Translations are garbled when using newer Java

Description

During development of 2.9.0 we ran into a problem of all translations being garbled when running Spark (they appear ok in the source on GitHub and IDE). One contributor found out this is due to some bug with unicode in Java 8. As we are still officially only supporting Java 8 in Spark and it is bundled with Java 8, we have applied the workaround patch provided by this contributor. It seems that this workaround is now causing Spark to garble translations when using it with Java newer than version 8.

I am not able to run Spark with Java 11 on Windows (probably because of how it is built with install4j), but using deb installer on Ubuntu and running with OpenJDK 11 i can reproduce this. Not sure if there is a way to make translations work with both Java 8 and newer. If not, for now we should keep the workaround and remove it when we stop support for Java 8.

Related tickets linked below.

Environment

None

Assignee

Unassigned

Reporter

wroot

Expected Effort

None

Components

Fix versions

Affects versions

Priority

Minor
Configure