torcs-devel Mailing List for TORCS - The Open Racing Car Simulator

Hi Bernie,
I want to simulate some specific motorway scenarios, perhaps 40-50 (to start
with) cars going round a 50 mile track (a subsection of a local highway).
I want to display/simulate the ones around the player for realism. 

Then a specific scenario would be a traffic jam on the highway, where there
are a bunch of cars around the corner  which have slowed down or stopped.
And I want to see how the player reacts to this (in clear conditions, and
then with low visibility).

The idea is to set up a bunch of these 'real-life' scenarios which can test
a driver and measure their reactions, to see how people react to bad
visibility and how they react with head-up displays to aid them (when they
can't see that far due to visibility).

Thanks
sachi
 

-----Original Message-----
From: tor...@li...
[mailto:tor...@li...] On Behalf Of Bernhard
Wymann
Sent: 29 August 2005 12:29
To: tor...@li...
Cc: tor...@li...
Subject: Re: [Torcs-devel] positioning robots

Hi Sachi

> Quick question, whats the best way (briefly) to setup a robot in a
> particular position on the track (instead of at the start of a race or
so).

As a hack you can add code to place them directly in 
src/libs/raceengineclient/raceinit.cpp. If this is suitable depends on your 
requirements, of course you could make it configurable via a configuration
file.

> So that when the race starts the human player does not experience these
> robots until they pass some point in the track (like if the robot is
waiting
> for them).

Hmm, first I try to understand this right: Because you want to simulate
highway 
driving, you need actually a lot of other cars (hundeds, maybe thousands),
so 
you want just to simulate/display the ones surrounding the player to make it

suitable for performance reasons, do you mean that?

Bye, Bernhard.


-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb



-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Torcs-devel mailing list
Tor...@li...
https://lists.sourceforge.net/lists/listinfo/torcs-devel

Hi Sachi

> Quick question, whats the best way (briefly) to setup a robot in a
> particular position on the track (instead of at the start of a race or so).

As a hack you can add code to place them directly in 
src/libs/raceengineclient/raceinit.cpp. If this is suitable depends on your 
requirements, of course you could make it configurable via a configuration file.

> So that when the race starts the human player does not experience these
> robots until they pass some point in the track (like if the robot is waiting
> for them).

Hmm, first I try to understand this right: Because you want to simulate highway 
driving, you need actually a lot of other cars (hundeds, maybe thousands), so 
you want just to simulate/display the ones surrounding the player to make it 
suitable for performance reasons, do you mean that?

Bye, Bernhard.


-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb

>
> Quick question, whats the best way (briefly) to setup a robot in a
> particular position on the track (instead of at the start of a race or so).
> So that when the race starts the human player does not experience these
> robots until they pass some point in the track (like if the robot is waiting
> for them).
>

Currently there is no way to do that easily. I would expect that the 
cleanest solution would be to create a new type of race, where
drivers are initially positioned at different places. This new type of 
race should also manage drivers that are too far away to be 
influencing the player, but I expect you have an idea of how to do 
that.



>
> I aim trying to simulate (slightly realistic) highway driving here as
> opposed to races. I apologise if what im asking is already laid down clearly
> in bernie's robot tutorial or a faq, (I assumed it wouldn't due to the
> different nature).thanks
>
>
>
> sachi
>
>

-- 
Christos Dimitrakakis
IDIAP (http://www.idiap.ch/~dimitrak/main.html)

Hi guys,

Quick question, whats the best way (briefly) to setup a robot in a
particular position on the track (instead of at the start of a race or so).
So that when the race starts the human player does not experience these
robots until they pass some point in the track (like if the robot is waiting
for them).

 

I aim trying to simulate (slightly realistic) highway driving here as
opposed to races. I apologise if what im asking is already laid down clearly
in bernie's robot tutorial or a faq, (I assumed it wouldn't due to the
different nature).thanks

 

sachi

Hi all

TORCS, The Open Racing Car Simulator is a highly portable multi platform car 
racing simulation. It is used as ordinary car racing game, as AI racing game and 
as research platform. It runs on Linux (x86, AMD64 and PPC), FreeBSD and Windows.

The 1.2.4 release highlights are reworked tracks and cars, updated sound and 
added OpenAL support, more clever and new opponents, support for texture 
compression and downscaling, heavily improved collision detection and response, 
a Windows debug project, lots of little improvements and bug fixes, and an 
updated track editor.

Full release notes: http://www.torcs.org
Recent screenshots: http://www.berniw.org/trb/gallery/gallery.php
Older screenshots: http://torcs.sourceforge.net/sections.php?op=listarticles&secid=3

Bye, have fun,

Bernhard.


-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb

With respect to the TODO list, I'll be away for one week or so. After 
that I'll be available as usual. (Vacation)

Good luck with the release

-- 
Christos Dimitrakakis
IDIAP (http://www.idiap.ch/~dimitrak/main.html)

Hi all

I have the "draft" TODO list for 1.3.0 ready. I think it needs to be 1.3.0 
(instead of 1.2.5) because of the SDL switch which could cause major grief and 
trouble in the beginning. I would like to switch to SDL because of portable 
threading and fullscreen problems (perhaps I will change mind after the first 
try, but we will see, a critical factor is how well it does on Windows).

The 1.3.0 should be still in the spirit of reducing the bug count and finishing 
started/broken features, below you find the current TODO list. Bye,

Bernhard.


TODO for 1.3.0 (estimate: goal 12.2005, real 2.2006)
--------------
- Fix sound volume option (ask Christos).
- Fix and reenable -s option in Windows (car rendering ignores missing multitex
   extension).
- fix temporary sound volume jump when changing camera with OpenAL (plib not
   tested, verify) (ask Christos).
- Missing doppler on Windows, investigate.
- SDL port.
- Check for memory leaks.
- Add valgrind option to "torcs".
- Include suppressions file for valgrind.
- (Fix parallel build with make -j n, breaks currently in SOLID) (Bernhard).
- Ongoing for every release: rework free car models (several holes, no
   emmissive color of lod (model becomes dark at a certain distance), single
   sided, add cockpit, lods).
- Ongoing for every release: Improve visual quality of some existing tracks
   (e.g. e-track-3, michigan, eventually wheel-1, e-track-2).
- Add cameras (alpine, spring)
- Texture clamp to edge needs explicit setting in nvidia driver tab in windows,
   have a look into that (otherwise the border is not taken as color around the
   pit door, pits are quite dark around the logo).
- check and merge Christos camera code if useful.
- Integrate documentation into CVS/distribution.
- merge berniw_2004 code into berniw's (rescheduled, a bit more testing will
   not hurt).
- merge bt_2004 into bt.
- review/comment simuv2 (e.g. rollres, fuel, collision, more).
- possibility of one pit for 2 cars of one module.
- fix tools (accc, trackgen, etc, these are way too touchy).
- make trackgen generate two sides for barriers and use just one sided
   polygons (better diffuse lighting).
- fix sound in split screen multiplayer. [in progress] (ask Christos).
- Investigate non reproducable (so far) crash after changing properties in
   multiple players (human) and then doing a race with 2 players. Happend just
   after a fresh install and just once so far (is one of the shipped xml files
   broken)? [note there are many other crashes when using the menus, seemingly
   at random times].
- Add remaining kcendra cars.
- Check and add Andrews tracks.
- Check and add Andrews car improvements.
- Check and add Erics new car class.
- Ongoing: Replace some defines with "static const" to be able to see the
   symbol name when debugging.
- berniw/bt/sparkle: make use of classes in lib/math.
- Investigate/fix crash on rainbowroad (happens as well with 1.2.3). 
RtTrackGlobal2Local
   seems to deliver nonsense, could be caused by trackloader through unusual turn
   angles (e.g. > 180deg). Check if this problem just occurs on version 3 tracks.
- add more races to championship.
- move berniw/bt spline code into math to share it.
- hunt down all glGet* crap during the simulation.
- verify gear switching flames for car, e.g for the 360.
- check all textures for the user selectable cars, they should be nicer.

- (get/write raceengine.xml to users local dir.)
- (Problem when driver list changes during championship.)
- (add proper init/shutdown to every module which is not bound to anything else
   but the loading/init/shutdown/unloading process.)
- (make use of pbuffers to render trackmap if available.)





-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb

Hi all

I released a new test release, visit www.torcs.org for details. Please grab the 
packages and test them. Thank you for your help, bye,

Bernhard.


-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb

>
>> WRT engine fuel consumption, this is a bit weird as when you accelerate so 
>> that you exactly cancel the engine braking torque, you consume 0 gas, while 
>> when you let go of the accelerator completely, you actually 
>
> Actually not, depending on the engine braking setup the injection will incect 
> more fuel and ignit earlier, IIRC (perhaps I am wrong on that, need to look 
> it up). Not too weird. But if it is wrong we can still change that after 
> test3, that should be quite harmless.
>

Oh, right. I know that my car always injects fuel a few seconds after 
I let go of the petal, but after that it stops injecting. And 
carburateur engines I think always consume a little bit.

My only concern was that with current implementation the minimal 
consumption occurs when accel = EngBrkK, which is 0.33 in most cars, 
and it actually equals 0 in that case. Before, without the fabs, it 
would go negative after with accel < 0.33.

So, basically my own solution in simuv3 is to simply add a constant so 
that it's always positive.

But it's not terribly important anyway.

>> Yeah, that's great.. what was the deal with that?
>
> Have a look into the sf.net bugtracker.

Peculiar that the tSituation struct is not reset when race restarts. 
Is it a static?

-- 
Christos Dimitrakakis
IDIAP (http://www.idiap.ch/~dimitrak/main.html)

A small mistake there. It should be:

  tdble Tq_pre = Tmax * car->ctrl->accelCmd * (1.0f + EngBrkK);
  engine->Tq = Tq_pre - Tmax * EngBrkK;
  car->fuel -= Tq_pre* engine->rads * engine->fuelcons * 0.0000001 *
              SimDeltaTime; // no fabs() needed




-- 
Christos Dimitrakakis
IDIAP (http://www.idiap.ch/~dimitrak/main.html)

Hi Christos

> Peculiar that the tSituation struct is not reset when race restarts. Is 
> it a static?

It becomes allocated in the CONFIG state in the state machine, if I remember 
correct (I am too lazy to look that up now;-) ).

Bye, Bernhard.

-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb

> Hi all
>
> I commited those changes into the CVS:
> - Prepared installation instructions for 1.2.4 (Bernhard).
> - Prepared credits for 1.2.4 (Bernhard).
> - Updated robot tutorial for 1.2.4 (Bernhard).
> - Applied reverse gear patch from Mart Kelder (Mart Kelder, Bernhard).
> - Make engine consume fuel on negative torque (Bernhard).

WRT engine fuel consumption, this is a bit weird as when you 
accelerate so that you exactly cancel the engine braking torque, 
you consume 0 gas, while when you let go of the accelerator 
completely, you actually consume something. More correctly, and 
without additional computation:

tdble Tq_pre = Tmax * car->ctrl->accelCmd * (1.0f + EngBrkK);
engine->Tq = Tq_pre - EngBrkK;
car->fuel -= Tq_pre* engine->rads * engine->fuelcons * 0.0000001 *
              SimDeltaTime; // no fabs() needed

This still does not consume any fuel at tickover... but maybe that's 
not important.

Simple way:
   tdble accelCmd2 = car->ctrl->accelCmd;
   if (engine->rads <= engine->tickover) {
     accelCmd2 = MAX(0.1f, accelCmd2);
   }


> - Fixed cars starting at odd heights, reported by Andrew Sumner (Bernhard).

Yeah, that's great.. what was the deal with that?

> - Commited some more docs from the website into the CVS (Bernhard).
>

Good work

-- 
Christos Dimitrakakis
IDIAP (http://www.idiap.ch/~dimitrak/main.html)

Hi Christos

> WRT engine fuel consumption, this is a bit weird as when you accelerate 
> so that you exactly cancel the engine braking torque, you consume 0 gas, 
> while when you let go of the accelerator completely, you actually 

Actually not, depending on the engine braking setup the injection will incect 
more fuel and ignit earlier, IIRC (perhaps I am wrong on that, need to look it 
up). Not too weird. But if it is wrong we can still change that after test3, 
that should be quite harmless.

> Yeah, that's great.. what was the deal with that?

Have a look into the sf.net bugtracker. I spent too much time with this, took me 
more that 3 hours I think (was distributed over three evenings). Tonight I will 
finishing the test3 packages and release them (diff and tar.bz2 are ready, but I 
still need to create the exe and the announcments).

Bye, thank you,

Bernhard.


-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb

Hi all

I commited those changes into the CVS:
- Prepared installation instructions for 1.2.4 (Bernhard).
- Prepared credits for 1.2.4 (Bernhard).
- Updated robot tutorial for 1.2.4 (Bernhard).
- Applied reverse gear patch from Mart Kelder (Mart Kelder, Bernhard).
- Make engine consume fuel on negative torque (Bernhard).
- Fixed cars starting at odd heights, reported by Andrew Sumner (Bernhard).
- Commited some more docs from the website into the CVS (Bernhard).

The most significant thing is the fix for the bug report of Andrew.

Bye, have fun,

Bernhard.


-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb

Hi

>For me, those crashes usually appear when in the race settings menus, when
>I click "accept" to skip to the next screen very fast, just after the
>screen got displayed. They doesn't seem to happen when I wait a couple of
>seconds before clicking accept, or when I change a setting before.

Uhm, that sounds like it will become a quite difficult story...

>Do you have any pointer to docs explaining building a debug exec to mention
>? I may as well run torcs from within a debugger till I catch one of those
>crashes again.

Under Linux it is trivial, perhaps this works as well in FreeBSD (I expect
it to work), try to run configure with the "--enable-debug" option. Previously
make sure that you have not set any CFLAGS, CPPFLAGS or CXXFLAGS in your
environment, it can be quite confusing to run an optimized binary with debugging
info.

Bye, thank you for your help, 

Bernhard. 

On 15-Aug-2005 Bernhard Wymann wrote:
> It would probably help if we could capture one or several reproducable 
> exemplars, do they perhaps happen on certain function keys/keys/mouse 
> sequences/combinations, can they caused with the mouse alone, are all
> screens 
> affected or just some of them, etc?
> 
[snip]
> So, if you ever hit one of them again, please hold on and try to
> reproduce it 
> immediately, perhaps you can remember what you have done exactly? Does it
> happen 
> again?

For me, those crashes usually appear when in the race settings menus, when
I click "accept" to skip to the next screen very fast, just after the
screen got displayed. They doesn't seem to happen when I wait a couple of
seconds before clicking accept, or when I change a setting before.

Perhaps some callbacks take a lil time before being actually registered ?

Anyway, hope that helps.


Do you have any pointer to docs explaining building a debug exec to mention
? I may as well run torcs from within a debugger till I catch one of those
crashes again.


Thanks


--
        Guy 'DeVice'

More on the Esc crash: it seems to happen at :

#1  0x0040f839 in GfuiScreenActivate (screen=0x192) at gui.cpp:465
#2  0x0040f147 in gfuiKeyboard (key=27 '\033') at gui.cpp:241

Both of which are callbacks, of course. Upon investigating, it seems 
that curKey->onPress is not pointing to a valid address at the moment 
of the crash. This normally does point to a proper address.. so..

-- 
Christos Dimitrakakis
IDIAP (http://www.idiap.ch/~dimitrak/main.html)

> So, if you ever hit one of them again, please hold on and try to reproduce it 
> immediately, perhaps you can remember what you have done exactly? Does it 
> happen again?

I can never reproduce it exactly, but it seems to usually happen when 
I run a race, go out to change something such as the player's car, or 
the simu version, then go to run a race again. It either happens when 
I actually start a race, or when I hit the Esc key. Come to think of 
it, it seems related to the Esc key. This is a recent traceback where 
TORCS crashed when I hit Esc:

#4  0x0036d67a in GfuiScreenActivate ()
    from /usr/local/lib/torcs/lib/libtgfclient.so
#5  0x0036d097 in gfuiKeyboard () from 
/usr/local/lib/torcs/lib/libtgfclient.so
#6  0x001214f6 in processEventsAndTimeouts () from 
/usr/local/lib/libglut.so.3
#7  0x001220f5 in glutMainLoop () from /usr/local/lib/libglut.so.3

I guess libtfd is compiled without -g flags? I'll recompile with 
appropriate flags and see if I can get some line numbers.

  >
> I will grep tonight through the sources to look at the callback 
> registering/unregistering, perhaps I find something (I fixed recently one 
> location where not all callbacks have been unregistered).
>
> Bye, thank you for your reports,
>
> Bernhard.
>
>
>

-- 
Christos Dimitrakakis
IDIAP (http://www.idiap.ch/~dimitrak/main.html)

Hi all

> #13 0x2849dff3 in glutMainLoopEvent () from /usr/local/lib/libglut.so.11
> #14 0x2849e3a5 in glutMainLoop () from /usr/local/lib/libglut.so.11

This is the only common thing I experience with all menu related crashes, they 
actually started with a callback from glut (ok, that is no surprise;-) ). I have 
the impression that callbacks are called where they should not, could be caused by:
- TORCS bugs in registering/unregistering callbacks.
- GLUT/FreeGLUT does not work like specified when changing the callbacks after 
the event loop has been started.
- Other bugs.

It would probably help if we could capture one or several reproducable 
exemplars, do they perhaps happen on certain function keys/keys/mouse 
sequences/combinations, can they caused with the mouse alone, are all screens 
affected or just some of them, etc?

Of course we should get rid of those if we can, they lurked around from the very 
beginning of my TORCS activities. But I had really the impression that the 
situation improved since 1.2.3, but when you guys experience those often there 
is somthing wrong:-( Hmm, perhaps just my menu navigation technique changed in a 
way which causes almost zero crashes...

So, if you ever hit one of them again, please hold on and try to reproduce it 
immediately, perhaps you can remember what you have done exactly? Does it happen 
again?

I will grep tonight through the sources to look at the callback 
registering/unregistering, perhaps I find something (I fixed recently one 
location where not all callbacks have been unregistered).

Bye, thank you for your reports,

Bernhard.


-- 

Visit my homepage http://www.berniw.org
Official TORCS racing: The TORCS Racing Board, http://www.berniw.org/trb

Bernhard, HI:

Thanks for your reply.  I asked the question below
when I read all the news which stated that debug build
does not work.  When I tried to compile the project in
debug build, I succeeded in compiling both release and
debug.  

relating to "Fatal Error C1001", my environment is 
windows 2000
visual studio enterprise 6.0(IIRC, maybe service pack
6.  For now, I don't know where to get this info). 
For this problem, it was resolved after I changed the
optimization setting as I said in my msg.  I don't
need any help for that.  thanks for asking again.





--- be...@bl... wrote:

> Hi
> 
> >Anybody know why the debug build does not work
> while
> >release build works?  
> 
> In 1.2.4-test2 both work, in 1.2.3 just the release
> worked like stated in
> the README. Probably you used the wrong compiler or
> did not follow the instructions.
> If you want help, please send a more specific report
> (Which version, OS/sp,
> Compiler/sp, 32/64bit, Memory, CPU(s),..., what did
> you do exactly, what
> error happend exactly, etc).
> 
> Bye, Bernhard.
> 
> 
> 
>
-------------------------------------------------------
> SF.Net email is Sponsored by the Better Software
> Conference & EXPO
> September 19-22, 2005 * San Francisco, CA *
> Development Lifecycle Practices
> Agile & Plan-Driven Development * Managing Projects
> & Teams * Testing & QA
> Security * Process Improvement & Measurement *
> http://www.sqe.com/bsce5sf
> _______________________________________________
> Torcs-devel mailing list
> Tor...@li...
>
https://lists.sourceforge.net/lists/listinfo/torcs-devel
> 


Thank you,

. tSb


		
____________________________________________________
Start your day with Yahoo! - make it your home page 
http://www.yahoo.com/r/hs 
 

Hi

Sorry, not sufficient info, read previous reply. What TORCS version? Compiler
vc++ 6.0sp6? 32 Bit version of Windows? Could you give a reosanable answer
with the given data -> no.

Bye, Bernhard.

>For anybody who is having similar compile problem.
>
>When I build TORCS - Win32 Release, I saw
>
>Fatal Error C1001
>
>for project ssggraph
>
>When I changed setting C/C++ | Optimizations from
>"Maximum speed" to "Default", the error went away.  My
>machine processor is Intel(R) Xeon(TM) CPU 2.00GHZ.
>
>
>Thank you,
>
>tSb
>
>
>		
>____________________________________________________
>Start your day with Yahoo! - make it your home page 
>http://www.yahoo.com/r/hs 
> 
>
>
>-------------------------------------------------------
>SF.Net email is Sponsored by the Better Software Conference & EXPO
>September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
>Agile & Plan-Driven Development * Managing Projects & Teams * Testing &
QA
>Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
>_______________________________________________
>Torcs-devel mailing list
>Tor...@li...
>https://lists.sourceforge.net/lists/listinfo/torcs-devel

Hi

>Anybody know why the debug build does not work while
>release build works?  

In 1.2.4-test2 both work, in 1.2.3 just the release worked like stated in
the README. Probably you used the wrong compiler or did not follow the instructions.
If you want help, please send a more specific report (Which version, OS/sp,
Compiler/sp, 32/64bit, Memory, CPU(s),..., what did you do exactly, what
error happend exactly, etc).

Bye, Bernhard.

For anybody who is having similar compile problem.

When I build TORCS - Win32 Release, I saw

Fatal Error C1001

for project ssggraph

When I changed setting C/C++ | Optimizations from
"Maximum speed" to "Default", the error went away.  My
machine processor is Intel(R) Xeon(TM) CPU 2.00GHZ.


Thank you,

tSb


		
____________________________________________________
Start your day with Yahoo! - make it your home page 
http://www.yahoo.com/r/hs 
 

Greetings, 

Anybody know why the debug build does not work while
release build works?  

Thank you,

tSb

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

On 13-Aug-2005 Christos Dimitrakakis wrote:
> A few crashes. Trace:
> 
> (gdb) #0  0x00d21a3a in reConfigureMenu ()
>     from /usr/local/lib/torcs/lib/libraceengine.so
>#1  0x0011da28 in gfuiButtonAction ()
>     from /usr/local/lib/torcs/lib/libtgfclient.so
>#2  0x0011c7eb in gfuiMouseAction ()
>     from /usr/local/lib/torcs/lib/libtgfclient.so
>#3  0x0011a485 in gfuiMouse () from 
> /usr/local/lib/torcs/lib/libtgfclient.so
>#4  0x00ce73ed in processEventsAndTimeouts () from 
> /usr/local/lib/libglut.so.3
>#5  0x00ce80f5 in glutMainLoop () from /usr/local/lib/libglut.so.3
>#6  0x08048de0 in main ()
> 
> I don't know if that's the same trace you get.
> 
> -- 
> Christos Dimitrakakis

Not the same as me anyway :

#0  0x28052fb0 in r_debug_state () from /libexec/ld-elf.so.1
#1  0x2805259f in dlopen () from /libexec/ld-elf.so.1
#2  0x08049881 in ?? ()
#3  0x2842bfc1 in GfModInfoDir () from /usr/local/lib/torcs/lib/libtgf.so
#4  0x280add8b in RmDriversSelect () from
/usr/local/lib/torcs/lib/libracescreens.so
#5  0x2846c53a in reConfigRunState () from
/usr/local/lib/torcs/lib/libraceengine.so
#6  0x2846c22b in configHookActivate ()
   from /usr/local/lib/torcs/lib/libraceengine.so
#7  0x2843d55a in GfuiScreenActivate ()
   from /usr/local/lib/torcs/lib/libtgfclient.so
#8  0x280abd7c in rmtsDeactivate () from
/usr/local/lib/torcs/lib/libracescreens.so
#9  0x280ac20f in rmtsSelect () from
/usr/local/lib/torcs/lib/libracescreens.so
#10 0x28440af9 in gfuiButtonAction () from
/usr/local/lib/torcs/lib/libtgfclient.so
#11 0x2843f735 in gfuiMouseAction () from
/usr/local/lib/torcs/lib/libtgfclient.so
#12 0x2843d2bd in gfuiMouse () from /usr/local/lib/torcs/lib/libtgfclient.so
#13 0x2849dff3 in glutMainLoopEvent () from /usr/local/lib/libglut.so.11
#14 0x2849e3a5 in glutMainLoop () from /usr/local/lib/libglut.so.11

(It also crash gdb, which i've never seen before... ?)

If you need further info or actions on my side, just ask.

--
        Guy 'DeVice'