You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am trying to integrate FCL into an existing scenegraph-system, that internally manages multiple levels of object-grouping. In order to use the broadphase algorithms of FCL I fill instances of fcl::DynamicAABBTreeCollisionManager for each grouping.
During the creation of a higher-level grouping I then attempt to add the list of collision objects in the subgroup that is currently empty.
The manager starts a tree traversal for insertion, and the final insertion method HierarchyTree::topdown_0(in file include/fcl/broadphase/hierarchy_tree.hxx, line 353) then tries to touch the first element of the empty list(line 386), causing a segfault.
The method topdown_0 even checks the number of elements before doing the main processing, so I assume it could be a reasonable place to catch the special case of an empty list. However, I can't really assess the affects of setting the root pointer of the tree to NULL (as would be my natural answer to no insertion).
I suspect that similar cases are also currently unhandled by the library, making the usage quite unrelieable.
If you could set a guideline on when special cases should be caught (As early as possible? Or process them until they have to be handled?), I'd be up to fixing issues I find myself and submitting pull requests.
The text was updated successfully, but these errors were encountered:
I am trying to integrate FCL into an existing scenegraph-system, that internally manages multiple levels of object-grouping. In order to use the broadphase algorithms of FCL I fill instances of fcl::DynamicAABBTreeCollisionManager for each grouping.
During the creation of a higher-level grouping I then attempt to add the list of collision objects in the subgroup that is currently empty.
The manager starts a tree traversal for insertion, and the final insertion method HierarchyTree::topdown_0(in file include/fcl/broadphase/hierarchy_tree.hxx, line 353) then tries to touch the first element of the empty list(line 386), causing a segfault.
The method topdown_0 even checks the number of elements before doing the main processing, so I assume it could be a reasonable place to catch the special case of an empty list. However, I can't really assess the affects of setting the root pointer of the tree to NULL (as would be my natural answer to no insertion).
I suspect that similar cases are also currently unhandled by the library, making the usage quite unrelieable.
If you could set a guideline on when special cases should be caught (As early as possible? Or process them until they have to be handled?), I'd be up to fixing issues I find myself and submitting pull requests.
The text was updated successfully, but these errors were encountered: