[jira] [Commented] (DIRAPI-396) Additional loading of a schema not complete

["Emmanuel Lécharny (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/DIRAPI-396?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17786106#comment-17786106 ] 

Emmanuel Lécharny commented on DIRAPI-396:
------------------------------------------

Hi!

There is something you need to do *after* having loaded the _Mozilla_ schema:
{code:java}
schemaManager.loadDisabled( "mozilla" ); 

// Regenerate the references
schemaManager.getRegistries().buildReferences(); 
...

{code}
This is the part where all the missing MatchingRule are loaded from their OID.

 

That should work.

> Additional loading of a schema not complete
> -------------------------------------------
>
>                 Key: DIRAPI-396
>                 URL: https://issues.apache.org/jira/browse/DIRAPI-396
>             Project: Directory Client API
>          Issue Type: Bug
>    Affects Versions: 2.1.5
>            Reporter: Thomas Gaebler
>            Priority: Critical
>
> Hi community!
> I have my own LDAP server implementation and use org.apache.directory.server libraries in version 2.0.0.AM27 and the API in version 2.1.5.
> The LDAP server is called by various mail clients, including Thunderbird. If I start a simplified address book search in Thunderbird, I receive the following filter in the SearchRequest:
> {code:java}
> (|(|(cn=*test*)(givenName=*test*)(sn=*test*)(xmozillanickname=*test*)(mail=*test*)(mozillaSecondEmail=*test*)(description=*test*)(o=*test*)(ou=*test*)(title=*test*)(workurl=*test*)(homeurl=*test*))(objectClass=referral))) {code}
> Not nice, but I have no influence on that.
> The 'mozilla' schema is not loaded by default. So far I have loaded this at server startup using the following code:
> {code:java}
> schemaManager.loadDisabled( "mozilla" ); {code}
> This has worked well so far. With the update to the latest versions, however, there is a problem. The received search query is now not parsed correctly. If I display it in the LOG, it now says:
> {code:java}
> (|(|(cn=*test*)(givenName=*test*)(sn=*test*)(Undefined:mozillaNickname)(mail=*test*)(Undefined:mozillaSecondEmail)(description=*test*)(o=*test*)(ou=*test*)(title=*test*)(Undefined:mozillaWorkUrl)(Undefined:mozillaHomeUrl))(objectClass=referral)) {code}
> All mozilla attributes are now Undefined. The reason for this can be found here:
> {code:java}
> org.apache.directory.server.core.api.normalization.FilterNormalizingVisitor.visitSubstringNode(SubstringNode)
> {
>         AttributeType attributeType = schemaManager.lookupAttributeTypeRegistry( node.getAttribute() );
>         MatchingRule substringMR = attributeType.getSubstring();
>         
>         if ( ( substringMR == null ) || ( substringMR.getNormalizer() == null ) )
>         {
>             // No normalizer for a Substring filter
>             return new UndefinedNode( node.getAttribute() );
>         }
>         ...
> }{code}
> The MatchingRule (substringMR) for the mozilla Attributes is null. Only the substringOid is set. If fixed this for some attributes, but of course, it's a hack:
> {code:java}
>         schemaManager.loadDisabled( "mozilla" );
>         
>         Registries registries = schemaManager.getRegistries();
>         MatchingRuleRegistry matchingRuleRegistry = registries.getMatchingRuleRegistry();
>         List<String> attributeNames = asList( "mozillaSecondEmail", "mozillaNickname", "mozillaHomeUrl", "mozillaWorkUrl" );
>         for ( var attributeName : attributeNames )
>         {
>             AttributeType attribute = schemaManager.lookupAttributeTypeRegistry( attributeName );
>             
>             attribute.unlock();
>             if ( isNull( attribute.getSubstring() ) && nonNull( attribute.getSubstringOid() ) )
>             {
>                 MatchingRule matchingRule = matchingRuleRegistry.lookup( attribute.getSubstringOid() );
>                 attribute.setSubstring( matchingRule );
>             }
>             
>             if ( isNull( attribute.getEquality() ) && nonNull( attribute.getEqualityOid() ) )
>             {
>                 MatchingRule matchingRule = matchingRuleRegistry.lookup( attribute.getEqualityOid() );
>                 attribute.setEquality( matchingRule );
>             }
>             
>             if ( isNull( attribute.getOrdering() ) && nonNull( attribute.getOrderingOid() ) )
>             {
>                 MatchingRule matchingRule = matchingRuleRegistry.lookup( attribute.getOrderingOid() );
>                 attribute.setOrdering( matchingRule );
>             }
>             attribute.lock();
>         }
> {code}
> If I have overlooked something in the implementation so far, I would be grateful for a hint. But it looks like a bug.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@directory.apache.org
For additional commands, e-mail: dev-help@directory.apache.org