Broker doctype string field contains unprintable characters

We are publishing a simple broker document to the 6.1 broker that IS 6.1 is configured to use.

This broker document is extremely simple. It contains one long and one unicode string. The unicode string field’s value is set to contain an XML document which IS should process.

We can publish this broker document from Developer to the broker and confirm that the trigger invokes the appropriate service with the correct content.

However, in one installation of IS and Broker (in a remote location), when the same broker document is published from a java-based web application or from DocTraker, a string variable in the broker document contains unprintable characters by the time it reaches the invoked service. This, of course, causes the XMLStringToXMLDocument built-in service to fail with a parse error.

In addition, when Developer at the remote location is used to publish the document in question other broker client’s receive it but it is empty. IS processes this same document just fine.

What in the IS / Broker connection could be causing this field of the document to get corrupted? Other broker clients can subscribe to the same document and process it successfully. I created a java test harness from the code the web app is using and could not recreate the issue on a different installation.

Ideas?

Mark

Mark, this usually is an encoding problem. You say that this only happens on one installation (remotely), is this server running on a different locale?