Buggy assert in print() function?

[histfuture]

Hi, I found a possible buggy assert in kernel_print function,

assert((level == 0 && node.hasStatus(ns_LAST_LEVEL)) || (level != 0 && hasStatus(ns_LAST_LEVEL))); // line 164

I think the line should be ,

assert((level == 1 && node.hasStatus(ns_LAST_LEVEL)) || (level != 1 && hasStatus(ns_LAST_LEVEL)));

Since nodes in level 0 are all leaf nodes , and in NTree.hpp,

`
// line 607

    if (level == 1)         
    {
      kernel_clearNodes<InnerNode, true> <<<numBlocks, num_threads_per_block>>>(branching_factor * numNodes, (InnerNode*) nodes );
      HANDLE_CUDA_ERROR(cudaDeviceSynchronize());
      kernel_setNodes<InnerNode, LeafNode, branching_factor> <<<numBlocks, num_threads_per_block>>>(
          D_PTR(d_voxels), num_points, level, D_PTR(nodeCount),
          (InnerNode*) nodes, D_PTR(nodeIds), (LeafNode*) childNodes);
  }

`

the above code set ns_LAST_LEVEL flag for nodes in level 1.

[Squelsh]

Thanks @nangong !!
From a first look you seem to be right. ns_LAST_LEVEL is not the Leaf-Level, but one level above.
It is used to trigger serial execution of Leaf-Node operations to prevent inefficient full parallelization.
So the assert is buggy.
I will check this again and correct it with an upcoming release (which should come soon).

[histfuture]

@Squelsh
Moreover , I found both level 0 and level 1 are set ns_LAST_LEVEL flag