Update MSTest.Test* packages from 3.6.4 to 3.7.0, follow best practices

[DaveTryon]

Dependabot tried to update MSTest.TestFramework from 3.6.4 to 3.7.0 in #839, and MSTest.TestAdapter in #841. These PR's both failed, as the 2 pieces need to move together for the tests to work correctly. In addition, the newer version of MSTest.TestFramework adds several new warnings, and since our builds treat warnings as errors, the build breaks.

This PR touches a lot of files, but since these packages are all test-specific, so are the code changes. It was done as a series of commits, and reviewing each commit in sequence will probably simplify the review process. Commits are as follows:

Commit	Description/Coment
1	Unchanged from #839
2	Suppress new warnings
3	Unchanged from #841 -- this allowed the tests to run
4	Added a suppression that was flagged once both packages were updated
5	MSTEST0005 We were using TestContext improperly, fix it
6	MSTEST0006 ExpectedException is no longer considered a best practice, so I changed them to ThrowsException or ThrowsExceptionAsync. Most of these changes were tool-driven, but some were manual and exposed some dead code in the unit tests
7	MSTEST0017 Assert.AreEqual wants the expected parameter first--there were many places in the code where we had the test parameter first. These changes were largely tool-driven
8	MSTEST0024 was fixed by 5, so the suppression can go away
9	MSTEST0034 This just adds a parameter to an attribute decoration
10	MSTEST0037 Assert.AreEqual, Assert.IsNull, and Assert.IsNotNull are all preferred over Assert.IsTrue becuase they are more descriptive. These changes were largely tool-driven.
11	I noticed a few more places where the parameters of Assert.AreEqual were in the wrong order, but the analyzer failed to catch the error. These changes were manual.
12	MSTEST0026 Null propagation in asserts is discouraged
13	MSTEST0032 Non-nullable types don't need to be asserted as non-null. Note that this undoes some of the changes from 12
14	JSON002 is no longer showing up as an error, so I removed the suppression
15	Remove a couple of using blocks that were accidentally added along the way
16	Restore a check was dropped in 6
17	Fix broken tests on macOS, add local JSON002 (only seems to impact Visual Studio, not command line)

Note: When running from inside Visual Studio, the net472 projects get skipped. They run correctly from the command line, so any potential regressions should get caught in the pipelines. The net80 versions of the tests work as expected in both Visual Studio and from the command line. I opened microsoft/testfx#4487 to report that issue. It was resolved as a dupe of microsoft/testfx#4426

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 0% with 128 lines in your changes missing coverage. Please review.

Project coverage is 46.50%. Comparing base (9ef64aa) to head (10b76c4).

Files with missing lines	Patch %	Lines
...bom.Api.Tests/PackageDetails/RubyGemsUtilsTests.cs	0.00%	16 Missing ⚠️
...s/Config/ConfigurationBuilderTestsForValidation.cs	0.00%	14 Missing ⚠️
...s/Config/ConfigurationBuilderTestsForGeneration.cs	0.00%	13 Missing ⚠️
...bom.Api.Tests/Executors/FileListEnumeratorTests.cs	0.00%	8 Missing ⚠️
...bom.Api.Tests/Entities/FileValidationResultTest.cs	0.00%	5 Missing ⚠️
...t.Sbom.Api.Tests/Executors/DirectoryWalkerTests.cs	0.00%	5 Missing ⚠️
...ft.Sbom.Api.Tests/Executors/PackagesWalkerTests.cs	0.00%	5 Missing ⚠️
.../Executors/ComponentToPackageInfoConverterTests.cs	0.00%	4 Missing ⚠️
...Utils/ComponentDetectionCliArgumentBuilderTests.cs	0.00%	4 Missing ⚠️
...Tests/Utils/ComponentDetectorCachedExecutorTest.cs	0.00%	4 Missing ⚠️
... and 23 more

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #848      +/-   ##
==========================================
+ Coverage   46.47%   46.50%   +0.02%     
==========================================
  Files         355      355              
  Lines       14151    14142       -9     
  Branches     1163     1161       -2     
==========================================
  Hits         6577     6577              
+ Misses       7026     7017       -9     
  Partials      548      548              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[alisonlomaka]

I saw in your comments that you used tooling to assist some of these refactors - might make an interesting short and sweet tech talk/demo for the team.

[Evangelink]

I would also strongly encourage you to enable test parallelization either from runsettings or using the folowing assembly level attribute [assembly: Parallelize(Scope = ExecutionScope.MethodLevel, Workers = 0)].

Finally, it'd be good to also consider moving to MSTest runner, which is faster, more ligtweight and the recommended way to run your tests.

[DaveTryon]

I would also strongly encourage you to enable test parallelization either from runsettings or using the folowing assembly level attribute [assembly: Parallelize(Scope = ExecutionScope.MethodLevel, Workers = 0)].

Finally, it'd be good to also consider moving to MSTest runner, which is faster, more ligtweight and the recommended way to run your tests.

@Evangelink, this entire PR was unexpected and frankly work that we hadn't scheduled. It's bloomed to 43 files, which is way too big for a single PR. Trying to add parallelization to this PR seems like it would only lead to pain and more churn. Please feel free to open an issue suggesting the parallelization feature and/or moving to MSTest runner, but I'll honestly say that it's unlikely to find much traction unless there's a compelling reason--our tests are fast enough that the inner dev loop won't be significantly improved by shaving a few seconds from the run time.