[
https://issues.apache.org/jira/browse/UIMA-6057?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16876973#comment-16876973
]
Marshall Schor commented on UIMA-6057:
--------------------------------------
a test case would be wonderful. It often exposes some flaw in my thinking :).
If you're seeing the problem in 4), and the patch would fix that, then that
implies to me that the code in "main" is organized so that it's following the 1
caveat I mentioned - that there is no reference to the JCas class in main that
would cause it to be loaded, before the the special pipeline is run.
It sounds like the intent is to be able to share (some?) JCas classes loaded by
a PEAR with the main class. Suppose we fix UIMA core to support this (I'm
imagining some API call that says for this PEAR, load the JCas classes (maybe a
list of class names, or *all* ), which would require that the JCas class
definitions be findable from the "main" pipeline, and those would be used by
the "special" pipeline.)
* Would that kind of approach more directly address this issue?
* Are there even better ways to directly address these use cases? (if so, what
are they? :) )
> Avoid falsely switching classloader
> -----------------------------------
>
> Key: UIMA-6057
> URL: https://issues.apache.org/jira/browse/UIMA-6057
> Project: UIMA
> Issue Type: Bug
> Components: Core Java Framework
> Reporter: Matthias Koch
> Priority: Major
> Attachments: UIMA-6057.diff
>
>
> In some cases the classloader is switched back, although it hasn't be
> switched before processing.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
