[
https://issues.apache.org/jira/browse/SOLR-2694?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13078787#comment-13078787
]
David Smiley commented on SOLR-2694:
------------------------------------
As an aside, I think it's important to document thread safety (or lack thereof)
in an API, particularly for classes that are extension points like
UpdateRequestProcessor. FWIW Joshua Bloch thinks so too as it's item #70 in
"Effective Java". This is good practice for object life-cycle information too
-- i.e. construction, and any applicable callbacks. As I'm working with coding
in Lucene and Solr I often wonder wether I need to be careful to be thread safe
or not and I always have to go deduce this question myself instead of the API
telling me. I propose that a comment be added to UpdateRequestProcessor saying:
bq. UpdateRequestProcessors are instantiated once per request, and thus don't
need to be thread safe.
> LogUpdateProcessor not thread safe
> ----------------------------------
>
> Key: SOLR-2694
> URL: https://issues.apache.org/jira/browse/SOLR-2694
> Project: Solr
> Issue Type: Bug
> Components: update
> Affects Versions: 1.4.1, 3.1, 3.2, 3.3, 4.0
> Reporter: Jan Høydahl
>
> Using the LogUpdateProcessor while feeding in multiple parallell threads does
> not work, as LogUpdateProcessor is not threadsafe.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
