beam-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Thomas Groh (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (BEAM-1657) DirectRunner should not call close twice in UnboundedSourceExecutorFactory
Date Thu, 09 Mar 2017 05:57:38 GMT

     [ https://issues.apache.org/jira/browse/BEAM-1657?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Thomas Groh updated BEAM-1657:
------------------------------
    Description: At the end of the processElement, in `finishRead` there is a 5% chance we
decide to `close` the reader. If that happens, it will null-out the reader in the returned
shard. But then the end of the `processElement` call will attempt too `close` the reader again,
violating the {{AutoCloseable}} spec.  (was: At the end of the processElement, in `finishRead`
there is a 5% chance we decide to `close` the reader. If that happens, it will null-out the
reader in the returned shard. But then the end of the `processElement` call will attempt too
`close` the reader again, violating the spec.)

> DirectRunner should not call close twice in UnboundedSourceExecutorFactory
> --------------------------------------------------------------------------
>
>                 Key: BEAM-1657
>                 URL: https://issues.apache.org/jira/browse/BEAM-1657
>             Project: Beam
>          Issue Type: Bug
>          Components: runner-direct
>            Reporter: Ben Chambers
>            Assignee: Thomas Groh
>
> At the end of the processElement, in `finishRead` there is a 5% chance we decide to `close`
the reader. If that happens, it will null-out the reader in the returned shard. But then the
end of the `processElement` call will attempt too `close` the reader again, violating the
{{AutoCloseable}} spec.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message