从Websphere 7升级到Websphere 8.5.5时的编码问题

Question

We recently moved an application from WAS 7.0 (on AIX) to WAS 8.5.5 (on Linux). It interfaces with a couple of applications that send data in the form of an xml

The XML is retrieved from the header using -

while ((i = request.getReader().read(buf, 0, buf.length)) != -1) {
            sb.append(buf, 0, i);
        }

However after transition, we noticed that the application was not handling special characters like è or © correctly - they are garbled.

This looks to me like an encoding issue. Can anyone point on what needs to be checked to understand the root cause?

I was reading further on this and i see that I can set the JVM arguments to

-Dclient.encoding.override=UTF-8

to always use UTF-8. Is this a good practice?

Edit :

Locale output in Linux

LANG=en_US.UTF-8
LC_CTYPE="en_US.UTF-8"
LC_NUMERIC="en_US.UTF-8"
LC_TIME="en_US.UTF-8"
LC_COLLATE="en_US.UTF-8"
LC_MONETARY="en_US.UTF-8"
LC_MESSAGES="en_US.UTF-8"
LC_PAPER="en_US.UTF-8"
LC_NAME="en_US.UTF-8"
LC_ADDRESS="en_US.UTF-8"
LC_TELEPHONE="en_US.UTF-8"
LC_MEASUREMENT="en_US.UTF-8"
LC_IDENTIFICATION="en_US.UTF-8"
LC_ALL=

Locale output on AIX
LANG=en_US
LC_COLLATE="en_US"
LC_CTYPE="en_US"
LC_MONETARY="en_US"
LC_NUMERIC="en_US"
LC_TIME="en_US"
LC_MESSAGES="en_US"
LC_ALL=

One application sends the xml as <?xml version="1.0" encoding="ISO-8859-1"?> and the other sends it as <?xml version="1.0">

After setting the above mentioned JVM setting, the <?xml version="1.0"> is treated correctly but the one with the encoding set to ISO-8859-1 is not. I am totally lost here.

Answer 1

As it seems you application is not written to use a specific encoding and therefore uses the default of your session.

Check on AIX and Linux the locale with locale . On Linux it might be something like LANG=en_US.UTF-8 .

To let your application behave on Linux the same as on AIX set on Linux the locale to the same value as on AIX.

Using unicode aware applications is not a bad idea in general. But there are exceptions where you need to stick to another encoding, eg LATIN-1 for some legacy systems. Then in your code you explicitly need to choose this encoding where it is needed.