• unicode conversions

    From Andreas Leitgeb@21:1/5 to All on Fri May 18 15:45:16 2018
    I have a strange problem with German umlaut characters.
    Every umlaut (ä,ö,ü) gets mapped to a '@' character before
    the method on the server (with TAO) sees it.


    I've got a Java(jacorb) client and a C++(tao) server.
    The server runs in an 8-bit locale (iso-8859-1).

    At some point, the client requests some string data,
    and the result string contains German umlauts.

    According to wireshark, the umlauts are correctly
    converted to unicode over the net, and the client
    displays them properly in its GUI.

    Then the string gets sent back as argument to another
    method call. Wireshark still sees correct unicodes in
    the GIOP packet of that call.

    But within the implementation of the method, the umlauts
    have each been replaced by a "@".

    Maybe someone could hint me towards what makes TAO convert
    unicode to "@" even though the umlaut is perfectly representable
    in the target locale? Maybe TAO has a different idea about what
    is server's locale. How can I specify the target 8-bit charset
    explicitly?

    Turning the server to unicode is (unfortunately) not an option.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)