[
https://issues.apache.org/jira/browse/UIMA-1715?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Marshall Schor resolved UIMA-1715.
----------------------------------
Resolution: Invalid
I think this is invalid - reading the code for CasCreationUtils shows that if
produceResource is called, then destroy() is later called (except if an
exception occurred - but in that case we have no resource to call destroy on,
anyways).
> uima-as client design contract for insuring proper cleanup not supported by
> core UIMA
> -------------------------------------------------------------------------------------
>
> Key: UIMA-1715
> URL: https://issues.apache.org/jira/browse/UIMA-1715
> Project: UIMA
> Issue Type: Bug
> Components: Async Scaleout
> Affects Versions: 2.3AS
> Reporter: Marshall Schor
> Assignee: Jerry Cwiklik
>
> The uima-as jms client design contract is to call "stop" after a call to
> "initialize", in order to clean up sockets, etc. However, base uima code has
> methods that call initialize just get client metadata, e.g.,
> CasCreationUtilities.getMetaDataList, which are general purpose and are
> supposed to work for any analysis engine resource.
> I think the jms client should figure out how to clean up after itself. Using
> finalize for this is not a good idea, for the standard reasons (see end of
> article http://java.sun.com/developer/technicalArticles/javase/finalization/
> ). Perhaps we need to call destroy after the call to produceResource, and
> insure that destroy is implemented by JmsAnalysisEngineServiceAdapter (and
> others) to release/cleanup
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
