For classes you do not control the source code for, you need to implement
an AppendConverter class, and register it with the AppenderHelper if you
want to have efficient appending of the string representation of objects
in that class. This is most easily demonstrated with an example. The following would
be an AppendConverter implementation for the java.util.Vector (not actually
needed as this is already handled by the AppenderHelper).
import java.util.Vector;
import jack.basics.AppenderHelper;
public class VectorAppender implements AppendConverter
{
private static final char[] COMMA_SPACE = {',',' '};
static {AppenderHelper.putInAppendMap(new VectorAppender(),"java.util.Vector");}
public void appendObj(AppenderHelper h, StringBuffer s, Object o, int depth)
{
if (depth <= 0)
AppenderHelper.appendDepth0To(s,o);
else
{
depth--;
int size = o.size();
s.append('[');
if (size != 0)
{
h.append(s, o.elementAt(0), depth);
for (int i = 1; i < size; i++)
{
s.append(COMMA_SPACE);
h.append(s, o.elementAt(i), depth);
}
}
s.append(']');
}
}
}
Note that unfortunately the AppenderHelper relies on the class of objects
passed to determine the converter object to use, and this does not carry
over to subclasses of that class if the object is passed in as type Object.
Objects of classes that are not registered
have their append call defaulted to the default StringBuffer append, so the
string representation will always be correct, but not necessarily the most
efficient append. For example, if there is a subclass of java.util.Vector
that you want to use the above example for, you should register that separately. |
