This buildpack can now use the tiny stack with reloadable processes

[joshuatcasey]

Summary

This buildpack can now use the tiny stack with reloadable processes. While #296 addressed this, there was some logic in the Detect method that was not removed.

See #296 (comment) for additional context.

Checklist

• I have viewed, signed, and submitted the Contributor License Agreement.
• I have linked issue(s) that this PR should close using keywords or the Github UI (See docs)
• I have added an integration test, if necessary.
• I have reviewed the styleguide for guidance on my code quality.
• I'm happy with the commit history on this PR (I have rebased/squashed as needed).

[fg-j]

I think the integration test is failing because of the WithPullPolicy("never"). The builder isn't present on the GHA runner docker daemon.

[joshuatcasey]

Good catch @fg-j . "It worked on my machine!"

Will push a fix in a minute.

[joshuatcasey]

@fg-j it looks like using the tiny builder is causing further integration test problems. Should I pull it out?

I admit that this integration test would have caught the logic in Detect that this PR fixes, but I feel like this PR addresses that and so there's little need to make sure an integration test verifies this.

[fg-j]

Hmm... I'm concerned that that integration test isn't passing. It should be, if the buildpack actually works with watchexec. This warrants more investigation.

[ForestEckhardt]

Grabbed the app locally and I get the following error

[[Error (not fatal)]]
io(spawning process group): No such file or directory (os error 2)

[joshuatcasey]

I've completely restored the integration test to its original state. Even though this has passed on my machine it's pretty flaky even locally.

I wonder if there's some nuance of how the builder is pulled or used that may be an issue. A quick grep through paketo codebases indicates that integration tests don't usually set the builder, although smoke tests often do (for samples and builders).

[fg-j]

@joshuatcasey You might get a clearer idea of what's failing in the container on startup if you turn up the verbosity on watchexec
See the watchexec man page

[joshuatcasey]

@ForestEckhardt would you mind taking a look at the PR? I think we've gotten past the integration test issue.