Re: Surefire at Java 9 and Java 10

2017-12-03 Thread Enrico Olivelli
Il dom 3 dic 2017, 17:34 Tibor Digana ha scritto: > 6 months is quite far away. > No reason to be afraid because I can check it nowadays and we will see how > much of the ITs is broken. > The problem in Java 9 was that Oracle broke backwards compatibility and > they did not give us a chance to re

Re: Surefire at Java 9 and Java 10

2017-12-03 Thread Tibor Digana
6 months is quite far away. No reason to be afraid because I can check it nowadays and we will see how much of the ITs is broken. The problem in Java 9 was that Oracle broke backwards compatibility and they did not give us a chance to re-introduce extensions (removed modules) back to JDK. I expecte

Re: Surefire at Java 9 and Java 10

2017-12-03 Thread Tibor Digana
Look, I do not want to force our contributors to rely on java 9 or java 10. They can use their default jdk. My strategy is to fallback to default jdk unless jdk.home system property is specified. I am fine with Jenkins multi-branch if it fails on java 9 before integrating a branch to master. It hap

Re: Surefire at Java 9 and Java 10

2017-12-03 Thread Enrico Olivelli
Tibor, There are IT tests which check for exaclty java.specification.version=9 How are we going to deal with them? It seems that a new major release will appear every 6 months IMHO it will be enough to drop that assertion Enrico Il dom 3 dic 2017, 13:21 Tibor Digana ha scritto: > As discussed

Surefire at Java 9 and Java 10

2017-12-03 Thread Tibor Digana
As discussed before about Java 9. I will try to run Surefire integration tests with JDK 9 and 10 on Linux and I will add JDK 8/9 to CI pipeline. I do not expect any bugs here. Maybe only enabling some modules for the ITs in worst case. Surefire supports Jigsaw already now in origin/master. There