Geo_pointconnector Class.. Any Examples?

[Daniel]

I'm trying to determine if 2 points are part of the same connected mesh, looks like the GEO_PointConnector class is what I want.. just can't figure out how to use it. Anyone know how to use this mysterious class? Or another away to achieve the same thing? This class has a method:

bool 	arePointsConnected (GEO_Point &pt0, GEO_Point &pt1)

Sounds Perfect!.. if I can just get it to work Looks like you need to prime it with a point and it then builds some sort of point tree from that before you can test connectivity, but I just can't get anything working.

Thanks for any help. I've mailed support at sidefx as well, so I'll post the answer if they mail me first.. though late on a Friday, I'm guessing I'll hear on Monday.

Daniel

[sibarrick]

Looks to me like you just prime it with a gdp, and then feed it 2 points to test if they are connected

GEO_PointConnector connections(gdp);

bool areConnected = connections.arePointsConnected(pt0, pt1);

is this what you already tried?

[Daniel]

Looks to me like you just prime it with a gdp, and then feed it 2 points to test if they are connected

GEO_PointConnector connections(gdp);

bool areConnected = connections.arePointsConnected(pt0, pt1);

is this what you already tried?

Hey Simon, yes that's the first thing I tried. It appeared to tell me that the pts were connected even when I knew they weren't. I'm dyamically breaking meshes apart in pops (a softbody pop where edges can be cut based on strain). So during the sim points that are connected on one frame may not be on another. Maybe since I'm doing it in pops and all points are part of the same particle prim it always returns true. Sort of doubt it, I'm guessing I'm not using it correctly there are those "build" methods that make me think more needs to be done to get it working

Daniel

[sibarrick]

You may well be right about the single particle prim, I really can't see it working any other way than you tried. Anyway I guess Sesi will know.

[Daniel]

You may well be right about the single particle prim, I really can't see it working any other way than you tried. Anyway I guess Sesi will know.

Hey Simon,

Yes, turns out the particle prim was an issue. I got very good support info back from Silvina at Sesi and he pointed me to another class as well as how to exclude the particle prim in the test, worked like a charm.

Here's my func (the edgeCache is my own class for storing edge constraints):

void
POP_SoftBody::deleteBadAngleConstraints( float time, POP_ContextData* data)
{
	edgeCache *e;
	GEO_Primitive *partprim;
	GEO_PointClassifier fclass;
	GB_PrimitiveGroup *partgrp = data->getDetail()->newPrimitiveGroup("__particle__", 1);

	FOR_MASK_PRIMITIVES(data->getDetail(), partprim, GEOPRIMPART)
	{
		partgrp->add(partprim);
	}

	fclass.classifyPoints(*data->getDetail(), partgrp);

	for(int i = 0; i < allEdges.size(); i++)
	{
		e=allEdges[i];
		if(e->active && e->type == ANGLE)
			if (fclass.getClass(e->p0->getNum()) != fclass.getClass(e->p1->getNum()))
				e->active = 0;
	}

	data->getDetail()->destroyPrimitiveGroup(partgrp);
}