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

see: http://www.yacy-forum.de/viewtopic.php?t=1416

Browse Source 
git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@1037 6c8d7289-2bf4-0310-a012-ef5d649a1542
pull/1/head
borg-0300 20 years ago
parent dd24f0252f
commit 17d2830394
1 changed files with 6 additions and 6 deletions
Show Stats Download Patch File Download Diff File

12
source/de/anomic/htmlFilter/htmlFilterContentScraper.java
Unescape View File

@ -110,7 +110,7 @@ public class htmlFilterContentScraper extends htmlFilterAbstractScraper implemen
} }
public static String urlNormalform(String us) { public static String urlNormalform(String us) {
// serverLog.logFiner("htmlFilter", "urlNormalform: IN=" + us); serverLog.logFiner("htmlFilter", "urlNormalform: IN=" + us);
if (us == null) { return null; } if (us == null) { return null; }
if (us.length() == 0) { return null; } if (us.length() == 0) { return null; }
@ -125,28 +125,28 @@ public class htmlFilterContentScraper extends htmlFilterAbstractScraper implemen
if (us.startsWith("https")) { if (us.startsWith("https")) {
if (us.endsWith(":443")) { if (us.endsWith(":443")) {
us = us.substring(0, us.length() - 4); us = us.substring(0, us.length() - 4);
// serverLog.logFinest("htmlFilter", "urlNormalform: :443=" + us); serverLog.logFinest("htmlFilter", "urlNormalform: :443=" + us);
} else { } else {
cpos = us.indexOf(":443/"); cpos = us.indexOf(":443/");
if (cpos >= 0) { if (cpos >= 0) {
us = us.substring(0, cpos).concat(us.substring(cpos + 4)); us = us.substring(0, cpos).concat(us.substring(cpos + 4));
// serverLog.logFinest("htmlFilter", "urlNormalform: :443/=" + us); serverLog.logFinest("htmlFilter", "urlNormalform: :443/=" + us);
} }
} }
} else if (us.startsWith("http")) { } else if (us.startsWith("http")) {
if (us.endsWith(":80")) { if (us.endsWith(":80")) {
us = us.substring(0, us.length() - 3); us = us.substring(0, us.length() - 3);
// serverLog.logFinest("htmlFilter", "urlNormalform: :80=" + us); serverLog.logFinest("htmlFilter", "urlNormalform: :80=" + us);
} else { } else {
cpos = us.indexOf(":80/"); cpos = us.indexOf(":80/");
if (cpos >= 0) { if (cpos >= 0) {
us = us.substring(0, cpos).concat(us.substring(cpos + 3)); us = us.substring(0, cpos).concat(us.substring(cpos + 3));
// serverLog.logFinest("htmlFilter", "urlNormalform: :80/=" + us); serverLog.logFinest("htmlFilter", "urlNormalform: :80/=" + us);
} }
} }
} }
if (((us.endsWith("/")) && (us.lastIndexOf('/', us.length() - 2) < 8))) us = us.substring(0, us.length() - 1); if (((us.endsWith("/")) && (us.lastIndexOf('/', us.length() - 2) < 8))) us = us.substring(0, us.length() - 1);
// serverLog.logFine("htmlFilter", "urlNormalform: OUT=" + us); serverLog.logFine("htmlFilter", "urlNormalform: OUT=" + us);
return us; return us;
} }

Write Preview
Loading…
Cancel
Save