Improve empty result handling

[skjolber]

Make sure console output and report is written when there are features detected, but no vulernabilities.

Specific error code for when no features have been detected in the image.

[skjolber]

@arminc seems like an error with the build?

$ goveralls -coverprofile=coverage.out -service=travis-ci -repotoken $COVERALLS_TOKEN
flag needs an argument: -repotoken
...
The command "goveralls -coverprofile=coverage.out -service=travis-ci -repotoken $COVERALLS_TOKEN" exited with 2.

[arminc]

The build works on master (release) but somehow goveralls broke in the PR's. I haven't had time to fix it.
Regarding the PR: it looks good but I don't know if people deliberately check for exit 1 instead of not 0, if it the first one adding the extra exit code will mess up their workflow. I do agree that having distinction between the two is interesting.

[skjolber]

I see your point about the error code, perhaps bump a minor version? The distinction between 'unsupported' and 'zero vulernabilities' is quite significant.

For your info, my use-case is the following:

• collect a lot of images from running services in kubernetes
• scan the images
• upload scan result to a log backend
• generate report in log backend