Skip to content

fix(cache): resolve .. before checking that a tracked path is in the workspace - #940

Merged
wan9chi merged 2 commits into
mainfrom
fix/cache-dotdot-paths
Nov 5, 2026
Merged

wan9chi merged 2 commits into
mainfrom
fix/cache-dotdot-paths

Conversation

@wan9chi

@wan9chi wan9chi commented Nov 5, 2026 •

Copy link
Copy Markdown
Member

Motivation

On Linux and macOS, a task that writes outside the workspace through a relative path with .., like cp -r dist ../site publishing a build next to the repo, fails with ✗ Cache update failed: paths in archives must not have ... fspy reports a relative open as the working directory joined with the path, .. included, so the write is <workspace>/../site/index.html. vt stripped the workspace root before resolving .., so the path passed the check that it's in the workspace and became the output ../site/index.html, which the archive rejects. A file read that way became an input outside the workspace, unlike the same file read through an absolute path.

Changes

Tracked paths have . and .. resolved lexically, as before, but before the workspace root is stripped, so such a path is outside the workspace and dropped like any other. A path that stays inside, like packages/sub-pkg/../shared/dist/output.js, is still packages/shared/dist/output.js. vt_path gets clean_and_strip_path_prefix, and fspy's IpcPath a method of the same name next to strip_path_prefix, which is unchanged. Windows resolves .. before fspy sees the path, so it wasn't affected. A unit test covers the order, and the malformed Windows drive path that the cleaning after the strip used to catch.

E2E case: write_outside_workspace_through_dotdot (synthetic, vtt fs mkdir ../site write ../site/index.html hi, where .. is the harness's folder holding the staged workspace). Both runs fail today; after, the first is cached and the second hits. The first commit runs it on Unix only, since Windows already caches it.

A standalone fix for main. The stack that lets automatic tracking cache tasks which read their own outputs and list the folders holding them, so they also hit on a clean checkout, sits on it and no longer checks for .. itself.

🤖 Generated with Claude Code


Note

Medium Risk
Changes how automatic file-access tracking classifies workspace inputs/outputs, which directly affects cache keys and archive contents; behavior is covered by new unit and e2e tests but touches core caching logic.

Overview
On Linux and macOS, cached tasks that write outside the repo via relative paths like ../site (e.g. cp -r dist ../site) no longer fail with Cache update failed: paths in archives must not have ... The second run can cache hit instead.

Tracked path normalization now lexically resolves . and .. before stripping the workspace root, so paths such as <workspace>/../site/index.html are treated as outside the workspace and dropped from inputs/outputs (same as absolute paths outside the repo). In-workspace paths like packages/sub-pkg/../shared/dist/output.js still normalize to packages/shared/dist/output.js.

Adds vt_path::clean_and_strip_path_prefix and IpcPath::clean_and_strip_path_prefix, wired through tracked_workspace_path; the old strip-then-clean() flow is removed. Unit tests cover ordering and the Windows malformed-path case; an e2e fixture asserts a ../site deploy task caches on repeat run.

Reviewed by Cursor Bugbot for commit 827bef8. Bugbot is set up for automated code reviews on this repo. Configure here.

@wan9chi wan9chi added ci: full Run the slow CI jobs (macOS x64, musl) on this pull request read-write overlap Caching tasks that read and write the same paths labels Nov 5, 2026
…` isn't cached

A case for a task that writes outside the workspace through a relative
path with `..`, like `cp -r dist ../site` publishing a build next to the
workspace. In `write_outside_workspace_through_dotdot`, the task makes
`../site` and writes `../site/index.html` through `vtt fs`; `..` is the
folder holding the staged workspace, which the harness removes.

Today fspy reports the write as `<workspace>/../site/index.html`. That
passes the check that a tracked path is in the workspace and becomes the
output `../site/index.html`, so every run fails to update the cache with
`paths in archives must not have ..`.

The case runs on Unix only for now. On Windows, `..` is resolved before
fspy reports the path, so the write is outside the workspace and the task
is already cached there.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…workspace

On Unix, fspy reports a relative path as the working directory joined
with it, `..` included, so a task that writes `../site/index.html` from
the workspace root, like `cp -r dist ../site`, is tracked as writing
`<workspace>/../site/index.html`. vt stripped the workspace root from
that path before resolving `..`, so it passed as `../site/index.html`,
and with automatic output tracking, archiving it failed the run with
`Cache update failed: paths in archives must not have ..`. A file read
that way became an input outside the workspace.

vt now resolves `.` and `..` lexically, as before, but before stripping
the workspace root, so such a path is outside the workspace and skipped
like any other. A path that stays inside, like
`packages/sub-pkg/../shared/dist/output.js`, is still
`packages/shared/dist/output.js`, and one that leaves the workspace and
comes back, like `<workspace>/../<workspace folder>/a`, is now `a`.
`vt_path` gets `clean_and_strip_path_prefix`, which fspy's `IpcPath`
offers next to `strip_path_prefix`.

`write_outside_workspace_through_dotdot` now runs on Windows too. There
`..` is resolved before fspy sees the path, so the task was cached
already, and the case now has the same result on every platform.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@wan9chi
wan9chi force-pushed the fix/cache-dotdot-paths branch from 5c87b9e to 827bef8 Compare November 5, 2026 08:53
@wan9chi
wan9chi added this pull request to stack #942 November 5, 2026 08:55
@github-actions

github-actions Bot commented Nov 5, 2026

Copy link
Copy Markdown

fspy benchmark

linux

dynamic/launch             change  -0.18%  [ -5.58% ..  +6.45%]  overhead   +57.28%
dynamic/access             change  -1.08%  [ -7.60% ..  +4.95%]  overhead   +19.73%
dynamic/access-relative    change  +0.25%  [ -4.98% ..  +5.64%]  overhead   +67.33%
dynamic/access-contended   change  -2.42%  [-16.19% ..  +9.18%]  overhead    +9.26%
static/launch              change  +1.42%  [ -6.49% ..  +9.24%]  overhead  +183.01%
static/access              change  +0.35%  [ -5.12% .. +15.05%]  overhead +1445.53%
static/access-relative     change  +0.41%  [ -2.19% ..  +2.74%]  overhead +1979.63%
static/access-contended    change  -1.27%  [ -6.48% ..  +3.12%]  overhead +1084.89%

macos

dynamic/launch             change  -0.41%  [ -5.06% ..  +4.72%]  overhead  +237.06%
dynamic/access             change  +0.35%  [ -1.76% ..  +2.91%]  overhead    +3.28%
dynamic/access-relative    change  -0.52%  [ -7.62% ..  +7.73%]  overhead  +267.22%
dynamic/access-contended   change  +1.72%  [ -6.93% ..  +5.93%]  overhead    +0.49%

windows

dynamic/launch             change  -0.48%  [ -6.92% ..  +4.27%]  overhead   +26.97%
dynamic/access             change  +0.00%  [ -1.10% ..  +1.47%]  overhead    +1.48%
dynamic/access-relative    change  +0.17%  [ -1.96% ..  +1.24%]  overhead    +1.27%
dynamic/access-contended   change  +1.26%  [ -9.82% .. +36.22%]  overhead    +1.74%

@wan9chi
wan9chi merged commit 6570fdc into main Nov 5, 2026
23 checks passed
@wan9chi
wan9chi deleted the fix/cache-dotdot-paths branch November 5, 2026 10:44
wan9chi added a commit that referenced this pull request Nov 5, 2026
## Motivation

Follow-ups to the rebuildable-inputs stack (#913–#919, #940, #941), now
merged:

- The real-world case for files renamed into place (#916) used Vitest's
`fsModuleCache`, which renames cache entries rather than outputs.
- #919's e2e cases test a listing entry like any other, so they add no
coverage of their own.
- The stack added eight changelog entries, most of them for what users
see as one feature.
- Socket flagged Astro 6.4.8, which #917 added for its real-world case:
critical advisory GHSA-26w7-cxv4-gfx2 (remote code execution through
AVIF image optimization), fixed in Astro 7.2.8.

## Changes

- New case `next_build_moved_pages`: Next.js 15's `next build` renders
each static page (`index.html`, `404.html`, `500.html`) into
`.next/export` and renames it into `.next/server/pages`. After `.next`
is removed, the next run is a cache hit that restores the pages. The
build log is redacted, since its progress lines, sizes and chunk hashes
vary. It replaces `vitest_fs_module_cache`, along with the SHA-1
redaction only that case used. `next`, `react` and `react-dom` join
`packages/tools`.
- `dist_appears_in_listed_folder` and `oxlint_dist` are removed.
- The changelog entries of #913–#918 are combined into one **Added**
item for rebuildable inputs; #919 (`dist` in listings) and #941 (input
exclusions apply to listings) stay their own short items, and so does
#940's fix.
- On Windows CI, the e2e harness stages fixtures in `RUNNER_TEMP`, on
the checkout's drive: the staged workspaces resolve `node_modules` into
the checkout, and Next.js builds its client entry from a path relative
to the project, which can't cross drives.
- Astro 6.4.8 → 7.3.8, the newest release that still supports the repo's
Node 22.19. Astro 7 needs upstream Vite 8, so its scoped override now
points at Vite 8.3.2. `astro_empty_collections` passes unchanged; its
comment now says Astro 7 stats the project root rather than listing it.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

<!-- CURSOR_SUMMARY -->
---

> [!NOTE]
> **Low Risk**
> Test harness, changelog, and dependency catalog changes only; no
changes to task caching or file-tracking behavior in this diff.
> 
> **Overview**
> Follow-up after the rebuildable-inputs stack lands: **no runtime cache
logic changes**—mostly release notes, snapshot tests, and toolchain
bumps.
> 
> The **changelog** folds the former #913–#918 narrative into one
**Added** “rebuildable inputs” entry and shortens the #919 / #941
**Changed** bullets; #940 stays separate.
> 
> **E2E snapshots** swap coverage: **`next_build_moved_pages`** (Next.js
15 renaming `.next/export` → `.next/server/pages`, cache hit after
deleting `.next`) replaces **`vitest_fs_module_cache`**;
**`dist_appears_in_listed_folder`** and **`oxlint_dist`** are dropped as
redundant with listing behavior elsewhere. Snapshot **redaction** drops
SHA-1 normalization (Vitest-only). On **Windows CI**, fixtures stage
under **`RUNNER_TEMP`** so Next’s client entry paths stay on the
checkout drive.
> 
> **Deps**: catalog adds **Next 15**, **React 19**, bumps **Astro 6 →
7.3.8** (security advisory), and scopes **`astro>vite`** to upstream
**Vite 8.3.2**. The Astro empty-collections fixture comment reflects
Astro 7 **stat**-ing the project root instead of listing it.
> 
> <sup>Reviewed by [Cursor Bugbot](https://cursor.com/bugbot) for commit
4c8f40c. Bugbot is set up for automated
code reviews on this repo. Configure
[here](https://www.cursor.com/dashboard/bugbot).</sup>
<!-- /CURSOR_SUMMARY -->

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
fengmk2 added a commit to voidzero-dev/vite-plus that referenced this pull request Nov 11, 2026
Vite+ 1.6.0 caches `vp run` tasks that rewrite files they read, previews
runs with `vp run --dry-run` and JSON output, lets `vp check
--continue-on-error` report every failed check, and sets `VP_USER_AGENT`
for the tools that `vp` starts.

### Features

- `vp run` caches tasks that read files they also write, such as `tsdown
--exports` rewriting `package.json` or `tsc --incremental` reading and
writing `tsconfig.tsbuildinfo`. The files a task writes, and folders
that hold only those files, may be missing on the next run, as on a
fresh clone, and a cache hit restores them with the task's outputs,
including the symbolic links the task created
([docs](https://viteplus.dev/guide/automatic-data-tracking#rebuildable-inputs))
([vite-task#913](voidzero-dev/vite-task#913),
[vite-task#914](voidzero-dev/vite-task#914),
[vite-task#915](voidzero-dev/vite-task#915),
[vite-task#916](voidzero-dev/vite-task#916),
[vite-task#917](voidzero-dev/vite-task#917),
[vite-task#918](voidzero-dev/vite-task#918)), by
@wan9chi.
- `vp run --dry-run` shows each command that a run would start and
whether it would hit the cache, without running anything or writing to
the cache. `vp run --json` prints every package and task in the
workspace, and `vp run --dry-run --json` prints the run plan with each
step's cache status, as one JSON document for scripts, CI, and coding
agents. `vp run -v` lists every change behind a cache miss, including
each changed input file, and `-vv` also lists the unchanged inputs and
the outputs that a hit restores
([docs](https://viteplus.dev/guide/run#dry-runs-and-json-output))
([vite-task#948](voidzero-dev/vite-task#948),
[vite-task#960](voidzero-dev/vite-task#960),
[vite-task#961](voidzero-dev/vite-task#961),
[vite-task#963](voidzero-dev/vite-task#963),
[vite-task#966](voidzero-dev/vite-task#966),
[vite-task#967](voidzero-dev/vite-task#967),
[vite-task#968](voidzero-dev/vite-task#968),
[vite-task#969](voidzero-dev/vite-task#969)), by
@wan9chi.
- `vp check --continue-on-error` runs the remaining enabled checks after
a failure, so one run reports formatting, lint, and type errors
together, and exits with the status of the first failed check. Without
the flag, `vp check` still stops at the first failed check
([#3367](#3367)), by
@chbaefront.
- `vp`, `vpr`, and `vpx` set `VP_USER_AGENT=vp/<version>` for their
child processes, so tools can detect that Vite+ started them. When the
global CLI delegates to a project's local CLI, the local version is
reported ([docs](https://viteplus.dev/guide/global-cli#vp-user-agent))
([#1958](#1958)), by
@fengmk2.
- First part of `VP_ENV_FORBIDDEN=1`, for systems such as NixOS that
provide Node.js and package managers themselves: with it set, Vite+
never downloads a Node.js or package-manager version, shims and `vpx`
use only the system tools on `PATH`, and installers and `vp upgrade` use
the system `node` and `pnpm` for Vite+'s own dependencies. In this
release, commands that delegate to a package manager, such as `vp
install` and `vp run`, fail with `DisabledByEnvForbidden` while it is
set. A follow-up change is planned to cover them and document the
variable ([#3233](#3233)),
by @myuron.
- Upgrade the bundled Oxfmt and Oxlint
([#3416](#3416)), by
@voidzero-guard[bot]:
- `oxfmt@0.76.0` to `oxfmt@0.78.0`. Oxfmt formats Vue files with its own
formatter instead of Prettier, so the output for Vue templates can
change. It also fixes Tailwind class sorting in several layouts, Less
escapes and `%` selectors, and TypeScript in Astro files.
- `oxlint@1.91.0` to `oxlint@1.93.0` and `oxlint-tsgolint@7.0.2005` to
`oxlint-tsgolint@7.0.2006`. `no-redeclare` gets an
`ignoreDeclarationMerge` option, and `promise/prefer-catch` offers a
suggestion.
- Oxlint and Oxfmt resolve an unset or relative `root` in
`vite.config.ts` from the directory of the config file instead of the
current directory, so editor integrations in multi-root workspaces load
the correct project.
- Vite+ installs its own dependencies during setup and `vp upgrade` with
`pnpm@10.34.9` (was `pnpm@10.34.7`).
- The new Oxfmt and Oxlint can flag code that passed before. If your CI
runs `vp check`, run `vp check --fix --continue-on-error` after you
upgrade: it formats files, applies the available lint fixes, formats
again, and reports every error that remains. For example,
`unicorn/prefer-type-error` now also reports a `throw` in an `if` branch
without braces, and this command rewrites it.

### Fixes & Enhancements

- While a Vite+ spinner is visible, such as `Installing global packages`
in `vp install -g`, Vite+ hides the cursor and discards typed keys.
Pressing Enter no longer leaves copies of the spinner row, and typed
text no longer runs in the shell after the command ends. Ctrl+C during a
spinner now ends the command with the usual signal status instead of
exit code 0, and the global package spinners show the elapsed time
([#3373](#3373),
[#3417](#3417)), by
@jong-kyung and @fengmk2.
- `vp run` loads a package's `vite.config.*` only when the run reaches
one of that package's tasks. A run in one package no longer waits for
the configs of other packages, a broken config in another package no
longer fails it, and `vp run <task>` in a package that does not define
the task reports `Task "<task>" not found` instead of another package's
config error
([vite-task#944](voidzero-dev/vite-task#944),
[vite-task#988](voidzero-dev/vite-task#988)), by
@wan9chi.
- `vp run nope#build`, where no workspace package has the name `nope`,
fails with `Package 'nope' not found in the workspace` and exits with
code 1, instead of running nothing and exiting with code 0. A short name
that matches several scoped packages fails as ambiguous
([vite-task#984](voidzero-dev/vite-task#984)),
by @wan9chi.
- Folder listings no longer skip entries named `dist`, so a `dist`
folder that appears or disappears in a folder a task lists makes the
next run a cache miss. Exclusions in a task's `input`, such as
`"!dist/**"`, now also apply to the folders that the task lists. In a
new monorepo from `vp create`, every task hits the cache from the third
`vp run ready`, because the first build creates the `dist` folders
([vite-task#919](voidzero-dev/vite-task#919),
[vite-task#941](voidzero-dev/vite-task#941)), by
@wan9chi.
- Fewer unnecessary cache misses and cache failures in `vp run`: on
Windows, files that a task deletes or moves are no longer among its
inputs. Tasks that read `node_modules/.vite/task-cache`, as Parcel and
Gatsby do, hit the cache. A task that writes outside the workspace
through `..`, like `cp -r dist ../site`, no longer fails with `Cache
update failed`
([vite-task#937](voidzero-dev/vite-task#937),
[vite-task#940](voidzero-dev/vite-task#940),
[vite-task#980](voidzero-dev/vite-task#980)), by
@wan9chi.
- A cache hit restores the symbolic links that a task's `output` globs
match as links, instead of dropping them
([vite-task#885](voidzero-dev/vite-task#885)),
by @jasonkuhrt.
- When several tasks become ready at the same time, `vp run` checks
their cache entries in parallel, so cache hits start sooner
([vite-task#982](voidzero-dev/vite-task#982)),
by @wan9chi.
- On macOS, task commands that use `sh` or `bash` behave as on Linux: a
`while read` loop with a redirect under `set -e` no longer exits the
script, an `EXIT` trap sees the exit status, so a failed task is no
longer reported as successful, and `$BASH` is set
([vite-task#939](voidzero-dev/vite-task#939)),
by @lifeiscontent.
- The helper files that `vp run` writes to the system temp directory go
in one folder that only the current user can write to, such as
`/tmp/vp_run-1000` on Linux. On Linux, a different Vite+ version run by
another user no longer fails
([vite-task#947](voidzero-dev/vite-task#947),
[vite-task#974](voidzero-dev/vite-task#974)), by
@wan9chi.
- A `getEnvs` query from `@voidzero-dev/vite-task-client` that matches
every name (`{ prefix: '' }` or `*`) returns nothing and no longer adds
every environment variable to the cache key
([vite-task#976](voidzero-dev/vite-task#976)),
by @wan9chi.
- `vp migrate` no longer writes package specs that the package manager
cannot resolve, such as a `@vitest/*` `catalog:` reference without a
catalog entry or a top-level catalog on Bun before 1.2.21. Bun projects
with an empty `workspaces` field get catalog references, and child
packages with a plain `vite-plus` range get the same concrete version as
the root when the project cannot use catalogs
([#3397](#3397)), by
@jong-kyung.
- `vp create` uses the package manager version that `VP_PACKAGE_MANAGER`
selects for the generated settings, the installation, and its summary.
Standalone pnpm projects get `packages: ['.']` in an existing
`pnpm-workspace.yaml` only for pnpm before 10.5.0
([#3197](#3197)), by
@evan-lc and @fengmk2.
- `vp` sends a `User-Agent` header on all HTTP requests, so Node.js
mirrors that reject requests without one, such as `mirrors.aliyun.com`,
work ([#3425](#3425)), by
@naokihaba.
- `vp env pin` prints `Updated` when it overwrites an existing
`.node-version` or `.nvmrc` file and `Created` when it writes a new one
([#3377](#3377)), by
@CheHyeonYeong.

### Docs

- Describe rebuildable inputs in the Automatic Data Tracking guide, and
dry runs, JSON output, and `-vv` in the `vp run` guide
([#3419](#3419),
[#3421](#3421)), by
@wan9chi.

### Chore

- Bump vite-task to `cb1ea8d`, with its user-facing changes listed
above, and list `--dry-run`, `--json`, and `-vv` in `vp run --help`
([#3418](#3418)), by
@wan9chi.
- The daily dependency upgrade also updates the pnpm version that the
installers pin, after it checks the release integrity
([#3407](#3407)), by
@fengmk2.
- Warm up the monorepo cache check in the `vp create` CI job under pnpm
and Bun ([#3423](#3423)),
by @wan9chi.
- Remove an unused CI job that the required `done` check did not cover
([#3403](#3403)), by
@jong-kyung.
- Remove an incorrect comment about export pattern order in the CLI
build script
([#3405](#3405)), by
@alexicum.

### Bundled Versions

| Tool | Version | Source | Changelog |
| --------------- | ---------- |
----------------------------------------------------------------------------------------------------
|
-----------------------------------------------------------------------------------------------------------------------------------------------
|
| vite | `8.3.7` |
[`5def936`](vitejs/vite@5def936)
| unchanged |
| rolldown | `1.2.18` |
[`f9a49d6`](rolldown/rolldown@f9a49d6)
| unchanged |
| oxc-node | `1.0.1` |
[`7ff14c3`](oxc-project/oxc-node@7ff14c3)
| unchanged |
| tsdown | `0.23.4` | [npm](https://npmx.dev/package/tsdown/v/0.23.4) |
unchanged |
| vitest | `5.1.0` | [npm](https://npmx.dev/package/vitest/v/5.1.0) |
unchanged |
| oxlint | `1.93.0` | [npm](https://npmx.dev/package/oxlint/v/1.93.0) |
[1.92.0](https://github.com/oxc-project/oxc/releases/tag/oxlint_v1.92.0),
[1.93.0](https://github.com/oxc-project/oxc/releases/tag/oxlint_v1.93.0)
|
| oxlint-tsgolint | `7.0.2006` |
[npm](https://npmx.dev/package/oxlint-tsgolint/v/7.0.2006) |
[7.0.2006](https://github.com/oxc-project/tsgolint/releases/tag/v7.0.2006)
|
| oxfmt | `0.78.0` | [npm](https://npmx.dev/package/oxfmt/v/0.78.0) |
[0.77.0](https://github.com/oxc-project/oxc/releases/tag/oxfmt_v0.77.0),
[0.78.0](https://github.com/oxc-project/oxc/releases/tag/oxfmt_v0.78.0)
|

### Upgrade

```bash
vp upgrade
```

If `vp upgrade` must download Node.js and `VP_NODE_DIST_MIRROR` points
to a mirror that rejects requests without a `User-Agent` header, such as
`mirrors.aliyun.com`, unset it or point it to another mirror for this
upgrade. The installed `vp` downloads Node.js and does not send the
header yet.

### New Contributors

@alexicum, @evan-lc, @CheHyeonYeong

**Full Changelog**:
v1.5.1...v1.6.0

---

For a stable release, complete the `HomebrewFormula/vp.rb` update after
publication by merging the generated `release/homebrew-v1.6.0` PR.

Merging this PR will trigger the release workflow.

---------

Co-authored-by: voidzero-guard[bot] <278573678+voidzero-guard[bot]@users.noreply.github.com>
Co-authored-by: MK (fengmk2) <fengmk2@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ci: full Run the slow CI jobs (macOS x64, musl) on this pull request read-write overlap Caching tasks that read and write the same paths

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant