The builds and deployments are handled automatically by TravisCI. Builds are triggered whenever a pull request is made or a commit is made to the main branch. Deployments are triggered and made to Heroku whenever commits are made to the main branch. The frontend app is deployed at https://use-zoomie-roomies.herokuapp.com/ and the backend is deployed at https://zoomie-roomies.herokuapp.com/.
For instructions on how to build the frontend and backend apps locally, refer to the README's in the frontend/react and backend folders.
This repo serves as a template for repositories in this organization. The following information describes how the native features/workflows of Github can be customized to work in a scrum development process.
An issue is a unit of tracking work. Issues can be classified into different classes using labels. This can be used to classify issues in the scrum process as follows.
An epic is an issue with the label epic
. It represents a large story that can be broken into stories, which can be addressed over multiple sprints. An epic issue references its story issues as a list in its description. A Github action has been added to automatically check/uncheck the story list items when they get closed/reopened.
A story is an issue with the label story
. It may represents a new feature, or an enhancement to an existing feature. A story issue can be broken into sub tasks, which are added as a list in the description of the story issue. These sub task items can be checked manually by the developer to indicate completion.
A bug is an issue with the label bug
. It represents a problem with the existing code that needs to be fixed.
A question is an issue with the label question
. It represents a question raised by any one and that may get converted into other types of issues.
In addition to the standard labels above, you can add new labels to issues to classify them into different classes like documentation
, frontend
, etc, or to add metadata like duplicate
, invalid
etc.
A milestone groups issues that are expected to be delivered at some point in time. It also allows ordering (prioritizing) theses issues and tracking their progress (percentage of issues completed so far). In the scrum context, a milestone can be used as a sprint. So, you can create your sprints and give them names like Sprint1, Sprint2, etc. and set their due dates respectively.
A project is a kanban-style board that can aggregate a set of issues for any purpose. In the scrum context, we can create one project called Scrum Board
and choose its template to be Automated kanban with reviews
. (This will create a set of initial notes that you can delete).
The master
branch (sometimes called the main
branch) is the main branch used for releases. Other branches can be created too. For example, a branch called gh-pages
is often used as a website for the repository (for more information check this link). Other branches can be created to address the issues of the repository, one branch per issue (called an issue
branch). Such branches can then be used to create pull requests, where they get peer reviewed and eventually merged into the master
branch. For more information on branches, check this link.
A pull request is a request to merge commits from one branch to another. This is typically used to merge commits from an issue
branch into the master
branch. A pull request is how the process of peer review is carried. Reviewers can comment on the code changes to show approval or request changes (which will need to be addressed by additional commits to the issue
branch). When a CI/CD pipeline is configured for a repository (see below), it will run on any issue
branch that is part of a pull request. When the peer review process has concluded, the new commits can merged into the master
branch. The recommended merge option to choose is Squash and merge
, (i.e., squash all commits into a single commit), since it makes the repository's history simpler.
Tags can be used to mark release points in a repository's commit history. Typically, after some work goal has been achieved, with a set of commits, a tag (typically a version number like 1.0.0, 1.0.1, etc.) is pushed to the respository to mark this point. This results in the tag showing up in the repository's tags page.
An issue can be created from the issues
tab of a repository. An issue type (bug, story, epic, question) is first chosen then its corresponding template can be sufficiently filled.
The Product Owner goes frequently to the Scrum Board
project and clicks on the Add Cards
link to triage new issues in the repository to the board's To do
column, which acts here as the Product Backlog
. Product Owner can also add unbaked ideas to the To do
column as notes, which are placeholders that can later be converted into issues (right click to do that). Issues and notes can then be ordered in the To do
column to show their priorities.
The Scrum Master creates a new milestone and gives it a suitable name (e.g., Sprint1) and a due date. Then, in the Scrum Board
, issues from the top of the To do
column (assuming they have been ordered based on priority) can be assigned to that milestone and to the developers who will work on them.
Developers go to the Scrum Board
where they can filter it for the issues assigned to them in a given milestone. They can pick ones to work on by moving them to the In progress
column (this is important since this is not automated).
In the daily standup, the Scrum Master
can review progress by going to the Scrum Board
and filtering it by the current milestone (sprint). Developers can then reference issues in the various columns when they answer the usual standup questions, e.g., issues they work on (In progress
), finsihed (Done
) or yet to work on (To do
).
Before developers can work on an issue, they should checkout and pull the master
branch to ensure that they have all the latest commits locally. Then, they should create a new local issue
branch and name it issue-[number]
(replacing [number]
by the issue number). Several issue
branches can be created concurrently, one for each issue, but it is important to make them independent from each other by checking out the master
branch before creating each of them. This allows them to be pushed and merged independently from each other (and with the least conflicts).
Each issue
branch can accumulate commits to address the issue. When ready, it can then be pushed to a corresponding remote branch that can then be used to create a pull request into the master
branch. The pull request template needs to be filled at this point. Once created, a pull request can be reviewed by a peer reviewer who may request changes. These changes can be made using new commits in the local issue
branch that can subsequently be pushed to the corresponding remote issue
branch. When all peer reviews have concluded, the pull request can then be squash merged
into the master
branch (read more here), and the issue
branch can be deleted. If the pull request description includes the words fixes #[number]
(where [number]
is an issue number), the issue with that number will automatically be closed.
it is recommeded to not push commits to the master branch directly but to always go through a peer review process using an
issue
branch.
It is recommended to create periodic releases from a repository, at least at the end of each sprint but can be more frequent. These releases should be working versions of the component(s) being developed in the repository. To create such releases, a new tag representing a version number (e.g., 1.0.0) is added to the local master
branch then pushed to the remote master
branch. A new release can then be created in Github using this tag.
Every repository needs to have a way to build its artifacts headlessly. It is a good idea to run tests as part of such build. Instructions on how to build the components in a repository needs to be documented in the repository's README.md.
A repository can also be setup to build continuously whenever a commit is pushed to the master
branch by setting up a CI/CD script (e.g., Travis CI) in its root folder. Such script will configure the build environment (as a virtual machine) and invoke the build script on the branch. If the script fails for some reason, the committer will be notified to fix it. It is a good practice to add a build badge to the README.md file to visibly indicate the status of the last CI/CD build (Travis CI provides such badges).
The CI/CD script will also be run when a new pull request is created or when more commits are pushed to its linked issue
branch. Such build assures peer reviewers that the new commits when accepted will not break the build. In fact, a successful CI/CD build can be a prerequisute for peer reviewers to look at the changes.
When a tag is pushed to the master
branch, the CI/CD script will additionally deliver and/or deploy the built artifact(s). The script can also be configured to create a Github release based on the tag.