(still) accented letters dropped

[Fred]

> On Wed, 2006-10-18 at 11:46 -0400, Fred wrote:
>
> > in /etc/cups/cupsd.conf, I've tried "utf-8" or "iso-8859-1" for
> > "DefaultCharset"
>
> The texttops filter in CUPS does not correctly handle the UTF-8
> character encoding.  You can try setting 'iso-8859-1' as the
> DefaultCharset, but make sure you also convert your (presumably UTF-8
> encoded) document to iso-8859-1 as well, like this:
>
> iconv -f utf-8 -t iso-8859-1 -o output.txt input.txt
>
> The Red Hat package of CUPS in Fedora Core 6 (shortly to be released)
> and in the next version of Red Hat Enterprise Linux will use a text->PS
> filter based on the paps package, which does correctly handle UTF-8 and
> uses fontconfig for rendering glyphs.  This means that documents such
> as /usr/share/doc/pango-*/HELLO.utf8 will print out correctly.
>
> Tim.

I've tried to print the HELLO.utf8 file, and it was fine.
In fact my files are neither utf-8 nor iso-8859-1, but old plain 8 bits characters (ANSI ?).
For example, the french word "déjà" is hexa coded by "64 E9 6A E0".
When converted to ps with texttops, the result is (in the ps file) "d\000j\000".
One big point I forgot to mention: there was no problem when I was in 1.1.17 version (I had to upgrade to 1.2.0 due to a bug that limit the number of pollable printers to 15) !
The only difference I see between the two versions is that the first one was install from a RPM, whereas the 2nd was compiled and installed from  sources (tarball).

Any idea ?