Batching Fails when Scenarios of same Feature are present in different files

Hi Team, If I put Scenarios belonging to the same Feature in separate files, then the batch execution fails. I tested this behaviour for serenity-core from 2.4.24 and 3.1.5 and its the same.

e.g.: Say there are two feature files Sceanrio1.feature and Scenario2.feature.

Scenario1.feature

Feature: Searching for a term

  Scenario: Searching for a term
    Given Sergey is researching things on the internet
    When he looks up "Cucumber"
    Then he should see information about "Cucumber"

Scenario2.feature

Feature: Searching for a term

  Scenario: Searching for a term again
    Given Sergey is researching things on the internet
    When he looks up "Serenity"
    Then he should see information about "Serenity"

when I execute it I get the error as stated in the below details.

gradlew clean test -Pdriver=firefox --info -Dserenity.batch.count=2 -Dserenity.batch.number=1

The error log

starter.CucumberTestSuite STANDARD_OUT
    14:29:44.334 [Test worker] INFO  i.c.junit.CucumberSerenityRunner - Running slice 2 of 2 using fork 1 of 1 from feature paths [file:/C:/Users/M1032630/Downloads/serenity-cucumber-starter-1.0.0/src/test/resources/features/]
    14:29:44.481 [Test worker] ERROR i.c.junit.CucumberSerenityRunner - Test failed to start
    java.lang.IllegalArgumentException: Can't find feature 'Searching for a term #2' in this slice
        at net.serenitybdd.cucumber.suiteslicing.WeightedCucumberScenarios.createFilterContainingScenariosIn(WeightedCucumberScenarios.java:69)
        at io.cucumber.junit.CucumberSerenityRunner.lambda$toPossibleFeatureRunner$5(CucumberSerenityRunner.java:328)
        at java.util.stream.ReferencePipeline$3$1.accept(ReferencePipeline.java:193)
        at java.util.stream.ReferencePipeline$2$1.accept(ReferencePipeline.java:175)
        at java.util.ArrayList$ArrayListSpliterator.forEachRemaining(ArrayList.java:1384)
        at java.util.stream.AbstractPipeline.copyInto(AbstractPipeline.java:482)

Observe that #⁠2 gets appended to the feature name java.lang.IllegalArgumentException: Can't find feature 'Searching for a term #2' in this slice

However if I put both the scenarios in a single file Scenarios.feature the batching and execution works perfectly fine.

Feature: Searching for a term

  Scenario: Searching for a term
    Given Sergey is researching things on the internet
    When he looks up "Cucumber"
    Then he should see information about "Cucumber"


  Scenario: Searching for a term again
    Given Sergey is researching things on the internet
    When he looks up "Serenity"
    Then he should see information about "Serenity"

For maintainability and reporting purpose, my project maintains each Scenario in a separate .feature file and all Scenarios of the same feature are kept in a directory.

Please suggest to fix this batching issue when maintaining scenarios of same feature in a differnet files.

[sudarsunperu]

Thanks @wakaleo , I will try to look into the Code and Provide a PR. Alternately, I will also look into parallel execution using Cucumber for gradle.

Because of OOM exceptions when running 350+ tests in the same machine, we used batching and executed each batch in a separate node. However I’m not sure if the same is possible using Cucumber parallel execution

[wakaleo]

Why don’t you just run the features in parallel using the Cucumber support for parallel execution?