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

Thank you, Bernhard. It works now.


Bernhard Wymann a =E9crit :

> Hi Speedy
>
> Try this before running configure (do you have NVidia headers?):
>
> export CPPFLAGS=3D"${CPPFLAGS} -DGL_GLEXT_PROTOTYPES"
>
> If that not helps send as well the config.log, an perhaps the gl.h.=20
> But the simplest choice would be to follow the suggestion of the error=20
> message, I think it is quite clear.
>
> Bye, Bernhard.



--=20
SpeedyChonChon

----------------------------------------------------------
AIM : /SpeedyChonChon/
ICQ : /302814595/
MSN : /ab...@ho.../
http://speedy.chonchon.free.fr
----------------------------------------------------------

Hi Speedy

Try this before running configure (do you have NVidia headers?):

export CPPFLAGS="${CPPFLAGS} -DGL_GLEXT_PROTOTYPES"

If that not helps send as well the config.log, an perhaps the gl.h. But 
the simplest choice would be to follow the suggestion of the error 
message, I think it is quite clear.

Bye, Bernhard.


> Hi all.
> 
> ./configure give me the following error :
> /checking for GL_VERSION_1_3... configure: error: GL_VERSION_1_3 not 
> defined via gl.h. Please get the GL header files from mesa3d.org./
> But in my /usr/include/GL/gl.h I have :
> /#define GL_VERSION_1_1   1
> #define GL_VERSION_1_2   1
> #define GL_VERSION_1_3   1
> #define GL_ARB_imaging   1
> /I don't know what to do. I had no problem with 1.2.3. Maybe must I 
> modify the ./configure script but I don't know how.
> 
> Thank.
> Bye


-- 

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

Hi all.

./configure give me the following error :
/checking for GL_VERSION_1_3... configure: error: GL_VERSION_1_3 not 
defined via gl.h. Please get the GL header files from mesa3d.org./
But in my /usr/include/GL/gl.h I have :
/#define GL_VERSION_1_1   1
#define GL_VERSION_1_2   1
#define GL_VERSION_1_3   1
#define GL_ARB_imaging   1
/I don't know what to do. I had no problem with 1.2.3. Maybe must I 
modify the ./configure script but I don't know how.

Thank.
Bye
-- 
SpeedyChonChon

----------------------------------------------------------
AIM : /SpeedyChonChon/
ICQ : /302814595/
MSN : /ab...@ho.../
http://speedy.chonchon.free.fr
----------------------------------------------------------

Hi all.

You thing you are a boss on the track ? You thing you are the best ?
So, show us ! ! !

I am happy to tell you that the TDC is born.
The TDC (for Torcs Drivers Championship) is a hotlap competition.
It's here <http://speedy.chonchon.free.fr/indextdc.html>.

Have Fun.
Bye
-- 
SpeedyChonChon

----------------------------------------------------------
AIM : /SpeedyChonChon/
ICQ : /302814595/
MSN : /ab...@ho.../
http://speedy.chonchon.free.fr
----------------------------------------------------------

Hi T.P. Reitzel

> -- I will try to fix that for test 2, but the problem is that the 
> vendors change their header files at chance, so it is a moving target. 
> If you have NVidia header files installed, you can set the following 
> before running configure (for bash):
> 
> export CPPFLAGS="${CPPFLAGS} -DGL_GLEXT_PROTOTYPES"  --

An attemt to fix this is in the CVS, I tested it with the NVidia and the 
Mesa headers... so there is still the possibility that it fails with the 
ATI ones.

> BTW, I want to thank you, Bernhard, for your work on TORCS over the 
> years along with Eric. I especially appreciate
> Christos' work on adding OpenAL because of slScheduler problems with 
> ALSA's incomplete OSS emulation. Also, I want to

Thank you for that:-)

> thank Charalampos for the new track editor which will attract new talent 
> to TORCS. TORCS is so looooooooooong overdue for the recognition it 
> deserves, and I think that time is rapidly approaching. Even with its 

I do not care much about that, I see the releases more as an opportunity 
to attract people which contribute cool stuff and share the same hobby:-)

> flaws, my favorite racing
> simulation is still Grand Prix 4, but it's not made for Linux. rFactor 
> is really being hyped, but open source
> initiatives like TORCS have great potential which is slowly being 
> fulfilled.

The problem is that there are simply too much free racing sim projects, 
so the manpower is too distributed. The most projects end up dead, so 
the whole effort is kind of wasted (I mean in a "global" sense, of 
course it is not wasted if the people had a good time and fun doing it). 
If those people would have contributed to an existing project... well, I 
think you have got my point:-)

I had to make the decision as well if I start my own sim and decided use 
TORCS (I did not know back then that I would become involved) instead of 
starting another project.

I think the initial effort of Eric and Christophe is really outstanding, 
it is a huge difference to talk/dream about a racing sim and actually 
doing it (e.g. look at the Motorsport project at RSC, in the beginning 
where around 100 people who talked about helping, finally it is one 
person which does actually real work).

There are areas where I think (of course I hope I am wrong) we will 
never beat commercial products, e.g. the quality of artwork. In other 
areas we have already won IMHO, e.g when I compare the stability of 
TORCS and GTR, well... I think we are far ahead (I actuall have GTR and 
drove more than 2500 laps in ~120 Online races). I dream of beating them 
as well in the networking department, but we will see.

Bye, Bernhard.


-- 

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

Bernhard,
 

-- I will try to fix that for test 2, but the problem is that the vendors 
 change their header files at chance, so it is a moving target. If you 
 have NVidia header files installed, you can set the following before 
 running configure (for bash):
 
 export CPPFLAGS="${CPPFLAGS} -DGL_GLEXT_PROTOTYPES"  --


Ok, I'll look forward to the fix if you can get it done. Yes, I'm using nVidia's OpenGL drivers, v. 7667, for Linux. Your suggested
fix works perfectly. 

BTW, I want to thank you, Bernhard, for your work on TORCS over the years along with Eric. I especially appreciate
Christos' work on adding OpenAL because of slScheduler problems with ALSA's incomplete OSS emulation. Also, I want to
thank Charalampos for the new track editor which will attract new talent to TORCS. TORCS is so looooooooooong overdue 
for the recognition it deserves, and I think that time is rapidly approaching. Even with its flaws, my favorite racing
simulation is still Grand Prix 4, but it's not made for Linux. rFactor is really being hyped, but open source
initiatives like TORCS have great potential which is slowly being fulfilled. 

THANKS guys and gals for TORCS!

Hi T.P. Reitzel

I will try to fix that for test 2, but the problem is that the vendors 
change their header files at chance, so it is a moving target. If you 
have NVidia header files installed, you can set the following before 
running configure (for bash):

export CPPFLAGS="${CPPFLAGS} -DGL_GLEXT_PROTOTYPES"

Bye, Bernhard.

-- 

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

Hi T.P. Reitzel

> Bernhard,
> 
> I get the following error during compilation of Test 1, TORCS 1.2.4:
> 
> g++ -I/usr/local/src/torcs-1.2.4/export/include 
> -I/usr/local/src/torcs-1.2.4 -g -O2 -Wall -fPIC -O2 -Wall -fPIC -O2  
> -D_SVID_SOURCE -D_BSD_SOURCE -DSHM -DHAVE_CONFIG_H  -c grmultitexstate.cpp
> grmultitexstate.cpp: In member function `virtual void
>   grMultiTexState::apply(int)':
> grmultitexstate.cpp:30: error: `glActiveTexture' undeclared (first use this
>   function)
> grmultitexstate.cpp:30: error: (Each undeclared identifier is reported only
>   once for each function it appears in.)
> make[5]: *** [grmultitexstate.o] Error 1
> make[5]: Leaving directory 
> `/usr/local/src/torcs-1.2.4/src/modules/graphic/ssggraph'
> make[4]: *** [subdirs] Error 1
> make[4]: Leaving directory `/usr/local/src/torcs-1.2.4/src/modules/graphic'
> make[3]: *** [subdirs] Error 1
> make[3]: Leaving directory `/usr/local/src/torcs-1.2.4/src/modules'
> make[2]: *** [subdirs] Error 1
> make[2]: Leaving directory `/usr/local/src/torcs-1.2.4/src'
> make[1]: *** [subdirs] Error 1
> make[1]: Leaving directory `/usr/local/src/torcs-1.2.4'
> make: *** [restart] Error 2

Hmm, my first answer was not that accurate, sorry for that (to the list: 
was sent private to T. P.), but I guess it is the same problem like the

4.2.6 Compiling stops with "error: `GL_TEXTURE0_ARB' undeclared".

from the FAQ. Which OpenGL header files do you use (vendor and version)? 
Replace the OpenGL header files with the ones from Mesa, you need at 
least gl.h and glext.h. Hope this helps.

Bye, Bernhard.


-- 

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

Bernhard,

I get the following error during compilation of Test 1, TORCS 1.2.4:

g++ -I/usr/local/src/torcs-1.2.4/export/include 
-I/usr/local/src/torcs-1.2.4 -g -O2 -Wall -fPIC -O2 -Wall -fPIC -O2  
-D_SVID_SOURCE -D_BSD_SOURCE -DSHM -DHAVE_CONFIG_H  -c grmultitexstate.cpp
grmultitexstate.cpp: In member function `virtual void
   grMultiTexState::apply(int)':
grmultitexstate.cpp:30: error: `glActiveTexture' undeclared (first use this
   function)
grmultitexstate.cpp:30: error: (Each undeclared identifier is reported only
   once for each function it appears in.)
make[5]: *** [grmultitexstate.o] Error 1
make[5]: Leaving directory 
`/usr/local/src/torcs-1.2.4/src/modules/graphic/ssggraph'
make[4]: *** [subdirs] Error 1
make[4]: Leaving directory `/usr/local/src/torcs-1.2.4/src/modules/graphic'
make[3]: *** [subdirs] Error 1
make[3]: Leaving directory `/usr/local/src/torcs-1.2.4/src/modules'
make[2]: *** [subdirs] Error 1
make[2]: Leaving directory `/usr/local/src/torcs-1.2.4/src'
make[1]: *** [subdirs] Error 1
make[1]: Leaving directory `/usr/local/src/torcs-1.2.4'
make: *** [restart] Error 2

Hi all

For the brave and curious there is a new source test release available, 
visit http://www.torcs.org for details. Bye, have fun,

Bernhard.


-- 

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

Hi

> Good work so far. Keep it up!  With your editor, I created my first 
> TORCS track. Before proceeding much further though, I'm having 
> difficulty matching the S(tart)/F(inish) lines so they're perfectly 
> even. Can this process be automated? The slight misalignment of the 
> track shows while driving across the S/F line.

I have auto adjustment in my todo list but is in low priority. Now i am
working in some kind of report of how much are you off. For now the only
way to work is to did a large zoom to connection area and fine adjust the
last couple of segments to make a good aligment.

Bye
Charalampos

Good work so far. Keep it up!  With your editor, I created my first 
TORCS track. Before proceeding much further though, I'm having 
difficulty matching the S(tart)/F(inish) lines so they're perfectly 
even. Can this process be automated? The slight misalignment of the 
track shows while driving across the S/F line.

TIA

I release this version to fix some bugs.

Hi Bernhard
> Btw. if you find such problems, could you please attach the track xml
> to 
> reproduce the problem, this would be helpful too.
Sure. I was trying to dont send large mails if there is no need.

Bye Charalampos 

Hi Charalampos

>>I recompile everything with the attached files in place of the defaults
>>but i still have segmentetion faults in trackgen.The problem appear
>>when i
>>try to create a track that exceeds a number of segments, depending of
>>straights and curve combination the number of segments that makes the
>>problem is around 25. If i remove a couple one segment then create the
>>track fine or alternatively works  if i remove the side, border,
>>barrier
>>data from segments.Can you please check it ?
> 
> 
> What i found is that the indices array is smaller than the number of
> indices the follow is part of the output of my test track:
> === Indices Array Size   = 15020
> === Vertex Array Size    = 45060
> === Tex Coord Array Size = 30040
> === Indices really used = 15024
> 
> I fix the problem by changing the line 435 in track.cpp, i change the
> safety margin from 1000 to 10000.
> 
> 431    nbvert++;
> 432    nbvert *= 30;
> 433    nbvert+=58; /* start bridge */
> 434    nbvert+=12 + 10 * Track->pits.driversPitsNb;
> 435    nbvert+=10000; /* safety margin */
> 436    printf("=== Indices Array Size   = %d\n", nbvert);
> 437    printf("=== Vertex Array Size    = %d\n", nbvert * 3);
> 438    printf("=== Tex Coord Array Size = %d\n", nbvert * 2);
> 439    trackindices  = (unsigned int*)malloc(sizeof(unsigned int)*nbvert);
> 440    trackvertices = (float*)malloc(sizeof(GLfloat)*(nbvert * 3));
> 441    tracktexcoord = (float*)malloc(sizeof(GLfloat)*(nbvert * 2));
> 
> this works for me but i dont know if it is the right solution. Please let
> me know if you make a better solution.

Thank you for figuring that out, this will help a lot. I will include 
your fix for test1 and if I find the time I will investigate the problem 
a bit more for test2. Thank you very much, great catch:-)

Btw. if you find such problems, could you please attach the track xml to 
reproduce the problem, this would be helpful too.

Bye, Bernhard.


-- 

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

> I recompile everything with the attached files in place of the defaults
> but i still have segmentetion faults in trackgen.The problem appear
> when i
> try to create a track that exceeds a number of segments, depending of
> straights and curve combination the number of segments that makes the
> problem is around 25. If i remove a couple one segment then create the
> track fine or alternatively works  if i remove the side, border,
> barrier
> data from segments.Can you please check it ?

What i found is that the indices array is smaller than the number of
indices the follow is part of the output of my test track:
=== Indices Array Size   = 15020
=== Vertex Array Size    = 45060
=== Tex Coord Array Size = 30040
=== Indices really used = 15024

I fix the problem by changing the line 435 in track.cpp, i change the
safety margin from 1000 to 10000.

431    nbvert++;
432    nbvert *= 30;
433    nbvert+=58; /* start bridge */
434    nbvert+=12 + 10 * Track->pits.driversPitsNb;
435    nbvert+=10000; /* safety margin */
436    printf("=== Indices Array Size   = %d\n", nbvert);
437    printf("=== Vertex Array Size    = %d\n", nbvert * 3);
438    printf("=== Tex Coord Array Size = %d\n", nbvert * 2);
439    trackindices  = (unsigned int*)malloc(sizeof(unsigned int)*nbvert);
440    trackvertices = (float*)malloc(sizeof(GLfloat)*(nbvert * 3));
441    tracktexcoord = (float*)malloc(sizeof(GLfloat)*(nbvert * 2));

this works for me but i dont know if it is the right solution. Please let
me know if you make a better solution.

Regards
Charalampos

Hi Bernhard

> attached the fixed files, should work with vanilla 1.2.3 sources, you 
> need to recompile the whole project:-(

I recompile everything with the attached files in place of the defaults
but i still have segmentetion faults in trackgen.The problem appear when i
try to create a track that exceeds a number of segments, depending of
straights and curve combination the number of segments that makes the
problem is around 25. If i remove a couple one segment then create the
track fine or alternatively works  if i remove the side, border, barrier
data from segments.Can you please check it ?

Bye 
Charalampos

Hi Charalampos

> Can't exec "/usr/bin/g++": No such file or directory at /usr/local/bin/g++ line 202.

You need to have g++ installed properly (more or less all tests fail 
because of this).

Bye, Bernhard.

-- 

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

> Weird (I have SuSE 8.2 and 9.3 in use, on 9.3 exactly the same
> package). 
> Can you attach your config.log?

Hi Bernhard 
Here is the config.log

Bye

Hi Charalampos

> I am trying to compile Torcs-1.2.3 from sources but the ./configure ends with:
> configure: error: Can't find GL/glut.h. freeglut can be found on
> http://freeglut.sourceforge.net/
> While glut.h is in the usr/include/GL
> Any advice ?
> My system is SuSe linux 9.3
> freeglut 2.2.0-83

Weird (I have SuSE 8.2 and 9.3 in use, on 9.3 exactly the same package). 
Can you attach your config.log?

Bye, Bernhard.

-- 

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

Hi
I am trying to compile Torcs-1.2.3 from sources but the ./configure ends with:
configure: error: Can't find GL/glut.h. freeglut can be found on
http://freeglut.sourceforge.net/
While glut.h is in the usr/include/GL
Any advice ?
My system is SuSe linux 9.3
freeglut 2.2.0-83

Than you in advance
VCharalampos Alexopoulos

Hi Charalampos

> Are this fix aply to Linux or it is just for Windows? because now i have
> problems in Linux.

Both. But the tgf fix is masked with #ifdefs and just has an effect on 
Windows. If you want the new Linux version simply get the CVS.

> "vanila 1.2.3 sources" Are you mean the CVS or what ?

I mean the version 1.2.3 of torcs, either the tarball of the release or 
a checkout of 1.2.3. The files should work with CVS HEAD as well, but 
HEAD does not yet compile in Windows.

A bit offtopic, the plan:

- Next Monday/Tuesday is razor day, I will decide finally what makes it 
into 1.2.4 (e.g. OpenAL, pit wall collisions, some more fixing, etc).

- Throw out 1.2.4-test1 till the end of the next week, Linux src only, 
might lack some parts.

- Finalizing, Windows port, more fixing -> test2, perhaps in 3-5 weeks, 
Linux src and Windows binary zipped.

- Waiting for feedback, fix showstoppers, depending on the discovered 
problems zero (ops, just dreaming), one or two more iterations (test3/4).

- Creating final packages/installer, release perhaps in 8 to 12 weeks.

Bye, Bernhard.

-- 

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

Hi Bernhard

> attached the fixed files, should work with vanilla 1.2.3 sources, you 
> need to recompile the whole project:-(

Are this fix aply to Linux or it is just for Windows? because now i have
problems in Linux.

"vanila 1.2.3 sources" Are you mean the CVS or what ?

Charalampos

Hi Bernhard

> attached the fixed files, should work with vanilla 1.2.3 sources, you 
> need to recompile the whole project:-(

Thank you for the response, i was thoughting that is the Xml that makes
trackgen crash, while the problem coming from trackgen its up to windows
users to install the patch. In my windows machine i install torcs just to
test and i have to remove it.

Bye
Charalampos

Cheers, this works.  The track editor is aaamazing!

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