Simu updates in CVS

The following updates:

wheel.cpp
---------

Removed all the complicated frame of reference change calculations and
also the not-so-necessary code that calculates the exact length of a
suspension.

The first bit was only of use for when the car was at an angle with
respect to the road surface, which is not very common and in any case
was never done very well.

The suspension length bit was useful in all cases where the road is 
not
flat. Without it the wheels can go a bit 'inside' the track, but I 
don't
think this is very important.

So, why the code was simplified?

Well, firstly, the 'calculation of forces when the wheels don't touch
the track properly' is a very tricky business. So it never worked 
properly.

Secondly, since the wheel code executes four times per car, any extra
code really has a strong impact. So, removed code.

Is there an impact in physics?
The quick asnwer is: yes, there is, but it is limited. In particular,

glut.h not found (although installed) on OpenSuSE 10.0

Hi,
the ./configure step fails on OpenSuSE 10.0 although
freeglut is installed (via YaST setup).

The last lines before ./configure quits:

checking for unistd.h... yes
checking GL/gl.h usability... yes
checking GL/gl.h presence... yes
checking for GL/gl.h... yes
checking GL/glut.h usability... no
checking GL/glut.h presence... no
checking for GL/glut.h... no
configure: error: Can't find GL/glut.h. freeglut can be found on http://freeglut.sourceforge.net/

glut.h itself is present at:

fe <at> garden:~/torcs/torcs> locate glut.h
/home/fe/torcs/torcs/src/windows/include/GL/glut.h
/usr/X11R6/include/GL/freeglut.h
/usr/X11R6/include/GL/glut.h

Could /usr/X11R6/include/ be added as additional path?

Thx,
Frieder

PS: gl.h is present (and found) at /usr/include/GL/gl.h

-------------------------------------------------------

Re: glut.h not found (although installed) on OpenSuSE 10.0

Hi Frieder

> Hi,
> the ./configure step fails on OpenSuSE 10.0 although
> freeglut is installed (via YaST setup).

Hmm, that's odd, I have exactly this setup here, and it works. Could you send as
well the config.log, perhaps this help to track it down.

> fe <at> garden:~/torcs/torcs> locate glut.h
> /home/fe/torcs/torcs/src/windows/include/GL/glut.h
> /usr/X11R6/include/GL/freeglut.h
> /usr/X11R6/include/GL/glut.h

Hmm, usually the GL headers reside in /usr/include/GL, have you updated your
system from an older version (I have fresh SuSE 10.0 and the headers are in
/usr/include/GL). Looks as well like you have somehow 2 versions on your system,
the problem could be caused through this.

Bye, Bernhard.

-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click

Bug Map

Re: Bug Map

Hi Speedy

> Sometimes I have a weirb display in the map zone. (behind the map I see
> something like the screen I have when the game is loading (see the
> image(It's a little image I hope it will pass on the mailing))).

Welcome at NVidia I think it is a driver "optimization" (I think it does not
draw obscured window parts in the back buffer, so the fetched map is not complete).

The fix will be to render the track map into a pbuffer and to fetch it from
there, but I think it is not very urgent (it never happend to me in fullscreen,
which most people will usually use in games, and if I remember correct it
happens just with X and not with Windows).

Bye, Bernhard.

-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click

Re: Bug Map

Am Sonntag, den 18.12.2005, 10:09 +0100 schrieb Bernhard Wymann:
> Hi Speedy
> 
> > Sometimes I have a weirb display in the map zone. (behind the map I see
> > something like the screen I have when the game is loading (see the
> > image(It's a little image I hope it will pass on the mailing))).
> 
> Welcome at NVidia I think it is a driver "optimization" (I think it does not
> draw obscured window parts in the back buffer, so the fetched map is not complete).

... FYI ...

This is not an NVidia feature. Most of the open source DRI drivers do
the same because all applications share a screen-sized back and depth
buffer. In this case it's not an optimization but it is necessary to
prevent windows in the background from corrupting the window in the
foreground. To make it even worse, the open source drivers don't support
pbuffers yet. FWIW I also get corruption in the map when the bottom of
the Torcs window is slightly off-screen (this is with ATI's proprietary
Linux driver).

> 
> The fix will be to render the track map into a pbuffer and to fetch it from
> there, but I think it is not very urgent (it never happend to me in fullscreen,
> which most people will usually use in games, and if I remember correct it
> happens just with X and not with Windows).

Fullscreen mode doesn't work very well for me on Linux with a
wide-screen display. 

Regards,
  Felix

> 
> Bye, Bernhard.
> 

--

-- 
| Felix Kühling <fxkuehl <at> gmx.de>                     http://fxk.de.vu |
| PGP Fingerprint: 6A3C 9566 5B30 DDED 73C3  B152 151C 5CC1 D888 E595 |

-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_idv37&alloc_id865&op=click

grskidmarks.cpp CVS update

Temporary fix, trying to make sure that
skids are above the road surface. (This had been there for some time,
but with a  z_adjust > 1, which had the opposite effect, namely, of
putting the skids under the road surface. Duh. Now I cleaned it up
slightly and corrected the constant.)

This is needed because the wheel position depends on the current
physical model road height, which is not exactly
the same as the gfx road height, even if it is so
on average, because the physical model adds some
sinewave to the road height to simulate uneveness
of the track.

So, does it work? There is a very good improvement, but on some tracks
(g-track-2 and g-track-3) skidmarks still flicker a bit. At least now we
are sure that the skidmarks are actually over the track surface and the
problem is due to z-buffering.

A better fix would be to add a routine grTrackHeightL(tTrkLocPos *p),
similar to  TrTrackHeightL(), but which aim to give the height
of the graphical track. This would also take care of the fact that the
wheels are slightly inside the track whenever the road is tilted (due to
simplifications in the suspension length computation).

--

-- 
Christos Dimitrakakis
Homepage: http://www.idiap.ch/~dimitrak/main.html
Music: http://olethros.dmusic.com

-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click

CVS update: Curbs fix

I have finally solved the problem with the curbs - update on CVS,
rttrack.cpp.

Attached is a test track that illustrates the problem and the fix.

(though it's better to try it under simuv3, it should work in simuv2 in
the parts of the track where the angles are not very steep)

--

-- 
Christos Dimitrakakis
Homepage: http://www.idiap.ch/~dimitrak/main.html
Music: http://olethros.dmusic.com

Segmentation Faults - Bad Category

Hi all.

I was working on the 962C when TORCS didn't want run any more : 
Segmentation Faults. I had tryed every thing I could. After several lost 
hours (uninstall, compile, reinstall... several times), I found that it 
was the category definition of the 962C which was bad (I put History 
instead of Historic).

Isn't the segmentation Fault avoidable in this case ?

Bye
Speedy

-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click