:toc: macro :toclevels: 5 :figure-caption!: :git_rebase_workflow_link: link:https://www.alchemists.io/articles/git_rebase[Git Rebase Workflow] = Git Lint Git Lint is a command line interface for linting Git commits by ensuring you maintain a clean, easy to read, debuggable, and maintainable project history. Having a consistent commit history leads to improved code reviews and is a perfect companion to tools like link:https://www.alchemists.io/projects/milestoner[Milestoner] for versioning and producing automated release notes of your deploys. toc::[] == Features * Enforces a {git_rebase_workflow_link}. * Enforces a clean and consistent Git commit history. * Supports Git default branch configuration. * Provides a customizable suite of analyzers. * Provides Git Hook support for local use. * Provides Continuous Integration (CI) support. == Requirements . link:https://git-scm.com[Git] . link:https://www.ruby-lang.org[Ruby] == Setup To install, run: [source,bash] ---- gem install git-lint ---- == Usage === Command Line Interface (CLI) From the command line, type: `git-lint --help` .... USAGE: -a, --analyze [options] Analyze current branch commits. -c, --config ACTION Manage gem configuration. Actions: edit or view. -h, --help Show this message. --hook PATH Hook for analyzing unsaved commits. -v, --version Show gem version. ANALYZE OPTIONS: --sha HASH Analyze specific commit SHA. .... To check if your Git commit history is clean, run: `git-lint --analyze`. It will exit with a failure if at least one issue with error severity is detected. This gem does not check commits on your default branch (i.e. `main`). This is intentional as you would, generally, not want to rewrite or fix commits on the `main` branch. This gem is best used on feature branches as it automatically detects all commits made since creation of the feature branch. Here is an example workflow, using gem defaults with issues detected: [source,bash] ---- cd example git checkout -b test touch text.txt git add --all . git commit --message "This is a bogus commit message that is also terribly long and will word wrap" git-lint --analyze ---- Output: .... Running Git Lint... 83dbad531d84a184e55cbb38c5b2a4e5fa5bcaee (Brooke Kuhlmann, 0 seconds ago): This is a bogus commit message that is also terribly long and will word wrap. Commit Body Presence Warning. Use minimum of 1 line (non-empty). Commit Subject Length Error. Use 72 characters or less. Commit Subject Prefix Error. Use: /Fixed/, /Added/, /Updated/, /Removed/, /Refactored/. Commit Subject Suffix Error. Avoid: /\./, /\?/, /\!/. 1 commit inspected. 4 issues detected (1 warning, 3 errors). .... === Configuration This gem can be configured via a global configuration: .... $HOME/.config/git-lint/configuration.yml .... It can also be configured via link:https://www.alchemists.io/projects/xdg[XDG] environment variables. The default configuration is: [source,yaml] ---- :analyzers: :commit_author_capitalization: :enabled: true :severity: :error :commit_author_email: :enabled: true :severity: :error :commit_author_name: :enabled: true :severity: :error :minimum: 2 :commit_body_bullet: :enabled: true :severity: :error :excludes: - "\\*" - "•" :commit_body_bullet_capitalization: :enabled: true :severity: :error :includes: "\\-" :commit_body_bullet_delimiter: :enabled: true :severity: :error :includes: "\\-" :commit_body_tracker_shorthand: :enabled: true :severity: :error :excludes: - "(f|F)ix(es|ed)?\\s\\#\\d+" - "(c|C)lose(s|d)?\\s\\#\\d+" - "(r|R)esolve(s|d)?\\s\\#\\d+" :commit_body_leading_line: :enabled: false :severity: :warn :commit_body_line_length: :enabled: true :severity: :error :maximum: 72 :commit_body_paragraph_capitalization: :enabled: true :severity: :error :commit_body_phrase: :enabled: true :severity: :error :excludes: - "absolutely" - "actually" - "all intents and purposes" - "along the lines" - "at this moment in time" - "basically" - "each and every one" - "everyone knows" - "fact of the matter" - "furthermore" - "however" - "in due course" - "in the end" - "last but not least" - "matter of fact" - "obviously" - "of course" - "really" - "simply" - "things being equal" - "would like to" - "\\beasy\\b" - "\\bjust\\b" - "\\bquite\\b" - "as\\sfar\\sas\\s.+\\sconcerned" - "of\\sthe\\s(fact|opinion)\\sthat" :commit_body_presence: :enabled: false :severity: :warn :minimum: 1 :commit_body_single_bullet: :enabled: true :severity: :error :includes: "\\-" :commit_subject_length: :enabled: true :severity: :error :maximum: 72 :commit_subject_prefix: :enabled: true :severity: :error :delimiter: " " :includes: - Fixed - Added - Updated - Removed - Refactored :commit_subject_suffix: :enabled: true :severity: :error :excludes: - "\\." - "\\?" - "\\!" :commit_trailer_collaborator_capitalization: :enabled: true :severity: :error :commit_trailer_collaborator_duplication: :enabled: true :severity: :error :commit_trailer_collaborator_email: :enabled: true :severity: :error :commit_trailer_collaborator_key: :enabled: true :severity: :error :includes: - "Co-Authored-By" :commit_trailer_collaborator_name: :enabled: true :severity: :error :minimum: 2 ---- Feel free to take this default configuration, modify, and save as your own custom `configuration.yml`. ==== Enablement By default, most analyzers are enabled. Accepted values are `true` or `false`. If you wish to disable a analyzer, set it to `false`. ==== Severity By default, most analyzers are set to `error` severity. If you wish to reduce the severity level of a analyzer, you can set it to `warn` instead. Here are the accepted values and what each means: * `warn`: Will count as an issue and display a warning but will not cause the program/build to fail. Use this if you want to display issues as reminders or cautionary warnings. * `error`: Will count as an issue, display error output, and cause the program/build to fail. Use this setting if you want to ensure bad commits are prevented. ==== Regular Expressions Some analyzers support _include_ or _exclude_ lists. These lists can consist of strings, regular expressions, or a combination thereof. Regardless of your choice, all lists are automatically converted to regular expression for use by the analyzers. This means a string like `"example"` becomes `/example/` and a regular expression of `"\\AExample.+"` becomes `/\AExample.+/`. If you need help constructing complex regular expressions for these lists, try launching an IRB session and using `Regexp.new` or `Regexp.escape` to experiment with the types of words/phrases you want to turn into regular expressions. _For purposes of the YAML configuration, these need to be expressed as strings with special characters escaped properly for internal conversion to a regular expression._ === Analyzers The following details the various analyzers provided by this gem to ensure a high standard of commits for your project. ==== Commit Author Capitalization [options="header"] |=== | Enabled | Severity | Defaults | true | error | none |=== Ensures author name is properly capitalized. Example: .... # Disallowed jayne cobb dr. simon tam # Allowed Jayne Cobb Dr. Simon Tam .... ==== Commit Author Email [options="header"] |=== | Enabled | Severity | Defaults | true | error | none |=== Ensures author email address exists. Git requires an author email when you use it for the first time too. This takes it a step further to ensure the email address loosely resembles an email address. .... # Disallowed mudder_man # Allowed jayne@serenity.com .... ==== Commit Author Name [options="header"] |=== | Enabled | Severity | Defaults | true | error | minimum: 2 |=== Ensures author name consists of, at least, a first and last name. Example: .... # Disallowed Kaylee # Allowed Kaywinnet Lee Frye .... ==== Commit Body Bullet [options="header"] |=== | Enabled | Severity | Defaults | true | error | excludes: `["\\*", "•"]` |=== Ensures commit message bodies use a standard Markdown syntax for bullet points. Markdown supports the following syntax for bullets: .... * - .... It's best to use dashes for bullet point syntax as stars are easier to read when used for _emphasis_. This makes parsing the Markdown syntax easier when reviewing a Git commit as the syntax used for bullet points and _emphasis_ are now, distinctly, unique. ==== Commit Body Bullet Capitalization [options="header"] |=== | Enabled | Severity | Defaults | true | error | includes: `["\\-"]` |=== Ensures commit body bullet lines are capitalized. Example: .... # Disallowed - an example bullet. # Allowed - An example bullet. .... ==== Commit Body Bullet Delimiter [options="header"] |=== | Enabled | Severity | Defaults | true | error | includes: `["\\-"]` |=== Ensures commit body bullets are delimited by a space. Example: .... # Disallowed -An example bullet. # Allowed - An example bullet. .... ==== Commit Body Leading Line [options="header"] |=== | Enabled | Severity | Defaults | true | error | none |=== Ensures there is a leading, empty line, between the commit subject and body. Generally, this isn't an issue but sometimes the Git CLI can be misused or a misconfigured Git editor will smash the subject line and start of the body as one run-on paragraph. Example: .... # Disallowed Curabitur eleifend wisi iaculis ipsum. Pellentque morbi-trist sentus et netus et malesuada fames ac turpis egestas. Vestibulum tortor quam, feugiat vitae, ultricies eget, tempor sit amet, ante. Donec eu_libero sit amet quam egestas semper. Aenean ultricies mi vitae est. Mauris placerat's eleifend leo. Quisque et sapien ullamcorper pharetra. Vestibulum erat wisi, condimentum sed, commodo vitae, orn si amt wit. # Allowed Curabitur eleifend wisi iaculis ipsum. Pellentque morbi-trist sentus et netus et malesuada fames ac turpis egestas. Vestibulum tortor quam, feugiat vitae, ultricies eget, tempor sit amet, ante. Donec eu_libero sit amet quam egestas semper. Aenean ultricies mi vitae est. Mauris placerat's eleifend leo. Quisque et sapien ullamcorper pharetra. Vestibulum erat wisi, condimentum sed, commodo vitae, orn si amt wit. .... ==== Commit Body Line Length [options="header"] |=== | Enabled | Severity | Defaults | true | error | maximum: 72 |=== Ensures each line of the commit body doesn't force you to scroll horizontally to read the message. This allows commit messages to remain readable and is especially handy when commit messages are read via the command line or email clients. ==== Commit Body Paragraph Capitalization [options="header"] |=== | Enabled | Severity | Defaults | true | error | none |=== Ensures each paragraph of the commit body is capitalized. Example: .... # Disallowed curabitur eleifend wisi iaculis ipsum. # Allowed Curabitur eleifend wisi iaculis ipsum. .... ==== Commit Body Phrase [options="header"] |=== | Enabled | Severity | Defaults | true | error | excludes: (see configuration) |=== Ensures non-descriptive words/phrases are avoided in order to keep commit message bodies informative and specific. The exclude list is case insensitive. Detection of excluded words/phrases is case insensitive as well. Example: .... # Disallowed Obviously, the existing implementation was too simple for my tastes. Of course, this couldn't be allowed. Everyone knows the correct way to implement this code is to do just what I've added in this commit. Easy! # Allowed Necessary to fix due to a bug detected in production. The included implementation fixes the bug and provides the missing spec to ensure this doesn't happen again. .... ==== Commit Body Presence [options="header"] |=== | Enabled | Severity | Defaults | false | warn | minimum: 1 |=== Ensures a minimum number of lines are present within the commit body. Lines with empty characters (i.e. whitespace, carriage returns, etc.) are considered to be empty. Automatically ignores _fixup!_ commits as they are not meant to have bodies. ==== Commit Body Single Bullet [options="header"] |=== | Enabled | Severity | Defaults | true | error | includes: `"\\-"` |=== Ensures a single bullet is never used when a paragraph could be used instead. Example: .... # Disallowed - Pellentque morbi-trist sentus et netus et malesuada fames ac turpis egestas. Vestibulum tortor quam, feugiat vitae, ultricies eget, tempor sit amet, ante. Donec eu_libero sit amet quam. # Allowed Pellentque morbi-trist sentus et netus et malesuada fames ac turpis egestas. Vestibulum tortor quam, feugiat vitae, ultricies eget, tempor sit amet, ante. Donec eu_libero sit amet quam. .... ==== Commit Body Tracker Shorthand [options="header"] |=== | Enabled | Severity | Defaults | true | error | excludes: (see configuration) |=== Ensures commit body doesn't use issue tracker shorthand. The exclude list defaults to GitHub Issues but can be customized for any issue tracker. There are several reasons for excluding issue tracker links from commit bodies: . Not all issue trackers preserve issues (meaning they can be deleted). This makes make reading historic commits harder to understand why the change was made when the reference no longer works. . When disconnected from the internet or working on a laggy connection, it's hard to understand why a commit was made when all you have is a shorthand issue reference with no supporting context. . During the course of a repository's life, issue trackers can be replaced (rare but does happen). If the old issue tracker service is no longer in use, none of the commit body shorthand will be of any relevance. Instead of using tracker shorthand syntax, take the time to write a short summary as to _why_ the commit was made. Doing this will make it easier to understand _why_ the commit was made, keeps the commit self-contained, and makes learning about/debugging the commit faster. ==== Commit Subject Length [options="header"] |=== | Enabled | Severity | Defaults | true | error | maximum: 72 |=== Ensures the commit subject length is no more than 72 characters in length. This default is more lenient than the link:http://tbaggery.com/2008/04/19/a-note-about-git-commit-messages.html[50/72 rule] as it gives one the ability to formulate a more descriptive subject line without being too wordy or suffer being word wrapped. Automatically ignores _fixup!_ or _squash!_ commit prefixes when calculating subject length. ==== Commit Subject Prefix [options="header"] |=== | Enabled | Severity | Defaults | true | error | includes: (see below) | | | delimiter: " " |=== Ensures each commit subject uses consistent prefixes that explain _what_ is being committed. The `includes` are _case sensitive_ and default to the following prefixes: * *Fixed* - Identifies what was fixed. The commit should be as small as possible and consist of changes to implementation and spec only. In some cases this might be a single line change. The important point is the change is applied to existing code which corrects behavior that wasn't properly implemented earlier. * *Removed* - Identifies what was removed. The commit should be as small as possible and consist only of removed lines/files from the existing implementation. This might also mean breaking changes requiring the publishing of a _major_ version release in the future. * *Added* - Identifies what was added. The commit should be as small as possible and consist of implementation and spec. Otherwise, it might be a change to an existing file which adds new behavior. * *Updated* - Identifies what was updated. The commit should be as small as possible and _not add or fix_ existing behavior. This can sometimes be a grey area but is typically reserved for updates to documentation, code comments, dependencies, etc. * *Refactored* - Identifies what was refactored. link:https://thoughtbot.com/blog/lets-not-misuse-refactoring[_Refactoring is for changing code structure without changing observable behavior_]. The commit should be as small as possible and not mix multiple kinds of changes at once. Refactored code should never break existing implementation behavior or corresponding specs because, if that happens, then one of the other four prefixes is what you want to use instead. In practice, it is quite rare to need a prefix other than what has been detailed above to explain _what_ is being committed. These prefixes are not only short and easy to remember but also have the added benefit of categorizing the commits for building release notes, change logs, etc. This becomes handy when coupled with another tool, link:https://www.alchemists.io/projects/milestoner[Milestoner], for producing consistent project milestones and Git tag histories. For a deeper dive on subject prefixes and good commit messages in general, please read about link:https://www.alchemists.io/articles/git_commit_anatomy[commit anatomy ] to learn more. 🎉 Each prefix is delimited by a space which is the default setting but can be customized if desired. Whatever you choose for a delimiter will not affect Git's special bang prefixes as described in the tip below. 💡 This analyzer automatically ignores _amend!_, _fixup!_, or _squash!_ commit prefixes when used as a Git Hook in order to not disturb interactive rebase workflows. ==== Commit Subject Suffix [options="header"] |=== | Enabled | Severity | Defaults | true | error | excludes: `["\\.", "\\?", "\\!"]` |=== Ensures commit subjects are suffixed consistently. The exclude list _is_ case sensitive and prevents the use of punctuation. This is handy when coupled with a tool, like link:https://www.alchemists.io/projects/milestoner[Milestoner], which automates project milestone releases. ==== Commit Trailer Collaborator Capitalization [options="header"] |=== | Enabled | Severity | Defaults | false | error | none |=== Ensures collaborator name is properly capitalized. Example: .... # Disallowed shepherd derrial book # Allowed Shepherd Derrial Book .... ==== Commit Trailer Collaborator Duplication [options="header"] |=== | Enabled | Severity | Defaults | false | error | none |=== Ensures collaborator trailers are not duplicated. Example: .... # Disallowed Co-Authored-By: Shepherd Derrial Book Co-Authored-By: Shepherd Derrial Book # Allowed Co-Authored-By: Malcolm Reynolds Co-Authored-By: Shepherd Derrial Book .... ==== Commit Trailer Collaborator Email [options="header"] |=== | Enabled | Severity | Defaults | false | error | none |=== Ensures collaborator email address is valid for commit trailer. .... # Disallowed Co-Authored-By: River Tam # Allowed Co-Authored-By: River Tam .... ==== Commit Trailer Collaborator Key [options="header"] |=== | Enabled | Severity | Defaults | false | error | includes: `["Co-Authored-By"]` |=== Ensures collaborator trailer key is correct format. .... # Disallowed Co-authored-by: River Tam # Allowed Co-Authored-By: River Tam .... ==== Commit Trailer Collaborator Name [options="header"] |=== | Enabled | Severity | Defaults | false | error | minimum: 2 |=== Ensures collaborator name consists of, at least, a first and last name. Example: .... # Disallowed Co-Authored-By: River # Allowed Co-Authored-By: River Tam .... === Git ==== Default Branch Your default branch configuration is respected no matter if it is set globally or locally. If the default branch is _not set_ then Git Lint will fall back to `master` for backwards compatibility. When the next major version is released, the default branch fallback will change from `master` to `main`. You can set your default branch at any time by running the following from the command line: [source,bash] ---- git config --add init.defaultBranch main ---- 💡 When setting your default branch, ensure you use a consistent Git configuration across all of your environments. ==== Hooks This gem supports link:https://git-scm.com/book/en/v2/Customizing-Git-Git-Hooks[Git Hooks]. It is _highly recommended_ you manage Git Hooks as global scripts as it'll reduce project maintenance costs for you. To configure global Git Hooks, add the following to your `$HOME/.gitconfig`: .... [core] hooksPath = ~/.git_template/hooks .... Then you can customize Git Hooks for all of your projects. link:https://github.com/bkuhlmann/dotfiles/tree/main/home_files/.config/git/hooks[Check out these examples]. If a global configuration is not desired, you can add Git Hooks at a per project level by editing any of the scripts within the `.git/hooks` directory of the repository. ===== Commit Message The _commit-msg_ hook -- which is the best way to use this gem as a Git Hook -- is provided as a `--hook` option. Usage: [source,bash] ---- git-lint --hook PATH ---- As shown above, the `--hook` command accepts a file path (i.e. `.git/COMMIT_EDITMSG`) which is provided to you by Git within the `.git/hooks/commit-msg` script. Here is a working example of what that script might look like: [source,bash] ---- #! /usr/bin/env bash set -o nounset set -o errexit set -o pipefail IFS=$'\n\t' if ! command -v git-lint > /dev/null; then printf "%s\n" "[git]: Git Lint not found. To install, run: gem install git-lint." exit 1 fi git-lint --hook "${BASH_ARGV[0]}" ---- Whenever you attempt to add a commit, Git Lint will check your commit for issues prior to saving it. ===== Post Commit The _post-commit_ hook is possible via the `--analyze` command. Usage: [source,bash] ---- git-lint --analyze --shas SHA ---- The _post-commit_ hook can be used multiple ways but, generally, you'll want to check the last commit made. Here is a working example which can be used as a `.git/hooks/post-commit` script: [source,bash] ---- #! /usr/bin/env bash set -o nounset set -o errexit set -o pipefail IFS=$'\n\t' if ! command -v git-lint > /dev/null; then printf "%s\n" "[git]: Git Lint not found. To install, run: gem install git-lint." exit 1 fi git-lint --analyze --shas $(git log --pretty=format:%H -1) ---- Whenever a commit has been saved, this script will run Git Lint to check for issues. === Rake This gem provides optional Rake tasks. They can be added to your project by adding the following requirement to the top of your `Rakefile`: [source,ruby] ---- require "git/lint/rake/setup" ---- Now, when running `bundle exec rake -T`, you'll see `git_lint` included in the list. If you need a concrete example, check out the link:https://github.com/bkuhlmann/git-lint/blob/main/Rakefile[Rakefile] of this project for details. === Continuous Integration (CI) This gem automatically configures itself for known CI build servers (see below for details). If you have a build server that is not listed, please log an issue or provide an implementation with support. Calculation of commits is done by reviewing all commits made on the feature branch since branching from `main`. ==== link:https://circleci.com[Circle CI] Detection and configuration happens automatically by checking the `CIRCLECI` environment variable. No additional setup required! ==== link:https://docs.github.com/en/free-pro-team@latest/actions[GitHub Actions] Detection happens automatically by checking the `GITHUB_ACTIONS` environment variable as supplied by the GitHub environment. The only configuration required is to add a `.github/workflows/git_lint.yml` to your repository with the following contents: [source,yaml] ---- name: Git Lint on: pull_request jobs: run: runs-on: ubuntu-latest container: image: ruby:latest steps: - name: Checkout uses: actions/checkout@v2 with: fetch-depth: '0' ref: ${{github.head_ref}} - name: Install run: gem install git-lint - name: Analyze run: git-lint --analyze ---- The above will ensure Git Lint runs as an additional check on each Pull Request. ==== link:https://www.netlify.com[Netlify CI] Detection and configuration happens automatically by checking the `NETLIFY` environment variable. No additional setup required! == Style Guide In addition to what is described above and automated for you, the following style guide is also worth considering: === General * Use a {git_rebase_workflow_link} instead of a Git Merge Workflow. * Use `git commit --amend` when fixing a previous commit, addressing code review feedback, etc. * Use `git commit --fixup` when fixing an earlier commit, addressing code review feedback, etc., and don't need to modify the original commit message. * Use `git commit --squash` when fixing an earlier commit, addressing code review feedback, etc., and want to combine multiple commit messages into a single commit message. _Avoid using squash to blindly combine multiple commit messages without editing them into a single, coherent message._ * Use `git rebase --interactive` when cleaning up commit history, order, messages, etc. This should be done prior to submitting a code review or when code review feedback has been addressed and you are ready to rebase onto `main`. * Use `git push --force-with-lease` instead of `git push --force` when pushing changes after an interactive rebasing session. * Avoid checking in development-specific configuration files (add to `.gitignore` instead). * Avoid checking in sensitive information (i.e. security keys, passphrases, etc). * Avoid "WIP" (a.k.a. "Work in Progress") commits and/or code review labels. Be confident with your code and colleagues' time. Use branches, stashes, etc. instead -- share a link to a feature branch diff if you have questions/concerns during development. * Avoid using link:https://git-scm.com/book/en/v2/Git-Tools-Submodules[Git Submodules]. This practice leads to complicated project cloning, deployments, maintenance, etc. Use separate repositories to better organize and split out this work. Sophisticated package managers, like link:https://bundler.io[Bundler], exist to manage these dependencies better than what multiple Git Submodules can accomplish. * Avoid using link:https://git-lfs.github.com[Git LFS] for tracking binary artifacts/resources. These files are not meant for version control and lead to large repositories that are time consuming to clone/deploy. Use storage managers like link:https://git-annex.branchable.com[Git Annex], link:https://aws.amazon.com/s3[Amazon S3], or link:https://lakefs.io[LakeFS] which are better suited for binary assets that don't change often. === Security Ensure signed commits, pushes, and tags are enabled within your global Git Configuration to reduce an link:https://blog.gruntwork.io/how-to-spoof-any-user-on-github-and-what-to-do-to-prevent-it-e237e95b8deb[attack vector]. Run the following commands to enable: [source,bash] ---- git config --global commit.gpgSign true git config --global push.gpgSign true git config --global tag.gpgSign true ---- ⚠️ GitHub, unfortunately, doesn't support signed pushes so you might need to leave that configuration disabled. === Commits * Use a commit subject that explains _what_ is being committed. * Use a commit message body that explains _why_ the commit is necessary. Additional considerations: ** If the commit has a dependency to the previous commit or is a precursor to the commit that will follow, make sure to explain that. ** Include links to dependent projects, stories, etc. if available. * Use small, atomic commits: ** Easier to review and provide feedback. ** Easier to review implementation and corresponding tests. ** Easier to document with detailed subjects (especially when grouped together in a pull request). ** Easier to reword, edit, squash, fix, or drop when interactively rebasing. ** Easier to combine together versus tearing apart a larger commit into smaller commits. * Use logically ordered commits: ** Each commit should tell a story and be a logical building block to the next commit. ** Each commit should, ideally, be the implementation plus corresponding test. Avoid committing changes that are a jumble of mixed ideas as they are hard to decipher and a huge insult not only to the reviewer but your future self. ** Each commit, when reviewed in order, should be able to explain _how_ the feature or bug fix was completed and implemented properly. * Keep refactored code separate from behavioral changes. This makes the review process easier because you don't have to sift through all the line and format changes to figure out what is new or changed. === Branches * Use feature branches for new work. * Maintain branches by rebasing upon `main` on a regular basis. === Tags * Use tags to denote link:https://www.alchemists.io/projects/milestoner[milestones]/releases: ** Makes it easier to record milestones and capture associated release notes. ** Makes it easier to compare differences between versions. ** Provides a starting point for debugging production issues (if any). === Rebases * Avoid rebasing a shared branch. If you must do this, clear communication should be used to warn those ahead of time, ensure that all of their work is checked in, and that their local branch is deleted first. === Hooks * Use hooks to augment and automate your personal workflow such as checking code quality, detecting forgotten debug statements, etc. * Use hooks globally rather than locally per project. Doing this applies the same functionality across all projects automatically, reduces maintenance per project, and provides consistency across all projects. This can best be managed via your link:https://github.com/bkuhlmann/dotfiles/tree/main/home_files/.config/git/hooks[Dotfiles]. * Avoid forcing global or local project hooks as a team-wide mandate. Hooks are a personal tool much like editors or other tools one choose to do their work. For team consistency, use a continuous integration build server instead. === Code Reviews For an in depth look at how to conduct code reviews, please read my link:https://www.alchemists.io/articles/code_reviews[article] on this subject to learn more. == Development To contribute, run: [source,bash] ---- git clone https://github.com/bkuhlmann/git-lint cd git-lint bin/setup ---- You can also use the IRB console for direct access to all objects: [source,bash] ---- bin/console ---- == Tests To test, run: [source,bash] ---- bundle exec rake ---- == link:https://www.alchemists.io/policies/license[License] == link:https://www.alchemists.io/policies/security[Security] == link:https://www.alchemists.io/policies/code_of_conduct[Code of Conduct] == link:https://www.alchemists.io/policies/contributions[Contributions] == link:https://www.alchemists.io/projects/git-lint/versions[Versions] == link:https://www.alchemists.io/community[Community] == Credits * Built with link:https://www.alchemists.io/projects/gemsmith[Gemsmith]. * Engineered by link:https://www.alchemists.io/team/brooke_kuhlmann[Brooke Kuhlmann].