archive-access-cvs Mailing List for Web Archive Access Utilities (Page 36)

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

Revision: 2930
          http://archive-access.svn.sourceforge.net/archive-access/?rev=2930&view=rev
Author:   bradtofel
Date:     2009-11-12 22:22:32 +0000 (Thu, 12 Nov 2009)

Log Message:
-----------
BUGFIX: now includes original named link (#name fragment at end of input URL) in resolved url

Modified Paths:
--------------
    trunk/archive-access/projects/wayback/wayback-core/src/main/java/org/archive/wayback/util/htmllex/ParseContext.java
    trunk/archive-access/projects/wayback/wayback-core/src/test/java/org/archive/wayback/util/htmllex/ParseContextTest.java

Modified: trunk/archive-access/projects/wayback/wayback-core/src/main/java/org/archive/wayback/util/htmllex/ParseContext.java
===================================================================

--- trunk/archive-access/projects/wayback/wayback-core/src/main/java/org/archive/wayback/util/htmllex/ParseContext.java	2009-11-11 00:25:13 UTC (rev 2929)
+++ trunk/archive-access/projects/wayback/wayback-core/src/main/java/org/archive/wayback/util/htmllex/ParseContext.java	2009-11-12 22:22:32 UTC (rev 2930)
@@ -24,7 +24,7 @@
  */
 package org.archive.wayback.util.htmllex;
 
-import java.net.MalformedURLException;
+import java.net.URISyntaxException;
 import java.net.URL;
 import java.util.HashMap;
 
@@ -54,15 +54,31 @@
 	private boolean inScriptText = false;
 	private HashMap<String,String> data = null;
 
+	/**
+	 * constructor
+	 */
 	public ParseContext() {
 		data = new HashMap<String, String>();
 	}
+	/**
+	 * Stores arbitrary key value pairs in this ParseContext
+	 * @param key for storage
+	 * @param value for storage
+	 */
 	public void putData(String key, String value) {
 		data.put(key, value);
 	}
+	/**
+	 * Retrieves previously stored data for key key from this ParseContext
+	 * @param key under which value was stored
+	 * @return previously stored value for key or null, if nothing was stored
+	 */
 	public String getData(String key) {
 		return data.get(key);
 	}
+	/**
+	 * @param url against which relative URLs should be resolved for this parse
+	 */
 	public void setBaseUrl(URL url) {
 		try {
 			baseUrl = UURIFactory.getInstance(url.toExternalForm());
@@ -70,23 +86,37 @@
 			e.printStackTrace();
 		}
 	}
-	public String resolve(String url) throws MalformedURLException {
+	/**
+	 * @param url which should be resolved against the baseUrl for this 
+	 * ParseContext.
+	 * @return absolute form of url, resolved against baseUrl if relative.
+	 * @throws URISyntaxException if the input URL is malformed
+	 */
+	public String resolve(String url) throws URISyntaxException {
+		int hashIdx = url.indexOf('#');
+		String frag = "";
+		if(hashIdx != -1) {
+			frag = url.substring(hashIdx);
+			url = url.substring(0,hashIdx);
+		}
 		try {
-			return baseUrl.resolve(url).toString();
+			return baseUrl.resolve(url).toString() + frag;
 		} catch (URIException e) {
 			e.printStackTrace();
 		}
 		return url;
-//		URL tmp = new URL(baseUrl,url);
-//		return tmp.toString();
 	}	
+	/**
+	 * @param url which should be resolved.
+	 * @return absolute form of input url, or url itself if javascript:
+	 */
 	public String contextualizeUrl(String url) {
 	    if(url.startsWith("javascript:")) {
 	    	return url;
 	    }
 		try {
 			return resolve(url);
-		} catch (MalformedURLException e) {
+		} catch (URISyntaxException e) {
 			e.printStackTrace();
 			return url;
 		}

Modified: trunk/archive-access/projects/wayback/wayback-core/src/test/java/org/archive/wayback/util/htmllex/ParseContextTest.java
===================================================================
--- trunk/archive-access/projects/wayback/wayback-core/src/test/java/org/archive/wayback/util/htmllex/ParseContextTest.java	2009-11-11 00:25:13 UTC (rev 2929)
+++ trunk/archive-access/projects/wayback/wayback-core/src/test/java/org/archive/wayback/util/htmllex/ParseContextTest.java	2009-11-12 22:22:32 UTC (rev 2930)
@@ -25,13 +25,9 @@
 
 package org.archive.wayback.util.htmllex;
 
-import java.net.MalformedURLException;
 import java.net.URI;
 import java.net.URL;
 
-import org.archive.net.UURI;
-import org.archive.net.UURIFactory;
-
 import junit.framework.TestCase;
 
 /**
@@ -46,6 +42,14 @@
 	public void testContextualizeUrl() {
 		ParseContext pc = new ParseContext();
 		try {
+			
+			URI tmp = new URI("http://base.com/foo.html#REF");
+			String ref = tmp.getFragment();
+			assertEquals("REF",ref);
+			tmp = new URI("http://base.com/foo.html");
+			assertNull(tmp.getFragment());
+
+			
 			pc.setBaseUrl(new URL("http://base.com/"));
 			assertEquals("http://base.com/images.gif",
 					pc.contextualizeUrl("/images.gif"));
@@ -57,6 +61,12 @@
 					pc.contextualizeUrl("/image/1s.gif"));
 			assertEquals("http://base.com/image/1s.gif",
 					pc.contextualizeUrl("../../image/1s.gif"));
+			assertEquals("http://base.com/image/1s.gif",
+					pc.contextualizeUrl("/../../image/1s.gif"));
+			assertEquals("http://base.com/image/1.html#REF",
+					pc.contextualizeUrl("/../../image/1.html#REF"));
+			assertEquals("http://base.com/image/1.html#REF FOO",
+					pc.contextualizeUrl("/../../image/1.html#REF FOO"));
 		} catch (Exception e) {
 			// TODO Auto-generated catch block
 			e.printStackTrace();


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.




2005	Jan	Feb	Mar	Apr	May	Jun	Jul (1)	Aug (10)	Sep (36)	Oct (339)	Nov (103)	Dec (152)
2006	Jan (141)	Feb (102)	Mar (125)	Apr (203)	May (57)	Jun (30)	Jul (139)	Aug (46)	Sep (64)	Oct (105)	Nov (34)	Dec (162)
2007	Jan (81)	Feb (57)	Mar (141)	Apr (72)	May (9)	Jun (1)	Jul (144)	Aug (88)	Sep (40)	Oct (43)	Nov (34)	Dec (20)
2008	Jan (44)	Feb (45)	Mar (16)	Apr (36)	May (8)	Jun (77)	Jul (177)	Aug (66)	Sep (8)	Oct (33)	Nov (13)	Dec (37)
2009	Jan (2)	Feb (5)	Mar (8)	Apr	May (36)	Jun (19)	Jul (46)	Aug (8)	Sep (1)	Oct (66)	Nov (61)	Dec (10)
2010	Jan (13)	Feb (16)	Mar (38)	Apr (76)	May (47)	Jun (32)	Jul (35)	Aug (45)	Sep (20)	Oct (61)	Nov (24)	Dec (16)
2011	Jan (22)	Feb (34)	Mar (11)	Apr (8)	May (24)	Jun (23)	Jul (11)	Aug (42)	Sep (81)	Oct (48)	Nov (21)	Dec (20)
2012	Jan (30)	Feb (25)	Mar (4)	Apr (6)	May (1)	Jun (5)	Jul (5)	Aug (8)	Sep (6)	Oct (6)	Nov	Dec

archive-access-cvs Mailing List for Web Archive Access Utilities (Page 36)

archive-access-cvs — CVS commits