At this point you need to take a step back and consider what you actually want to achieve. To help that process take a look at Unicode#Code_point_planes_and_blocks and consider that large areas of the code space are undefined and that even the areas where there are defined characters it is very likely that your system doesn't have font support for the characters anyway. Also take a look at ISO-8859-1 - if all you are doing is checking conversion from latin 1 to UTF-8 then you have fewer than 256 characters to bother about.

The other part of the problem is that Windows essentially does not do UTF-8. Internally Windows uses UTF-16LE, but most often what gets rendered is the Windows-1252 code page. Emacs seems to be doing the right thing - render a rectangle when the character is missing from whatever font is being used.