chenlin/Mujoco_WASM
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Re-implement VFS internals in C++, removing the constraints of the previous implementation.

Browse Source 
PiperOrigin-RevId: 649433087
Change-Id: Icaa3e6b7f2ef14f56b04fef2f4360b46d8b5f022
This commit is contained in:
Kyle Bayes
2024-07-04 09:21:18 -07:00
committed by Copybara-Service
parent 118b6810b3
commit 57e6760ec9
19 changed files with 513 additions and 717 deletions
Download Patch File Download Diff File Expand all files Collapse all files
doc/changelog.rst
+22 -19
View File
@@ -10,11 +10,12 @@ General
.. admonition:: Breaking API changes
:class: attention
1. Removed deprecated ``mj_makeEmptyFileVFS`` and ``mj_findFileVFS`` functions.
1. Removed deprecated ``mj_makeEmptyFileVFS`` and ``mj_findFileVFS`` functions. The constants ``mjMAXVFS`` and
``mjMAXVFSNAME`` are also removed as they are no longer needed.
**Migration:** Use :ref:`mj_addBufferVFS` to copy a buffer into a VFS file directly.
2. Calls to:ref:`mj_defaultVFS` may allocate memory inside VFS, and the corresponding
2. Calls to :ref:`mj_defaultVFS` may allocate memory inside VFS, and the corresponding
:ref:`mj_deleteVFS` must be called to deallocate any internal allocated memory.
3. Deprecated :ref:`mju_rotVecMat` and :ref:`mju_rotVecMatT` in favor of :ref:`mju_mulMatVec3` and
@@ -22,7 +23,9 @@ General
The older functions have been removed from the Python bindings and will be removed from the C API in the next
release.
4. Added a new API for :doc:`procedural model manipulation<programming/modeledit>`. Fixes :github:issue:`364`.
4. The :ref:`VFS<Virtualfilesystem>` implementation has been rewritten in C++ and is now considerably more efficient in
speed and memory footprint.
5. Added a new API for :doc:`procedural model manipulation<programming/modeledit>`. Fixes :github:issue:`364`.
Still missing:
- Detailed documentation.
@@ -31,27 +34,27 @@ General
:align: right
:width: 240px
5. Added support for orthographic cameras. This is available for both fixed cameras and the free camera, using the
6. Added support for orthographic cameras. This is available for both fixed cameras and the free camera, using the
:ref:`camera/orthographic<body-camera-orthographic>` and :ref:`global/orthographic<visual-global-orthographic>`
attributes, respectively.
6. Added :ref:`maxhullvert<asset-mesh-maxhullvert>`, the maximum number of vertices in a mesh's convex hull.
7. Added :ref:`mj_setKeyframe` for saving the current state into a model keyframe.
8. Added support for ``ball`` joints in the URDF parser ("spherical" in URDF).
9. Replaced ``mjUSEDOUBLE`` which was previously hard-coded in
`mjtnum.h <https://github.com/google-deepmind/mujoco/blob/main/include/mujoco/mjtnum.h>`__
with the build-time flag ``mjUSESINGLE``. If this symbol is not defined, MuJoCo will use double-precision floating
point, as usual. If ``mjUSESINGLE`` is defined, MuJoCo will use single-precision floating point. See :ref:`mjtNum`.
7. Added :ref:`maxhullvert<asset-mesh-maxhullvert>`, the maximum number of vertices in a mesh's convex hull.
8. Added :ref:`mj_setKeyframe` for saving the current state into a model keyframe.
9. Added support for ``ball`` joints in the URDF parser ("spherical" in URDF).
10. Replaced ``mjUSEDOUBLE`` which was previously hard-coded in
`mjtnum.h <https://github.com/google-deepmind/mujoco/blob/main/include/mujoco/mjtnum.h>`__
with the build-time flag ``mjUSESINGLE``. If this symbol is not defined, MuJoCo will use double-precision floating
point, as usual. If ``mjUSESINGLE`` is defined, MuJoCo will use single-precision floating point. See :ref:`mjtNum`.
Relatedly, fixed various type errors that prevented building with single-precision.
10. Quaternions in ``mjData->qpos`` and ``mjData->mocap_quat`` are no longer normalized in-place by
Relatedly, fixed various type errors that prevented building with single-precision.
11. Quaternions in ``mjData->qpos`` and ``mjData->mocap_quat`` are no longer normalized in-place by
:ref:`mj_kinematics`. Instead they are normalized when they are used. After the first step, quaternions in
``mjData->qpos`` will be normalized.
MJX
~~~
10. Added support for :ref:`elliptic friction cones<option-cone>`.
11. Fixed a bug that resulted in less-optimal linesearch solutions for some difficult constraint settings.
12. Fixed a bug in the Newton solver that sometimes resulted in less-optimal gradients.
12. Added support for :ref:`elliptic friction cones<option-cone>`.
13. Fixed a bug that resulted in less-optimal linesearch solutions for some difficult constraint settings.
14. Fixed a bug in the Newton solver that sometimes resulted in less-optimal gradients.
.. youtube:: P83tKA1iz2Y
@@ -60,14 +63,14 @@ MJX
Simulate
^^^^^^^^
13. Added improved tutorial video.
14. Improved the Brownian noise generator.
15. Added improved tutorial video.
16. Improved the Brownian noise generator.
|br| |br| |br| |br|
Python bindings
^^^^^^^^^^^^^^^
15. Fixed a memory leak when using ``copy.deepcopy()`` on a ``mujoco.MjData`` instance (:github:issue:`1572`).
17. Fixed a memory leak when using ``copy.deepcopy()`` on a ``mujoco.MjData`` instance (:github:issue:`1572`).
Version 3.1.6 (Jun 3, 2024)
---------------------------