Fail Action if before-build or after-build Return a Non-Zero Exit Code

[olly]

In a similar fashion to the other commands, after running the before-build or after-build steps we should fail the action if they return with a non-zero exit code.

---

In our setup we use simplecov, of which versions after 0.17, produce a coverage report whose format is incompatible with cc-test-reporter. This produces failures in the after-build command, for example:

2020-08-18T10:02:02.6528716Z [command]/home/runner/work/cms/cms/cc-reporter after-build --exit-code 0
2020-08-18T10:02:02.7785615Z Error: json: cannot unmarshal object into Go struct field input.coverage of type []formatters.NullInt
2020-08-18T10:02:02.7792113Z Usage:
2020-08-18T10:02:02.7792927Z   cc-test-reporter after-build [flags]
2020-08-18T10:02:02.7793147Z 
2020-08-18T10:02:02.7793360Z Flags:
2020-08-18T10:02:02.7793764Z   -s, --batch-size int               batch size for source files (default 500)
2020-08-18T10:02:02.7794312Z   -e, --coverage-endpoint string     endpoint to upload coverage information to (default "https://api.codeclimate.com/v1/test_reports")
2020-08-18T10:02:02.7794860Z   -t, --coverage-input-type string   type of input source to use [clover, cobertura, coverage.py, excoveralls, gcov, gocov, jacoco, lcov, simplecov, xccov]
2020-08-18T10:02:02.7795281Z       --exit-code int                exit code of the test run
2020-08-18T10:02:02.7796525Z   -r, --id string                    reporter identifier (default "***")
2020-08-18T10:02:02.7797282Z       --insecure                     send coverage insecurely (without HTTPS)
2020-08-18T10:02:02.7797996Z   -p, --prefix string                the root directory where the coverage analysis was performed (default "/home/runner/work/cms/cms")
2020-08-18T10:02:02.7798276Z 
2020-08-18T10:02:02.7798482Z Global Flags:
2020-08-18T10:02:02.7798946Z   -d, --debug   run in debug mode

In our specific case, as we use dependabot to update dependencies, we'd like this to fail the action to prevent the dependency from being updated. More generally I think it makes sense for this to fail the action in order to alert the user to fix the issue, although this could potentially be a breaking change for some user's setups.

I suspect this would fix #181

[paambaati]

@olly Thanks a lot for this PR, I'll be rebasing this against latest master and adding tests in #231 if that's okay with you.

[olly]

@paambaati of course that's OK with me. Thanks for merging. 🙏