From a637923e6ea2bd6d53180b7e843b02a191b6d6e5 Mon Sep 17 00:00:00 2001
From: =?utf8?q?David=20=E2=80=98Bombe=E2=80=99=20Roden?=
 <bombe@pterodactylus.net>
Date: Tue, 10 Mar 2009 23:11:13 +0100
Subject: [PATCH] Use a default parser factory in the core and hand it in to
 every URL fetcher.

---
 src/net/pterodactylus/arachne/core/Core.java       |  6 +++++-
 src/net/pterodactylus/arachne/core/URLFetcher.java | 18 +++++++++++++-----
 2 files changed, 18 insertions(+), 6 deletions(-)

diff --git a/src/net/pterodactylus/arachne/core/Core.java b/src/net/pterodactylus/arachne/core/Core.java
index 9e3f1a5..9efc960 100644
--- a/src/net/pterodactylus/arachne/core/Core.java
+++ b/src/net/pterodactylus/arachne/core/Core.java
@@ -13,6 +13,7 @@ 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;
@@ -41,6 +42,9 @@ 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-"));
 
@@ -137,7 +141,7 @@ public class Core extends AbstractService {
 			URLFetcher urlFetcher;
 			try {
 				logger.log(Level.INFO, "Fetching “" + nextURL + "”...");
-				urlFetcher = new URLFetcher(nextURL);
+				urlFetcher = new URLFetcher(parserFactory, nextURL);
 				urlFetcherExecutor.execute(urlFetcher);
 			} catch (NoSuchAlgorithmException nsae1) {
 				logger.log(Level.SEVERE, "Could not get “SHA-256” message digest!", nsae1);
diff --git a/src/net/pterodactylus/arachne/core/URLFetcher.java b/src/net/pterodactylus/arachne/core/URLFetcher.java
index d2ce831..cdccd2b 100644
--- a/src/net/pterodactylus/arachne/core/URLFetcher.java
+++ b/src/net/pterodactylus/arachne/core/URLFetcher.java
@@ -12,7 +12,8 @@ import java.util.List;
 import java.util.logging.Level;
 import java.util.logging.Logger;
 
-import net.pterodactylus.arachne.parser.HtmlEditorKitParser;
+import net.pterodactylus.arachne.parser.Parser;
+import net.pterodactylus.arachne.parser.ParserFactory;
 import net.pterodactylus.arachne.parser.ParserListener;
 import de.ina.util.io.MessageDigestInputStream;
 import de.ina.util.validation.Validation;
@@ -28,6 +29,9 @@ class URLFetcher implements Runnable, ParserListener {
 	/** The logger. */
 	private static final Logger logger = Logger.getLogger(URLFetcher.class.getName());
 
+	/** The parser factory. */
+	private final ParserFactory parserFactory;
+
 	/** The URL to fetch. */
 	private final URL url;
 
@@ -46,14 +50,18 @@ class URLFetcher implements Runnable, ParserListener {
 	/**
 	 * Creates a new fetcher for the given URL.
 	 *
+	 * @param parserFactory
+	 *            The parser factory that is used to create content-type
+	 *            specific parsers
 	 * @param url
 	 *            The URL to fetch
 	 * @throws NoSuchAlgorithmException
 	 *             if no {@link MessageDigest} instance with an
 	 *             <code>SHA-256</code> algorithm can be created
 	 */
-	public URLFetcher(URL url) throws NoSuchAlgorithmException {
-		Validation.begin().isNotNull("url", url).check();
+	public URLFetcher(ParserFactory parserFactory, URL url) throws NoSuchAlgorithmException {
+		Validation.begin().isNotNull("parserFactory", parserFactory).isNotNull("url", url).check();
+		this.parserFactory = parserFactory;
 		this.url = url;
 		messageDigest = MessageDigest.getInstance("SHA-256");
 	}
@@ -112,8 +120,8 @@ class URLFetcher implements Runnable, ParserListener {
 			logger.log(Level.INFO, "Type is “" + contentType + "”, length is " + contentLength + ".");
 			urlInputStream = urlConnection.getInputStream();
 			hashInputStream = new MessageDigestInputStream(urlInputStream, messageDigest);
-			HtmlEditorKitParser htmlEditorKitParser = new HtmlEditorKitParser();
-			htmlEditorKitParser.parse(this, hashInputStream, "UTF-8");
+			Parser parser = parserFactory.getParser(contentType);
+			parser.parse(this, hashInputStream, "UTF-8");
 			hash = messageDigest.digest();
 		} catch (IOException ioe1) {
 			logger.log(Level.WARNING, "Could not fetch “" + url + "”.", ioe1);
-- 
2.7.4