I don’t know if my mind just fools me or this is really not working.
I need different type of Logging-classes so I created a abstract-class, the only definition that all classes will have the same is the way the writeToLog is handled:
public abstract class LoggerTemplate {
protected String filename ="log/";
protected File logfile;
protected FileWriter fw;
public void writeToLog(String message) {
if(fw != null) {
try {
message = new SimpleDateFormat("dd-MM-hh:mm").format(new Date()) + " " + message;
fw.write(message);
} catch (IOException e) {
e.printStackTrace();
}
}
}
}
The concrete sub-classes will implement rest of the logic in their constructor, ie one of them:
public class JitterBufferLogger extends LoggerTemplate {
public JitterBufferLogger() {
super();
filename += new SimpleDateFormat("yyyyddMMhhmm'.log'").format(new Date());
if(!new File("log/").exists())
new File("log").mkdir();
logfile = new File(filename);
try {
logfile.createNewFile();
} catch (IOException e) {
e.printStackTrace();
}
try {
fw = new FileWriter(logfile);
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
}
But when I debug I can see that when calling the writeToLog for a specific logger, it jumps into the LoggerTemplate method, and therefore fw and logfile are null. So it’s not working.
Isn’t it supposed to work or do I just mess something a bit up and should go into weekend 😉
It should work, it is normal, that the debugger stepped into the
LoggerTemplateclass upon entering thewriteToLog()method. What is strange that the attributes in the base class havenullvalues.I have tested your code with the following short test program:
After adding
fw.flush()to theLoggerTemplate.writeToLog()method just after thefw.write()call, it worked for me, the log file had been created and it contained the log message.Maybe the
new File("log").mkdir()or some other calls throw an exception which you cannot see, because stderr had been redirected somewhere.