- 09 Oct, 2026 18 commits
-
-
next-js-bot[bot] authored
-
Jiwon Choi authored
### Why? Upgrade outcomes lack version and runtime context, so failures and missing results cannot be compared by upgrade path or environment. ### How? Add metadata to the existing run-linked events: - Run start: the app's starting Next.js version, the invoking CLI's Node version, and the package manager and its version when available. - CLI result: the starting and approved target Next.js versions. - Agent result: the installed app Next.js version when the result is reported. Keep existing run IDs, agent detection/selection, and policy attribution so analyses can group outcomes by version path, Node version, agent, policy, and package manager. Reuse the upgrade commands' package-manager selection. Version lookup failures warn and leave metadata null without stopping the upgrade. Targets are recorded only after preparation succeeds; copied or printed prompts may have no identified agent. Analyses should retain unknown metadata in their denominators. The receiver and compression schemas must be updated and deployed to retain the new fields. Historical events cannot gain metadata they never recorded. Full CLI execution verification is blocked by missing workspace dependencies and build outputs. <!-- NEXT_JS_LLM -->
-
Aurora Scharff authored
## Summary `printTreeView` copied generated-route lists several times and created a temporary object for every route before printing a small preview in the build output. This change writes the preview directly to the output table. When routes need to be reordered, it sorts one list of numeric indexes instead of copying the routes into objects. The build output does not change. The function creates less temporary data and puts less pressure on garbage collection during builds with many generated routes. ## Impact In a focused A/B of the compiled `printTreeView` with 50,000 generated routes, median temporary heap growth per call dropped: - from 460 KB to 68 KB when preserving route order (-85%) - from 4.53 MB to 2.14 MB when ordering slow routes first (-53%) These results were repeated across five fresh Node.js processes with 50 samples per process. This is a memory-allocation improvement; no CPU-time improvement is claimed. ## Verification - `pnpm build-all` - `HEADLESS=true pnpm test-start-turbo test/e2e/app-dir/build-output-truncated-generated-paths/build-output-truncated-generated-paths.test.ts` - `HEADLESS=true pnpm test-start-webpack test/e2e/app-dir/build-output-truncated-generated-paths/build-output-truncated-generated-paths.test.ts` <!-- NEXT_JS_LLM -->
-
Tobias Koppers authored
## Summary The `create-pr` skill is shared by every coding agent working in this repo, but it hard-coded the `codex/` branch prefix, so Claude, Fleet, and other agents were told to name branches as if they were Codex. The skill now uses an agent-agnostic `<agent-name>/<short-topic>` convention: agents prefix branches with their own name as a single lowercase word, with `codex/`, `claude/`, and `fleet/` given as examples. The user can still override the name. The frontmatter description is updated to match so skill matching still covers branch naming. The Codex-app git directive guidance is left as is, since it is genuinely Codex-specific. ## Verification - `prettier --check .agents/skills/create-pr/SKILL.md` - `git diff --check` <!-- NEXT_JS_LLM --> <!-- fleet e712205b-2273-4aec-9b25-14aeb34efaba --> Co-authored-by:
vercel-fleet-prod[bot] <318278635+vercel-fleet-prod[bot]@users.noreply.github.com> Co-authored-by:
Tobias Koppers <1365881+sokra@users.noreply.github.com>
-
Sebastian "Sebbie" Silbermann authored
This is based on a vulnerability report we got without a reproduction. The test shows `GoogleTagManager` is not vulnerable to XSS attacks. `<GoogleTagManager dataLayer={...}>` serializes user-controlled values into its inline init script via `JSON.stringify` without HTML escaping. If such a value reaches an HTML document (or an HTML fragment parser in a non-script context), a `</script>` sequence breaks out of the script tag and the injected markup executes (XSS). These assertions encode the safe behavior: the payload must never appear unescaped in the served HTML, and it must never execute in the browser. Closes https://linear.app/vercel/issue/NSEC-263 Closes https://github.com/vercel/next-js-mirror/pull/181 -
Sebastian "Sebbie" Silbermann authored
Vercel's proxy service runs the image optimizer transform step in a child process and imports `next/dist/server/image-optimizer/transform` directly, which pins an internal file layout that should be free to change. This PR exposes the module as the public subpath `next/image-optimizer-transform`, following the existing `next/og` pattern of a root-level pass-through plus a type stub registered in the package `files` array. The new entry also re-exports `ImageError` and `getMaxAge` (both already in the module's dependency graph) so consumers can drop every deep `next/dist/server/image-optimizer/*` import, and the module header now states that its location and export surface are a stable contract. The existing unit test imports the module through the new subpath to keep it exercised, and a new API reference page documents `imageOptimizerTransform`, `optimizeImage`, `getSharp`, `getMaxAge`, and `ImageError`. Verified with `pnpm test-unit test/unit/image-optimizer/transform.test.ts` (passes through the new subpath), `pnpm --filter=next types`, and an `npm pack --dry-run` confirming both stub files ship in the tarball. The 22 failures in the broader unit run reproduce on clean canary and are unrelated to this change.
-
Tobias Koppers authored
### What? Registers chunk groups in the module graph exactly as output code chunks them, in two places: - **Pages layout segments**: `_document` and `_app` are registered as their own `Entry` groups (`layout_chunk_group_entry`), in both the per-page (dev) graph chain and the whole-app `entries()`. `internal_ssr_chunk` chunks them as `ChunkGroup::Entry(vec![layout])` instead of `ChunkGroup::Shared(layout)`. - **turbopack-cli dev**: `create_web_entry_source` registers one `Entry` group per entry, with the same entries `DevHtmlAsset` chunks: the runtime entries followed by the entry for evaluatable modules, or the module alone otherwise. It passes `runtime_entries` without `with_entry(entry)`, because `DevHtmlAsset` appends the entry itself. ### Why? A chunk group is identified by its key, and for an `Entry` group that is its ordered entries. The next PR in the stack looks every chunked group up in `ChunkGroupInfo` by key, so a chunked group the graph didn't register fails. - Pages layouts were declared `Shared(layout)`. The whole-app graph also registers the standalone `/_document` and `/_app` endpoints' `Entry([module])` group for the same module and keeps only one root group per module, so `Shared(layout)` wasn't registered there. - turbopack-cli registered a single `Entry` group with all entries and runtime entries, while `DevHtmlAsset` chunks one group per entry (with the entry appended twice). No output change is intended: the same modules are chunked, but the registered groups now match them. This is PR 3 of an eight-PR stack (on top of #99100, already merged): 1. #98733 — canonical chunk-group identity 2. #99686 — chunk several chunk groups in one `chunk_group` call 3. **This PR** — register the chunk groups that are chunked 4. #98734 — use graph-derived merged groups 5. #99750 — chunking takes a `ChunkGroupKey` 6. #98823 — Pages Router dev client graph chaining 7. #98685 — chunk-group availability bitmaps 8. #99050 — keyed chunk-group index lookups ### Verification - Workspace `cargo clippy --all-targets -D warnings`, `cargo check --workspace --all-targets`, `cargo fmt --check` - Turbopack snapshots: 132 passed, no snapshot changes - `turbo-tasks` + `turbopack-core` + `next-core` unit tests: 362 passed - Turbopack production and dev `404-page-router` (custom `_document` / `_app`): 25/25, 12/12 <!-- NEXT_JS_LLM --> <!-- fleet 4f67c6f7-bbe2-454b-9a17-822a60ef8f7e --> Co-authored-by:
vercel-fleet-prod[bot] <318278635+vercel-fleet-prod[bot]@users.noreply.github.com> Co-authored-by:
Tobias Koppers <1365881+sokra@users.noreply.github.com>
-
Tobias Koppers authored
### What? `ChunkingContext::chunk_group` takes a list of chunk groups and chunks them together, as one chunk group, in a single chunking task. It produces one `ChunkGroupResult` (one `make_chunks` run) for the whole list. `evaluated_chunk_group`, `entry_chunk_group` and the `ChunkingContextExt` helpers still take a single group and wrap it in a one-element list. All existing callers pass a one-element list, so **this PR changes no output** (zero snapshot changes). ### Why? Review feedback on #98734 ([1](https://github.com/vercel/next.js/pull/98734#discussion_r4182977091), [2](https://github.com/vercel/next.js/pull/98734#discussion_r4182978846)): a layout segment's client references can derive several isolated merged chunk groups, and #98734 chunked each one with a separate `chunk_group` call. That costs an extra chunking task per group and threads availability by hand. This PR adds the list form; #98734 then calls `chunk_group` once per segment and side with all of the segment's derived groups. ### How? `make_chunk_group` / `chunk_group_content` / `chunk_group_content_operation` take the list as a `SmallVec<[ChunkGroup; 1]>`: almost every call passes a single group, so the common case doesn't heap-allocate. The list crosses turbo-task boundaries, so this PR also adds `impl TaskInput for SmallVec<A>` in turbo-tasks, mirroring the `Vec` impl. Semantics: - **Traversal:** the entries of each group, in list order, with one shared traversal state. A module already reached from an earlier group isn't collected again, which matches what the previous one-call-per-group sequence produced (CSS order unchanged). - **Availability:** every listed group is added to the resulting availability info. - **Entry groups:** if several `Entry` groups are listed, the active page entries and the resulting entry group are the union of their modules, in list order. - **Edge cases:** an empty list is an error; groups listed more than once (same key) are chunked once. Unit tests cover the edge cases, the `Entry` union, and `SmallVec` as a task input (inline and spilled). This is PR 2 of an eight-PR stack (on top of #99100, already merged): 1. #98733 — canonical chunk-group identity 2. **This PR** — chunk several chunk groups in one `chunk_group` call 3. #99749 — register the chunk groups that are chunked 4. #98734 — use graph-derived merged groups 5. #99750 — chunking takes a `ChunkGroupKey` 6. #98823 — Pages Router dev client graph chaining 7. #98685 — chunk-group availability bitmaps 8. #99050 — keyed chunk-group index lookups <!-- NEXT_JS_LLM --> <!-- fleet 4f67c6f7-bbe2-454b-9a17-822a60ef8f7e --> Co-authored-by:
vercel-fleet-prod[bot] <318278635+vercel-fleet-prod[bot]@users.noreply.github.com> Co-authored-by:
Tobias Koppers <1365881+sokra@users.noreply.github.com>
-
Tobias Koppers authored
### What? Makes `ChunkGroupInfo` resolve chunk groups through their canonical `ChunkGroupKey` identity. ### Why? Merged chunk groups carry an entry list whose order is intentionally unspecified. Looking up the full `ChunkGroup` value therefore makes identity depend on incidental entry ordering, even though the module graph already indexes groups by parent and merge tag. This is PR 1 of an eight-PR stack (on top of #99100, already merged): 1. **This PR** — canonical chunk-group identity 2. #99686 — chunk several chunk groups in one `chunk_group` call 3. #99749 — register the chunk groups that are chunked 4. #98734 — use graph-derived merged groups 5. #99750 — chunking takes a `ChunkGroupKey` 6. #98823 — Pages Router dev client graph chaining 7. #98685 — chunk-group availability bitmaps 8. #99050 — keyed chunk-group index lookups It provides stable identity for the graph-derived chunk-group work above it. ### How? Adds a canonical key conversion for every chunk-group variant and uses the existing `chunk_group_keys` index for lookup. Merged keys intentionally contain only their parent group and merge tag. ### Verification - `cargo test -p turbopack-core` — 160 passed - `cargo test -p turbopack-tests --test snapshot` — 125 passed - Affected Cargo checks and Rust formatting <!-- NEXT_JS_LLM --> <!-- fleet 4f67c6f7-bbe2-454b-9a17-822a60ef8f7e --> Co-authored-by:
vercel-fleet-prod[bot] <318278635+vercel-fleet-prod[bot]@users.noreply.github.com> Co-authored-by:
Tobias Koppers <1365881+sokra@users.noreply.github.com>
-
Tobias Koppers authored
## Summary Adds zstd compression flags for `NEXT_TURBOPACK_TRACING`, next to `gz` and `gz-best`: - `zstd`: level 3. Recommended: about the same writer CPU as `gz`, about 25% smaller files. - `zstd-fast`: level 1, the lowest-CPU option, still smaller than `gz`. - `zstd-best`: level 19. The trace writer drops the encoder on exit (`auto_finish`), so the zstd frame is complete. If several compression flags are given, the last one wins, as before. The default stays uncompressed, and the trace format is unchanged. `zstd` was already in the dependency graph, so no new crate is compiled. `contributing/turbopack/tracing.md` documents the flags. It also notes that `zstd-best` and `gz-best` can fall behind on large traces: the data waiting to be compressed is kept in memory, and the process waits for it on exit. Stack (on top of #99765): 1. #99777: `no-memory` flag 2. #99778: Merge allocation counters into Enter/Exit rows 3. #99779: Delta-encode allocation counters, skip unchanged ones 4. #99780: Omit Exit+Enter pairs with (almost) no gap 5. #99808: **This PR**: zstd flags 6. #99809: Delta-encoded timestamps 7. #99781: Bump the trace header to `TRACEv1` 8. #99810: `upload-trace` accepts compressed traces ## Verification `next build` of `bench/heavy-npm-deps` with a release binding built from this branch: | flag | `NEXT_TURBOPACK_TRACING=1` trace | | --- | --- | | none | 23.8 MB | | `zstd-fast` | 7.2 MB | | `zstd` | 6.1 MB | | `zstd-best` | 5.1 MB | | `turbo-tasks,zstd` | 27.8 MB (raw: 155.7 MB) | - Every zstd capture passes `zstd -t`, is fully decoded by `turbo-trace-size` and loads completely in the trace server. - `1,gz,zstd` writes zstd, and `1,zstd,gz` writes gzip. <!-- NEXT_JS_LLM --> <!-- fleet 4a52842b-f1a2-43cb-9779-3655caada6a0 --> Co-authored-by:
vercel-fleet-prod[bot] <318278635+vercel-fleet-prod[bot]@users.noreply.github.com> Co-authored-by:
Tobias Koppers <1365881+sokra@users.noreply.github.com>
-
Tobias Koppers authored
## Summary When a span is exited and then entered again on the same thread almost immediately, the writer drops both rows. Async spans do this on every poll. It cuts the number of Enter/Exit rows almost in half. A pair is only omitted when all of these hold: - same span, same thread, and the `Exit` is still the last row in the thread-local buffer (not sent to the writer thread yet) - nothing was (de)allocated between the Exit and the Enter (otherwise the Enter would need to report it for whatever ran in between) - the gap is **≤ 1µs**, or **≤ 0.1%** of the time between the last `Enter` row written on this thread (of any span; omitted `Enter` rows don't count) and the Exit. For nested spans the last `Enter` can be a nested span's, which only shrinks the allowance. Memory samples are still taken when the `Enter` row is omitted. Allocations reported with the dropped Exit aren't lost. The baseline is rolled back, so the next row of the thread reports them, while the span is on top of the stack again. The viewer then shows the span as entered during the gap. Its self time can be slightly too large, by at most 1µs or 0.1% per gap. This is documented in `contributing/turbopack/tracing.md`. Implementation: - `TraceWriter`'s `WriteGuard` can mark exactly what a callback writes (`mark(marker, |guard| ...)`) and remove that row again with `remove_last_row(marker)`, as long as it still ends the thread local buffer. Any other write, or sending the buffer, invalidates it. `Exit` rows are marked with the span id. - `RawTraceLayer` keeps per-thread state (last Enter timestamp, last exit, allocation baseline). It is reset when the thread id changes. The allowed gap is only computed in `on_enter`, after the cheaper checks pass. Allocations made by this bookkeeping are excluded from the counters. - The clock and the allocation counters can be injected in tests. Stack (on top of #99765): 1. #99777: `no-memory` flag 2. #99778: Merge allocation counters into Enter/Exit rows 3. #99779: Delta-encode allocation counters, skip unchanged ones 4. #99780: **This PR**: omit Exit+Enter pairs with (almost) no gap 5. #99781: Bump the trace header to `TRACEv1` ## Measurements `next build` of `bench/heavy-npm-deps`, release native binding built from source: | | trace size | Enter rows | | --- | --- | --- | | PR 3 | 31.6 MiB | 1,138,850 | | same-timestamp pairs only | 27.2 MiB | 875,258 | | this PR (≤ 1µs or ≤ 0.1%) | 22.9 MiB | 626,956 | A dev build of the binding has almost no same-µs pairs (it's slower), so measure this with a release build. ## Verification - Unit tests for each condition: same/different span, thread and timestamp; allocations in between; an intervening row; a sent buffer; thread slot reuse; the 1µs fast path; exactly at and just over the 0.1% bound; chained omissions; allocation totals preserved - `TraceWriter` tests for marking and removing rows - Captured a trace with a locally built release binding and loaded it in the trace server <!-- NEXT_JS_LLM --> <!-- fleet 4a52842b-f1a2-43cb-9779-3655caada6a0 --> --------- Co-authored-by:
vercel-fleet-prod[bot] <318278635+vercel-fleet-prod[bot]@users.noreply.github.com> Co-authored-by:
Tobias Koppers <1365881+sokra@users.noreply.github.com>
-
Tobias Koppers authored
## Summary The allocation counters on `Enter`/`Exit` rows are now the **delta** since the previous report of the same thread, not cumulative per-thread totals. When nothing was (de)allocated, the delta is omitted (`None`). Cumulative counters are large numbers, so their varints take many bytes. Deltas are mostly small or zero: in a typical trace, half of all reports had an unchanged counter. - The writer keeps the last reported counters per thread. The first report of a thread only sets the baseline (`None`). That state also stores the thread id, because `thread_local` slots are reused after a thread exits, so a new thread never diffs against an old thread's counters. - The trace server attributes the delta directly, with no per-thread cumulative state. Stack (on top of #99765): 1. #99777: `no-memory` flag 2. #99778: Merge allocation counters into Enter/Exit rows 3. #99779: **This PR**: delta-encode allocation counters, skip unchanged ones 4. #99780: Omit Exit+Enter pairs with (almost) no gap 5. #99781: Bump the trace header to `TRACEv1` ## Measurements `next build` of `bench/heavy-npm-deps`, release native binding built from source: | | trace size | | --- | --- | | original | 68.2 MiB | | PR 2 | 58.8 MiB | | this PR | 31.6 MiB | ## Verification - Unit tests in `raw_trace.rs` for deltas, omitted zero deltas, the per-thread baseline and a reused thread slot - Captured a trace with a locally built binding and loaded it in the trace server <!-- NEXT_JS_LLM --> <!-- fleet 4a52842b-f1a2-43cb-9779-3655caada6a0 --> Co-authored-by:
vercel-fleet-prod[bot] <318278635+vercel-fleet-prod[bot]@users.noreply.github.com> Co-authored-by:
Tobias Koppers <1365881+sokra@users.noreply.github.com>
-
Tobias Koppers authored
## Summary Removes the separate `AllocationCounters` (and the unused `Allocation`) trace rows. Their data is attached to the `Enter`/`Exit` rows as `allocations: Option<Allocations>` instead. Before, the writer emitted one `AllocationCounters` row before every Enter and Exit. Each carried its own tag, timestamp and thread id. Those rows were **66%** of a typical trace. Since the counters are always paired with an Enter/Exit, the duplicated tag, timestamp and thread id can go. - `TraceRow::Enter`/`Exit` get an `allocations` field (`None` when memory isn't tracked). The new `Allocations` struct is shared by writer and readers. - Trace server: the counters on Enter/Exit go through the same diff-and-attribute logic as before. The span on top of the thread's stack before the row is applied gets the allocations: for Enter, the previously running span; for Exit, the exited span. - `turbo-trace-size`: the `AllocationCounters`/`Allocation` row kinds are removed. Enter/Exit now have an `allocations` component. This is an incompatible format change. The `TRACEv1` header bump comes last in the stack (PR 5), and older traces are then rejected with an error. The stack is meant to land together. Stack (on top of #99765): 1. #99777: `no-memory` flag 2. #99778: **This PR**: merge allocation counters into Enter/Exit rows 3. #99779: Delta-encode allocation counters, skip unchanged ones 4. #99780: Omit Exit+Enter pairs with (almost) no gap 5. #99781: Bump the trace header to `TRACEv1` ## Measurements `next build` of `bench/heavy-npm-deps`, release native binding built from source: | | trace size | rows | | --- | --- | --- | | before (PR 1) | 68.2 MiB | 4.65 M | | this PR | 58.8 MiB | 2.37 M | ## Verification - New trace server test checks allocation attribution for Enter/Exit rows through `TurbopackFormat` - Captured a trace with a locally built binding and loaded it in the trace server <!-- NEXT_JS_LLM --> <!-- fleet 4a52842b-f1a2-43cb-9779-3655caada6a0 --> Co-authored-by:
vercel-fleet-prod[bot] <318278635+vercel-fleet-prod[bot]@users.noreply.github.com> Co-authored-by:
Tobias Koppers <1365881+sokra@users.noreply.github.com>
-
Aurora Scharff authored
### What Fix the build route table so a collapsed generated-path row uses the shared symbol of the concrete paths it represents. The Maintainer Agent regression test now expects the Static symbol. ### Before The generated paths are static, but the collapsed row inherits the Partial Prerender symbol from the dynamic route pattern: ```text └ /blog/[...slug] ├ ◐ /blog/[...slug] ├ ○ /blog/post-1 ├ ○ /blog/post-2 └ ◐ [+10 more paths] ``` ### After The collapsed row uses the shared Static symbol from the generated paths it represents: ```text └ /blog/[...slug] ├ ◐ /blog/[...slug] ├ ○ /blog/post-1 ├ ○ /blog/post-2 └ ○ [+10 more paths] ``` The route pattern remains Partial Prerender because it can still handle paths outside `generateStaticParams`. Only the collapsed generated-path row changes. ### Why With Cache Components, Partial Prefetching, and `ensureStatic = 'navigation'`, a catch-all route can have a Partial Prerender route pattern while every generated path is complete and static. The synthetic `[+N more paths]` label has no `pageInfos` entry, so it previously fell back to the route pattern and printed Partial Prerender for those static paths. ### How Compute the hidden paths’ symbols in both the regular and duration-sorted truncation branches. Use the shared child symbol when every hidden path agrees; if the hidden paths have mixed symbols, preserve the conservative route-pattern fallback. The regression test asserts the real `next build` CLI output and verifies that every generated path produces complete static HTML. ### Documentation The Building guide introduces the collapsed `[+N more paths]` notation at first use and shows the correct symbol as the rendering classification changes. ### Validation - `NEXT_SKIP_ISOLATE=1 pnpm test-start-turbo test/e2e/app-dir/build-output-truncated-generated-paths/` — 2 passed - `NEXT_SKIP_ISOLATE=1 pnpm test-start-turbo test/production/build-output/build-output.test.ts` — 10 passed, 2 skipped - `pnpm --filter=next build` — passed - Pre-commit Prettier and ESLint checks passed - Prettier check for the documentation — passed Maintainer investigation: https://next-maintainer-agent.vercel.tools/investigations/feedback/64 DX Agent feedback: https://dxagent.labs.vercel.dev/dashboard/feedback/3fbc3904-2843-4eb6-9033-312f4306b66b Prepared in fork: https://github.com/vercel-labs/next.js-experimental/pull/237 --------- Co-authored-by:
vercel-gh-bot-3[bot] <282332853+vercel-gh-bot-3[bot]@users.noreply.github.com>
-
next-js-bot[bot] authored
Co-authored-by:next-js-bot[bot] <279046576+next-js-bot[bot]@users.noreply.github.com>
-
Sebastian "Sebbie" Silbermann authored
Release branches no longer carry the version number. Two long-lived, manually-created branches replace them: `releases/lts/active` (current major's release line) and `releases/lts/maintenance` (previous major's). From then on, triggering a stable release moves the refs automatically, via the release GitHub App (next.js bot) credentials already used for the signed release commit: - **patch**: no ref moves — the release commit already advances the branch it was cut on - **minor**: routed by the released version's major against the latest major (the one `releases/lts/active` currently points at) — equal moves `releases/lts/active` to the new tag, latest − 1 moves `releases/lts/maintenance` to the new tag, anything else moves nothing - **major**: `releases/lts/maintenance` takes `releases/lts/active`'s previous position, then `releases/lts/active` moves to the new tag Supporting changes: - `scripts/release-branches.json` lists the fixed branch set (`canary` + both LTS branches) so `version-bump.js` allows bumps on them - `build_and_test.yml` / `build_and_deploy.yml` run on pushes to the LTS branches, and the staging deploy-target condition covers them (parity with the per-branch rewrite the removed script used to produce) - `backport-pr` skill, `AGENTS.md`, `release-channels-publishing.md`, and a stale `version-bump.js` comment updated for the new branch model Co-authored-by:Claude Code <noreply@anthropic.com>
-
Niklas Mischkulnig authored
We already set `package.json#packageManager` <img width="539" height="388" alt="Bildschirmfoto 2026-07-01 um 16 26 30" src="https://github.com/user-attachments/assets/8f6f3e8f-60a6-484d-ac57-f8dc261f8b85" /> But the Vercel deployment tests still used npm to install dependencies because you need to set `ENABLE_EXPERIMENTAL_COREPACK` so that `packageManager` applies. pnpm should be hopefully faster, and also improves consistency across the test environments Full deploy test run: ~~https://github.com/vercel/next.js/actions/runs/28528683809~~ ~~https://github.com/vercel/next.js/actions/runs/28531873334~~ ~~https://github.com/vercel/next.js/actions/runs/28542029071~~ ~~https://github.com/vercel/next.js/actions/runs/29311003094~~ ~~https://github.com/vercel/next.js/actions/runs/32731722428~~ ~~https://github.com/vercel/next.js/actions/runs/37841713568~~ https://github.com/vercel/next.js/actions/runs/37900216932
-
Niklas Mischkulnig authored
Imported https://github.com/vercel/next.js/pull/99872 by @pieh as a local branch to be able to run CI Closes https://github.com/vercel/next.js/pull/99872 --- <!-- Thanks for opening a PR! Your contribution is much appreciated. To make sure your PR is handled as smoothly as possible we request that you follow the checklist sections below. Choose the right checklist for the change(s) that you're making: ## For Contributors We encourage you to use AI to assist you in researching, creating, and reviewing changes. However, you must review and deeply understand the contributions you are making. For this reason, **pull request descriptions from external contributors must be written by a human**. ### Improving Documentation - Run `pnpm prettier-fix` to fix formatting issues before opening the PR. - Read the Docs Contribution Guide to ensure your contribution follows the docs guidelines: https://nextjs.org/docs/community/contribution-guide ### Fixing a bug - Related issues linked using `fixes #number` - Tests added. See: https://github.com/vercel/next.js/blob/canary/contributing/core/testing.md#writing-tests-for-nextjs - Errors have a helpful link attached, see https://github.com/vercel/next.js/blob/canary/contributing.md ### Adding a feature - Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR. (A discussion must be opened, see https://github.com/vercel/next.js/discussions/new?category=ideas) - Related issues/discussions are linked using `fixes #number` - e2e tests added (https://github.com/vercel/next.js/blob/canary/contributing/core/testing.md#writing-tests-for-nextjs) - Documentation added - Telemetry added. In case of a feature if it's used or not. - Errors have a helpful link attached, see https://github.com/vercel/next.js/blob/canary/contributing.md ### Signed commits - This repository requires verified commit signatures on protected branches. - If this pull request is blocked for unsigned commits, re-sign the commits and force-push the branch. - A `Signed-off-by` line in the commit message is not enough. ## For Maintainers - Minimal description (aim for explaining to someone not on the team to understand the PR) - When linking to a Slack thread, you might want to share details of the conclusion - Link both the Linear (Fixes NEXT-xxx) and the GitHub issues - Add review comments if necessary to explain to the reviewer the logic behind a change ### What? ### Why? ### How? Closes NEXT- Fixes # --> ### What? Adjust existing `cache-control` assertions in e2e tests running in deploy mode to compare directives in a way that whitespace changes don't fail a test, but still assert that exact set of expected directives is received. To minimize the change I avoided changing tests that do not run in deploy mode. ### Why? Tests asserting on exact string are failing like this when testing Netlify deployments: ``` Error: expect(received).toBe(expected) // Object.is equality Expected: "public, max-age=0, must-revalidate" Received: "public,max-age=0,must-revalidate" at Object.toBe (/home/runner/work/opennextjs-netlify/opennextjs-netlify/nextjs/test/e2e/app-dir/metadata-dynamic-routes/index.test.ts:26:48) at processTicksAndRejections (node:internal/process/task_queues:95:5) ``` which still has same overall behavior, just not the same exact formatting ### How? Extract pre-existing local `expectDirectives` helper from one of the the [tests](https://github.com/vercel/next.js/blob/ea119d2702ba9252233c678ca74b4a310b5eedd0/test/e2e/app-dir/ppr-full/ppr-full.test.ts#L70-L83) to shared utils, generalize it to allow regex rules and adjust deploy tests to use it instead of strict string comparisons. Co-authored-by:
pieh <misiek.piechowiak@gmail.com>
-
- 08 Oct, 2026 13 commits
-
-
next-js-bot[bot] authored
-
Benjamin Woodruff authored
Example failure: https://github.com/vercel/next.js/actions/runs/37806862569/job/113451942869#step:35:458 Caused by: https://github.com/vercel/vercel-internal/pull/1050
-
Benjamin Woodruff authored
Most of the time the `next.start()` call happens as part of `nextTestSetup`, but here it happens inside of the `it` block, so extend the timeout. It looks like this started timing out because of https://github.com/vercel/next.js/pull/98776 ? The webpack deployment tests frequently fail out: | Deployment | Remote build start → deployment complete | |---|---:| | [canary.4, first execution](https://vercel.com/vtest314-next-e2e-tests/vtest314-e2e-tests/9BSVkxJfRGBxgLjXWDK1sGVyfqwJ) | **61\.0s** | | [canary.4, second execution](https://vercel.com/vtest314-next-e2e-tests/vtest314-e2e-tests/2hwJUWduf2F1qMtsdC3CnwJGSQoA) | **53\.8s** | | [canary.62](https://vercel.com/vtest314-next-e2e-tests/vtest314-e2e-tests/FefU9FREV4BCmEo1dsu2to5Wx1uf) | **55\.7s** |
-
Benjamin Woodruff authored
It sounds like the script tag for GTM 404ed in the past if given a bad tag id, and now it 200s. We need to only check for the events we care about instead of asserting exact sets of events.
-
next-js-bot[bot] authored
-
Aurora Scharff authored
## Summary Remove parentheses from the `navigation` and `prefetch` sidebar labels so they match the other function API references. ## Verification - `pnpm prettier --with-node-modules --ignore-path .prettierignore --check docs/01-app/03-api-reference/04-functions/navigation.mdx docs/01-app/03-api-reference/04-functions/prefetch.mdx` - Confirmed no function API reference has a `nav_title` ending in `()` <!-- NEXT_JS_LLM -->
-
Jiwon Choi authored
### Why? Telemetry fails on Node.js 24 with this error before any HTTP request is sent: ```text TypeError: RequestInit: Expected signal ("AbortSignal {}") to be an instance of AbortSignal. ``` Native `fetch` checks that `RequestInit.signal` is an instance of Node's native `AbortSignal`. The Edge Runtime ponyfill creates its own `AbortSignal` class: the name is the same, but the constructor is different, so the signal fails the instance check. The telemetry transport retries once and suppresses the final error, so the CLI can exit successfully without delivering events. The controller was introduced when Next.js supported Node.js 14.6. Next.js now requires Node.js 20.9 or later, where native `AbortController` is available on all supported versions. ### How? Use Node's native `AbortController` so its signal matches `fetch`, preserving the existing timeout. Verified real SDK HTTP delivery on Node.js 20.9, 22.6, and 24.1, and the five-second cancellation deadline. The patched CLI received HTTP 204 for all three upgrade lifecycle events; Snowflake persistence is not yet verified. The full repository build and test suite were not run. <!-- NEXT_JS_LLM --> -
Sebastian "Sebbie" Silbermann authored
Deploy tests now authenticate the Vercel CLI via the `VERCEL_TOKEN` env var instead of writing a shared `$HOME/.vercel/auth.json` and passing `--global-config` to every CLI invocation. This reduces the chance of a read/write race on the shared auth.json. I've never seen this with our current concurrency limits (2) but we saw it immediately when bumping to 8 (https://github.com/vercel/next.js/actions/runs/37704518620/job/113083223389#step:35:386) --------- Co-authored-by:
Claude Code <noreply@anthropic.com>
-
Jan Kassens authored
## Summary `@types/react` 19.2.10 deprecates `FormEvent`/`FormEventHandler` (they don't correspond to a real DOM event). This switches `onSubmit` handlers to `SubmitEvent<HTMLFormElement>` in test fixtures, the docs (`authentication.mdx`, `preserving-ui-state.mdx`), and `examples/with-supabase`. It also removes an unused `FormEvent` import from a JSX docs sample. Examples still pinned to `@types/react` 18 are unchanged, since `SubmitEvent` only exists in `@types/react` >= 19.2.10. ## Verification - Type-checked the updated fixtures and the docs snippet patterns with `tsc` (strict) against `@types/react` 19.2.18, plus a negative check confirming the event type is enforced against `onSubmit`. The fixture files aren't in the root `tsconfig.json` include list, so `pnpm typescript` doesn't cover them. <!-- NEXT_JS_LLM -->
-
Tobias Koppers authored
### What? Stop the Turbopack disk watcher from printing `watch error … NotFound` for paths that vanish while notify is scanning them, and send the remaining watcher errors to stderr instead of stdout. ### Why? In polling mode, notify's `PollWatcher` walks the tree with walkdir. When a directory is deleted between being listed and being opened or stat'ed, notify reports an `Io` error with `NotFound`. Deleting a large tree (`rm -rf`, a package install) produces a burst of these. A filesystem fuzzer saw about 4000 in two minutes. Each one was printed with `println!`, so a library was spamming the user's stdout. These errors don't signal a correctness problem. The watcher already invalidates the reported path recursively, and the next scan reports the removal anyway. The fuzzer runs had no missed invalidations. ### How? - A small private classifier recognizes notify errors of kind `Io` wrapping `std::io::ErrorKind::NotFound`. It matches on the structured error kinds, not on message text. - The watcher thread skips the print for those errors. Batch invalidation and batch-delay scheduling still run for every error, so invalidation behavior doesn't change. - Other notify error kinds, including `PathNotFound`, are still reported, now on stderr. ### Verification - `cargo test -p turbo-tasks-fs --lib watcher::tests`: 19 passed, covering native and polling watchers in both recursive and non-recursive modes - `cargo clippy -p turbo-tasks-fs --tests -- -D warnings` - Not run: a deletion stress test reproducing the race (the fuzzer finding above). <!-- NEXT_JS_LLM --> <!-- fleet 89bd34bb-fe6a-4ea4-abcc-abb97db76087 --> --------- Co-authored-by:
vercel-fleet-prod[bot] <318278635+vercel-fleet-prod[bot]@users.noreply.github.com> Co-authored-by:
Tobias Koppers <1365881+sokra@users.noreply.github.com>
-
Benjamin Woodruff authored
Reverts vercel/next.js#99770 We're seeing a lot of deployment test failures due to rate limits. We can try bumping the rate limits and trying again, but reverting for right now to unblock deployment tests.
-
Luke Sandberg authored
## Summary Bumps `next` in three upgrade eval fixtures that were pinned to versions affected by CVE-2026-94483 (Image Optimization SSRF via allow-listed remote patterns; fixed in 16.3.8): - `future-cache-components-nudge`: 16.3.5 → 16.4.0 - `future-cache-components-same-version`: 16.3.5 → 16.4.0 - `latest-same-major`: 16.2.12 → 16.3.8 A few things needed to be updated to match - `latest/setup.ts` and `future/setup.ts` mock `registry.npmjs.org/next/latest` with a hardcoded `target`. That target is now 16.4.0 (the current `latest`), so the same-version scenarios still sit on "latest" and the same-major scenario still has a newer target in the same major. - `latest-same-major` uses 16.3.8 rather than 16.4.0 because it has to start below the target in the same major. 16.3.8 is the oldest patched release; there is no patched 16.2.x. - The `EVAL.ts` exact-version assertions are updated to match. The `latest-cross-major`, `latest-nudge` and `future-cache-components` scenarios now also upgrade to 16.4.0. Their starting versions are unchanged. Fixes VULN-16485 Fixes VULN-16486 Fixes VULN-16487 <!-- NEXT_JS_LLM -->
-
Luke Sandberg authored
## What Builds with the Turbopack filesystem cache didn't give useful output when the cache couldn't be written during shutdown, for example because of a read-only directory or a full disk. The build succeeded, but it printed a bare `Shutting down failed: <message>` which was not useful Before: ``` Persisting failed during shutdown: Unable to write SST file 00000020.sst Caused by: : Permission denied (os error 13)te/var/folders/p6/ytr7tcts5ndf72xw136ky30m0000gn/T/next-install-6cb7b5bc6e3b95a18ae07aa856827db93efc662fe73934ecebae8de778735e0c/.next/cache/turbopack/v16.5.0-canary.2-2c48f075ef/00000020.sst` Shutting down failed: A previous write operation failed with an unrecoverable error; no further writes are possible ``` Now: ``` WARNING: Saving the filesystem cache failed: Unable to write SST file 00000020.sst: failed to create file `/Users/lukesandberg/projects/next_2/next.js/test/tmp/next-test-1791392696220-115/.next/cache/turbopack/unversioned/00000020.sst`: Permission denied (os error 13) ``` The "Shutting down failed" error is gone and the filesystem cache error is slightly improved. Also: - **Compaction is skipped once a write has already failed.** That avoids a second, redundant failure. - **Database shutdown can't fail anymore.** Compaction errors during shutdown are logged as warnings. --------- Co-authored-by:vercel-gh-bot-5[bot] <312521305+vercel-gh-bot-5[bot]@users.noreply.github.com>
-
- 07 Oct, 2026 9 commits
-
-
next-js-bot[bot] authored
-
Will Binns-Smith authored
## Summary Mark `next analyze export` and `next-bundle-optimizer` as added in v16.4. Add version history entries to the AI agents guide, package bundling guide, and CLI reference. ## Verification - Prettier and ESLint passed. - Runtime tests not run (docs only). <!-- NEXT_JS_LLM --> --------- Co-authored-by:Jan Kassens <kassens@vercel.com>
-
Aurora Scharff authored
### What When `experimental.agentFeedback` is enabled, create-next-app now writes the `nextjs-agent-feedback` block into `AGENTS.md` along with the `nextjs-agent-rules` block. ### Why Before this change, create-next-app enabled the flag in `next.config.ts` but only wrote the rules block. The feedback block was added later by `next dev`, and only when `@vercel/detect-agent` found a coding agent (`syncAgentFeedbackForDev` in `app-info-log.ts`). As a result: - An agent that read `AGENTS.md` before running `next dev` never saw the feedback instructions. - The block then appeared as an uncommitted `AGENTS.md` change on top of the "Initial commit from Create Next App". ### How - `generateAgentFiles(root, { agentFeedback })` adds the feedback block after the rules block. The text is copied from `packages/next/src/server/lib/generate-agent-files.ts`, with a comment saying to keep them in sync. `next dev` compares the block byte-for-byte. - `create-app.ts` passes `agentFeedback: !example && agentFeedback`, because examples don't get `experimental.agentFeedback` written to their config. - `--no-agents-md` still writes no `AGENTS.md`. ### Verification I ran the new helper next to `next dev`'s `writeAgentFiles` + `writeAgentFeedbackFiles`, with the flag on and off: - The output is byte-identical in both cases. - `hasCurrentAgentRules` / `hasCurrentAgentFeedback` return true on create-next-app's output. - Running `next dev`'s writers again on that output returns `unchanged`. Tests in `test/production/create-next-app/index.test.ts`: - `--agents-md --agent-feedback` writes the rules block, then the feedback block, and `next dev`'s `writeAgentFiles` / `writeAgentFeedbackFiles` both report `unchanged`. This catches drift between the two copies of the text. - `--agents-md --no-agent-feedback` writes only the rules block. - `--no-agents-md --agent-feedback` creates no `AGENTS.md`.🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by:Claude Opus 5.5 <noreply@anthropic.com>
-
Jan Kassens authored
These were breaking some GitHub builds for me for some reason. ## Test Plan GitHub CI + local cargo check validation
-
Aurora Scharff authored
## Summary Make each skill named by `next-bundle-optimizer` directly installable when it is unavailable, matching the inline handoff pattern used by the other Next.js skills. ## Verification - `pnpm prettier --with-node-modules --ignore-path .prettierignore --write skills/next-bundle-optimizer/SKILL.md` - Not run: runtime tests (skill text only) <!-- NEXT_JS_LLM -->
-
Sebastian "Sebbie" Silbermann authored
Increases test concurrency for the deploy test jobs to 8 (prev: 2 in the full suite, 1 in new build-and-test jobs). For deploy tests, the CPU-intensive work that should be as stable as possible (the Next.js build and server) does not happen on the runner. It happens remotely when the deployment is built at Vercel. The runner mostly handles I/O and setup: pushing the deployment, waiting for it to become ready, polling logs, and making requests against it. So the usual reason to cap concurrency (runner CPU contention) barely applies. Adjusted rate-limits: - 10x `api-projects-get` (prev failure: https://github.com/vercel/next.js/actions/runs/37617890690/job/112780953612#step:35:1159) Trade-offs to watch: - Peak concurrent Vercel deployments per run rises 4× (shards × 2 → shards × 8; 64 → 256 concurrent test files at today's shard counts). If we hit Vercel account deployment-concurrency limits, increase the limits - maybe we still need a bit beefier runners if the concurrent setup/tests starve runner CPU ## test plan - [x] [deploy tests with debug logging on this branch takes 30 minutes wall time without hitting Vercel rate limits](https://github.com/vercel/next.js/actions/runs/37627918534); usual flake on the 1st try (down from [97 minutes on canary](https://github.com/vercel/next.js/actions/runs/37588790388/attempts/1)) --------- Co-authored-by:
Claude Code <noreply@anthropic.com>
-
next-js-bot[bot] authored
-
Tobias Koppers authored
### What? Fixes the `rust check` (clippy) and `test cargo unit` jobs on canary, which fail with: ``` error[E0063]: missing field `active_worker_threads` in initializer of `TraceRow<'_>` --> turbopack/crates/turbopack-trace-size/src/lib.rs:936:13 ``` The fix adds the missing field to the `TraceRow::MemorySample` row in the `turbo-trace-size` unit test. ### Why? This was a semantic merge conflict between two PRs that each passed CI on their own: - #99233 added `active_worker_threads` to `TraceRow::MemorySample`. - #99765 added the `turbo-trace-size` crate, whose test builds a `MemorySample` row without that field. Only test code is affected. Every other `MemorySample` constructor already sets the field. ### How? The test row now sets `active_worker_threads: 3`. The size assertions are computed from the serialized rows, so they don't change. Verified locally on top of `db03f83f`: before the change, `cargo test -p turbopack-trace-size` fails with the error above. After it, the following pass: - `cargo test -p turbopack-trace-size` - `cargo clippy --all-targets -- -D warnings` for `turbopack-trace-size`, `turbopack-trace-utils` and `turbopack-trace-server` - tests of `turbopack-trace-utils` and `turbopack-trace-server` <!-- NEXT_JS_LLM --> <!-- fleet 4a52842b-f1a2-43cb-9779-3655caada6a0 --> Co-authored-by:
vercel-fleet-prod[bot] <318278635+vercel-fleet-prod[bot]@users.noreply.github.com> Co-authored-by:
Tobias Koppers <1365881+sokra@users.noreply.github.com>
-
Niklas Mischkulnig authored
**View diff without whitespace** Call these just once (per module), not once per referenced import - `unused_references()` - `reference.get_referenced_asset()` - `imported_module.get_ident(export)`
-