A tool for managing JavaScript projects with multiple packages.
Splitting up large codebases into separate independently versioned packages is extremely useful for code sharing. However, making changes across many repositories is messy and difficult to track, and testing across repositories gets complicated really fast.
To solve these (and many other) problems, some projects will organize their codebases into multi-package repositories (sometimes called monorepos). Projects like Babel, React, Angular, Ember, Meteor, Jest, and many others develop all of their packages within a single repository.
Lerna is a tool that optimizes the workflow around managing multi-package repositories with git and npm.
Lerna can also reduce the time and space requirements for numerous copies of packages in development and build environments - normally a downside of dividing a project into many separate NPM package. See the hoist documentation for details.
There's actually very little to it. You have a file system that looks like this:
my-lerna-repo/
package.json
packages/
package-1/
package.json
package-2/
package.json
The two primary commands in Lerna are lerna bootstrap
and lerna publish
.
bootstrap
will link dependencies in the repo together.
publish
will help publish any updated packages.
The instructions below are for Lerna 2.x. We recommend using it instead of 1.x for a new Lerna project. Check the wiki if you need to see the 1.x README.
Let's start by installing Lerna globally with npm.
$ npm install --global lerna
Next we'll create a new folder:
$ mkdir lerna-repo
$ cd lerna-repo
And now let's turn it into a Lerna repo:
$ lerna init
This will create a lerna.json
configuration file as well as a packages
folder, so your folder should now look like this:
lerna-repo/
packages/
package.json
lerna.json
Lerna allows you to manage your project using one of two modes: Fixed or Independent.
Fixed mode Lerna projects operate on a single version line. The version is kept in the lerna.json
file at the root of your project under the version
key. When you run lerna publish
, if a module has been updated since the last time a release was made, it will be updated to the new version you're releasing. This means that you only publish a new version of a package when you need to.
This is the mode that Babel is currently using. Use this if you want to automatically tie all package versions together. One issue with this approach is that a major change in any package will result in all packages having a new major version.
Independent mode Lerna projects allows maintainers to increment package versions independently of each other. Each time you publish, you will get a prompt for each package that has changed to specify if it's a patch, minor, major or custom change.
Independent mode allows you to more specifically update versions for each package and makes sense for a group of components. Combining this mode with something like semantic-release would make it less painful. (There is work on this already at atlassian/lerna-semantic-release).
The
version
key inlerna.json
is ignored in independent mode.
If you encounter any issues while using Lerna please check out our Troubleshooting document where you might find the answer to your problem.
See FAQ.md.
$ lerna init
Create a new Lerna repo or upgrade an existing repo to the current version of Lerna.
Lerna assumes the repo has already been initialized with
git init
.
When run, this command will:
- Add
lerna
as adevDependency
inpackage.json
if it doesn't already exist. - Create a
lerna.json
config file to store theversion
number.
Example output on a new git repo:
$ lerna init
lerna info version v2.0.0
lerna info Updating package.json
lerna info Creating lerna.json
lerna success Initialized Lerna files
$ lerna init --independent
This flag tells Lerna to use independent versioning mode.
$ lerna init --exact
By default, lerna init
will use a caret range when adding or updating
the local version of lerna
, just like npm install --save-dev lerna
.
To retain the lerna
1.x behavior of "exact" comparison, pass this flag.
It will configure lerna.json
to enforce exact match for all subsequent executions.
{
"command": {
"init": {
"exact": true
}
},
"version": "0.0.0"
}
$ lerna bootstrap
Bootstrap the packages in the current Lerna repo. Installs all of their dependencies and links any cross-dependencies.
When run, this command will:
npm install
all external dependencies of each package.- Symlink together all Lerna
packages
that are dependencies of each other. npm run prepublish
in all bootstrapped packages.npm run prepare
in all bootstrapped packages.
lerna bootstrap
respects the --ignore
, --ignore-scripts
, --scope
and --include-filtered-dependencies
flags (see Flags).
Pass extra arguments to npm client by placing them after --
:
$ lerna bootstrap -- --production --no-optional
May also be configured in lerna.json
:
{
...
"npmClient": "yarn",
"npmClientArgs": ["--production", "--no-optional"]
}
This runs lerna bootstrap
with npm ci
as opposed to npm install
. The specifics of this command can be found in the NPM documentation here
Let's use babel
as an example.
babel-generator
andsource-map
(among others) are dependencies ofbabel-core
.babel-core
'spackage.json
lists both these packages as keys independencies
, as shown below.
$ lerna add <package>[@version] [--dev]
Add local or remote package
as dependency to packages in the current Lerna repo.
When run, this command will:
- Add
package
to each applicable package. Applicable are packages that are notpackage
and are in scope - Bootstrap packages with changes to their manifest file (
package.json
)
lerna add
respects the --ignore
, --scope
and --include-filtered-dependencies
flags (see Flags).
lerna add module-1 --scope=module-2 # Install module-1 to module-2
lerna add module-1 --scope=module-2 --dev # Install module-1 to module-2 in devDependencies
lerna add module-1 # Install module-1 in all modules except module-1
lerna add babel-core # Install babel-core in all modules
// babel-core package.json
{
"name": "babel-core",
...
"dependencies": {
...
"babel-generator": "^6.9.0",
...
"source-map": "^0.5.0"
}
}
- Lerna checks if each dependency is also part of the Lerna repo.
- In this example,
babel-generator
can be an internal dependency, whilesource-map
is always an external dependency. - The version of
babel-generator
in thepackage.json
ofbabel-core
is satisfied bypackages/babel-generator
, passing for an internal dependency. source-map
isnpm install
ed (oryarn
ed) like normal.
- In this example,
packages/babel-core/node_modules/babel-generator
symlinks topackages/babel-generator
- This allows nested directory imports
Notes:
- When a dependency version in a package is not satisfied by a package of the same name in the repo, it will be
npm install
ed (oryarn
ed) like normal. - Dist-tags, like
latest
, do not satisfy semver ranges. - Circular dependencies result in circular symlinks which may impact your editor/IDE.
Webstorm locks up when circular symlinks are present. To prevent this, add node_modules
to the list of ignored files and folders in Preferences | Editor | File Types | Ignored files and folders
.
$ lerna publish
Publish packages in the current Lerna project. When run, this command does the following:
Creates a new release of the packages that have been updated. Prompts for a new version. Creates a new git commit/tag in the process of publishing to npm.
More specifically, this command will:
- Run the equivalent of
lerna updated
to determine which packages need to be published. - If necessary, increment the
version
key inlerna.json
. - Update the
package.json
of all updated packages to their new versions. - Update all dependencies of the updated packages with the new versions, specified with a caret (^).
- Create a new git commit and tag for the new version.
- Publish updated packages to npm.
Lerna won't publish packages which are marked as private (
"private": true
in thepackage.json
).
Note: to publish scoped packages, you need to add the following to each package.json
:
"publishConfig": {
"access": "public"
}
$ lerna publish --exact
When run with this flag, publish
will specify updated dependencies in updated packages exactly (with no punctuation), instead of as semver compatible (with a ^
).
For more information, see the package.json dependencies documentation.
$ lerna publish --npm-tag=next
When run with this flag, publish
will publish to npm with the given npm dist-tag (defaults to latest
).
This option can be used to publish a prerelease
or beta
version.
Note: the
latest
tag is the one that is used when a user runsnpm install my-package
. To install a different tag, a user can runnpm install my-package@prerelease
.
$ lerna publish --canary
$ lerna publish --canary=beta
When run with this flag, publish
publishes packages in a more granular way (per commit). Before publishing to npm, it creates the new version
tag by taking the current version
, bumping it to the next minor version, adding the provided meta suffix (defaults to alpha
) and appending the current git sha (ex: 1.0.0
becomes 1.1.0-alpha.81e3b443
).
The intended use case for this flag is a per commit level release or nightly release.
$ lerna publish --conventional-commits
When run with this flag, publish
will use the Conventional Commits Specification to determine the version bump and generate CHANGELOG
$ lerna publish --conventional-commits --changelog-preset angular-bitbucket
By default, the changelog preset is set to angular
.
In some cases you might want to change either use a another preset or a custom one.
Presets are names of built-in or installable configuration for conventional changelog.
Presets may be passed as the full name of the package, or the auto-expanded suffix
(e.g., angular
is expanded to conventional-changelog-angular
).
$ lerna publish --git-remote upstream
When run with this flag, publish
will push the git changes to the specified remote instead of origin
.
$ lerna publish --skip-git
When run with this flag, publish
will publish to npm without running any of the git commands.
Only publish to npm; skip committing, tagging, and pushing git changes (this only affects publish).
$ lerna publish --skip-npm
When run with this flag, publish
will update all package.json
package
versions and dependency versions, but it will not actually publish the
packages to npm.
This flag can be combined with --skip-git
to just update versions and
dependencies, without committing, tagging, pushing or publishing.
Only update versions and dependencies; don't actually publish (this only affects publish).
$ lerna publish --force-publish=package-2,package-4
# force publish all packages
$ lerna publish --force-publish=*
When run with this flag, publish
will force publish the specified packages (comma-separated) or all packages using *
.
This will skip the
lerna updated
check for changed packages and forces a package that didn't have agit diff
change to be updated.
$ lerna publish --canary --yes
# skips `Are you sure you want to publish the above changes?`
When run with this flag, publish
will skip all confirmation prompts.
Useful in Continuous integration (CI) to automatically answer the publish confirmation prompt.
$ lerna publish --cd-version (major | minor | patch | premajor | preminor | prepatch | prerelease)
# uses the next semantic version(s) value and this skips `Select a new version for...` prompt
When run with this flag, publish
will skip the version selection prompt (in independent mode) and use the next specified semantic version.
You must still use the --yes
flag to avoid all prompts. This is useful when build systems need to publish without command prompts. Works in both normal and independent modes.
If you have any packages with a prerelease version number (e.g. 2.0.0-beta.3
) and you run lerna publish
with --cd-version
and a non-prerelease version increment (major / minor / patch), it will publish those packages in addition to the packages that have changed since the last release.
$ lerna publish --cd-version=prerelease
# uses the next semantic prerelease version, e.g.
# 1.0.0 => 1.0.0-0
$ lerna publish --cd-version=prepatch --preid=next
# uses the next semantic prerelease version with a specific prerelease identifier, e.g.
# 1.0.0 => 1.0.1-next.0
When run with this flag, lerna publish --cd-version
will
increment premajor
, preminor
, prepatch
, or prerelease
versions using the specified prerelease identifier.
$ lerna publish --repo-version 1.0.1
# applies version and skips `Select a new version for...` prompt
When run with this flag, publish
will skip the version selection prompt and use the specified version.
Useful for bypassing the user input prompt if you already know which version to publish.
$ lerna publish -m "chore(release): publish %s"
# commit message = "chore(release): publish v1.0.0"
$ lerna publish -m "chore(release): publish %v"
# commit message = "chore(release): publish 1.0.0"
$ lerna publish -m "chore(release): publish" --independent
# commit message = "chore(release): publish
#
# - package-1@3.0.1
# - package-2@1.5.4"
When run with this flag, publish
will use the provided message when committing the version updates
for publication. Useful for integrating lerna into projects that expect commit messages to adhere
to certain guidelines, such as projects which use commitizen and/or semantic-release.
If the message contains %s
, it will be replaced with the new global version version number prefixed with a "v".
If the message contains %v
, it will be replaced with the new global version version number without the leading "v".
Note that this only applies when using the default "fixed" versioning mode, as there is no "global" version when using --independent
.
This can be configured in lerna.json, as well:
{
"command": {
"publish": {
"message": "chore(release): publish %s"
}
}
}
Lerna allows you to specify a glob or an array of globs in your lerna.json
that your current branch needs to match to be publishable.
You can use this flag to override this setting.
If your lerna.json
contains something like this:
{
"command": {
"publish": {
"allowBranch": "master"
}
}
}
{
"command": {
"publish": {
"allowBranch": ["master", "feature/*"]
}
}
}
and you are not on the branch master
lerna will prevent you from publishing. To force a publish despite this config, pass the --allow-branch
flag:
$ lerna publish --allow-branch my-new-feature
$ lerna updated
Check which packages
have changed since the last release (the last git tag).
Lerna determines the last git tag created and runs git diff --name-only v6.8.1
to get all files changed since that tag. It then returns an array of packages that have an updated file.
Note that configuration for the publish
command also affects the
updated
command. For example command.publish.ignoreChanges
$ lerna updated --json
When run with this flag, updated
will return an array of objects in the following format:
[
{
"name": "package",
"version": "1.0.0",
"private": false
}
]
$ lerna clean
Remove the node_modules
directory from all packages.
lerna clean
respects the --ignore
, --scope
, and --yes
flags (see Flags).
$ lerna diff [package?]
$ lerna diff
# diff a specific package
$ lerna diff package-name
Diff all packages or a single package since the last release.
Similar to
lerna updated
. This command runsgit diff
.
$ lerna ls
List all of the public packages in the current Lerna repo.
lerna ls
respects the --ignore
and --scope
flags (see Flags).
$ lerna ls --json
When run with this flag, ls
will return an array of objects in the following format:
[
{
"name": "package",
"version": "1.0.0",
"private": false
}
]
$ lerna run <script> -- [..args] # runs npm run my-script in all packages that have it
$ lerna run test
$ lerna run build
# watch all packages and transpile on change, streaming prefixed output
$ lerna run --parallel watch
Run an npm script in each package that contains that script. A double-dash (--
) is necessary to pass dashed arguments to the script execution.
lerna run
respects the --concurrency
, --scope
, --ignore
, --stream
, and --parallel
flags (see Flags).
$ lerna run --scope my-component test
Note: It is advised to constrain the scope of this command (and
lerna exec
, below) when using the--parallel
flag, as spawning dozens of subprocesses may be harmful to your shell's equanimity (or maximum file descriptor limit, for example). YMMV
$ lerna exec -- <command> [..args] # runs the command in all packages
$ lerna exec -- rm -rf ./node_modules
$ lerna exec -- protractor conf.js
Run an arbitrary command in each package.
A double-dash (--
) is necessary to pass dashed flags to the spawned command, but is not necessary when all the arguments are positional.
lerna exec
respects the --concurrency
, --scope
, --ignore
, --stream
and --parallel
flags (see Flags).
$ lerna exec --scope my-component -- ls -la
To spawn long-running processes, pass the --parallel
flag:
# transpile all modules as they change in every package
$ lerna exec --parallel -- babel src -d lib -w
You may also get the name of the current package through the environment variable LERNA_PACKAGE_NAME
:
$ lerna exec -- npm view \$LERNA_PACKAGE_NAME
You may also run a script located in the root dir, in a complicated dir structure through the environment variable LERNA_ROOT_PATH
:
$ lerna exec -- node \$LERNA_ROOT_PATH/scripts/some-script.js
Hint: The commands are spawned in parallel, using the concurrency given (except with
--parallel
). The output is piped through, so not deterministic. If you want to run the command in one package after another, use it like this:
$ lerna exec --concurrency 1 -- ls -la
$ lerna exec --bail=<boolean> <command>
This flag signifies whether or not the exec
command should halt execution upon encountering an error thrown by one of the spawned subprocesses. Its default value is true
.
$ lerna import <path-to-external-repository>
Import the package at <path-to-external-repository>
, with commit history,
into packages/<directory-name>
. Original commit authors, dates and messages
are preserved. Commits are applied to the current branch.
This is useful for gathering pre-existing standalone packages into a Lerna
repo. Each commit is modified to make changes relative to the package
directory. So, for example, the commit that added package.json
will
instead add packages/<directory-name>/package.json
.
$ lerna link
Symlink together all Lerna packages
that are dependencies of each other in the current Lerna repo.
$ lerna link --force-local
When passed, this flag causes the link
command to always symlink local dependencies regardless of matching version range.
Lerna will log to a lerna-debug.log
file (same as npm-debug.log
) when it encounters an error running a command.
Lerna also has support for scoped packages.
Running lerna
without arguments will show all commands/options.
{
"version": "1.1.3",
"command": {
"publish": {
"ignoreChanges": [
"ignored-file",
"*.md"
]
},
"bootstrap": {
"ignore": "component-*"
}
},
"packages": ["packages/*"]
}
version
: the current version of the repository.command.publish.ignoreChanges
: an array of globs that won't be included inlerna changed/publish
. Use this to prevent publishing a new version unnecessarily for changes, such as fixing aREADME.md
typo.command.bootstrap.ignore
: an array of globs that won't be bootstrapped when running thelerna bootstrap
command.command.bootstrap.scope
: an array of globs that restricts which packages will be bootstrapped when running thelerna bootstrap
command.packages
: Array of globs to use as package locations.
Most devDependencies
can be pulled up to the root of a Lerna repo.
This has a few benefits:
- All packages use the same version of a given dependency
- Can keep dependencies at the root up-to-date with an automated tool such as GreenKeeper
- Dependency installation time is reduced
- Less storage is needed
Note that devDependencies
providing "binary" executables that are used by
npm scripts still need to be installed directly in each package where they're
used.
For example the nsp
dependency is necessary in this case for lerna run nsp
(and npm run nsp
within the package's directory) to work correctly:
{
"scripts": {
"nsp": "nsp"
},
"devDependencies": {
"nsp": "^2.3.3"
}
}
Lerna allows target versions of local dependent packages to be written as a git remote url with a committish
(e.g., #v1.0.0
or #semver:^1.0.0
) instead of the normal numeric version range.
This allows packages to be distributed via git repositories when packages must be private and a private npm registry is not desired.
Please note that lerna does not perform the actual splitting of git history into the separate read-only repositories. This is the responsibility of the user. (See this comment for implementation details)
// packages/pkg-1/package.json
{
name: "pkg-1",
version: "1.0.0",
dependencies: {
"pkg-2": "github:example-user/pkg-2#v1.0.0"
}
}
// packages/pkg-2/package.json
{
name: "pkg-2",
version: "1.0.0"
}
In the example above,
lerna bootstrap
will properly symlinkpkg-2
intopkg-1
.lerna publish
will update the committish (#v1.0.0
) inpkg-1
whenpkg-2
changes.
Options to Lerna can come from configuration (lerna.json
) or on the command
line. Additionally options in config can live at the top level or may be
applied to specific commands.
Example:
{
"version": "1.2.0",
"exampleOption": "foo",
"command": {
"init": {
"exampleOption": "bar"
}
}
}
In this case exampleOption
will be "foo" for all commands except init
,
where it will be "bar". In all cases it may be overridden to "baz" on the
command-line with --example-option=baz
.
How many threads to use when Lerna parallelizes the tasks (defaults to 4
)
$ lerna publish --concurrency 1
Scopes a command to a subset of packages.
$ lerna exec --scope my-component -- ls -la
$ lerna run --scope toolbar-* test
When executing a script or command, scope the operation to packages that have been updated since the specified ref
. If ref
is not specified, it defaults to the latest tag.
List the contents of packages that have changed since the latest tag:
$ lerna exec --since -- ls -la
Run the tests for all packages that have changed since master
:
$ lerna run test --since master
List all packages that have changed since some-branch
:
$ lerna ls --since some-branch
This can be particularly useful when used in CI, if you can obtain the target branch a PR will be going into, because you can use that as the ref
to the --since
option. This works well for PRs going into master as well as feature branches.
When importing repositories with merge commits with conflicts, the import command will fail trying to apply all commits. The user can use this flag to ask for import of "flat" history, i.e. with each merge commit as a single change the merge introduced.
$ lerna import ~/Product --flatten
Excludes a subset of packages when running a command.
$ lerna bootstrap --ignore component-*
The ignore
flag, when used with the bootstrap
command, can also be set in lerna.json
under the command.bootstrap
key. The command-line flag will take precedence over this option.
Example
{
"version": "0.0.0",
"command": {
"bootstrap": {
"ignore": "component-*"
}
}
}
Hint: The glob is matched against the package name defined in
package.json
, not the directory name the package lives in.
When used with the bootstrap
command it won't run any lifecycle scripts in bootstrapped packages.
$ lerna bootstrap --ignore-scripts
Used in combination with any command that accepts --scope
(bootstrap
, clean
, ls
, run
, exec
). Ensures that all dependencies (and dev dependencies) of any scoped packages (either through --scope
or --ignore
) are operated on as well.
Note: This will override the
--scope
and--ignore
flags.i.e. A package matched by the
--ignore
flag will still be bootstrapped if it is depended on by another package that is being bootstrapped.
This is useful for situations where you want to "set up" a single package that relies on other packages being set up.
$ lerna bootstrap --scope my-component --include-filtered-dependencies
# my-component and all of its dependencies will be bootstrapped
$ lerna bootstrap --scope "package-*" --ignore "package-util-*" --include-filtered-dependencies
# all package-util's will be ignored unless they are depended upon by a
# package matched by "package-*"
What level of logs to report. On failure, all logs are written to lerna-debug.log in the current working directory.
Any logs of a higher level than the setting are shown. The default is "info".
Set a max buffer length for each underlying process call. Useful for example
when someone wants to import a repo with a larger amount of commits while
running lerna import
. In that case the built-in buffer length might not
be sufficient.
By default, all tasks execute on packages in topologically sorted order as to respect the dependency relationships of the packages in question. Cycles are broken on a best-effort basis in a way not guaranteed to be consistent across Lerna invocations.
Topological sorting can cause concurrency bottlenecks if there are a small number of packages with many dependents or if some packages take a disproportionately long time to execute. The --no-sort
option disables sorting, instead executing tasks in an arbitrary order with maximum concurrency.
This option can also help if you run multiple "watch" commands. Since lerna run
will execute commands in topologically sorted order, it can end up waiting for a command before moving on. This will block execution when you run "watch" commands, since they typically never end. An example of a "watch" command is running babel
with the --watch
CLI flag.
Install external dependencies matching glob
at the repo root so they're
available to all packages. Any binaries from these dependencies will be
linked into dependent package node_modules/.bin/
directories so they're
available for npm scripts. If the option is present but no glob
is given
the default is **
(hoist everything). This option only affects the
bootstrap
command.
$ lerna bootstrap --hoist
For background on --hoist
, see the hoist documentation.
Note: If packages depend on different versions of an external dependency, the most commonly used version will be hoisted, and a warning will be emitted.
Do not install external dependencies matching glob
at the repo root. This
can be used to opt out of hoisting for certain dependencies.
$ lerna bootstrap --hoist --nohoist=babel-*
Install external dependencies using [client] install
. Must be an executable
that knows how to install npm dependencies.
$ lerna bootstrap --npm-client=yarn
May also be configured in lerna.json
:
{
...
"npmClient": "yarn"
}
Fail immediately if a cycle is found (in bootstrap
, exec
, publish
or run
).
$ lerna bootstrap --reject-cycles
Enables integration with Yarn Workspaces (available since yarn@0.27+).
The values in the array are the commands in which Lerna will delegate operation to Yarn (currently only bootstrapping).
If --use-workspaces
is true then packages
will be overridden by the value from package.json/workspaces.
May also be configured in lerna.json
:
{
...
"npmClient": "yarn",
"useWorkspaces": true
}
The root-level package.json must also include a workspaces
array:
{
"private": true,
"devDependencies": {
"lerna": "^2.2.0"
},
"workspaces": ["packages/*"]
}
This list is broadly similar to lerna's packages
config (a list of globs matching directories with a package.json),
except it does not support recursive globs ("**"
, a.k.a. "globstars").
Stream output from child processes immediately, prefixed with the originating package name. This allows output from different packages to be interleaved.
$ lerna run watch --stream
Similar to --stream
, but completely disregards concurrency and topological sorting, running a given command or script immediately in all matching packages with prefixed streaming output. This is the preferred flag for long-running processes such as babel src -d lib -w
run over many packages.
$ lerna exec --parallel -- babel src -d lib -w
When run with this flag, forwarded npm commands will use the specified registry for your package(s).
This is useful if you do not want to explicitly set up your registry configuration in all of your package.json files individually when e.g. using private registries.
When passed, this flag will alter the default publish process by first publishing
all changed packages to a temporary dist-tag (lerna-temp
) and then moving the
new version(s) to the default dist-tag (latest
).
This is not generally necessary, as Lerna will publish packages in topological order (all dependencies before dependents) by default.
Using Lerna? Add a README badge to show it off:
[![lerna](https://img.shields.io/badge/maintained%20with-lerna-cc00ff.svg)](https://lernajs.io/)
If you prefer some guidance for cli (in case you're about to start using lerna or introducing it to a new team), you might like lerna-wizard. It will lead you through a series of well-defined steps: