My pgsql database was set to use UTF8. dbmail: 2.1.3 When tried to move a message in Outlook Express from one mailbox(non-dbmail) into dbmail, through dbmail-IMAP, I got database error when it tries to insert the body into dbmail_messageblks:
ERROR: invalid UTF-8 byte sequence detected near byte 0xfe This absolutely becase the data fed to database was not valid UTF8 sequence. By checking the email header content-type, it appears to use iso-8859-9: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; CHARSET=iso-8859-9; FORMAT=flowed Content-Transfer-Encoding: 8BIT Content-ID: <[EMAIL PROTECTED]> So my guess about how IMAPd working is it took whatever encoding it saw when startup and never change it. Is it true? And if so, is it true to other daemons? I think the daemons should be able to detect what kind of encoding the source is and set it accordingly. The same sql statement failed when I run from psql with default encoding. By switching encoding to LATIN1, I was able to insret the data into the database. Do you think this a bug? Thanks, Kevin
