RE: Brain cramp: reading an IFS file from program running on PC

But that doesn't quite make sense. Converting to Unicode should generate
a stream of characters, that's what "Unicode" means in Java. And for
that you would use a Reader, not an InputStream.

If this InputStream were going to convert from EBCDIC, it would have to
convert to Unicode characters ***in some encoding***, and that would
produce a stream of bytes. I haven't gone to the API documentation to
see if it's intended to do anything like that, or if so what that
unspecified encoding might be, but as I say I don't think that you can
just expect EBCDIC-to-Unicode conversion.

PC2

-----Original Message-----
From: java400-l-bounces@xxxxxxxxxxxx
[mailto:java400-l-bounces@xxxxxxxxxxxx] On Behalf Of David Gibbs
Sent: May 4, 2009 12:33
To: Java Programming on and around the iSeries / AS400
Subject: Re: Brain cramp: reading an IFS file from program running on PC

Gary L Peskin wrote:

Oh, I think I see what you're talking about now. Properties.load()
assumes that its argument is a stream using ISO 8859-1 encoding which
is a kind of ASCII. Your input stream ("is") is presenting characters

based on the native byte representation on the 400 since InputStreams
don't take the CCSID into account. I think you can create your own
input stream class to accomplish what you want. I haven't tested this

(or even compiled it) so take that into account. Also, you'd need to
make adjustments where the CCSID was not a single byte character set.

This is what I assumed the IFSTextFileInputStream class would do.
Reading from a text file in Java should convert from EBCDIC to Unicode
automatically.

Otherwise it seems to be a waste of code.

I'm 95% sure it DOES do that ... but I just can't figure out what I'm
doing wrong.