Dry run is ignored for ConditionalPerformable

[nbarrett]

Ever since upgrading from 2.0.0.20 to 2.0.27 dry run is ignored, if a conditional performable exists at the top level. For instance in the following example, Navigate.toTaskThatDoesSomething() is always run, rather than being skipped as expected:

    @Test
    public void dryRunBroken() {
        System.setProperty("serenity.dry.run", "true");
        Actor.named("nick").attemptsTo(
            Check.whether(true)
                .andIfSo(Navigate.toTaskThatDoesSomething()));
    }

Whereas this version of the test would obey the dry run setting:

    @Test
    public void dryRunGood() {
        System.setProperty("serenity.dry.run", "true");
        Actor.named("nick").attemptsTo(
            Navigate.toTaskThatDoesSomething());
    }

[nbarrett]

Interestingly, a workaround using Task.where() works:

    @Test
    public void dryRunWorkaround() {
        System.setProperty("serenity.dry.run", "true");
        Actor.named("nick").attemptsTo(
            Task.where("Temporary fix to enable dry run to work correctly until https://github.com/serenity-bdd/serenity-core/issues/1503 fixed",
                       Check.whether(true)
                           .andIfSo(Navigate.toTaskThatDoesSomething())));
    }

But using SilentTask doesn't work:

    @Test
    public void dryRunWorkaroundDoesntWork() {
        System.setProperty("serenity.dry.run", "true");
        Actor.named("nick").attemptsTo(
            SilentTask.where("Temporary fix to enable dry run to work correctly until https://github.com/serenity-bdd/serenity-core/issues/1503 fixed",
                       Check.whether(true)
                           .andIfSo(Navigate.toTaskThatDoesSomething())));
    }
``

[wakaleo]

Hmm, DryRun isn't really an official feature (more for testing the framework) - can you propose a PR to reinstate it for tasks? Not sure what would have changed.

[nbarrett]

Yes, I understand your point about Dry run not being an official feature! However on more than one client project, I've used it very effectively to skip potentially time-consuming sections of data setup in tests which provides significant performance benefits. I've been meaning to propose a PR to add this 'skippable concept' to Serenity or maybe write an article about it because it's a really ground-breaking feature, allowing parts of tests to be completely skipped over like so:

The really cool thing is that the reporting already works perfectly and shows the parts that have been skipped. Of course 'leveraging' capabilities like this is risky as they can be broken by new releases and this has happened on more than one occasion especially around the area of Silent Performables... I'll see what I can do 😄