diff options
| author | Sébastien Crozet <sebcrozet@dimforge.com> | 2024-05-25 11:36:14 +0200 |
|---|---|---|
| committer | Sébastien Crozet <sebastien@crozet.re> | 2024-05-25 12:00:15 +0200 |
| commit | 798406e00de86b78d86ae8903b2c598938fa1089 (patch) | |
| tree | 25aa2fd00f5f04ea56800c548f149c2adbd01a66 | |
| parent | cdec395d094bfcc1e5c6ea7d49972f7d2ae3e11b (diff) | |
| download | rapier-798406e00de86b78d86ae8903b2c598938fa1089.tar.gz rapier-798406e00de86b78d86ae8903b2c598938fa1089.tar.bz2 rapier-798406e00de86b78d86ae8903b2c598938fa1089.zip | |
chore: update changelog
| -rw-r--r-- | CHANGELOG.md | 9 |
1 files changed, 9 insertions, 0 deletions
diff --git a/CHANGELOG.md b/CHANGELOG.md index a7c1efe..f9ec4d1 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -14,12 +14,21 @@ ### Modified +- The contact constraints regularization parameters have been changed from `erp/damping_ratio` to + `natural_frequency/damping_ratio`. This helps define them in a timestep-length independent way. The new variables + are named `IntegrationParameters::contact_natural_frequency` and `IntegrationParameters::contact_damping_ratio`. +- The `IntegrationParameters::normalized_max_penetration_correction` has been replaced + by `::normalized_max_corrective_velocity` + to make the parameter more timestep-length independent. It is now set to a non-infinite value to eliminate aggressive + "popping effects". - The `Multibody::forward_kinematics` method will no longer automatically update the poses of the `RigidBody` associated to each joint. Instead `Multibody::update_rigid_bodies` has to be called explicitly. - The `Multibody::forward_kinematics` method will automatically adjust the multibody’s degrees of freedom if the root rigid-body changed type (between dynamic and non-dynamic). It can also optionally apply the root’s rigid-body pose instead of the root link’s pose (useful for example if you modified the root rigid-body pose externally and wanted to propagate it to the multibody). +- Remove an internal special-case for contact constraints on fast contacts. The doesn’t seem necessary with the substep + solver. ## v0.19.0 (05 May 2024) |