On Fri, Jan 28, 2011 at 11:40:42PM +0100, Reindl Harald wrote:
> what i really not understand in this post is why the OP
> doens not change all his charset/collation to UTF8
> to avoid mixing?
The input may not be valid UTF8. Not all octet strings are valid
UTF8. Someone posted a solution using
attribute = _latin1 '%s'
that's probably the most robust, without changing the schema to
model the columns as binary data not text.
The highly hypothetical discussion of UTF8 in SMTP is largely
off-topic here, and mostly pointless.
--
Viktor.
