Re: JLDAP LdapSearchResults getCount strange behaviour

[Jon Roberts <jon@jonanddeb.net>]

Rob Shepherd wrote:
>             LDAPSearchResults searchResults =
>                 lc.search(  searchBase,
>                             searchScope,
>                             searchFilter,
>                             new String[]{"uid"},          // return all 
> attributes
>                             false);        // return attrs and values
>                 //try{Thread.sleep(200);}
>                 //catch(InterruptedException ie){}
>         System.out.println(System.nanoTime() + "  Num results: 
> "+searchResults.getCount());

The batch size constraint for searches on a connection defaults to 1 
because the assumption is you'll want to do *something* with all that 
data before counting the results. Try setting the batchsize constraint 
on your search to 0. Default will block an entry at a time.

> If I uncomment the Thread.sleep(200) I never see such bizarre results.

Because the search completes.

> Qs. Is getCount() not for this purpose? 

Yes and no. It will tell you where the search is at the moment you ask. 
Recalling the batch size parameter, reread the javadocs:

http://developer.novell.com/documentation/jldap/jldapenu/api/com/novell/ldap/LDAPSearchResults.html#getCount()

> I thought search(...) was a 
> synchronous method... should it not block until everything is returned? 

It is a synchronous search, but it's still designed for efficiency and 
configured by default to assume you will iterate on the results.

> Is there something else I should be doing?

No, but since you're  *I* should probably put some clarification on the 
JLDAP javadocs on my lengthy todo list.

Jon Roberts
www.mentata.com