The husky command to add JIRA ticket ID into the scope of conventional commit message if it is missed.
The JIRA ticket ID is taken from a git branch name.
Installing Jira prepare commit msg hook into your project will mean everyone contributing code to your project will automatically tag each commit with it's associated issue key based off the branch name.
So if your branch name is feature/TEST-123-new-feature
, then when you commit with a message "initial commit"
it will automatically become "chore(TEST-123): initial commit"
.
Why would you want this? Well, Jira has many hidden goodies, and this is one of them! If you include an issue key in your commit messages AND you have your deployment pipeline connected to Jira this will unlock many bonus features, such as the Deployments view, Cycle time report, Deployment frequency report and etc.
Additionally, when you use Conventional Commits you will get a nice commit history and clean changelog.
Install the package using NPM
npm install husky jira-prepare-conventional-commit-msg --save-dev && npx husky install
For Husky 5:
Execute command
npx husky add .husky/prepare-commit-msg 'npx jira-prepare-conventional-commit-msg $1'
For Husky 2-4:
Inside your package.json add a standard husky npm script for the git hook
{
"husky": {
"hooks": {
"prepare-commit-msg": "jira-prepare-conventional-commit-msg"
}
}
}
Starting with v1.3 you can now use different ways of configuring it:
jira-prepare-commit-msg
object in yourpackage.json
.jirapreparecommitmsgrc
file in JSON or YML formatjira-prepare-commit-msg.config.js
file in JS format
See cosmiconfig for more details on what formats are supported.
{
"jira-prepare-commit-msg": {
"jiraTicketPattern": "([A-Z]+-\\d+)",
"conventionalCommitPattern": "^([a-z]+)(?:\\(([a-z0-9.,-_ ]+)\\))?!?: ([\\w \\S]+)$",
"allowEmptyCommitMessage": false,
"ignoredBranchesPattern": "^(master|main|dev|develop|development|release)$",
"ignoreBranchesMissingTickets": false
}
}
jira-prepare-conventional-commit-msg
allows using custom regexp string pattern to search JIRA ticket number.
Pattern ([A-Z]+-\\d+)
is currently supported by default.
NOTE: to search JIRA ticket pattern flag i
is used: new RegExp(pattern, i')
{
"jira-prepare-commit-msg": {
"jiraTicketPattern": "([A-Z]+-\\d+)"
}
}
The commit message might be empty after cleanup or using -m ""
, jira-prepare-conventional-commit-msg
might insert the JIRA ticket number anyway if this flag is set.
{
"jira-prepare-commit-msg": {
"allowEmptyCommitMessage": true
}
}
Branches can be ignored and skipped by regex pattern string
{
"jira-prepare-commit-msg": {
"ignoredBranchesPattern": "^main|develop|(maint-.*)$"
}
}
Moreover, this can be solved by replacing the Husky hook. Put in your prepare-commit-msg file (husky git hook):
#!/bin/sh
. "$(dirname "$0")/_/husky.sh"
if [[ "$(git rev-parse --abbrev-ref HEAD)" =~ YOUR_BRANCH_REGEX ]]; then
npx --no-install jira-prepare-conventional-commit-msg $1
fi
where YOUR_BRANCH_REGEX
e.g. ^(feature|(bug|hot)fix)\/[A-Z]+-[0-9]+$
Be silent and skip any branch with missing jira ticket
{
"jira-prepare-commit-msg": {
"ignoreBranchesMissingTickets": true
}
}
jira-prepare-conventional-commit-msg
natively works with conventional commit.
If the commit message is fix(test)!: important changes
then at result will be fix(test, JIRA-1234)!: important changes
.
Additionally, you can customize the conventional commit format with the following setting:
{
"jira-prepare-commit-msg": {
"conventionalCommitPattern": "^([a-z]+)(?:\\(([a-z0-9.,-_ ]+)\\))?!?: ([\\w \\S]+)$"
}
}
The above regular expression is the default conventional commit pattern so, if you don't provide this property, jira-prepare-conventional-commit-msg
will use this by default.
In the default regular expression, from left to right:
([a-z]+)
is the committype
.(?:\\(([a-z0-9.,-_ ]+)\\))?!?
is the commitscope
.- And
([\\w \\S]+)
is the commitsubject
.
With this setting you can change how jira-prepare-conventional-commit-msg
reads your custom conventional commit message and rewrites it properly adding the Jira ticket id in the scope
.
You can allow the scope to have capital letters adding A-Z to the regular expression above. If the configuration is:
{
"jira-prepare-commit-msg": {
"conventionalCommitPattern": "^([a-z]+)(?:\\(([a-zA-Z0-9.,-_ ]+)\\))?!?: ([\\w \\S]+)$"
// ^^^
// Now we can use capital letters in the conventional commit scope
}
}
and commit message is "test(E2E): some end-to-end testing stuff
" then at result will be "test(E2E, JIRA-1234): some end-to-end testing stuff
"
Be aware that if you leave the default conventionalCommitPattern
value (that it not allows capital letters in the commit scope) in the example above, your resulting message will be "chore(JIRA-1234): test(E2E): some end-to-end testing stuff
". Maybe, this is not the result you are expecting and you can have problems using other tools like commitlint.
MIT