win32: UTF8_OUTPUT: reject overlong encoding

Overlong UTF8 encoding means that a codepoint is encoded using more
bytes than necessary, and it's disallowed.

Until now we didn't reject it, e.g. 2-bytes sequence of 0xc0 0x9b
was incorrectly decoded as codepoint 0x1b (ESC), but it's overlong
because 1-byte 0x1b is enough for this codepoint.

Now we reject such sequences and print '?' (CONFIG_SUBST_WCHAR).

Additionally, we now also reject 4-bytes sequences which end up above
the maximum valid codepoint value (0x10ffff). No such issue with 1/2/3
bytes UTF-8 because technically only 4 can encode a too big value.

0 files changed, 0 insertions, 0 deletions