kennknowles commented on PR #36615:
URL: https://github.com/apache/beam/pull/36615#issuecomment-3446809830
I would prefer a semantic tag that justifies why this is OK. For example,
- Maybe it doesn't validate the runner but just validates the SDK harness
so we can remove the tag.
- Maybe it is always the same whether it is batch or streaming, so it could
be `@Category(SameInBatchAndStreaming.class)`
But the problem is that the existence of a batch vs streaming mode is
actually quite under-the-hood. In fact it would be better for the execution
mode to be chosen at finer granularity than "whole job". And if there really is
a potentially different codepath in batch and streaming we can't safely disable
it.
The real solution to the issue is to combine multiple ValidatesRunner
pipelines into a single Dataflow job so that we can run 100 at a time or more.
This is just sort of hard to do with JUnit. Or at least it requires hackery
that I didn't figure out yet.
If it is a hack to make just Dataflow work better I would rather make it in
the Dataflow build.gradle than tag on the test, since it isn't a property of
the test.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
