Re: XML Parser Error 12

No PC systems has CCSID's so when you transfer an IFS file til IBM I it will
always have default 819 or 1252 as CCSID even if it is a UTF-8 file - the
same
goes for FTP.

That is a setting in NetServer. You could change that to 1208, but then all files coming from 'outside' would get CCSID 1208.

Whether you should worry about characters not supported by EBCDIC depends on whether you use EBCDIC. RPG supports Unicode (UCS-2. data type C) and so does DB2. So if you define the fields in the receiving data structure as type C, you should have no problems with those characters in RPG.

Joep Beckeringh



Op 29 aug. 2013, om 17:01 heeft Henrik Rützou <hr@xxxxxxxxxxxx> het volgende geschreven:

Let me give you a little breifing on UTF-8 ...

No PC systems has CCSID's so when you transfer an IFS file til IBM I it will
always have default 819 or 1252 as CCSID even if it is a UTF-8 file - the
same
goes for FTP.

Since you have charset=UTF-8 in your XML header not all XML readers are
forgiving when it meets UTF-8 encoding if the reader thinks it meets ASCII.

If you know the file is in UTF-8 you have to make a work around to change
the file CCSID - but that may not always be enough since UTF-8 i full
UNICODE
and may contain characters that isn't allowed in SBCS EBCDIC causing the
XML
reader will crash. Typically you will see this scenario when receiving XML
from
European or DBCS countries and try to read it into a CSSID 37 (US) EBCDIC
machine.

If that's the case - it can be a tricky affair because you have to clean it
up first
either by replacing special non supported characters or by dropping them or
by defining the fields as DBCS fields where you actually just moves the
problem
to ICONV if you have to end up with an EBCDIC SBCS field.

It would be nice if you could attach the original XML file so we could look
at it