UA-FI and UNOC

[Hans Klok]

Hi,

I am working on an application using the Maxware UA-FI product to send and
receive EDI messages. The Syntax identifier used here is UNOC (extended
ASCII set with accents and umlauts). However, incoming messages are
translated by the software to a small ASCII set (only upper/lowercase), and
this means that my special characters are lost somewhere.

Since I send outgoing messages as "BILATERALLY DEFINED" (and not as IA5),
the other party can receive my messages in a correct manner. Their response
messages are recognized as IA5, but without using the UNOC.

Anyone has experience with this and might know what to do?

Hans.

[Tomas]

Is the sending party using P35-protocol?
P35 = A standard for sending EDI messages over X.400 (also called Pedi)

"Hans Klok" <ha...@nospam.com> wrote in message news:<3f2509ca$0$137$edd6...@news.versatel.net>...

[Hans Klok]

No, they are using IA5. Pedi gives big troubles. The problem is that IA5
uses 7bits, and a mapping will translate all extended characters to their
normal equivalent. I found a solution: defining the National Character Set
for the user to a set not recognized by UAFI (in the call CREATEUSER). In
that case, no mapping is performed at all and the message comes through the
way I want. A bit tricky, but it does the job. Any better solutions are
welcome.

Hans

"Tomas" <tngo...@sverige.nu> schreef in bericht
news:fa50dd8b.03080...@posting.google.com...

[Torsten Kirschner]

"Hans Klok" <ha...@nospam.com> skrev i melding
news:3f2509ca$0$137$edd6...@news.versatel.net...

EDIFACT UNOC is based on ISO 8859-1, i.e. everything but the non-readable
control characters.

We've been doing exactly what You're after for several years. Later today
I'll post our setup. From what I recall, we use attachement type
BILATERALLY_DEFINED with both P2 and P22 without any problems.

Since You mention converters: I am under the impression that X.400 as such
is and always has been "8 bit clean". Any conversion takes place in either
the MTA or the UA - if they are configured to do so.
Our MTAs are ISOCOR ISOPLEX running in 84/88 compatibility mode.
All conversion is turned off.
I am not 100% sure whether I changed or set any conversion routines or
"national language settings" when configuring UA-FI. After all, UA-FI is
or at least used to be a Norwegian product, so everything worked out of the
box.

Which software is it that You suspect to convert incoming messages? The original
sender, Your/some other MTA or Your UA-FI?

[Hans]

I know that UA-FI performs the conversion, the logfile shows the correct
messages. But UA-FI creates a textfile with the message, and my characters
are gone.

Hans.

"Torsten Kirschner" <torsten....@sandbox.no> schreef in bericht
news:3f31a945$1...@news.broadpark.no...