Git Commit Message Conventions

I want to take a moment to elaborate on what makes a well formed commit message. I think the best practices for commit message formatting is one of the little details that makes Git great.

Goals

Good commit messages should satisfy the below goals.

  • allow generating CHANGELOG.md by script
  • allow ignoring commits by git bisect (not important commits like formatting)
  • provide better information when browsing the history

Generating CHANGELOG.md

We use these three sections in changelog: new features, bug fixes, breaking changes.
This list could be generated by script when doing a release. Along with links to related commits.
Of course you can edit this change log before actual release, but it could generate the skeleton.

List of all subjects (first lines in commit message) since last release:
>> git log <last tag> HEAD --pretty=format:%s

New features in this release
>> git log <last release> HEAD --grep FEAT

Recognizing unimportant commits

These are formatting changes (adding/removing spaces/empty lines, indentation), missing semi colons, comments. So when you are looking for some change, you can ignore these commits - no logic change inside this commit.

When bisecting, you can ignore these by:
>> git bisect skip $(git rev-list --grep irrelevant <good place> HEAD)

Provide more information when browsing the history

This would add kinda “context” information.
Look at these messages (taken from last few angular’s commits):

1
2
3
4
5
6
Fix small typo in docs widget (tutorial instructions)
Fix test for scenario.Application - should remove old iframe
docs - various doc fixes
docs - stripping extra new lines
Replaced double line break with single when text is fetched from Google
Added support for properties in documentation

All of these messages try to specify where is the change. But they don’t share any convention…

Look at these messages:

1
2
3
4
5
fix comment stripping
fixing broken links
Bit of refactoring
Check whether links do exist and throw exception
Fix sitemap include (to work on case sensitive linux)

Are you able to guess what’s inside ? These messages miss place specification…
So maybe something like parts of the code: docs, docs-parser, compiler, scenario-runner, …

I know, you can find this information by checking which files had been changed, but that’s slow. And when looking in git history I can see all of us tries to specify the place, only missing the convention.

Format of the commit message

1
2
3
4
5
<type>(<scope>): <subject>
<BLANK LINE>
<body>
<BLANK LINE>
<footer>

Any line of the commit message cannot be longer 72 characters! This allows the message to be easier to read on github as well as in various git tools.

Subject line

Subject line contains succinct description of the change.

Allowed

  • feet (feature)
  • fix (bug fix)
  • docs (documentation)
  • style (formatting, missing semi colons, …)
  • refactor (refactoring)
  • test (when adding missing tests)
  • chore (maintain)

Allowed

Scope could be anything specifying place of the commit change. For example $location, $browser, $compile, $rootScope, ngHref, ngClick, ngView, etc…

text

Subject should use imperative, present tense: “change” not “changed” nor “changes”
don’t capitalize first letter
no dot (.) at the end

Message body

just as in , the body should provide a meaningful commit message, which use imperative, present tense: “change” not “changed” nor “changes” and includes motivation for the change that contrasts its implementation with previous behavior. This convention matches up with commit messages generated
by commands like git merge and git revert.

More detailed explanatory text, if necessary. Wrap it to about 72 characters or so.
The blank line separating the summary from the body is critical (unless you omit the body entirely); tools like rebase can get confused if you run the two together.

Further paragraphs come after blank lines.

  • Bullet points are okay, too
  • Typically a hyphen or asterisk is used for the bullet, followed by a single space, with blank lines in between, but conventions vary here
  • Use a hanging indent

Referencing issues

Closed bugs should be listed on a separate line in the footer prefixed with “Closes” keyword like this:

Closes #234

or in case of multiple issues:

Closes #123, #245, #992

Breaking changes

All breaking changes have to be mentioned in footer with the description of the change, justification and migration notes

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
BREAKING CHANGE: isolate scope bindings definition has changed and
    the inject option for the directive controller injection was removed.
    To migrate the code follow the example below:
    Before:
    scope: {
      myAttr: 'attribute',
      myBind: 'bind',
      myExpression: 'expression',
      myEval: 'evaluate',
      myAccessor: 'accessor'
    }
    After:
    scope: {
      myAttr: '@',
      myBind: '@',
      myExpression: '&',
      // myEval - usually not useful, but in cases where the expression is assignable, you can use '='
      myAccessor: '=' // in directive's template change myAccessor() to myAccessor
    }
    The removed `inject` wasn't generaly useful for directives so there should be no code using it.

Examples

1
2
3
4
5
6
7
8
FEAT($browser): onUrlChange event (popstate/hashchange/polling)
Added new event to $browser:
- forward popstate event if available
- forward hashchange event if popstate not available
- do polling when neither popstate nor hashchange available
Breaks $browser.onHashChange, which was removed (use onUrlChange instead)
1
2
3
4
5
6
7
8
FIX($compile): couple of unit tests for IE9
Older IEs serialize html uppercased, but IE9 does not...
Would be better to expect case insensitive, unfortunately jasmine does
not allow to user regexps for throw expectations.
Closes #392
Breaks foo.bar api, foo.baz should be used instead
1
2
3
4
5
6
FEAT(directive): ng:disabled, ng:checked, ng:multiple, ng:readonly, ng:selected
New directives for proper binding these attributes in older browsers (IE).
Added coresponding description, live examples and e2e tests.
Closes #351
1
2
3
4
5
6
7
8
9
STYLE($location): add couple of missing semi colons
DOCS(guide): updated fixed docs from Google Docs
Couple of typos fixed:
- indentation
- batchLogbatchLog -> batchLog
- start periodic checking
- missing brace
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
FEAT($compile): simplify isolate scope bindings
Changed the isolate scope binding options to:
  - @attr - attribute binding (including interpolation)
  - =model - by-directional model binding
  - &expr - expression execution binding
This change simplifies the terminology as well as
number of choices available to the developer. It
also supports local name aliasing from the parent.
BREAKING CHANGE: isolate scope bindings definition has changed and
the inject option for the directive controller injection was removed.
To migrate the code follow the example below:
Before:
scope: {
  myAttr: 'attribute',
  myBind: 'bind',
  myExpression: 'expression',
  myEval: 'evaluate',
  myAccessor: 'accessor'
}
After:
scope: {
  myAttr: '@',
  myBind: '@',
  myExpression: '&',
  // myEval - usually not useful, but in cases where the expression is assignable, you can use '='
  myAccessor: '=' // in directive's template change myAccessor() to myAccessor
}
The removed `inject` wasn't generaly useful for directives so there should be no code using it.

Kommentare: