LDAP re-encode a utf-8 string

-- YangCheng - 07 Sep 2016

I happend to found that all my chinese words downloaded from my LDAP server is wrongly encoded to utf-8 even if it's already utf-8

Here is example:

1.When i get an entry from LDAP using linux command "ldapsearch",

dn='CN=chengyang [\xc3\xa7\xc2\xa8\xc2\x8b\xc3\xa6\xc2\xb4\xc2\x8b],OU=Employee,OU=Mioffice,DC=mioffice,DC=cn'

It's already a utf-8 string.

2. But when i get the same entry from LDAP contrib logs, i got

dn='CN=chengyang [\xc3\xa7\xc2\xa8\xc2\x8b\xc3\xa6\xc2\xb4\xc2\x8b],OU=Employee,OU=Mioffice,DC=mioffice,DC=cn'

You will see it's wrongly encoded again as utf-8.

And i found, before cacheGroupFromEntry is called, the string is already like that, is it NET::LDAP's bug?

But when i just write a test.pm to query using NET::LDAP, everything is alright

-- YangCheng - 07 Sep 2016

use Net::LDAP;

$ldap = Net::LDAP->new('********') or die "$@";

$mesg = $ldap->bind('***************', password => '********') ; # an anonymous bind $mesg = $ldap->search( # perform a search base => "OU=Mioffice,DC=mioffice,DC=cn", filter => "sAMAccountName=chengyang" );

$mesg->code && die $mesg->error;

foreach $entry ($mesg->entries) { # $entry->dump; $dn = $entry->dn(); print STDOUT "dn is $dn\n"; $dn =~ s/(.)/sprintf("0x%x",ord($1))/eg; print STDOUT "dn hex is $dn\n"; }

It's really wierd, Above codes, if i execute in the perl script just using command, i will get correct code. But if i run in LdapContrib.pm, the apache error log will show [Thu Sep 08 10:38:17.676292 2016] [cgi:error] [pid 49315] [client 10.237.130.70:49742] AH01215: cytest dn is CN=chengyang [\xc3\xa7\xc2\xa8\xc2\x8b\xc3\xa6\xc2\xb4\xc2\x8b],OU=Employee,OU=Mioffice,DC=mioffice,DC=cn, referer: http://wiki.bsp.xiaomi.com/Main/WikiGroups

Why? This is a really damn bug

-- YangCheng - 08 Sep 2016

the encoding is already wrong with NET::LDAP itself, but is affected by wiki Plugin environment

-- YangCheng - 08 Sep 2016