_Hello Next.js team! First PR here, I hope I've followed the right practices._
### What's in there?
It has been decided to only support the following uses cases in Next.js' middleware:
- rewrite the URL (`x-middleware-rewrite` response header)
- redirect to another URL (`Location` response header)
- pass on to the next piece in the request pipeline (`x-middleware-next` response header)
1. during development, a warning on console tells developers when they are returning a response (either with `Response` or `NextResponse`).
2. at build time, this warning becomes an error.
3. at run time, returning a response body will trigger a 500 HTTP error with a JSON payload containing the detailed error.
All returned/thrown errors contain a link to the documentation.
This is a breaking feature compared to the _beta_ middleware implementation, and also removes `NextResponse.json()` which makes no sense any more.
### How to try it?
- runtime behavior: `HEADLESS=true yarn jest test/integration/middleware/core`
- build behavior : `yarn jest test/integration/middleware/build-errors`
- development behavior: `HEADLESS=true yarn jest test/development/middleware-warnings`
### Notes to reviewers
The limitation happens in next's web adapter. ~The initial implementation was to check `response.body` existence, but it turns out [`Response.redirect()`](https://github.com/vercel/next.js/blob/canary/packages/next/server/web/spec-compliant/response.ts#L42-L53) may set the response body (https://github.com/vercel/next.js/pull/31886). Hence why the proposed implementation specifically looks at response headers.~
`Response.redirect()` and `NextResponse.redirect()` do not need to include the final location in their body: it is handled by next server https://github.com/vercel/next.js/blob/canary/packages/next/server/next-server.ts#L1142
Because this is a breaking change, I had to adjust several tests cases, previously returning JSON/stream/text bodies. When relevant, these middlewares are returning data using response headers.
About DevEx: relying on AST analysis to detect forbidden use cases is not as good as running the code.
Such cases are easy to detect:
```js
new Response('a text value')
new Response(JSON.stringify({ /* whatever */ })
```
But these are false-positive cases:
```js
function returnNull() { return null }
new Response(returnNull())
function doesNothing() {}
new Response(doesNothing())
```
However, I see no good reasons to let users ship middleware such as the one above, hence why the build will fail, even if _technically speaking_, they are not setting the response body.
## Feature
- [x] Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR.
- [ ] Related issues linked using `fixes #number`
- [x] Integration tests added
- [x] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
- [x] Errors have helpful link attached, see `contributing.md`
## Documentation / Examples
- [x] Make sure the linting passes by running `yarn lint`
BeforeInteractive inline script in v12.1.7-canary.8 don't run. Beacause the script has unknow src.
![image](https://user-images.githubusercontent.com/17813559/169257330-4419228a-6d10-4815-9451-d9a5dd7f011b.png)
Fixes https://github.com/vercel/next.js/issues/31275
## Bug
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
## Feature
- [ ] Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR.
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
- [ ] Errors have helpful link attached, see `contributing.md`
## Documentation / Examples
- [ ] Make sure the linting passes by running `yarn lint`
The Edge SSR server and the client bundle should share the same Styled JSX instance to ensure the context can be passed correctly, same with the way we handle `next/head`.
## Bug
- [ ] Related issues linked using `fixes #number`
- [x] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
## Feature
- [ ] Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR.
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
- [ ] Errors have helpful link attached, see `contributing.md`
## Documentation / Examples
- [ ] Make sure the linting passes by running `yarn lint`
This PR deprecates declaring a middleware under `pages` in favour of the project root naming it after `middleware` instead of `_middleware`. This is in the context of having a simpler execution model for middleware and also ships some refactor work. There is a ton of a code to be simplified after this deprecation but I think it is best to do it progressively.
With this PR, when in development, we will **fail** whenever we find a nested middleware but we do **not** include it in the compiler so if the project is using it, it will no longer work. For production we will **fail** too so it will not be possible to build and deploy a deprecated middleware. The error points to a page that should also be reviewed as part of **documentation**.
Aside from the deprecation, this migrates all middleware tests to work with a single middleware. It also splits tests into multiple folders to make them easier to isolate and work with. Finally it ships some small code refactor and simplifications.
Hello,
This is an iteration after first work at https://github.com/vercel/next.js/pull/36478.
What that PR missed is a way to just get a cookie value. Well, this PR adds two new things:
`cookies.get` returns the cookie value that could be `string | undefined`:
```js
const response = new NextResponse()
response.cookies.set('foo', 'bar', { path: '/test' })
const value = response.cookies.get('foo')
console.log(value) // => 'bar'
```
Additionally, if you want to know all the cookie details, you can use `cookies.getWithOptions`:
```js
const response = new NextResponse()
response.cookies.set('foo', 'bar', { path: '/test' })
const { value, options } response.cookies.getWithOptions('foo')
console.log(value) // => 'bar'
console.log(options) // => { Path: '/test' }
```
This PR changes the experimental `layout=raw` images to use the native lazy loading behavior (as opposed to the IntersectionObserver).
This will (eventually) lead to smaller client bundles and faster image loading since there is no JS needed to load the image.
However, we'll lose the `lazyRoot` and `lazyBoundary` behavior since those are specific to the IntersectionObserver implementation.
This fixes the interop default from https://github.com/vercel/next.js/pull/36877 on the jest `object-proxy` as it currently causes the below error when running tests in our `with-jest` example:
```sh
TypeError: 'get' on proxy: property '__esModule' is a read-only and non-configurable data property on the proxy target but the proxy did not return its actual value (expected 'true' but got 'false')
```
## Bug
- [ ] Related issues linked using `fixes #number`
- [x] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
x-ref: https://github.com/vercel/next.js/pull/36877
## What's in there?
Partially fixes https://github.com/vercel/edge-functions/issues/82
Relates to #36715
Our webpack plugin for middleware leverages static analysis to detect Dyanamic code evaluation in user `_middleware.js` file (and depedencies). Since edge function runtime do not allow them, the build is aborted.
The use of `Function.bind` is considered invalid, while it is legit. A customer using `@aws-sdk/client-s3` reported it.
This PR fixes it.
Please note that this check is too strict: some dynamic code may be in the bundle (despite treeshaking), but may never be used (because of code branches). Since this point is under discussion, this PR adds tests covering some false positives (`@apollo/react-hook`, `qs` and `has`), but does not change the behavior (consider them as errors).
## Notes to reviewer
I looked for test facilities allowing to download the required 3rd party modules. `createNext()` in production context made my day, but showed two issues:
- `cliOutput` is not cleaned in between tests. While clearance during `stop()` would be annoying, I hope that clearance during `start()` is better.
- if `start()` fails while building, the created instance can never be stopped. This is because we don't clear `childProcess` after `build`.
## Bug
- [x] Related issues linked using `fixes #number`
- [x] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
## Feature
- [ ] Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR.
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
- [ ] Errors have helpful link attached, see `contributing.md`
## Documentation / Examples
- [x] Make sure the linting passes by running `yarn lint`
Applies additional escaping to flight data written to script tags during RSC. A test was added. I'm not aware of any issues reported for this and there are no new errors
## Bug
- [ ] Related issues linked using `fixes #number`
- [x] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
Implements the first part of #33227
- Applies browserslist to JS transforms when `experimental.browsersListForSwc` is enabled.
- You don't have to use browserslist, there's also `legacyBrowsers: false` which will be the new default in Next.js 13. See #33227 for which browsers and why. `legacyBrowsers` requires `browsersListForSwc: true` to function until it is the default.
```js
module.exports = {
experimental: {
legacyBrowsers: false,
browsersListForSwc: true,
}
}
```
I only implemented the JS part of the RFC, the CSS part should be handled in a follow-up PR.
## Bug
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
## Feature
- [x] Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR.
- [x] Related issues linked using `fixes #number`
- [x] Integration tests added
- [ ] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
- [x] Errors have helpful link attached, see `contributing.md`
## Documentation / Examples
- [ ] Make sure the linting passes by running `yarn lint`
Co-authored-by: JJ Kasper <22380829+ijjk@users.noreply.github.com>
This ensures different lockfile versions are handled and we skip patching when the version isn't supported. This also adds an env variable to allow skipping this check if desired.
## Bug
- [x] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
Closes: https://github.com/vercel/next.js/issues/36816
Currently, if you try the following code while developing with Next.js inside Middleware:
```ts
const url = new URL(MY_URL)
const res = await fetch(url)
```
The app will work as expected, and so will the dom types for TS 4.5+. However when running `next build` or `tsc` typescript fails with:
<img width="837" alt="image" src="https://user-images.githubusercontent.com/4278345/168647502-06b8b223-e0cf-4e8b-9a82-cdac51748789.png">
The types for `node-fetch` have been updated to include `URL` so this PR updates the dependency to fix it.
## Bug
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
## Bug
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
## Feature
- [ ] Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR.
- [x] Related issues linked using [19693](https://github.com/vercel/next.js/discussions/19693)
- [ ] Integration tests added
- [ ] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
- [ ] Errors have helpful link attached, see `contributing.md`
## Documentation / Examples
- [ x] Make sure the linting passes by running `yarn lint`
Closes#29959
* remove the experimental web vital hook api
* remove the exported flush effects api and only error on development, keep only usage to styled-jsx
for web vital hook API: The usage is not widly adopted since the existing exported vital api could do the same work. In the future we'll deprecate the `_app.server` in favor of `_app` in server component pages. so that this api won't be required.
for flush effects api: other css-in-js libs are not using the same approach like styled-jsx which holding a style registry and could flush it during streaming. emotion-js and styled-components are still relying on `Document.getInitialProps` atm and we have supported it in latest canary
## Bug
- [x] Related issues linked using `fixes #number`
- [x] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
The PR fixes#30300 and #36855.
The corresponding integration test case has been added.
## Bug
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
## Feature
- [ ] Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR.
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
- [ ] Errors have helpful link attached, see `contributing.md`
## Documentation / Examples
- [ ] Make sure the linting passes by running `yarn lint`
Co-authored-by: Tim Neutkens <6324199+timneutkens@users.noreply.github.com>
Follow up for #35888 to re-enable more test, and re-enable post processors after #36792 has better support for document.gIP with react 18. Apply post-pocessing when the the shell chunk is fully buffered.
re-enabled integration tests for react 18:
- amphtml
- amphtml-custom-optimizer
- app-document
- font-optimization
Fixes#35835
## Bug
- [x] Related issues linked using `fixes #number`
- [x] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
This adds an experimental config for testing `basePath` handling on the client.
x-ref: [slack thread](https://vercel.slack.com/archives/CLDDX2Y0G/p1652221605742559)
## Bug
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
## Feature
- [ ] Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR.
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
- [ ] Errors have helpful link attached, see `contributing.md`
## Documentation / Examples
- [ ] Make sure the linting passes by running `yarn lint`
Follow-up to https://github.com/vercel/next.js/pull/36612 this updates to eagerly load the swc bindings unless babel is being used so that we don't wait for the transform calls to initialize swc. Eagerly loading in jest also allows us to fallback to the wasm bindings when previously we couldn't since they needed to wait for the import.
## Bug
- [x] Related issues linked using `fixes #number`
- [x] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
The PR fixes#30300.
The previous integration test case only checks if `/out/404.html` exists. However, the test passes since `/out/404.html/index.html` is being exported instead.
The PR changes that by checking if a given path exists and is a file.
Stack trace disappears when error is converted to string.
I changed the types in `log.ts` to match `console.log`/`console.error`/`console.warn`.
fixes#31591
When getInitialProps is customized with react 18, since gIP requires to return `html` as doc property which could be used by user-land customization, we do blocking-rendering there and passdown the `html` to document
Fixes#36675Closes#36419
* move FlightManifestPlugin to server compilers
* revert loader condition
* fix module id
* fix test and refactor
Co-authored-by: kodiakhq[bot] <49736102+kodiakhq[bot]@users.noreply.github.com>
This aligns the output with what npm outputs. Without this change,
Next.js causes unwanted changes in package-lock.json.
## Bug
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
## Feature
- [ ] Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR.
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
- [ ] Errors have helpful link attached, see `contributing.md`
## Documentation / Examples
- [x] Make sure the linting passes by running `yarn lint`
fixes#36763fixes#36590
## Feature
- [x] Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR.
- It hasn't been accepted for implementation, although that process isn't clear, and this is a pretty trivial fix.
- [x] Related issues linked using `fixes #number`
- [x] Integration tests added
- [x] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
- This is somewhat inherent in the error log
- [x] Errors have helpful link attached, see `contributing.md`
## Documentation / Examples
- [x] Make sure the linting passes by running `yarn lint`
This PR introduces a more predictable API to manipulate cookies in an Edge Function context.
```js
const response = new NextResponse()
// set a cookie
response.cookies.set('foo, 'bar') // => set-cookie: 'foo=bar; Path=/'`
// set another cookie
response.cookies.set('fooz, 'barz') // => set-cookie: 'foo=bar; Path=/, fooz=barz; Path=/'`
// delete a cookie means mark it as expired
response.cookies.delete('foo') // => set-cookie: 'foo=; Path=/; Expires=Thu, 01 Jan 1970 00:00:00 GMT, fooz=barz; Path=/'`
// clear all cookies means mark all of them as expired
response.cookies.clear() // => set-cookie: 'fooz=; Path=/; Expires=Thu, 01 Jan 1970 00:00:00 GMT, foo=; Path=/; Expires=Thu, 01 Jan 1970 00:00:00 GMT'`
```
This new cookies API uses [Map](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Map) interface, and it's available for `NextRequest` and `NextResponse`.
Additionally, you can pass a specific cookies option as a third argument in `set` method:
```js
response.cookies.set('foo', 'bar', {
path: '/',
maxAge: 60 * 60 * 24 * 7,
httpOnly: true,
sameSite: 'strict',
domain: 'example.com'
}
```
**Note**: `maxAge` it's in seconds rather than milliseconds.
Any cookie manipulation will be reflected over the `set-cookie` header, transparently.
closes#31719
## Bug
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Errors have helpful link attached, see `contributing.md`
## Feature
- [ ] Implements an existing feature request or RFC. Make sure the feature request has been accepted for implementation before opening a PR.
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
- [ ] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
- [ ] Errors have helpful link attached, see `contributing.md`
## Documentation / Examples
- [ ] Make sure the linting passes by running `yarn lint`