Changeset 12504

Timestamp:

02/23/10 14:34:45 (12 years ago)

Author:

Mark Evenson

Message:

Implement HTTP HEAD Last-Modified checking for ZipCache? objects.

Since it seems that no Sun-derived JVM ever invalidates its cache of
URLConnection objects, we have to check for modification "manually".
This implementation is "better than nothing", but not expected to be
especially robust. Most notably, this implementation does not attempt
to use http proxies if they have been specified to the JVM by the
'http.proxy' system property.

Location:

trunk/abcl/src/org/armedbear/lisp

Files:

• ZipCache.java (modified) (5 diffs)
• util/HttpHead.java (added)

trunk/abcl/src/org/armedbear/lisp/ZipCache.java

@@ -33 +33 @@
 package org.armedbear.lisp;
 
-
+import org.armedbear.lisp.util.HttpHead;
 import static org.armedbear.lisp.Lisp.*;
 
 import java.io.File;
-import java.io.InputStream;
 import java.io.IOException;
 import java.net.JarURLConnection;
@@ -43 +42 @@
 import java.net.URL;
 import java.net.URLConnection;
-import java.util.Enumeration;
+import java.text.ParseException;
+import java.text.SimpleDateFormat;
+import java.util.Date;
 import java.util.HashMap;
 import java.util.zip.ZipException;
 import java.util.zip.ZipFile;
-import java.util.zip.ZipEntry;
 
 /**
@@ -68 +68 @@
     // out, not allowing ZipFile.close() to succeed until that count
     // has been reduced to 1 or the finalizer is executing.
-    // Unfortunately the relatively simple strategy of extended
+    // Unfortunately the relatively simple strategy of extending
     // ZipFile via a CachedZipFile does not work because there is not
     // a null arg constructor for ZipFile.
@@ -101 +101 @@
         return get(Pathname.makeURL(arg));
     }
+
+    static final SimpleDateFormat RFC_1123
+        = new SimpleDateFormat("EEE, dd MMM yyyy HH:mm:ss zzz");
 
     synchronized public static ZipFile get(final URL url) {
@@ -146 +149 @@
                     }
                 }
-            } else {
+            } else if (url.getProtocol().equals("http")) {
                 // Unfortunately, the Apple JDK under OS X doesn't do
                 // HTTP HEAD requests, instead refetching the entire
-                // resource, so the following code is a waste.  I assume
-                // this is the case in all Sun-dervied JVMs. We'll have
-                // to implement a custom HTTP lastModified checker.
-
-                // URLConnection connection;
-                // try {
-                //     connection = url.openConnection();
-                // } catch (IOException ex) {
-                //     Debug.trace("Failed to open "
-                //                 + "'" + url + "'");
-                //     return null;
-                // }
-                // long current = connection.getLastModified();
-                // if (current > entry.lastModified) {
-                //     try {
-                //         entry.file.close();
-                //     } catch (IOException ex) {}
-                //     entry = fetchURL(url, false);
-                // }
+                // resource, and I assume this is the case in all
+                // Sun-derived JVMs.  So, we use a custom HEAD
+                // implementation only looking for Last-Modified
+                // headers, which if we don't find, we give up and
+                // refetch the resource.n
+                String dateString = HttpHead.get(url, "Last-Modified");
+                Date date = null;
+                try {
+                    date = RFC_1123.parse(dateString);
+                    long current = date.getTime();
+                    if (current > entry.lastModified) {
+                        entry = fetchURL(url, false);
+                        zipCache.put(url, entry);
+                    }
+                } catch (ParseException e) {
+                   Debug.trace("Failed to parse HTTP Last-Modified field: " + e);
+                   entry = fetchURL(url, false);
+                   zipCache.put(url, entry);
+                }
+           } else { 
+                entry = fetchURL(url, false);
+                zipCache.put(url, entry);
             }
         } else {