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

Hi Brian

> What are the plans for the SDL patch?  Is ever going

It will go in, its the next thing on my TODO list, see README.

> to be used? Will a new patch need to be made if it
> used because of code changes since the patch was made?
>  
> 
> I miss having fullscreen in Linux :(

With FreeGLUT it should work perfectly... I want SDL more for the portable
threading.

Bye, thank you very much,

Bernhard.


-- 

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

What are the plans for the SDL patch?  Is ever going
to be used? Will a new patch need to be made if it
used because of code changes since the patch was made?
 

I miss having fullscreen in Linux :(

Brian



 
____________________________________________________________________________________
Sponsored Link

$420k for $1,399/mo. 
Think You Pay Too Much For Your Mortgage? 
Find Out! www.LowerMyBills.com/lre

Hi Brian

> It is caused by having a space in the path name where
> wtorcs.exe is installed (ie "C:\Program Files")
> 
> The bug is in the file screen.cpp in the function
> GfScrReinit(void * /* dummy */)

Confirmed and applied to CVS, good catch:-) Thank you very much, this never hit
me because I have a German Windows version, where the "Program Files" folder is
named "Programme". Funny:-)

Bye, Bernhard.

-- 

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

The attached patch should fix this bug.

[ 1432114 ] i cannot change resolution

It is caused by having a space in the path name where
wtorcs.exe is installed (ie "C:\Program Files")

The bug is in the file screen.cpp in the function
GfScrReinit(void * /* dummy */)

The bug will only occur in the windows version.

Brian


 
____________________________________________________________________________________
Do you Yahoo!?
Everyone is raving about the all-new Yahoo! Mail beta.
http://new.mail.yahoo.com

2006/11/12, Bernhard Wymann <be...@bl...>:
> > this is a small fix to correct the "checking for XOpenDisplay in
> > -lX11" check if the version of XOrg is 7.x on the host system
>
> Is this really required, does supplying the parameters --x-libraries and
> --x-includes not do the trick?

thanks for your hint, using "--x-libraries=/usr/lib
--x-includes=/usr/include" also works

udv / greetings,
VMiklos

-- 
Developer of Frugalware Linux, to make things frugal - http://frugalware.org

Hi,

 I am preparing a Pardus[1] package[2] for Torcs 1.3.0.

 Unlike most distributions, Pardus doesn't use /usr/share/games.
Instead data files of games are put to /usr/share too.

 When I wanted to install Torcs data files to /usr/share/torcs, I
noticed that the "games" directory in datadir is hard-coded in
various .in files and the configure parameter --datadir doesn't override
it, "games" suffix is appended nevertheless.

 At the current revision of the Pardus package, I had to sed 20 files in
the torcs source package to remove the games directory for
installation. It's not a solution that I prefer since I would have to
revise it when the next torcs version is released.

 Is there a way to automatically generate these .in files to exclude
the "games" suffix?

 Thanks in advance.


[1] http://www.pardus.org.tr/eng/
[2] http://svn.pardus.org.tr/pardus/playground/applications/games/torcs/

                   Doruk

--
FISEK INSTITUTE - http://www.fisek.org.tr

Hi VMiklos

> this is a small fix to correct the "checking for XOpenDisplay in
> -lX11" check if the version of XOrg is 7.x on the host system

Is this really required, does supplying the parameters --x-libraries and
--x-includes not do the trick?

Bye, thank you for the report,

Bernhard.	

-- 

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

Hi all,

I attached the list of features I implemented in the past period as 
attachment. I can isolate any subset of this list if the dependencies hold and 
make a patch of it.

I post this now because:
- in an earlier mail I announced to repost it later;
- I think the best time to post this is at the start of a release cycle.

Regards,

Mart

hi

http://ftp.frugalware.org/pub/frugalware/frugalware-current/source/games-extra/torcs/torcs-1.3.0-xorg7.diff

this is a small fix to correct the "checking for XOpenDisplay in
-lX11" check if the version of XOrg is 7.x on the host system

udv / greetings,
VMiklos

-- 
Developer of Frugalware Linux, to make things frugal - http://frugalware.org

On Tue, 07 Nov 2006 17:20:05 +0100, Lewis Foster <do...@gm...> wrote:

> Hi, my first post to the list.
>
> I'm looking at doing a physics type final year project for my Bsc  
> Computer
> Science. Torcs is looking nice and it would be great if I could put my  
> time
> into a racing sim to try and improve some aspect of the game.
>
> Does anyone have any ideas as to what I could try and improve?
>
> Also, what physics code does torcs use?
>
> Thanks,
>
> Lewis


Hi,

not really what you had in mind but you may like to consider this idea.

For some time I have been toying with the idea of making a braking  
distance awareness program based on Torcs.

Apart from being an intersting test to run at home it would be very useful  
in driving tution to make people aware of just how far behind you NEED to  
be to avoid a collision.

The simplest form is pretty much already doable: circulate at a normal  
road speed , then the instructor or a random event in torcs signals an  
emergency . You hit the brakes and then review your reactions in a play  
back.

Rolling start , empty road just to appreciate braking distances and  
reaction time to hit the brakes.

Program mods to take it further:

1/ on-screen display of speed, reaction time and breaking distance.
2/ child, dog , ball, pram, sudden car from dirt track... events.
3/ new urban tracks more suited to the context of driving. Addition of  
some urban robots (the torcs automatic cars) to make roads busier  
+ motorbikes, pushbikes.
4/ more urban car specs: clios and escorts rather than race tuned subarus.

To judge by most road driving, people do not realise they would not even  
have the time to close thier eyes and say "shit!" before hitting the car  
in front, let alone actually touching the brake pedal.

Although this may not be as effective as Tony Blair getting the police to  
falsify the road accident statistics , the lives saved would be real  
rather than statistical fiction.


In sticking to open source and with a couple of privisos, the end result  
could be distributed on a knoppix or similar live CD that could work out  
of the box without installation on most PCs.

There would be the need for a seriously scientific approach to measure car  
parameters that affect braking simulation and comparison with real car  
data. This will almost certainly lead to some applied physics and work on  
the modelling to provide meaningful results.


That's a fairly broad outline of what would be involved if that seems  
interesting or relevant to you.

regards.

Lewis Foster wrote:

> Hi, my first post to the list.
>
> I'm looking at doing a physics type final year project for my Bsc 
> Computer Science. Torcs is looking nice and it would be great if I 
> could put my time into a racing sim to try and improve some aspect of 
> the game.
>
> Does anyone have any ideas as to what I could try and improve?
>
We have two major ptoblems: Collisions and Friction.  This is not 
surprising given that both these problems are an active research topic 
in modelling.

So, for Collisions, we want to do:
a) Semi-elastic body collisions with of cars with track/border/other cars.
c) Handling the case when the suspension is fully compressed, which is a 
sort of collision
d) Handling the case when the car is tilted at an extreme angle relative 
to the road but the wheels are still touching the track - or when the 
wheels hit a barrier - which is also a type of collision. 

These problems are very big and I am going to re-write the whole of the 
collision engine in order to solve them - anything you think might be 
feasible would be nice.

For friction, the problems are not very big right now -
a) scaling of friction forces depending on speed - there are some 
oscillations at very low speeds.
b) when the car is on a tilted track, with brakes on, it very very 
slowly slides downwards.
Both of these are value-propagation problems.

> Also, what physics code does torcs use?
>
It's a mix of things.  Most of it is a first-order approximation of 
differential equations.  Most of the dependencies variables are resolved 
with linear
value propagation.. For the tyres a simple relative-velocity friction 
model is used, which explains the 2 friction problems above.

The actual simu engine is maintained in 2 separate parts now:

- simuv2, which is stable and for which only serious bug fixes are released
- simuv3, which is experimental and which works in full 3d.

For simuv3:
Car rotation: moments of inertia
Car position: rotation matrix
Car movement in general uses conservation of momentum as the central 
principle.
Suspension: Simplified projection on track (for position)
Wheels: Semi-independent
Aerodynamics: Three models: An ad-hoc very simple one, a planar one, and 
one based on conservation of energy.  Still, they are all very simple.
Drivetrain: Uses conservation of momentum as the central principle, 
frictive forces for some differentials, while others rely on the 
action-reaction principle.

I think that the only changes in the setup for the future might be made 
at the interface between wheels and the track and the interface between 
the car body and other objects.

Cheers, Christos

Hi, my first post to the list.

I'm looking at doing a physics type final year project for my Bsc Computer
Science. Torcs is looking nice and it would be great if I could put my time
into a racing sim to try and improve some aspect of the game.

Does anyone have any ideas as to what I could try and improve?

Also, what physics code does torcs use?

Thanks,

Lewis

Hi all

I have modified some piece of code to import an ac3d to shape wheels. I 
had to make a system of LOD not to loose some fps. I have also added an 
option "3d wheel" on graph.xml use this mode of "wheel shaping".

I join patch files generated with the "diff" command but I am not 
familiar with patches, so I have made a package with the whole files 
(even using patches you'll need this package to have the wheel I've made 
for the F1).


The package is : 
http://speedy.chonchon.free.fr/fichiers/torcs/torcs_3D_wheel.tar.gz.

I hope you'll like this.

Bye
Speedy