Hi Paul, this multiple cenversations decrease the oerformance of body search again, because at the meoment the search process has to load every messagepart of every message the performance is not really good. I'm thinking a little bit about a universal solution, but I have now idea up to now. (for me with postgresql I can add a new messagetext table, fill them with a trigger an d search with simple selects on this table, but that is no universal solution for postgresql, mysql and sqlite)
do you have think about a solution where you save the message twice if a mail is inserted into the system? One time in the existing format and additional in another table with converted (UTF8?) content that you can use for search? Jochen > The problem is ü can be encoded in utf-8 (like I did just now) or in > iso-8859-XX. Matching during body searches should convert the body and > search key to utf-8 before searching... > > Jochen Schroer wrote: > >> Hi Paul, >> >> fine, your fix work :-) >> Multilingual code is not so easy, I know. >> >> Kind regards, >> >> Jochen >> >>> Please try again. It should work now, but please note that ü appears to >>> be encoded differently in different charsets (iso-8859-15 versus >>> iso-8859-1). This is (at the moment) not dealt with properly. >>> >>> >> > > -- ECS-Solution GmbH Raiffeisenstr. 26 67817 Imsbach Germany Web: http://www.ecs-solution.de Tel: +49 (6302) 609 71-0 Fax: +49 (6302) 609 71-1
