pnpm 11.7.0

pnpm 11.7

on Node.js NPM

Minor Changes

• Added a new setting frozenStore (--frozen-store) that lets pnpm install run against a package store on a read-only filesystem (e.g. a Nix store, a read-only bind mount, an OCI layer). When enabled, pnpm opens the store's SQLite index.db through the immutable=1 URI — bypassing the WAL/-shm sidecar creation that otherwise fails on a read-only directory — and suppresses every store-write path (the index.db writer and the project-registry write). Pair it with --offline --frozen-lockfile against a fully-populated store. Under the global virtual store, package directories live inside the store, so if the store is missing the build output of a package whose lifecycle scripts are approved (or that has a patch), pnpm fails up front with ERR_PNPM_FROZEN_STORE_NEEDS_BUILD rather than crashing mid-build on a read-only write — seed the store with those builds first. Incompatible with --force and with a configured pnpr server, since both write into the store; the side-effects cache is likewise not written under frozenStore. If the store is missing its content directory, the install fails fast with ERR_PNPM_FROZEN_STORE_INCOMPLETE rather than attempting to initialize it. The read-only immutable=1 open requires Node.js >=22.15.0, >=23.11.0, or >=24.0.0; on older runtimes --frozen-store fails with a clear ERR_PNPM_FROZEN_STORE_UNSUPPORTED_NODE error. Bin-linking also tolerates a read-only store: under the global virtual store a package's bin source lives inside the store, so the chmod that makes it executable would be refused — with EPERM/EACCES, or with EROFS on a genuinely read-only filesystem. That chmod is redundant when the seed already ships its bins executable with a normalized shebang, so it is now skipped in that case, while a non-executable bin (or one still carrying a Windows CRLF shebang) on a read-only store still errors.

• When pacquet (the Rust port of pnpm) is declared in configDependencies, pnpm now delegates dependency resolution to it too — not just materialization — provided the installed pacquet is new enough to support full resolving installs (>= 0.11.7).

  Previously pacquet only ran in frozen-install mode: pnpm always resolved the dependency graph itself (writing pnpm-lock.yaml) and handed pacquet a finished lockfile to fetch / import / link. With pacquet >= 0.11.7, a non-frozen pnpm install (default isolated nodeLinker, plain install) is delegated to pacquet end-to-end in a single pass — pacquet resolves the manifests, writes the lockfile, and materializes node_modules. pnpm detects the capability from the installed pacquet's version; older pacquet releases keep the resolve-then-materialize split, and add / update / remove still resolve in pnpm (it has to mutate the manifests first). This remains an opt-in preview of the Rust install engine #11723.

• Added a new opt-in --batch flag to pnpm publish --recursive that sends all selected packages to the registry in a single PUT /-/pnpm/v1/publish request instead of one request per package. The target registry has to implement the batch publish endpoint (pnpr does); registries that don't are reported with a clear ERR_PNPM_BATCH_PUBLISH_UNSUPPORTED error. The batch is processed all-or-nothing by pnpr: if any package in the batch fails validation, none of the packages are published.

Patch Changes

• Reject path-traversal and reserved dependency aliases (such as ../../../escape, .bin, .pnpm, or node_modules) that come from a lockfile rather than a freshly resolved manifest. A crafted lockfile alias could otherwise be joined directly under a hoisted node_modules directory, letting package files be written outside the intended install root or overwrite pnpm-owned layout.

  The fix adds two layers:

  • The nodeLinker: hoisted graph builder now validates each alias at the directory sink (safeJoinModulesDir), matching the validation pnpm already performs when resolving aliases from manifests.
  • The lockfile verification gate (verifyLockfileResolutions) now runs an always-on, policy-independent check that rejects any importer or snapshot dependency alias that is not a valid package name, failing the install early — before any fetch or filesystem work — for every node linker at once.

• Made shared package child resolution deterministic when the same package is reached through multiple contexts. pnpm now chooses the shallowest occurrence, then importer order, then parent path, instead of letting request timing decide the child context and missing-peer report pnpm/pnpm#12358.

• Fix garbled summary line after submitting pnpm update -i and pnpm audit --fix -i. The interactive checkbox prompt previously printed every selected choice's full table row (label, current/target versions, workspace, URL) joined by commas, producing a wall of text after pressing Enter. The summary now lists only the selected package names (or vulnerability keys) by setting an explicit short per choice; the in-progress selection UI is unchanged.

• Prevent pnpm patch-remove from removing files outside the configured patches directory.

• Fixed pnpm publish ignoring strictSsl: false when publishing to registries with self-signed certificates. The strictSSL option is now forwarded to libnpmpublish / npm-registry-fetch so that strict-ssl=false in .npmrc or strictSsl: false in pnpm-workspace.yaml is respected during publish, the same way it is for pnpm install pnpm/pnpm#12012.

• Fixed Cannot destructure property 'manifest' of 'manifestsByPath[rootDir]' as it is undefined regression introduced in 11.6.0 when running pnpm add <pkg> outside a workspace on Windows. selectProjectByDir was keying the resulting ProjectsGraph by opts.dir instead of project.rootDir, so downstream manifestsByPath lookups missed when the two paths normalized differently (typically drive-letter casing). pnpm/pnpm#12379

• Git dependencies that point to a subdirectory of a repository (repo#commit&path:/sub/dir) keep their path in the lockfile again. Since the integrity of git-hosted tarballs started being pinned in the lockfile, any install that actually downloaded the tarball rebuilt the lockfile resolution as { integrity, tarball, gitHosted } and dropped the path field, while installs served from the store kept it — so the field disappeared seemingly at random. Without path, later installs from that lockfile silently unpacked the repository root instead of the subdirectory #12304.

• Fixed nondeterministic lockfile output that made pnpm dedupe --check fail intermittently in CI. When a locked peer provider was pinned for a dependency that has no child dependencies of its own, the pinned provider leaked into the shared parent scope, so siblings resolved after it could pick up an optional peer they should not see. Which siblings were affected depended on resolution order, which varies with network timing.

• Sped up pnpm install with a frozen lockfile by running lockfile verification (the policy revalidation gate added for minimumReleaseAge/trustPolicy and the tarball-URL anti-tamper check) concurrently with fetching and linking instead of blocking the whole install on it. Dependency lifecycle scripts are still held back until verification succeeds, so no script runs on an unverified lockfile: if verification fails the install aborts before any dependency build, and if linking finishes first the install waits for the verification verdict before completing.

• User-defined npm_config_* environment variables are now preserved during lifecycle script execution. Previously, all npm_-prefixed env vars were stripped, which caused user-set variables like npm_config_platform_arch to be lost pnpm/pnpm#12399.

• pnpm can now use different auth tokens for different package scopes, even when those scopes use the same registry URL.

  Previously, auth was selected only by registry URL. If @org-a and @org-b both used https://npm.pkg.github.com/, they had to share the same token. This caused problems for registries that issue tokens per organization or per scope.

  Configure a scope-specific token by adding the package scope after the registry URL in the auth key:

  @org-a:registry=https://npm.pkg.github.com/
  @org-b:registry=https://npm.pkg.github.com/
  
  //npm.pkg.github.com/:@org-a:_authToken=${ORG_A_TOKEN}
  //npm.pkg.github.com/:@org-b:_authToken=${ORG_B_TOKEN}
  
  //npm.pkg.github.com/:_authToken=${FALLBACK_TOKEN}

  pnpm login --registry=https://npm.pkg.github.com --scope=@org-a writes the token to the same scope-specific auth key.

  When installing or publishing @org-a/*, pnpm uses ORG_A_TOKEN. For @org-b/*, pnpm uses ORG_B_TOKEN. Packages without a matching scope continue to use the registry-wide fallback token.

• pnpm setup no longer prompts to approve build scripts for @pnpm/exe when installing the standalone executable. pnpm links the platform-specific binary itself, so the package's install scripts are skipped during the global self-install #12377.

• Close lockfile reads deterministically before rewriting lockfiles and keep pacquet's virtual store directory length aligned with pnpm on Windows.

• A 304 Not Modified answer from the registry now renews the cached metadata file's mtime, so the minimumReleaseAge freshness shortcut keeps serving resolutions from the cache. Previously, once a cached packument grew older than minimumReleaseAge, every subsequent install re-validated it against the registry forever, because a 304 never rewrites the file.

• Updated dependency ranges. Notably:

  • @pnpm/logger peer dependency range moved to ^1100.0.0.
  • msgpackr 1.11.8 → 2.0.4 (store index files remain byte-compatible in both directions).
  • open ^7.4.2 → ^11.0.0, memoize ^10 → ^11, cli-truncate ^5 → ^6, pidtree ^0.6 → ^1.
  • @yarnpkg/core 4.5.0 → 4.8.0, @rushstack/worker-pool 0.7.7 → 0.7.18, @cyclonedx/cyclonedx-library 10.0.0 → 10.1.0, @pnpm/config.nerf-dart ^1 → ^2, @pnpm/log.group 3.0.2 → 4.0.1, @pnpm/util.lex-comparator ^3 → ^4.

• Updated @zkochan/cmd-shim to v9.0.6.

• Fixed a Windows-only hang where a failed command could take 20–46 seconds to exit. On error, pnpm enumerates descendant processes (via pidtree) to terminate them, which on Windows shells out to wmic/PowerShell Get-CimInstance Win32_Process — a lookup that is extremely slow on some machines. The lookup is now bounded by a short timeout so it can no longer stall the process exit.

Platinum Sponsors

Gold Sponsors