1. 11 Dec, 2025 6 commits
    • vercel-release-bot's avatar
      v15.0.7 · fc9d731d
      vercel-release-bot authored
      fc9d731d
    • Hendrik Liebau's avatar
      Backport facebook/react#35351 for 15.0.6 (#87092) · e2110e15
      Hendrik Liebau authored
      facebook/react#35351
      Co-authored-by: 's avatarJosh Story <story@hey.com>
      e2110e15
    • vercel-release-bot's avatar
      v15.0.6 · 24793ec1
      vercel-release-bot authored
      24793ec1
    • Hendrik Liebau's avatar
      Update React Version (#46) · 53abcfa2
      Hendrik Liebau authored
      Co-authored-by: 's avatarJosh Story <story@hey.com>
      53abcfa2
    • Josh Story's avatar
      Update React Version (#35) · 9068d90e
      Josh Story authored
      9068d90e
    • Zack Tanner's avatar
      Backport Next.js changes to v15.0.6 (#28) · 84f7ed4c
      Zack Tanner authored
      * remove urlencoded server action handling (#14)
      
      * remove urlencoded server action handling
      
      * urlencoded POSTs should still flow through action handler
      
      * bailout early for urlencoded
      
      * tweak handling to 404 in feetch action case
      
      * error when there are no server actions (#15)
      
      * remove custom error
      
      * remove dev gate
      
      * error when there are no server actions
      
      * gate against dev
      
      * Check whether action ids are valid before parsing (#16)
      
      * Check whether action ids are valid before parsing
      
      In an effort to narrow access to parsing methods Next.js now prequalifies MPA form parsing by ensuring there are only valid action IDs as part of the submission. This is not meant to be a strong line of defense against malicious payloads since action IDs are not private and are for many apps relatively easy to acquire. It does however provide some limited narrowing of code paths so that action decoding only happens for plausible actions
      
      Additionally all other branches that use next-action header now consistently check the header before proceeding with parsing. This is also a perf improvement.
      
      * fix test assertion
      
      ---------
      Co-authored-by: 's avatarZack Tanner <1939140+ztanner@users.noreply.github.com>
      
      * fix test
      
      ---------
      Co-authored-by: 's avatarJosh Story <story@hey.com>
      84f7ed4c
  2. 10 Dec, 2025 1 commit
  3. 03 Dec, 2025 4 commits
  4. 02 Dec, 2025 1 commit
  5. 05 Dec, 2024 3 commits
  6. 07 Nov, 2024 19 commits
  7. 06 Nov, 2024 6 commits
    • Jam Balaya's avatar
      chore: add `syntax` parser directive to Dockerfile (#72380) · d8c18758
      Jam Balaya authored
      ## What?
      Add [`syntax`](https://docs.docker.com/reference/dockerfile/#syntax)
      parser directive to the first line of the
      [Dockerfile](https://docs.docker.com/reference/dockerfile/).
      
      ## Why?
      To declare the Dockerfile syntax version to use for the build.
      If unspecified, [BuildKit](https://docs.docker.com/build/buildkit/) uses
      a bundled version of the Dockerfile frontend.
      
      ## How?
      Append `# syntax=docker.io/docker/dockerfile:1` to the top of the
      [Dockerfile](https://docs.docker.com/reference/dockerfile/).
      Co-authored-by: 's avatarJJ Kasper <jj@jjsweb.site>
      d8c18758
    • Sam Ko's avatar
      docs(output): update standalone docs with cp and start commands (#72432) · 3b5b5a25
      Sam Ko authored
      ## Why?
      
      We're not documenting a way to manually move files to `.next/standalone` for `output: 'standalone'`.
      3b5b5a25
    • Andrew Clark's avatar
      [Prefetching] Fix: Read segment data from correct source (#72427) · 3d29c270
      Andrew Clark authored
      In the last PR (#72367) I got confused and incorrectly assumed that
      initialRSCPayload was a client Flight value, not a server value. Which
      meant it was actually re-rendering the Server Components again, rather
      than reusing the result from the page render.
      
      Instead, I need to decode this value from the full page Flight stream,
      like I did originally. (Except we still don't need to do it once per
      segment, just once per page.)
      
      To avoid another unecessary decoding of the page stream, I've moved the
      segment rendering tasks so that they are spawned from inside the render
      that generates the tree metadata response.
      
      The size of the diff is bit misleading; mostly this involves just
      rearranging the code so that it executes in a slightly different order.
      3d29c270
    • Janka Uryga's avatar
      fix(font): don't break directives with inserted next/font import (#72350) · 7a6b20c1
      Janka Uryga authored
      the `next_font_loaders` transform replaces imports from `next/font` with
      slightly different ones. unfortunately it was inserting these imports
      **at the top** of the file, which broke the `server_actions` transform
      because it expects "use server"/"use cache" directives to be the first
      item in the file.
      
      this PR changes the logic in the font transform to insert the import in
      the place where the original `next/font` import occurred instead of at
      the top of the file.
      
      (i suppose that we could just reorder the transforms and run
      `server_actions` before `next_font_loaders`, but this feels more robust)
      
      Fixes #72179
      7a6b20c1
    • Jiachi Liu's avatar
      f6e9bd9c
    • Andrew Clark's avatar
      Prevent client prefetch stream from closing (#72420) · c3c5bf89
      Andrew Clark authored
      Based on:
      - #72418 
      
      ---
      
      When PPR is enabled, prefetch streams may contain references that never
      resolve, because that's how we encode dynamic data access. In the
      decoded object returned by the Flight client, these are reified into
      hanging promises that suspend during render, which is effectively what
      we want. The UI resolves when it switches to the dynamic data stream
      (via useDeferredValue(dynamic, static)).
      
      However, the Flight implementation currently errors if the server closes
      the response before all the references are resolved. As a cheat to work
      around this, we wrap the original stream in a new stream that never
      closes, and therefore doesn't error.
      c3c5bf89