javax.tools.JavaFileManager.getJavaFileForOutput(ANNOTATION_PROCESSOR_PATH, "file", HTML ) doesn't throw expected exception UnsupportedOperationException.

As far as I understand only SOURCE and CLASS kinds are suitable for use in getJavaFileFor{Input/Output}
If I'm wrong - the spec should be clarified.

see a test
-----------
import javax.tools.*;
import java.io.*;
import java.util.Collections;

public class test11 {

    public static void main(String[] argv){
        JavaCompilerTool compiler = ToolProvider.getSystemJavaCompilerTool();
        StandardJavaFileManager mgr = compiler.getStandardFileManager( new DiagnosticCollector<JavaFileObject>() );
        System.out.println( new File( new File(".").toURI() ).getAbsolutePath() );
        JavaFileManager.Location loc = StandardLocation.ANNOTATION_PROCESSOR_PATH;
        mgr.setLocation(loc, Collections.singleton(new File( new File(".").toURI())));

        try {
            JavaFileObject f = mgr.getJavaFileForOutput( loc, "file", JavaFileObject.Kind.HTML, null );
            System.out.println( "result " + f.toUri() );
            System.out.println( "getJavaFileForOutput should throw UnsupportedOperationException - if the specified kind (HTML) is not supported" );
        } catch( Exception x ){
            x.printStackTrace(System.out);
        }
    }
}
-----------
result:
-----------
Z:\tests>z:/lnks/jdk6/bin/java.exe -cp . test11
Z:\tests\.
result file:/Z:/tests/./file.html
getJavaFileForOutput should throw UnsupportedOperationException - if the specified kind (HTML) is not supported
-----------
tests failed:
api/javax_tools/FileManager/index.html#All[getJavaFileForInput0003]
api/javax_tools/FileManager/index.html#All[getJavaFileForOutput0003]