Hi everybody,

I wake up this thread because following this discussion and several point to point discussions, I gave a presentation @ EclipseCon France about the importance of "unit tests which are not launching the Workbench or an OSGi platform".

Here is the recording: https://www.youtube.com/watch?v=IGkFy2H-d60&index=33&list=PLy7t4z5SYNaRJff0KBMbubOaj8gevvML4

On Mon, Feb 1, 2016 at 7:43 PM, Mickael Istria <mistria@redhat.com> wrote:
On 02/01/2016 07:25 PM, Aurelien Pupier wrote:
Le 1/27/2016 11:53 AM, Mickael Istria a écrit :
Yes, but it requires an effort in moving the code to the right plugin or turning fragment into a bundle everytime one identifies something to reuse; whereas using bundles only doesn't require any effort from the "producer" side.
Considering this code as first citizen, it seems to me an acceptable effort ;-)
Do you put your actual "first citizen" business code into fragments too?

When running locally, I may want to run only unit tests first.  Without launching the OSGi platform it will be faster. It will also allow to use some tools such as Infinitest to have continuous feedback while developing.
I don't know much of Infinitest, but I believe it doesn't rely on how Maven runs tests. Eclipse already provide the ability to run a test class in a bundle in plain Java without starting the workbench; it's "Run As > JUnit Test" instead of "Run As > JUnit Plugin Test". I guess Infinitest can rely on that, can't it?
Yes Infinitest can rely on it but it means that you are able to launch as "JUnit test" and so not starting the workbench.
Yes, and if you're able to do it from the Eclipse IDE for your test class, isn't it enough?
So in your IDE, you can run "unit tests" without the workbench/Platform actually started, just the API available; and with Maven and tycho-surefire-plugin, you get a Platform/workbench started, like it will happen in real-life, to find real issues that wouldn't happen in plain Java environment.

Just to be clear, I'm not saying it's wrong to change that; I'm more wonderint how much profitable is it, what changing test structure would provide better than the current one does. The Infinitest story seems the only one "worth it" IMO, and it doesn't seem to be correlated to tycho-surefire vs maven-surefire.
InfiniTest or only launching test of a single plugin very fast (less than a second versus tens seconds) it is the difference between keeping concentrated on the task and have our brain switching to another idea.
It seems to me that only tests launched with surefire will be able and ensured - to run with Junit test.
I believe this sentence illustrates our divergence: you mainly think about tests as a tool for the developer, and they have to be fast for the developer to use them. That's right.
I see them as an armor for the application, I want them to find as many bugs possible in the actual environment where the code will actually run, and if it has a cost for the developer, so be it, as long as tests are still run and checked by continuous integration. That's right too.

Overall, both are right and useful. Depending on the test, you might want one or the other.

What's possible and that seems like the best thing to me is to have the unit-tests in a bundle that run fast with maven-surefire-plugin, and have integration tests in another bundle also running the unit-test suite (+ some other tests) with tycho-surefire-plugin. So you get the best of both: unit tests are run fast and keep running fast because they are automatically tested using plain JUnit and maven-surefire-plugin; and you also have them running as part of the integration tests to also detect bugs that depend on the Eclipse runtime mechanism.
Mickael Istria
Eclipse developer at JBoss, by Red Hat
My blog - My Tweets

jbosstools-dev mailing list

Aurelien Pupier
Senior Software Engineer in JBoss Fuse Tooling team