Update MSTest to metapackage and MTP

[Youssef1313]

FYI @Evangelink

[DaveTryon]

/azp run

[DaveTryon]

@Youssef1313 and @Evangelink, tests are failing in all environments with these changes. Please see https://dev.azure.com/mseng/1ES/_build/results?buildId=29527577&view=results for details.

[Youssef1313]

@DaveTryon Navigating to the build shows "Build not found." for me.

[DaveTryon]

@Youssef1313, I've emailed the logs to you and @Evangelink. They're compressed and less than 1MB, so hopefully they'll make it through your email gateway,

[Youssef1313]

@DaveTryon I see you are adding coverage support for coverlet in #895. For MTP, Microsoft CodeCoverage is better supported. Are you okay with switching from coverlet to Microsoft CodeCov?

[Evangelink]

/azp run

[azure-pipelines]

Commenter does not have sufficient privileges for PR 881 in repo microsoft/sbom-tool

[DaveTryon]

/azp run

[azure-pipelines]

Pull request contains merge conflicts.

[Youssef1313]

@DaveTryon I just fixed the conflict

[DaveTryon]

/azp run

[DaveTryon]

@DaveTryon I see you are adding coverage support for coverlet in #895. For MTP, Microsoft CodeCoverage is better supported. Are you okay with switching from coverlet to Microsoft CodeCov?

I'm not married to coverlet, but I'd like to see the MTP version before making the final call. I'll hold off on merging the PR for a bit so we can do a side-by-side comparison.

[DaveTryon]

/azp run

[DaveTryon]

@DaveTryon it seems that E2E tests were not run before and now they are run causing the timeout. Is this expected that they run or not?

That's an interesting question. the unit tests only target net472 on Windows, but the e2e tests target net472 everywhere. The test runner, though, ignores the non-Windows net472 tests, so whether it's intentional or not, the net472 tests are effectively limited to Windows. I've sent a message to the team to see if they're OK with codifying the de facto behavior.

I did some experimenting in a branch based on your fork, and made 2 commits that you can backport to get past the test failures:

• This commit follows the pattern from the unit tests
• This commit gives MTP a test that succeeds, since apparently it treats it as a failure if all tests are skipped. There may also be a command line way to do this

After these commits are added, we still get an error because each stage is trying to upload to the same artifact name, and only the first one can win. Let's just go with code coverage results from Windows, since it's the platform with the broadest coverage

[Evangelink]

@dave

[Youssef1313]

@DaveTryon Could you run the pipeline again? Thanks!

[DaveTryon]

/azp run

[DaveTryon]

@DaveTryon Could you run the pipeline again? Thanks!

@Youssef1313, it looks like the yaml is broken. The pipeline failed immediately, generating the 2 following errors but no logs:

The 'stages' parameter is not a valid StageList

/pipelines/build-test-tool-template.yaml (Line: 34, Col: 5): 'condition' is already defined

[DaveTryon]

/azp run

[DaveTryon]

@Youssef1313, the build is green again, so I think we're ready for the final clean-up phase. I'll email you the logs

[DaveTryon]

/azp run

[DaveTryon]

@Youssef1313 This is super close, but now the code coverage isn't showing up in the pipeline like it used to. Happy to merge this as-is and have the code coverage be a separate PR if that makes more sense than adding it to this PR

[DaveTryon]

/azp run

[DaveTryon]

Thanks, @Youssef1313!