I see this in our log files a lot:
Oct 13 19:05:09 roundcube1 roundcube: DB Error: MDB2 Error: unknown error Query: _doQuery: [Error message: Could not execute statement] [Last executed query: INSERT INTO contacts (user_id, changed, del, "name", "email", "firstname", "surname") VALUES (122994, now(), 0, 'zzzzzzzzz', 'xxxxxxxx@yyyyyyyy.com', '', 'Mo(nique)<A0>Be cking')] [Native message: ERROR: invalid byte sequence for encoding "UTF8": 0xa0 HINT: This error can also happen if the byte sequence does not match the encoding expe cted by the server, which is controlled by "client_encoding".] in /data/WWW/release-0.5/program/include/rcube_mdb2.php on line 642
We use postgres, and the DB is definitely set to UTF-8. We're running a 0.5.x SVN, so this may have been fixed later. Or it may be some local problem. Before I start digging into this, anyone have an idea what this could be?
Name | Owner | Encoding
------------------+------------------+----------- roundcubemail | roundcube | UTF8
Regards,
Cor
List info: http://lists.roundcube.net/dev/ BT/aba52c80
On 13.10.2011 20:56, Cor Bosman wrote:
[Native message: ERROR: invalid byte sequence for encoding "UTF8": 0xa0
\xa0 is used as non-breaking space, but for utf8 it should be \xc2a0. We need some input validation.
On Fri, Oct 14, 2011 at 08:22, A.L.E.C alec@alec.pl wrote:
On 13.10.2011 20:56, Cor Bosman wrote:
[Native message: ERROR: invalid byte sequence for encoding "UTF8": 0xa0
\xa0 is used as non-breaking space, but for utf8 it should be \xc2a0. We need some input validation.
get_input_value() would be the right place to add such. Currently we rely on the fact, that the browser sends data with the same charset as the Roundcube pages are loaded.
~Thomas _______________________________________________ List info: http://lists.roundcube.net/dev/ BT/aba52c80