Figure out how the CODEOWNERS Linter is supposed to deal with changes to the baseline errors file

[JimSuplizio]

This is in response to the following comment

The scenario here is that updates were made to the baseline errors file. What happened was a bunch of baseline errors were removed in a PR. PR processing generates a base branch baseline file against the base branch and some of the errors removed from the PR's baseline errors were in the baseline, which is used to do further filtering, allowed the PR through and didn't start failing until the nightly runs.

@weshaggard - I don't know how this is supposed to work if the CODEOWNERS_baseline_file.txt is updated as part of the PR.

• If we generate the base branch baseline file - Changes made in the PR, like removing an error, would still get filtered out by the generated one. This would allow the PR through, and the failures would start failing in the nightly run. This is where we are today
• If we don't generate the base branch baseline file - Any existing failures in nightly runs would now be the responsibility of the PR owner to fix in order to get the PR though. This is similar to where we were prior to generating the base branch baseline file

In theory, the files changed in the PR could be inspected and we could skip the base branch generation if and only if, the CODEOWNERS_baseline_file.txt file is part of the PR. That would have the same problem in that the PR owner would be responsible for fixing whatever existing failures are in the nightly runs in order to get the PR through.

[weshaggard]

If they change the baseline file then I would suggest we don't use the baseline file from main at all. Yes if there are existing failures in main they will need to either fix or baseline them as part of the validation.