PYTHON-4609: [1/x] Migrate all pre and most post shell.exec to subprocess.exec

[Jibola]

This PR focuses on primarily converting the pre/post functions with a few additional artifact and bootstrapping functions included as well.

Functions with converted shell.exec --> subprocess.exec

• "prepare resources"
• "upload mo artifacts"
• "bootstrap mongo orchestration"
• "bootstrap data lake"
• "stop mongo-orchestration"
• "run aws ECS auth test"
• "windows fix"
• "make files executable"
• "init test-results"
• "install dependencies"

Functions Remaining (to be done in more PRs)

• "run tests" -- combined both shell.exec calls
• "download and merge coverage"
• "run mod_wsgi tests"
• "run mockupdb tests"
• "run doctests"
• "run enterprise auth tests"
• "run atlas tests"
• "run aws auth test with regular aws credentials"
• "run aws auth test with assume role credentials"
• "run aws auth test with aws EC2 credentials"
• "run aws auth test with aws web identity credentials"
• "run oidc auth test with aws credentials as environment variables"
• "run aws auth test with aws credentials and session token as environment variables"
• "cleanup"
• "run-ocsp-test"
• run-valid-ocsp-server
• run-revoked-ocsp-server
• run-valid-delegate-ocsp-server
• run-revoked-delegate-ocsp-server
• "run load-balancer"
• "stop load-balancer"
• "teardown_docker"
• "teardown_aws"
• "run perf tests"

[Jibola]

Re-working this to use some drivers-evergreen-tools version of scripts!

[caseyclements]

Linting errors in static tests are definitely real.

Not sure about missing log files in other tests. Seem like the type of difference one might find from making a shell change like this but should be picked up in all tests. Let's look through when you're back.

[blink1073]

Superseded by #1997