Wrong encoding for special characters (Swedish language)

[Pooriafd]

Describe the bug
In the latest release (20.0.0.1), our application started to handle Swedish characters in wrong way

Steps to Reproduce
Print something in swedish like äppel or översätt, and it will be question marks instead of right characters

Expected behavior
to be written correctly in responses

Diagnostic information:

• OpenLiberty Version: 20.0.0.1
• Java Version: 11.0.6+10-LTS

Additional context
The problem got fixed by setting LANG=en_US.UTF-8 in linux, however we still have the same problem in windows which seems not to solve the problem by LANG env variable

[chooty]

This seems to happen as a result of the change in JsonBProvider in commit 548bcf2 where getBytes is called without a charset specified.

[andymc12]

I'm taking a look - it seems like we should be setting the charset based on the Accept-Charset header instead of just assuming the default charset.

[chooty]

According to RFC8259,
JSON text exchanged between systems that are not part of a closed ecosystem MUST be encoded using UTF-8

So to me it sounds like always using UTF-8 could be ok, or at least the default if no charset is specified.

[c-koell]

We are facing the same problems after upgrading. German umlauts are not readable

[andymc12]

I've just integrated a change that uses UTF-8 as the default, but allows users to override the default charset for the JSON-B entity provider by setting com.ibm.ws.jaxrs.jsonbprovider.defaultCharset=<differentCharset>. The change will also attempt to respect the Accept-Charset HTTP header from the client.