This pull request adds "Vary: Accept" header to responses from the image optimizer (i.e. the /_next/image endpoint).
The image optimizer prefers re-encoding JPG files to WebP, but some browsers (such as Safari 14 on Catalina) do not yet support WebP. In such cases the optimizer uses the Accept header sent by the browser to send out a JPG response. Thus the optimizer's response may depend on the Accept header.
Potential caching proxies can be informed of this fact by adding "Vary: Accept" to the response headers. Otherwise WebP data may be served to browsers that do not support it, for example in the following scenario:
* A browser that supports WebP requests the JPG. The optimizer re-encodes it to WebP. The proxy caches the WebP data.
* After this another browser that doesn't support WebP requests the JPG. The proxy sends the WebP data to the browser.
- [x] Integration tests added
- [x] Make sure the linting passes
This solves the main use case from Issue #19914.
Previously, we would set the `Cache-Control` header to a constant and rely on the server cache. This would mean the browser would always request the image and the server could response with 304 Not Modified to omit the response body.
This PR changes the behavior such that the `max-age` will propagate from the upstream server to the Next.js Image Optimization Server and allow browser caching. ("upstream" meaning external server or just an internal route to an image)
This PR does not change the `max-age` for static imports which will remain `public, max-age=315360000, immutable`.
#### Pros:
- Fewer HTTP requests after initial browser visit
- User configurable `max-age` via the upstream image `Cache-Control` header
#### Cons:
- ~~Might be annoying for `next dev` when modifying a source image~~ (solved: use `max-age=0` for dev)
- Might cause browser to cache longer than expected (up to 2x longer than the server cache if requested in the last second before expiration)
## Bug
- [x] Related issues linked using `fixes #number`
`next-dev-server` having its own implementations of `renderToHTML` and `renderErrorToHTML` has historically made reasoning about streaming hard, as it adds additional places where status codes are explicitly set and the full HTML is blocked on.
Instead, this PR simplifies things considerably by moving the majority of the custom logic for e.g. hot reloading and on-demand compilation to when we're resolving the page to be loaded, rather than upfront when handling the request. It also cleans up a few other details (e.g. default error page rendering) that managed to creep into the base implementation over time.
One unfortunate side effect is that this makes compilation errors slightly more difficult. Previously, we'd render them directly. Now, we have to rethrow them. But since they've already been logged (by the watch pipeline), we have to make sure they don't get logged again.
Fixes#24056.
## Bug
- [x] Related issues linked using `fixes #number`
- [x] Integration tests added
## 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.
## Documentation / Examples
- [ ] Make sure the linting passes
This expands on https://github.com/vercel/next.js/issues/24070 and ensures we show the dev overlay for additional cases like where `_app` or `_document` have syntax errors causing compilation to not be able to complete. This achieves showing the dev overlay even when compilation fails from a syntax error by doing a third minimal compilation in development with the needed client-side assets to render the dev overlay.
## Bug
- [x] Related issues linked using `fixes #number`
- [x] Integration tests added
x-ref: https://github.com/vercel/next.js/issues/24070
* import next-server logic during the time the configuration is loaded
* load minimizer plugins only when used
* load ReactDevOverlay only when used
* load only meta information of tsconfig for validation
* make worker for configuration loading lighter
* only load runTypeCheck when used
* load postcss config only when used
Tobias has fixed the `compiler.hooks.invalid.call()` bug in webpack 5 so this is no longer needed
## Bug
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
## 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.
## Documentation / Examples
- [ ] Make sure the linting passes
This pull request **temporarily** removes ESLint, as it was not landed in accordance with our standard experimental policies. We are fully committed to landing this change again.
This is being reverted because:
- Next.js has very strict goals for its install size. This feature resulted in adding over 17MB, or a 43.6% increase.
- The feature was not first landed under the `experimental` key in `next.config.js`, rather, it was added under the stable namespace (top-level)
- Using the feature doesn't do a "guided setup" like TypeScript, it should ask you to "bring your own" dependencies for ESLint
- It uses a undesirable ESLint plugin name: `plugin:@next/next/recommended`. This should read out as strictly `next`, or as short as we can get it.
- Does not provide actionable warnings (missing link to resolve issue)
- Does not follow appropriate console output styling. We need to revisit how these are presented.
To re-land this, we need to ensure the following minimums are met:
- Very minor change in install size
- Fully experimental (i.e. flagged) with warnings
- Finalized package name and configuration shape, preferably so we can do ` { extends: 'next' } `.
This adds support for returning an object from `rewrites` in `next.config.js` with `beforeFiles`, `afterFiles`, and `fallback` to allow specifying rewrites at different stages of routing. The existing support for returning an array for rewrites is still supported and behaves the same way. The documentation has been updated to include information on these new stages that can be rewritten and removes the outdated note of rewrites not being able to override pages.
## Bug
- [ ] Related issues linked using `fixes #number`
- [ ] Integration tests added
## 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`
- [x] Integration tests added
- [x] Documentation added
- [ ] Telemetry added. In case of a feature if it's used or not.
## Documentation / Examples
- [ ] Make sure the linting passes
For #22228
This PR:
- Adds ESLint to toolchain
- Included by default for builds (`next build`)
- Can be enabled for development (`next dev`)
- Custom formatter built for output
- Adds appropriate tests
- Adds two documentation pages
This PR upgrades `jest-worker` and `jest-cli` to the latest pre-release version, also removed `jest-circus` which is included in Jest by default. `jest-worker@next` includes a fix for memory leak that we need (https://github.com/facebook/jest/pull/11187).
Fixes#22925. This will also improve the OOM issue for `next dev` #15855.
A number of changes here. I recommend viewing the diff with the <a href="?w=1">whitespace flag enabled</a>.
- OpenTelemetry is replaced with a custom and lightweight tracing solution.
- Three trace targets are currently supported: console, Zipkin, and NextJS.
- Tracing is now governed by environment variables rather than `--require instrument.js`.
+ `TRACE_TARGET`: one of `CONSOLE`, `ZIPKIN`, or `TELEMETRY`; defaults to `TELEMETRY` if unset or invalid.
+ `TRACE_ID`: an 8-byte hex-encoded value used as the Zipkin trace ID; if not provided, this value will be randomly generated and passed down to subprocesses.
Other sundry:
- I'm missing something, probably a setup step, with the Zipkin target. Traces are captured successfully, but you have to manually enter the Trace ID in order to view the trace - it doesn't show up in queries.
- I'm generally unhappy with [this commit](235cedcb3e). It is... untidy to provide a telemetry object via `setGlobal`, but I don't have a ready alternative. Is `distDir` strictly required when creating a new Telemetry object? I didn't dig too deep here.
As noted, there are a lot of changes, so it'd be great if a reviewer could:
- [ ] pull down the branch and try to break it
- [ ] check the Zipkin traces and identify possible regressions in the functionality
Closes#22570Fixes#22574
This adds generating a static 500 status page when a `pages/500.js` file is added similar to how we handle generating static 404 pages when `pages/404.js` is present. This allows showing a customized error page when a 500 error occurs in an optimal way.
This removes `import type` usage from our core files since `import type` requires a higher TypeScript version than currently expected.
Fixes: https://github.com/vercel/next.js/issues/19300
This upgrades to ncc@0.25.0 and fixes the previous bugs including:
* ncc not referenced correctly in build
* Babel type errors
* node-fetch, etag, chalk and raw-body dependencies not building with ncc - these have been "un-ncc'd" for now. As they are relatively small dependencies, this doesn't seem too much of an issue and we can follow up in the tracking ncc issue at https://github.com/vercel/ncc/issues/612.
* `yarn dev` issues
Took a lot of bisecting, but the overall diff isn't too bad here in the end.
This adds the initial changes outlined in the [i18n routing RFC](https://github.com/vercel/next.js/discussions/17078). This currently treats the locale prefix on routes similar to how the basePath is treated in that the config doesn't require any changes to your pages directory and is automatically stripped/added based on the detected locale that should be used.
Currently redirecting occurs on the `/` route if a locale is detected regardless of if an optional catch-all route would match the `/` route or not we may want to investigate whether we want to disable this redirection automatically if an `/index.js` file isn't present at root of the pages directory.
TODO:
- [x] ensure locale detection/populating works in serverless mode correctly
- [x] add tests for locale handling in different modes, fallback/getStaticProps/getServerSideProps
To be continued in fall-up PRs
- [ ] add tests for revalidate, auto-export, basePath + i18n
- [ ] add mapping of domains with locales
- [ ] investigate detecting locale against non-index routes and populating the locale in a cookie
x-ref: https://github.com/vercel/next.js/issues/17110
Prior to this pull request, Next.js would immediately decode all URLs sent to its server (via `path-match`).
This was rarely needed, and Next.js would typically re-encode the incoming request right away (see all the `encodeURIComponent`s removed in PR diff). This adds unnecessary performance overhead.
Long term, this will also help prevent weird encoding edge-cases like #10004, #10022, #11371, et al.
---
No new tests are necessary for this change because we've extensively tested these edge cases with existing tests.
One test was updated to reflect that we skip decoding in a 404 scenario.
Let's see if all the existing tests pass!
This makes sure to the page path is the expected version to trigger refreshing on the client and adds additional tests to make sure it is working properly with these page variants.
Closes: https://github.com/vercel/next.js/issues/16938
This adds initial support for reloading the page when `getStaticProps`, `getStaticPaths`, or `getServerSideProps` were changed for a page by triggering a reload when the server output for a page has changed but the client output has not since these methods aren't included in the client output.
Closes: https://github.com/vercel/next.js/issues/13949
By popular request, this pull request adds support for returning `fallback: 'blocking'` from `getStaticPaths`.
This new mode will cause unknown paths to be rendered on-demand ("SSR") without the static (placeholder) fallback.
This feature is **currently experimental and should not be used in production yet**. It's currently flagged behind `unstable_`:
```
fallback: 'unstable_blocking'
```
TODO:
- [x] Next.js tests
- [ ] Add Vercel support
- [ ] Vercel tests
---
Fixes#15637
- Using `namedChunks` where possible, this will also allow for faster access to the chunks as we no longer have to look them up like we did before using `find`
- Using the new asset hooks introduced in the latest webpack beta
- Using the new externals function signature
Chunks already have a normalized path.
Not sure if there are other chunks that require this change, I did a global search and didn't find similar cases.
This adds handling for custom-routes with `basePath` to automatically add the `basePath` for custom-routes `source` and `destination` unless `basePath: false` is set for the route.
Closes: https://github.com/vercel/next.js/issues/14782
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 builds off of @timneutkens's PR instead of updating it because it's his `canary` branch.
Updated to still `stat`, as it's the only way to test the difference between a file and directory.
---
Closes#13506Fixes#12235
Extracted from https://github.com/vercel/next.js/pull/13333, the same exact code lives in that PR as well, but we can merge this separately if it makes reviewing https://github.com/vercel/next.js/pull/13333 easier
This PR does 3 things
- deduplicate code from build and next-dev-server that loads custom routes from next.config.js (`loadCustomRoutes`)
- in `loadCustomRoutes`, load these rewrites, headers and redirects configs concurrently instead of sequentially.
- in next-server, make `this.customRoutes` always defined, this allows us to remove the big `if` around its initialization code in `generateRoutes`, which in turn makes it possible to reuse this code for other routing than user defined routes, which is how https://github.com/vercel/next.js/pull/13333 adds its redirects.
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
**First, apologies for a second PR on the same issue but I was working on this already so I thought I'd push it and let you decide which you want to merge.**
The PR is related to [13466](https://github.com/vercel/next.js/issues/13466).
Based on my research, the error happens if the options are empty, null, or undefined. That's why I have decided that the most proper check would be using the! post-fix expression operator may assert that its operand is non-null and non-undefined. ``if (options == null)``
(Optional)
I have also added a warning, which warns the user if the passed "dev" variable is not a boolean.
It's my first PR on the "packages" part of the repo so I'd be glad to receive all kinds of critics. If you want me to change or remove anything, I'm open to suggestions.
---
Fixes#13466
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.
This code existed mostly to work around webpack 2 (yes 2.x) limitations where it crashed in certain cases where files didn't exist anymore. We have tests for that behavior and latest webpack has fixed these. Hence why this can be removed 👍
This removes `fork-ts-checker-webpack-plugin` and instead directly calls the TypeScript API.
This is approximately 10x faster.
Base build: 7s (no TypeScript features enabled)
- `fork-ts-checker-webpack-plugin@3.1.1`: 90s, computer sounds like an airplane
- `fork-ts-checker-webpack-plugin@4.1.6`: 84s, computer did **not** sound like an airplane
- `fork-ts-checker-webpack-plugin@5.0.0-alpha.14`: 90s, regressed
- `npx tsc -p tsconfig.json --noEmit`: 12s (time: `18.57s user 0.97s system 169% cpu 11.525 total`)
- **This PR**: 22s, expected to get better when we run this as a side-car
All of these tests were run 3 times and repeat-accurate within +/- 0.5s.
As discussed this adds bundling of `.env` files in `serverless` mode so that the environment values are also available when deploying with this target
closes: https://github.com/zeit/next.js/issues/13332
* fix(debugging): do not pass NODE_OPTIONS='--inspect' to subprocesses
fixes#11030
* fix(debugger): use a regex to remove bad NODE_OPTIONS flags
Co-authored-by: Alec Larson <alec.stanford.larson@gmail.com>