y1chi commented on a change in pull request #11993:
URL: https://github.com/apache/beam/pull/11993#discussion_r442420147
##########
File path:
sdks/java/core/src/test/java/org/apache/beam/sdk/transforms/ParDoTest.java
##########
@@ -3808,15 +3817,24 @@ public void onTimer(
TestStream<KV<String, Integer>> stream =
TestStream.create(KvCoder.of(StringUtf8Coder.of(), VarIntCoder.of()))
// See GlobalWindow,
- // END_OF_GLOBAL_WINDOW is
TIMESTAMP_MAX_VALUE.minus(Duration.standardDays(1))
-
.advanceWatermarkTo(BoundedWindow.TIMESTAMP_MAX_VALUE.minus(Duration.standardDays(1)))
+ // END_OF_GLOBAL_WINDOW is
TIMESTAMP_MAX_VALUE.minus(Duration.standardDays(1)),
Review comment:
My understanding is that it's trying to test setting a timer with
timestamp that exceed END_OF_GLOBAL_WINDOW and the fire timestamp gets altered
to the END_OF_GLOBAL_WINDOW(gc time). How the original exceeding-limit
timestamp is obtained shouldn't make a difference?
Since END_OF_GLOBAL_WINDOW is with millis granularity, advance the watermark
of the TestStream to this time will not be processed by dataflow.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
