plib-users

Re: [Plib-users] Memory Leaks

[Steve B. <sjb...@ai...>]

"Alejo, Larry W." wrote:
> 
> Has anyone tried to incorporate plib into a MSVC++ MFC application?

Hmmm - I doubt it.  The whole idea is to use PLIB to make PORTABLE
applications - using MFC utterly guarantees that your code will never
run on a non-Windoze platform and I'll be forced to flame you mercilessly
for idiotically (and unnecessarily) tying your all your hard work to the
worst operating system on the planet.  (Not that I'm biassed or anything :-).

Hence, you won't get *any* help/sympathy from me to get that going.

To the contrary - if there is anything I can do to PLIB to utterly
ensure 100% incompatibility with MFC, I'll do it.   It's not that I'm
opposed to running PLIB programs under Windoze - it's that I'm opposed
to people mindlessly writing code that *won't* run elsewhere.  That's
THE ENTIRE REASON that I wrote PLIB and chose to give it away for nothing
and spend thousands of hours of my time maintaining it.  It would have
been much easier to write a Linux-only package!
 
> I am in the process of developing an MFC application utilizing plib
> (mainly ssg) and the current snapshot of plib in cvs.sourceforge.net
> produces a multiple number of memory leaks, as reported by BoundsChecker.
> 
> I have made an attempt, on my version of plib, to squash a few of these
> memory leaks and respectfully offer the following code changes to the
> contributors of plib for consideration and inclusion into the library.

Thanks!  Good catch.

> I also discovered conflicting multiple definitions for macro _SSG_PUBLIC:
> ssglocal.h:
> #define _SSG_PUBLIC  public
> 
> ssg.h:
> #ifndef _SSG_PUBLIC
> #define _SSG_PUBLIC  protected
> #endif

That's 100% intentional.  Those definitions are not "conflicting" - the
code in ssg.h is checking to see if the definition in ssgLocal.h is present
and explicitly NOT defining _SSG_PUBLIC if it is.

I *hate* the fact that C++ doesn't provide a way to have selected members
be private as far as the application is concerned and public as far as
other functions in the same library is concerned (you can use 'friend' but
that's an awfully blunt instrument because it makes *ALL* fields accessible
in that way).

This mechanism (in effect) creates a new security class for C++ members
'_SSG_PUBLIC' - which is public for code within SSG and protected as
far as the application is concerned.

It works because only members of the SSG library #include "ssgLocal.h"
(hence the name 'Local') and hence #define _SSG_PUBLIC as 'public'. For
general applications code (who *don't* define _SSG_PUBLIC), they'll get
those same fields declared as 'protected'.

Kinda hackish - but then that's the kind of person I am!

----------------------------- Steve Baker -------------------------------
HomeMail : <sjb...@ai...>   WorkMail: <sj...@li...>
HomePage : http://web2.airmail.net/sjbaker1
Projects : http://plib.sf.net     http://tuxaqfh.sf.net  http://tuxkart.sf.net
           http://agtoys.sf.net   http://prettypoly.sf.net
           http://freeglut.sf.net http://toobular.sf.net

[Plib-users] Memory Leaks

[Alejo, L. W. <al...@ws...>]

Has anyone tried to incorporate plib into a MSVC++ MFC application?

I am in the process of developing an MFC application utilizing plib 
(mainly ssg) and the current snapshot of plib in cvs.sourceforge.net
produces a multiple number of memory leaks, as reported by BoundsChecker. 

I have made an attempt, on my version of plib, to squash a few of these
memory leaks and respectfully offer the following code changes to the
contributors of plib for consideration and inclusion into the library.

Changes:

In ssg.h:

virtual ~ssgLoaderOptions()
{ 
  if (model_dir)
  {
    delete model_dir;
    model_dir = 0;
  }

  if (texture_dir)
  {
    delete texture_dir;
    texture_dir = 0;
  }
}


In ssgTexture.cxx:

ssgTexture::~ssgTexture (void)
{
  if (filename != NULL)
  {
    delete [] filename;
    filename = NULL;
  }

  free_handle () ;
}


In ssgLoadSGI.cxx, added dtor:

ssgSGIHeader::~ssgSGIHeader(void)
{
  if (start != NULL)
    delete [] start;

  if (leng != NULL)
    delete [] leng;
}

I also discovered conflicting multiple definitions for macro _SSG_PUBLIC:
ssglocal.h:
#define _SSG_PUBLIC  public
 
ssg.h:
#ifndef _SSG_PUBLIC
#define _SSG_PUBLIC  protected
#endif

-Larry

RE: [Plib-users] sl.h and ul.h in cygwin

[Norman V. <nh...@ca...>]

You need to add -DWIN32 or -mwin32
to your programs compiler FLAGS

< cygwin does not do this aytomatically like the other
  windows compilers in that it might be building an X program>

This should be done automagically for you by the configure script
when building the library and / or the demos

Norman
  -----Original Message-----
  From: pli...@li...
[mailto:pli...@li...]On Behalf Of Claude Eckel
  Sent: Tuesday, May 22, 2001 1:26 PM
  To: pli...@li...
  Subject: [Plib-users] sl.h and ul.h in cygwin


  I just compiled PLIB1.3.1 with cygwin. everything worked fine. But then I
wanted to compile a game by me and i got the following errors:
  In file included from........ul.h:61: GL/glx: No such file or directory

  In file included from........sl.h:20: #error "Port me !"

[Plib-users] sl.h and ul.h in cygwin

[Claude E. <pin...@we...>]

I just compiled PLIB1.3.1 with cygwin. everything worked fine. But then =
I wanted to compile a game by me and i got the following errors:
In file included from........ul.h:61: GL/glx: No such file or directory

In file included from........sl.h:20: #error "Port me !"

Re: [Plib-users] Compile error `DBL_MAX' undeclared?

[Michael L. S. <ms...@mi...>]

Thank You! That helped!

Norman Vine wrote:

> Michael Larry Strean writes:
> 
>> I'm compiling plib-1.3.1 & I get
>> 
>> c++ -DPACKAGE=\"plib\" -DVERSION=\"1.3.1\" -DHAVE_LIBDL=1 
>> -DHAVE_LIBGL=1 
>> -DHAVE_LIBGLU=1 -DHAVE_LIBGLUT=1 -DSTDC_HEADERS=1 -DHAVE_GL_GL_H=1 
>> -DHAVE_GL_GLU_H=1 -DLINUX_JOYSTICK_IS_PRESENT=1 
>> -DGLUT_IS_PRESENT=1  -I. 
>> -I.  -I../../src/sg -I../../src/fnt -I../../src/util  
>> -I/usr/local/include  -g -O2 -O6 -Wall -c pu.cxx
>> In file included from pu.h:4,
>>                from puLocal.h:2,
>>                from pu.cxx:2:
>> ../../src/sg/sg.h: In method `void sgBox::empty ()':
>> ../../src/sg/sg.h:844: `FLT_MAX' undeclared (first use this function)
> 
> 
> you will need to add
> #include <float.h> 
> 
> in ul.h
> 
> This is fixed in the current CVS 
> and will be in the upcoming release.
> 
> Norman VIne
> 
> _______________________________________________
> plib-users mailing list
> pli...@li...
> http://lists.sourceforge.net/lists/listinfo/plib-users
> 
> 

RE: [Plib-users] Compile error `DBL_MAX' undeclared?

[Norman V. <nh...@ca...>]

Michael Larry Strean writes:
>
>I'm compiling plib-1.3.1 & I get
>
>c++ -DPACKAGE=\"plib\" -DVERSION=\"1.3.1\" -DHAVE_LIBDL=1 
>-DHAVE_LIBGL=1 
>-DHAVE_LIBGLU=1 -DHAVE_LIBGLUT=1 -DSTDC_HEADERS=1 -DHAVE_GL_GL_H=1 
>-DHAVE_GL_GLU_H=1 -DLINUX_JOYSTICK_IS_PRESENT=1 
>-DGLUT_IS_PRESENT=1  -I. 
>-I.  -I../../src/sg -I../../src/fnt -I../../src/util  
>-I/usr/local/include  -g -O2 -O6 -Wall -c pu.cxx
>In file included from pu.h:4,
>                from puLocal.h:2,
>                from pu.cxx:2:
>../../src/sg/sg.h: In method `void sgBox::empty ()':
>../../src/sg/sg.h:844: `FLT_MAX' undeclared (first use this function)

you will need to add
#include <float.h> 

in ul.h

This is fixed in the current CVS 
and will be in the upcoming release.

Norman VIne

[Plib-users] Compile error `DBL_MAX' undeclared?

[Michael L. S. <ms...@mi...>]

Hi!

I'm compiling plib-1.3.1 & I get

c++ -DPACKAGE=\"plib\" -DVERSION=\"1.3.1\" -DHAVE_LIBDL=1 -DHAVE_LIBGL=1 
-DHAVE_LIBGLU=1 -DHAVE_LIBGLUT=1 -DSTDC_HEADERS=1 -DHAVE_GL_GL_H=1 
-DHAVE_GL_GLU_H=1 -DLINUX_JOYSTICK_IS_PRESENT=1 -DGLUT_IS_PRESENT=1  -I. 
-I.  -I../../src/sg -I../../src/fnt -I../../src/util  
-I/usr/local/include  -g -O2 -O6 -Wall -c pu.cxx
In file included from pu.h:4,
                from puLocal.h:2,
                from pu.cxx:2:
../../src/sg/sg.h: In method `void sgBox::empty ()':
../../src/sg/sg.h:844: `FLT_MAX' undeclared (first use this function)
../../src/sg/sg.h:844: (Each undeclared identifier is reported only
once for each function it appears in.)
../../src/sg/sg.h: In method `void sgdBox::empty ()':
../../src/sg/sg.h:2016: `DBL_MAX' undeclared (first use this function)
make[2]: *** [pu.o] Error 1
make[2]: Leaving directory `/home/Michael/plib-1.3.1/src/pui'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/Michael/plib-1.3.1/src'
make: *** [all-recursive] Error 1
[Michael@localhost plib-1.3.1]$

I'm compiling on a new install of RH 7.1 (the only other thing I've done 
to this
system is install Mesa 3.4.2)
here's the output of ./configure

loading cache ./config.cache
checking for a BSD compatible install... (cached) /usr/bin/install -c
checking whether build environment is sane... yes
checking whether make sets ${MAKE}... (cached) yes
checking for working aclocal... found
checking for working autoconf... found
checking for working automake... found
checking for working autoheader... found
checking for working makeinfo... found
includedir changed to ${prefix}/include/plib libdir is ${exec_prefix}/lib
checking for gcc... (cached) gcc
checking whether the C compiler (gcc  ) works... yes
checking whether the C compiler (gcc  ) is a cross-compiler... no
checking whether we are using GNU C... (cached) yes
checking whether gcc accepts -g... (cached) yes
checking how to run the C preprocessor... (cached) gcc -E
checking for c++... (cached) c++
checking whether the C++ compiler (c++  ) works... yes
checking whether the C++ compiler (c++  ) is a cross-compiler... no
checking whether we are using GNU C++... (cached) yes
checking whether c++ accepts -g... (cached) yes
checking how to run the C++ preprocessor... (cached) c++ -E
checking for a BSD compatible install... /usr/bin/install -c
checking for ranlib... (cached) ranlib
checking for windows.h... (cached) no
checking for X... (cached) libraries /usr/X11R6/lib, headers 
/usr/X11R6/include
checking for dnet_ntoa in -ldnet... (cached) no
checking for dnet_ntoa in -ldnet_stub... (cached) no
checking for gethostbyname... (cached) yes
checking for connect... (cached) yes
checking for remove... (cached) yes
checking for shmat... (cached) yes
checking for IceConnectionNumber in -lICE... (cached) yes
checking for dlclose in -ldl... (cached) yes
checking for glNewList in -lGL... (cached) yes
checking for gluLookAt in -lGLU... (cached) yes
checking for glutGetModifiers in -lfreeglut... (cached) no
checking for glutGetModifiers in -lglut... (cached) yes
checking for ALopenport in -laudio... (cached) no
checking for ANSI C header files... (cached) yes
checking for GL/gl.h... (cached) yes
checking for GL/glu.h... (cached) yes
checking for windows.h... (cached) no
checking for joystick.h... (cached) no
checking for linux/joystick.h... (cached) yes
checking for GL/glut.h... (cached) yes
checking for GL/freeglut.h... (cached) no
creating ./config.status
creating Makefile
creating src/Makefile
creating src/js/Makefile
creating src/util/Makefile
creating src/pui/Makefile
creating src/sg/Makefile
creating src/sl/Makefile
creating src/ssg/Makefile
creating src/fnt/Makefile

Re: [Plib-users] compiling with mingw32

[Steve B. <sjb...@ai...>]

Claude Eckel wrote:
> 
> Has anybody ever tried to compile plib with mingw32?

Someone did try it - but it was more than a year ago and they never did
get it done (or if they did, they didn't send me any updates as a result).

> can anybody send me
> makefiles? because "configure" doesn't work under window$ I can't generate
> them.

Well, you could always download the Cygwin package - configure works just
fine under that - and if mingw32 is your default C++ compiler, I guess it'll
find it and Do The Right Thing.

We can't send the Makefile's as a part of the distribution because that would
screw up the autoconfiguration stuff for every other C++ system.

> And if I try to generate them with dev-c++ (mybe some of you know it)
> nothing does work anymore. Borland 5.02 doesn't work, too. After I spent an
> eternity with adding all the files I became an error in every file. I was
> too lazy to change the damned configuration of BCC so I ask you.

Get Cygwin - it seems to work just fine.

----------------------------- Steve Baker -------------------------------
HomeMail : <sjb...@ai...>   WorkMail: <sj...@li...>
HomePage : http://web2.airmail.net/sjbaker1
Projects : http://plib.sf.net     http://tuxaqfh.sf.net  http://tuxkart.sf.net
           http://agtoys.sf.net   http://prettypoly.sf.net
           http://freeglut.sf.net http://toobular.sf.net

RE: [Plib-users] compiling with mingw32

[Norman V. <nh...@ca...>]

Claude Eckel writes:
>
>Has anybody ever tried to compile plib with mingw32? 

Not to my knowledge

>can anybody send me makefiles? 
>because "configure" doesn't work under window$ 

Configure works fine under Windows with the Cygwin Toolkit
ask over on the MingW32 list how to set it up for MinGW.

Min...@li...
http://lists.sourceforge.net/lists/listinfo/mingw-users

Cheers

Norman

[Plib-users] compiling with mingw32

[Claude E. <pin...@we...>]

Has anybody ever tried to compile plib with mingw32? can anybody send me
makefiles? because "configure" doesn't work under window$ I can't generat=
e
them. And if I try to generate them with dev-c++ (mybe some of you know i=
t)
nothing does work anymore. Borland 5.02 doesn't work, too. After I spent =
an
eternity with adding all the files I became an error in every file. I was
too lazy to change the damned configuration of BCC so I ask you.

--=20

Lass's wackeln Baby, und mail mir wieder du St=FCck du!

ThE KiNg a.k.a. Gahre
formerly known as:
Thomas Gahr

pin...@we...
da....@we...
http://kickme.to/gahre
http://kickme.to/glinteractive

EmailExpress: 110...@pa...
ICQ Communication Center Address: http://wwp.icq.com/110882469

Re: [Plib-users] transparent material with texture ?

[Steve B. <sjb...@ai...>]

Marc Escher wrote:

> By the way, my scene is composed of translucent and opaque objects that
> can intersect each other and I have some problems of drawing order with
> the transparent objects. Is there a way in PLIB to manage that ? Can we
> enable the GL_DEPTH_TEST for the opaque object and disable it for the
> tranparent ones ?

PLIB renders all translucent polygons in no particular order *after* it's
rendered all opaque polygons in no particular order.

We have been discussing a mechanism to do fairly crude range-sorting
of the translucent polygons - but that's never a perfect solution.

Disabling depth testing for translucent objects does not help
the problem - that's a common myth.

You should read my FAQ on the subject (which is not PLIB-specific).

     http://web2.airmail.net/sjbaker1/alpha_sorting.html

The bottom line is that PLIB should allow translucent objects
to interact perfectly with opaque objects regardless.  HOWEVER
when multiple translucent object overlap, PLIB does not get it
right every time.

-- 
Steve Baker   HomeEmail: <sjb...@ai...>
              WorkEmail: <sj...@li...>
              HomePage : http://web2.airmail.net/sjbaker1
              Projects : http://plib.sourceforge.net
                         http://tuxaqfh.sourceforge.net
                         http://tuxkart.sourceforge.net
                         http://prettypoly.sourceforge.net
                         http://freeglut.sourceforge.net

Re: [Plib-users] transparent material with texture ?

[Marc E. <ma...@fa...>]

Steve Baker wrote:
>=20
> Marc Escher wrote:
>=20
> > I'm trying to render an object with a bmp texture (no alpha) and a
> > tranparent material associated to it, but I can't get the material
> > properties to be mixed with the texture. The object is displayed with
> > the texture but with no transparency or any color for the material. W=
hat
> > is the right way to do it ?
>=20
> You are using setColourMaterial (GL_AMBIENT_AND_DIFFUSE) - which tells
> the system to use the *polygon* colour for the Ambient and Diffuse comp=
onent
> of the lighting and to use the setMaterial settings only for the remain=
ing
> components (Specular and Emission).
>=20
> Now, one of the oddities of OpenGL (and hence of SSG) is that the Alpha
> produced by this process is ALWAYS the alpha of the DIFFUSE component...
> which in this case (because of the setColourMaterial) is coming from th=
e
> polygon itself.
>=20
> So, you either need to disable GL_COLOR_MATERIAL so that all the lighti=
ng
> components come from the material settings *OR* put the 0.7 alpha on th=
e
> polygon vertex colours instead of the material.

Right, with these modifications and a few internal bug fixes, it seems
to be working well.
Thanks a lot.

By the way, my scene is composed of translucent and opaque objects that
can intersect each other and I have some problems of drawing order with
the transparent objects. Is there a way in PLIB to manage that ? Can we
enable the GL_DEPTH_TEST for the opaque object and disable it for the
tranparent ones ?


Marc.

--=20
___________Escher Marc____________________Alternet Fabric, sarl_______

rue de Langallerie, 6            Tel    : ++41-21-351.10.20   =20
CH-1003 Lausanne                 Fax    : ++41-21-351.10.20
                                 mailto : ma...@fa...  =20
http://www.fabric.ch             smsto  : http://www.flashsms.com/marc
______________________________________________________________________
--- Pingouins et manchots dans les champs, hiver tr=E8s m=E9chant. ---

Re: [Plib-users] transparent material with texture ?

[Steve B. <sjb...@ai...>]

Marc Escher wrote:
 
> I'm trying to render an object with a bmp texture (no alpha) and a
> tranparent material associated to it, but I can't get the material
> properties to be mixed with the texture. The object is displayed with
> the texture but with no transparency or any color for the material. What
> is the right way to do it ?

You are using setColourMaterial (GL_AMBIENT_AND_DIFFUSE) - which tells
the system to use the *polygon* colour for the Ambient and Diffuse component
of the lighting and to use the setMaterial settings only for the remaining
components (Specular and Emission).

Now, one of the oddities of OpenGL (and hence of SSG) is that the Alpha
produced by this process is ALWAYS the alpha of the DIFFUSE component...
which in this case (because of the setColourMaterial) is coming from the
polygon itself.

So, you either need to disable GL_COLOR_MATERIAL so that all the lighting
components come from the material settings *OR* put the 0.7 alpha on the
polygon vertex colours instead of the material.

> --- Pingouins dans les champs, hiver méchant. ---

:-)

Strictly, the French word 'Pingouin' refers to *northern* arctic seabirds
like the Puffin and the Auk. The correct French work for Penguin is 'Manchot'.

Note this discussion:

    http://linuxfr.org/2001/02/05/2198,0,0,1.php3

-- 
Steve Baker   HomeEmail: <sjb...@ai...>
              WorkEmail: <sj...@li...>
              HomePage : http://web2.airmail.net/sjbaker1
              Projects : http://plib.sourceforge.net
                         http://tuxaqfh.sourceforge.net
                         http://tuxkart.sourceforge.net
                         http://prettypoly.sourceforge.net
                         http://freeglut.sourceforge.net

RE: [Plib-users] Particle Systems

[Claude E. <pin...@we...>]

Thanx a lot!!!

[Plib-users] transparent material with texture ?

[Marc E. <ma...@fa...>]

I'm trying to render an object with a bmp texture (no alpha) and a
tranparent material associated to it, but I can't get the material
properties to be mixed with the texture. The object is displayed with
the texture but with no transparency or any color for the material. What
is the right way to do it ?

The state setting are:

topo1 -> setTexture ( "sat.bmp",0,0 );
 topo1 -> enable (GL_TEXTURE_2D);
 topo1 -> disable (GL_LIGHTING);        //the texture gives the shade
 topo1 -> enable (GL_COLOR_MATERIAL);
 topo1 -> disable (GL_CULL_FACE);
 topo1 -> setColourMaterial (GL_AMBIENT_AND_DIFFUSE);
 topo1 -> setMaterial (GL_SPECULAR, 0.1, 0.1, 0.5, 0.9);
 topo1 -> setMaterial (GL_AMBIENT, 0.5, 0.3, 0.3, 0.7);
 topo1 -> setMaterial (GL_DIFFUSE, 0.5, 0.3, 0.3, 0.7);
 topo1 -> setShininess (0.5);
 topo1 -> enable (GL_BLEND);
 topo1 -> enable (GL_ALPHA_TEST);
 topo1 -> setAlphaClamp (0.001);  /* 0..1 */
 topo1 -> setTranslucent();

with a blending function:
 glBlendFunc (GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);

Have I done something wrong ?


Marc.

--
___________Escher Marc____________________Alternet Fabric, sarl_______

rue de Langallerie, 6            Tel    : ++41-21-351.10.20
CH-1003 Lausanne                 Fax    : ++41-21-351.10.20
                                 mailto : ma...@fa...
http://www.fabric.ch             smsto  : http://www.flashsms.com/marc
______________________________________________________________________
--- Pingouins dans les champs, hiver m=E9chant. ---

Re: [Plib-users] Particle Systems

[Steve B. <sjb...@ai...>]

Norman Vine wrote:
> 
> Steve Baker writes:
> >
> >Go to http://toobular.sourceforge.net - then right-click on the
> >"Totally Toobular: Version 0.0.0." link at the bottom of the
> >page...and...
> >Dammit!  It's "gone away" again!
> >
> >I'm on the case - I'll reply again when it's fixed (again!?!).
> 
> Steve as of May 1 the directory structure of SourceForge changed
> 
> It is now
> 
> < external name >
> ftp://plib.sourceforge.net/pub/plib
> 
> < internal name >
> /home/groups/p/pl/plib/public_ftp

Yep - I knew that - but none of my projects use the FTP area - I
just create a directory 'dist' (distribution) under the usual
HTTP area:

  /home/groups/p/pl/plib/htdocs/dist

...since the web page uses relative links, everything *should*
still work.

No - I don't know what happened to Toobular - still, it's all OK
now.

-- 
Steve Baker   HomeEmail: <sjb...@ai...>
              WorkEmail: <sj...@li...>
              HomePage : http://web2.airmail.net/sjbaker1
              Projects : http://plib.sourceforge.net
                         http://tuxaqfh.sourceforge.net
                         http://tuxkart.sourceforge.net
                         http://prettypoly.sourceforge.net
                         http://freeglut.sourceforge.net

RE: [Plib-users] Particle Systems

[Norman V. <nh...@ca...>]

Steve Baker writes:
>
>Go to http://toobular.sourceforge.net - then right-click on the
>"Totally Toobular: Version 0.0.0." link at the bottom of the 
>page...and...
>Dammit!  It's "gone away" again!
>
>I'm on the case - I'll reply again when it's fixed (again!?!).

Steve as of May 1 the directory structure of SourceForge changed

It is now

< external name >
ftp://plib.sourceforge.net/pub/plib

< internal name >
/home/groups/p/pl/plib/public_ftp

there should be a login banner on your shell account to this effect.

Cheers

Norman

Re: [Plib-users] Particle Systems

[Steve B. <sjb...@ai...>]

Claude Eckel wrote:
> 
> Where are the Toobular sources? I read that this game uses a simole
> particle system with PLIB so I wanted to downlaod it. But there is no
> tar.gz - File on the page / ftp-space.

All better now!  Just download it from the link on the bottom of the web
site.  Sorry!

-- 
Steve Baker   HomeEmail: <sjb...@ai...>
              WorkEmail: <sj...@li...>
              HomePage : http://web2.airmail.net/sjbaker1
              Projects : http://plib.sourceforge.net
                         http://tuxaqfh.sourceforge.net
                         http://tuxkart.sourceforge.net
                         http://prettypoly.sourceforge.net
                         http://freeglut.sourceforge.net

Re: [Plib-users] Particle Systems

[Steve B. <sjb...@ai...>]

Claude Eckel wrote:
> 
> Where are the Toobular sources? I read that this game uses a simole
> particle system with PLIB so I wanted to downlaod it. But there is no
> tar.gz - File on the page / ftp-space.

Go to http://toobular.sourceforge.net - then right-click on the
"Totally Toobular: Version 0.0.0." link at the bottom of the page...and...
Dammit!  It's "gone away" again!

I'm on the case - I'll reply again when it's fixed (again!?!).

-- 
Steve Baker   HomeEmail: <sjb...@ai...>
              WorkEmail: <sj...@li...>
              HomePage : http://web2.airmail.net/sjbaker1
              Projects : http://plib.sourceforge.net
                         http://tuxaqfh.sourceforge.net
                         http://tuxkart.sourceforge.net
                         http://prettypoly.sourceforge.net
                         http://freeglut.sourceforge.net

[Plib-users] Particle Systems

[Claude E. <pin...@we...>]

Where are the Toobular sources? I read that this game uses a simole
particle system with PLIB so I wanted to downlaod it. But there is no
tar.gz - File on the page / ftp-space.

[Plib-users] Re: .MDL problems

[Wolfram K. <Wol...@3D...>]

Argh! I wanted to send that to the PPE users mailing list. Sorry.

Bye bye,
Wolfram.

[Plib-users] .MDL problems

[Wolfram K. <w_...@rz...>]

This is an answer to an E-Mail several of us got.

=46irst of all Dave, you are right that you shouldn't
send mails with attachments to the list.
The normal thing to do as long as one doesn't know the developers=20
is IMHO to ask on the list and send the asttachment to the people=20
that request it. This week (starting tomorrow) I will be out of=20
town 3 days, but apart from that you can send questions with=20
attachments to me.=20

>I downloaded the Windows binary of PPE (Version 0.0.1=20
>pre-pre-Alpha from ppewinbin.zip) since I wanted to do a=20
>few quick touchups to a Microsoft Flight Simulator MDL file. =20
>I was unable to do the touchups.

Your main problem is that PPE can't save as .MDL, it can=20
only load .MDL. You have to save as, for example .AC or .OBJ
or .FLT and would then need a converter.=20
Thomas Sevaldrud (sp?) has one in the works, see=20
http://home5.swipnet.se/~w-50954/hv/mdl/dxf2mdl.htm
You may E-Mail him to ask.
I can also try to dig out the old forum post, but this=20
will take some time.

Also, there was at least one person on the aircraft=20
development forum at www.flightsim.com, but I don't think=20
there is a mature one.

Now to your questions:

>1. I could not save (see 'no save.bmp' and ppe_output.txt)

Saving as .MDL doesn't work. You can see all supported formats here:

http://plib.sourceforge.net/ssg/non_class.html

>2. Parts of the plane get cutoff while rotating it until I=20
>select 'Make Everything Visible' (see 'before cutoff.bmp'=20
>and 'after cutoff.bmp')

Yes, thats normal. That's what 'Make Everything Visible' is for.
But sometimes you want to either view only a small part (for example=20
front wheel) or want to look into things (for example fuselage).

>3. Is there a keyboard shortcut to switch between 'Move Camera'=20
>and 'Move' mouse modes?  It would be nice for all options to=20
>have keyboard shortcuts.

I think 'c'. I have a three button mouse and therefore almost
never go into camera mode. With the middle mouse button,
you can rotate the camera around the object.

>4. The cursor position indicator does not change.

Strange. Probably you are not in the correct mode.

>5. I could not align the axis between two vertices=20
>(see 'no axis alignment.bmp')

I don't really understand this. But we know there are still=20
lots of modelling functions missing.

>I get approximately 9 fps without texture and 3 fps=20
>with texture while zoomed into the plane in the bitmaps. =20
>Does this seem slow for my computer?

Which model do you edit? How large is the .MDL / how many=20
trias does it have (open structure window, click on highest=20
level node, look into console).

>How do I know PPE is using the right OpenGL driver=20
>(nvopengl.dll not Glu32.dll)? =20

I had this in PPE once, but I think only locally. There=20
is a preogram to test it, its part of the flightgear=20
binaries, maybe someone will send it to you or has a link.

>How do I maximize performance?

With the defualt_binsdings.py from my home page

http://wolfram.kuss.bei.t-online.de

simply press F8.

>7. Is rendering using wireframe with hidden=20
>lines possible?

Yes. There is a widget in the main window.
Sorry, I am out of time!

>8. Should I be using (or compiling) a different version=20
>of PPE that is more stable and/or a different OS (I use=20
>RedHat Linux 7.0 too)?

No.

>9. What other freeware/shareware graphical editors could=20
>I use for Microsoft Flight Simulator MDL files?

None.=20
Commercial, there is FSDS, but I don't know whether it reads .MDL

Bye bye,
Wolfram.

Re: [Plib-users] Geometry library?

[Steve B. <sjb...@ai...>]

Marlin Mixon wrote:

> >You should be able to form a parametric equation for each line that takes
> >the value zero at one end of the line and 1.0 at the other...finding the
> >intersection in parametric space of infinite lines is then quite easy and
> >you can just test the value of the parameter for each line.
> 
> That sounds interesting, but I'm not familiar with the parametric techniques
> you are describing.  Do you mean like for the general
> form of the line AX + BY + C = 0, the parameters are A,B & C?  Or is it
> something completely different?

You can express a line as:

  x = x1 + p * dx
  y = y1 + p * dy
 
Where (x,y) is any point on the line, (x1,y1) is a point on the line and
(dx,dy) is the direction vector of the line.

If (dx,dy) is a unit vector then 'p' is just the distance of (x,y) from (x1,y1)
which is quite useful - but you don't *have* to make (dx,dy) be a unit vector,
you'd just wind up with different values of 'p' - right?

So, if you have a line *segment*  (x1,y1)...(x2,y2) and you say:

  dx = x2 - x1 ;
  dy = y2 - y1 ;

...then the parameter 'p' is no longer a 'distance' in normally understood
terms.  But (usefully), it is zero at (x1,y1) and 1.0 at (x2,y2) - it's negative
'before' (x1,y1) and >1 for points 'after' (x2,y2).  That's a useful thing to
know when you are doing intersection tests!

So, for two different lines A and B, you have:

  x = x1A + pA * dxA
  y = y1A + pA * dyA
  x = x1B + pB * dxB
  y = y1B + pB * dyB
 
...four equations with four unknowns: (x,y) the point at which they
intersect, pA and pB are the parameteric positions along each line at
which they intersect.

Solving those four equations (which is left as an exercise for the
reader - as all the best math books say), gets you pA and pB.

Now, you can just test pA and pB and if they are both in the range
0..1 then the line segments intersect - if either lies outside
that range - then they don't.  You don't even need to compute
(x,y) if you don't need to - but it drops out of the equations
very easily of you do.

> One feature about the sweep method that I found is it's more than just
> intersection, it also helps you considerably narrow down the lines to
> compare: normally you want to compare each line with every other line but
> the sweep method has you comparing neighbors based on a vertical sweep line,
> so the actual intersection determination is just one part of a solution.

Yes - there are other cute things you can do if you know more about the
problem than just "please intersect these two lines".

>  + <-   sorry, purring cat walked on the keyboard.

Hmmm - typical dumb cat  '+<-' ...plus is less than minus?? Geez! Get
a dog!

> Also, I will gladly contribute any code that I write that is within the
> scope of PLIB.

Excellent!

-- 
Steve Baker   HomeEmail: <sjb...@ai...>
              WorkEmail: <sj...@li...>
              HomePage : http://web2.airmail.net/sjbaker1
              Projects : http://plib.sourceforge.net
                         http://tuxaqfh.sourceforge.net
                         http://tuxkart.sourceforge.net
                         http://prettypoly.sourceforge.net
                         http://freeglut.sourceforge.net

Re: [Plib-users] Geometry library?

[Marlin M. <mar...@ho...>]

>From: Steve Baker <sjb...@ai...>
>Marlin Mixon wrote:
> > I needed to compute 2-D line segment intersections.  I looked through 
>plib
> > and saw some 3D intersection routines, but no 2D.
>
[SNIP]
>
>   3) Contribute the resulting function to the library (please!).
>
>Things are a little simpler for 2D lines than for 3D because you *know*
>that two infinite 2D lines will definitely intersect (unless they are
>parallel).
>
>You should be able to form a parametric equation for each line that takes
>the value zero at one end of the line and 1.0 at the other...finding the
>intersection in parametric space of infinite lines is then quite easy and
>you can just test the value of the parameter for each line.

That sounds interesting, but I'm not familiar with the parametric techniques 
you are describing.  Do you mean like for the general
form of the line AX + BY + C = 0, the parameters are A,B & C?  Or is it 
something completely different?

One feature about the sweep method that I found is it's more than just 
intersection, it also helps you considerably narrow down the lines to 
compare: normally you want to compare each line with every other line but 
the sweep method has you comparing neighbors based on a vertical sweep line, 
so the actual intersection determination is just one part of a solution.+ <- 
sorry, purring cat walked on the keyboard.

Also, I will gladly contribute any code that I write that is within the 
scope of PLIB.

Marlin
_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com

Re: [Plib-users] Geometry library?

[Steve B. <sjb...@ai...>]

Marlin Mixon wrote:
> 
> I needed to compute 2-D line segment intersections.  I looked through plib
> and saw some 3D intersection routines, but no 2D. 

Well, it *is* a 3D library!...If you see a 3D routine that does what you want,
can't you just use the 3D version and just set all the Z coordinates to zero?

But the routine that I recollect in PLIB/SG does *infinite* line versus
*infinite* line...so you need to do a little more:

  Step 1: Intersect two *infinite* lines to find the intersection point.

  Step 2: Does that point lie between the end-points of both lines?

So, if you have the endpoints of your two line segments:

  1) Use sgIsectInfLineInfLine (or a make a 2D version of it) to compute the
     intersection point.

  2) For 2D lines, it's enough to ask whether the X coordinate of that
     point lies between the X coordinates of the ends of both lines...but
     for numerical precision reasons with near-vertical lines, you probably
     ought to check that the Y coordinate also lies between the Y coordinate
     of the end-points.

  3) Contribute the resulting function to the library (please!).

Things are a little simpler for 2D lines than for 3D because you *know*
that two infinite 2D lines will definitely intersect (unless they are
parallel).

You should be able to form a parametric equation for each line that takes
the value zero at one end of the line and 1.0 at the other...finding the
intersection in parametric space of infinite lines is then quite easy and
you can just test the value of the parameter for each line.

-- 
Steve Baker   HomeEmail: <sjb...@ai...>
              WorkEmail: <sj...@li...>
              HomePage : http://web2.airmail.net/sjbaker1
              Projects : http://plib.sourceforge.net
                         http://tuxaqfh.sourceforge.net
                         http://tuxkart.sourceforge.net
                         http://prettypoly.sourceforge.net
                         http://freeglut.sourceforge.net