Revision: 64372
http://sourceforge.net/p/brlcad/code/64372
Author: starseeker
Date: 2015-03-12 00:54:45 +0000 (Thu, 12 Mar 2015)
Log Message:
-----------
Start outlining a strategy for building boolean trees.
Modified Paths:
--------------
brlcad/trunk/src/libbrep/shape_recognition.cpp
Modified: brlcad/trunk/src/libbrep/shape_recognition.cpp
===================================================================
--- brlcad/trunk/src/libbrep/shape_recognition.cpp 2015-03-12 00:18:03 UTC
(rev 64371)
+++ brlcad/trunk/src/libbrep/shape_recognition.cpp 2015-03-12 00:54:45 UTC
(rev 64372)
@@ -171,9 +171,25 @@
}
}
-// TODO - this approach is insufficient/incorrect. Need to assemble
subtracted subbreps into
-// combs and subtract those as a unit - this approach doesn't correctly
subtract everything
-// from objects that need the subtractions.
+/* This is the critical point at which we take a pile of shapes and actually
reconstruct a
+ * valid boolean tree. The stages are as follows:
+ *
+ * 1. Identify the top level union objects (they will have no fil faces in
their structure).
+ * 2. Using the fil faces, build up a queue of subbreps that are
topologically connected
+ * by a loop to the top level union objects.
+ * 3. Determine the boolean status of the 2nd layer of objects.
+ * 4. For each sub-object, if that object in turn has fil faces, queue up
those objects and goto 3.
+ * 5. Continue iterating until all objects have a boolean operation assigned.
+ * 6. For each unioned object, build a set of all subtraction objects
(topologically connected or not.)
+ * 7. add the union pointer to the subbreps_tree, and then add all all
topologically connected subtractions
+ * and remove them from the local set.
+ * 8. For the remaining (non-topologically linked) subtractions, get a
bounding box and see if it overlaps
+ * with the bounding box of the union object in question. If yes, add the
pointer to subbreps_tree.
+ * If not, no action is needed. Once evaluated, remove the subtraction
pointer from the set.
+ *
+ * Initially the test will be axis aligned bounding boxes, but ideally we
should use oriented bounding boxes
+ * or even tighter tests.
+ */
struct bu_ptbl *
find_top_level_hierarchy(struct bu_ptbl *subbreps)
{
@@ -191,6 +207,18 @@
// shares a face with the comb will have to subtract the whole comb,
unless we do some sort
// of bounding box based subset identification... Probably necessary
to keep the booleans
// reasonable. nist 2 is a good test case for this...
+ //
+ // Will have to do minimal bbox and possibly convex hull testing, may
even grow to SSI.
+ // We need to know if subtraction volumes intrude on union volumes, and
there is no way
+ // to know that short of testing.
+ //
+ // Only have to worry about testing fil faces in unions against
subtracted objects, assuming
+ // original B-Rep is non-self-intersecting. fil surfaces are new outer
faces in new brep
+ // objects, and only in those faces is there a possibility of new
object intersections.
+ //
+ // NEEDED: better organization of negative/positive testing and
assignment
+ // robust convex hull/minimum bounding arb logic -
GeometricTools port
+ // arb/arb intersection testing - collision detection perhaps?
if (obj->fil_cnt == 0) {
std::cout << "Top union found: " << bu_vls_addr(obj->key) << "\n";
}
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
BRL-CAD Source Commits mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/brlcad-commits
