Remove XBAP-related dead code from Application/BrowserInteropHelper, finish cleanup

[h3xds1nz]

Description

This continues the work started in #9855, removing rest of the internal properties in BrowserInteropHelper and thanks to this we can remove dead code in other places of the codebase (Application, AppSecurityManager, etc.) I've separated the two as those deletions might be a bit harder to grasp, while the first one is easier to understand imho.

• SiteOfOriginContainer.BrowserSource is never set, thus it is always NULL.
  • This means the public property Uri in BrowserInteropHelper in always NULL.
  • GetReferer method in BindUriHelper depends on this value, if BrowserSource is NULL, it returns NULL.
  • So the value in httpRequest.Referer of WpfWebRequestHelper would end up being NULL anyways, no-op / dead code.
• Application.MimeType is never set, thus it is always MimeType.Unknown.
  • This means IsViewer of BrowserInteropHelper is always false.
  • That also means that IsInitialViewerNavigation is always false and its underlying value does not matter at all for any operations. So we can remove any setters; and then we can remove it from both conditions; which makes it unused.
• Application.ServiceProvider is never set (classes doing it don't exist anymore), so it is always NULL.
• Classes calling the GetService also don't exist anymore, so the whole method does not matter.
• Any methods removed in AppSecurityManager just call themselves, besides that we were still calling ClearSecurityManager from Application but it is never initialized anymore so there's nothing to clear.
• As a result, SecurityMgrSite is never initialized anymore, so we can remove it as well.

Customer Impact

Decreased size of assemblies, tiny bit of perf, cleaner codebase for developers.

Regression

No.

Testing

Local build, app run.

Risk

Low, the code is a dead code or it doesn't do anything.

Microsoft Reviewers: Open in CodeFlow

[h3xds1nz]

Looks like one of the changes in #9220 was here. Rebased and resolved merge conflicts.

[h3xds1nz]

Resolved merge conflicts 01

[h3xds1nz]

Fix merge conflicts after #9521

[h3xds1nz]

Resolved merge conflicts after #9855

@harshit7962 Can we go with this one and #9943 next?

[harshit7962]

@h3xds1nz, We are done with the test pass of #9943 and will prioritize it. I am yet to review this PR and will include it in the next test run.

[h3xds1nz]

@harshit7962 Awesome, let me know at which part you'd like me to resolve the conflicts.

I feel its counter-productive to do it now should a style-PR get merged meanwhile as that will mess up the resolutions.

[harshit7962]

Sure, will let you know when the PR is ready to be merged. It would be better to resolve the conflicts at that time.

[harshit7962]

@h3xds1nz, we are done with the test pass and its all green. We are ready to take this change in.

[h3xds1nz]

@h3xds1nz, we are done with the test pass and its all green. We are ready to take this change in.

@harshit7962 Great, should be merged/rebased properly I hope.
I can't compile right now so I have to verify with CI I did pick the correct usings.

[harshit7962]

@h3xds1nz, thanks for the contribution.

We are working on getting in the PRs that have passed the tests. Please let us know if any of these you would like us to prioritize.

Also, feel free to mention any PR that you would like us to include in the next test run.

[h3xds1nz]

@harshit7962 Great, glad to see this one in.

The untested ones in order of importance (perf gains):

1. Remove allocations on all base converters, improve TokenizerHelper #9364
2. Update PropertyValues in TriggerBase/FrameworkElementFactory via reference #10003
3. Remove double lookups on SizeLimitedCache<K, V>, improve performance #9949
4. Replace ArrayList in DataFormats with List<DataFormat> and refactor the class #9199
5. Simplify CriticalCopyPixels in BitmapSource by removing duplicate type checks #9395
6. Optimize FontSourceCollection creation from a filesystem directory, reduce allocs #9844 (this is a big gain but the customer impact is lower as its more of an edge case)

Do note there are two unit test PRs that could be merged imho:

1. Add test coverage for public API surface of TextDecorationCollectionConverter #9895
2. Add test coverage for public API surface of MouseActionConverter #9891 (this one includes a fix that slipped in through the WPF tests)

Regarding the ones that have passed tests so I can do further work/would be good to merge:

1. Optimize ResourceContainer/AssemblyLoadEventHandler methods, remove allocs #9822
   1.1) Optimize retrieval of simple name from Assembly.FullName, fix up faulty methods #9739
2. Optimize parsing of color strings into KnownColor enum, remove allocations #9669
3. Replace occurrences of params T[] with params ReadOnlySpan<T> where possible #9481 (@dipeshmsft was reviewing it)
4. Modify AvTrace call chain to use params ReadOnlySpan<object> instead of an array #9468 (this provides a regression fix)