V2.0: Pass output from the step definition on to the gherkin formatters (JSON formatter) #706
Conversation
The legacy_formatter need also to make sure that the puts calls from the step definitions are not passed to the formatter until it got information about the corresponding step.
|
You've been busy @brasmusson! |
|
@mattwynne Yep, and basically all of it has been spin-offs from #700. The back story there is that people are building third party tools that use the JSON output. And that a plugin is originally named "cucumber-jvm reports plugin", does not stop users from using it with Cucumber(-Ruby), and complain when it does not work. The writing on the wall I see, is that to support the Cucumber eco-system best, the JSON output should be treated as a first class API, so that given a feature file, regardless of it it executed with Cucumber(-Ruby), Cucumber-JVM or Cucumber-JS, the exact same JSON file should be produced - when using the JSON formatter that is. |
|
This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs. |
The output from the step definition is not passed on in the GherkinFormatterAdapter, which cases the output to be missing the the generated JSON report.
This RP fixes that. To make it work the legacy_formatter delays the calls to puts from the step definitions until the formatter has been informed of the corresponding step (similar to the handling of embed calls in #705).
The corresponding PR for v1.3.x is #701