Clone wiki

org.openmarkov / Working_methodology

Working methodology

Getting started with Git

OpenMarkov's code is stored in Bitbucket using Git as the control version system.

Atlassian, the company that hosts the Bitbucket repositories, offers a tutorial that explains the basic operations with Git.

The basic operations are the following:

  • Commit generates a changeset locally, a version of the repository, so to say.
  • Push is used to submit these changesets to the central repository. The "Team Synchronize" of Eclipse shows the changesets waiting to be submitted; they are marked as "Outgoing". To upload your changes to the central repository, you will have to do both operations, Commit and Push.
  • Pull downloads the changesets from the remote repository to your local repository.
  • Update updates the copy in the local working directory with the changesets we pulled.

For simplicity, we will assume that the Git client you are using (in general, you will use the client included in your IDE) is configured to perform an update after every pull. For this reason in the rest of this document we will just say "pull" instead of "pull and update".

Therefore, when you wish to upload to Bitbucket's server the changes you have made, the correct workflow is as follows:

  1. Pull
  2. Resolve conflicts, if any
  3. Commit
  4. Push

Daily work with OpenMarkov

The first thing a developer should do every day before working on OpenMarkov's code is to download the latest changes from the central repository. To do this, just right-click on the Eclipse project and go to Team > Pull. The first time you do it you will need to specify the URL of the Bitbucket repository and your user information. Leave the box "Update after Pull" checked. If you wish to see the changesets you are going to download, click "Next"; otherwise, click "Finish".

Once you do this, there might be conflicts (i.e. a file that you have changed has been changed by some other user and submitted to the central repository). This conflicts will be marked in red in the "team synchronize view" (right click on the Java project and go to Team > Synchronize) Go through the files that have conflicts by double-clicking on them (and clicking "No" on the pop up). You'll see a comparison of the local and remote versions of the file so that you can download the changes in the remote version without losing your changes. Check that resulting code should compile and pass the jUnit tests. Once we are done with the conflict resolution, we have to right click on the files and choose "Mark as resolved".

Submiting code changes to the central repository

Before submitting any local change to the server, the developer must pull the latest changes from the server first. After doing so, she might need to resolve the conflicts as explained in the previous version. Once the latest changes have been pulled and all the conflicts resolved, the developer can proceed to commit (Team > Commit) the local changes. Please write a short but clear description of the changes contained in the changeset you are about to commit. Committing will generate a changeset that should immediately be pushed (Team > Push) to the central repository. In case some changeset is pushed since the last time a developer pulled and she has already committed a changeset, the push will fail. To solve this, a pull must be done, which will then cause a conflict with the commits of the developer. Select "rebase on tip" and resolve the conflicts (if any) comparing the local version with the commited versions of the conflicting file. Once the conflicts have been resolved, mark them as such, commit the changes done during the conflict resolution and push them all.

It is not allowed to force a push without having resolved the conflicts because it may damage the work of other developers.

Updated