[
https://issues.apache.org/jira/browse/RANGER-5245?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18008649#comment-18008649
]
Vyom Mani Tiwari commented on RANGER-5245:
------------------------------------------
Given the logic of the {{dedupTags}} method, where it always replaces the
higher tag ID with the lower one if it exists in {{{}cachedTags{}}}, it's
unlikely to enter a situation where there's a cycle in {{{}replacedIds{}}}.
The logic ensures that the {{replacedIds}} map will always point to a lower tag
ID, which reduces the possibility of cycles. Since the replacement is always
directed towards the lower ID, it's more like a directed acyclic graph (DAG)
where each node (tag ID) points to a lower node.
Given this logic, it's reasonable to conclude that cycles in {{replacedIds}}
are unlikely to occur. The method's design seems to prevent cycles by always
favoring the lower tag ID.
> Investigate the potential loop in dedupTags method
> ---------------------------------------------------
>
> Key: RANGER-5245
> URL: https://issues.apache.org/jira/browse/RANGER-5245
> Project: Ranger
> Issue Type: Bug
> Components: Ranger
> Affects Versions: 2.6.0
> Reporter: Vyom Mani Tiwari
> Assignee: Vyom Mani Tiwari
> Priority: Major
>
> for (Long replacerTagId = replacedIds.get(tagId); replacerTagId != null;
> replacerTagId = replacedIds.get(mappedTagId)) {
> mappedTagId = replacerTagId;
> }
> Investigate if it possible for dedupTags to enter infinite loop.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
