Previously when harnessing missing keys, the identity function was
supplied, returning the job it was passed. This is very useful for
uniformly harnessing a set of jobs which may or may not define the
particular lifecycle being harnessed.
However, identity (of arity 1) is only sufficient for 2/3 cases -
pre-process and process. post-process has arity 2 (job res) and we
need to use an appropriate identity function in that case.
(The code looks good but the test case actually confuses me--is there a simpler explanation or possibly a slightly different phrasing of the test case?)
Previously when harnessing missing keys, the identity function was supplied, returning the job it was passed. This is very useful for uniformly harnessing a set of jobs which may or may not define the particular lifecycle being harnessed.
However, identity (of arity 1) is only sufficient for 2/3 cases - pre-process and process. post-process has arity 2 (job res) and we need to use an appropriate identity function in that case.