy to identify the error
:-)

Your 'save after x reps' is probably set to 0. I'll modify the code to prevent
the spinner from going down all the way to 0.

I doubt this has anything to do with the other hangs though.

Peter

Peter Bienstman

unread,
Jul 23, 2012, 3:44:06 AM7/23/12
to mnemosyne-...@googlegroups.com
Oisín, thanks for pointing out process hacker tool!

Chris, how much memory do you have available if Mnemosyne starts to hang?
Process hacker will probably also give you this type of information.

I did some googling around and found this thread:
http://www.mail-archive.com/sqlite...@sqlite.org/msg65666.html

Here's a test version with a reduced cache size:

http://users.ugent.be/~pbienst/pub/mnemosyne-2.0.1-test3-setup.exe

Peter

Scott Youngman

unread,
Jul 23, 2012, 1:06:33 PM7/23/12
to mnemosyne-...@googlegroups.com
I uninstalled test2, then installed test3 (Win7-64). Still hangs essentially the same as before. On five successive runs, it successfully showed 1, 4, 14, 10, and 4 cards (hanging on the next card). Autosave after 20 cards, so those runs were on the same cards in the same order each time. When Mnemosyne was hung, CPU usage was 28-30%, compared to 0-6% without Mnemosyne. Memory usage was 38% either when hung or not hung (4 Gb RAM). The cards reviewed did not include any images or sounds; total deck size 1,170 cards. Hang occurs when Show Answer is requested. I'll try it again with Process Hacker in the next day or so.

Peter Bienstman

unread,
Jul 23, 2012, 1:11:37 PM7/23/12
to mnemosyne-...@googlegroups.com
Looks like I'll have to update the sqlite version anyhow. This will
require a change in Python middleware however, so it could take some
time.

Peter
> --
> You received this message because you are subscribed to the Google
> Groups "mnemosyne-proj-users" group.
> To post to this group, send email to mnemosyne-...@googlegroups.com.
> To unsubscribe from this group, send email to
> mnemosyne-proj-u...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/mnemosyne-proj-users/-/bKJmw7JPNsgJ.

Chris

unread,
Jul 24, 2012, 12:45:33 AM7/24/12
to mnemosyne-...@googlegroups.com
On Monday, July 23, 2012 12:44:06 AM UTC-7, Peter Bienstman wrote:
Oisín, thanks for pointing out process hacker tool!

Chris, how much memory do you have available if Mnemosyne starts to hang?
Process hacker will probably also give you this type of information.

I'm not exactly sure what that means I should be looking for in process hacker.  Oisín, if you know the layout of process hacker, could you let me know where to look for this information Peter has requested?

   - Chris

Peter Bienstman

unread,
Jul 24, 2012, 1:41:44 AM7/24/12
to mnemosyne-...@googlegroups.com
In your screenshot of process hacker, I saw a tab called 'memory'.
That should give you the info.

Cheers,

Peter
> --
> You received this message because you are subscribed to the Google
> Groups "mnemosyne-proj-users" group.
> To post to this group, send email to mnemosyne-...@googlegroups.com.
> To unsubscribe from this group, send email to
> mnemosyne-proj-u...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/mnemosyne-proj-users/-/h0Ka98FrhQsJ.

Chris

unread,
Jul 24, 2012, 1:55:33 AM7/24/12
to mnemosyne-...@googlegroups.com
If I remember right, when I checked that graph, nothing seemed to change on freeze; it stayed steady at about 50%.  I'll check again tomorrow when I'm reviewing my cards to be sure.

  - Chris
> To post to this group, send email to mnemosyne-proj-users@googlegroups.com.
> To unsubscribe from this group, send email to  
> mnemosyne-proj-users+unsub...@googlegroups.com.

Oisín

unread,
Jul 24, 2012, 10:07:20 AM7/24/12
to mnemosyne-...@googlegroups.com
On 23 July 2012 18:06, Scott Youngman <syo...@gmail.com> wrote:
I uninstalled test2, then installed test3 (Win7-64). Still hangs essentially the same as before. On five successive runs, it successfully showed 1, 4, 14, 10, and 4 cards (hanging on the next card). Autosave after 20 cards, so those runs were on the same cards in the same order each time. When Mnemosyne was hung, CPU usage was 28-30%, compared to 0-6% without Mnemosyne. Memory usage was 38% either when hung or not hung (4 Gb RAM). The cards reviewed did not include any images or sounds; total deck size 1,170 cards. Hang occurs when Show Answer is requested. I'll try it again with Process Hacker in the next day or so.


Do you have a 4-core CPU? If you hit the System Information button in the main Process Hacker window, on the CPU tab there's a checkbox called "Show one graph per CPU".
If the overall CPU usage is around 25-30% you'll often see that actually three cores are idle while one core is stuck in a busy loop, near 100% usage.
Also, if you hover over one of the individual core graphs with the mouse, it'll tell you which process was using that core at that time.

I'd suspect that when Mnemosyne hangs, the main thread goes into an infinite loop, either by Mnemosyne repeatedly calling an sqlite function which fails, or by calling an sqlite function once which itself goes into an infinite loop, trying to malloc over and over again.
Maybe it's a bug in the sqlite driver - I wonder why it only affects some machines though. Apparently, some query orderings cause sqlite to load the whole database into memory, but a 1170-card database shouldn't be very large, right? The cache option Peter found seems more likely to help on embedded devices with very little RAM, but your machine has 4 gigs plus virtual memory, which is a ridiculously large amount for most purposes.

Another possibility, to do with SQL cursors in sqlite:
http://supportforums.blackberry.com/t5/Java-Development/out-of-memory-in-sqlite/td-p/1617333

Does it make any difference if you use random order scheduling rather than most urgent first? If we can replicate the conditions that cause this hang, the problem will be 80% solved already.
It would be nice if you could get a log of every SQL statement before Mnemosyne executes it.

Oisín
 

Oisín

unread,
Jul 24, 2012, 10:45:45 AM7/24/12
to mnemosyne-...@googlegroups.com
Chris, just another small test to be sure that another sqlite3.dll is somehow overriding the one supplied with Mnemosyne:
1. Open the Mnemosyne process in Process Hacker and go to the modules tab
2. Find sqlite3.dll, right-click and select "Properties"
3. Check the path to the DLL in "Location".

In my case this is "C:\Program Files (x86)\Mnemosyne" - is it the same in yours?

Peter Bienstman

unread,
Jul 25, 2012, 3:46:10 PM7/25/12
to mnemosyne-...@googlegroups.com
OK, I've ripped out the database layer and replaced it with something more up-to-date. Hopefully, this version will finally solve the Windows hangs!


Peter

Scott Youngman

unread,
Jul 25, 2012, 7:26:09 PM7/25/12
to mnemosyne-...@googlegroups.com
I really hate to say this, but I uninstalled test3, installed test4, and it still hangs. 5 successive runs showed 22, 23, 28, 2, and 106 cards before hanging (the last run included 15 unlearned cards). Autosave after 20 repetitions. sqlite3.dll was running from C:\Program Files (x86)\Mnemosyne. When it hangs, overall CPU use is ~28%, but the individual graphs show mnemosyne is spiking to 98% on any one of the four cores. However, any and all four cores may be used by Mnemosyne while it is hung, including the high spikes of usage. I can send screenshots of the CPU graphs if you want. We'll see what Chris reports.

What else should I look at in Process Hacker?

Scott Youngman

unread,
Jul 25, 2012, 7:30:15 PM7/25/12
to mnemosyne-...@googlegroups.com
... and the final run of 106 cards did not hang.

Peter Bienstman

unread,
Jul 26, 2012, 3:08:39 AM7/26/12
to mnemosyne-...@googlegroups.com
On Wednesday, July 25, 2012 04:26:09 PM Scott Youngman wrote:
> sqlite3.dll was running from C:\Program Files (x86)\Mnemosyne.

Small packaging problem, as it shouldn't rely on sqlite3.dll anymore.

This one should fix it:

http://users.ugent.be/~pbienst/pub/mnemosyne-2.0.1-test5-setup.exe

(although you might still need to delete C:\Program Files
(x86)\Mnemosyne\sqlite3.dll manually after the install to be really sure).

Peter

Murray James Morrison

unread,
Jul 26, 2012, 4:14:39 AM7/26/12
to mnemosyne-...@googlegroups.com
Peter,
 
Should all users be downloading these new versions? I'm not having the freezing problems others describe--are the new versions specifically for them?
 
--murrayjames
--
You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group.
To post to this group, send email to mnemosyne-...@googlegroups.com.
To unsubscribe from this group, send email to mnemosyne-proj-u...@googlegroups.com.

Peter Bienstman

unread,
Jul 26, 2012, 4:20:22 AM7/26/12
to mnemosyne-...@googlegroups.com
On Thursday, July 26, 2012 04:14:39 PM Murray James Morrison wrote:
> Peter,
>
> Should all users be downloading these new versions? I'm not having the
> freezing problems others describe--are the new versions specifically for
> them?

Yes, these are test version to help debug the problem. Other users should not
bother with them.

Peter

Chris

unread,
Jul 26, 2012, 5:53:58 AM7/26/12
to mnemosyne-...@googlegroups.com
When I started with this one, I had a hunch it would be the one to work.  It then proceeded to freeze after about 50 cards, then 3 on my second try.  Process hacker's thread looked quite different.  As I quickly refreshed it as the program was frozen, every few seconds the thread would be really long, then shrink back down.  If I have time tomorrow I will try and post this series of images.

  - Chris S.

Scott Youngman

unread,
Jul 26, 2012, 7:50:34 AM7/26/12
to mnemosyne-...@googlegroups.com
I uninstalled test4 and installed test5. Unfortunately, it is still hanging (on successive runs, it hung after cards numbered 39, 22, 57, 1, 2, 8). Autosave after 20 cards.

Peter Bienstman

unread,
Jul 26, 2012, 7:52:00 AM7/26/12
to mnemosyne-...@googlegroups.com
What does the process hacker stack trace say in this case?

Peter

Quoting "Scott Youngman" <syo...@gmail.com>:

> --
> You received this message because you are subscribed to the Google
> Groups "mnemosyne-proj-users" group.
> To post to this group, send email to mnemosyne-...@googlegroups.com.
> To unsubscribe from this group, send email to
> mnemosyne-proj-u...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/mnemosyne-proj-users/-/2YcAZGxj6d4J.

Scott Youngman

unread,
Jul 26, 2012, 8:12:08 AM7/26/12
to mnemosyne-...@googlegroups.com
I'm having trouble uploading an image to our Google group. I'm sending to you it in an email attachment instead.

Peter Bienstman

unread,
Jul 26, 2012, 8:23:25 AM7/26/12
to mnemosyne-...@googlegroups.com
On Thursday, July 26, 2012 05:12:08 AM Scott Youngman wrote:
> I'm having trouble uploading an image to our Google group. I'm sending to
> you it in an email attachment instead.

OK, I made a trivial change to the code, but with far-reaching consequences,
namely ripping out the entire card browser, which also contains some database
code.

http://users.ugent.be/~pbienst/pub/mnemosyne-2.0.1-test6-setup.exe

To be really sure, delete the contents of C:\program files\Mnemosyne before
install, and make sure that afterwards there is no filename with sql in it
somewhere.

Peter
It is loading more messages.
0 new messages