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

This is what I get from trackgen on my track:
Delta X   = -0.045227
Delta Y   = 0.000065
Delta Z   = 0.000000
Delta Ang = -0.000000 (-0.000010)

Ok, the Delta X is a bit high, but I couldn't bother to be more
precise that day :-)

I have attached the XML file of my track. I'm working on the 3D and
graphical appearance, so I hope you don't mind the default terrain for
the time being. Apart from the surface of some side sections, it
should be definitive. If you want to use it for any purpose, here it
is.

Cheers and happy holidays,

Simone

2008/12/27 hj- schuessler <hj-...@we...>:
> Am Thu, 18 Dec 2008 20:15:43 +0100 schrieb Sk <sk...@gm...>:
>
>> 5) In my tracks, many bots keep crushing on the interior of said curves.
>> It's like they wanted to cut the corner, but they hit the barrier instead.
>> In some of the corners, they keep stepping on the inner curb even if it's
>> quite high, with the result that on every lap they invariantly take off on
>> those curbs. Am I the one with a crazy mind (to have those curbs) or is
>> this
>> simply a scenario the bots weren't thought for? If so, will they in the
>> future, or my tracks are simply "not the expected kind", so to say?
>
> Hello Simone, hello Mario
>  For your graphics problems I cannot help unfortunately…. With me folded
> transferred from 1.3.0 on 1.3.1.test1 without problems (Windows PC) But
> your new bent tracks would interest me much. I assume the problems with
> some bots in an inaccurate conclusion of the track. The differences
> between starting and goal straight lines must be nearly zero. In
> particular the angle differenz is crucial and should be almost zero. Bots of
> Andrew (locus and  hymie) are there particularly sensitive.
> To the problem of Mario with the cars from 1.3.0 I can say the following:
> The simplest solution would be the changeing the simuV2.dll from the
> version 1.3.0, then all cars run, but a substantial progress is lost in
> the suspension comfort when the over driving edgings and flanks. The
> difficult way is, with the appropriate cars, it is not all, the attitudes
> of the suspension to be adapted: The values of bump and rebound are partly
> 800 or 1000, after change on 80 and 100 ran with me all cars problem-free.
> With the control system problems cannot harm an attempt with the old human
> DLL…. Also I used many rounds, in order to accustom me to the new human
> DLL….
>
> I hope that I could help you
>  Wish far much fun
> Bye, Jürgen
>

Hello,

after run this:

# cd /usr/src/torcs/torcs-1.2.4
# ./configure

I receive the following error message: "Can't find libm."


Can you help me to solve the problem?
Enclosed I send you the log file.


Marc
-- 
Sensationsangebot verlängert: GMX FreeDSL - Telefonanschluss + DSL 
für nur 16,37 Euro/mtl.!* http://dsl.gmx.de/?ac=OM.AD.PD003K1308T4569a

-- 
Sensationsangebot verlängert: GMX FreeDSL - Telefonanschluss + DSL 
für nur 16,37 Euro/mtl.!* http://dsl.gmx.de/?ac=OM.AD.PD003K1308T4569a

-- 
Psssst! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger

-- 
Psssst! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger

Am Thu, 18 Dec 2008 20:15:43 +0100 schrieb Sk <sk...@gm...>:

> 5) In my tracks, many bots keep crushing on the interior of said curves.
> It's like they wanted to cut the corner, but they hit the barrier  
> instead.
> In some of the corners, they keep stepping on the inner curb even if it's
> quite high, with the result that on every lap they invariantly take off  
> on
> those curbs. Am I the one with a crazy mind (to have those curbs) or is  
> this
> simply a scenario the bots weren't thought for? If so, will they in the
> future, or my tracks are simply "not the expected kind", so to say?

Hello Simone, hello Mario
   For your graphics problems I cannot help unfortunately…. With me folded
transferred from 1.3.0 on 1.3.1.test1 without problems (Windows PC) But
your new bent tracks would interest me much. I assume the problems with
some bots in an inaccurate conclusion of the track. The differences
between starting and goal straight lines must be nearly zero. In
particular the angle differenz is crucial and should be almost zero. Bots  
of Andrew (locus and  hymie) are there particularly sensitive.
To the problem of Mario with the cars from 1.3.0 I can say the following:
The simplest solution would be the changeing the simuV2.dll from the
version 1.3.0, then all cars run, but a substantial progress is lost in
the suspension comfort when the over driving edgings and flanks. The
difficult way is, with the appropriate cars, it is not all, the attitudes
of the suspension to be adapted: The values of bump and rebound are partly
800 or 1000, after change on 80 and 100 ran with me all cars problem-free.
With the control system problems cannot harm an attempt with the old human
DLL…. Also I used many rounds, in order to accustom me to the new human  
DLL….

I hope that I could help you
   Wish far much fun
Bye, Jürgen

Hi everyone,

I'm experiencing unexpected graphic behaviour. I've attached two
screenshots to illustrate the artifacts: some of the polygons (those
far away from the camera) are flickering. I remember having similar
results with OpenGL when setting inadequate values for the far and
near clipping distance; may this be related somehow?

I never got this problem with v.1.3.0. It first emerged with 1.3.1.

Thanks and cheers,

Simone

Hi Bernhard,


> It is basically the ATI driver which hangs the box then, an "ordinary"

Agreed, but torcs 1.3.1 seems to change the resolution in a different
way than 1.3.0. I think 1.3.0 sets a new size in screen.xml and restarts
itself, and I never saw it crashing. But then, I admit I didn't try
recently to change the resolution in 1.3.0.

> application like TORCS in user mode cannot accomplish this. Try using a newer
> version of the driver, use the open source driver or/and search the web for
> similar reports and eventual solutions.

I'm using the latest one offered by Ubuntu. Anyway, I'll mark that as a
"don't do that" thing for myself.

> Maybe you have a 64-bit system and did not compile with -fPIC (have a look into
> the FAQ). I know there is one last segmentation fault in TORCS (but this is
> there since ages), this happens when you reconfigure the human driver and go
> back and start a race under certain conditions (just in the Linux version, I
> guess it has something to do with the "non dll" architecture of plib, for the
> Windows version the graphics engine is linked statically).

Speaking of plib, I remember I violated the torcs requirements here. The
doc says it wants 1.8.5, and Ubuntu 8.04 only provides 1.84. However,
configure didn't report an error and so I just went on compiling.

Could this be the cause of all the trouble I have?

>> I'm missing many cars from the Track-RWD-GrA group, like the Porsche
>> GT1, McLaren F1 (my favorites) etc. Now there are only these
>> car<x>-trb<y> cars. I copied the others over from the 1.3.0 installation
>> tree (share/games/torcs/cars directory) but they don't work. Did I miss
>> something here?
> 
> Don't work, what does this mean? With the "old" cars the simulation can
> "expolode" when the parameters are not adjusted.

Don't work means that the track is displayed from a strange view (I'm
looking down to the track in a relatively steep angle), and the cars
don't really drive, they only make strange movements.

Have these other cars been left out deliberately?

Is there a tutorial on how to adopt the cars to the new version? Which
parameters need to be asjusted?

>> The frame rate is less than 20 fps, and relatively constant. With torcs
>> 1.3.0 the frame rate usually varied between 40 and 50 fps.
> 
> Depends on the cars/tracks which you select, some stuff has more details, so
> yes, it eats more performance (use the cg-nascar-rwd cars and a simple track to
> save performance, reduce the maximum texture size in the OpenGL options).

Well, with torcs 1.3.0 I could start a race with 18 cars, and the frame
rate was still between 20 and 50 fps, depending on how close all the
other cars were. With 1.3.1 the frame rate is never above 20 fps, even
if only my car is on the track.

>> Steering feels completely different now. The response of the cars to
>> every key stroke is delayed maybe .2 to .4 sec. This is the same for
> 
> Weird, no problem here, in fact you can steer much more agressive now.

I think I will try again with plib-1.8.5 and see if it is any better then.

Thanks and regards
Mario

Hello!
I have a problem with texmapper: here I have run an example command.

~/Desktop/torcs-1.3.1/data/cars/models/155-DTM$ texmapper -f 155-DTM.ac 
-o 155-DTM-out.ac -p texmapper.xml -s skin.rgb
0 branches found
Bounding boxes:
/usr/local/bin/texmapper: line 32:  9628 Segmentation fault      
$LIBDIR/texmapper-bin $*

What is the problem? Thanks for any help.
Regards,
Ocirne

Hi Mario

> Changing the screen resolution in Options -> Display does freeze my PC
> (HP NC6400 with ATI graphics running Ubuntu 8.04 with the fglrx driver).
> It is not only the display, I cannot even login from a second box using
> ssh. I need to power-cycle it for a restart.

It is basically the ATI driver which hangs the box then, an "ordinary"
application like TORCS in user mode cannot accomplish this. Try using a newer
version of the driver, use the open source driver or/and search the web for
similar reports and eventual solutions.

> However, I was able to change the screen resolution manually in
> ~/.torcs/config/screen.xml, and then torcs starts up with the desired
> window size.
> 
> 2)
> torcs crashes with a segmentation fault after I have selected a new car
> and start a race. After restarting torcs I can start the same race with
> the same car without a problem. I saw that mostly  with the "Spring"
> track(my favorite) and the car<x>-trb<y> cars:

Maybe you have a 64-bit system and did not compile with -fPIC (have a look into
the FAQ). I know there is one last segmentation fault in TORCS (but this is
there since ages), this happens when you reconfigure the human driver and go
back and start a race under certain conditions (just in the Linux version, I
guess it has something to do with the "non dll" architecture of plib, for the
Windows version the graphics engine is linked statically).

> I'm missing many cars from the Track-RWD-GrA group, like the Porsche
> GT1, McLaren F1 (my favorites) etc. Now there are only these
> car<x>-trb<y> cars. I copied the others over from the 1.3.0 installation
> tree (share/games/torcs/cars directory) but they don't work. Did I miss
> something here?

Don't work, what does this mean? With the "old" cars the simulation can
"expolode" when the parameters are not adjusted.

> The frame rate is less than 20 fps, and relatively constant. With torcs
> 1.3.0 the frame rate usually varied between 40 and 50 fps.

Depends on the cars/tracks which you select, some stuff has more details, so
yes, it eats more performance (use the cg-nascar-rwd cars and a simple track to
save performance, reduce the maximum texture size in the OpenGL options).

> Steering feels completely different now. The response of the cars to
> every key stroke is delayed maybe .2 to .4 sec. This is the same for

Weird, no problem here, in fact you can steer much more agressive now.

> every car I tried. I hardly get around any somewhat tighter curve the
> same way I did it with 1.3.0.

Try another car, most updated cars are set up to be neutral or to slightly
untersteer.

Bye, Bernhard.

Hi Bernhard,


thanks for releasing the new version. I compiled and ran it yesterday
and want to reply with my impressions.

First, what I really appreciate is that the cpu is no longer running
with full power while I am in a menu or have stopped the race. But there
are a few things that keep me from really wanting to use the new version:

1)
Changing the screen resolution in Options -> Display does freeze my PC
(HP NC6400 with ATI graphics running Ubuntu 8.04 with the fglrx driver).
It is not only the display, I cannot even login from a second box using
ssh. I need to power-cycle it for a restart.

However, I was able to change the screen resolution manually in
~/.torcs/config/screen.xml, and then torcs starts up with the desired
window size.

2)
torcs crashes with a segmentation fault after I have selected a new car
and start a race. After restarting torcs I can start the same race with
the same car without a problem. I saw that mostly  with the "Spring"
track(my favorite) and the car<x>-trb<y> cars:

$ /opt/torcs-1.3.1/bin/torcs
Visual Properties Report
------------------------
Compatibility mode, properties unknown.
WARNING: grscene:initBackground Failed to open shadow2.rgb for reading
WARNING:         no shadow mapping on cars for this track
WARNING: grscene:initBackground Failed to open shadow2.rgb for reading
WARNING:         no shadow mapping on cars for this track
GfParmCheckHandle: parameter "Car/initial fuel" out of bounds: min:1
max:87 val:90.8152 in ("default car" -
"/home/mstreibe/.torcs/drivers/human/car.xml")
WARNING: grscene:initBackground Failed to open shadow2.rgb for reading
WARNING:         no shadow mapping on cars for this track
/opt/torcs-1.3.1/bin/torcs: line 53: 10441 Segmentation fault
$LIBDIR/torcs-bin -l $LOCAL_CONF -L $LIBDIR -D $DATADIR $*
$

3)
I'm missing many cars from the Track-RWD-GrA group, like the Porsche
GT1, McLaren F1 (my favorites) etc. Now there are only these
car<x>-trb<y> cars. I copied the others over from the 1.3.0 installation
tree (share/games/torcs/cars directory) but they don't work. Did I miss
something here?

4)
The frame rate is less than 20 fps, and relatively constant. With torcs
1.3.0 the frame rate usually varied between 40 and 50 fps.

5)
Steering feels completely different now. The response of the cars to
every key stroke is delayed maybe .2 to .4 sec. This is the same for
every car I tried. I hardly get around any somewhat tighter curve the
same way I did it with 1.3.0.


Did anyone else try 1.3.1 yet? Am I the only one with these experiences?
Or is my graphics card too slow for the new version?


Regards
Mario


Bernhard Wymann schrieb:
> Hi all
> 
> TORCS 1.3.1 is now available for download, visit www.torcs.org for details. Do
> not miss the links in the announcment, especially the content exchange.
> 
> Bye, have fun,
> 
> Bernhard.
> 
> ------------------------------------------------------------------------------
> SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
> The future of the web can't happen without you.  Join us at MIX09 to help
> pave the way to the Next Web now. Learn more and register at
> http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
> _______________________________________________
> Torcs-users mailing list
> Tor...@li...
> https://lists.sourceforge.net/lists/listinfo/torcs-users
> 

2008/12/18 mayeul.kauffmann <may...@fr...>
> Le jeudi 18 décembre 2008 à 20:15 +0100, Sk a écrit :
> > 1) I had started some time ago a track editor written in C#. The
> > project was dropped, partly because the Mono user base was not too
> > large. The Java trackeditor, though, has a lot of shortcomings
> > (including the fact that it is Java): the interface is awkward, and
> > there are too many options that a track maker has to set by hand.
> > Could this issue by solved? Or, if enough people accept a .NET track
> > editor, I may try to restore that project of mine.
>
> I guess the software will not run under Linux (would it work with wine
> out of the box). I have a Linux system and several of my TOCS co-players
> have Linux boxes has well. We are developing a long track of a mountain
> road in the Alps and a Java software is better for us because it works
> on many OS's.

Well, Mono *should* run .NET applications under Linux (and Windows
too, for that matter). I haven't tested it a lot, though. The Mono
Migration Analyzer told about problems with the OpenGL library (due to
TaoFramework), but somebody said they ran it ok under Linux.

Both Java and .NET/Mono have the problem of requiring the framework.
Personally, I'd like to avoid all the stuff that Java installs (which
I don't remember seeing in Linux distros, btw). If the trackeditor was
to be improved, though, I may re-install Java just for it.

> > 3) When I try to edit the track ac file in Blender then export it,
> > TORCS keeps crashing during track load. I saw in several places that
> > there is (was?) an export plugin for Blender that was designed
> > specifically for TORCS tracks, but all the links are dead. Is the
> > plugin still available?
> I am not sure. Here it is, attached to this email (I not sure it will be
> kept by the list engine).
> You should put it in your Blender script folder (on
> Linux: /home/user_name/.blender/scripts/ac3d_torcs_export.py
> I'm not sure on Windows).
> The files says it's covered by the "GNU General Public License
> as published by the Free Software Foundation; either version 2 of the
> License, or (at your option) any later version" so I think anybody is
> allowed to put this on any website.

Thank you A LOT! I've been searching for this script for so long! Now
I can create tracks rather than asphalt ribbons :-)

> > 5) In my tracks, many bots keep crushing on the interior of said
> > curves. It's like they wanted to cut the corner, but they hit the
> > barrier instead. In some of the corners, they keep stepping on the
> > inner curb even if it's quite high, with the result that on every lap
> > they invariantly take off on those curbs. Am I the one with a crazy
> > mind (to have those curbs)
> No, you are not crazy, I've seen that on several very curved curves in
> our mountain track. Some robots manage to turn well however. I guess the
> answer is in the AI of some bots.
>
> >  or is this simply a scenario the bots weren't thought for? If so,
> > will they in the future, or my tracks are simply "not the expected
> > kind", so to say?
>
> Hope it helps !
> Mayeul
>

Of course it must be the bots - some do and some don't, so it's not
due to TORCS nor to my tracks. I started to wonder whether I was doing
something that TORCS wasn't made for.

Thanks for the replies!

Hi all

TORCS 1.3.1 is now available for download, visit www.torcs.org for details. Do
not miss the links in the announcment, especially the content exchange.

Bye, have fun,

Bernhard.

Hi,
Thanks for asking some interesting questions, a few of them are my
concern too. AMong them, I'm concerned by the lack of the ability to
change the slope in trackeditor, because I'm working on mountain roads.
More below.

Le jeudi 18 décembre 2008 à 20:15 +0100, Sk a écrit :
> 1) I had started some time ago a track editor written in C#. The
> project was dropped, partly because the Mono user base was not too
> large. The Java trackeditor, though, has a lot of shortcomings
> (including the fact that it is Java): the interface is awkward, and
> there are too many options that a track maker has to set by hand.
> Could this issue by solved? Or, if enough people accept a .NET track
> editor, I may try to restore that project of mine. 

I guess the software will not run under Linux (would it work with wine
out of the box). I have a Linux system and several of my TOCS co-players
have Linux boxes has well. We are developing a long track of a mountain
road in the Alps and a Java software is better for us because it works
on many OS's.


> 3) When I try to edit the track ac file in Blender then export it,
> TORCS keeps crashing during track load. I saw in several places that
> there is (was?) an export plugin for Blender that was designed
> specifically for TORCS tracks, but all the links are dead. Is the
> plugin still available?
I am not sure. Here it is, attached to this email (I not sure it will be
kept by the list engine).
You should put it in your Blender script folder (on
Linux: /home/user_name/.blender/scripts/ac3d_torcs_export.py
I'm not sure on Windows).
The files says it's covered by the "GNU General Public License
as published by the Free Software Foundation; either version 2 of the
License, or (at your option) any later version" so I think anybody is
allowed to put this on any website.


> 5) In my tracks, many bots keep crushing on the interior of said
> curves. It's like they wanted to cut the corner, but they hit the
> barrier instead. In some of the corners, they keep stepping on the
> inner curb even if it's quite high, with the result that on every lap
> they invariantly take off on those curbs. Am I the one with a crazy
> mind (to have those curbs)
No, you are not crazy, I've seen that on several very curved curves in
our mountain track. Some robots manage to turn well however. I guess the
answer is in the AI of some bots.

>  or is this simply a scenario the bots weren't thought for? If so,
> will they in the future, or my tracks are simply "not the expected
> kind", so to say?

Hope it helps !
Mayeul

Hi everybody!

I've been fiddling around with track making for TORCS for a while now, and I
would like to know more about the tracks.

1) I had started some time ago a track editor written in C#. The project was
dropped, partly because the Mono user base was not too large. The Java
trackeditor, though, has a lot of shortcomings (including the fact that it
is Java): the interface is awkward, and there are too many options that a
track maker has to set by hand. Could this issue by solved? Or, if enough
people accept a .NET track editor, I may try to restore that project of
mine.

2) Is it possible to have a better documentation for the track XML file? For
example, what do the border and barrier "style" attributes do? Or, is it
possible to have different segment widths? And again, do segments start and
end with a z-derivative equal to zero? I know most of these can be
experimented with, but it'd be nice to have some official documentation.

3) When I try to edit the track ac file in Blender then export it, TORCS
keeps crashing during track load. I saw in several places that there is
(was?) an export plugin for Blender that was designed specifically for TORCS
tracks, but all the links are dead. Is the plugin still available?

4) I made two road tracks that are actually fictional city tracks, i.e. they
are modeled on the actual streets of two real-life cities near me. These
include several sharp turns, mainly 90 degrees curves with short radii.
Because of the track format, a short radius means that one has to put
barrier and border close to the track, so that they don't include the center
of the curve (for example, if the track is 10m wide and the radius is 7m in
a curve to the left, I can't put anything further than 2m from the left side
of the track). Is this a known problem (a feature?), or will it be addressed
in the future?

5) In my tracks, many bots keep crushing on the interior of said curves.
It's like they wanted to cut the corner, but they hit the barrier instead.
In some of the corners, they keep stepping on the inner curb even if it's
quite high, with the result that on every lap they invariantly take off on
those curbs. Am I the one with a crazy mind (to have those curbs) or is this
simply a scenario the bots weren't thought for? If so, will they in the
future, or my tracks are simply "not the expected kind", so to say?

Thanks everyone for the patience... quite a long mail, isn't it?

Simone

Hi,
had the same problem.

The libXi should be installed.
On my system though, it was called libXi.so.6.0.0 (with having a symbolic link from libXi.so.6 > libXi.so.6.0.0)
And configure seems to search for libXi.so
So I added a symolic link via:
ln -s libXi.so.6.0.0 libXi.so

After that I got the same error for a couple of other libraries (libXmu,  libXxf86vm, libXrender),
so I did the same thing for those libraries.

They all were found then. But now I am stuck with libpng not being found. And adding a sym-link doesn't fix it.

Does someone know what the underlying problem is? Or what to do with libpng?

The solution pointed out before with removing the -L in the configure File on line 6454 pointed out by Mario didn't work for me, as the libX11 wasn't found then at all.

Best Regards,
Thorsten

Re: [Torcs-users] torcs compilation problem
From: Rob Allen <y2kprawn@gm...> - 2008-05-10 19:56

Attachments: Message as HTML     
Interesting, when I use this option you listed below I get a different error


Can't find libXi.

Is this related or is there another related library I need as I cannot find
this in the synaptic package manager ?

Hello Friends!
I'm trying to create something like a car, but I have problems painting 
it: how do I create the file (carname).rgb? When I look at it in other 
cars, I can't understand the position of textures. I know I have to use 
GIMP, but I don't know how to set the shape and the position of pictures.
I have also another question: what is texmapper tool for?
Thanks for your answers,
Cheers,
             Ocirne

Hi Tim

> -has anyone made a version of Torcs that allows the use of a DC electric
> motor powered by a battery?

Not as far as I know, but it should be easy to do, just replace the current
engine model (engine.cpp).

> -can a three wheeled trike (driven single rear wheel) be setup in torcs that
> will give realistic handling?

Not currently, but you can try Christos suggestion or dig into the code and
modify it accordingly, this should be straightforward as well.

> -Instead of fuel used, torcs would have to be modified to show electric
> current used.  The fuel tank level would be replaced by battery charge
> level.  Has anyone got any advice as to how to do this?

Implement you own engine/battery model in engine.cpp.

> -Could a robot be the best way to take data from torcs and display current,
> amp hours used from battery etc?

It could be a quite good way, but I guess you need to fit the model you
implement/parametrize in TROCS with real world measured data to get accurate
results (e.g. rolling resistance, general efficiency, drag, etc.).

> -has anyone found realistic tyre data for bicycle tyres?

No idea, but I guess it should be possible to use the current tire model with
appropriate parameters, but I think you will require real measurements to
calibrate the model.

> -has anyone successfully modelled a go-kart?

Nobody has ever tried officially, so not as far as I know of.

Sincerely yours

Bernhard