[
https://issues.apache.org/jira/browse/SOLR-14088?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16996871#comment-16996871
]
Andrzej Bialecki commented on SOLR-14088:
-----------------------------------------
The exception and the stacktrace that was mentioned in the other issue
(\{{Could not initialize...}}) doesn't actually show the underlying first cause
of the issue. Was it a {{ClassNotFoundException}}? Or was it rather a
{{ClassCastException}}?
If it was the former then it means that the loading was attempted from a
classloader that didn't have access to all dependencies.
If it's the latter it probably means that the class was already loaded by an
incompatible (sibling or child) classloader and it became visible to the parent
classloader that also loaded the same class. This happens if a common
dependency is accidentally put both into a child classloader and a parent
classloader, and if the class loading is executed child-first - then the common
dependency classes are loaded twice by different classloaders, and by
definition they are incompatible (even though the underlying bytecode is the
same).
> Tika and commons-compress dependency in solr core causes classloader issue
> --------------------------------------------------------------------------
>
> Key: SOLR-14088
> URL: https://issues.apache.org/jira/browse/SOLR-14088
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Components: contrib - Solr Cell (Tika extraction)
> Reporter: Kevin Risden
> Priority: Major
>
> SOLR-14086 found that if commons-compress is in core ivy.xml as a compile
> dependency, it messes up the classloader for any commons-compress
> dependencies. It causes issues with items like xz being loaded.
> This is problematic where dependencies shouldn't matter based on classloader.
> This jira to to determine if there is something wrong w/ Solr's classloader
> or if its a commons-compress issue only.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
