Fix Encoding.Default regressions

**.NET Core Version:**
any .NET Core version

**Have you experienced this same bug with .NET Framework?:**
no, regression against Desktop Framework

**Problem description:**

.NET Core changed the behavior of `Encoding.Default` to return UTF8 encoding, while Desktop (from which WinForms is derived) returned the Windows ANSI encoding (which is also used for interop). Any usage of `Encoding.Default` in WinForms is now very likely to be incorrect and must be reviewed and updated.

See issue #3032 and PR #3100, #3141, #3210 for the previously done corrections, but [more remain](https://github.com/dotnet/winforms/search?q=%22Encoding.Default%22&unscoped_q=%22Encoding.Default%22) to be fixed.

The correct pattern for porting Desktop Framework usage of `Encoding.Default` is now to use `CodePagesEncodingProvider.Instance.GetEncoding(0) ?? Encoding.UTF8` instead, which retrieves the ANSI encoding Windows uses for interop. (The codepage based encoding is null if Windows was configured to be using UTF8 instead of a normal codepage, see discussion in #3032.)

For call sites which aren't related to interop there needs to be discussion and case-by-case decision of whether to replace `Encoding.Default` or whether to use UTF8 going forward. They could still be updated to refer to the UTF8 encoding explicitly for clarity.

**Expected behavior:**
WinForms is updated to respect the changed behavior of `Encoding.Default`

**Minimal repro:**
n/a (issue raised due to manual inspection of source code, analysis of scenarios should happen as part of the issue, if required)