- 07 Jan, 2025 8 commits
-
-
vercel-release-bot authored
-
Janka Uryga authored
Fixes #74062 (`jotai` ran into this error [when they added `"type": "commonjs"` to their package.json](https://github.com/pmndrs/jotai/discussions/2579#discussioncomment-9812492)) > this is a bug in the transform we do when a `'use client'` directive is encountered. I think what's happening is that we're creating a virtual file that uses ESM import/export syntax, but it's called proxy.js (not proxy.mjs), so the `"type": "commonjs" `makes turbopack "correctly" upset at the ESM syntax. https://github.com/vercel/next.js/issues/74062#issuecomment-2555922867 The (slightly kludgy) solution is to use `proxy.mjs` or `proxy.cjs` to force the module format, bypassing the issue where `proxy.js` changes meaning depending on `package.json#type`.
-
JJ Kasper authored
This backports #74577 which ensures we don't lose tags added to a page before notFound() is called.
-
JJ Kasper authored
This backports https://github.com/vercel/next.js/pull/74164 to leverage built-in waitUntil if available instead of using the approach that keeps the stream open until the waitUntil promise resolves. x-ref: [slack thread](https://vercel.slack.com/archives/C02K2HCH5V4/p1736211642221149?thread_ts=1734707275.666089&cid=C02K2HCH5V4)
-
Steven authored
Backport https://github.com/vercel/next.js/pull/74257 to 15.1.x
-
Steven authored
Backport https://github.com/vercel/next.js/pull/73925 to 15.1.x
-
Steven authored
Backport https://github.com/vercel/next.js/pull/73923 to 15.1.x
-
Steven authored
Backport https://github.com/vercel/next.js/pull/73909 to 15.1.x
-
- 26 Dec, 2024 1 commit
-
-
vercel-release-bot authored
-
- 24 Dec, 2024 2 commits
-
-
Zack Tanner authored
Backports: - #74154
-
Zack Tanner authored
Backports: - #73900 Co-authored-by:Hendrik Liebau <mail@hendrik-liebau.de>
-
- 19 Dec, 2024 2 commits
-
-
vercel-release-bot authored
-
Zack Tanner authored
[diff facebook/react@7283a213...65e06cb7](https://github.com/facebook/react/compare/7283a213...65e06cb7) <details> <summary>React upstream changes</summary> - [Merge 4af40315c864c4467aa1a2fb61b9c05b7d180e3e into 7283a213dbbc31029e65005276f12202558558fc facebook/react@65e06cb77](https://github.com/facebook/react/commit/65e06cb77c084a11e22142e5f663c0969bf7edfa) (Josh Story) - https://github.com/facebook/react/pull/31840 </details>
-
- 17 Dec, 2024 3 commits
-
-
vercel-release-bot authored
-
Zack Tanner authored
-
Zack Tanner authored
backport: fix(turbo): sassOptions silenceDeprecations was not overwritten with user options (#74005) Backports: - https://github.com/vercel/next.js/pull/73937Co-authored-by:
Jiwon Choi <devjiwonchoi@gmail.com> Co-authored-by:
Donny/강동윤 <kdy1997.dev@gmail.com>
-
- 16 Dec, 2024 3 commits
-
-
Zack Tanner authored
Backports: - https://github.com/vercel/next.js/pull/73908
-
eps1lon authored
- 10 Dec, 2024 21 commits
-
-
vercel-release-bot authored
-
vercel-release-bot authored
-
Jiachi Liu authored
Fix type error introduced in #73540
-
Benjamin Woodruff authored
docs(turbopack): Better document the Vc type, with references to ResolvedVc and VcOperation (#72524)    
-
Andrew Clark authored
During a navigation, if all the data has been prefetched, and the target route does not contain any dynamic data, then we should skip a request to the server. This uses the `isPartial` field I added in the previous PRs to track whether the prefetched data is complete or not.
-
vercel-release-bot authored
-
Zack Tanner authored
This was fixed upstream, so we can re-enable the deployment tests for it.
-
Andrew Clark authored
Currently, when the client prefetches a segment, the server responds with a 404 if it cannot fulfill the request. This updates it to respond with a 204 No Content instead, since it's not an error for the client to request a segment whose prefetch hasn't been generated. When responding with 204, the server also sends the 'x-nextjs-postponed' header, but only if PPR is enabled for the route. The client can use the absence of this header to distinguish between a regular cache miss and a miss due to PPR being disabled. In a later PR, I will update the client to use this information to fall back to the non-PPR behavior.
-
Steven authored
We regressed in https://github.com/vercel/next.js/pull/73483 since the LRU calculates the value based on the value, not the key (and we were not setting the value).
-
vercel-release-bot authored
-
Tobias Koppers authored
### What? This adds a workaround for a Node.js bug on MacOS which causes closing file watchers to be really slow: https://github.com/nodejs/node/issues/29949 It was fixed before here: https://github.com/vercel/next.js/pull/51826 But accidentically removed here: https://github.com/vercel/next.js/commit/c5a8e0989ea13af8d04bf6dda8d741c79b964abb#diff-0ff576bd02e76f96680accff48ecbe4126a7f6bc4eafa547b6d44dee49bab770L88
-
Zack Tanner authored
Returning an empty array from `generateStaticParams` should still be deployed as an ISR route to support the `dynamicParams = true` case, where a route can be generated & cached on demand. At the moment the only way to achieve this behavior is with `export const dynamic = 'force-static'`. As a result, these routes are being treated as dynamic and never return a cached response. This PR resolves the bug by fixing several distinct things: - `prerenderedRoutes` does not need to be an array > 0 to mark a route as being a static path. This will ensure the build output shows the correct symbol and adds the path to the list of prerenders. - we initialize `prerenderedRoutes` to an empty array if we determine that the leaf-most segment (the thing rendering the current page) has a `generateStaticParams` function - the function responsible for collecting segments associated with a page only considered the `children` branch, but for something like a parallel/interception route, those segments would be inside of a different parallel route key - the regex for determining if a function was dynamic didn't consider that the dynamic route might also contain an interception marker (eg `(.)`) This appears to have regressed in #68125 Closes NEXT-3905 --------- Co-authored-by:Wyatt Johnson <accounts+github@wyattjoh.ca>
-
Niklas Mischkulnig authored
Closes PACK-3537
-
Delba de Oliveira authored
**Do not merge until stable `after` lands in canary (#73605).** Closes: https://linear.app/vercel/issue/DOC-3818/[stable]-nextafter Redirect: https://github.com/vercel/front/pull/38704 --------- Co-authored-by:
Lee Robinson <me@leerob.io>
-
Tobias Koppers authored
### What? Fixes persistent caching for webpack in development Add a test case for persistent caching on Turbopack and webpack.
-
Delba de Oliveira authored
**Do not merge until 15.1 is released.** Currently these features are tagged as `canary`, but once 15.1 becomes stable, they'll be `experimental`. Closes: https://linear.app/vercel/issue/DOC-3928/update-forbidden-docs-to-add-experimental-tag
-
Delba de Oliveira authored
Going through the getting started pages with fresh eyes, picked up some nits.
-
Niklas Mischkulnig authored
Closes PACK-3535 - Adds some infrastructure for the single-graph-traversal stuff, ~~focusing on dev with layout segment optimization for now~~ - Migrate next/dynamic collection to it - This shouldn't make any observable change to the output
-
Tobias Koppers authored
### What? * avoid flagging "thread" span as incomplete * improve error message * add comments * fixup print_stats * fix lmdb feature
-
Tobias Koppers authored
### What? Fixes the lmdb feature and makes it compile fine
-
Delba de Oliveira authored
Closes: https://linear.app/vercel/issue/DOC-3929/allow-request-apis-in-after Adds example for using request APIs in `after`. Adds note about Server Component restriction. related: https://github.com/vercel/next.js/pull/73345
-