2.0 sometimes hangs on show answer or grade

[Scott Youngman]

2.0 sometimes hangs (stops responding) on "show answer" or inputting a
grade. The problem usually happens after displaying 3 to 5 cards. The
cause appears to be random: it is not any particular card (an
individual card sometimes works correctly and at other times hangs)
nor card type (all cards are either front-to-back or cloze types). The
program does not recover, so I have to force it to close.

This began with beta 10, but still occurs after reverting to beta 9.
And previously, it never happened with beta 9 so I suspect this is a
problem in Windows, not Mnemosyne. But I wonder if anyone has an idea
of what is going on. (Win 7/64)

[Gnome]

I also use win 7/64 but no such problems.

[Peter Bienstman]

Strange..

Do the cards have sounds, images, or non-standard fonts?

Do you perhaps have a new virus scanner which interferes with the database?

Is that database on a network drive?

You can also try the development snapshot here:

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

It will upgrade your database and perform a sanity check in the process.
(doesn't hurt to backup C:/Users/Scott/Application Data/mnemosyne first,
though).

Peter

[Dan Schmidt]

I've had a few hangs (2 or 3) with beta 9, none yet with beta 10. They
were all while I was just going through my deck, nothing special. I
don't remember whether it was on "show answer" or while transitioning
to the next card. I'm also on 64-bit Windows 7. I didn't bother to
mention it before because I don't have any useful information, but
there it is.

[Dan Schmidt]

And to answer Peter's questions to Scott:
- The majority of my cards have images; none have sounds. A few have
Japanese characters.
- I use Microsoft Security Essentials.
- The database is on c:, just like the program is.

[Scott Youngman]

The cards are text only; no sounds, images, or non-standard fonts.
I am using Microsoft Security Essentials, but that has been a constant
before and after the problem appeared.
The database is in the default location, C:\Users\Scott\AppData\Roaming
\Mnemosyne.

I will try the development version. Thanks, Peter.

[Peter Bienstman]

Strange... If the development version does not work, this might turn out to be
yet another of these frustrating non-reproducible Qt bugs...

Peter

[Scott Youngman]

I've been using the development version for several days. It does not
hang as frequently as beta 10, but still does hang at odd times.

Two observations to add:
1. Excluding the process mnemosyne.exe from Microsoft Security
Essentials did not solve the problem.
2. If I change configuration settings, they are not remembered for the
next session if the program hangs (hence is forced to close). The new
configuration is only remembered if the program exits normally. I
suppose this is normal, but I'm just making the observation and not
suggesting it is related to my problem.

[Peter Bienstman]

On Sunday 26 Feb 2012 05:32:28 Scott Youngman wrote:
> I've been using the development version for several days. It does not
> hang as frequently as beta 10, but still does hang at odd times.
>
> Two observations to add:
> 1. Excluding the process mnemosyne.exe from Microsoft Security
> Essentials did not solve the problem.

Too bad...

> 2. If I change configuration settings, they are not remembered for the
> next session if the program hangs (hence is forced to close). The new
> configuration is only remembered if the program exits normally. I
> suppose this is normal, but I'm just making the observation and not
> suggesting it is related to my problem.

Indeed as expected...

Peter

[Scott Youngman]

Wow. I changed the configuration to autosave after every repetition
(entered a value of 1 in the box), and then was able to go through all
scheduled reviews plus 15 new or failed cards (total of ~60 cards)
without the program hanging once. (Database is in default user folder,
Win 7/64.)

[Peter Bienstman]

Good to know! So that also fixes the issue with map images?

I don't really have an explanation for this, apart from the fact that the
saving of the database causes a small delay which gives the Qt messages more
time to arrive where they should, thereby avoiding a subtle Qt bug..

Peter

[Scott Youngman]

No, the maps image problem is still the same as before. I had big
hopes for that :^|

And there is still some background conflict in reviewing my non-image
cards. Today the program hung after several cards -- say card D
reviewed successfully and card E wouldn't show the answer. I restarted
several times in quick succession, and each time repeated the same
expeience with the exact same cards: D succeeding, then E hanging
before showing the answer. After waiting a longer period before
restarting, the program finally got past card E. But it hung some 10
cards later.

I agree that the problem appears to be related somehow to the save
step. But it is more likely in the _retention_ function rather than in
the actual save, since I increased the frequency of save (or reduced
the retention time).

On Mar 2, 2:35 am, Peter Bienstman <Peter.Bienst...@ugent.be> wrote:
> On Thursday, March 01, 2012 12:37:41 PM Scott Youngman wrote:

> ...

[Peter Bienstman]

OK, so it seems that the workaround is not full-proof then.

This is really frustrating. I hope that the latest version of Qt solves this,
but we cannot upgrade to it yet, because the sound doesn't work there :-(

Peter

--
Peter Bienstman
Ghent University, Dept. of Information Technology
Sint-Pietersnieuwstraat 41, B-9000 Gent, Belgium
tel: +32 9 264 34 46, fax: +32 9 264 35 93
WWW: http://photonics.intec.UGent.be
email: Peter.B...@UGent.be

[Gnome]

Perhaps you should try to run the database on another computer and see
if the problem still occurs. Then you might have a better clue if it's
your database or the program that has issues.

[Scott Youngman]

Thanks for the suggestion. But it seems that would just test the same
beta release and the same database on a different computer -- hence
only testing the influence of the computer (either hardware or OS).

[Gnome]

Your right, but still it could narrow it down for where to look for
problems. Also there could be another program on your computer that
causes the problem. For example both google chrome and mnemosyne uses
web-kit. Here is an list of some of the programs using webkit:
http://trac.webkit.org/wiki/Applications%20using%20WebKit

[Scott Youngman]

That's interesting. Are you suggesting that the mere installation of
another web-kit application could cause a conflict?

I may be able to test on a different computer (although it too has
Chrome installed). In the meantime, I will note any correlation
between Mnemosyne hanging and Chrome running.

[Scott Youngman]

Gnome, I read your discussion at
http://groups.google.com/group/mnemosyne-proj-users/browse_thread/thread/300fd0bfc9978471?hl=en
about how a crash in Google Chrome cause Mnemosyne to crash because
both programs use webkit. Ok, now I understand a bit better what sort
of problem you were referring to in this thread. In my case, the
hanging problem in Mnemosyne is not due to any other program having
problems. But thanks for passing on the advice.

On Mar 4, 3:24 am, Gnome <geirr_n...@hotmail.com> wrote:

[Scott Youngman]

A new observation on this problem, and possibly a new clue to its
cause...

I can usually go for 15 to 30 cards before Mnemosyne hangs (beta 11
development version). Usually I just restart the program and get
moving again. But twice now the program has reopened in an endless
loop wherein a sequence of two cards continues to reappear in the same
order as in the just-prior session(s): the first card gets graded,
then the second card hangs the program -- again and again with the
same two cards. (But a different pair of cards in each of the two days
this has happened.)

This would seem to indicate a problem with one or the other cards.
Today I deleted the first of the two cards, and the loop was broken. I
could then grade and continue past the second of the original two
cards in the loop.

That would seem to indicate that hanging on the second card was
somehow related to the first card. Just for the record, this is the
Cloze-deletion card I deleted. Is there anything in its structure
which would be difficult for the program? Unfortunately I did not note
which two cards were involved in my first experience of the closed
loop. I will watch this in the future.

-------
*Meta-story of the Bible*
1. [Creation] (by God)
2. [Corruption] (by sin)
3. [Rescue] (by Christ)
4. [Restoration] (new heavens & earth)
-------

[Scott Youngman]

Ok, we may be on to something here. I added the above cloze-deletion
card back into my database. (I also deleted the parend symbols to
simplify its structure, but that had no effect on the problem.) I then
observed that the program _always_ hung on this card or its daughters
-- but in either of two patterns. (1) Sometimes it hung on this card
by not accepting a grade (not showing the answer). (2) But if a grade
was accepted for this card, then the next card would _always_ hang.
The only common factor in either case is this particular card. So this
points strongly toward it as the source of the problem -- that is,
something about its structure.

On Mar 6, 3:41 pm, Scott Youngman <syo...@gmail.com> wrote:
> ... Just for the record, this is the
> Cloze-deletion card I deleted. >

[Scott Youngman]

And I should add that during the period when the problematic card was
not in the database (and therefore not being shown), the program never
hung. Granted, that was not very many repetitions. But that is another
way to say that that card was always involved when the program hung,
and never when the program ran properly. I'll continue to watch the
pattern.

[Peter Bienstman]

Hi Scott,

Thanks for your detailed analysis!

I'm afraid we are mixing up two things, though:

1) Mnemosyne keeping on repeating the same two cards. This I should have
solved in the development version last week
2) Mnemosyne sometimes hanging, i.e. not accepting any input

I tried with your cloze card, but I could not see any issues..

From your cloze card, it seems you are using the fast format plugin, though.
Does disabling that plugin have any effect?

Cheers,

Peter

[Scott Youngman]

Sorry, I didn't mean to imply that the endless loop of the same two
cards was part of the problem. It was just the motivation to delete
the first card in order to break the cycle. I think the loop was due
to the program hanging before it could record any change in the
database, hence it restarted from the exact same condition (grades and
ordering of cards for review) as before the crash.

And that is a characteristic which I have not mentioned previously, so
I will generalize it here: if C, D, and E are sequential cards in
review and the program hangs on E, then the last review grade or
status for D will not be registered in the database. The evidence for
that is when the program is restarted, the number of scheduled and not
memorised cards shown in the status bar will revert to the number
shown after card C was reviewed.

(The development version I am using is dated February 23 13:50.)

[Scott Youngman]

Good question about the fast format plugin. But no, disabling it has
no effect on the problem.

I think we are back to square one. Anything I have observed as a
pattern turns out to have exceptions:
- May hang even though Autosave is set to 1.
- The suspect card does not always cause the program to hang in the
pattern described above.

Nevertheless, it seems that the suspect card is very often "nearby"
when the program hangs, though that may be coincidental. Still, it is
notable that it was the first card on multiple restarts today, and the
program always hung on it or the next card over 10-15 restarts. So
that does not seem random. For now I have deleted that card and will
see what happens over the next several days.

[Peter Bienstman]

On Wednesday 07 Mar 2012 08:17:10 Scott Youngman wrote:

> I will generalize it here: if C, D, and E are sequential cards in
> review and the program hangs on E, then the last review grade or
> status for D will not be registered in the database.

That is not unexpected, because of the delayed writes in libmnemosyne.

Peter

[Peter Bienstman]

On Wednesday 07 Mar 2012 08:39:42 Scott Youngman wrote:
> Good question about the fast format plugin. But no, disabling it has
> no effect on the problem.
>
> I think we are back to square one. Anything I have observed as a
> pattern turns out to have exceptions:

I don't think there is any pattern to be discovered here. It's almost
certainly a bug in Qt which seems to pop up in certain circumstances, which
seems to get triggered on your setup, but not really in other people's...

Peter

[Michael Campbell]

Delaying writes (say, until the program exits "normally") for things like preference changes is a pattern I see often, and universally dislike.  What's the upside, really?  I'm getting the impression from previous emails that mnemosyne does exactly this, and I am curious why.

[Peter Bienstman]

The idea is that Mnemosyne already shows you the next question, so
that you can start thinking about it, before it saves your database.
Helps with the perceived speed of the program :-) So, the delay is
just the time it takes to show the question. (or in case of the
hanging, forever...)

Peter

Quoting "Michael Campbell" <michael....@unixgeek.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.
> For more options, visit this group at
> http://groups.google.com/group/mnemosyne-proj-users?hl=en.
>
>

[Scott Youngman]

Peter wrote:
"I don't think there is any pattern to be discovered here. It's almost
certainly a bug in Qt which seems to pop up in certain circumstances,
which seems to get triggered on your setup, but not really in other
people's..."

That is an encouraging point. In further support of the lack of
pattern, I just noticed that my database contains a cloze-type card
with essentally the same structure as the "problematic" card reported
above, yet as far as I can remember, it has not been been in the
immediate sequence of reviews when the program has hung. That all may
be just a matter of coincidence, but like the problem itself, it is
strange.

[Scott Youngman]

I'm not having the hanging problem any more. It appears to have
"solved itself" in beta 11-b. Whew. I'll keep monitoring to confirm
this.

[Peter Bienstman]

Good that it's gone away, although I must confess I have no idea what caused
it.

Thanks for the update!

Peter

[J Clark]

This has started for me (in 11-b) and I'm also on win7. 

The cards have no images/sounds but are latex'd up.  Like Scott I'm noticing that certain cards seem to trigger the issue and that by juggling/deleting/varying-the-pace I can sometimes "push through" a problem card, only to run into another. 

I also notice that after force-closing and reopening the program the problem card is not displayed normally as the next card to review, but appears blank until I edit it, upon which time the question will appear (after closing the edit-card box).  Not sure if that's relevant.

Jeff

[Peter Bienstman]

On Wednesday 25 Apr 2012 17:15:11 J Clark wrote:
> This has started for me (in 11-b) and I'm also on win7.

Problem is that in Scott's case, we never really figured out what caused this
and what fixed it eventually...

Hopefully a new Qt version will help.

Peter

[J Clark]

Yeah, I was just confirming that 11-b wasn't the explicit fix!

[JW]

I'd just like to report that my beta 11b and 2.0 RC1 installations
appear to hang after cloze cards containing at least 2 cloze deletions
are graded at 2 or higher. After the program hangs, Task Manager
indicates that the program is using exactly 50% of the CPU--no more
and no less--until it is force closed.

I am using Windows XP Pro (Service Pack 3) with a Centrino Duo
Processor.

Hope this helps. The cloze deletion plugin is perhaps the most
valuable addition to Mnemosyne yet.

[Peter Bienstman]

Feel free to mail me your database file privately, together with a
mention of which card causes problems, so that I can see if I can
reproduce the issue.

Cheers,

Peter

[JW]

It should be in your mailbox now.

On May 12, 2:43 am, "Peter Bienstman" <Peter.B...@UGent.be>
wrote:

> Feel free to mail me your database file privately, together with a
> mention of which card causes problems, so that I can see if I can
> reproduce the issue.
>
> Cheers,
>
> Peter
>
> Quoting JW <digita...@gmail.com>:
>
>
>
>
>
>
>
> > I'd just like to report that my beta 11b and 2.0 RC1 installations
> > appear to hang afterclozecards containing at least 2clozedeletions
> > are graded at 2 or higher. After the program hangs, Task Manager
> > indicates that the program is using exactly 50% of the CPU--no more
> > and no less--until it is force closed.
>
> > I am using Windows XP Pro (Service Pack 3) with a Centrino Duo
> > Processor.
>

> > Hope this helps. Theclozedeletion plugin is perhaps the most

[Peter Bienstman]

I've done about 50 reviews on an XP system using your database without any
problems...

Have you tried uninstalling Mnemosyne completely and then reinstalling it?

If that doesn't fix it (and it also does not fix Henrik's crashes), I will have
to take more drastic measures, like rebuilding Mnemosyne with the latest
version of Qt and see if that helps.

However, this will break sound support and therefore I would have to implement
a new sound system, so I'd prefer to only take that route once all the other
options are exhausted.

Cheers,

Peter

[JW]

Yes, I have uninstalled the program and deleted the Mnemosyne folder
in the Documents and Settings directory. After reinstallation, the
program still presents the same problem.

On May 14, 2:36 am, Peter Bienstman <Peter.Bienst...@ugent.be> wrote:
> I've done about 50 reviews on an XP system using your database without any
> problems...
>
> Have you tried uninstalling Mnemosyne completely and then reinstalling it?
>
> If that doesn't fix it (and it also does not fix Henrik's crashes), I will have
> to take more drastic measures, like rebuilding Mnemosyne with the latest
> version of Qt and see if that helps.
>
> However, this will break sound support and therefore I would have to implement
> a new sound system, so I'd prefer to only take that route once all the other
> options are exhausted.
>
> Cheers,
>
> Peter
>
> On Sunday, May 13, 2012 04:36:46 PM you wrote:
>
>
>
>
>
>
>
> > It should be in your mailbox now.
>

> > On May 12, 2:43 am, "Peter Bienstman" <Peter.Bienst...@UGent.be>

[donald...@gmail.com]

On Wednesday, February 22, 2012 9:12:21 PM UTC-7, Scott Youngman wrote:
> 2.0 sometimes hangs (stops responding) on "show answer" or inputting a
> grade. The problem usually happens after displaying 3 to 5 cards. The
> cause appears to be random: it is not any particular card (an
> individual card sometimes works correctly and at other times hangs)
> nor card type (all cards are either front-to-back or cloze types). The
> program does not recover, so I have to force it to close.
>
> This began with beta 10, but still occurs after reverting to beta 9.
> And previously, it never happened with beta 9 so I suspect this is a
> problem in Windows, not Mnemosyne. But I wonder if anyone has an idea
> of what is going on. (Win 7/64)

After upgrading to 2.0, I seem to be having the exact problem described here. This is on Win XP and Win 7 both. I can't detect a pattern.

[Peter Bienstman]

On Tuesday, June 19, 2012 04:44:01 PM donald...@gmail.com wrote:

> After upgrading to 2.0, I seem to be having the exact problem described
> here. This is on Win XP and Win 7 both. I can't detect a pattern.

Hopefully, a future version of Qt fixes this problem...

Peter

[Scott Youngman]

Just to update... the problem disappeared for me in beta 11 but has reappeared in the Release Candidates and 2.0. Sometimes I can go 50 cards without a hang, but at other times it will hang after only a card or two. It may hang repeatedly on the same card, but later pass that particular card and hang on a different card. There is no pattern. Let's hope for a fix in Qt.