X-Git-Url: https://git.pterodactylus.net/?a=blobdiff_plain;f=src%2Fnet%2Fpterodactylus%2Farachne%2Fcore%2FCore.java;h=53dedb80c0e87e3d2c24454227e5b279a1ea571e;hb=b3f6d8f1de06593b1266d0432f77f684dd79bb4f;hp=9e3f1a55a1d0d1143b792474429eef63013b7a3e;hpb=c85307a32e0957c15233b4d9ead5f58d8d167244;p=arachne.git

diff --git a/src/net/pterodactylus/arachne/core/Core.java b/src/net/pterodactylus/arachne/core/Core.java
index 9e3f1a5..53dedb8 100644
--- a/src/net/pterodactylus/arachne/core/Core.java
+++ b/src/net/pterodactylus/arachne/core/Core.java
@@ -7,12 +7,15 @@ import java.net.MalformedURLException;
 import java.net.URL;
 import java.security.NoSuchAlgorithmException;
 import java.util.ArrayList;
+import java.util.HashSet;
 import java.util.List;
+import java.util.Set;
 import java.util.concurrent.Executor;
 import java.util.concurrent.Executors;
 import java.util.logging.Level;
 import java.util.logging.Logger;
 
+import net.pterodactylus.arachne.parser.ParserFactory;
 import de.ina.util.service.AbstractService;
 import de.ina.util.thread.DumpingThreadFactory;
 import de.ina.util.validation.Validation;
@@ -41,12 +44,18 @@ public class Core extends AbstractService {
 	// INTERNAL MEMBERS
 	//
 
+	/** The parser factory. */
+	private ParserFactory parserFactory = new ParserFactory();
+
 	/** Thread pool for the URL fetches. */
-	private Executor urlFetcherExecutor = Executors.newFixedThreadPool(1, new DumpingThreadFactory("URLFetcher-"));
+	private Executor urlFetcherExecutor = Executors.newSingleThreadExecutor(new DumpingThreadFactory("URLFetcher-"));
 
 	/** The current list of URLs to crawl. */
 	private final List<Page> pages = new ArrayList<Page>();
 
+	/** Already crawled pages. */
+	private final Set<Page> crawledPages = new HashSet<Page>();
+
 	//
 	// ACCESSORS
 	//
@@ -73,6 +82,7 @@ public class Core extends AbstractService {
 	 */
 	public void addPage(URL url) {
 		Validation.begin().isNotNull("url", url).check().isEqual("url.getHost()", url.getHost(), (Object) nodeHost).isEqual("url.getPort()", url.getPort(), nodePort).check();
+		addPage(Page.fromURL(url));
 	}
 
 	/**
@@ -96,8 +106,10 @@ public class Core extends AbstractService {
 	 */
 	public void addPage(Page page) {
 		Validation.begin().isNotNull("page", page).check();
-		pages.add(page);
-		notifySyncObject();
+		if (!crawledPages.contains(page) && !pages.contains(page)) {
+			pages.add(page);
+			notifySyncObject();
+		}
 	}
 
 	//
@@ -105,15 +117,15 @@ public class Core extends AbstractService {
 	//
 
 	/**
-	 * {@inheritdoc}
+	 * {@inheritDoc}
 	 *
 	 * @see de.ina.util.service.AbstractService#serviceRun()
 	 */
 	@Override
-	@SuppressWarnings("null")
 	protected void serviceRun() {
 		while (!shouldStop()) {
-			Page nextPage = null;
+			final Page nextPage;
+			Page page = null;
 			synchronized (syncObject) {
 				while (!shouldStop() && pages.isEmpty()) {
 					try {
@@ -123,22 +135,33 @@ public class Core extends AbstractService {
 					}
 				}
 				if (!shouldStop()) {
-					nextPage = pages.remove(0);
+					page = pages.remove(0);
 				}
 			}
 			if (shouldStop()) {
 				break;
 			}
+			nextPage = page;
 			URL nextURL = nextPage.toURL(nodeHost, nodePort);
 			if (nextURL == null) {
 				logger.log(Level.INFO, "Skipping “" + nextPage + "”.");
 				continue;
 			}
-			URLFetcher urlFetcher;
+			final URLFetcher urlFetcher;
 			try {
 				logger.log(Level.INFO, "Fetching “" + nextURL + "”...");
-				urlFetcher = new URLFetcher(nextURL);
-				urlFetcherExecutor.execute(urlFetcher);
+				urlFetcher = new URLFetcher(parserFactory, nextURL);
+				urlFetcherExecutor.execute(new Runnable() {
+
+					@SuppressWarnings("synthetic-access")
+					public void run() {
+						urlFetcher.run();
+						crawledPages.add(nextPage);
+						for (Page page : urlFetcher.getCollectedPages()) {
+							addPage(page);
+						}
+					}
+				});
 			} catch (NoSuchAlgorithmException nsae1) {
 				logger.log(Level.SEVERE, "Could not get “SHA-256” message digest!", nsae1);
 			}