In #309 @rmisev noted that the extra flags were not set in the "valid domain" algorithm's call to ToUnicode. I pushed a fixup that adds them. But maybe it would be better to deduplicate the calls to ToUnicode, similar to what is already done for the calls to ToASCII.
This is not obviously straightforward since I am not sure who uses "domain to Unicode", and I'm not sure why it doesn't return failure on errors.
In #309 @rmisev noted that the extra flags were not set in the "valid domain" algorithm's call to ToUnicode. I pushed a fixup that adds them. But maybe it would be better to deduplicate the calls to ToUnicode, similar to what is already done for the calls to ToASCII.
This is not obviously straightforward since I am not sure who uses "domain to Unicode", and I'm not sure why it doesn't return failure on errors.