s2i steps: cannot check logs when s2i build fails

[jmtd]

s2i builds all take place via the helper routine s2i_inner

behave-test-steps/steps/s2i_steps.py

Line 14 in e7e9c72

def s2i_inner(context, application, path='.', env="", incremental=False, tag="master", runtime_image=""):

which calls the s2i binary via the helper routine _execute

behave-test-steps/steps/steps.py

Line 18 in e7e9c72

def _execute(command, log_output=True):

_execute returns False if the command fails, otherwise it (sometimes) returns the command's output. s2i_inner saves the command output returned by _execute if the command succeeds, but not if it fails (since it hasn't been returned).

We provide a test step to check for a failing s2i build:

behave-test-steps/steps/s2i_steps.py

Line 72 in e7e9c72

@given(u'failing s2i build {application} from {path} using {tag}')

We also provide steps to inspect the logs, e.g.,

behave-test-steps/steps/s2i_steps.py

Line 80 in e7e9c72

@then(u's2i build log should contain {phrase}')

However, none of these follow-on steps can be used in conjunction with the failing s2i build case, because the build logs are not saved.

I think we should capture and save the build logs in either the success or failure case. It looks like the only caller of _execute is s2i_inner, luckily.