FMSLogo Code

Explicitly write characters in UTF-8, instead of leaving it to the whim of the CRT.

Carry the character encoding of loadstream along with loadstream.
This makes it possible for TO, which continue reading from loadstream, to read the procedure definitions correctly.

Add more explanation to a comment

Apply the BOM sniffing logic to fileopen(), which is used by File -> Open.

Try to fix the problems with committing a UTF-16LE test case.

Update the metadata for the unicode tests to work on Windows.

Get LOAD to work for UTF16LE-encoded files.

Extend reader() and rd_getc() to be parameterized by wxBOM.
This brings the code base closer to being able to read both
legacy ANSI coded files (ones without a BOM) and Unicode
files (ones with a BOM).

Parameterize the function which reads Unicode from a FILE* by character encoding.
This revision does not include any functional changes, but it's a step on the way to
supporting BOM-sniffing to determine character encoding when reading Logo code
from files.

Fix build breaks when debugutils.cpp is compiled with Unicode.