directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Emmanuel Lecharny (JIRA)" <j...@apache.org>
Subject [jira] Commented: (DIRSERVER-1173) Delete operation with a PersistentSearch returns the deleted entry
Date Mon, 12 May 2008 09:49:55 GMT

    [ https://issues.apache.org/jira/browse/DIRSERVER-1173?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12596020#action_12596020
] 

Emmanuel Lecharny commented on DIRSERVER-1173:
----------------------------------------------

Actually, it's even worse...

The testPsearchAbandon is not consistent at all. Sometime it fails, sometime it succeeds.

It seems that when it fails, an AbandonRequest is received _before_ the modify request is
run, leading to a removal of the PersistentSearch handler.

Depending on which test is run before the testPsearchAbandon, the result may be different.

I suggest either we fix it in 1.5.3, or we remove the tests and consider PersistentSearch
broken and to be fixed in 1.5.4, but in the mean time, the tests should be commented.

> Delete operation with a PersistentSearch returns the deleted entry
> ------------------------------------------------------------------
>
>                 Key: DIRSERVER-1173
>                 URL: https://issues.apache.org/jira/browse/DIRSERVER-1173
>             Project: Directory ApacheDS
>          Issue Type: Bug
>            Reporter: Emmanuel Lecharny
>
> While debugging a failure in PersistentSearch I found that we have an inconsistant behavior
when deleting entries :
> testPsearchDelete :
>   ctx.destroySubcontext( RDN ); // RDN = "cn=Tori Amos"
>   ...
>   assertNotNull( listener.result );  // Should be null, but is not
>   assertEquals( RDN, listener.result.getName() );   // Contains the deleted entry...
> Another test :
> testPsearchAbandon :
>   ctx.destroySubcontext( "cn=Jack Black" );
>   ...
>   // there seems to be a race condition here
>   //assertNull( listener.result ); // Has been commented as otherwise, the test would
fail
>   ...
> Note the comment...
> While looking into the PersistentSearchListener code, here is what we have :
>     public void objectRemoved( NamingEvent evt )
>     {
>         // send the entry back
>         sendEntry( evt );
>     }
> This sendEntry method simply return the deleted entry, and is supposed to set the PersistentSearchControl,
so the test is incorrect. We should test that the Control contains the correct ChangeType
>   

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message