[
https://issues.apache.org/jira/browse/DIRSERVER-1121?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Emmanuel Lecharny updated DIRSERVER-1121:
-----------------------------------------
Fix Version/s: (was: 1.5.2)
1.5.3
Didn't had time to evaluate this issue. Postponed.
> Filter mutated by optimizer passed to other partitions
> ------------------------------------------------------
>
> Key: DIRSERVER-1121
> URL: https://issues.apache.org/jira/browse/DIRSERVER-1121
> Project: Directory ApacheDS
> Issue Type: Bug
> Components: core
> Affects Versions: 1.5.1
> Reporter: Shon Vella
> Assignee: Alex Karasulu
> Priority: Minor
> Fix For: 1.5.3
>
>
> org.apache.directory.server.core.authz.GroupCache.initialize() creates a
> filter and subsequently performs a search using the same filter object on
> different suffixes. When passed to a JDBMPartition,
> org.apache.directory.server.core.partition.impl.btree.DefaultOptimizer.annotate()
> adds annotations to nodes in the filter. The same filter is then passed with
> all the annotations to subsequent searches on other parititions. When passing
> to an alternate custom partition implementation, these annotations get in the
> way because the printToBuffer() functionality of each filter node includes
> annotation information in the serialized form of the filter making it
> unusable.
> Providing a way to recursively remove annotations would provide a workable
> bandaid for this specific problem, though it is probably a bad idea to reuse
> a filter that has been previously annotated, so either the partition should
> make a copy before annotating or GroupCache (and any other code that uses the
> same filter reuse paradigm) shouldn't be reusing the filter.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
