CJK codecs are incomplete.

Description

For CJK codecs (cp932 for example), the "encoding" sysmodule requires some complements and they are slightly complex.
In CPython, these complements are implemented in C.

• Modules/cjkcodecs
  • e.g.) _codecs_jp.c for japanese locale (cp932, euc-jp, ...)

Steps to Reproduce

1. Run ipy.exe without arguments.

Expected behavior:
IronPython.Runtime.Operations.PythonOps.LookupEncoding() does not throw any exceptions.
This method is called by importing "site" module.

Actual behavior:
IronPython.Runtime.Operations.PythonOps.LookupEncoding() throws following exception:
PythonOps.LookupError("unknown encoding: {0}", encoding)

Versions

IronPython 3.4.0a0 DEBUG (3.4.0.0000)
[.NETFramework,Version=v4.5 on .NET Framework 4.8.4075.0 (64-bit)] on win32

$ git rev-parse HEAD
342e49f69d9a967606d49394df51a059a55a4381

Workarounds (for developers)

The immediate issue is that make.ps1 test-all fails on CJK environment, but this can be workaround by (locally) applying patch as follows:

-                return "cp" + CultureInfo.CurrentCulture.TextInfo.ANSICodePage.ToString();
+                return "cp" + (new CultureInfo("en-US", false)).TextInfo.ANSICodePage.ToString();

Start at IronPython.Runtime.Operations.PythonOps.LookupEncoding() and reproduce the failure by running ipy.exe without arguments, then run make.ps1 test-all in a CJK environment. Compare the relevant CPython Modules/cjkcodecs files, including _codecs_jp.c, to identify the missing complements. Done means importing site and the test suite no longer fail with an unknown encoding error.