Since I wrote several different JUnit runners (and it is very painful experience unfortunately...) I can tell you that you have 2 options:
1. Write your own plugin. Most robust but requires lots of work and it will bind you to specific IDE. This way you can fully control how the feedback from scenarios execution is provided. You can even provide a way of running a text scenario without creating a corresponding Scenario class.
2. Write jUnit runner. This way you can control how 'TestCases' are created and you can create a TestCase per scenario, then a test method per step. This way all IDEs will show very nicely which scenario failed, which step failed, etc.
Since I wrote several different JUnit runners (and it is very painful experience unfortunately...) I can tell you that you have 2 options:
1. Write your own plugin. Most robust but requires lots of work and it will bind you to specific IDE. This way you can fully control how the feedback from scenarios execution is provided. You can even provide a way of running a text scenario without creating a corresponding Scenario class.
2. Write jUnit runner. This way you can control how 'TestCases' are created and you can create a TestCase per scenario, then a test method per step. This way all IDEs will show very nicely which scenario failed, which step failed, etc.