Closed m8pple closed 2 years ago
(Though looking for "unintelligible" is a great detector of certain classes of failure).
We will be more consistent on making sure that every commands produces a logged output with the correct I/W/E/etc. flag.
Closing. Idealistic, and not really critical for anything.
(This is based on
1.0.0-alpha
(6669d06a9d2a7c4e6b732d8a719a0e6b69f2226d), but may also apply todevelopment
).(This is related to working around #197 and #198, and trying to do automated testing with the orchestrator.)
When trying to develop scripts that will test whether a given orchestrator passed or failed, it is difficult to work out whether a common succeeded or failed in a uniform way.
For example, the command
load /app
will return either201(E)
or65(I)
, so we can watch for those two results and check whether it completed.However, other commands do not print anything out. For example,
tlink /app = *
prints nothing if it succeeds (or maybe it fails - not sure). I guess you can look in the microlog?When trying to adapt the interactive interface for batch testing it would help to have a really un-ambiguous result for each command: did it succeed, or did it fail?
Ideally there would be something that was consistent for all commands, but at least having two distinct outputs/codes in
OrchestratorMessages.ocfg
would make it much easier to script.