Item12365: Wrong encoding when accessing AD from utf-8 Foswiki

When using Foswiki (with {Site}{CharSet} set to utf-8) to access an Active Directory server (with {Ldap}{CharSet} set to utf-8), the resulting data is not correctly encoded, i.e. special characters are not displayed properly (just showing a question mark). By removing the deconding step, this can be fixed (but of course it will not working for other encodings):

--- Core.pm.orig        2012-11-29 11:03:32.000000000 +0100
+++ Core.pm     2013-01-22 12:12:12.000000000 +0100
@@ -175,7 +175,8 @@
       if ($blobAttrs{$attr}) {
         $data{$attr} = $ldap->cacheBlob($entry, $attr, $theRefresh);
       } else {
-        $data{$attr} = $ldap->fromLdapCharSet($entry->get_value($attr));
+        #$data{$attr} = $ldap->fromLdapCharSet($entry->get_value($attr));
+        $data{$attr} = $entry->get_value($attr);
       }
     }
     push @results, expandVars($theFormat, %data);
@@ -422,7 +423,7 @@
     next unless defined $value && $value ne '';

     # SMELL: do we need to encode it to utf-8?
-    $value = Encode::decode($ldapCharSet, $value);
+    #$value = Encode::decode($ldapCharSet, $value);

     my $label = $personAttributes->{$attr};
     #print STDERR "$label: $value\n";

-- MartinKaufmann - 22 Jan 2013

The same holds true for LdapContrib. The WikiNames are not translated properly (broken umlaut etc) unless the decoding is commented out:

--- LdapContrib.pm.orig 2012-12-13 15:56:15.000000000 +0100
+++ LdapContrib.pm      2013-01-24 16:14:01.000000000 +0100
@@ -1222,7 +1222,7 @@
         next unless $value;
         $value =~ s/^\s+//o;
         $value =~ s/\s+$//o;
-        $value = $this->fromLdapCharSet($value);
+        #$value = $this->fromLdapCharSet($value);
         #writeDebug("$attr=$value");
         push @wikiName, $value;
       }

-- MartinKaufmann - 24 Jan 2013

This has been fixed in fromLdapCharSet directly not to recode chars if both encodings are the same.

-- MichaelDaum - 18 Mar 2014

Commenting is disabled while not logged in