*) Removing url hash from loaded list before inserting the url into the crawling queue

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@1296 6c8d7289-2bf4-0310-a012-ef5d649a1542
pull/1/head
theli 19 years ago
parent 99861ff2c1
commit 94e54c4450

@ -14,6 +14,7 @@ import de.anomic.http.httpc;
import de.anomic.plasma.plasmaCrawlProfile;
import de.anomic.plasma.plasmaParser;
import de.anomic.plasma.plasmaSwitchboard;
import de.anomic.plasma.plasmaURL;
import de.anomic.server.serverCore;
import de.anomic.server.serverHandler;
import de.anomic.server.logging.serverLog;
@ -170,6 +171,12 @@ public class urlRedirectord implements serverHandler {
reqURL,
header.mime())
) {
// first delete old entry, if exists
String urlhash = plasmaURL.urlHash(this.nextURL);
switchboard.urlPool.loadedURL.remove(urlhash);
switchboard.urlPool.noticeURL.remove(urlhash);
switchboard.urlPool.errorURL.remove(urlhash);
// enqueuing URL for crawling
reasonString = switchboard.sbStackCrawlThread.stackCrawl(
this.nextURL,

Loading…
Cancel
Save