com.atlassian.crowd.util.connector
Class DNStandardiser

java.lang.Object
  com.atlassian.crowd.util.connector.DNStandardiser

public class DNStandardiser
extends java.lang.Object

Constructor Summary

DNStandardiser()

Method Summary

static java.lang.String	standardise(org.springframework.ldap.core.DistinguishedName dn, boolean forceProperStandard)
static java.lang.String	standardise(java.lang.String dn, boolean forceProperStandard) Converts a DN string into a "standard" DN string representation consummable by caches.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

DNStandardiser

public DNStandardiser()

Method Detail

standardise

public static java.lang.String standardise(java.lang.String dn,
                                           boolean forceProperStandard)

Converts a DN string into a "standard" DN string representation consummable by caches. This is particularly useful when matching group memberDNs to entity DNs as the two DN formats need to be equivalent. Unfortunately, DN parsing and standardisation is an expensive process (if done over 100,000+ iterations). This is because it involves a Java CC parse followed by a serialise. This is 100 times slower than simply calling toLowerCase on the DN! AD returns DN's that have mixed case attribute labels, eg. member attribute might look like: CN=user-1100,CN=Users,DC=sydney,DC=atlassian,DC=com for AD. If you call: new DistinguishedName("CN=user-1100,CN=Users,DC=sydney,DC=atlassian,DC=com").toString() you get: "cn=user-1100,cn=Users,dc=sydney,dc=atlassian,dc=com" ie. SpringLDAP does some automagic foo to convert the DN attribute names into lower case, causing the DN to be monged. Since everything goes through a dirContextAdapter (thanks for coming in: SpringLDAP), when we call context.getDn() in the UserContextMapper/GroupContextMapper, the DN we get back is a SpringLDAP parsed DN (ie. forces lower-casing of attribute names). However, when we get back memberDNs for a group, these attributes values are pure String representaions of the DN returned from the server. This will cause the memberDN to not match the DN of any object previously seen as the memberDN will not be forced toLowerCase or forced to compaction (",") or comma-space-delimmition (", ") for RDN components of the memberDN. It is critical that the DN and memberDN are comparable based on a raw String comparison (String.equals) for things like group membership determination and cache consistency. Therefore, if it is known that the directory server always returns DNs that do not have ", " but have "," separating the naming components, then we do not need to force "proper" (ie. complete + expensive) standardisation. Thus if forceProperStandard is false, then the compact (spaceless) comma delimited DNs are transformed to lower case. This method assumes that the DNs are case-insensitive. If it is not known whether the directory server will return spaceless or spaced commas, we need to use the full DN standardisation. Set forceProperStandard to true for the significantly slower, 100% effective, standardisation. This method also assumes nothing about the case of DNs.

Parameters:

dn - original DN.

forceProperStandard - true if you want to enforce the slow but effective standardisation process.

Returns:

DN in standard form.

standardise

public static java.lang.String standardise(org.springframework.ldap.core.DistinguishedName dn,
                                           boolean forceProperStandard)