[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: Are aliases experimental?

On Mon, Nov 04, 2002 at 12:00:19PM -0800, Kurt D. Zeilenga wrote:
> At 02:09 PM 2002-11-03, Hallvard B Furuseth wrote:
> >Are aliases still experimental?
> 
> Yes.  Aside not implementing the complete aliasing semantics,
> slapd(8) has problems dealing with large number of aliases.
> It likely not all that hard to resolve these issues, but it
> would require some redesign of basic search logic in the
> backends.  Volunteers welcomed.

To me the biggest problem seems to be that indexes are not used.
I think one should build index entries for the alias entries,
based on the entries they point to. It will however require an
extra check whenever entries are modified. When an entry is
modified one will have to check if it has aliases (this can be
pretty fast with proper indexing), and if so, update the alias
index entries.

Stig