GitLab

Now correctly rejects UTF-8 encoded Unicode surrogates and does not support
5 and 6 byte UTF-8 encodings.

For reference:
https://github.com/openssl/openssl/commit/ba64e5a92a6f009e311ad1c3565817820a1632a4

Replaced unicode_is_zerowidth() with unicode_width(), in preparation for
"fullwidth" char support.
Added UNICODE_UNDEFINED definition (0x0000)
(UNICODE_NULL is already defined, at least in MSVC).

utf8.c: In function 'utf8_getc':
utf8.c:327:7: warning: this 'else' clause does not guard... [-Wmisleading-indentation]
     } else
       ^~~~
utf8.c:329:2: note: ...this statement, but the latter is misleadingly indented as if it is guarded by the 'else'
  if(val != NULL) {
  ^~

New function (derived from sbbs_t::utf8_to_cp437()): unicode_to_cp437()
New utf8 functions: utf8_replace_chars(), utf8_str_is_valid().
utf8_getc() enhancement: val arg may be NULL (for length/validation uses).

UNICODE -> CP437 conversions (no string/memory compares needed), so
cp437_utf8_tbl.* is being replaced by cp437_unicode_tbl.*.

Added utf8 module which includes:
- utf8_normalize_str() - normalize US-ASCIIZ string, from mailsrvr.c
- utf8_putc() - encoder / length calculator
- utf8_getc() - decoder

The latter 2 were swiped from OpenSSL's a_utf8.c.