I've just created a patch to update branch-2 to Java 8. Not sure what's
going to break, but hopefully all problems can be done with some selective
changes. We know that Hadoop branch-2 builds and runs on Java 8, it being
the java version we are already using for our branch-2 build and test runs.
It's time to move on; maintenance is getting harder and harder, as well as
more inconvenient. And while I know of clusters being deployed, they are,
AFAIK, branch-2 on java 8 JVMs. This is just changing the language to
reflect the reality.
Before anyone asks, yes, our compatibility guidelines cover this; we had
the same issue in the java 6 -> 7 move, remember?
---------- Forwarded message ---------
From: Steve Loughran (JIRA) <j...@apache.org>
Date: Thu, Mar 28, 2019 at 7:33 PM
Subject: [jira] [Created] (HADOOP-16219) Hadoop branch-2 to set java
language version to 1.8
To: <common-...@hadoop.apache.org>
Steve Loughran created HADOOP-16219:
---------------------------------------
Summary: Hadoop branch-2 to set java language version to 1.8
Key: HADOOP-16219
URL: https://issues.apache.org/jira/browse/HADOOP-16219
Project: Hadoop Common
Issue Type: Improvement
Components: build
Affects Versions: 2.10.0
Reporter: Steve Loughran
Java 7 is long EOL; having branch-2 require it is simply making the release
process a pain (we aren't building, testing, or releasing on java 7 JVMs
any more, are we?).
Staying on java 7 complicates backporting, JAR updates for CVEs (hello
Guava!) &c are becoming impossible.
Proposed: increment javac.version = 1.8
