Summary

• The v1.4 fixed-width bool-mask reset compatibility layer is removed in v1.5 instead of v1.6, per the decision on the driving issue.
• unilab.base.reset_selectors.resolve_reset_env_ids now fails closed with TypeError on any bool mask; the DeprecationWarning passthroughs in ResetMaskBuffer.resolve / resolve_reset_mask are gone.
• EventManager loses the mask-form reset-term boundary: dispatch_reset, _detect_mask_form_reset, _reset_term_mask_form, and the mask dispatch branch in _call_reset_term are removed. Reset event terms are only called as func(env, env_ids, **params).
• Internal mask consumers are unchanged: backend transactions (SimBackend.apply_reset / packed reset), the protected CommandTerm._resample_command(reset_mask) hook, NoiseModel.reset(reset_mask), and the motion-tracking tensor runtime keep the fixed-width mask form.
• Tests migrate to env_ids selectors, including stale slow-lane fixtures that referenced the already-removed __call_mask__ / reset_mask APIs (they were red on main); changelog and both 1.4→1.5 migration guides now record the immediate removal.

User-facing impact: passing a bool mask to any public reset boundary (TorchEnv.reset, ManagerBase.reset and all manager/term reset hooks, EventManager.apply(mode="reset", ...)) now raises TypeError; EventManager.dispatch_reset no longer exists.

Linked Work

• Issue: Work: shim/compatibility audit report for UniLab and unisim v2.0.0 #2474
• Parent roadmap (when applicable): roadmap: reset 公共契约恢复 env_ids（mjlab 式分层），mask 下沉为 backend 实现细节 #2459
• Roadmap declared base (when applicable): main
• Milestone: v1.5
• Base branch: main

Validation

• make test-all passed on the final local head before this PR was created or updated
• Additional task-specific validation listed below

Commands actually run:

make check        # mypy / pyright / ruff: all clean
make test         # 2760 passed, 0 failed
make test-all     # check + test-cov + benchmark smoke: passed on 334a7ed1
uv run pytest tests/envs/locomotion/g1/test_manager_terms.py -q -m slow   # 15 passed
uv run pytest tests/scripts/test_train_scripts.py::test_run_native_viewer_play_loop_runs_without_physics_state -q -m slow  # passed
uv run pytest tests/scripts/test_check_docs.py -q  # 20 passed

Remote CI route:

• Base main: current-head CI link and status to be recorded once the run starts.

Impact

• Backend impact: none (backend transaction mask form unchanged)
• Platform impact: both (pure contract change, platform-independent)
• Training effect expected: no

Artifacts

• W&B: n/a
• benchmark result: n/a
• video / screenshot: n/a
• ONNX / checkpoint: n/a

Checklist

• Added or updated tests where needed
• Updated docs if behavior or workflow changed
• Linked the driving issue
• Noted any follow-up work explicitly