feat: add range option to MeshBVH #660
Conversation
minitoine
force-pushed
the
feat/add-range-option
branch
from
664a351
to
4913b32
Compare
minitoine
force-pushed
the
feat/add-range-option
branch
from
4913b32
to
510e99a
Compare
|
Will the index rearrangement only occur for the given range ? Or do I need to put restriction there as well ? |
There was a problem hiding this comment.
Great, thanks! Can we update the README, as well?
Will the index rearrangement only occur for the given range ? Or do I need to put restriction there as well ?
Yeah the indices are only arranged within the necessary range.
One memory issue currently, though, is that the "indirect" buffer is created for the full length of the geometry buffers even if a small subset is needed. But we can track that in another issue.
Another thing we may want to consider for the future is keeping an "offset" field on the BVH so that it can be retained if the subgeometry in the BatchedMesh is moved (ie when "optimizing" or repacking the BatchedMesh).
| // If given, the MeshBVH will be computed for the given range on the geometry. | ||
| // Default range is { start: 0, count: Infinity } | ||
| range: { start: number, count: number } | ||
|
|
There was a problem hiding this comment.
nit: Can we fix the indentation here
| let workingRange = geo.drawRange; | ||
| if ( range ) { | ||
|
|
||
| workingRange = getRangesIntersection( workingRange, range ); |
There was a problem hiding this comment.
I think it's okay if we just use the passed in "range" here instead of trying to find the intersection of drawRange and the passed in range. It's simpler and in this way a user can override the use of draw range if needed, as well.
That means we'd also have to default the "range" option to null. Or we could default "range" to the geometries draw range to be more clear. What do you think?
|
|
||
| } ); | ||
|
|
||
| it( 'should respect the range option without groups.', () => { |
There was a problem hiding this comment.
I think we'll want to make sure this works with the indirect argument, as well. Do you mind adding another test or two for that case?
Closes #513