How to specify encoding of arguments for external program call in java

问题

I want to start an external program with an argument which contains German letters, like this:

ProcessBuilder pb = new ProcessBuilder("myScript.sh", "argument_with_letters_äöü");     
Process p = pb.start();

My JVM (in my case a JBoss AS) is started with encoding ISO 8859-15. The external program 'myScript.sh' however expects UTF-8.

Is there a way to send my argument encoded as UTF-8? I serched the web, but didn't find an answer.

回答1:

Looking at the code for java.lang.ProcessImpl (the package-private class that's responsible for starting processes in the non-Windows JRE - I assume you're not on Windows given the .sh extension), process arguments are always converted to bytes using the default encoding of the running JRE:

byte[][] args = new byte[cmdarray.length-1][];
// ...
for (int i = 0; i < args.length; i++) {
    args[i] = cmdarray[i+1].getBytes();

so there's no way to do this directly. However, you may be able to work around it by using the xargs command provided you don't need to pass any standard input to myScript.sh. The purpose of xargs is to take data from its standard input and turn it into command line arguments to another executable:

// xargs -0 expects arguments on stdin separated by NUL characters
ProcessBuilder pb = new ProcessBuilder("xargs", "-0", "myScript.sh");
pb.environment().put("LANG", "de_DE.UTF-8"); // or whatever locale you require
Process p = pb.start();
OutputStream out = p.getOutputStream();
out.write("argument_with_letters_äöü".getBytes("UTF-8")); // convert to UTF-8
out.write(0); // NUL terminator
out.close();

(or if you have control of myScript.sh, then modify it to expect its file names on stdin rather than as arguments)

来源：https://stackoverflow.com/questions/26827546/how-to-specify-encoding-of-arguments-for-external-program-call-in-java

标签

java

encoding