PyChess dumps core on Ubuntu 8.10 (stable)

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Load a game or press "start game" on start screen
2. Client freezes before even seeing the board
3.

What is the expected output? What do you see instead?
I expect to see a chess board. What I see instead is nothing.

Does it happen every time?
Yes.

What version of the product are you using?
"Philidor 0.82". I am running on Archlinux with an english-localized GNOME.

Did you use an installed version of PyChess or did you run in from a
tarball/svn?
Tried downloading first from website, and then from distro's repository
with same result

Please provide any additional information below.
This is the first time I tried pychess so it might be some package I am
missing etc. 

Please attach the latest pychess logfile.
13:08:48 Default Debug: Started

Original issue reported on code.google.com by regns...@gmail.com on 4 Aug 2008 at 11:11

[GoogleCodeExporter]

I have the same problem, 
Distro debian lenny, updated,p
Python 2.5.2
and both in 0.8.2 and SVN pychess

Original comment by kave...@gmail.com on 6 Aug 2008 at 9:51

[GoogleCodeExporter]

This is strange - and you get no errors in neither the logfile nor the terminal?

Original comment by lobais on 28 Aug 2008 at 11:27

[GoogleCodeExporter]

Original comment by lobais on 17 Oct 2008 at 2:16

• Changed state: Need-Info

[GoogleCodeExporter]

I would like to confirm this.

Ubuntu Linux 8.10 64bit Beta
Python 2.5.2
Pychess Philidor 0.8.2

No output in the Log Viewer, Pychess just stalls.

Tried Starting a new game from the "start view" as well as from the menu ("New
game...") with no success.

Any hints on where to look to provide more helpful information?

Original comment by jens.kuelpmann@gmail.com on 22 Oct 2008 at 8:26

[GoogleCodeExporter]

Hi kuelpmann, thanks for your response.

Can you try to run the developing version and see if it as well stalls?
http://code.google.com/p/pychess/source/checkout

If it does, I'll try to come up with some ideas to pinpoint the problem.
Thanks, Thomas

Original comment by lobais on 22 Oct 2008 at 9:08

[GoogleCodeExporter]

Running the development version resulted in a slightly different behaviour.

Pychess still stalled, but at some earlier point - only one small window got 
drawn
without even the menus. The "start screen" was not drawn. Also my Gnome session 
was
affected, new windows were drawn "empty" without any controls and the already 
open
ones became unresponsive. I had to restart the X server.

I then disabled compiz/visual effects and switched from murrine 
gtk-controls-theme to
clearlooks (ubuntu standard).

Starting Pychess again then resulted in a crash, without affecting the Gnome 
session.

Terminal:

jens@ubuntu-notebook:~$ cd svn/pychess-read-only/
jens@ubuntu-notebook:~/svn/pychess-read-only$ PYTHONPATH=lib/ ./pychess
/home/jens/svn/pychess-read-only/lib/pychess/System/uistuff.py:365: GtkWarning:
GtkSpinButton: setting an adjustment with non-zero page size is deprecated
  self.glade = gtk.glade.XML(addDataPrefix("glade/%s" % filename))
/home/jens/svn/pychess-read-only/lib/pychess/System/uistuff.py:365: Warning:
IA__g_object_newv: object class `GtkWindow' has no property named `climb_rate'
  self.glade = gtk.glade.XML(addDataPrefix("glade/%s" % filename))
/home/jens/svn/pychess-read-only/lib/pychess/System/uistuff.py:365: Warning:
IA__g_object_newv: object class `GtkWindow' has no property named `digits'
  self.glade = gtk.glade.XML(addDataPrefix("glade/%s" % filename))
/home/jens/svn/pychess-read-only/lib/pychess/System/uistuff.py:365: Warning:
IA__g_object_newv: object class `GtkWindow' has no property named `numeric'
  self.glade = gtk.glade.XML(addDataPrefix("glade/%s" % filename))
/home/jens/svn/pychess-read-only/lib/pychess/System/uistuff.py:365: Warning:
IA__g_object_newv: object class `GtkWindow' has no property named 
`update_policy'
  self.glade = gtk.glade.XML(addDataPrefix("glade/%s" % filename))
/home/jens/svn/pychess-read-only/lib/pychess/System/uistuff.py:365: Warning:
IA__g_object_newv: object class `GtkWindow' has no property named 
`snap_to_ticks'
  self.glade = gtk.glade.XML(addDataPrefix("glade/%s" % filename))
/home/jens/svn/pychess-read-only/lib/pychess/System/uistuff.py:365: Warning:
IA__g_object_newv: object class `GtkWindow' has no property named `wrap'
  self.glade = gtk.glade.XML(addDataPrefix("glade/%s" % filename))
/home/jens/svn/pychess-read-only/lib/pychess/System/uistuff.py:365: Warning:
IA__g_object_newv: object class `GtkWindow' has no property named `adjustment'
  self.glade = gtk.glade.XML(addDataPrefix("glade/%s" % filename))
**
Gtk:ERROR:/build/buildd/gtk+2.0-2.14.4/gtk/gtkcontainer.c:2673:IA__gtk_container
_propagate_expose:
assertion failed: (child->parent == GTK_WIDGET (container))
Aborted (core dumped)
jens@ubuntu-notebook:~/svn/pychess-read-only$

I also have attached two logfiles.

Original comment by jens.kuelpmann@gmail.com on 22 Oct 2008 at 11:42

Attachments:

• 2008-10-23_01-19-53.log
• 2008-10-23_01-27-40.log

[GoogleCodeExporter]

Ah, I forgot trunk wasn't runnable with a clean install :P
It should be fixed now. Please try svn update and run it again.
Thanks for your time, Thomas

Original comment by lobais on 23 Oct 2008 at 8:23

[GoogleCodeExporter]

There still seems to be some problem with the trunk version and the ui - logfile
enclosed.

Regards
Jens

Original comment by jens.kuelpmann@gmail.com on 24 Oct 2008 at 5:20

Attachments:

• 2008-10-24_19-08-36.log

[GoogleCodeExporter]

These are just warnings. Do you also encounter actual problems when trying to 
use the
application?

Original comment by lobais on 24 Oct 2008 at 7:41

[GoogleCodeExporter]

Sorry, I forgot to mention those ... Pychess got as far as the very first small
window, no menu, then stalled. And it messed up Gnome badly again, as described 
in
comment 6.

Thanks,
Jens

Original comment by jens.kuelpmann@gmail.com on 24 Oct 2008 at 8:00

[GoogleCodeExporter]

I'm sorry to hear that.
If you start it up three times, will it stall the same place every time, with 
the
same errors?

Original comment by lobais on 24 Oct 2008 at 8:47

[GoogleCodeExporter]

Just started it three times in a row:

1. Stalled after having drawn the first small window without menus. One 
GTK-related
warning in terminal. Gnome messed up, had to relog/restart X.

2. Stalled seemingly at the same place as in 1. but lots of GTK-related 
warnings in
terminal. Gnome not affected, able to force-quit Pychess and head for 3. time.

3. Stalled like in 1.

I attached the logs.

Thanks,
Jens

Original comment by jens.kuelpmann@gmail.com on 24 Oct 2008 at 9:11

Attachments:

• 2008-10-24_22-59-17.log
• 2008-10-24_23-02-07.log
• 2008-10-24_23-02-50.log

[GoogleCodeExporter]

Hi Jens, thanks a lot for your time.
I have tried ubuntu 8.10 (only 32bit though), but haven't been able to 
reproduce.
Your problem looks slightly like something we fought with pre 0.8, with 
threading. At
that time it helped to try the tests on:
http://www.pardon-sleeuwaegen.be/antoon/python/page0.html . They are different
implementations of the same application. An application, that just shows some 
bars
flickering.
If it is possible, I'd be interested to know which of these, if any, 
implementations
that freeze your system.

Thanks,
Thomas

Original comment by lobais on 27 Oct 2008 at 7:04

[GoogleCodeExporter]

demo1a.py - started/stopped all threads consecutively and changed their mode; 
runs fine

demo1b.py - crashes shortly after starting any thread, no freeze
jens@ubuntu-notebook:~/Desktop$ python demo1b.py
/home/jens/.themes/Murrine Wise/gtk-2.0/gtkrc:107: Murrine configuration option
"hilight_ratio" will be deprecated in future releases. Please update this theme 
to
get rid of this warning.
python: ../../src/xcb_lock.c:77: _XGetXCBBuffer: Assertion `((int) ((xcb_req) -
(dpy->request)) >= 0)' failed.
Aborted

demo1c.py - freezes or crashes shortly after starting any thread, got different
messages in the console on several runs; program freeze does not affect Gnome 
session
as with Pychess (force quitting possible)
jens@ubuntu-notebook:~/Desktop$ python demo1c.py
/home/jens/.themes/Murrine Wise/gtk-2.0/gtkrc:107: Murrine configuration option
"hilight_ratio" will be deprecated in future releases. Please update this theme 
to
get rid of this warning.
python: ../../src/xcb_lock.c:77: _XGetXCBBuffer: Assertion `((int) ((xcb_req) -
(dpy->request)) >= 0)' failed.
Aborted
jens@ubuntu-notebook:~/Desktop$ python demo1c.py
/home/jens/.themes/Murrine Wise/gtk-2.0/gtkrc:107: Murrine configuration option
"hilight_ratio" will be deprecated in future releases. Please update this theme 
to
get rid of this warning.
The program 'demo1c.py' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadIDChoice (invalid resource ID chosen for this connection)'.
  (Details: serial 2544 error_code 14 request_code 155 minor_code 4)
  (Note to programmers: normally, X errors are reported asynchronously;
   that is, you will receive the error a while after causing it.
   To debug your program, run it with the --sync command line
   option to change this behavior. You can then get a meaningful
   backtrace from your debugger if you break on the gdk_x_error() function.)
jens@ubuntu-notebook:~/Desktop$ python demo1c.py
/home/jens/.themes/Murrine Wise/gtk-2.0/gtkrc:107: Murrine configuration option
"hilight_ratio" will be deprecated in future releases. Please update this theme 
to
get rid of this warning.
python: ../../src/xcb_lock.c:33: _XCBUnlockDisplay: Assertion
`xcb_get_request_sent(dpy->xcb->connection) == dpy->request' failed.
Aborted

demo2.py - see demo1a.py

demo3a.py - freezes shortly after starting any thread; unlike demo1c.py the 
started
thread's "counter" does not stop but goes on, although kind of sluggish and 
taking a
lot of cpu power. Program gets unresponsive though, no clicking of buttons 
possible
any more. Force quitting possible, Gnome session unaffected.
jens@ubuntu-notebook:~/Desktop$ python demo3a.py
/home/jens/.themes/Murrine Wise/gtk-2.0/gtkrc:107: Murrine configuration option
"hilight_ratio" will be deprecated in future releases. Please update this theme 
to
get rid of this warning.
Killed

demo3b.py - see demo1a.py; while clicking the buttons the "thread counters" got 
quite
slow sometimes for a short period of time while the program consumed lots of 
cpu power 

-- This was a totally relaxing experience, a bit like watching my favourite
screensaver :)

Thanks,
Jens

Original comment by jens.kuelpmann@gmail.com on 27 Oct 2008 at 8:53

[GoogleCodeExporter]

Hi Jens, I'm sorry about the much too later answer, but while I've been a bit 
busy,
I've thought a lot about how we could debug the problem. I didn't find any good
solutions though.
So eventually I tried to download the ubuntu 8.10 cd myself. I didn't really 
think it
would help anything - these errors tend to be very hard to reproduce. However 
just a
moment ago my doubt was made shame of, and I see the problem with my own eyes. 
I'll
see if I can get it fixed soon, but I'm a bit busy the next two weeks as well, 
so it
depends on its strength :)

Thanks a lot,
Thomas

Original comment by lobais on 14 Nov 2008 at 10:16

• Changed title: PyChess dumps core on Ubuntu 8.10 (stable)
• Changed state: Accepted
• Added labels: Priority-Critical
• Removed labels: Priority-Medium

[GoogleCodeExporter]

Well, should be fixed in rev 1270. Please verify.

Original comment by lobais on 14 Nov 2008 at 11:00

[GoogleCodeExporter]

I upgraded to 1270 pychess_read_only on the Ubuntu 8.10 64bit system and the 
problem
still persists. I have got a 32bit Ubuntu 8.10 system, too,  where I will try 
the SVN
version tomorrow and post the result here as well.

Logfile attached, but there's almost nothing in it. Didn't get anything besides 
some
gtk-related warnings on the console also, before Gnome stalled and I had to
strg+alt+backspace.

Thanks,
Jens

Original comment by jens.kuelpmann@gmail.com on 15 Nov 2008 at 9:21

Attachments:

• 2008-11-15_22-07-49.log

[GoogleCodeExporter]

That warning shouldn't mean anything. I get that one too.
But you say, that the program still freezes or crashes in the exact same spot?

Original comment by lobais on 15 Nov 2008 at 9:36

[GoogleCodeExporter]

Yes it still freezes after drawing a small window in the left upper corner of 
the
screen partly (just the frame without menu or contents) and messes up Gnome 
(other
windows become unresponsive, new windows have no menus/contents either). The 
freeze
occurs before the main pychess window is drawn.

I also tried the 1270 pychess_read_only SVN version on Ubuntu 8.10 32bit and it 
ran
just fine. So it looks like this may only affect 64bit systems.

I could see the very small window with the working version on the 32bit system, 
too,
but only very briefly. It is then replaced by or becomes the pychess main window
which appears centered on the screen.

I've got Eclipse with PyDev installed on the 64bit system since I use it for
development (haven't got any experience with PyGTK though). Would it be of any 
help
to run pychess from the PyDev debugging mode?

Original comment by jens.kuelpmann@gmail.com on 16 Nov 2008 at 8:16

[GoogleCodeExporter]

Running the code in a debugger is probably hard, when you don't have an idea on 
where
to put breakpoints.

Can I ask you have many cores you have in your 64 and 32 bit machines?
Is the 64bit machine running ubuntu beta or stable?
Also, did you remember to remove all files and directories in your homedirectory
starting with .pychess?

Original comment by lobais on 16 Nov 2008 at 10:01

[GoogleCodeExporter]

> Can I ask you have many cores you have in your 64 and 32 bit machines?
64bit system:
jens@ubuntu-notebook:~$ sudo cat /proc/cpuinfo | grep "model name"
model name  : Intel(R) Core(TM)2 Duo CPU     T8100  @ 2.10GHz
model name  : Intel(R) Core(TM)2 Duo CPU     T8100  @ 2.10GHz

Will post specs of the 32bit system later, when I am at that pc. Almost sure 
it's a
dual core cpu also, though.

> Is the 64bit machine running ubuntu beta or stable?
jens@ubuntu-notebook:~$ sudo uname -a
Linux ubuntu-notebook 2.6.27-8-generic #1 SMP Thu Nov 6 17:38:14 UTC 2008 x86_64
GNU/Linux
jens@ubuntu-notebook:~$ sudo cat /etc/lsb-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=8.10
DISTRIB_CODENAME=intrepid
DISTRIB_DESCRIPTION="Ubuntu 8.10"
That is the "stable" version, out since end of october

> Also, did you remember to remove all files and directories in your 
homedirectory
starting with .pychess?
I had no idea that you had to do that. Tried it right away and got a bit further
before pychess stalled while detecting the available chess engines. No messing 
up of
Gnome either, I could simply force quit it and return to desktop. I attached a
screenshot and the log.

I also tried the ubuntu-provided Philidor 0.8.2 again with having deleted
/home/.pychess beforehand. The startup process of the two versions seems to be
different. But it is still impossible to start a game. I attached a screen and 
the
log as well.

Thanks,
Jens

Original comment by jens.kuelpmann@gmail.com on 17 Nov 2008 at 7:27

Attachments:

• Screenshot1.png
• 2008-11-17_20-20-31.log
• Screenshot2.png
• 2008-11-17_20-22-57.log

[GoogleCodeExporter]

I just realized there was a file named .pychessconf in my home directory also. 
By
deleting both the .pychess folder and the .pychessconf file I could reproduce 
the
behaviour where only one small window is drawn and then pychess stalls messing 
up the
Gnome session. Had to strg-alt-backspace, so no screen. The log just has a
GTK-warning in it.

BTW: cpu in the 32bit system is an Intel(R) Xeon(R) 3050 @ 2.13GHz, two cores.

Thanks,
Jens

Original comment by jens.kuelpmann@gmail.com on 17 Nov 2008 at 7:54

[GoogleCodeExporter]

I also get this problem on currently updated ubuntu 32bit intrepid, version 
0.8.2-1.
I start pychess and the start window appears but it hangs after selecting "start
game". Usually it hangs then but I've seen a few times where the board will 
appear
but hang later, or the board will start to spin around for black but hang 
during the
spin. I have a dual core laptop.

Problem occurs slightly differently in the current svn version. The start window
appears then the "discovering chess engines" window appears and then that window
hangs after the 3 chess engines are discovered (originally had crafty installed 
and
it hung after the 4th engine but after deleting crafty it made no difference). 
I've
deleted ~/.pychess* but that makes no difference.

I have a standard laptop with a standard ubuntu install so I'm surprised more 
people
are not reporting this drastic bug.

Original comment by blakeney...@gmail.com on 30 Dec 2008 at 3:07

[GoogleCodeExporter]

Should also not that this problem has also hung my ubuntu gnome session a 
couple of
times. Once, a "killall nautilus" even failed to recover it and I had to reboot 
my
laptop.

Original comment by blakeney...@gmail.com on 30 Dec 2008 at 3:11

[GoogleCodeExporter]

I know this kind of problem from older version of PyChess, but I haven't been 
able to
barely reproduce it on any Ubuntu/other system, that I've been able to get to.

I hope, that it will just magically vanish, or it we will be able to better 
trace it,
once more people gets their hands on Staunton.

Thanks for all your time helping to solve this issue.

Original comment by lobais on 9 Jan 2009 at 10:34

[GoogleCodeExporter]

I am getting the same problem, it just completely stalls after hitting "start 
game."

I'm running Ubuntu Intrepid 32 bit.

Original comment by dobner...@gmail.com on 4 Mar 2009 at 8:46

[GoogleCodeExporter]

What is the status of this issue?
Do you need any testing, debug info?

Original comment by ldr...@gmail.com on 14 Mar 2009 at 8:13

[GoogleCodeExporter]

I haven't been able to reproduce this on any of my own Ubuntu systems.

Rereading this thread I can see different issues, and I'm not sure if different 
people 
are getting different issues, or if people with an issue get a combination:

1) Starting PyChess displays a small window with no content, and occasinally 
crashes X.
2) PyChess hangs during engine discovery. No feedback, but not frozen and fully
closeable.
3) PyChess goes through engine discovery, but then hangs during start game. 
Perhaps
only people who have encountered (2) and restarts PyChess get this..
4) PyChess fails to start and write 'Aborted (core dumped)' to the terminal. 
I'm not
sure if anybody is still seeing this.

Original comment by lobais on 15 Mar 2009 at 7:55

[GoogleCodeExporter]

ok, so could you please define what do you need exactly?

For begining, here's what I have:

Ubuntu 8.10 amd64
proc: amd turion 64 x2 (TL-58)
PyChess version: pychess Philidor 0.8.2 (from ubuntu rep.)

After starting pychess from Application menu everything is ok, pychess starts 
and I
get main window, menus are working(about,log viewer).
After hitting 'Start game' main window hangs and nothing happens. Content of 
pychess
window disappears - probably it's not refreshed. After killing and restarting 
app
situations looks in the same way. I don't get anything from stdout in console.
After rerunning with --gst-debug=5 I've got quite big debug file which I've 
attached
to this comment.

Original comment by ldr...@gmail.com on 29 Mar 2009 at 3:42

Attachments:

• debug.out.bz2

[GoogleCodeExporter]

Hey, I ve been using PyChess for a while and never had this problem, but now, 
since
some revision 1441 (svn) or so, I cannot make it start. It actually opens and 
hangs
or opens and shows engine discovery screen and then hangs. Often it even hangs 
the
whole gnome so I have to restart X server.

It happens with latest revision 1448 as well as with repository last version now
(which worked fine before). I have several engines in my system, but it had 
worked
fine with them before.

Here are the logs attached (chronologically first is when it hanged without 
"engine
discovery screen" the two others are from one run when it hanged after "engine
discovery screen" showed.

Original comment by pe3...@gmail.com on 4 Apr 2009 at 8:55

Attachments:

• 2009-04-04_10-45-45.log
• 2009-04-04_10-51-35.log
• 2009-04-04_10-51-40.log

[GoogleCodeExporter]

This happens every time?

Do you see any additional output if you run it from a terminal window like this:
$ ~/pychess> PYTHONPATH=lib/ pychess >& ~/Desktop/pychess.log

Original comment by mattgatto on 4 Apr 2009 at 9:30

[GoogleCodeExporter]

I'm getting the same "freeze when clicking start game" issue under Fedora 10.  I
tried Fedora's 0.8.2-1 RPM and the 0.8.4 tarball with the same outcome.

FYI, this still occurs even after doing all of the following:
1. Uninstalling GNUChess.
2. Selecting "Human Being" for the opponent.
3. Unchecking *everything* in the preferences dialog and selecting "No 
animation"
does not help.

Original comment by dean.brettle on 19 Jun 2009 at 11:03

[GoogleCodeExporter]

@comment 32:
Have you tried the 0.10beta1 tarball or running from svn?

Original comment by mattgatto on 19 Jun 2009 at 11:29

[GoogleCodeExporter]

0.10beta1 tarball - hangs while detecting chess engines.
svn - haven't tested yet.

FYI, if I modify the 0.8.4 ionest.py to look like the attached file I can get a 
chess
board to appear but the workerfunc exits and the UI still freezes with the 
chessboard
showing.

Original comment by dean.brettle on 19 Jun 2009 at 1:07

Attachments:

• ionest.py

[GoogleCodeExporter]

After looking through the code, I think you need to identify all methods of 
Human,
BoardControl, and BoardView that manipulate the GUI and could possibly be 
called from
any other class (e.g. GameModel) on the worker thread, and change them to use
idle_add() to perform their function on the gtk thread, as described at:

http://www.pardon-sleeuwaegen.be/antoon/python/page3.html

Original comment by dean.brettle on 19 Jun 2009 at 2:27

[GoogleCodeExporter]

I think I've fixed the "hang while detecting chess engines" issue in revision 
1501.

dean.brettle or anyone else that was experiencing this, could you try running 
svn and
see if it's still hanging?

svn checkout http://pychess.googlecode.com/svn/trunk/ pychess-read-only

Original comment by mattgatto on 20 Jun 2009 at 7:36

[GoogleCodeExporter]

Dean: We turned away from idle_add in 0.6 -> 0.8 because the large use of it 
created some nasty crashes 
elsewhere. We made a test on a lot of peoples machines, and generally the 
thread_enter/thread_leave 
approached turned out most successfull.
I'm not sure which of the two approaches we are going to pull for 1.0, but no 
matter which one, we'll need to 
refactor things to make it much more clear what uses threading and what does 
not.

Original comment by lobais on 20 Jun 2009 at 11:18

[GoogleCodeExporter]

Here's what I see with revision 1501:

First try exited while finding chess engines with the following output:
<quote>
[brettle@localhost pychess-read-only]$ PYTHONPATH=lib/ ./pychess
14:28:05 Default Warning: engineNest: no version attribute found

/home/brettle/src/pychess-svn/pychess-read-only/lib/pychess/System/uistuff.py:36
8:
GtkWarning: GtkSpinButton: setting an adjustment with non-zero page size is 
deprecated
  self.glade = gtk.glade.XML(addDataPrefix("glade/%s" % filename))
Alarm clock
</quote>

Second try hung with:
<quote>
[brettle@localhost pychess-read-only]$ PYTHONPATH=lib/ ./pychess
14:30:15 Default Warning: engineNest: no version attribute found

/home/brettle/src/pychess-svn/pychess-read-only/lib/pychess/System/uistuff.py:36
8:
GtkWarning: GtkSpinButton: setting an adjustment with non-zero page size is 
deprecated
  self.glade = gtk.glade.XML(addDataPrefix("glade/%s" % filename))
</quote>

I tried a bunch more times and the "alarm clock" exit seems to be much more 
common
for me.

Original comment by dean.brettle on 20 Jun 2009 at 9:33

[GoogleCodeExporter]

Issue 424 has been merged into this issue.

Original comment by mattgatto on 21 Jun 2009 at 11:49

[GoogleCodeExporter]

The "Alarm clock" crash, reported in issue 459, should be fixed now with 
revision 1503.

Original comment by mattgatto on 23 Jun 2009 at 10:33

[GoogleCodeExporter]

Upgradaed to revision 1504 and can confirm that the "Alarm clock" crashes have
disappeared, but the hangs still occur while (or sometime before) the 
"discovering
engines" window is on screen.

Also, I'm now running Fedora 11.  Both revision 1504 and 0.8.2 still hang.

Original comment by dean.brettle on 23 Jun 2009 at 11:30

[GoogleCodeExporter]

Thanks for your help Dean. Could you try running this command a bunch of times 
and
see if there's any problems:
~/pychess-read-only$ PYTHONPATH=lib/ python -u lib/pychess/Players/PyChess.py

Type "quit" after the engine finishes initialization to terminate it.

Original comment by mattgatto on 24 Jun 2009 at 12:12

[GoogleCodeExporter]

I always get "feature done=0", no GUI.  I assume that is what you were 
expecting.

Original comment by dean.brettle on 24 Jun 2009 at 12:35

[GoogleCodeExporter]

What happens if you do this:

~/pychess-read-only$ PYTHONPATH=lib/ python -u lib/pychess/Players/PyChess.py 
<<EOF
> xboard
> protover 2
> quit
> <ctrl-d>

Original comment by mattgatto on 24 Jun 2009 at 1:27

[GoogleCodeExporter]

The most recent ~/.pychess/2009-*.log created after each of these command-line 
runs
might have some relevant information, assuming the problem lies in running 
PyChess.py
as a subprocess.

Original comment by mattgatto on 24 Jun 2009 at 1:39

[GoogleCodeExporter]

Log file says:

18:42:13 stdout Debug: feature done=0
18:42:13 stdout Debug: feature draw=1 reuse=0 usermove=1 setboard=1 sigterm=1
analyze=1 variants='normal,nocastle,fischerandom' myname='PyChess 0.10beta1' 
done=1

which corresponds to the terminal output...

Original comment by dean.brettle on 24 Jun 2009 at 1:43

[GoogleCodeExporter]

What mattgatto is testing is wether done=1 appears. It seams like 
PyChess-engine runs
fine in your terminal, but refuses to start as a subprocess under 
PyChess-client.

When PyChess hangs while (or before) the discover-window, I guess your log 
contains
very few lines, but that one of them includes "done=0" and something about 
"Adds 10
minutes timeout". It would also be of some interest, if you could let PyChess 
hang
for those 10-15 minutes, and see if it wakes up.

Thanks in advance,
Thomas

Original comment by lobais on 18 Aug 2009 at 3:03

[GoogleCodeExporter]

Yeah. And when it was apparent that the problem didn't lie in the PyChess player
subprocess I also tried to reproduce this under Fedora 11 but was unable to. My 
next
plan was to add a bunch of relevant debugging output to the relevant areas of 
the
code and have you run it I didn't get around to it, yet.

Original comment by mattgatto on 18 Aug 2009 at 10:08

[GoogleCodeExporter]

Another solution I've been concidering is to port PyChess engine to UCI, simply
because we tend to have much less trouble with these engines.
Likely this would let us access more features of the engine.

Original comment by lobais on 19 Aug 2009 at 6:05

[GoogleCodeExporter]

Hey guys, Im back and unfortunately almost nothing changed for better. :-(

At rev 1521. Still cannot make pychess start (for half a year or so). I would 
say it
is some kind of python config problem or so because when I tried to do fresh 
install
of Ubuntu 9.04 and downloaded svn revision there, everything went just fine.

But here, under 9.04 (originaly 8.04, then 8.10, having pychess 0.8 from 
distribution
now svn) cannot make running any version at all (not even distribution now). 
What I
want to express here is that my system might be a bit cluttered with config 
files and
versions of libraries.

What I did lately was "rm -rf /root/.cpan". Also I previously removed all chess
engines so the only one checked was pychess engine. But after that it hangs
nevertheless. Sometimes (rarely) it shows even "main program screen" after 
engine
check but most of the time (99 %) it hangs at engine check screen.
Absolutely no output to terminal (no debugging turned on, do not know how, but 
may
try when advised), creates two log files, both attached below.

Sorry for a bit unclean description, will answer questions when needed.
Its a pitty it does not work and I assume there is no sign of solving this issue
soon, as I understand it to be quite mysterious, is it so? If it only was on 
win I
would reinstall os and be happy, but I just keep telling myself: "hey, you are 
on
linux, you are not supposed to actually >solve< problems by 
rebooting/reinstalling"
so I am waiting for salvation. :-)

Any advice or help will be greatly appriciated.

Original comment by pe3...@gmail.com on 17 Sep 2009 at 8:57

Attachments:

• 2009-09-17_22-49-34.log
• 2009-09-17_22-49-42.log