-
Notifications
You must be signed in to change notification settings - Fork 19
Add Support for Cucumber using @split[feature:treatment] tags #247
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
<dependency> | ||
<groupId>io.cucumber</groupId> | ||
<artifactId>cucumber-java</artifactId> | ||
<version>6.10.4</version> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This should probably be test scope, no?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
probably
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've thought a bit more about this. I think there are two options here - compile
scope or provided
scope. (See maven docs for details).
The junit
dependency is already compile
scope, so using the same scope for cucumber-java
and cucumber-junit
would be consistent with that.
Alternatively, making it provided
scope would indicate that dependent modules would have to explicitly add the dependencies themselves. I think that's ok too, if you don't want consumers of the library transitively download cucumber unless they opt into it with an explicit dependency in their own pom/gradle file.
I'll defer to you to decide what's most appropriate.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If someone is already using this framework (our Java-testing) without cucumber and I merge this PR. Regardless of the provided or not, I believe the build will fail for not finding the dependency in the classpath, right?
I need time to sync internally but if that's the case we may need to split this into a separate submodule.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I believe the build will fail for not finding the dependency in the classpath, right?
No it won't.
In Java, classes are only loaded when they are referenced by another class. Classes are not loaded by mere presence in a jar file.
No other classes in this library reference io.split.client.testing.cucumber.CucumberSplit
, so it won't be loaded.
@@ -0,0 +1,19 @@ | |||
@split[cappuccino:off] | |||
@split[dollars:on] | |||
Feature: Split |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Would you mind educating me how the unit test binds to the split listed in this file?
also, is the idea to have a single descritptor file or one per test class or test case?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm not sure what you mean by unit test or descriptor file. I'll clarify some of the terminology used by Cucumber (which is closer to an acceptance testing tool than a unit testing tool).
A .feature
file (sometimes called gherkin document) contains multiple scenarios (sometimes called examples).
Each Scenario is an acceptance test and consists of the following parts:
- One or more Steps (
Given
,When
,Then
). - Step Definitions, which are Java methods annotated with
@Given
,When
or@Then
.
Cucumber creates a new instance of every class with step definitions before the scenario runs. This is similar to how JUnit creates a new instance of a test class for every @Test
method it executes. However, the step definition is not a test - think of it as a method that is "invoked" by one of the scenario's steps.
For each step in the scenario, Cucumber finds a matching step definition (based on the expression on the annotation) and invokes the method.
Step definition classes aren't scoped to a particular .feature
file. The steps in a .feature
file can invoke step definitions from any step definition class.
Scenarios can have tags. The code in this library looks for tags of the format @split[feature:treatment]
and uses that to register the appropriate treatments. This is done on a new instance of SplitClientForTest
for every scenario.
public static void configureSplit(SplitClientForTest splitClient, Scenario scenario) { | ||
Collection<String> tags = scenario.getSourceTagNames(); | ||
for (String tag : tags) { | ||
Matcher matcher = SPLIT_TAG_PATTERN.matcher(tag); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Question here:
- if more than 1 tag matches this, do you end up overwriting the registered (feature, treatment) pair?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes. This is intentional.
Tags that are placed at the top level Feature
level are inherited by all the Scenario
s. The getSourceTagNames()
method returns a Collection
of the tags from both the Feature
and the current Scenario
. The inherited Feature
tags are always before the Scenario
tags in the collection (it would have been more appropriate to declare the return type as List
rather than Collection
since it's ordered).
We're taking advantage of this here to allow users to define "default" treatments at the feature level, and (optionally) override them at the scenario level. I have illustrated this in the last scenario where the default dollars=off
treatment from the feature level is overridden to dollars=on
for that particular scenario.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've pushed some more changes to (hopefully) illustrate better how this works.
@aslakhellesoy The only dependency with Split testing module is on the SplitClientForTest correct? just triple checking. |
That is correct. The only class I have added to the module is import io.cucumber.java.Scenario;
import io.split.client.testing.SplitClientForTest;
import java.util.Collection;
import java.util.regex.Matcher;
import java.util.regex.Pattern; |
hey @aslakhellesoy are you able to allow me to edit your fork pull request? I'd like to create a new module so you have complete isolation for iterations and so I can run a git checkout on this branch to do a release. |
Let's continue in #248 |
I've now moved this code to https://github.com/cucumber/split-java/ and added some documentation. Please remove the code I have contributed now that Cucumber-Split has a new home. |
Hi Splitsters,
This PR adds a small utility that would make it easier to use Split from Cucumber Scenarios. Here is a short example:
In order to enable the integration with Split, a Before Hook must be defined:
I hope you'll consider adding this to the
io.split.client:java-client-testing
library!Many thanks,
Aslak Hellesøy
Creator and lead developer of Cucumber Open