Re: [gts-general] Cutting a model up...

[Paul <pau...@st...>]

I can't store the pointers properly and I get errors... here is my code

8<------------------------------------------------------------

static void cut_model( GtsTriangle *t, vector<GtsTriangle*> tv )
{
	GtsPoint * p1, * p2, * p3;
 
	gts_triangle_vertices (t,
				(GtsVertex **) &p1,
				(GtsVertex **) &p2,
				(GtsVertex **) &p3);
	if (p1->y < 0 || p2->y < 0 || p3->y < 0) {
		tv.push_back(t);
				
		cout << tv.size() << " " << p1->y << " " << p2->y <<
" " << p3->y << endl;		
	}
}

void SimpleWindow::cut()
{
	vector<GtsTriangle*> tri_vec;
	tri_vec.push_back(NULL);
	
	gts_surface_foreach_face(s, (GtsFunc) cut_model, &tri_vec);

	for (int i = 1; i < (int)tri_vec.size(); i++) {
		gts_object_destroy(GTS_OBJECT(tri_vec[i]));
	}
	tri_vec.clear();
}

------------------------------------------------------------>8

Is there any problem in sending a STL vector?  It causes a memory
leak.  Also in case you are wondering why I put in the NULL at the start,
I did it because the data seems to get corrupted on the first element.

Does anyone know what is happening?  Or have any suggestion as to what I
should store the values in?

Thanks for all the help so far and in advance for this problem,

Paul

PS A message to Stephane:

The CVS does not contain gts_surface_foreach_face_remove() in surface.c
Also, do you know when the next will be (roughly)?


On Mon, 20 Aug 2001, Stephane Popinet wrote:

> Hi Paul,
> 
>  > I didn't know whether I should be receiving a GtsTriangle or a GtsFace, so
>  > I tried both but the program fails on the assert - keep_faces ==
>  > FALSE.  This is set in the GTS function gts_surface_foreach_face.  
> 
> Ok, what you do is allright in principle but the function
> gts_surface_foreach_face() does not allow you to destroy faces of the
> surface you are currently traversing (it confuses the hash table used
> to store the faces of a given surface, hence the test on keep_faces in 
> gts_surface_foreach_face).
> 
> In the next release (due soon), there will be a
> gts_surface_foreach_face_remove() function which will allow you to do that.
> 
> Go and have a look in the CVS repository in the surface.c file you
> will find this new function.
> 
>  > I have
>  > also tried passing an array to store all the triangles in that I want to
>  > delete and then deleteing them afterwards but this doesn't seem to work
>  > either.
> 
> Hmm, this is more surprising, this approach should work, you probably
> want to check that there isn't another problem in your code.
> 
>  > So any pointers as to what I should be doing would be very helpful, but
>  > here are a few questions...
>  > 
>  > (1) can I just delete a GtsTriangle, and
>  > (2) does it update the information associated with the surrounding
>  > triangles?
> 
> Yes, it does.
> 
>  > (3) can I just delete a GtsFace, and
>  > (4) does it update the information associated with the surrounding
>  > triangles?
> 
> Yes, in the same way.
> 
>  > (5) which should I use - and what is the difference?
> 
> There is no difference. A GtsFace is also (via inheritance) a
> GtsTriangle, if you call the gts_object_destroy() function on these
> objects, there destroy() methods will be called (recursively if
> necessary), if this is a GtsFace, its destroy() method will be called
> and update the parent GtsSurface, then the destroy() method for
> GtsTriangle will be called (because a GtsFace is also a GtsTriangle)
> which will update the GtsEdge used by this GtsTriangle, and so on
> until the destroy() method for a GtsObject is called which will free
> the memory associated with this object.
> 
>  > (6) is there something better to use than gts_surface_foreach_face to get
>  > the information I need.
> 
> See my point above.
> 
> Hope this helps,
> 
> Stephane
> 
> 
> _______________________________________________
> Gts-general mailing list
> Gts...@li...
> http://lists.sourceforge.net/lists/listinfo/gts-general
>