No subject

    <dtml-call "RESPONSE.setHeader('Content-Type','text/html; charset=3Dutf=
-8')">

into standard_header_html and

    REQUEST.RESPONSE.setHeader('Content-Type','text/html; charset=3Dutf-8')=


in any python script which forced at least most strings inside the scripts
to be displayed correctly.  Strangely enough I have to put strings like
Title if I want to use it inside the code very strangely.  For instance if
I want to get the Umlaut '=FC' inside the title I have to write it as
'=C3=BC' which is what you get if you try

   $ echo '=FC' | recode l1..u8
   =C3=BC

If I use a real '=FC' it is displayed as '?' on the rendered page.  But thi=
s
problem can somehow circumvented by the described workaround.

The hardest part are the strings coming from the database.  For historical
reasons I'm using zope-psycopgda on one machine which somehow works - perha=
ps
because the opder adapter has an extra property "Backend encoding:".  I set=
ted
this to UTF-8 and so database strings are OK.

But on the more important live server I'm using zope-psycopgda2 which is
lacking this property (at least it is not displayed in the ZMI.  If I tried=

to replace zope-psycopgda2 by zope-psycopgda Zope seems to have a serios
problem because I can not open any Z Psycopda Database Connection any more
(which meand my old connection object vanished from ZMI and the selection b=
ox
to add new objects is missing this entry).  So there seems to be a conflict=

when trying to switch back from zope-psycopgda2 to zope-psycopgda.

Now I have really no idea how to convince Zope to full UTF-8 compatibility.=


Any hint will be warmly welcome

       Andreas.


[1] http://lists.alioth.debian.org/pipermail/pkg-zope-developers/2007-Febru=
ary/002948.html
--=20
http://fam-tille.de

--8323328-941574151-1225284583=:8921--