[sync] Re: syncml and p910i

[Tim Nijenbrink]

On Thursday 17 March 2005 17:14, Chuck Hagenbuch wrote:
> Quoting Tim Nijenbrink <horde at magni.nl>:
> > cvs is up-to-date, turning debugging on gives me the following sql
> > -queries right before the error:
>
> Please try to filter stuff like this...
>
> > Mar 17 12:51:53 HORDE [debug] [nag] SQL Query by DataTree_sql::_load():
> > SELECT datatree_id, datatree_name, datatree_parents, datatree_order
> > FROM horde_datatree WHERE group_uid = 'horde.history' AND () [on line
> > 113 of "/usr/share/pear/Horde/DataTree/sql.php"]
> > Mar 17 12:51:53 HORDE [debug] [nag] SQL Query by
> > DataTree_sql::getByAttributes():  [on line 1093 of
> > "/usr/share/pear/Horde/DataTree/sql.php"]
> > ---------------
>
> Here's the bad query. Do you have Nag configured? Can you use it
> individually?
>
> -chuck
>
> --
> "But she goes not abroad in search of monsters to destroy." - John
> Quincy Adams

Setting aside my previous answer, I looked into it a little bit anyway. 
Seems that with kolab drivers (but SQL datatree) there is no history entry 
added when a new calendar entry is entered, and with sql driver for kronolith 
and datatree there is but datatree_data field of the added row is always 
NULL. 
Neither of these seem correct to me.  

$changes in Horde_SyncML_Sync_TwoWaySync::handlesync is an empty array with 
each rpc.php response, no matter the timestamp used with the listBy call.

tim