Runtime evaluate method does not set previous pipeline run id correctly

In the runtime's evaluate method, it seems that the previous pipeline run id is being set correctly. I have not yet identified the specific cause. It might be caused by setting the previous_pipeline_run_id in the constructor and then the previous pipeline run is being modified by combine_pipeline_runs as part of evaluate.

See branch previous-pipeline-run-id for a reproducible example. Run python3 tests/test_cli_runtime.py TestRuntimeCmd.test_evaluate to see the failure.