add ability to fail watch if a spawned process exits with an error code#2

at the moment we are not treating non-0 exit codes as an error, which ensures that watchrow does not break if a spawned process produces an error. However, this also means that it is easy not to notice a failing trigger. Need to research how other tools are handling this scenario Regardless, the current implementation is flawed because it does not interrupt broken workflow, e.g. ```ts await spawn`docker build ...`; await spawn`docker run ...`; ``` If `docker build` fails, `docker run` still runs.