----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/4152/#review5557 -----------------------------------------------------------
flume-ng-core/src/main/java/org/apache/flume/SinkProcessorType.java <https://reviews.apache.org/r/4152/#comment12021> For consistency with the other types (SinkType, SourceType, ...), can we please put OTHER at the top of the enum listing? Also for same consistency, can we have the comment read: /** * Place holder for custom sink processors not part of this enumeration. */ flume-ng-core/src/main/java/org/apache/flume/sink/SinkProcessorFactory.java <https://reviews.apache.org/r/4152/#comment12035> Before this line: Preconditions.checkNotNull(context); Preconditions.checkNotNull(sinks); logger.debug("Creating instance of sink processor name {}, type {}", name, type); flume-ng-core/src/main/java/org/apache/flume/sink/SinkProcessorFactory.java <https://reviews.apache.org/r/4152/#comment12032> => "Sink processor type {} is a custom type" (this string output matches the convention used by DefaultSourceFactory) flume-ng-core/src/main/java/org/apache/flume/sink/SinkProcessorFactory.java <https://reviews.apache.org/r/4152/#comment12041> add "sink", remove comma: =>"Unable to create sink processor type: " flume-ng-core/src/test/java/org/apache/flume/sink/TestSinkProcessorFactory.java <https://reviews.apache.org/r/4152/#comment12028> Thank you for this test. It looks like it confirms that the FQDN and the enum name of the sink processor each return return the same SinkProcessor instance. I see that the intent of using the FQDN is to test the "OTHER" functionality. - Will On 2012-03-02 07:51:49, Juhani Connolly wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/4152/ > ----------------------------------------------------------- > > (Updated 2012-03-02 07:51:49) > > > Review request for Flume. > > > Summary > ------- > > Made custom processors possible in the same way as sink and source: the type > can represent a classname or a shorthand name from the typedef. > > > This addresses bug FLUME-1001. > https://issues.apache.org/jira/browse/FLUME-1001 > > > Diffs > ----- > > flume-ng-core/src/main/java/org/apache/flume/SinkProcessorType.java be1891b > flume-ng-core/src/main/java/org/apache/flume/sink/SinkProcessorFactory.java > 10f9f4e > > flume-ng-core/src/test/java/org/apache/flume/sink/TestSinkProcessorFactory.java > PRE-CREATION > > Diff: https://reviews.apache.org/r/4152/diff > > > Testing > ------- > > Existing tests pass, except I'm having an issue with TestNetcatSource which > seems to be totally unrelated... It turns up on flume-728 head too, probably > something in my environment > > Added a new test to verify that processors created by shorthand name and by > full class create the same class, verifying that creation by classname also > works > > > Thanks, > > Juhani > >
