[HtmlUnit] CVS Commit: src/xdocs: Encode path part of urls containing illegal

Log Message:
-----------
Encode path part of urls containing illegal characters (fix for bug 1437068)

Modified Files:
--------------
    htmlunit/src/test/java/com/gargoylesoftware/htmlunit:
        WebClientTest.java
        
(http://cvs.sourceforge.net/viewcvs.py/htmlunit/htmlunit/src/test/java/com/gargoylesoftware/htmlunit/WebClientTest.java)
    htmlunit/src/java/com/gargoylesoftware/htmlunit:
        WebClient.java
        
(http://cvs.sourceforge.net/viewcvs.py/htmlunit/htmlunit/src/java/com/gargoylesoftware/htmlunit/WebClient.java)
    htmlunit/src/xdocs:
        changes.xml
        
(http://cvs.sourceforge.net/viewcvs.py/htmlunit/htmlunit/src/xdocs/changes.xml)

Revision Data
-------------
Index: WebClient.java
===================================================================
RCS file: 
/cvsroot/htmlunit/htmlunit/src/java/com/gargoylesoftware/htmlunit/WebClient.java,v
retrieving revision 1.118
retrieving revision 1.119
diff -Lsrc/java/com/gargoylesoftware/htmlunit/WebClient.java 
-Lsrc/java/com/gargoylesoftware/htmlunit/WebClient.java -u -d -r1.118 -r1.119
--- src/java/com/gargoylesoftware/htmlunit/WebClient.java
+++ src/java/com/gargoylesoftware/htmlunit/WebClient.java
@@ -1392,49 +1392,66 @@
     }
 
     /**
-     * Encodes illegal parameter in query string (if any) as done by browsers.
-     * Example: change "http://first?a=b c" to "http://first?a=b%20c";  
+     * Encodes illegal parameter in path or query string (if any) as done by 
browsers.
+     * Example: changes "http://first?a=b c" to "http://first?a=b%20c";  
      * @param url the url to encode
      * @return the provided url if no change needed, the fixed url else
      * @throws MalformedURLException if the new URL could note be instantiated
      * @throws URIException if the default protocol charset is not supported
      */
     protected URL encodeUrl(final URL url) throws MalformedURLException, 
URIException {
-        // just look at urls with query string (better test?)
-        final String str = url.toExternalForm();
-        final int queryStart = url.toExternalForm().indexOf('?');
-        if (queryStart != -1) {
-            // extract query string: browsers seem not to encode everything, 
for instance not "#"
-            final String query;
-            final int anchorStart = str.indexOf('#');
-            if (anchorStart < queryStart) {
-                query = str.substring(queryStart);
+        final String path = url.getPath();
+        final String fixedPath = encode(path, URI.allowed_abs_path);
+        final String query = url.getQuery();
+        final String fixedQuery = encode(query, URI.allowed_query);
+
+        if (!StringUtils.equals(path, fixedPath) || !StringUtils.equals(query, 
fixedQuery)) {
+            final StringBuffer newUrl = new StringBuffer();
+            newUrl.append(url.getProtocol());
+            newUrl.append("://");
+            newUrl.append(url.getHost());
+            if (url.getPort() != -1) {
+                newUrl.append(":");
+                newUrl.append(url.getPort());
             }
-            else {
-                query = str.substring(queryStart, anchorStart);
+            newUrl.append(fixedPath);
+            if (url.getUserInfo() != null) {
+                newUrl.append(url.getUserInfo());
             }
-
-            // url may be partially encoded like "http://first?a=b%20c&d=e f"
-//          // don't re-encode the %'s from already encoded items
-            final BitSet partiallyEncodedQuery = new BitSet(256);
-            partiallyEncodedQuery.set('%');
-            partiallyEncodedQuery.or(URI.allowed_query);
-            final String fixedQuery = URIUtil.encode(query, 
partiallyEncodedQuery);
-            if (query.equals(fixedQuery)) {
-                return url;
+            if (fixedQuery != null) {
+                newUrl.append("?");
+                newUrl.append(fixedQuery);
             }
-            else {
-                final StringBuffer newUrl = new StringBuffer(str);
-                newUrl.replace(queryStart, queryStart + query.length(), 
fixedQuery);
-                return new URL(newUrl.toString());
+            if (url.getRef() != null) {
+                newUrl.append("#");
+                newUrl.append(url.getRef());
             }
+
+            return new URL(newUrl.toString());
         }
         else {
             return url;
         }
+    }
 
+    /**
+     * Encodes unallowed characters in a string
+     * @param str the string to encode
+     * @param allowed the allowed characters
+     * @return the encoded string
+     * @throws URIException if encoding fails
+     */
+    private String encode(final String str, final BitSet allowed) throws 
URIException {
+        if (str == null) {
+            return null;
+        }
+        final BitSet bits = new BitSet(str.length());
+        bits.set('%');
+        bits.or(allowed);
+        return URIUtil.encode(str, bits);
     }
 
+
     /**
      * Remove the focus to the specified component.  This will trigger any 
relevant javascript
      * event handlers.
Index: WebClientTest.java
===================================================================
RCS file: 
/cvsroot/htmlunit/htmlunit/src/test/java/com/gargoylesoftware/htmlunit/WebClientTest.java,v
retrieving revision 1.67
retrieving revision 1.68
diff -Lsrc/test/java/com/gargoylesoftware/htmlunit/WebClientTest.java 
-Lsrc/test/java/com/gargoylesoftware/htmlunit/WebClientTest.java -u -d -r1.67 
-r1.68
--- src/test/java/com/gargoylesoftware/htmlunit/WebClientTest.java
+++ src/test/java/com/gargoylesoftware/htmlunit/WebClientTest.java
@@ -626,10 +626,10 @@
     }
 
     /**
-     * Test that the query string is encoded to be valid.
+     * Test that the path and query string are encoded to be valid.
      * @throws Exception If something goes wrong.
      */
-    public void testLoadPage_EncodeQueryString() throws Exception {
+    public void testLoadPage_EncodeRequest() throws Exception {
         final String htmlContent
             = "<html><head><title>foo</title></head><body>"
             + "</body></html>";
@@ -642,34 +642,31 @@
 
         // with query string not encoded
         HtmlPage page = (HtmlPage) client.getPage(new URL("http://first?a=b 
c&d=" + ((char) 0xE9) + ((char) 0xE8)));
-        assertEquals(
-            "http://first?a=b%20c&d=%C3%A9%C3%A8";,
-            page.getWebResponse().getUrl());
-
+        assertEquals("http://first?a=b%20c&d=%C3%A9%C3%A8";, 
page.getWebResponse().getUrl());
 
         // with query string already encoded
         page = (HtmlPage) client.getPage(new 
URL("http://first?a=b%20c&d=%C3%A9%C3%A8";));
-        assertEquals(
-            "http://first?a=b%20c&d=%C3%A9%C3%A8";,
-            page.getWebResponse().getUrl().toExternalForm());
+        assertEquals("http://first?a=b%20c&d=%C3%A9%C3%A8";, 
page.getWebResponse().getUrl());
 
         // with query string partially encoded
         page = (HtmlPage) client.getPage(new URL("http://first?a=b%20c&d=e 
f"));
-        assertEquals(
-            "http://first?a=b%20c&d=e%20f";,
-            page.getWebResponse().getUrl().toExternalForm());
+        assertEquals("http://first?a=b%20c&d=e%20f";, 
page.getWebResponse().getUrl());
 
         // with anchor
         page = (HtmlPage) client.getPage(new URL("http://first?a=b 
c#myAnchor"));
-        assertEquals(
-            "http://first?a=b%20c#myAnchor";,
-            page.getWebResponse().getUrl().toExternalForm());
+        assertEquals("http://first?a=b%20c#myAnchor";, 
page.getWebResponse().getUrl());
 
         // with query string containing encoded "&", "=", "+", ",", and "$"
         page = (HtmlPage) client.getPage(new 
URL("http://first?a=%26%3D%20%2C%24";));
-        assertEquals(
-            "http://first?a=%26%3D%20%2C%24";,
-            page.getWebResponse().getUrl().toExternalForm());
+        assertEquals("http://first?a=%26%3D%20%2C%24";, 
page.getWebResponse().getUrl());
+
+        // with character to encode in path
+        page = (HtmlPage) client.getPage(new URL("http://first/page 1.html"));
+        assertEquals("http://first/page%201.html";, 
page.getWebResponse().getUrl());
+
+        // with character to encode in path
+        page = (HtmlPage) client.getPage(new URL("http://first/page 1.html"));
+        assertEquals("http://first/page%201.html";, 
page.getWebResponse().getUrl());
     }
 
     /**
Index: changes.xml
===================================================================
RCS file: /cvsroot/htmlunit/htmlunit/src/xdocs/changes.xml,v
retrieving revision 1.492
retrieving revision 1.493
diff -Lsrc/xdocs/changes.xml -Lsrc/xdocs/changes.xml -u -d -r1.492 -r1.493
--- src/xdocs/changes.xml
+++ src/xdocs/changes.xml
@@ -7,6 +7,9 @@
 
     <body>
         <release version="next">
+            <action type="update" dev="mguillem" id="1437068">
+                               Encode path part of urls containing illegal 
characters.
+                       </action>
             <action type="update" dev="mguillem" id="1436102" due-to="Brad 
Murray">
                                Added support for firing onkeydown events.
                        </action>


-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642