[
https://issues.apache.org/jira/browse/OAK-2814?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chetan Mehrotra updated OAK-2814:
---------------------------------
Attachment: OAK-2814-v1.patch
[patch|^OAK-2814-v1.patch] for the same.
Earlier code in ObservationManager had a bug where the includePaths.remove
triggers a ConcurrentModificationException as the iterator for that set is
currently being used in outer loop
{code:linenumbers=true}
private static void optimise(Set<String> includePaths, Set<String>
excludedPaths) {
Set<String> retain = newHashSet();
for (String include : includePaths) {
for (String exclude : excludedPaths) {
if (exclude.equals(include) || isAncestor(exclude, include)) {
includePaths.remove(include);
} else if (isAncestor(include, exclude)) {
retain.add(exclude);
}
}
}
excludedPaths.retainAll(retain);
}
{code}
[~mduerig] Can you review the patch? Specially the contract examples mentioned
in javadoc to ensure all aspects are covered
Also I was not able to create a scenario now for second clause in as now post
your suggestion redundant includes would be removed anyway.
bq. if (exclude.equals(include) || isAncestor(exclude, include))
> Refactor the optimize logic regarding path include and exclude to avoid
> duplication
> -----------------------------------------------------------------------------------
>
> Key: OAK-2814
> URL: https://issues.apache.org/jira/browse/OAK-2814
> Project: Jackrabbit Oak
> Issue Type: Improvement
> Components: core
> Reporter: Chetan Mehrotra
> Assignee: Chetan Mehrotra
> Priority: Minor
> Labels: technical_debt
> Fix For: 1.4
>
> Attachments: OAK-2814-v1.patch
>
>
> {{ObservationManagerImpl}} has a optimize method which process the list of
> includes and excludes and removes redundant clauses. That logic is now also
> being used in index filtering (OAK-2599) and is getting duplicated.
> Going forward we need to refactor this logic so that both places can use it
> without copying. Possibly making it part of PathUtils
> [~mduerig] Also suggested to further optimize
> bq. Also PathFilter#optimise could be further optimised by removing entries
> that subsume each other (e.g. including /a/b, /a is the same as including
> (/a.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
