Disables articulation tests in Isaac Sim CI#5557
Conversation
Greptile SummaryThis PR removes the
Confidence Score: 5/5Safe to merge; the change only removes pytest markers from test files and cannot affect production code or runtime behaviour. Both files are fully cleaned — zero isaacsim_ci markers remain after the change. The removal is mechanically correct and consistent across both packages. The only open question is the missing tracking mechanism for restoring the tests later, but that is a process concern rather than a defect in the code itself. No files require special attention; changes are limited to pytest marker annotations in two test files. Important Files Changed
Flowchart%%{init: {'theme': 'neutral'}}%%
flowchart TD
A[CI Pipeline triggers pytest] --> B{Marker filter: isaacsim_ci?}
B -- Before PR --> C[Collect all tests with @pytest.mark.isaacsim_ci ~50 articulation tests selected]
B -- After PR --> D[No tests match isaacsim_ci in test_articulation.py files 0 articulation tests run]
C --> E[Tests run → timeouts in CI]
D --> F[CI unblocked; articulation tests skipped until root cause is resolved]
Reviews (1): Last reviewed commit: "add changelog fragments" | Re-trigger Greptile |
NewtonArticulationData._create_simulation_bindings rebound
_sim_bind_root_com_vel_w / _sim_bind_body_com_vel_w on every reset,
including when get_root_velocities / get_link_velocities returned None
for fixed-base articulations. The first call left a wp.zeros fallback
allocated by _create_buffers, but every subsequent sim.reset() then
overwrote that fallback with None, so _pin_proxy_arrays crashed:
TypeError: ProxyArray expects a warp.array, got NoneType.
Reproduces with any panda test that calls sim.reset() twice (e.g.
test_write_joint_frictions_to_sim) once the Newton articulation tests
were re-enabled in the regular CI suite by isaac-sim#5557.
Only assign the binding when the solver actually exposes one; otherwise
let the existing wp.zeros fallback persist.
Also wire model.articulation_count / max_joints_per_articulation /
max_dofs_per_articulation / joint_dof_count / body_count into the mock
NewtonManager used by test_articulation_iface; the task-space scratch
buffers added in this PR consume those sizes during data init and would
otherwise pick up MagicMock placeholders, triggering reshape mismatches
in wp.zeros / wp.array.reshape.
Description
articulation tests have been timing out in CI. disabling it to unblock isaac sim MR merging until we figure out why the timeouts are happening.
Type of change
Checklist
pre-commitchecks with./isaaclab.sh --formatconfig/extension.tomlfileCONTRIBUTORS.mdor my name already exists there