register_pipeline
def register_pipeline(mut self: Scheduler, binary: String, now: Int64, flags: RegisterFlags = RegisterFlags(), file: ProjectConfig = ProjectConfig()) -> StringRecords the pipeline binary describes, with its path; its schedule counts from now the first time, and a sensor from now the first time it's one, with the snapshots its datasets have now seen. Queued runs move to the new code, and so do failed ones waiting for a retry with no step done. How its runs run is flags over what file, the project's orca.toml (orca.config), says for it: for a pipeline the file names, exactly that, a setting neither gives going back to its default; for one it doesn't, the flags given, each left out keeping what was registered (--workers MIN..MAX scales an EC2 run's workers, --worker-procs is the dista worker processes per EC2 worker instance, --secrets replace the pipeline's, --placement is control, ec2 or default). Returns the pipeline's name.