I have an application that reads in an XML document, inserts some dynamic content, and outputs it as a PDF using the UJAC utilities.
This works perfectly on a Windows server, but on a Linux machine it replaces all the French characters with symbols.
I get the same issue when reading text files from the server into strings, but I solved that by specifying the charset when creating the string. I know that there must be a better solution and that some setting on the server or in Tomcat/Catalina must be causing my problem, but I have no idea where to look!
I know enough to check the locale of the server - I get en_US.UTF-8 which seems right. Similarly, calling System.getProperty("file.encoding") returns UTF-8.
I know the problem occurs before UJAC, as the echoed content of the XML has the symbols rather than French characters.
I didn't write the code that actually gets the file contents. It uses a RequestDispatcher to read in a JSP that contains the XML with the JSP tags. Changing that is probably not an option.