Wietse:
> It's not possible to convert from characterset X to UTF-8, without
> knowing what X is.
>
> With smtputf8_enable=yes, Postfix >= 3.0 will log UTF-8 text as
> UTF-8 text. It will not convert other text to UTF-8, because there
> is no characterset information.
King Cao:
> Maybe BASE64 or HEX encode is another option.
BASE64/HEX forms are not helpful if you want to enter a malformed
name into Postfix a configuration file or access rule. To match
a malformed name, you need to specify a regular expression that
contains "." at the locations where Postfix logs "?".
Adding string sanitizer support that increases the length of text
would require invasive changes to Postfix source code such as
converting C strings into managed VSTRINGs, or making other changes
in the way that memory is managed, The cost of those changes does
not justify the benefit.
Wietse
