[
https://issues.apache.org/jira/browse/DIRSERVER-872?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12495974
]
Emmanuel Lecharny commented on DIRSERVER-872:
---------------------------------------------
We have limited the DN normalization to three places :
- when receiving the request (and it's done in two passes, the first one just
normalize the string, the second one use tha Schema to normalize values)
- when we receive a value which is a DN (like memberOf, seeAlso)
- and when reading data from the partitions to compare them whith the received
DN
There are some more places where we are normalizing a DN, notably when we are
manipulating Dn (concatenating two DNs leads to a re-normalization, for
instance). We should have special cases where we don't want to normalize the
DNs, if it has already been noprmalized.
We should be able to get rid of the last kind of normaliztion, if we store a
serialized version of the DN on disk.
To be continued...
> Still too many DN normalization
> -------------------------------
>
> Key: DIRSERVER-872
> URL: https://issues.apache.org/jira/browse/DIRSERVER-872
> Project: Directory ApacheDS
> Issue Type: Improvement
> Reporter: Emmanuel Lecharny
> Assigned To: Emmanuel Lecharny
> Fix For: 1.5.1
>
>
> This is not a bug, but as I was doing some profiling session, I found that we
> are still normalizing 5 times more than what we should. In fact, the
> PrincipalDN is normalized again and again, when it should have been
> normalized when the user bound, and should never be changed. All the
> following normalizations are just used to obtain better numbers when we will
> remove them to make the server run faster :)
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
