Learnt from: lenroe-nv
Repo: newton-physics/newton PR: 1248
File: newton/_src/geometry/contact_data.py:47-49
Timestamp: 2025-12-13T17:26:39.791Z
Learning: In ContactData (newton/_src/geometry/contact_data.py), the fields contact_stiffness, contact_damping, and contact_friction_scale use 0.0 as the "unset" sentinel value. This is intentional: Warp struct fields initialize to 0.0 by default, and solver-side code interprets 0.0 as "use default/unscaled" rather than "disable." This design avoids explicit initialization in contact generation kernels.

Learnt from: lenroe-nv
Repo: newton-physics/newton PR: 1248
File: newton/_src/geometry/contact_data.py:47-49
Timestamp: 2025-12-13T17:26:39.791Z
Learning: In ContactData (newton/_src/geometry/contact_data.py), the fields contact_stiffness, contact_damping, and contact_friction_scale use 0.0 as the "unset" sentinel value. This is intentional: Warp struct fields initialize to 0.0 by default, and solver-side code interprets 0.0 as "use default/unscaled" rather than "disable." This design avoids explicit initialization in contact generation kernels.

Learnt from: adenzler-nvidia
Repo: newton-physics/newton PR: 1107
File: newton/_src/solvers/mujoco/kernels.py:973-974
Timestamp: 2025-11-24T08:05:21.390Z
Learning: In Newton's MuJoCo solver integration (newton/_src/solvers/mujoco/), the mapping between Newton joints and MuJoCo joints is not 1-to-1. Instead, each Newton DOF maps to a distinct MuJoCo joint. This means that for multi-DOF joints (like D6 with 6 DOFs), there will be 6 corresponding MuJoCo joints, each with its own properties (margin, solimp, solref, etc.). The mapping is done via dof_to_mjc_joint array, ensuring each DOF's properties are written to its own MuJoCo joint without overwriting.

Learnt from: dylanturpin
Repo: newton-physics/newton PR: 806
File: newton/examples/ik/example_ik_franka.py:121-123
Timestamp: 2025-09-22T21:08:31.901Z
Learning: In the newton physics framework, when creating warp arrays for IK solver joint variables using wp.array(self.model.joint_q, shape=(1, coord_count)), the resulting array acts as a reference/pointer to the original model's joint coordinates, so updates from the IK solver automatically reflect in the model's joint_q buffer used for rendering, despite the general warp documentation suggesting copies are made by default.

Learnt from: dylanturpin
Repo: newton-physics/newton PR: 634
File: newton/examples/ik/example_ik_franka.py:0-0
Timestamp: 2025-08-25T20:10:59.536Z
Learning: In Newton's IK examples, when creating solver arrays with `wp.array(source_array, shape=new_shape)`, this creates a view into the same underlying memory rather than a copy. This means updates to the reshaped array are automatically reflected in the original array without needing explicit copy operations like `wp.copy()`.

Learnt from: preist-nvidia
Repo: newton-physics/newton PR: 579
File: newton/examples/example_mujoco.py:350-354
Timestamp: 2025-08-19T12:27:30.629Z
Learning: In Newton examples, Example.use_mujoco is a high-level attribute that controls whether to use MuJoCo solver vs other solvers (like XPBD), while SolverMuJoCo.use_mujoco_cpu controls the backend within MuJoCo (CPU vs Warp). These are separate concepts serving different purposes - the PR rename only applies to the solver parameter, not the Example class attributes.

Learnt from: preist-nvidia
Repo: newton-physics/newton PR: 579
File: newton/examples/example_mujoco.py:350-354
Timestamp: 2025-08-19T12:27:30.629Z
Learning: In Newton examples, there's a distinction between solver parameters and Example class attributes. The Example class can have its own use_mujoco attribute for controlling example-level behavior (like CUDA graphs, rendering logic), while the solver uses use_mujoco_cpu for backend selection. These serve different purposes and should not be conflated during API renames.

Learnt from: nvtw
Repo: newton-physics/newton PR: 899
File: newton/tests/test_rigid_contact.py:214-268
Timestamp: 2025-10-10T10:47:41.082Z
Learning: In `newton/tests/test_rigid_contact.py`, using `add_shape_plane(width=0, length=0)` for an infinite plane works correctly with the collision system and does not cause degenerate AABB issues in the test scenario `test_shape_collisions_gjk_mpr_multicontact`.

Learnt from: nvlukasz
Repo: newton-physics/newton PR: 519
File: newton/tests/test_ik.py:25-26
Timestamp: 2025-08-12T17:58:16.929Z
Learning: In the Newton physics engine codebase, it's acceptable for tests to import and use private `_src` internal modules and functions when needed for testing purposes, even though this breaks the public API boundary. This is an intentional project decision for test code.