I believe our usage is high enough on Java 8 that we wouldn't be close to
ceasing support for it.
I actually don't really know much about the possibility of distributing
multi-version artifacts for Java but that would probably be better than
what we do today. Today we just distribute Java 8 artifa
Thank you everyone for responding.
Replying to Kenn:
1. This explanation helps me understand more.
2. I realize the drive for the breaking change PR in question is to support
Java 17. Would a potential path be to target a more gentle ramp from 8, to
11, and up with a communicated and anticipated
Seems like https://github.com/apache/beam/pull/27606 fixes this (at least
for me locally on Java 11). +Liam Miller-Cushon
On Fri, Jul 21, 2023 at 9:51 AM Reuven Lax via dev
wrote:
> Curious why these failing tests didn't block submission.
>
Sounds like we don't run (all ?) unit tests for affec
Curious why these failing tests didn't block submission.
For now rollback seems to be the simplest option. However is there a path
forward on Java 11, or is our model irretrievably broken on Java 11?
On Fri, Jul 21, 2023 at 8:57 AM Kenneth Knowles wrote:
> This is a tricky situation that I don'
This is a tricky situation that I don't know how to resolve best. Here are
some pieces of information I know:
1. The reason we put certain generated classes in the same package is
because of Java's package-private access restriction. If they are in
another package the generated wrapper won't be ab
Good day, everyone,
For clarity, I organize the following into situation, background,
assessment, and proposal.
Best,
Damon
-
Situation
Issue #26981 reports an IllegalArgumentException associated with the
ByteBuddy dependency throwing the message " must
