IBM Mainframe Forum

Or an FCB with the correct characters?

Hello,

Another option might be to alter the program inserting the offending characters into the output to insert the "correct" characters.

As this is some 3rd Party software, might be tough.

dick scherrer wrote:Hello,

Another option might be to alter the program inserting the offending characters into the output to insert the "correct" characters.

As this is some 3rd Party software, might be tough.

True. I didn't pickup there was vendor provided software here. It should be possible to discuss this matter with the vendor. Can the software create Postscript or PCL data directly, or is the volume so large it really needs to go to something like a 3900 printer? Pardon my ignorance about printers, but is there something that competes with the 3900 for workstation type computers?

steve-myers wrote:

Robert Sample wrote:... you really need to contact the site support group to find out what translation tables have been set up for FTP already -- and especially if you need one added.

Ah, but the IP claims he is doing a binary transfer, which does not do character translation in FTP. Another translation issue is AFP data is a mix of EBCDIC character data and true binary data, so one does not want to translate this data unless it is clearly understood where binary ends and and character starts and for how long.

You got it!
EBCDIC is directly made on the server, AFP is a mix of EBCDIC and binary, so the transfer must be binary.

Some more explanations:

I'm just installing and configuring the software on the server.
So I'm looking for the right parameters in the server software and do not want to change something on the host.

I found something about codepage converter:
http://publib.boulder.ibm.com/infocente ... epage.html
which looks like the parameter in the software and what it should do.

Thanks to all, I think I can test tomorrow some parameters,
even if I still not know, where to find the CCSID of the host.

best regards
Mike1304

Good luck

We'll be interested to hear what happens,

d

googling for IBM EBCDIC CODEPAGES or ZOS EBCDIC CODEPAGES might come up with hints
usually, IIRC, the common codepages for the IBM mvs family were/are 037 or 500

intermediate report:

seams to be good:
The AFP-Parameter in the vendor software showed CP1148 instead of CP1141, which was a good explanation for the "}" instead of a "ü" .
But unfortunately, the change to CP1141 has no effect, maybe it's a bug in the softeware

...to be continued...

next finding:

The change to CP1141 produces the right hex-codes for the special characters.
My irritation was caused from the AFP-Browser, which showed the wrong characters for the right hex-codes.

Now, I have to test with the AFP2PDF-converter and with a real 3800 printer.

If they do not work proper, I'll wonder again if 1141 is the right codepage.