Cucumber-JVM is built with Maven.
mvn clean install
File -> Open Project -> path/to/cucumber-jvm/pom.xml
Your .feature
files must be in a folder that IDEA recognises as source or test. You must also tell IDEA to copy your .feature
files to your output directory:
Preferences -> Compiler -> Resource Patterns -> Add `;?*.feature`
If you are writing step definitions in a scripting language you must also add the appropriate file extension for that language as well.
Just load the root pom.xml
To hack on Cucumber-JVM you need a JDK, Maven and Git to get the code. You also need to set your IDE/text editor to use:
- UTF-8 file encoding
- LF (UNIX) line endings
- No wildcard imports
- Curly brace on same line as block
- 4 Space indent (no tabs)
- Java
- XML
- 2 Space indent (no tabs)
- Gherkin
Please do not add @author tags - this project embraces collective code ownership. If you want to know who wrote some code, look in git. When you are done, send a pull request. If we get a pull request where an entire file is changed because of insignificant whitespace changes we cannot see what you have changed, and your contribution might get rejected.
All Cucumber implementations (cucumber-ruby, cucumber-jvm, cucumber-js) share a common set of Cucumber features to ensure all implementations support the same basic features. To run these you need to clone the cucumber-tck repo into your cucumber-jvm working copy:
git submodule update --init
Now you can run the cross-platform Cucumber features:
gem install bundler
bundle install
rake
Below are some common problems you might encounter while hacking on Cucumber-JVM - and solutions.
This can be solved by changing the Compiler settings: Preferences -> Compiler -> Java Compiler
:
- Use compiler:
Javac
- Additional command line parameters:
-target 1.6 -source 1.6 -encoding UTF-8
You should also use JDK 6: Project Structure... -> Project -> Project SDK
:
- *Use 1.6", not 1.7
If you still have problems, try building the project with Maven, using Java 6:
export JAVA_HOME=/System/Library/Java/JavaVirtualMachines/1.6.0.jdk/Contents/Home
Note that even though development is sometimes easier to do with 1.6, releasing should be done with 1.7.
This is a reminder to the developers:
First, make sure you have the proper keys set up - in your ~/.m2/settings.xml
- for example:
<settings>
<servers>
<server>
<id>cukes.info</id>
<username>yourcukesinfouser</username>
<privateKey>fullkeypath</privateKey>
</server>
<!-- See https://docs.sonatype.org/display/Repository/Sonatype+OSS+Maven+Repository+Usage+Guide -->
<server>
<id>sonatype-nexus-snapshots</id>
<username>yoursonatypeuser</username>
<password>TOPSECRET</password>
</server>
<server>
<id>sonatype-nexus-staging</id>
<username>yoursonatypeuser</username>
<password>TOPSECRET</password>
</server>
</servers>
</settings>
Replace version numbers in:
- examples/java-helloworld/build.xml
- examples/java-helloworld/pom.xml
- examples/java-gradle/build.gradle
- README.md (this file)
- History.md
Run git commit -am "Release X.Y.Z"
Now release everything:
mvn release:clean
mvn --batch-mode -P release-sign-artifacts release:prepare -DautoVersionSubmodules=true -DdevelopmentVersion=1.1.3-SNAPSHOT
mvn -P release-sign-artifacts release:perform
Post release the API docs must be generated for each module and manually copied over to a working copy of the cucumber.github.com which must be a sibling of cucumber-jvm
(this repo):
./doc/genapi.sh
After that's done, commit and push cucumber.github.com
Code coverage is collected mainly to identify code that can be deleted or needs to be tested better. To generate a report, run:
COBERTURA_HOME=/some/where ./cobertura.sh
This technique to collect coverage for a multi-module Maven project is based on a blog post by Thomas Sundberg.