Problems with special characters!

[Former Member]

Hi experts,

Scenario: WebSphere <> XI <> IS-U

I have a problem with xml encoding: when WebSphere sends a message in utf-8 with some characters like (ã, ç, é, etc.) to XI, it sends it back to WebSphere with some changes on these characters, something like (Ã, ##, etc). In XI monitoring I can see these characters like they are, but when XI sends a message back to WebSphere the “special” characters change.

What happens?

Does someone know if XI processes xml messages with a different encoding as utf-8? Is it possible to send a message with “iso-8859-1” encoding???

Thanks in advance.

Cheers,

Ricardo.

[Former Member · Accepted solution]

Hi,

even we had the same problem in file to idoc scenario.In our scenario these characters like(ã, ç, é) were sent from our flat file which when posted in idoc was changed to a#...

so inorder to avoid it,in the communication channel we gave the

file type as text and encoding type as UTF-8

and it worked.The change in those characters didnt occur when those characters were posted to idoc.

so when u send back the payload to websphere from XI,in that communication channel u try giving the encoding type as UTF-8.hope it solves ur problem.

i dont think iso 8859-1 encoding technique wont solve ur problem.we tried our scenario with that encoding scheme i didnt work.

regards

jithesh

[MichalKrawczyk]

hi Ricardo,

>>>Is it possible to send a message with “iso-8859-1”

probably yes but you have to change it

in the adapter (in standard or in the adapter module)

of in a java proxy if you use one

Regards,

michal

[Former Member]

Hi again guys,

I’ve found the problem!

WebSphere is in “iso-8859-1” and XI is in “utf-8” and when a portal sends a message via form/html or file upload, it sends using “iso-8859-1” encoding. XI receives the message in “iso-8859-1” and during the mapping the new version of the message mapped to be sent is changed to “utf-8 by XI”… and WebSphere expects “iso-8859-1” encoding and not utf-8!!!

So, the WebSphere team has found a way to turn around this issue. With a routine they can convert the utf-8 encoding to iso-8859-1 encoding.

It works this way … I think problems are solved!

For XML keywords like (ex. & > <) I’m using placeholders (ex. &amp; &gt; &lt;) and with tests that I have made, I almost affirm that XI doesn’t have any problems to handle this kind of issues.

Thanks a lot for all

Cheers,

Ricardo.

[Former Member]

Hi Ricardo,

YES! It is possible to use ISO-8859-1 encoding in Xi. Infact, in the case of the characters that u have mentioned here (like ã,é)... try seeing if ISO-8859-16 works.

Actually, UTF-8 encoding is supposed to work just fine with these characters, but I do not know what is causing the # symbols.

Please try it out and let me know.

cheers,

Prashanth