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-987) TestStream.advanceWatermarkToInfinity should perhaps also advance processing time
Date Mon, 03 Apr 2017 21:56:41 GMT

     [ https://issues.apache.org/jira/browse/BEAM-987?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Thomas Groh updated BEAM-987:
-----------------------------
    Labels: starter  (was: )

> TestStream.advanceWatermarkToInfinity should perhaps also advance processing time
> ---------------------------------------------------------------------------------
>
>                 Key: BEAM-987
>                 URL: https://issues.apache.org/jira/browse/BEAM-987
>             Project: Beam
>          Issue Type: Improvement
>          Components: runner-direct, testing
>            Reporter: Eugene Kirpichov
>              Labels: starter
>
> I ran into this when writing a test for Splittable DoFn whose input was a TestStream.
I constructed a TestStream that didn't call advanceProcessingTime, and as a result, the SDF's
timers didn't fire and the test got stuck.
> I think the meaning of "advanceWatermarkToInfinity" is "don't add any more elements to
the stream and see what happens eventually", and "eventually" includes "eventually in processing
time domain", not just in event-time domain (watermark).



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

Mime
View raw message