boxtec
/
yacy_search_server
mirror of https://github.com/yacy/yacy_search_server
4
1
Fork 0
Code Issues Projects Releases Wiki Activity

fix for Malformed URL Exception in url migration

Browse Source 
git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@2825 6c8d7289-2bf4-0310-a012-ef5d649a1542
pull/1/head
orbiter 19 years ago
parent 9eecc9a888
commit 11843bba7f
1 changed files with 31 additions and 28 deletions
Show Stats Download Patch File Download Diff File

59
source/yacy.java
Unescape View File

@ -1122,25 +1122,29 @@ public final class yacy {
}
private static void migratelurls(File root, File urlHash) {
plasmaURLPool pool = new plasmaURLPool(new File(root, "DATA/PLASMADB"), new File(root, "DATA/INDEX"), 16000, true, 1000, true, 1000, true, 10000);
kelondroTree oldindex = null;
try {
plasmaURLPool pool = new plasmaURLPool(new File(root, "DATA/PLASMADB"), new File(root, "DATA/INDEX"), 16000, true, 1000, true, 1000, true, 10000);
kelondroTree oldindex = new kelondroTree(urlHash, 1000, -1, kelondroTree.defaultObjectCachePercent, plasmaCrawlLURLOldEntry.rowdef);
oldindex = new kelondroTree(urlHash, 1000, -1, kelondroTree.defaultObjectCachePercent, plasmaCrawlLURLOldEntry.rowdef);
} catch (IOException e) {
System.out.println("ERROR: CANNOT OPEN OLD INDEX: " + e.getMessage());
}
long start = System.currentTimeMillis();
long last = start;
int tc = oldindex.size(), c = 0;
Iterator eiter = oldindex.contentRows(-1);
kelondroRow.Entry oldrow;
plasmaCrawlLURLEntry oldentry;
plasmaCrawlLURLEntry newentry;
plasmaCrawlLURLEntry.Components comp;
byte[] dummymd5 = new byte[0];
while (eiter.hasNext()) {
oldrow = (kelondroRow.Entry) eiter.next();
if (oldrow != null) {
oldentry = new plasmaCrawlLURLOldEntry(oldrow, null);
comp = oldentry.comp();
newentry = pool.loadedURL.newEntry(
long start = System.currentTimeMillis();
long last = start;
int tc = oldindex.size(), c = 0;
Iterator eiter = oldindex.contentRows(-1);
kelondroRow.Entry oldrow;
plasmaCrawlLURLEntry oldentry;
plasmaCrawlLURLEntry newentry;
plasmaCrawlLURLEntry.Components comp;
byte[] dummymd5 = new byte[0];
while (eiter.hasNext()) {
oldrow = (kelondroRow.Entry) eiter.next();
if (oldrow != null) try {
oldentry = new plasmaCrawlLURLOldEntry(oldrow, null);
comp = oldentry.comp();
newentry = pool.loadedURL.newEntry(
comp.url(),
comp.descr(),
"",
@ -1157,20 +1161,19 @@ public final class yacy {
new bitfield(4),
oldentry.language(),
0, 0, 0, 0, 0, 0);
pool.loadedURL.store(newentry);
}
pool.loadedURL.store(newentry);
c++;
if (System.currentTimeMillis() - last > 60000) {
System.out.println("Migrated " + c + " from " + tc + " urls. Estimated remaining time: " + ((System.currentTimeMillis() - start) * (tc - c) / c / 60000) + " minutes");
last = System.currentTimeMillis();
}
} catch (IOException e) {
// ignore
}
if (System.currentTimeMillis() - last > 60000) {
System.out.println("Migrated " + c + " from " + tc + " urls. Estimated remaining time: " + ((System.currentTimeMillis() - start) * (tc - c) / c / 60000) + " minutes");
last = System.currentTimeMillis();
}
pool.close();
oldindex.close();
System.out.println("MIGRATION OF " + c + " URLs FINISHED");
} catch (IOException e) {
e.printStackTrace();
}
pool.close();
try { oldindex.close(); } catch (IOException e) { }
System.out.println("MIGRATION OF " + c + " URLs FINISHED");
}
private static String[] shift(String[] args, int pos, int count) {

Write Preview
Loading…
Cancel
Save