Update startup log tests to account for updated Node.js behavior

[bm1549]

Motivation

• fix(startup-log): emit config early at init, integrations on first ag… dd-trace-js#7643 the trace flush is no longer needed for Node.js to trigger config logs
• feat(config): enable startup logs by default dd-trace-js#7474 in v6, startup logs are enabled by default

Changes

• Removes the trace flush from Node.js
• Updates the manifest for Node.js default behavior for v6

Workflow

1. ⚠️ Create your PR as draft ⚠️
2. Work on you PR until the CI passes
3. Mark it as ready for review
   • Test logic is modified? -> Get a review from RFC owner.
   • Framework is modified, or non obvious usage of it -> get a review from R&P team

🚀 Once your PR is reviewed and the CI green, you can merge it!

🛟 #apm-shared-testing 🛟

Reviewer checklist

• Anything but tests/ or manifests/ is modified ? I have the approval from R&P team
• A docker base image is modified?
  • the relevant build-XXX-image label is present
• A scenario is added, removed or renamed?
  • Get a review from R&P team

[github-actions]

CODEOWNERS have been resolved as:

manifests/nodejs.yml                                                    @DataDog/dd-trace-js
tests/parametric/test_startup_logs.py                                   @DataDog/system-tests-core @DataDog/apm-sdk-capabilities

[simon-id]

Did you run the test locally to make sure the test passes ? The CI is not going to run the test you activated right now because the declaration is >=6.0.0 and the tested dev version is 6.0.0-pre, and in npm semver, prerelease versions do not match with non prerelease versions.
Just saying this test will not be run at all until we release v6, then it might start failing out of nowhere and break CI for everyone in system tests.

[bm1549]

Did you run the test locally to make sure the test passes ? The CI is not going to run the test you activated right now because the declaration is >=6.0.0 and the tested dev version is 6.0.0-pre, and in npm semver, prerelease versions do not match with non prerelease versions. Just saying this test will not be run at all until we release v6, then it might start failing out of nowhere and break CI for everyone in system tests.

@simon-id - yep, just confirmed it xpass'd locally for dd-trace-js v6

[simon-id]

@bm1549 that's good, but since that doesn't add it to the CI, the risk of it starting to silently fail after you merge this, and start failing the whole system test CI once we release v6 is suboptimal

[bm1549]

@simon-id misinterpreted your prior feedback - just added updated the ref in the manifest so it runs on 6.x

[datadog-official]

✅ Tests

🎉 All green!

❄️ No new flaky tests detected
🧪 All tests passed

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 50dd378 | Docs | Datadog PR Page | Was this helpful? React with 👍/👎 or give us feedback!}

[nccatoni]

Detail: I would suggest using the last failing version with > (something like >5.90.0) or the current pre-release version directly so it is more readable

[simon-id]

@nccatoni this isn't possible actually, prerelease versions cannot be compared with non-prerelease versions, and this change will only apply when v6.0.0 comes out, we don't know when

[nccatoni]

@simon-id - I'm not sure what you mean by that. Pre release versions can be compared to any other version without problem otherwise the manifests wouldn't work. We follow the semver standard for version comparison. You can checkout the ruby, go or php manifests for exemples if you want

[simon-id]

@nccatoni ah i just tried locally and our implementation does match what you're saying, it's the npm implementation that is different which is why i always get confused with prerelease stuff. The semver standard used to be ambiguous about prerelease comparison, which looks like it's been fixed in the latest draft of the spec.

With our ST custom implem:

range = SemverRange('>=6.0.0.pre')
>>> semver.Version('7.0.0-pre') in range
True

In node.js:

> require('semver').satisfies('6.0.0', '>=6.0.0.pre')
false

So we can use >=6.0.0-pre like you said, unless I'm missing another detail.

(btw your idea of using >5.90.0 is not possible because we will never know which version of the v5.x branch will be the final one, as we operate on multiple release lines at the same time)

[simon-id]

LGTM but once CI passes, double check it's correctly running the test in dev, but not in prod 👍

[bm1549]

@simon-id just double checked. Merging now

Thanks for all the feedback!