This PR removes a legacy router method that was used for old-style HMR, now replaced by Fast Refresh.
This method was not public:
```tsx
export type NextRouter = BaseRouter &
Pick<
Router,
| 'push'
| 'replace'
| 'reload'
| 'back'
| 'prefetch'
| 'beforePopState'
| 'events'
| 'isFallback'
>
```
Even if someone found this method, it's highly unlikely they could use it successfully—it required the full module object.
This PR adds a second experimental post-processing step for the framework introduced by @prateekbh in #14746. The image post-processing step scans the rendered document for the first few images and uses a simple heuristic to determine if the images should be automatically preloaded.
Analysis of quite a few production Next apps has shown that a lot of sites are taking a substantial hit to their [LCP](https://web.dev/lcp/) score because an image that's part of the "hero" element on the page is not preloaded and is getting downloaded with lower priority than the JavaScript bundles. This post-processor should automatically fix that for a lot of sites, without causing any real performance effects in cases where it fails to identify the hero image.
This feature is behind an experimental flag, and will be subject to quite a bit of experimentation and tweaking before it's ready to be made a default setting.
In terms of url rewriting, `trailingSlash` supports everything `exportTrailingSlash` does. We can just share all other code paths and deprecate `exportTrailingSlash`.
This PR shows a deprecation warning when `exportTrailingSlash` is used.
Also fixes https://github.com/vercel/next.js/issues/15774
We can update the tests now or later. (I kept them the same to prove it's non-breaking)
To do:
- [x] Do we want to keep this? => nope 841d4efc51/packages/next/next-server/lib/router/router.ts (L329)
- [x] I kept `exportTrailingSlash` here. Do we want to rename that as well? => nope 2d9d649d49/packages/next/build/index.ts (L959)
Fixes https://github.com/vercel/next.js/issues/15188
`parseRelativeUrl` was used on urls that weren't always relative. It was used to generate a cache key, but we actually don't need these cache keys to be relative if the urls aren't relative.
Also took a look at the overall static data fetching logic and found a few things:
- [x] cache key is unnecessarily transformed through `prepareRoute`, we can just cache by resolved `dataHref` and remove that function. Pretty sure that `prepareRoute` was also introducing edge cases with `assetPath` and `delBasePath`
- [x] there is [a bug in the caching logic](ebdfa2e7a3/packages/next/next-server/lib/router/router.ts (L898)) that made it fail on the second visit: it should be `Promise.resolve(this.sdc[pathname])` instead of `Promise.resolve(this.sdc[dataHref])`. Also added a test for this
- [x] ~converted to async await to improve stacktraces and readability.~ I assumed this was fine since I saw some async/awaits in that file already but it seems to just blow up the size of the non-modern bundle.
- [x] extracted nested `getResponse` function and define it top level. this should improve runtime performance
- [x] convert `_getStaticData` and `_getServerData` to class methods instead of properties. Not sure why they were defined as properties but I think they belong on the prototype instead.
- [x] remove `cb` property from `fetchNextData`, it's unnecessary and makes the async flow hard to understand. The exact same logic can go in the `.then` instead.
- [ ] data fetching logic [retries on 5xx errors](ebdfa2e7a3/packages/next/next-server/lib/router/router.ts (L157)), but not on network level errors. It should also retry on those. It should also not retry on every 5xx, probably only makes sense on 502, 503 and 504. (e.g. 500 is a server error that I wouldn't expect to succeed on a retry)
The overall result also is a few bytes smaller in size
Replace `url.parse` and `url.resolve` logic with whatwg `URL`, Bring in a customized `format` function to handle the node url objects that can be passed to router methods. This eliminates the need for `url` (and thus `native-url`) in core. Looks like it shaves off about 2.5Kb, according to the `size-limits` integration tests.
This adds additional checks against the routeKeys used to build the named regexes for dynamic routes to ensure they follow PCRE rules for named capture groups
x-ref: https://github.com/vercel/vercel/pull/4813
Discovered while working on https://github.com/vercel/next.js/pull/14848
when asPath is the same but href is different it should use `replaceState` instead of `pushState`, so that browser back/forward behavior is preserved. Currently it's comparing a path that includes basepath with one that excludes it, so `pushState` is always used. This makes sure the behavior is the same as when running next.js without a basepath
This updates the scroll position saving to occur as the scroll position changes instead of trying to do it when the navigation is changing since the `popState` event doesn't allow us to update the leaving history state once the `popState` has occurred.
The order of events that was previously attempted to save scroll position on a `popState` event (back/forward navigation)
1. history.state is already updated with state from `popState`
2. we replace state with the currently rendered page adding scroll info
3. we replace state again with the `popState` event state overriding scroll info
Using this approach the above event order is no longer in conflict since we don't attempt to populate the state with scroll position while it's leaving the state and instead do it while it is still the active state in history
This approach resembles existing solutions:
https://www.npmjs.com/package/scroll-behaviorhttps://twitter.com/ryanflorence/status/1029121580855488512
Fixes: https://github.com/vercel/next.js/issues/13990Fixes: #12530
x-ref: https://github.com/vercel/next.js/pull/14075
Fixes: #13512
Defined and exported type for `metric` used in [reportWebVitals](https://nextjs.org/docs/advanced-features/measuring-performance)
```
export function reportWebVitals(metric) {
if (metric.label === 'custom') {
console.log(metric) // The metric object ({ id, name, startTime, value, label }) is logged to the console
}
}
```
One can now do
```
import { NextWebVitalsMetric } from 'next/app'
export function reportWebVitals(metric: NextWebVitalsMetric ) {
if (metric.label === 'custom') {
console.log(metric) // The metric object ({ id, name, startTime, value, label }) is logged to the console
}
}
```
* avoid pulling code in the bundle for `trailingSlash` logic when it's not enabled
* avoid cloning the url an extra time if normalizing the path doesn't change it
Avoid trailing slashes on urls that look like files. The redirect for `trailingSlash: true` will now look like:
```
Redirects
┌ source: /:path*/:file.:ext/
├ destination: /:path*/:file.:ext
└ permanent: true
┌ source: /:path*/:notfile([^/.]+)
├ destination: /:path*/:notfile/
└ permanent: true
```
The default still looks like:
```
Redirects
┌ source: /:path+/
├ destination: /:path+
└ permanent: true
```
After this gets merged, I have a few optimizations planned on the normalization code that should reduce the client bundle a little and that consolidates the `trailingSlash` and `exportTrailingSlash` options
This updates `fetchNextData` to re-use the `getDataHref` function from `page-loader` which has more verbose handling to ensure the correct `/_next/data` URL is built. Re-using this logic ensures the `/_next/data` URL can still be built even when a mismatching `href` and `as` value is provided to `next/link`.
This also fixes a case in `getDataHref` where optional values that weren't provided would fail to build the data href since the check requiring the param be present while interpolating the route values hasn't been updated to allow missing params for optional values.
An additional test case has been added to the prerender suite to ensure the `/_next/data` URL is built correctly when mismatching `href` and `as` values are provided
x-ref: https://github.com/vercel/next.js/discussions/14536
x-ref: https://github.com/vercel/next.js/discussions/9081#discussioncomment-31160
Closes: https://github.com/vercel/next.js/issues/14668
Add tests and fix for when the url contains query parameters.
`router` now uses the same method for formatting url+as pair as `Link`, will be able to share code after https://github.com/vercel/next.js/pull/14633 is merged
* Avoid adding basePath when it's not needed
When using the `basePath` setting, on pages with params it will fire a router change. This will pass the url pathname in the `as` param using the `getUrl()` function. This means the `as` path will be sent through already including the `basePath`, leading to `/basePath/basePath/path` which will cause the router to throw an error.
* lint
* Add test case and ensure removal
* Make sure to re-add before changeState
Co-authored-by: JJ Kasper <jj@jjsweb.site>
Noticed this while reviewing https://github.com/vercel/next.js/pull/14376. After having done https://github.com/vercel/next.js/pull/13699, this code didn't feel right to me:
```js
function prepareRoute(path: string) {
path = delBasePath(path || '')
// this /index rewrite is problematic, it makes pages/index.js
// and pages/index/index.js point to the same thing:
return toRoute(!path || path === '/' ? '/index' : path)
}
```
Added a nested index page to the prerender tests and found it was rendering the `/` route on navigation. This uncovered 2 more places around the dataroute where the index path was not translated correctly.
**edit:**
Just to note that there was nothing wrong with https://github.com/vercel/next.js/pull/14376, the issue was already there, I just noticed it while reading that PR
Noticed while working on https://github.com/vercel/next.js/pull/14400 that the optional catch-all handling was missing in `namedRegex`.
This whole file also seemed quite regex heavy so I took a look at the overall logic and changed a few things. It worked by regex escaping the whole route then unescape the dynamic parts. I changed it to only regex escape the static parts, this eliminates unnecessary back and forth escaping. It also makes the dynamic parts handling more readable. The whole logic is less reliant on regexes and just uses simple string manipulation to translate the route into a regex, I didn't measure anything but as an effect this should make it more performant.
Updates the way filenames are generated for browser compilation.
Notably:
- All entry bundles now have hashes in production, this includes pages (previously pages used a buildId in the path)
- The AmpFiles no longer depends on hardcoded bundle names, it uses the buildManifest instead (internals)
- All cases where we match the page name from the chunk/entrypoint name now use the same function `getRouteFromEntrypoint` (internals)
- In development we no longer include the "faked" `buildId` set to `development` for page files, instead we just use the `/_next/static/pages` path (was `/_next/static/development/pages`). This was changed as it caused unneeded complexity and makes generating the bundles easier (internals)
- Updated tons of tests to be more resilient to these changes by relying on the buildManifest instead of hardcoded paths (internals)
Follow up of these PRs:
https://github.com/vercel/next.js/pull/13759https://github.com/vercel/next.js/pull/13870https://github.com/vercel/next.js/pull/13937https://github.com/vercel/next.js/pull/14130https://github.com/vercel/next.js/pull/14176https://github.com/vercel/next.js/pull/14268Fixes#6303Fixes#12087Fixes#1948Fixes#4368Fixes#4255Fixes#2548
This corrects the `/_next/data` path generated when using `basePath` with `getStaticProps` in a `pages/index.js` file which was previously stripping the `basePath` without checking if `/index` needed to be appended after stripping. This also adds additional checks to the `basePath` test suite to prevent regressing
x-ref: https://github.com/vercel/next.js/pull/9872#issuecomment-646841260
This updates the named regexes output in the `routes-manifest` and the associated `routeKeys` to not use any non-word characters as this breaks the named regexes e.g. `"Invalid regular expression: "^/(?<data\-provider\-id>[^/]+?)(?:/)?$"`
x-ref: https://github.com/zeit/now/pull/4355
To make `asPath` consistent with `basePath` handling this makes sure it is always stripped including on the client under the `asPath` value and from `req.url` in the `serverless-loader`. Additional tests have been added for this behavior to ensure we don't regress on this
Closes: https://github.com/vercel/next.js/issues/14037
Closes: https://github.com/vercel/next.js/issues/14039
This correctly strips the `basePath` before generating the `route-matcher` for dynamic routes and adds regression tests to ensure these work correctly with the `basePath` feature
Closes: https://github.com/vercel/next.js/issues/13966
This adds scroll restoration handling to make sure the correct scroll position is restored after navigating back/forward to a page and the rendering hasn't completed by the time the default browser scroll restoration has taken place.
An initial failing test case was added which is working with the changes in this PR, if there are any other cases that should be added let me know and I can make sure we have them to ensure we don't regress on this behavior
---
Fixes#12530
As discussed, this streamlines the handling for `basePath` to not automatically strip and add the `basePath` when provided to `next/link` or `router.push/replace` and only automatically adds the `basePath` and when it is manually provided it will cause a 404 which ensures `href` still matches to the pages directory 1-to-1.
This also adds additional test cases that we discussed to ensure this behavior is working as intended
---
Fixes#13902
So I can't *entirely* explain why, but I believe this fixes#13132. 🙈 I basically ended up looking around at other `_next` URLs (are those asset URLs?) around the project and seeing that they tended to use `delBasePath()` to remove the base path from the current page's path whenever it was used.
When testing locally with the [repo submitted with the issue](https://github.com/robertovg/next-base-path-example), I no longer experience the constant page-reloading in dev mode when adding a query string to the URL.
Disambiguate between pages/index.js and pages/index/index.js so that they resolve differently.
It all started with a bug in pagesmanifest that propagated throughout the codebase. After fixing pagesmanifest I was able to remove a few hacks here and there and more logic is shared now. especially the logic that resolves an entrypoint back into a route path. To sum up what happened:
- `getRouteFromEntrypoint` is the inverse operation of `getPageFile` that's under `pages/_document.tsx`
- `denormalizePagePath` is the inverse operation of `normalizePagePath`.
Everything is refactored in terms of these operations, that makes their behavior uniform and easier to update/patch in a central place. Before there were subtle differences between those that made `index/index.js` hard to handle.
Some potential follow up on this PR:
- [`hot-reloader`](https://github.com/vercel/next.js/pull/13699/files#diff-6161346d2c5f4b7abc87059d8768c44bR207) still has one place that does very similar behavior to `getRouteFromEntrypoint`. It can probably be rewritten in terms of `getRouteFromEntrypoint`.
- There are a few places where `denormalizePagePath(normalizePagePath(...))` is happening. This is a sign that `normalizePagePath` is doing some validation that is independent of its rewriting logic. That should probably be factored out in its own function. after that I should probably investigate whether `normalizePagePath` is even still needed at all.
- a lot of code is doing `.replace(/\\/g, '')`. If wanted, that could be replaced with `normalizePathSep`.
- It looks to me like some logic that's spread across the project can be centralized in 4 functions
- `getRouteFromEntrypoint` (part of this PR)
- its inverse `getEntrypointFromRoute` (already exists in `_document.tsx` as `getPageFile`)
- `getRouteFromPageFile`
- its inverse `getPageFileFromRoute` (already exists as `findPageFile ` in `server/lib/find-page-file.ts`)
It could be beneficial to structure the code to keep these fuctionalities close together and name them similarly.
- revise `index.amp` handling in pagesmanifest. I left it alone in this PR to keep it scoped, but it may be broken wrt nested index files as well. It might even make sense to reshape the pagesmanifest altogether to handle html/json/amp/... better
Was going through _document and noticed some variable shadowing going on. Added a rule for it to our eslint configuration and went through all warnings with @Timer.
Fixes https://github.com/vercel/next.js/issues/13524
To do:
- [x] fix dev mode
- [x] there's a ~route ordering or~ trailing slash issue with top level catch-all (current tests reflect that)
- [x] in this test `/get-static-paths/whatever` should fall back to `/[[...optionalName]].js` since `fallback` is `false` in its `getStaticPaths` method. ~Currently seems to 500~ must have been a glitch
- [x] add tests for `null`, `undefined` ~and `false`~ behavior as well (if decided these are valid)
- [x] ~add tests for string params as well~ this is not allowed for catch-all routes
- [x] test behavior when fallback is enabled and a top level catch-all exists
This waits for the render to be committed to DOM before we render the route change complete event (no longer sync in new React).
We have tests that ensure this resolves.
---
Closes#12938
* Generate sourcemaps for core files that pass through Babel
* Run files through Babel instead of tsc
* Get rid of wildcard helper
* Get rid of wildcard helper
* Remove unused file
* Update wildcard imports
* Add exclude
* Get rid of object-assign helper
* Use Object.assign as it gives better output
Co-authored-by: Joe Haddad <joe.haddad@zeit.co>
* Fix type on NextApiHandler
The example [here](https://nextjs.org/docs/api-routes/api-middlewares) shows that `NextApiHandler` can be async, so the return type should be `void | Promise<void>`.
* Add TS integration tests for API
This allows a page to be fully static (no runtime JavaScript) on a per-page basis.
The initial implementation does not disable JS in development mode as we need to figure out a way to inject CSS from CSS imports / CSS modules without executing the component JS. This restriction is somewhat similar to https://www.gatsbyjs.org/packages/gatsby-plugin-no-javascript/. All things considered that plugin only has a usage of 600 downloads per week though, hence why I've made this option unstable/experimental initially as I'd like to see adoption patterns for it first.
Having a built-in way to do this makes sense however as the people that do want to adopt this pattern are overriding Next.js internals currently and that'll break between versions.
Related issue: #5054 - Not adding `fixes` right now as this implementation needs more work. If anyone wants to work on this feel free to reach out on https://twitter.com/timneutkens
* Add basePath in link component and add/remove it consistently
* Update to not use regex for delBasePath
* Expose addBasePath as router method
* Revert "Expose addBasePath as router method"
This reverts commit 40fed596195c6affabf837e42d472452768e13a3.
* Expose basePath as router field
* Apply suggestion
* Expose basePath as router field
* remove un-used vars
* Update externals
* Apply lint fix
* Update size-limit test
* Update prefetch
* Add initial support for new env config file
* Fix serverless processEnv call when no env is provided
* Add missing await for test method
* Update env config to .env.json and add dotenv loading
* ncc dotenv package
* Update type
* Update with new discussed behavior removing .env.json
* Update hot-reloader createEntrypoints
* Make sure .env is loaded before next.config.js
* Add tests for all separate .env files
* Remove comments
* Add override tests
* Add test for overriding env vars based on local environment
* Add support for .env.test
* Apply suggestions from code review
Co-Authored-By: Joe Haddad <joe.haddad@zeit.co>
* Use chalk for env loaded message
* Remove constant as it’s not needed
* Update test
* Update errsh, taskr, and CNA template ignores
* Make sure to only consider undefined missing
* Remove old .env ignore
* Update to not populate process.env with loaded env
* Add experimental flag and add loading of global env values
Co-authored-by: Tim Neutkens <timneutkens@me.com>
Co-authored-by: Joe Haddad <joe.haddad@zeit.co>