Author: sshafroi
Date: 2009-01-28 12:11:58 +0100 (Wed, 28 Jan 2009)
New Revision: 7176
Modified:
branches/2.18/core-api/src/main/java/no/sesat/search/result/Boomerang.java
Log:
SEARCH-5144 - ALFA: URL med ?\195?\184 vises ikke i front og logo f?\195?\165r
da feil treffside
Punycode url as a safeguard. We will add a punycode directive to do this
explicit. This should fix broken links to sites containing non ascii
characters.
Modified:
branches/2.18/core-api/src/main/java/no/sesat/search/result/Boomerang.java
===================================================================
--- branches/2.18/core-api/src/main/java/no/sesat/search/result/Boomerang.java
2009-01-28 10:35:30 UTC (rev 7175)
+++ branches/2.18/core-api/src/main/java/no/sesat/search/result/Boomerang.java
2009-01-28 11:11:58 UTC (rev 7176)
@@ -1,4 +1,4 @@
-/* Copyright (2005-2007) Schibsted Søk AS
+/* Copyright (2005-2009) Schibsted Søk AS
* This file is part of SESAT.
*
* SESAT is free software: you can redistribute it and/or modify
@@ -16,6 +16,7 @@
*/
package no.sesat.search.result;
+import java.net.MalformedURLException;
import java.util.StringTokenizer;
import no.sesat.search.site.Site;
import no.sesat.search.site.config.SiteConfiguration;
@@ -74,11 +75,25 @@
}
// append the original destination url
- toUrl.append(orgUrl);
+ toUrl.append(punyencode(orgUrl));
return toUrl.toString();
}
+ private static String punyencode(String url) {
+ String host = null;
+ try {
+ host = new java.net.URL(url).getHost();
+ } catch (MalformedURLException ex) {
+ LOG.warn("Invalid url in boomerang: " + url, ex);
+ }
+
+ if (host != null) {
+ url = url.replace(host, java.net.IDN.toASCII(host));
+ }
+ return url;
+ }
+
// Constructors -------------------------------------------------
// Public --------------------------------------------------------
_______________________________________________
Kernel-commits mailing list
Kernel-commits@sesat.no
http://sesat.no/mailman/listinfo/kernel-commits
