Fixed bug where -getSnippet in clj.clj received only the step parameter.

[shaolang]

cucumber.runtime.UndefinedStepsTracker actually calls the Backend
instances with 2 parameters--Step and FunctionNameGenerator--but
Clojure's Backend implementation only accepts one.

Instead of testing generation of Clojure Snippet with a test-only
ClojureSnippet instance, ClojureSnippetTest now tests the real
implementation in clj.clj.

Prior this fix, cucumber-clojure throws the following exceptions:

Exception in thread "main" clojure.lang.ArityException: Wrong number of args (3) passed to: clj$-getSnippet
        at clojure.lang.AFn.throwArity(AFn.java:437)
        at clojure.lang.AFn.invoke(AFn.java:47)
        at cucumber.runtime.clj.Backend.getSnippet(Unknown Source)
        at cucumber.runtime.UndefinedStepsTracker.getSnippets(UndefinedStepsTracker.java:32)
        at cucumber.runtime.Runtime.getSnippets(Runtime.java:183)
        at cucumber.runtime.snippets.SummaryPrinter.printSnippets(SummaryPrinter.java:33)
        at cucumber.runtime.snippets.SummaryPrinter.print(SummaryPrinter.java:18)
        at cucumber.runtime.Runtime.printSummary(Runtime.java:125)
        at cucumber.runtime.Runtime.run(Runtime.java:114)
        at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
        at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
        at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)

[brasmusson]

@aslakhellesoy I have checked this PR and think it is OK to merge (and at the same time remove the now obsolete comment about keeping the snippet templates in clj.clj and the removed ClojureSnippet.java in sync).

• The getSnippet function of the Clojure Backend needs to conform to the signature of the Backend interface in core (even though the FunctionNameGenerator argument is not used, since Clojure step definitions does not have a function name separate from the step definition regexp).
• Testing the generated snippets in isolation using a copy of the snippet template in ClojureSnippet is faster (~0.3s instead of ~3s), but testing a copy instead of the real thing also has problems (including hiding this bug), so the solution is the PR seems like an improvement to me.

OK?

[aslakhellesoy]

Ok - go ahead and merge!

[brasmusson]

Merged. Thanks for your contribution @shaolang.

[shaolang]

@brasmusson You're welcome. :)

[lock]

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.