Name: poR10007 Date: 04/07/2004



The getMBeansFromURL() methods of the class javax.management.loading.MLet
use the system default encoding when reading an m-let text file. This may
cause a compatibility problem between systems with different default
charsets.

The InputStreamReader which is used to read m-let files in the MLetParser class
is created using a default charset. According to java.nio.charset.Charset spec:

  "Every instance of the Java virtual machine has a default charset,
   which may or may not be one of the standard charsets. The default
   charset is determined during virtual-machine startup and typically
   depends upon the locale and charset being used by the underlying
   operating system."

Thus, if one system uses, for example, Cp037 as the default charset, and
it tries to load an m-let file from another system which uses ASCII default
charset, then such operation will fail.

The m-let service should use the charset defined in the spec. See also
the corresponding spec bug 5028083 "JMX spec should specify character
encoding for m-let text file".

======================================================================