Add -logger option, to specify name of appender to use

(console or DRFA currently).

Use inline layout when setting up loop directory logging,
versus relying on "console" appender that might not have
been set up, depending on which log4j.properties file
got used.

src/main/java/bixo/tools/SimpleCrawlTool.java

@@ -9,6 +9,7 @@
 import org.apache.hadoop.mapred.JobConf;
 import org.apache.log4j.FileAppender;
 import org.apache.log4j.Logger;
+import org.apache.log4j.PatternLayout;
 import org.kohsuke.args4j.CmdLineException;
 import org.kohsuke.args4j.CmdLineParser;
 
@@ -100,14 +101,13 @@ private static void setLoopLoggerFile(String outputDirName, int loopNumber) {
         Logger rootLogger = Logger.getRootLogger();
 
         String filename = String.format("%s/%d-SiteCrawlTool.log", outputDirName, loopNumber);
-        FileAppender appender = (FileAppender) rootLogger.getAppender("loop-logger");
+        FileAppender appender = (FileAppender)rootLogger.getAppender("loop-logger");
         if (appender == null) {
             appender = new FileAppender();
             appender.setName("loop-logger");
-            appender.setLayout(rootLogger.getAppender("console").getLayout());
-
-            // We have to do this before calling addAppender, as otherwise Log4J
-            // warns us.
+            appender.setLayout(new PatternLayout("%d{yy/MM/dd HH:mm:ss} %p %c{2}:%L - %m%n"));
+
+            // We have to do this before calling addAppender, as otherwise Log4J warns us.
             appender.setFile(filename);
             appender.activateOptions();
             rootLogger.addAppender(appender);
@@ -147,6 +147,11 @@ public static void main(String[] args) {
             System.setProperty("bixo.root.level", "INFO");
         }
 
+        if (options.getLoggingAppender() != null) {
+            // Set console vs. DRFA vs. something else
+            System.setProperty("bixo.appender", options.getLoggingAppender());
+        }
+
         try {
             JobConf conf = new JobConf();
             Path outputPath = new Path(outputDirName);

src/main/java/bixo/tools/SimpleCrawlToolOptions.java

@@ -10,6 +10,7 @@ public class SimpleCrawlToolOptions {
     private static final int DEFAULT_NUM_LOOPS = 1;
 
     private boolean _debugLogging = false;
+    private String _loggingAppender = null;
 
     private String _outputDir;
     private String _agentName;
@@ -19,17 +20,22 @@ public class SimpleCrawlToolOptions {
     private int _maxThreads = DEFAULT_MAX_THREADS;
     private int _numLoops = DEFAULT_NUM_LOOPS;
 
-
+    
     @Option(name = "-domain", usage = "domain to crawl (e.g. cnn.com)", required = true)
     public void setDomain(String domain) {
         _domain = domain;
     }
 
     @Option(name = "-d", usage = "debug logging", required = false)
     public void setDebugLogging(boolean debugLogging) {
-        this._debugLogging = debugLogging;
+        _debugLogging = debugLogging;
     }
 
+    @Option(name = "-logger", usage = "set logging appender (console, DRFA)", required = false)
+    public void setLoggingAppender(String loggingAppender) {
+        _loggingAppender = loggingAppender;
+    }
+
     @Option(name = "-outputdir", usage = "output directory", required = true)
     public void setOutputDir(String outputDir) {
         _outputDir = outputDir;
@@ -83,6 +89,10 @@ public boolean isDebugLogging() {
         return _debugLogging;
     }
 
+    public String getLoggingAppender() {
+        return _loggingAppender;
+    }
+
     @Override
     public String toString() {
         return ReflectionToStringBuilder.toString(this, ToStringStyle.MULTI_LINE_STYLE);