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

Bernhard Kaindl <ber...@gm...> writes:

> There are some possilities to check what might be causing the
> messages from slScheduler when sound is on.

In fact, Eric told me about a patch to PLIB-1.6.0 that fixed the issue. 

I enclose in this email the needed patch. It will probably be included
in next stable release on PLIB.

Now, TORCS is working perfectly!! :)

Best regards,
d.

This problem has been fixed in PLIB by a patch commited last week
(but the patch was much older than that).

David sent me the patch he applied on the PLIB 1.6.0 to have the
sound work on his powerpc system:

  Index: sl.h
  ===================================================================
  RCS file: /cvsroot/plib/plib/src/sl/sl.h,v
  retrieving revision 1.22
  diff -u -p -r1.22 sl.h
  --- sl.h	4 Jan 2003 02:55:55 -0000	1.22
  +++ sl.h	16 May 2003 09:04:21 -0000
  @@ -645,6 +645,7 @@ class slScheduler : public slDSP
     float safety_margin ;

     int mixer_buffer_size, mixer_gain ;
  +  float seconds_per_buffer;

     Uchar *mixer_buffer  ;
     Uchar *mixer_inputs [ SL_MAX_MIXERINPUTS + 1 ] ;
  Index: slScheduler.cxx
  ===================================================================
  RCS file: /cvsroot/plib/plib/src/sl/slScheduler.cxx,v
  retrieving revision 1.9
  diff -u -p -r1.9 slScheduler.cxx
  --- slScheduler.cxx	2 Sep 2002 06:05:47 -0000	1.9
  +++ slScheduler.cxx	16 May 2003 09:04:21 -0000
  @@ -84,6 +84,10 @@ void slScheduler::initBuffers ()
     setMaxConcurrent ( 0 );

     mixer_buffer_size = getDriverBufferSize () ;
  +  int div = 1;
  +  if ( getStereo() )     div *= 2 ;
  +  if ( getBps() == 16 )  div *= 2 ;
  +  seconds_per_buffer = (float)mixer_buffer_size / ( div * getRate() ) ;

     mixer_buffer = new Uchar [ mixer_buffer_size ] ;
     memset ( mixer_buffer, 0x80, mixer_buffer_size ) ;
  @@ -132,7 +136,7 @@ void slScheduler::realUpdate ( int dump_

     int i ;

  -  while ( secondsUsed() <= safety_margin )
  +  while ( secondsUsed() <= safety_margin && secondsRemaining() >= 
seconds_per_buffer )
     {
       slPlayer *psp [ SL_MAX_MIXERINPUTS ] ;
       int       pri [ SL_MAX_MIXERINPUTS ] ;
  @@ -174,7 +178,9 @@ void slScheduler::realUpdate ( int dump_
              ( mypri > pri[lowest] )
            )
         { /* Not enough room; get rid of someone */
  -        player[lowest] -> preempt ( mixer_buffer_size ) ;
  +        if ( player[lowest] != NULL )
  +          player[lowest] -> preempt ( mixer_buffer_size ) ;
  +
   	psp[lowest] = player[i] ;
   	pri[lowest] = mypri ;
           int j;


Bye,
Eric.

Bernhard Kaindl wrote:
> On Sun, 25 May 2003, David MENTRE wrote:
> 
> 
>>David MENTRE <dav...@wa...> writes:
>>
>>
>>>However, I have know another error. When launching a race, all the data
>>>is loaded correctly, the sound is available but the screen freezes
>>>without display the car, with the word Ready.
>>>
>>>On the console, I have the following messages:
>>>david@microbe:~$ torcs
>>>WARNING: slScheduler: Too many pending callback events!
>>>WARNING: slScheduler: Too many pending callback events!
>>
>>As a side not, without sound, the display is correct (with about 40
>>fps).
> 
> 
> Very interesting, good to know that the display works without sound.
> 
> There are some possilities to check what might be causing the
> messages from slScheduler when sound is on.
> 
> The man difference of the ppc processor is that it's bigendian,
> and this makes a difference when using addresses with different
> variable types.
> 
> While it works in little endian archs to store a big type e.g.
> int/long at some address and reading then only the lower bits
> by using the same address but a smaller data type, e.g. char/int
> it does not work with bigendian because the value is in the
> upper bits and so the upper bits are read instead of the lower.
> 
> The an other differnece between linux-i386 and linux-ppc is
> that the PPC instruction set obviosly can do unsigned 8 bit
> arithmetic better(with less instructions possibly) so the
> gcc of linux-ppc defined "char" as "unsigned char" while
> most other archichtiectures define "char" as "signed char".
> (ISO C lets this open to the to architecture)
> 
> You can force gcc to define "char" as "signed char" by adding
> -fsigned-char to the gcc compile flags. But you should be
> recompiling all object files which might just use "char"
> for external data types and don't specify signed/unsigned.
> 
> You could also try to add -Wall -W to force all warnings on,
> log the output of both runs, grep the warings into separate
> files and compare the warnings beween i386 and ppc when the
> same compiler options were used.
> 
> An other thing you could try is to use -O0 -g as compilation
> options in addition, which disables all optimisations and add
> debugging space to work around possible problems...
> 
> I hope this includes some helpful hint to find it, or get
> it working, besides checking the code itself.
> Bernd
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: ObjectStore.
> If flattening out C++ or Java code to make your application fit in a
> relational database is painful, don't do it! Check out ObjectStore.
> Now part of Progress Software. http://www.objectstore.net/sourceforge
> _______________________________________________
> Torcs-users mailing list
> Tor...@li...
> https://lists.sourceforge.net/lists/listinfo/torcs-users
> 
> 


-- 
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=
                        TORCS
             The Open Racing Car Simulator
         AKA The Other Release Coming Soon (Skin'r)
                  http://torcs.org
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=

On Sun, 25 May 2003, David MENTRE wrote:

> David MENTRE <dav...@wa...> writes:
>
> > However, I have know another error. When launching a race, all the data
> > is loaded correctly, the sound is available but the screen freezes
> > without display the car, with the word Ready.
> >
> > On the console, I have the following messages:
> > david@microbe:~$ torcs
> > WARNING: slScheduler: Too many pending callback events!
> > WARNING: slScheduler: Too many pending callback events!
>
> As a side not, without sound, the display is correct (with about 40
> fps).

Very interesting, good to know that the display works without sound.

There are some possilities to check what might be causing the
messages from slScheduler when sound is on.

The man difference of the ppc processor is that it's bigendian,
and this makes a difference when using addresses with different
variable types.

While it works in little endian archs to store a big type e.g.
int/long at some address and reading then only the lower bits
by using the same address but a smaller data type, e.g. char/int
it does not work with bigendian because the value is in the
upper bits and so the upper bits are read instead of the lower.

The an other differnece between linux-i386 and linux-ppc is
that the PPC instruction set obviosly can do unsigned 8 bit
arithmetic better(with less instructions possibly) so the
gcc of linux-ppc defined "char" as "unsigned char" while
most other archichtiectures define "char" as "signed char".
(ISO C lets this open to the to architecture)

You can force gcc to define "char" as "signed char" by adding
-fsigned-char to the gcc compile flags. But you should be
recompiling all object files which might just use "char"
for external data types and don't specify signed/unsigned.

You could also try to add -Wall -W to force all warnings on,
log the output of both runs, grep the warings into separate
files and compare the warnings beween i386 and ppc when the
same compiler options were used.

An other thing you could try is to use -O0 -g as compilation
options in addition, which disables all optimisations and add
debugging space to work around possible problems...

I hope this includes some helpful hint to find it, or get
it working, besides checking the code itself.
Bernd

David MENTRE <dav...@wa...> writes:

> However, I have know another error. When launching a race, all the data
> is loaded correctly, the sound is available but the screen freezes
> without display the car, with the word Ready.
> 
> On the console, I have the following messages:
> david@microbe:~$ torcs 
> WARNING: slScheduler: Too many pending callback events!
> WARNING: slScheduler: Too many pending callback events!

As a side not, without sound, the display is correct (with about 40
fps).


Regards,
d.
-- 
 dav...@wa...

David MENTRE <dav...@wa...> writes:

> However, on the powerpc platform, I have no sound. Here is the error
> message:
> david@microbe:~$ torcs
> slDSP: open: Device or resource busy
> WARNING: slScheduler: soundcard init failed.

I have been able to fix that. The esd daemon under gnome was keeping the
audio device.


However, I have know another error. When launching a race, all the data
is loaded correctly, the sound is available but the screen freezes
without display the car, with the word Ready.

On the console, I have the following messages:
david@microbe:~$ torcs 
WARNING: slScheduler: Too many pending callback events!
WARNING: slScheduler: Too many pending callback events!
WARNING: slScheduler: Too many pending callback events!
WARNING: slScheduler: Too many pending callback events!
WARNING: slScheduler: Too many pending callback events!


Any idea how to fix that?

Best regards,
david
-- 
 dav...@wa...

Hello,

I've successfully compiled and run Torcs 1.2.1 on two platforms:

 - i386 (debian 3.0r1 woody)
 - powerpc (debian 3.0r1 woody)

However, on the powerpc platform, I have no sound. Here is the error
message:
david@microbe:~$ torcs
slDSP: open: Device or resource busy
WARNING: slScheduler: soundcard init failed.

I am able to launch another program playing sound. Any idea to dig into
this issue? I have done an strace but it did not shown the opened audio
device.

I can play audio with other programs, so I think access rights on audio
devices are correct.


BTW, in the sources of torcs-1.2.1, the file
src/libs/txml/gennmtab/gennmtab.o should not be included in the tar
file. This is a i386 binary file. :) Erasing it allows to continue the
compilation on PowerPC platform.


Best regards,
david
-- 
 dav...@wa...

Hi all

You can download a "pre-release" of an improved tutorial robot on
http://n.ethz.ch/student/hwymann//aboutme/publications/bt20030518.tar.gz

The tutorial chapter 8 will probably follow till the 29.6.2003.

Changes to bt chapter 7:
------------------------
- Pit stops and strategy
- Some improvments mentioned in the tutorial
- 5 pairs of cars

bye, have fun, bernhard.


-- 

visit my homepage http://n.ethz.ch/student/hwymann

Hi all

today I've uploaded the robot tutorial chapter 7. It features the 
implementation of collision avoidance and overtaking. As usual you can 
find it on http://n.ethz.ch/student/hwymann in the torcs section.

bye, have fun

bernhard.


-- 

visit my homepage http://n.ethz.ch/student/hwymann

Am Mon, 2003-05-12 um 00.42 schrieb Pavel:

> 2. Why my opponents drive like me? (i'm go right, and they go right, etc.), 
> may be i need compile my own driver? I'm try, but not happened... 

Maybe you should download the torcs-robots an compile torcs with them?

Gernot

Pavel wrote:
>>>i'm use torcs v1.2.1 and can't find rally cars. May be i need some package?
>>
>>The rally cars are in the "data-cars-Patwo-Design" package.
> 
> Already download, and install. Thanks!
> 
> I have else questions: (Sorry for my english, i'm russian...)
> 1. Torcs have a replays? It's my like thing. I Like to see for my driving... 
> :)

It's not yet done.

> 2. Why my opponents drive like me? (i'm go right, and they go right, etc.), 

Have you selected multiple players ? I did not protected the game from
selecting multiple players... so every player do the same thing.

Real opponents are the robots.

> may be i need compile my own driver? I'm try, but not happened... 

The "Berniw's robot tutorial" is a very good start.

> 
> PS: Who played on Sony Playstation in game "Gran Tourismo"? The engine of 
> torcs very like with it.  	

:-)))

> 

bye,
Eric.
-- 
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=
                        TORCS
             The Open Racing Car Simulator
         AKA The Other Release Coming Soon (Skin'r)
                  http://torcs.org
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=

>> i'm use torcs v1.2.1 and can't find rally cars. May be i need some package?
> The rally cars are in the "data-cars-Patwo-Design" package.
Already download, and install. Thanks!

I have else questions: (Sorry for my english, i'm russian...)
1. Torcs have a replays? It's my like thing. I Like to see for my driving... 
:)
2. Why my opponents drive like me? (i'm go right, and they go right, etc.), 
may be i need compile my own driver? I'm try, but not happened... 

PS: Who played on Sony Playstation in game "Gran Tourismo"? The engine of 
torcs very like with it.  	

-- 
Kmail v1.4.1

Hi,

The rally cars are in the "data-cars-Patwo-Design" package.
Check the download and install instructions:
<http://torcs.sourceforge.net/sections.php?op=viewarticle&artid=3>
Note that the rally cars are not licensed under GPL.

The robots using rally cars are in the "robots-base" and the "robots-berniw"
packages.

Bye,
Eric.

Pavel wrote:
> Hi all!
> i'm use torcs v1.2.1 and can't find rally cars. May be i need some package?
> 



-- 
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=
                        TORCS
             The Open Racing Car Simulator
         AKA The Other Release Coming Soon (Skin'r)
                  http://torcs.org
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=

On Sun, 11 May 2003, Pavel wrote:

> Hi all!
> i'm use torcs v1.2.1 and can't find rally cars. May be i need some package?

Did you install the src-robots-base package/tarball?
Bernhard

Hi all!
i'm use torcs v1.2.1 and can't find rally cars. May be i need some package?

-- 
Kmail v1.4.1

to...@to... wrote:
> I can't get at torcs.org right now so I can't check with the latest 
> version. However for 1.1.0 there were a couple of project files in 
> German. I'm guessing VC doesn't understand German unless it's the 
> German version, because these wouldn't load on mine. (An example is 
> berniw2.dsp.)

berniw2 project is not in the TORC-1.2.1-src.tgz archive. So
Visual Studio should be complaining about missing projects (at least
for the missing robots) and not about corrupted ones.

> 
> You can fix this by loading up a .dsp file you've saved yourself, 
> loading up the offending .dsp file, and copying the native message 
> (in English, the top 25 lines or so of comments and !MESSAGE) over 
> the foreign one. Seems VC actually checks this message for 
> correctness
> 
> --Tom
> 

Bye,
Eric.
-- 
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=
                        TORCS
             The Open Racing Car Simulator
         AKA The Other Release Coming Soon (Skin'r)
                  http://torcs.org
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=

Dheeman Basu wrote:
> Hi,
> 
>   I have downloaded the binaries of this project first and was very 
> impressed with the quality of the software. Next thing I wanted is to 
> look at the code.
> 
> I have downloaded TORC-1.2.1-src.tgz  and uncopressed it with Winzip. I 
> tried to open the torcs-1.2.1\TORC.dsw with Visual Studio.net. The 
> studio complains about most of the projects contained inside to be 
> corrupt. Am I doing something wrong? Has anyone faced similar problem?

Can you list the projects corrupted ?

Thanks,
Eric.
-- 
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=
                        TORCS
             The Open Racing Car Simulator
         AKA The Other Release Coming Soon (Skin'r)
                  http://torcs.org
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=

I can't get at torcs.org right now so I can't check with the latest 
version. However for 1.1.0 there were a couple of project files in 
German. I'm guessing VC doesn't understand German unless it's the 
German version, because these wouldn't load on mine. (An example is 
berniw2.dsp.)

You can fix this by loading up a .dsp file you've saved yourself, 
loading up the offending .dsp file, and copying the native message 
(in English, the top 25 lines or so of comments and !MESSAGE) over 
the foreign one. Seems VC actually checks this message for 
correctness!

--Tom

On 3 May 2003 at 5:11, Dheeman Basu wrote:

> Hi,
> 
>    I have downloaded the binaries of this project first and was very 
> impressed with the quality of the software. Next thing I wanted is to look 
> at the code.
> 
> I have downloaded TORC-1.2.1-src.tgz  and uncopressed it with Winzip. I 
> tried to open the torcs-1.2.1\TORC.dsw with Visual Studio.net. The studio 
> complains about most of the projects contained inside to be corrupt. Am I 
> doing something wrong? Has anyone faced similar problem?
> 
> regards,
> 
> 
> Dheeman Basu
> 
> I do have downloaded the glut and plib and got them compiled with Studio.net 
> before attempting to open TORC. I think these two are not related, but still 
> prefered to mention.
> 
> 
> 
> 
> 
> _________________________________________________________________
> Visit Latin America. Discover! Explore! 
> http://www.coxandkingsusa.com/L2/latin_amer/latin_amer_L2.html Experience 
> it!
> 
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Torcs-users mailing list
> Tor...@li...
> https://lists.sourceforge.net/lists/listinfo/torcs-users
> 


-- 
--Tom

Bernhard Wymann wrote:
> Hi
> 
>> I have downloaded TORC-1.2.1-src.tgz  and uncopressed it with Winzip. 
>> I tried to open the torcs-1.2.1\TORC.dsw with Visual Studio.net. The 
>> studio complains about most of the projects contained inside to be 
>> corrupt. Am I doing something wrong? Has anyone faced similar problem?

May be because I have the bad habit to update the project files directly
with xemacs under Linux each time the files or functions change.
VC++ 6.0 is not complaining provided the lines end with ^M ;-)
I don't know what to do... I don't know how to force VC++ to
regenerate the project files.

> 
> 
> Hmm, like it looks it should (!!!) be a piece of cake to import the 
> projects... look at 
> http://www.dotnetextreme.com/articles/MigratingVs6VS7.asp
> 
> bye, bernhard
> 

Bye,
Eric.
-- 
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=
                        TORCS
             The Open Racing Car Simulator
         AKA The Other Release Coming Soon (Skin'r)
                  http://torcs.org
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=

Hi

> I have downloaded TORC-1.2.1-src.tgz  and uncopressed it with Winzip. I 
> tried to open the torcs-1.2.1\TORC.dsw with Visual Studio.net. The 
> studio complains about most of the projects contained inside to be 
> corrupt. Am I doing something wrong? Has anyone faced similar problem?

Hmm, like it looks it should (!!!) be a piece of cake to import the 
projects... look at 
http://www.dotnetextreme.com/articles/MigratingVs6VS7.asp

bye, bernhard

-- 

visit my homepage http://n.ethz.ch/student/hwymann

Hi

> I have downloaded TORC-1.2.1-src.tgz  and uncopressed it with Winzip. I 
> tried to open the torcs-1.2.1\TORC.dsw with Visual Studio.net. The 
> studio complains about most of the projects contained inside to be 
> corrupt. Am I doing something wrong? Has anyone faced similar problem?

The projects are for VC++ 6.0. I don't own the .NET bersion, so I can't 
tell you how to import the project, but perhaps there is a hint in the 
online help or one internet (google?).

> I do have downloaded the glut and plib and got them compiled with 
> Studio.net before attempting to open TORC. I think these two are not 
> related, but still prefered to mention.

Glut and plib are included in the TORCS sources.

bye, bernhard.

-- 

visit my homepage http://n.ethz.ch/student/hwymann

Hi,

   I have downloaded the binaries of this project first and was very 
impressed with the quality of the software. Next thing I wanted is to look 
at the code.

I have downloaded TORC-1.2.1-src.tgz  and uncopressed it with Winzip. I 
tried to open the torcs-1.2.1\TORC.dsw with Visual Studio.net. The studio 
complains about most of the projects contained inside to be corrupt. Am I 
doing something wrong? Has anyone faced similar problem?

regards,


Dheeman Basu

I do have downloaded the glut and plib and got them compiled with Studio.net 
before attempting to open TORC. I think these two are not related, but still 
prefered to mention.





_________________________________________________________________
Visit Latin America. Discover! Explore! 
http://www.coxandkingsusa.com/L2/latin_amer/latin_amer_L2.html Experience 
it!

Hi,

Toby Haynes wrote:
[...]
> 
> Staring into the simuv2 directory is pretty interesting. There is plenty 
> here I don't pretend to be able to follow (the code is a little light on 
> comments - at least in the cpp files - thank god for the comments in the 
> headers!) but most of the equations ring bells from my physics degree. I 
> can't find the weight transfer equations for the action of the car on 

The weight transfer is done in the car.cpp::SimCarUpdateForces lines 174 to
179, the weight transfer is expressed as moments.

> the suspension - maybe I'm just being blind. I can find the spring 
> equations for each wheel in susp.cpp. In wheel.cpp I can see that the 

The moments act on the car by modifying the pitch and roll angles.
This will determine the ride height of the wheels, and the wheel vertical force
is computed by the spring/damper reaction.
I agree it's not easy to see in the code ;-)

> weight on the wheel is considered as wheel->opLoad. This appears to come 
> direct from one of the parameters - is this altered dynamically 
> somewhere else? I can see that wheel->opLoad is derived from 
> wheel->weight0 and that wheel->weight0 is derived from the car mass and 
> the static car balance, but that car balance should be a function of the 
> forces acting on the car through the tyres I think?

All that are static data. wheel->weight0 is the initial weight and
wheel->opLoad is the operating load (the one indicated by the manufacturer),
this is used to compute the wheel behaviour in function of the load.

> 
> Am I just missing something here? What I think is happening is that when 
> braking the wheels are locking up and I'm not steering because of the 
> wheels locking up (I'm not using ABS). Without weight transfering 
> through onto the front wheels, the car understeers, whereas I think the 
> car should lose the back end in a big way.

to increase the weight transfer you'll have to set the gravity center higher
in the car definitions.

> 
> Thanks for any insight (and corrections!!)
> 
> Cheers,
> Toby Haynes
> 


Bye,
Eric.
-- 
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=
                        TORCS
             The Open Racing Car Simulator
         AKA The Other Release Coming Soon (Skin'r)
                  http://torcs.org
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=

Bernhard Wymann wrote:

> The reasons are the characteristics and the setup of the cars. For 
> racing cars the center of gravity is near the middle and the setup is 
> optimized for neutral car behaviour. If you want your rear slide out 
> you have to:
> - brake or accelarate tough within the turn
> - move the brake balance towards the rear
> - reduce the downforce of the rear spoiler
> - did you switch off ABS and ASR?
> - are you racing on semi-pro or pro?
>
> You can find the "human" driver setups in 
> /usr/local/share/games/torcs/drivers/human if you want to change them.

Changing the settings to pro certainly makes things act more as I would 
expect. I'm still puzzled by the noticeable understeer when, say, 
slamming the wheel full left and engaging the break on a corner. I'd 
expect most cars to press down the front tyres hard under braking and 
increase the maximum friction that they could apply, with the back tyres 
going 'light' and the rear of the car therefore attempting to overtake 
the front. This was observed on the unaltered Ford Focus OffRoad on one 
of the dirt tracks.

Staring into the simuv2 directory is pretty interesting. There is plenty 
here I don't pretend to be able to follow (the code is a little light on 
comments - at least in the cpp files - thank god for the comments in the 
headers!) but most of the equations ring bells from my physics degree. I 
can't find the weight transfer equations for the action of the car on 
the suspension - maybe I'm just being blind. I can find the spring 
equations for each wheel in susp.cpp. In wheel.cpp I can see that the 
weight on the wheel is considered as wheel->opLoad. This appears to come 
direct from one of the parameters - is this altered dynamically 
somewhere else? I can see that wheel->opLoad is derived from 
wheel->weight0 and that wheel->weight0 is derived from the car mass and 
the static car balance, but that car balance should be a function of the 
forces acting on the car through the tyres I think?

Am I just missing something here? What I think is happening is that when 
braking the wheels are locking up and I'm not steering because of the 
wheels locking up (I'm not using ABS). Without weight transfering 
through onto the front wheels, the car understeers, whereas I think the 
car should lose the back end in a big way.

Thanks for any insight (and corrections!!)

Cheers,
Toby Haynes