[rt-users] 3.0.9, Perl 5.8.3, and UTF-8 problem still

[Toni Mueller]

Hello,

I've got a problem finding tickets with accented characters. To test,
I've generated a ticket with a subject containing my last name,
"Müller". When searching it, I get the correct display in the web page
search mask, namely

Status = open [delete]
Subject LIKE müller [delete]
Status = new [delete]

But I don't find the ticket. This is what I find in the query log
(PostgreSQL 7.4.1 on Debian unstable for the whole thing including the
browser):

Mar  2 16:12:23 spruce postgres[4294]: [16-1] LOG:  statement: SELECT count(main.id) FROM Tickets main WHERE ((main.EffectiveId
Mar  2 16:12:23 spruce postgres[4294]: [16-2]  = main.id)) AND ((main.Status != 'deleted')) AND ((main.Type = 'ticket')) AND ( ( (main.Subject ILIKE '%m\203ller%') ) AND (
Mar  2 16:12:23 spruce postgres[4294]: [16-3]  (main.Status =
'open')OR(main.Status = 'new') ) ) 


Sorry, this is garbled. The string around ILIKE above reads like this in
Emacs hexl-mode:

00000130: 6d61 696e 2e53 7562 6a65 6374 2049 4c49  main.Subject ILI
00000140: 4b45 2027 256d c35c 3230 33c2 bc6c 6c65  KE '%m.\203..lle
00000150: 7225 2729 2029 2041 4e44 2028 0a4d 6172  r%') ) AND (.Mar

Or, with od:

1776600   e  sp   =  sp   '   t   i   c   k   e   t   '   )   )  sp   A
1776620   N   D  sp   (  sp   (  sp   (   m   a   i   n   .   S   u   b
1776640   j   e   c   t  sp   I   L   I   K   E  sp   '   %   m   C   \
1776660   2   0   3   B   <   l   l   e   r   %   '   )  sp   )  sp   A
1776700   N   D  sp   (  nl   M   a   r  sp  sp   2  sp   1   6   :   1
1776720   2   :   2   3  sp   s   p   r   u   c   e  sp   p   o   s   t


My database is configured to have a standard encoding of UNICODE.

If I search for a string with no accented characters, I find the
ticket(s) no problem.


TIA!



Best,
--Toni++