parry3d-f64

Added

• Debug for TriMesh now prints a summary instead of just the type name (#177).
• query::contact now supports heightfields, instead of returning Err(Unsupported) (#274).

Fixed

• Fix GJK reporting overlapping axisymmetric shapes as disjoint when the simplex stalls on a degenerate support direction (#396).
• Fix wrong shape-cast normals and distances for stalled GJK simplices, now classified as touching or penetrating from the solver's own distance bound (#193).
• Fix shape-casts missing grazing hits at large coordinates, by scaling the GJK ray-cast tolerance with the support magnitude (#180).
• Fix shape-cast time of impact being short by an amount scaling with the shape's extent (#429).
• Fix ShapeCastStatus::PenetratingOrWithinTargetDist being reported for merely touching support-map casts (#106).
• Fix contact between exactly-touching shapes returning witness points at the shapes' centers (#315).
• Fix Triangle::area returning wildly wrong values for degenerate triangles (#111).
• Fix point projection on degenerate triangles reporting the point as inside (#76).
• Fix segments_intersection2d classifying exact segment endpoints as OnEdge instead of OnVertex (#109).
• Fix composite-shape point projection panicking on non-finite points (#395).
• Fix HeightField::project_local_point iterating on every element, making it O(rows * cols) (rapier#332).
• Fix single-point contact manifolds on cylinder and cone caps, by orienting the cap's polygonal approximation toward the contact (rapier#810).
• Fix missing speculative contacts between a Voxels shape and a shape within the prediction distance (#404).
• Fix Voxels shape-cast witness1 not being expressed in the voxels shape's local frame (#373).
• Fix an out-of-bounds panic in the binned BVH builder on degenerate leaf AABBs (rapier#961).
• Fix out-of-bounds indices in to_outline for round shapes with a zero border radius (rapier#969).
• Fix the quadratic polygon removal in hertel_mehlhorn (#408).
• Require spade 2.15, the minimum version providing try_bulk_load_cdt (#428).

Modified

• Document the frame conventions of the ShapeCastHit fields (rapier#933).
• Document that the Voxels iterators only yield non-empty voxels (#382).
• Document that BoundingVolume::merge/merged don't guarantee strict containment of their inputs (#260).

Fixed

• Fix incorrect closest_points_cuboid_cuboid closest points (affecting query::distance too) in degenerate configurations where the separating axis is ambiguous, e.g., axis-aligned cuboids closest at a vertex or along parallel edges.

Release v0.30.0 (#435)

Breaking changes

• The simd-stable and simd-nightly features were removed. 4-lane SIMD is now always enabled (it falls back to scalar code on targets without SIMD support). The new opt-in simd8 feature widens SIMD to 8 lanes for f32 builds; it requires an AVX-enabled target to actually emit 256-bit instructions, and is incompatible with enhanced-determinism.
• ContactManifold::subshape_pos1/subshape_pos2 are no longer public fields. They are replaced by a single boxed subshape_poses: Option<Box<SubshapePoses>> field (to shrink ContactManifold), accessed through the new subshape_pos1()/subshape_pos2() getters and set_subshape_pos1/set_subshape_pos2 setters.

Added

• query::sweep_toi: sweep-based time-of-impact queries. A timestep is modeled as a Sweep between two endpoint poses (linear translation + rotation nlerp), and sweep_time_of_impact (plus sweep_time_of_impact_composite for composite shapes) computes the earliest time the swept shapes reach a slop-based target separation, using conservative advancement. Also exports ToiProxy, SimplexCache, SweepToiOutput, SweepToiStatus, and SweepCompositeFastShape.
• Bvh gains incremental and parallel update APIs: refit_partial, flag-preserving refit_without_resolve variants, refit_parallel, parallel BVTT traversal, and batched parallel leaf updates (insert_or_update_batch_partially_parallel, reinsert_or_update_with_change_detection, reinsert_or_update_if_present, update_partially_if_present, and the BvhLeafUpdateStatus enum).

Modified

• Cuboid support-face feature ids are now computed with bit operations instead of lookup tables.

Release v0.29.0 (#427)

Breaking changes

• The <Polyline as TypedCompositeShape>::PartNormalConstraints is now SegmentPseudoNormals instead of ().

Fixed

• Fix enhanced-determinism not producing cross-platform deterministic results for parry3d/parry3d-f64. The feature now enables glamx/scalar-math to disable architecture-specific SIMD (NEON on arm64, SSE2 on x86_64) which caused floating-point non-associativity in Vec3/Vec4/Quat dot products.
• Oriented (one-sided) 2D polylines (#254). Build a Polyline with PolylineFlags::ORIENTED via Polyline::with_flags, or toggle it with Polyline::set_flags, to clamp contact normals to the outward side. This removes the spurious sideways push a body gets at a convex corner of a thin double-sided polyline. A single 2D flag covers what TriMesh splits across TriMeshFlags::ORIENTED (assume an outward orientation and compute pseudo-normals) and TriMeshFlags::FIX_INTERNAL_EDGES (use them to clamp contact normals). Also adds Polyline::flags, Polyline::segment_normal_constraints, and the PolylineFlags / SegmentPseudoNormals types. (#425)

• feat: improve publish-script + fix building parry3d-f64 standalone

• chore: add publish dry-run to CI

Breaking changes

• CompositeShapeRef::project_local_point and project_local_point_and_get_feature now return Option<...> instead of unwrapping internally, and take an additional max_dist parameter that bounds the search distance.

Modified

• Bump glamx to 0.3 (built on glam 0.33), simba to 0.10, rstar to 0.13, hashbrown to 0.17, rand to 0.10, and kiss3d (visual examples) to 0.42.
• Many shape, AABB, and voxel internals were reworked to avoid bracket-indexing of glam vectors, which is not supported on SPIR-V targets. A new VectorExt::vget/vset API and a for_each_dim! macro are now used in place of v[i]. parry3d's alloc feature also now gates smallvec, downcast-rs, rstar, and glamx/approx, so the no-alloc build can target SPIR-V.

Fixed

• Fix Bvh::scale/BvhNode::scale producing an inverted AABB on negative scale components, which silently broke queries on shapes scaled with reflections (e.g. TriMesh::scaled). Scale components must now be non-zero (previously: positive) (#406).
• Fix Voxels::scaled not scaling the internal chunk BVH, causing queries on scaled voxels to miss geometry (#419).
• Fix Bvh::optimize_incremental panicking after multiple removes without an intervening refit when the root collapses to a partial root, leaving orphaned wide nodes (#409).
• Correct the assertion message on Aabb/BoundingSphere's loosen/tighten (and their *_ed variants) to "must be non-negative" (#394).
• Fix multiple EPA failures on large coordinate magnitudes by scaling the face-rejection tolerance relative to the simplex vertex magnitudes (#415).
• Fix CCD edge cases (in both ball-vs-ball and support-map-vs-support-map casts) where casts starting at (or very close to) the contact boundary with a near-tangent direction would return an unreliable normal. The fallback now uses the contact query to recover a robust closest-point normal, and the small-TOI threshold was relaxed from 1e-5 to 1e-4.
• Fix Aabb::cast_local_ray_and_get_normal panicking with a subtraction overflow when casting a zero-direction ray starting inside the AABB (#383).
• Fix infinite loop in TriMesh::intersection_with_local_plane on degenerate adjacency graphs that don't loop cleanly back to the starting index (#398).
• Fix panic in mesh_intersection when constraint edges overlap (e.g. for co-planar triangles). Overlapping constraints are now skipped instead of crashing the CDT (#389).
• Fix panic in 3D voxelization's internal convex-hull step by falling back to try_convex_hull and returning an empty hull on failure (#347).
• Fix WSign::copy_sign_to on non-CUDA targets by using the native copysign (the bit-twiddling workaround is now scoped to nvptx64, where cuda_std's copysign does not compile).

• Switched to glamx 0.2 (glam 0.32).

Release v0.26.1 (#417)

Added

• Added Bvh::traverse_indexed to traverse only a part of the subtree, and to retrieve the node’s indices (#416).

Fixed

• Fix Bvh::optimize_incremental being non-deterministic when restored from a serialized snapshot. The incremental optimization state is now stored on the Bvh itself (instead of the non-serializable BvhWorkspace) so it survives serialization/deserialization (#403).

Migrate codebase to glam + release v0.26.0 (#401)

Breaking changes

This release migrates parry from nalgebra to glam (via the glamx crate) for future compatibility with rust-gpu. This is a major breaking change affecting almost all public APIs.

Type renames

• Isometry → Pose (using glamx::Pose2/Pose3)
• Rotation → Rot2/Rot3 (using glamx::Rot2/Rot3)
• IsometryOps → PoseOps
• IsometryOpt → PoseOpt

Removed types

• Point<Real> - use Vector instead. Points and vectors are now unified.
• UnitVector<Real> - use Vector instead. Normalization is no longer encoded in the type.
• Translation<Real> - use Vector for translations.

Math type changes

• Vector is now glam::Vec2/Vec3/DVec2/DVec3 depending on dimension and precision features
• Matrix is now glam::Mat2/Mat3/DMat2/DMat3
• The math module now re-exports glam types and provides dimension-agnostic aliases

API signature changes

• Many functions that previously took &Point<Real> or &Vector<Real> now take Vector by value
• Functions taking &Isometry<Real> now take &Pose or Pose by value
• HalfSpace::new now takes Vector instead of Unit<Vector<Real>>
• Shape constructors like Segment::new, Triangle::new, Capsule::new now take Vector instead of Point
• Aabb::mins and Aabb::maxs are now Vector instead of Point<Real>

Migration guide

If your codebase currently relies on nalgebra, note that nalgebra and glamx provide type conversion. Enable the corresponding features:

• nalgebra = { version = "0.34", features = [ "convert-glam030" ] }
• glamx = { version = "0.1", features = ["nalgebra"] } then you can convert between glam and nalgebra types using .into().

// Before (nalgebra)
use parry3d::na::{Point3, Vector3, Isometry3, Unit};
let point = Point3::new(1.0, 2.0, 3.0);
let vector = Vector3::new(1.0, 0.0, 0.0);
let normal = Unit::new_normalize(vector);
let pos = Isometry3::translation(1.0, 2.0, 3.0);

// After (glam)
use parry3d::math::{Vector, Pose};
let point = Vector::new(1.0, 2.0, 3.0);  // Points are now Vector
let vector = Vector::X;
let normal = vector.normalize();          // No Unit wrapper
let pos = Pose::translation(1.0, 2.0, 3.0);

Common patterns:

• Point3::origin() → Vector::ZERO
• Vector3::x() → Vector::X
• point.coords → just use the vector directly
• Unit::new_normalize(v) → v.normalize()
• Isometry3::identity() → Pose::IDENTITY
• isometry.translation.vector → pose.translation

Modified

• Re-export glamx instead of nalgebra as the public linear algebra dependency
• Migrate visual examples from macroquad to kiss3d
• Renamed rkyv-serialize feature to just rkyv

Release v0.25.3 (#393)

• Significantly improve performances of Voxels::combine_voxel_states.

Release v0.25.2 (#391)

• Add the encase feature to enable implementation of encase::ShaderType for primitive shapes.
• Fix regression in shape-vs-composite-shape contact calculation.

• Fix point-projection of a ball on a voxels shape containing empty chunks (e.g. after voxels removal).
• General documentation improvements.

• The Voxels shape now uses a sparse storage internally.
• Remove Voxels::try_set_voxel. Use ::set_voxel instead.
• The Voxels::domain function returns values instead of references now.
• Removed Voxels::extents and Voxels::domain_center. They can be obtained from Voxels::local_aabb.

• Fixed same-machine determinism of the re-exported hashmap when building without the enhanced-determinism feature.
• Replace all the square-root angular inertia tensors by the same tensor without taking it square root. As a result MassProperties::world_inv_inertia_sqrt has been renamed to world_inv_inertia, MassProperties::inv_principal_inertia_sqrt renamed to inv_principal_inertia.
• Modified the type aliases from the simd module when neither the simd-stable nor simd-nightly are enabled. Without these simd features all type aliases are single-lane non-simd types (f32, f64, bool) instead of 4-lanes auto-simd types.
• Moved the SimdAabb type behind the simd-stable or simd-nightly cargo feature.

• Update to nalgebra 0.34.

Added

• Add Bvh::insert_with_change_detection which runs an insertion unless the inserted aabb is enclosed in the existing one. This differs from insert_or_update_partially because the former will update parent AABBs so they enclose their descendants.

Fixed

• Fix bug in BVH tree node removal.
• Fix bug in BVH tree building from exactly two leaves.

Fixed

• Fix invalid BVH state that could be reached after a removal resulting in a partial root.

Added

• Add a new Bvh acceleration data-structure. It replaces Qbvh entirely. It supports:
  • Traversals (best-first, depth-first, BVTT, leaf iterators, and leaf pairs iterator).
  • It can be constructed either incrementally by inserting nodes, or from a set of leaves using either the binned building strategy or the PLOC (without parallelism) strategy.
  • Dynamic leaf insertion, update, removal.
  • Incremental tree rebalancing.

Fixed

• Fix clip_aabb_line crashing when given incorrect inputs (zero length direction or NAN).
• Fix Segment::intersects_ray returning false-positive when the segment is zero-length. (#31).
• Expose utils::sort3 and utils::sort2.

Modified

• The local_point_cloud_aabb, point_cloud_aabb, and Aabb::from_points now takes an iterator over point values instead of an iterator to point references. Variants taking point references still exist and are named local_point_cloud_aabb_ref, point_cloud_aabb_ref and Aabb::from_points_ref.
• Renamed SimdCompositeShape and TypedSimdCompositeShape to CompositeShape and TypedCompositeShape`.
• The TypedCompositeShape trait now derives from CompositeShape.
• Removed every *Visitor structures. Instead, either call Bvh::traverse (or Bvh::search_best, or Bvh::leaves, or bvh::leaf_pairs), or wrap your composite shape into CompositeShapeRef to access some generic implementation of various geometric queries for any composite shape.
• All composite shapes now rely on the new Bvh acceleration structure instead of Qbvh.
• The Qbvh has been removed. Use Bvh instead.

Added

• Add Voxels::combine_voxel_states to merge the voxel state of two Voxels shapes, updating their internal voxels state as if both shapes were part of a single one. In particular, this will prevent any internal edge that would arise at the boundaries of both shapes if they were adjacent.
• Add Voxels::propagate_voxel_change to propagate a single-voxel modification from one Voxels shape to another, in order to update their internal neighborhood information as if both were part of the same Voxels shape.

Added

• Add TriMesh::cast_ray_with_culling and TriMesh::cast_local_ray_with_culling for casting rays on a triangle mesh but with the possibility to prevents hits on front-faces or back-faces.
• Add contact manifold calculation between two voxels shapes, or a voxels shape and compound shape.
• Add intersection check between voxels and other shapes.
• Add MassProperties::from_voxels to compute the mass and angular inertia tensor from a voxels shape.

Modified

• Add new variants to TypedWorkspaceData for voxels-voxels and voxels-compound collision detection workspace data.
• The Voxels shape now only support cuboids as their leaf geometries (pseudo-balls were removed).

Fixed

• Fix infinite loop in Voxels::set_voxel.

Added

• Rework the Voxels shape API to use better method names.
• Added implementations for linear and non-linear shape-casting involving Voxels shapes.

Added

• Added the Voxels type: a dedicated shape for voxel models. This is currently experimental because some features are still missing (in particular: shape-casting, mass properties, and collision-detection against non-convex shapes).
• Added SharedShape::voxels, SharedShape::voxels_from_points, and ::voxelized_mesh for creating a voxels shape from grid coordinates, points, or automatic voxelization of a triangle mesh.

Added

• Derive Copy for VHACDParameters.
• Add spade default feature for algorithms using Delaunay triangulation from spade.
• Add SharedShape::from_convex_polyline_unmodified and ConvexPolygon::from_convex_polyline_unmodified to initialize a polyline from a set of points assumed to be convex, and without modifying this set even if some points are collinear.
• Add TriMesh::pseudo_normals_if_oriented that returns Some only if the mesh has the TriMeshFlags::ORIENTED flag enabled.

Modified

• The TriMeshFlags::FIX_INTERNAL_EDGES flag no longer automatically enable the TriMeshFlags::ORIENTED flag (but the mesh pseudo-normals will still be computed).
• Improve no_std compatibility.
  • Everything is now compatible, except mesh_intersections, split_trimesh, convex hull validation, and computation of connected components for TriMesh.
  • Add the alloc_instead_of_core, std_instead_of_alloc, and std_instead_of_core Clippy lints to the workspace.
  • Use core and alloc directly rather than using an std alias.
  • Use hashbrown instead of rustc-hash when enhanced-determinism is not enabled.
  • Make spade optional.

Fix

• Fix trimesh inertia tensor computation #331.
• Fix shifted inertia tensor computation #334.

Added

• Implement ::to_trimesh in 2d for Cuboid and Aabb.
• Fix some edge-cases in point_in_poly2d for self-intersecting polygons.
• Fix some edge-cases in mesh/mesh intersection that could result in degenerate triangles being generated.

Fix

• Fix panic in epa3::EPA::closest_points and epa2::EPA::closest_points. Related issues: #253, #246

Modified

• Propagate error information while creating a mesh and using functions making use of it (See #262):
  • TriMesh::new
  • TriMesh::intersection_with_aabb
  • SharedShape::trimesh
  • SharedShape::trimesh_with_flags
• point_cloud_bounding_sphere and point_cloud_bounding_sphere_with_center now returns a BoundingSphere.
• Removed IntersectionCompositeShapeShapeBestFirstVisitor (which had been deprecated for a while): use IntersectionCompositeShapeShapeVisitor instead.

Fix

• Always compute connected-components from union-find instead of topology. It is faster and the function based on topology could result in a crash for non-manifold meshes.

Added

• Add TriMeshConnectedComponents::to_meshes and ::to_mesh_buffers to easily extract individual meshes from the set of connected components.
• Add TriMesh::connected_component_meshes to get the connected components as meshes directly.

Modified

• Connected-components extraction will never fail now, and no longer require the successful calculation of the mesh’s half-edge topology.

Fix

• Fix compiling with enhanced-determinism feature enabled.
  • This is now checked on CI.

Added

• Implement Shape::feature_normal_at_point for TriMesh to retrieve the normal of a face, when passing a FeatureId::Face.
• Add convex_polygons_intersection_points_with_tolerances, convex_polygons_intersection_with_tolerances, and intersect_meshes_with_tolerances that let the user specify tolerances value for the collinearity check.

Fix

• Fix some robustness issues in mesh/mesh intersection when parts of both meshes overlap perfectly.
• Improve robustness of convex polygons intersections when all the vertices of one polygon are located in either the edges or vertices of the other polygon.
• Fix incorrect orientation sometimes given to the polygon output by the convex polygon intersections when one of the polygon is completely inside the other.

Modified

• Improve convergence of epa algorithm in degenerate configurations.
• Fix bug in the mesh/mesh intersection algorithm that didn’t properly take mesh transforms into account.

Added

• Add Triangle::robust_scaled_normal and Triangle::robust_normal as a more robust way to compute the triangles normal for thin triangles that generally cause numerical instabilities.
• Add Triangle::angle_closest_to_90 to find the triangle’s vertex with an angle closest to 90 degree.
• Add the wavefront feature that enables TriMesh::to_obj_file for exporting a mesh as an obj file.
• Add Shape::scale_dyn for scaling a shape as a trait-object.

Modified

• TypedShape::Custom(u32) is now TypedShape::Custom(&dyn Shape).
• AabbSetsInterferencesCollector::tolerence is now spelled correctly as tolerance.
• Real is now exposed through a use statement, so that an indirection is removed in documentation: previous occurrences of Real now show f32 or f64.
• Significantly improved the general stability of mesh/mesh intersection calculation.
• Rename Shape::clone_box to Shape::clone_dyn (the clone_box method still exists but has been deprecated).
• Make try_convex_hull return an error instead of panicking if less than 3 input points are given.
• Make Triangle::normal and Triangle::scaled_normal only available in 3D instead of panicking in 2D.

Fix

• Fix occasional crash in mesh/mesh intersection if some of the vertex coordinates are very small.

Fix

• Fix edge case where some of the principal angular inertia are clamped to zero for decimeter-sized objects.
• Have ball-ball shape casting take into account the stop_on_penetration flags.
• Don’t panic in EPA for a corner case that needs some additional debugging. Show a debug log instead.

Added

• Implement concave polygons intersections: polygons_intersection_points, polygon_intersection.

Modified

• Update bitflags to version ^2.3
• Update nalgebra to 0.33.
• Update indexmap to 2.

Fix

• Fix a regression in ball vs. convex shape contact manifold calculation.

Added

• Add ShapeCastOptions that includes two new options for (linear) shape-casting. ShapeCastOptions::target_distance which will return a hit as soon as the moving shapes are closer than this distance; and compute_impact_geometry_on_penetration which forces the calculation of proper witness points and normals even if the shapes are initially intersecting (time_of_impact == 0.0).

Modified

This version modifies many names related to shape-casting:

• Renamed TOI to ShapeCastHit.
• Renamed TOIStatus to ShapeCastStatus.
• Rename RayIntersection::toi to RayIntersection::time_of_impact.
• More generally, all occurrences of the word toi have been replaced by time_of_impact for better clarity.
• Rename query::time_of_impact to query::cast_shapes. More generally, all the functions prefixed with time_of_impact_ (e.g. time_of_impact_ball_ball) are now prefixed with cast_shapes_ (e.g. cast_shapes_ball_ball).
• Rename QueryDispatcher::time_of_impact to QueryDispatcher::cast_shapes.
• The (linear) shape-casting functions like query::cast_shapes (previously named query::time_of_impact) now take a ShapeCastOptions instead of the max_toi and stop_at_penetration arguments.
• Rename query::nonlinear_time_of_impact to query::cast_shapes_nonlinear.
• Rename QueryDispatcher::nonlinear_time_of_impact to QueryDispatcher::cast_shapes_nonlinear.
• Rename NonlinearTOIMode to NonlinearShapeCastMode, and NonlinearTOIMode::DirectionalTOI to NonlinearShapeCastMode::Directional.
• Rename TimeOfImpactStatus::Penetrating to ShapeCastStatus::PenetratingOrWithinTargetDist.

Modified

• Remove CUDA support to break free from the toolchain restriction required by cust.
• Rework internal edges resolution using normal cones. This implies the modification of the SimdCompositeShape::map_part_at, TypedSimdCompositeShape::map_typed_part, and TypedSimdCompositeShape::map_untyped_part trait functions so that the closure argument takes an extra argument for the (optional) normal constraints. This argument can be safely ignored by user code unless applying the normal collection is relevant to your use-case.
• Contact manifolds will now retain all contacts (including the ones further than the specified prediction distance) whenever any contact is actually closer than this prediction distance.
• Typo fix: renamed TopologyError::BadAdjascentTrianglesOrientation to BadAdjacentTrianglesOrientation.

Fixed

• Fix contacts between convex shapes being occasionally ignored due to some rounding errors.
• Remove crash when entering unreachable code in non-linear TOI calculation.
• Fix accuracy issue in triangle-mesh center-of-mass calculation when the mesh isn’t manifold.

Added

• Add SdpMatrix2::inverse_and_get_determinant_unchecked. This is useful for computing the inverse in a AoSoA SIMD setting.
• Add Aabb::intersects_moving_aabb to perform a swept test between two moving aabbs.

Added

• Add Qbvh::traverse_depth_first_with_context, Qbvh::traverse_depth_first_node_with_stack_and_context, and the related SimdVisitorWithContext trait to allow parent nodes to pass a custom context to its children during recursion.

Modified

• The point_in_poly2d now handles arbitrary (convex and non-convex) polygons. The previous implementation that only supported convex polygons has been renamed point_in_convex_poly2d.

Fixed

• Fix a crash in Qbvh::refit that results from the QBVH tree becoming increasingly imbalanced.

Added

• Add Aabb::scaled_wrt_center to scale an AABB while keeping its center unchanged.

Fixed

• Fix ball-convex manifolds missing contacts in some corner cases.
• Fix panic in TriMesh::intersection_with_plane.

Fixed

• When using rkyv, fix CheckBytes implementation for types archived as themselves.
• Fix occasional crash in the QBVH incremental update.

Fixed

• Fix Polyline::flat_indices that returned an incorrectly sized slice.
• Fix serialization of SimdAabb into map-like formats like JSON, YAML, RON.

Added

• Add validation when using rkyv safe API whenever applicable to parry types.

Modified

• Improved performance of intersection checks involving composite shapes (compound shapes, trimeshes, polylines, etc.)

This version was yanked. See the release notes for 0.13.3 instead.

Fixed

• Add workaround to address jitter issue due to incorrectly empty contact manifolds generated sometimes for convex
  polyhedron.

Modified

• About rkyv support: most POD structs (Aabb, Ball, Cuboid, etc.) are now archived as themselves instead of being archived as different types (for example Aabb is archived as Aabb itself instead of ArchivedAabb).

Added

• In 3D, add transformation::try_convex_hull for a convex hull calculation that will return an error instead of panicking on unsupported inputs.

Fixed

• Fixed duplicate faces in the connected components returned by TriMesh::connected_components.

Added

• Add TriMesh::canonical_intersection_with_plane for intersecting with planes aligned with one of the coordinate axes.
• Add TriMesh::intersection_with_plane for intersecting with arbitrary planes.
• Add TriMesh::intersection_with_local_plane for intersecting with arbitrary planes in the same space as the mesh
• Add IntersectResult as the output type for the above functions.
• Add Polyline::extract_connected_components which splits a compound polyline into its connected components.
• Add implementations of bytemuck::Pod and bytemuck::Zeroable for all the simple shapes that allow it (Cuboid, Ball, Cone, etc.), and for bounding volumes (BoundingSphere and Aabb).

Modified

• Qbvh::leaf_data now requires &self instead of &mut self.
• Replace the Qbvh::leaf boolean by a bitflags.

Added

• Add Qbvh::remove, Qbvh::pre_update_or_insert, Qbvh::refit, Qbvh::rebalance to allow modifying a Qbvh without having to rebuild it completely.
• Add QbvhNode::is_leaf to get if a node is a leaf or not.
• Add SharedShape::trimesh_with_flags for building a trimesh with specific pre-processing flags.

Fixed

• Fix Triangle::contains_point.

Added

• Add SharedShape::trimesh_with_flags for constructing a triangle mesh with flags specified by the user.

Modified

• Rename AABB to Aabb to comply with Rust’s style guide.
• Rename QBVH to Qbvh to comply with Rust’s style guide.

Added

• Add ConvexPolygon::offsetted to dilate a polygon.
• Add CudaTriMesh and CudaTriMeshPtr for triangle-meshes usable with CUDA.
• Add a no-std implementation of point-projection on a triangle mesh.

Fixed

• Fix ghost collisions on internal edges on flat 3D meshed and flat 3D heightfields.
• Fix pseudo-normals calculation that could generate invalid normals for triangles with some small vertex angles.
• Fix Aabb::bounding_sphere which returned a bounding sphere that was too big.

Modified

• Add to query::time_of_impact a boolean argument stop_at_penetration. If set to false the linear shape-cast won’t immediately stop if the shape is penetrating another shape at its starting point and its trajectory is such that it’s existing that penetration configuration.

Added

• Add 2D Heightfield::to_polyline to get the explicit vertices/indices of a 2D heightfield seen as a polyline.
• Add the support for linear shape-cast (query::time_of_impact) for heightfields.
• Make the convex polyhedron scaling more forgiving regarding normals to avoid frequent unjustified panics.
• Fix panic happening when building a convex polyhedron with empty inputs.
• Add the support of Heightfields on CUDA kernels written in Rust using the cust crate.
• Add the rkyv-serialize feature that enables the implementation of rkyv serialization/deserialization for most shapes.
• Add the parallel feature that enables methods for the parallel traversal of Qbvh trees: Qbvh::traverse_bvtt_parallel, Qbvh::traverse_bvtt_node_parallel, Qbvh::traverse_depth_first_parallel, Qbvh::traverse_depth_first_node_parallel.

Fixed

• Fix the application of non-uniform scaling to balls.

Modified

• Remove &self argument from Compound::decompose_trimesh.
• Switch to cust 0.3 (for partial CUDA support).
• Rename RoundShape::base_shape to RoundShape::inner_shape.

Added

• Allow custom balancing strategies for the Qbvh construction. Some strategies are allowed to generate new leaves during the splitting process.
• Allow using point projection on heightfields from a CUDA kernel.
• Add the simultaneous traversal of two Qbvhs.
• Add computation of MassProperties for a TriMesh.
• Add .to_outline methods to compute the outline of a 3D shape (useful for debug-rendering).
• Add method to apply a scaling factor to some shapes. Shapes not supporting non-uniform scaling (like balls) will return a convex approximation of the scaled result.
• Add methods to split (into up to two parts) a Cuboid, Segment, or TriMesh with an axis-aligned plane.
• Add the computation of the intersection mesh between two TriMesh, or between a Cuboid and a TriMesh.

Modified

• Until now, the orientation of the polygon computed by 2D convex hull computation parry2d::transformation::convex_hull and parry2d::transformation::convex_hull_idx wasn't specified (and was generally in clockwise order). Now, this orientation is explicitly specified in the documentation and is set to counter-clockwise order (which is coherent with orientation expected by, e.g., the ConvexPolygon type).

Added

• Add parry::utils::obb which computes a (possibly sub-optimal) OBB for a set of points.
• Add Polyline::project_local_point_assuming_solid_interior_ccw which projects a point on the polyline contour, and is able to detect if that points is located inside of the polyline, assuming that the polyline is closed and oriented counter-clock-wise.
• Add (3D only) TriMesh::compute_pseudo_normals. If this is called, and if the trimesh respects some constraints ( oriented with outward normals, manifold almost everywhere, etc.) any subsequent point queries with the solid argument set to true will properly set the PointProjection::is_inside to true when the point lies in the interior of the trimesh.
• Added the implementation of the ear-clipping and Hertel-Mehlhorn algorithm that can be used for 2D triangulation and convex decomposition.
• Add the ability to use a subset of Parry’s features in a no-std context. If the new cuda cargo feature of Parry is enabled, all features compatible with no-std can be used inside of a CUDA kernel written in Rust thanks to the rust-cuda ecosystem.

Fixed

• Fix the orientation of the polygons generated by the 2D convex polygon decomposition (they are now always oriented counter-clockwise as expected by the ConvexPolygon type).
• Fix the intersection test between a 2D ball and a 2D compound shape.

Added

• Add the method Aabb::volume to compute the volume of an Aabb.

Modified

• Update the codebase to use nalgebra v0.29.

Fixed

• Fix a bug where the normal returned by ray-casting on polylines would not be normalized.

Added

• Implement Debug, Clone, PartialEq for VHACDParameters.
• Add a method to reverse the order of a polyline.
• Add a method to remove duplicate vertices form a TriMesh (and adjusting the index buffer accordingly).
• Add a method to iterate through all the lean data stored by a Qbvh.
• Implement the Interval Newton Method for computing all the roots of a non-linear scalar function.
• Implement the intersection test between a spiral and an Aabb.

Modified

• Rename all occurrences of quadtree to qbvh. Using the term quadtree was not representative of the actual acceleration structure being used (which is a BVH).

Fixed

• Fix a bug where query::contact would return None for a all intersecting a cuboid in such a way that one of the cuboid's vertices coincides exactly with the ball's center.

Modified

• Updated all dependencies to their latest version.

Fixed

• Fix ray-casting against solid triangles.
• Fix NaN when adding mass-properties with a zero mass.