npm-packlist

What is npm-packlist?

The npm-packlist package is used to generate a list of files that would be included in an npm package publish. This includes filtering out files that are not meant to be published, such as those specified in `.npmignore` or the `.gitignore` files, and including those specified in the `files` array in `package.json`. It's useful for package authors to understand and control what gets included in their published npm package.

What are npm-packlist's main functionalities?

Generate packlist

This feature allows you to generate a list of files that would be included if you were to publish the current package. The code sample demonstrates how to use npm-packlist to get an array of file paths that are included in the package's publish list.

const packlist = require('npm-packlist');
packlist().then(files => {
  console.log(files)
});

Customize packlist with package.json

While not a direct feature of npm-packlist, the package respects the `files` field in `package.json`. This allows users to explicitly specify which files or directories should be included in the publish. The code sample shows how to specify a list of files and directories in `package.json` that should be included in the npm package.

{
  "files": [
    "lib/**/*",
    "bin/*",
    "README.md"
  ]
}

Other packages similar to npm-packlist

glob

The 'glob' package provides functionality to match files using the patterns the shell uses, like stars and stuff. It's similar to npm-packlist in that it can be used to select files to include in a package, but it does not directly integrate with npm's packaging logic or respect `.npmignore` and `files` in `package.json`.

ignore

The 'ignore' package is used to filter file paths according to the rules found in `.gitignore` and `.npmignore` files. It's similar to npm-packlist in its purpose of filtering out files, but it's more focused on the ignore rules and doesn't directly generate a list of files for npm packaging.

README

npm-packlist

Get a list of the files to add from a folder into an npm package.

These can be handed to tar like so to make an npm package tarball:

const Arborist = require('@npmcli/arborist')
const packlist = require('npm-packlist')
const tar = require('tar')
const packageDir = '/path/to/package'
const packageTarball = '/path/to/package.tgz'

const arborist = new Arborist({ path: packageDir })
arborist.loadActual().then((tree) => {
  packlist(tree)
    .then(files => tar.create({
      prefix: 'package/',
      cwd: packageDir,
      file: packageTarball,
      gzip: true
    }, files))
    .then(_ => {
      // tarball has been created, continue with your day
    })
  })

This uses the following rules:

1. If a package.json file is found, and it has a files list, then ignore everything that isn't in files. Always include the root readme, license, licence and copying files, if they exist, as well as the package.json file itself. Non-root readme, license, licence and copying files are included by default, but can be excluded using the files list e.g. "!readme".

2. If there's no package.json file (or it has no files list), and there is a .npmignore file, then ignore all the files in the .npmignore file.

3. If there's no package.json with a files list, and there's no .npmignore file, but there is a .gitignore file, then ignore all the files in the .gitignore file.

4. Everything in the root node_modules is ignored, unless it's a bundled dependency. If it IS a bundled dependency, and it's a symbolic link, then the target of the link is included, not the symlink itself.

5. Unless they're explicitly included (by being in a files list, or a !negated rule in a relevant .npmignore or .gitignore), always ignore certain common cruft files:

   1. .npmignore and .gitignore files (their effect is in the package already, there's no need to include them in the package)
   2. editor junk like .*.swp, ._* and .*.orig files
   3. .npmrc files (these may contain private configs)
   4. The node_modules/.bin folder
   5. Waf and gyp cruft like /build/config.gypi and .lock-wscript
   6. Darwin's .DS_Store files because wtf are those even
   7. npm-debug.log files at the root of a project

   You can explicitly re-include any of these with a files list in package.json or a negated ignore file rule.

Only the package.json file in the very root of the project is ever inspected for a files list. Below the top level of the root package, package.json is treated as just another file, and no package-specific semantics are applied.

Interaction between package.json and .npmignore rules

In previous versions of this library, the files list in package.json was used as an initial filter to drive further tree walking. That is no longer the case as of version 6.0.0.

If you have a package.json file with a files array within, any top level .npmignore and .gitignore files will be ignored.

If a directory is listed in files, then any rules in nested .npmignore files within that directory will be honored.

For example, with this package.json:

{
  "files": [ "dir" ]
}

a .npmignore file at dir/.npmignore (and any subsequent sub-directories) will be honored. However, a .npmignore at the root level will be skipped.

Additionally, with this package.json:

{
  "files": ["dir/subdir"]
}

a .npmignore file at dir/.npmignore will be honored, as well as dir/subdir/.npmignore.

Any specific file matched by an exact filename in the package.json files list will be included, and cannot be excluded, by any .npmignore files.

API

Same API as ignore-walk, except providing a tree is required and there are hard-coded file list and rule sets.

The Walker class requires an arborist tree, and if any bundled dependencies are found will include them as well as their own dependencies in the resulting file set.