Skip to content

Latest commit

 

History

History
492 lines (352 loc) · 18.9 KB

github_projects_guidelines.md

File metadata and controls

492 lines (352 loc) · 18.9 KB
Raw

GitHub project guidelines 📚

This document describes guidelines and recommendations that should be used when working with projects hosted on GitHub.

Table of Contents 📜

Repository naming scheme 📝

General rules

  • Names are written in lowercase letters.
  • Names consist of several fields separated by dashes, which means that the fields themselves can not contain dashes.
  • Some fields are mandatory, and some of them are optional, as per project requirements.
  • Underscores are not allowed.

Naming scheme

Use this scheme when naming GitHub repositories:

{client}-{project}-[{specifier}-]{repo_type}

Explanation of fields:

  • client - Client's name, such as irnas, fabrikor, companyinc.
  • project - Project name, such as controller, beacon, robot.
  • specifier - Used to specify chip name, sensor name, or similar, such as nrf92, nrf52, scp40. This field is optional. Use only when specifying extra information that distinguishes codebases, such as two MCUs on the same PCB.
  • repo_type - Possible options are:
    • firmware - C/C++ project for microcontrollers, projects for RPi,
    • driver - C/C++ project for a sensor, communication module driver,
    • hardware - PCB hardware project,
    • mechanical - mechanical CAD project,
    • software - software project, for example, Python tooling or a web app,
    • application - Android/iOS application project,
    • docs - Documentation project.
    • master - Master repo which ties together all other repos.

It can happen that your new GitHub project does not fit the preceding repo_type options. In that case, describe your repo_type with one, short word.

When creating fields client and project for a new project long names should be avoided, the soft limit is around 12 characters. Character shortening techniques such as abbreviations, internal codenames and concatenations of multiple words are allowed to satisfy the limit, however, overall clarity should not be compromised.

Some examples include:

  • irnas-blebeacon-firmware
  • irnas-userapp-application
  • irnas-lis2-driver
  • fabrikor-3dprinter-nrf91-firmware
  • fabrikor-3dprinter-nrf52-firmware
  • irnas-guidelines-docs

Versioning 1️⃣0️⃣0️⃣

The versioning scheme consists of a letter v, followed by 3 numbers separated by dots: v{major}.{minor}.{bugfix}

The following versions are all valid examples:

  • v0.1.4
  • v0.9.3
  • v0.9.10
  • v1.0.0
  • v1.1.0
  • v1.12.45
  • v5.0.0

Numbering was inspired by SemVer convention. We deviate from their rules on version incrementing, as the SemVer convention makes more sense when you are developing libraries that are consumed by the developers and not in the sense of our product development.

Starting version and how it is incremented depends on the type of the project.

Software/Firmware projects

Given the preceding versioning scheme increment:

  1. amajor number when releasing a new, initial stable version of a product (v1.0.0), making any conceptual changes, major rewrite, major documentation changes, a new generation of a product or any other change which requires additional human involvement.
  2. a minor number when adding new features, enhancements, and documentation in a backwards-compatible manner and
  3. a bugfix number when you make backwards-compatible bug fixes.

New projects should start with version v0.1.0 and continue from there.

Hardware projects

Given the preceding versioning scheme increment:

  1. a major number for major layout and schematic changes such as changes to board dimensions, components placement or general functionality and
  2. a minor number for minor layout and schematic changes such as component value changes, layout routing, copper fills, etc.

The number bugfix is never incremented and is always set to 0. New projects should start with version v1.0.0 and continue from there.

Mechanical projects

Given the preceding versioning scheme increment:

  1. a major number for major changes such as changes to dimensions, components placement or general functionality and
  2. a minor number for minor changes such as fixes, etc.

The number bugfix is never incremented and is always set to 0. New projects should start with version v1.0.0 and continue from there.

Changelog 📋

A changelog is a file that contains a curated, chronologically ordered list of notable changes for each version of a project. Changelog makes it easier for users and engineers to see precisely what notable changes have been made between each version of the project.

IRNAS's changelog format is based on the Keep a Changelog's format, we follow it is almost to a point with some minor additions and modifications.

This means that:

  • We follow its overall structure in markdown, which can be seen at the top of the page.
  • We follow its guiding principles, however, we define our versioning scheme.
  • Each version entry needs to be linkable, where the link points to a page showing the comparison between that version and the previous one.
  • Each bullet line should finish with a period dot.

The same types of changes should be grouped under one of the following groups:

  • Added for new features,
  • Changed for changes in existing functionality,
  • Deprecated for soon-to-be removed features,
  • Removed for now removed features,
  • Fixed for any bug fixes and
  • Security in case of vulnerabilities.

Content of changelog notes

Changelog notes can be seen as a common point between development and business.

Below are a few points that will help you construct clear and concise changelog notes:

  • Use plain language, without technical jargon. Write your release notes like you are explaining them to a friend.
  • Keep them short.
  • If you’ve implemented a new feature, your changelog notes should contain a high-level summary of what it can do. But, of course, some more explanation on how to use it might be required, so provide a link to the detailed stuff, like a user guide, step-by-step instructions, etc.

In general, your changelog notes should answer the following questions:

  • What has changed in the latest version of your product?
  • Why has that thing changed?
  • How does this change impact the user?
  • What does the user need to do differently as a result?

Keep in mind that in most cases a project manager will notify a customer about the new release and will also need to create a customer-understandable abstract from the changelog notes. Write changelog notes in such a way that is easy to write an abstract that communicates to the customer what value a specific the release brings to him.

Zephyr drivers

When creating a version entry for a Zephyr driver add a section which says for which NCS version was the driver built. This information should be visible to the developer who is deciding which version of the driver to use so that it will work with its NCS version of the project. Use the below template:

### Compatibility

- This release was built and tested on the NCS version <version>.

Releases 🚀

Taken from https://docs.github.com/en/repositories/releasing-projects-on-github/about-releases:

Releases are deployable software iterations you can package and make available for a wider audience to download and use.

Releases are based on Git tags, which mark a specific point in your repository's history. A tag date may be different from a release date since they can be created at different times.

GitHub release consists of a git tag, some text (commonly referred to as Release Notes) and artefacts (binaries, executables, any kind of documentation, etc.)

Tagging and naming releases

Release tags and names consist of the letter v followed by a version number.

The below tags are all valid examples:

  • v0.1.4
  • v0.9.3
  • v1.0.0
  • v1.1.0
  • v1.12.45
  • v5.0.0

The name of a GitHub release should be identical to the tag that the release was created from.

Release notes

Release notes are identical to the changelog notes for that specific version of a project and should be always kept in sync.

Creating releases

Since we are following the [trunk-based development] branching model, a GitHub release is created directly from a commit that was tagged with the release tag.

This process is automated with GitHub Actions with the "Basic" group of workflow files. More about this here.

Release artefacts naming scheme 📦

Important: this section is relevant only for software and firmware projects as the release process for mechanical and hardware projects still yet needs to be defined.

General rules

  • Names are written in lowercase letters.
  • Names consist of several fields separated by dashes, which means that the fields themselves can not contain dashes.
  • Some fields are mandatory and some are optional, as per project requirements.
  • Underscores are not allowed.

Naming scheme

The naming scheme that should be used for release objects:

{project}-{firmware_type}-{board_name}-{hardware_version}-{version}-{qualifiers}.{file_extension}

Fields project and firmware_version are the only mandatory ones, others should be added to avoid any confusion when dealing with generated files. If the repo_type of your GitHub project is firmware then the hardware_version the field is mandatory.

Explanation of fields:

  • project - Project name, such as blebeacon or tracker
  • firmware_type - The type of firmware, needs to be added if the build system produces applications and bootloader firmware. It can be app (for application firmware or software) or bl (for bootloader firmware).
  • board_name - If a project supports multiple boards, this field should be used to distinguish between different hardware boards, such as, VYPER_GO and VYPER_30.
  • hardware_version - Hardware version of the board which consists of hv and a version number. Hardware versions such as hv1.2.0, hv4.0.1 or hv0.5.1 are all valid options.
  • version - Version of the software/firmware which consists of v and a version number. Versions such as v1.2.0, v4.0.1 or v0.5.1 are all valid options.
  • qualifiers - Optional field, can be repeated. See the explanation below.
  • file_extension - Depends on a generated object, could be bin, elf, hex or something else.

Qualifiers

Qualifiers come in several forms:

  • Misc qualifiers - examples: log, dbg, dbgble, rf
  • 7 char Git hash - examples: 57fb962, a982467, 6b3089c

Misc qualifiers

These are special qualifiers that indicate that some special set of build flags was used to build a release artefact. Release artefacts that are meant to be used in the production do not contain any misc qualifiers.

Example scenario: you could be developing firmware that is used:

  • in production, with debug logs turned off,
  • in development, with debug logs turned on and
  • in RF compliance tests, where the device behaves completely differently for testing purposes.

In that scenario, production artefact would have no misc qualifier, development the artefact could have log qualifier and RF compliance artefact would have rf qualifier.

A project that uses misc qualifiers should have its meaning and usage documented in a visible place, such as the project's README.

Git hash

Git hash qualifiers are useful for internal testing processes of the product and where later identification is required. The version that precedes the qualifier should be a version of the release that was already released.

Important: release artefacts should never contain git hash qualifiers.

Valid release objects names

Below release names are all valid examples:

Simple firmware project:

  • GitHub repo name: irnas-blebeacon-firmware
  • Release artifacts: blebeacon-hv1.4.0-v1.3.3.hex

Firmware project with application and bootloader firmware:

  • GitHub repo name: irnas-robot-firmware
  • Release artefacts:
    • robot-app-hv1.4.0-v1.0.0.hex
    • robot-bl-hv1.0.0-v1.0.0.hex

Firmware project with application and bootloader firmware, and various sets of build flags:

  • GitHub repo name: irnas-largerobot-firmware
  • Release artifacts:
    • largerobot-app-hv1.4.0-v1.0.0.hex
    • largerobot-bl-hv1.0.0-v1.0.0.hex
    • largerobot-app-hv1.4.0-v1.0.0-log.hex
    • largerobot-bl-hv1.0.0-v1.0.0-log.hex
    • largerobot-app-hv1.4.0-v1.0.0-rf.hex
    • largerobot-bl-hv1.0.0-v1.0.0-rf.hex

GitHub labels management 🏷️

What are GitHub labels? From GitHub's docs:

You can manage your work on GitHub by creating labels to categorize issues, pull requests, and discussions. You can apply labels in the repository the label was created in. Once a label exists, you can use the label on any issue, pull request, or discussion within that repository.

To organize and categorize issues we use a set of labels that are defined in the irnas-project-template repository.

When creating a new project labels are automatically transferred if the new the the project used the irnas-project-template repository as a template.

If you want to use the labels in a project that was not created from the above the template then read the next section.

Usage

Below are some guidelines regarding the usage of labels:

  • An Issue/PR without labels should not require labels to attract attention, therefore the default state should be label-less.
  • Most of the labels have prefixes that organize them into groups:
    • priority - describes the immediacy of the attention required.
    • state - describes the decision state of the issue or pull request.
    • type - describes the type of the issue or pull request.
  • Issue should have a maximum of one label per group.
  • Some labels do not have prefixes, specifically pull request and release as they are created by the templates/CI.

Please note that the above guidelines are exactly that, guidelines, and not rules. If your project requires a different set of labels, uses additional groups, etc. then feel free to break the guidelines.

Documentation 📖

There are two general places where project-related documentation can exist:

  • GitBook
  • Github

This section does not try to address documentation in the source code. That is a separate topic.

Gitbook

Documentation on the Gitbook is meant to be mostly written by the management.

Things like:

  • project-related notes and ideas,
  • mockups,
  • project timeline projections,
  • meeting notes,
  • concepts and
  • field research notes

all belong on the Gitbook.

Engineers and developers can also write to the GitBook due to project-specific requests.

Github

Documentation on GitHub is meant to be mostly written by engineers and developers. It is written in the form of markdown files either in README.md or in the docs folder.

Things like:

  • setting up the build environment,
  • instruction on how to create a release,
  • customer-facing technical documentation,
  • getting started guide,
  • API documentation,
  • architecture diagrams,
  • block and flow diagrams and
  • power consumption reports

all belong on GitHub.

When looking at the documentation from a point of a specific release, the documentation should be up to date and should reflect the behaviour and implementation of the project at that point.

Resources 🤓

The below section is a hot mess of links to various topics related to this document, embedded systems and more.

Branching models/strategies/Git related:

Versioning:

Release naming:

Changelog:

Awesome projects: