Comment #3 on issue 964 by [email protected]: WARN test result
http://code.google.com/p/robotframework/issues/detail?id=964
I get the point and accept the conclusion. Thanks for your time.
Nevertheless, as Bryan points out, the same test may have more outcomes, so
it is not just a matter of tagging it as noncritical, and a decision should
not be taken from this assumption.
I agree also that WARN is unusual, yet we encountered a situation when the
presence of some logs MAY indicate a problem, an this must be put in
evidence so that an engineer can check what happened. This is due to
inherent multiprocessing of the SUT and the happening of possibly
unexpected conditions.
In general, a flexible outcome result set would be the fit for all...
