directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alex Karasulu <>
Subject Re: Alias cycle detection
Date Wed, 15 Jun 2011 09:26:54 GMT
On Wed, Jun 15, 2011 at 8:33 AM, Emmanuel Lecharny <> wrote:
> On 6/15/11 2:28 AM, Alex Karasulu wrote:
>>> In that case we can also get rid of all the alias indices (aliasIdx,
>>> oneAliasIdx, subAliasIdx).
>> This would be a big mistake. Then you'd be better off just getting rid
>> of aliases all together. You don't know how many aliases you'll have
>> at the end of the day and presuming some number is making presumptions
>> on behalf of the user even if aliases in practice are not used.
> Detecting if an entry is an alias is just a matter of checking the 'alias'
> ObjectClass. This is already indexed.
> Sorry, but can you refresh my memory and tell what are we using those index
> for ? I was wondering yesterday if we needed to keep an 'aliased Object' ->
> 'alias' reverse index.

OK let me do this. Seems this page is missing information here so I
will add it so we're all on the same page with these indices. Will be
good for others to see as well:

After I fill this out perhaps we can discuss this in more depth. But
let's make sure that check to prevent alias cycles is put back and
improved otherwise we'll see odd behavior in search with alias
dereferencing enabled. We cannot just go and do this without modifying
a few other things.

> Maybe I'm overlooking this aspect. Anyway, the index are stillpresent atm.

I will update this. Also seems the people that removed the updn and
ndn indices introducing the rdn indices never updated the
documentation. This is a core concept and it should be done too for
those who want to come up to speed with ADS internals.


View raw message