Changing url-check so that it allows different schemes but still requires the same domain

(cherry picked from commit d4ff78918dd9317a5686b9675cdade3cb6c2007e)
Signed-off-by: Dan Fabulich <dan.fabulich@redfin.com>

src/main/java/com/redfin/sitemapgenerator/AbstractSitemapGeneratorOptions.java

@@ -7,7 +7,7 @@ import java.net.URL;
 // It makes sense, I swear! http://madbean.com/2004/mb2004-3/
 abstract class AbstractSitemapGeneratorOptions<THIS extends AbstractSitemapGeneratorOptions<THIS>> {
 	File baseDir;
-	String baseUrl;
+	URL baseUrl;
 	String fileNamePrefix = "sitemap";
 	boolean allowMultipleSitemaps = true;
 	W3CDateFormat dateFormat;
@@ -19,7 +19,7 @@ abstract class AbstractSitemapGeneratorOptions<THIS extends AbstractSitemapGener
 		if (baseDir == null) throw new NullPointerException("baseDir may not be null");
 		if (baseUrl == null) throw new NullPointerException("baseUrl may not be null");
 		this.baseDir = baseDir;
-		this.baseUrl = baseUrl.toString();
+		this.baseUrl = baseUrl;
 	}
 	
 	/** The prefix of the name of the sitemaps we'll create; by default this is "sitemap" */
@@ -62,6 +62,7 @@ abstract class AbstractSitemapGeneratorOptions<THIS extends AbstractSitemapGener
 		this.gzip = gzip;
 		return getThis();
 	}
+	
 	@SuppressWarnings("unchecked")
 	THIS getThis() {
 		return (THIS)this;

src/main/java/com/redfin/sitemapgenerator/SitemapGenerator.java

@@ -17,7 +17,7 @@ abstract class SitemapGenerator<U extends ISitemapUrl, THIS extends SitemapGener
 	/** 50000 URLs per sitemap maximum */
 	public static final int MAX_URLS_PER_SITEMAP = 50000;
 	
-	private final String baseUrl;
+	private final URL baseUrl;
 	private final File baseDir;
 	private final String fileNamePrefix;
 	private final String fileNameSuffix;
@@ -56,7 +56,7 @@ abstract class SitemapGenerator<U extends ISitemapUrl, THIS extends SitemapGener
 	 */
 	public THIS addUrl(U url) {
 		if (finished) throw new RuntimeException("Sitemap already printed; you must create a new generator to make more sitemaps"); 
-		UrlUtils.checkUrl(url.getUrl().toString(), baseUrl);
+		UrlUtils.checkUrl(url.getUrl(), baseUrl);
 		if (urls.size() == maxUrls) {
 			if (!allowMultipleSitemaps) throw new RuntimeException("More than " + maxUrls + " urls, but allowMultipleSitemaps is false.  Enable allowMultipleSitemaps to split the sitemap into multiple files with a sitemap index.");
 			if (mapCount == 0) mapCount++;
@@ -169,12 +169,8 @@ abstract class SitemapGenerator<U extends ISitemapUrl, THIS extends SitemapGener
 	public void writeSitemapsWithIndex() {
 		if (!finished) throw new RuntimeException("Sitemaps not generated yet; call write() first");
 		File outFile = new File(baseDir, "sitemap_index.xml");
-		SitemapIndexGenerator sig;
-		try {
-			sig = new SitemapIndexGenerator.Options(baseUrl, outFile).dateFormat(dateFormat).autoValidate(autoValidate).build();
-		} catch (MalformedURLException e) {
-			throw new RuntimeException("bug", e);
-		}
+		SitemapIndexGenerator sig;		
+		sig = new SitemapIndexGenerator.Options(baseUrl, outFile).dateFormat(dateFormat).autoValidate(autoValidate).build();		
 		sig.addUrls(fileNamePrefix, fileNameSuffix, mapCount).write();
 	}
 	

src/main/java/com/redfin/sitemapgenerator/SitemapIndexGenerator.java

@@ -17,8 +17,7 @@ import org.xml.sax.SAXException;
  *
  */
 public class SitemapIndexGenerator {
-	private final URL baseUrl;
-	private final String baseUrlString;
+	private final URL baseUrl;	
 	private final File outFile;
 	private final ArrayList<SitemapIndexUrl> urls = new ArrayList<SitemapIndexUrl>();
 	private final int maxUrls;
@@ -115,8 +114,7 @@ public class SitemapIndexGenerator {
 	}
 	
 	private SitemapIndexGenerator(Options options) {
-		this.baseUrl = options.baseUrl;
-		this.baseUrlString = baseUrl.toString();
+		this.baseUrl = options.baseUrl;		
 		this.outFile = options.outFile;
 		this.maxUrls = options.maxUrls;
 		W3CDateFormat dateFormat = options.dateFormat;
@@ -128,7 +126,7 @@ public class SitemapIndexGenerator {
 	
 	/** Adds a single sitemap to the index */
 	public SitemapIndexGenerator addUrl(SitemapIndexUrl url) { 
-		UrlUtils.checkUrl(url.url.toString(), baseUrlString);
+		UrlUtils.checkUrl(url.url, baseUrl);
 		if (urls.size() >= maxUrls) {
 			throw new RuntimeException("More than " + maxUrls + " urls");
 		}

src/main/java/com/redfin/sitemapgenerator/UrlUtils.java

@@ -1,13 +1,19 @@
 package com.redfin.sitemapgenerator;
 
+import java.net.URL;
 import java.util.HashMap;
 
 class UrlUtils {
 
-	static void checkUrl(String url, String baseUrl) {
+	static void checkUrl(URL url, URL baseUrl) {
 		// Is there a better test to use here?
-		if (!url.startsWith(baseUrl)) {
-			throw new RuntimeException("Url " + url + " doesn't start with base URL " + baseUrl);
+		
+		if (baseUrl.getHost() == null) {
+			throw new RuntimeException("base URL is null");
+		}
+		
+		if (!baseUrl.getHost().equalsIgnoreCase(url.getHost())) {
+			throw new RuntimeException("Domain of URL " + url + " doesn't match base URL " + baseUrl);
 		}
 	}
 

src/test/java/com/redfin/sitemapgenerator/SitemapGeneratorTest.java

@@ -161,6 +161,21 @@ public class SitemapGeneratorTest extends TestCase {
 			fail("wrong domain allowed to be added");
 		} catch (RuntimeException e) {}
 	}
+
+	public void testSameDomainDifferentSchemeOK() throws Exception {
+		wsg = new WebSitemapGenerator("http://www.example.com", dir);
+			
+		wsg.addUrl("https://www.example.com/index.html");
+		
+		String expected = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n" + 
+				"<urlset xmlns=\"http://www.sitemaps.org/schemas/sitemap/0.9\" >\n" + 
+				"  <url>\n" + 
+				"    <loc>https://www.example.com/index.html</loc>\n" + 
+				"  </url>\n" + 
+				"</urlset>";
+		String sitemap = writeSingleSiteMap(wsg);
+		assertEquals(expected, sitemap);		
+	}
 	
 	public void testDoubleWrite() throws Exception {
 		testSimpleUrl();