Jiayi-Liao commented on pull request #16744: URL: https://github.com/apache/flink/pull/16744#issuecomment-895030964
> @Jiayi-Liao , let me repeat my question if I understand the problem correctly: > > > This is basically the same bug as https://issues.apache.org/jira/browse/FLINK-9857 but this time in test code? > > That the `TestProcessingTimeService` is firing the timers 1ms too soon, but unfortunately a lot of tests depend on that behaviour? If so, I would prefer to fix it, as that would be simply a bug in tests that can come and bite us in the future anyway? > > I see that it requires a lot of modifications, but have you tried to automise it? If you are using IntelliJ (or some other clever IDE) this trick might work: > > 1. create a temporary method that will increase `time + 1`: > > ``` > public void setProcessingTime(long time) throws Exception { > setProcessingTimeV2(time + 1); > } > > public void setProcessingTimeV2(long time) throws Exception { > processingTimeService.setCurrentTime(time); > } > ``` > > 1. Fix a couple of tests that were passing `MAX_VALUE` > 2. Inline `setProcessingTime(time)` > 3. Rename `setProcessingTimeV2(time)` back to `setProcessingTime(time)` > > ? > > Btw, if this change is really that big, then having a separate ticket for it would be a bit better. @pnowojski Yes. I've tried your way before, but the real situation is very complex and this way can only fix part of the failed tests. In most tests which are invoking `setCurrentTime(long)` for several times, the fix doesn't work. (e.g. StreamSourceOperatorWatermarksTest.java) -- 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. To unsubscribe, e-mail: issues-unsubscr...@flink.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
