* docs: lint prose with Vale Add Vale with an in-repo style, .vale/styles/pnpm, that turns the writing rules in AGENTS.md and the Google developer documentation style guide into lint rules. The stock Vale packages (Google, Microsoft, RedHat, write-good) were too noisy to block on, for example 1,300 passive-voice hits in the v12 docs alone, so the rules are written for this repo, following the GitLab model. Each one links the guideline it enforces. Levels: - error, fails CI: dashes, plain-word substitutions, condescending words, "please", "note that", hedges, puffery, optional plurals, exclamation marks, ellipses, curly quotes, generic link text, product spellings, sentence-case headings, repeated words, double spaces, implementation history in comments; in changesets, bold or italic emphasis, internal names and release-line names. - warning: future tense, sentences over 30 words, "instead of" and "so that" tails in changesets. - suggestion: passive voice. .vale.ini scopes the rules. Docs get every rule. Contributor Markdown and code comments keep dashes and headings as warnings until their roughly 8,100 em dashes are rewritten. Released changesets stay in .changeset/ and are not rewritten, so lint-prose.mjs lints only the changesets added since PROSE_BASE (the PR base in CI). The prose workflow installs a pinned, checksum-verified Vale release and runs `pnpm lint:prose`. The rest of this change rewrites the existing docs, READMEs, contributor Markdown, and code comments to pass the error-level rules. Only prose and comment text changed. Renamed headings keep their anchors. Spelling stays with cspell, which checks identifiers that Vale does not see. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: keep prose lint from flattening meaning --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
4.8 KiB
id, title
| id | title |
|---|---|
| symlinked-node-modules-structure | Symlinked `node_modules` structure |
:::info
This article only describes how pnpm's node_modules are structured when
there are no packages with peer dependencies. For the more complex scenario of
dependencies with peers, see how peers are resolved.
:::
pnpm's node_modules layout uses symbolic links to create a nested structure of
dependencies.
Every file of every package inside node_modules is a hard link to the
content-addressable store. Let's say you install foo@1.0.0 that depends on
bar@1.0.0. pnpm will hard link both packages to node_modules like this:
node_modules
└── .pnpm
├── bar@1.0.0
│ └── node_modules
│ └── bar
│ ├── index.js -> <store>/001
│ └── package.json -> <store>/002
└── foo@1.0.0
└── node_modules
└── foo
├── index.js -> <store>/003
└── package.json -> <store>/004
These are the only "real" files in node_modules. Once all the packages are
hard linked to node_modules, symbolic links are created to build the nested
dependency graph structure.
As you might have noticed, both packages are hard linked into a subfolder inside
a node_modules folder (foo@1.0.0/node_modules/foo). This is needed to:
- allow packages to import themselves.
fooshould be able torequire('foo/package.json')orimport * as package from "foo/package.json". - avoid circular symlinks. Dependencies of packages are placed in the same
folder in which the dependent packages are. For Node.js it doesn't make a
difference whether dependencies are inside the package's
node_modulesor in any othernode_modulesin the parent directories.
The next stage of installation is symlinking dependencies. bar is going to be
symlinked to the foo@1.0.0/node_modules folder:
node_modules
└── .pnpm
├── bar@1.0.0
│ └── node_modules
│ └── bar -> <store>
└── foo@1.0.0
└── node_modules
├── foo -> <store>
└── bar -> ../../bar@1.0.0/node_modules/bar
Next, direct dependencies are handled. foo is going to be symlinked into the
root node_modules folder because foo is a dependency of the project:
node_modules
├── foo -> ./.pnpm/foo@1.0.0/node_modules/foo
└── .pnpm
├── bar@1.0.0
│ └── node_modules
│ └── bar -> <store>
└── foo@1.0.0
└── node_modules
├── foo -> <store>
└── bar -> ../../bar@1.0.0/node_modules/bar
This is a very simple example. However, the layout will maintain this structure regardless of the number of dependencies and the depth of the dependency graph.
Let's add qar@2.0.0 as a dependency of bar and foo. This is how the new
structure will look:
node_modules
├── foo -> ./.pnpm/foo@1.0.0/node_modules/foo
└── .pnpm
├── bar@1.0.0
│ └── node_modules
│ ├── bar -> <store>
│ └── qar -> ../../qar@2.0.0/node_modules/qar
├── foo@1.0.0
│ └── node_modules
│ ├── foo -> <store>
│ ├── bar -> ../../bar@1.0.0/node_modules/bar
│ └── qar -> ../../qar@2.0.0/node_modules/qar
└── qar@2.0.0
└── node_modules
└── qar -> <store>
As you may see, even though the graph is deeper now (foo > bar > qar), the
directory depth in the file system is still the same.
This layout might look weird at first glance, but it is completely compatible
with Node's module resolution algorithm. When resolving modules, Node ignores
symlinks, so when bar is required from foo@1.0.0/node_modules/foo/index.js,
Node does not use bar at foo@1.0.0/node_modules/bar, but instead, bar is
resolved to its real location (bar@1.0.0/node_modules/bar). As a consequence,
bar can also resolve its dependencies which are in bar@1.0.0/node_modules.
A great bonus of this layout is that only packages that are really in the
dependencies are accessible. With a flattened node_modules structure, all
hoisted packages are accessible. To read more about why this is an advantage,
see "pnpm's strictness helps to avoid silly bugs"
Unfortunately, many packages in the ecosystem are broken. They use dependencies that are not listed in their package.json. To minimize the number of issues new users encounter, pnpm hoists all dependencies by default into node_modules/.pnpm/node_modules. To disable this hoisting, set hoist to false.