On Aug 22, 3:15=A0pm, uuilly <willy.li...@[EMAIL PROTECTED]
> wrote:
> On Aug 21, 6:15=A0pm, uuilly <willy.li...@[EMAIL PROTECTED]
> wrote:
>
>
>
> > On Aug 21, 5:45=A0pm, uuilly <willy.li...@[EMAIL PROTECTED]
> wrote:
>
> > > On Aug 21, 5:35=A0pm, uuilly <willy.li...@[EMAIL PROTECTED]
> wrote:
>
> > > > On Aug 21, 4:58=A0pm, "jbwest" <jbw...@[EMAIL PROTECTED]
> wrote:
>
> > > > > "uuilly" <willy.li...@[EMAIL PROTECTED]
> wrote in message
>
> > > >
>news:b204512e-7fb7-4fdc-9d46-369db9277154@[EMAIL PROTECTED]
> > > > > On Aug 21, 9:11 am, "jbwest" <jbw...@[EMAIL PROTECTED]
> wrote:
>
> > > > > > <willy.p...@[EMAIL PROTECTED]
> wrote in message
>
> > > > >
>news:c45bd917-ce46-4f04-82e7-4cf6508e6e15@[EMAIL PROTECTED]
> > > > > > On Aug 20, 5:40 pm, "jbwest" <jbw...@[EMAIL PROTECTED]
> wrote:
>
> > > > > > [...]
>
> > > > > > > Gosh, it's been a while... Is it a closure problem? Do you
ne=
ed to add a
> > > > > > > last point =3D 1st in the input data to close up the polygon
=
?
>
> > > > > > > jbw
> > > > > > >I've tried w/ and w/o a last=3D=3Dfirst. It's funny, it fixes
=
some of the
> > > > > > >counties and breaks just as many others. Any other thoughts?
M=
y next
> > > > > > > move is to write my own tessellation algorithm.
>
> > > > > > That's definitely a closure, off-by-one or last-triangle logic
=
problem.
> > > > > > I've
> > > > > > never had problems with the much-maligned tesselator with very
=
complex
> > > > > > geometries. Surely you can just trace what's being emitted and
=
see why the
> > > > > > last triangle isn't being created (or has the wrong winding
?).
>
> > > > > > jbw
> > > > > >I think I get closure, first =3D=3D last, but what is an: "Off
b=
y one" or
> > > > > >"last triangle logic problem"? =A0Thanks for the advice btw.
>
> > > > > Your sample data doesn't even invoke Combine on my Linux glu
Tess=
elator
> > > > > sample app. Just a combination of tri-strips and tri-fans,
result=
ing in a
> > > > > rectangle. One or the other of those types of BEGIN appears to
be
> > > > > implemented wrong in your app. Otherwise, does this note apply
to=
you ?
>
> > > > > "Notes
> > > > > It is a common error to use a local variable for location or
data=
and store
> > > > > values into it as part of a loop. For example: for (i =3D 0; i <
=
NVERTICES;
> > > > > ++i) {
> > > > > =A0 GLdouble data[3];
> > > > > =A0 data[0] =3D vertex[i][0];
> > > > > =A0 data[1] =3D vertex[i][1];
> > > > > =A0 data[2] =3D vertex[i][2];
> > > > > =A0 gluTessVertex(tobj, data, data);
> > > > > =A0 }
>
> > > > > This doesn't work. Because the pointers specified by location
and=
data might
> > > > > not be dereferenced until gluTessEndPolygon is executed, all the
=
vertex
> > > > > coordinates but the very last set could be overwritten before
tes=
sellation
> > > > > begins.
>
> > > > > Two common symptoms of this problem are consists of a single
poin=
t (when a
> > > > > local variable is used for data) and a
GLU_TESS_NEED_COMBINE_CALL=
BACK error
> > > > > (when a local variable is used for location."
>
> > > > > jbw
>
> > > > Wow. =A0Can you paste all your gluCallbacks here?
>
> > > I am not even using callbacks for begin and end. =A0I'm just passing
=
the
> > > raw gl routines. =A0Combine callback and tessError are my only two
> > > custom callbacks.
>
> > > =A0 =A0 =A0 =A0 gluTessCallback( _tessPtr, GLU_TESS_BEGIN, ( GLvoid(
=
* )
> > > ( ... ) )glBegin );
> > > =A0 =A0 =A0 =A0 gluTessCallback( _tessPtr, GLU_TESS_END, ( GLvoid( *
=
)
> > > ( ... ) )glEnd );
> > > =A0 =A0 =A0 =A0 gluTessCallback( _tessPtr, GLU_TESS_VERTEX, (
GLvoid(=
* )
> > > ( ... ) )glVertex3fv );
> > > =A0 =A0 =A0 =A0 gluTessCallback( _tessPtr, GLU_TESS_COMBINE, (
GLvoid=
( * )
> > > ( ... ) )combineCallback );
> > > =A0 =A0 =A0 =A0 gluTessCallback( _tessPtr, GLU_TESS_ERROR, ( GLvoid(
=
* )
> > > ( ... ) )tessError );
>
> > > same code for linux ****t:
> > > =A0 =A0 =A0 =A0 gluTessCallback( _tessPtr, GLU_TESS_BEGIN, ( void( *
=
)() )glBegin );
> > > =A0 =A0 =A0 =A0 gluTessCallback( _tessPtr, GLU_TESS_END, ( void( *
)(=
) )glEnd );
> > > =A0 =A0 =A0 =A0 gluTessCallback( _tessPtr, GLU_TESS_VERTEX, ( void(
*=
)
> > > () )glVertex3fv );
> > > =A0 =A0 =A0 =A0 gluTessCallback( _tessPtr, GLU_TESS_COMBINE, ( void(
=
* )
> > > () )combineCallback );
> > > =A0 =A0 =A0 =A0 gluTessCallback( _tessPtr, GLU_TESS_ERROR, ( void( *
=
)
> > > () )tessError );
>
> > And the verts are not getting destroyed before endTess().
>
> Here is the output from my tessellation:
>
> Raw Verts:
> -2.2583 0.801231 0
> -2.25511 0.801228 0
> -2.25496 0.801229 0
> -2.25203 0.801228 0
> -2.24635 0.801228 0
> -2.24635 0.798129 0
> -2.24636 0.796031 0
> -2.24636 0.791673 0
> -2.25056 0.791665 0
> -2.25095 0.791663 0
> -2.2557 0.791649 0
> -2.25728 0.791648 0
> -2.25829 0.791646 0
> -2.25828 0.791834 0
> -2.25829 0.792248 0
> -2.25829 0.799885 0
> -2.2583 0.80077 0
> -2.2583 0.801231 0
>
> Output from Callbacks:
>
> Combine Callback Vertex Data
> -2.25828 0.791834 0
> -2.25829 0.791646 0
> -2.2557 0.791649 0
> -2.25728 0.791648 0
>
> Combine Callback Vertex Data
> -2.25496 0.801229 0
> -2.25203 0.801228 0
> -2.24635 0.801228 0
> -2.24635 0.798129 0
>
> begin: "GL_TRIANGLE_FAN"
> vert -2.25496 0.801229 0
> vert -2.24635 0.798129 0
> vert -2.25829 0.799885 0
> vert -2.2583 0.80077 0
> vert -2.25511 0.801228 0
> vert -2.25496 0.801229 0
> end
>
> begin: "GL_TRIANGLE_FAN"
> vert -2.24636 0.796031 0
> vert -2.24636 0.791673 0
> vert -2.25828 0.791834 0
> vert -2.25829 0.792248 0
> vert -2.25829 0.799885 0
> vert -2.24635 0.798129 0
> end
>
> begin: "GL_TRIANGLE_FAN"
> vert -2.25828 0.791834 0
> vert -2.24636 0.791673 0
> vert -2.25056 0.791665 0
> vert -2.25095 0.791663 0
> vert -2.2557 0.791649 0
> vert -2.25828 0.791834 0
> end
>
> begin: "GL_TRIANGLES"
> vert -2.25511 0.801228 0
> vert -2.2583 0.80077 0
> vert -2.2583 0.801231 0
> vert -2.24635 0.801228 0
> vert -2.25203 0.801228 0
> vert -2.25496 0.801229 0
> vert -2.25828 0.791834 0
> vert -2.25829 0.791646 0
> vert -2.25728 0.791648 0
> end
>
> Any thoughts?
A little more info
With this signature:
void combineCallback(GLfloat coords[ 3 ],
GLfloat *vertex_data[ 4 ],
GLfloat weight[4], GLfloat **dataOut )
The variable "coords," that is supposed to contain the location of the
newly created point, is printing total junk. Ie: numbers that are way
too big and seemingly out of left field. So somehow I'm giving GLU
something that is causing it to puke.
Is it a problem that I'm using GLfloats instead of GLdoubles? I was
always under the impression that GPU's don't have double precision so
they're no point in using them.
Thanks,
Willy
|
