Preserve xml file encoding to prevent non-ascii chars from disappearing

[carlossanlop]

For example, System/Math.xml and MathF.xml have several unicode characters that get lost after the tool runs.

[danmoseley]

Previously the tool loaded in 1252
https://github.com/carlossanlop/DocsPortingTool/blob/bcbb8d36dc3c4e264b2d37b063cdb737ba0e1796/Libraries/Docs/DocsCommentsContainer.cs#L179-L182
so I changed the save code to save in 1252 based on that comment
https://github.com/carlossanlop/DocsPortingTool/blob/bcbb8d36dc3c4e264b2d37b063cdb737ba0e1796/Libraries/Docs/DocsCommentsContainer.cs#L50-L51

I'm now unsure what is correct. @gewarren is asking. (UTF-8 everywhere would be wonderful)

[carlossanlop]

Thanks for that. I chose that encoding because 99% of the xml files had that particular encoding.

Ideally (and I tried to do this in many ways) we would extract the original encoding of the file, and use that when loading the file, then when saving the file. Unfortunately, it wasn't as straighforward as I hoped: whenever a non-ascii character showed up, the XML APIs would either throw an unexpected exception, or the characters would get unexpectedly replaced with the square character □.