Support dictionaries saved in UTF-8 without BOM

[psoukie]

Update reading and writing dictionaries to allow users to work with dictionaries saved as UTF-8 without BOM.

This would prevent issues like #149.

It could be done by using FileOpen() after figuring out the encoding. Technically, the detection without BOM cannot be 100% but the method of comparing text length would work in all practical situations. See the discussion:

https://www.autohotkey.com/board/topic/95945-find-encoding-format-of-file/page-2#entry604453