set revisit profile on 304 not modified

songfj · Jul 11, 2014 · 810ad91 · 810ad91
1 parent 905a98c
commit 810ad91
Showing 1 changed file with 7 additions and 1 deletion.
diff --git a/modules/src/main/java/org/archive/modules/recrawl/FetchHistoryProcessor.java b/modules/src/main/java/org/archive/modules/recrawl/FetchHistoryProcessor.java
@@ -33,6 +33,7 @@
 import org.archive.modules.CrawlURI;
 import org.archive.modules.Processor;
 import org.archive.modules.revisit.IdenticalPayloadDigestRevisit;
+import org.archive.modules.revisit.ServerNotModifiedRevisit;
 
 /**
  * Maintain a history of fetch information inside the CrawlURI's attributes. 
@@ -105,7 +106,12 @@ protected void innerProcess(CrawlURI puri) throws InterruptedException {
 
         curi.getData().put(A_FETCH_HISTORY, history);
 
-        if (hasIdenticalDigest(curi)) {
+        if (curi.getFetchStatus() == 304) {
+            ServerNotModifiedRevisit revisit = new ServerNotModifiedRevisit();
+            revisit.setETag((String) latestFetch.get(A_ETAG_HEADER));
+            revisit.setLastModified((String) latestFetch.get(A_LAST_MODIFIED_HEADER));
+            curi.setRevisitProfile(revisit);
+        } else if (hasIdenticalDigest(curi)) {
             curi.getAnnotations().add("duplicate:digest");
             IdenticalPayloadDigestRevisit revisit = 
             		new IdenticalPayloadDigestRevisit((String)history[1].get(A_CONTENT_DIGEST));