I'm not certain this will have a major effect on jobs, but just in case, this reorders so that the timeout of a job is done after the processing of commands. This should allow a client to stop a job being destroyed just outside of the timeout.
Coverage increased (+0.0006%) to 99.821% when pulling 830524f2982ea69130f5c7f9a4ae1b8cbb6d06fe on reorder_timeout into 48ebf24e9c42577e95d8c1a7e2270cacbe84052b on master.
I'm not certain this will have a major effect on jobs, but just in case, this reorders so that the timeout of a job is done after the processing of commands. This should allow a client to stop a job being destroyed just outside of the timeout.