lots of exercises in java... from https://github.com/exercism/java

CONTRIBUTING.md 13KB

Table of Contents

Overview

This guide covers contributing to the Java track. If you are new, this guide is for you.

If, at any point, you're having any trouble, pop in the Gitter exercism/java room or the Gitter exercism/dev room for help.

For general guidelines about contributing to Exercism see the Exercism contributing guide.

Before Making Your Pull Request

Hi! Thanks for contributing to the Exercism Java track!

Before opening your pull request, please review the track policies and make sure your changes comply with them all. This helps us focus our review time on the more important aspects of your changes.

Also please only address one issue per pull request. This makes it easier for us to review it, and it means that if we request changes to the fix for one issue, it won't prevent to a fix for another issue being merged.

It's perfectly fine to have more than one pull request open at a time. In that case it's important to keep the work for each pull request on a separate branch to prevent unrelated commits being added to your pull request. This is good practice to do always, even if you only have one pull request open.

One last thing to note before you get started. When you fork the repository and you want to sync your fork, you can perform a git rebase. This is preferred over merging the changes because merging leads to a dirty commit history whereas performing a rebase adds in those changes without making extra commit messages. However, this is only preferred, so don't worry about it too much.

Contributing With Minimal Setup

First things first: by contributing to Exercism, you are making this learning tool that much better and improving our industry as a whole... thank you!!!

To submit a fix for an existing exercise or port an exercise to Java with the least amount of setup:

  1. Ensure you have the basic Java tooling installed: JDK 1.8+, an editor and Gradle 2.x.

(see exercism.io: Installing Java)

See GitHub Help: Forking. Use those instructions (in conjunction with the Git Basics doc) to:

  • "fork" a repository on GitHub;
  • install git;
  • "clone" a copy of your fork;
  • configure an "upstream remote" (in this case, exercism/java);
  • create a branch to house your work
  • Write the codes. Do your work on that branch you just created.

The Getting Familiar With the Codebase section, below, is an orientation.

  • Commit, push and create a pull request.

Something like:

   $ git add .
   $ git commit -m "(An intention-revealing commit message)"
   $ git push

The Git Basics doc has a section on commit messages that provides practical advice on crafting meaningful commit messages.

  • Verify that your work passes all tests. When you create a pull request (PR), GitHub triggers a build on Travis CI. Your PR will not be merged unless those tests pass.

Getting Familiar With the Codebase

There are two objectives to the design of this build:

  1. when a problem is built from within the exercism/java repo (i.e. when you, the contributor, are developing the exercise), the tests run against the reference solution;
  2. when a problem is built outside the exercism/java repo (when a participant is solving the exercise), the tests run against the "main" code.

This repo is a multi-project gradle build.

The exercises Module

This is the top-level module, contained in the exercises directory. It is a container for the problem sub-modules.

  • its build.gradle points the "main" sourceset to the reference solution.
  • its settings.gradle names each of the subprojects, one for each problem in the set.

The Problem Submodules

The exercises subdirectory contains all of the problem submodules. Each problem/submodule is a subdirectory of the same name as its slug.

  • its build.gradle names dependencies required to work that problem.
  • its README.md describes the exercise.

Each problem/submodule has three source sets:

  • src/test/java/ — a test suite defining the edges of the problem
  • .meta/src/reference/java/ — a reference solution that passes all the tests
  • src/main/java/ — starter source files, if required/desired (this directory usually only has a .keep file in it).

Advanced: Complete Local Setup

Prerequisites

Before you proceed, please ensure that you have jq (library that parses JSON) & ruby installed on your machine.

Debian Linux

sudo apt-get install jq ruby-full

Mac OSX

brew install jq ruby

If you are going to make significant contribution(s) to the track, you might find it handy to have a complete local install of exercism on your computer. This way, you can run the full suite of tests without having to create/update a PR.

The easiest way to achieve this is simply use the bin/journey-test.sh script. However, you may want to perform other tests, depending on what you are doing. You can do so by duplicating the setup performed by the bin/journey-test.sh script.

Tip: gradle clean before exercism fetch

If you exercism fetch after doing a build, the CLI will fail with the following error message:

$ exercism fetch java bob
2015/09/06 15:03:21 an internal server error was received.
Please file a bug report with the contents of 'exercism debug' at: https://github.com/exercism/exercism.io/issues

and if you review the logs of your x-api, you'll find:

127.0.0.1 - - [06/Sep/2015:15:20:56 -0700] "GET /v2/exercises/java/bob HTTP/1.1" 500 514949 0.2138
2015-09-06 15:21:01 - JSON::GeneratorError - source sequence is illegal/malformed utf-8:

This is because some files generated by the build can't be served from the x-api. This is by design: the CLI does not serve binaries. To fix this, simply make sure you do a clean in your exercism/java repo before you fetch:

cd ~/workspace/exercism/java/exercises
gradle clean
cd ~/workspace/exercism/exercises
exercism fetch java bob

Adding a New Exercise

The easiest way to add a new exercise to the Java track is to port an exercise from another track. That means that you take an exercise that has already been implemented in another language and you implement it in this track.

To add a completely new exercise you need to open a pull request to the problem specifications repository. Any completely new exercise needs to be added and accepted there before it can be added to the Java track.

There is a general Exercism guide for porting an exercise to a new language. Please review this before porting an exercise to the Java track.

See here for a list of exercises that have yet to be implemented on the Java track and can therefore be ported to this track. Please make sure no one else has a pull request open to implement your chosen exercise before you start.

It might also be a good idea to open a WIP pull request to make it clear to others that you are working on this exercise. This can just be a pull request with an empty commit that states which new exercise you're working on, with WIP (work in progress) in the title so that the maintainers know that it's not ready for review yet.

The Java specific details you need to know about adding an exercise are:

  • Please add an entry to the exercises array in config.json. You can find details about what should be in that entry here. You can also look at other entries in config.json as examples and try to mimic them.

  • Please add an entry for your exercise to settings.gradle. This should just be include 'exercise-name'. This list is in alphabetical order so please add your exercise so that it maintains this order.

  • Please add an exercise submodule for your exercise. See The Problem Submodules section for what needs to be in this. See the POLICIES doc for an explanation of when you need to add a starter implementation. The build.gradle file can just be copied from any other exercise submodule. The README.md file can be generated using configlet. You can do this by:

    1. Download configlet and put it somewhere in your PATH

    2. Clone the problem-specifications repository.

    3. Run configlet generate . --only name_of_new_exercise --spec-path path_to_problem_specifications from the root of this repository.

  • Check if there is canonical data for the exercise you're adding. This can be found at https://github.com/exercism/problem-specifications/tree/master/exercises/EXERCISE-SLUG/canonical-data.json. If there is canonical data for your exercise then you should follow this when making the tests. We aim to follow the canonical data as closely as possible in our tests to ensure thorough test coverage. If there is canonical data available you also need to create a file at exercises/exercise-slug/.meta/version specifying the canonical data version you have implemented (e.g. 1.0.0). The canonical data version can be found at the top of the canonical data file for that exercise. See other exercises, e.g. acronym, for an example version file.

  • Make sure you've followed the track policies, especially the ones for exercise added/updated.

Hopefully that should be enough information to help you port an exercise to the Java track. Feel free to open an issue or post in the Gitter exercism/java room if you have any questions and we'll try and answer as soon as we can.

Updating the READMEs

The README.md files are generated from the exercise descriptions in problem specifications. They need to be regenerated regularly so that any changes to the descriptions in problem specifications propagate to our READMEs. This can be done using configlet:

  1. Download configlet and put it somewhere in your PATH

  2. Clone the problem-specifications repository.

  3. Run configlet generate . --spec-path path_to_problem_specifications from the root of this repository.

Checking tests are up to date

The tests for each exercise should follow the canonical data in problem specifications as closely as possible. The canonical data can change quite regularly, in which case the canonical data version for that exercise will be updated.

We keep track of which version of the canonical data each exercise implements in a version file, for example: https://github.com/exercism/java/blob/master/exercises/two-fer/.meta/version. Not all exercises have canonical data in problem specifications. For those that don't we don't add a version file.

We have a script which can check if these version are up to date with the ones in problem specification. This script can be used to check if any version files, tests and reference implementations need updating.

To run this script:

  1. Clone the problem-specifications repository.

  2. Run ./scripts/canonical_data_check.sh -t . -s --spec-path path_to_problem_specifications from the root of this repository.