Changeset 17235 for gli

Timestamp:

2008-09-10T19:25:45+12:00 (16 years ago)

Author:

ak19

Message:

To bypass wget's unintuitive way of dealing with URLs that refer to dir-listings where the final slash is missing (and still continue to work properly with URLs like nzdl.org/niupepa which redirect to entirely different URLs), added method getRedirectURL()--along with helper functions--which is consulted by DownloadButtonListener.actionPerformed() to work out the real URL of the resource, prior to launching Wget with the URL to be retrieved.

File:

• gli/trunk/src/org/greenstone/gatherer/gui/DownloadPane.java (modified) (2 diffs)

gli/trunk/src/org/greenstone/gatherer/gui/DownloadPane.java

@@ -223 +223 @@
         String launch = launch_str.toString();      
         System.err.println("*** launch = " + launch);
-        
+
         URL launch_url = new URL(launch);
         URLConnection launch_connection = launch_url.openConnection();
@@ -495 +495 @@
     implements ActionListener {
     public void actionPerformed(ActionEvent event) {
-        
+
         if(checkURL(true) && checkProxy() == true) {
         
+        // Proxy settings are now set. Check that the url is not a redirect, else get
+        // redirect url (we do this step in order to avoid some unintuitive behaviour from wget)
+        Download current_download = (Download)download_map.get(mode);
+        Argument arg_url = current_download.getArgument("url");
+        String url_str = arg_url.getValue();
+        String redirect_url_str = getRedirectURL(url_str);
+        
+        // only update the Argument and its GUI ArgumentControl if the URL
+        // has in fact changed
+        if(!url_str.equals(redirect_url_str)) {
+            arg_url.setValue(redirect_url_str);
+            updateArgument(arg_url, redirect_url_str);
+        }
+    
         getter.newDownloadJob((Download)download_map.get(mode) ,mode,proxy_url);
         }
     }
     }
+
+    /**
+     * The Java code here will retrieve the page at the given url. If the response code is 
+     * a redirect, it will get the redirect url so that wget may be called with the proper url. 
+     * This preprocessing of the URL is necessary because:
+     * Wget does not behave the way the browser does when faced with urls of the form
+     * http://www.englishhistory.net/tudor/citizens and if that page does not exist.
+     * The directory listing with a slash at the end (http://www.englishhistory.net/tudor/citizens/)
+     * does exist, however. In order to prevent wget from assuming that the root URL
+     * to traverse is http://www.englishhistory.net/tudor/ instead of the intended
+     * http://www.englishhistory.net/tudor/citizens/, we need give wget the redirect location
+     * that's returned when we initially make a request for http://www.englishhistory.net/tudor/citizens
+     * The proper url is sent back in the Location header, allowing us to bypass wget's 
+     * unexpected behaviour.
+     * This method ensures that urls like http://www.nzdl.org/niupepa also continue to work:
+     * there is no http://www.nzdl.org/niupepa/ page, because this url actually redirects to an 
+     * entirely different URL.
+     * @return the redirect url for the given url if any redirection is involved, or the
+     * url_str.
+     */
+    private String getRedirectURL(String url_str) {
+    HttpURLConnection connection = null;
+    if(url_str.startsWith("http:")) { // only test http urls
+        try {
+        URL url = new URL(url_str); 
+        connection = (HttpURLConnection)url.openConnection(); //new HttpURLConnection(url);
+        // don't let it automatically follow redirects, since we want to
+        // find out whether we are dealing with redirects in the first place
+        connection.setInstanceFollowRedirects(false);
+        
+        // now check for whether we get a redirect response
+        // HTTP Codes 3xx are redirects, http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html
+        int responseCode = connection.getResponseCode();
+        if(responseCode >= 300 && responseCode < 400) { 
+            String responseMsg = connection.getResponseMessage();
+
+            // Get the Location header since this specifies the new location of the resource
+            String location = connection.getHeaderField("Location");
+            
+            // this becomes the url that wget should download from
+            url_str = location.trim();
+        }
+        
+        connection.disconnect();
+        } catch(Exception e) {
+        if(connection != null) {
+            connection.disconnect();
+        }
+        System.err.println("Checking redirection. Tried to connect to "
+                   + url_str + ",\nbut got exception: " + e);
+        }       
+    } 
+
+    return url_str;
+    }
+
     
+    /** For a string-based Argument whose value has changed, this method
+     * updates the GUI ArgumentControl's value correspondingly. */
+    private void updateArgument(Argument arg, String value) {
+    for(int i = 0; i < options_pane.getComponentCount(); i++) {
+        Component component = options_pane.getComponent(i);
+        if(component instanceof ArgumentControl) {
+        ArgumentControl control = (ArgumentControl)component;
+        if(control.getArgument() == arg) {
+            control.setValue(value);
+            control.repaint();
+        }
+        }
+    }
+    }
 
     private boolean checkURL(boolean checkRequired){