How do you structure your hooks in your project?

[Speedydanzalez]

I've been using Reqnroll for a few months but never really dug into the hooks. I've watched a few videos and read the Reqnroll documentation but I don't see much in the way of how to structure them in a project and the high level approach to using them. Up until now I've been putting the before/after scenario hooks in the step def file and then just scope binding with the name of the feature, but I've seen that people have Hook folders with global hooks and then more business specific hooks perhaps.

Let's say I'm initialising a driver and tearing it down in each test, so then I create a global hook to handle this. I create 100 features that use this hook since it's global and will run for each test.

But now let's say I have a simple backend test that uses .NET to search for the existence of a file in a folder, this doesn't need a setup and teardown of a driver but the global hook will run for this test unnecessarily.

Okay, so now we use a tag or scope binding against the hook to tell it to run only for those specifics features. That means I now need to write 100 feature tag/scope bindings with that hook, just because that one backend test didn't require that hook.

Would like to get an idea about how people are structuring and approaching hooks in their projects. Any advice would be really appreciated. Cheers.

[gasparnagy]

I'm not sure I fully understand the situation, but if you have a driver that is initialized by Reqnroll using context injection, you can simply implement IDisposable and then Reqnroll will call the dispose after the end of the scenario if the driver was used.

The other approach that comes to my mind is to tag only the one that does not need the hook and add a filter condition to the global hook:

[AfterScenario]
public void ReleaseResources(ScenarioContext scenarioContext)
{
  if (scenarioContext.CombinedTags.Contains("skip-release-resources"))
    return;
  //...
}

[Speedydanzalez]

Thanks @gasparnagy.

Apologies, a driver was probably a bad example. For context 90% of my tests I have an application ("Launcher") that needs to be configured and launched at set up, this application is a landing page (or launcher) for other applications so it made sense to put it in a hook. Seeing as a lot of the time the application is simply an entry point, it shouldn't really be included in any Background or Given-When-Then steps unless the application is itself being tested.

So sometimes the test may not require this setup so we wouldn't want that hook to run in this case.

[gasparnagy]

So sometimes the test may not require this setup so we wouldn't want that hook to run in this case.

I see. I think then the exclusion with the if is the best you can do what I have mentioned in my comment above.

In Cucumber the hooks can have a "tag expression" as a condition, not just a simple tag list. So there you could say [AfterScenario("not @skip-release-resources")]. It is on my wish list to support this in Reqnroll too... I think that would be helpful for you.