* Expose CSS Issuer When Error
This introduces a friendlier error for invalid CSS imports by telling you the offending file.
---
Fixes#9847
* fix test matches
* Add check: true behavior to custom routes
* Update adding dev routes
* Add checking of pages and dynamic routes for check: true
* Fix hasPage binding
* Add tests for check: true behavior
* Update regex checking
* Make changes based on review
* Update to handle rewrites in serverless loader
* Update to not change req.url
* Make sure to always parse dynamic route params
* Export all of pathToRegexp from path-match
Co-authored-by: Joe Haddad <timer150@gmail.com>
* CSS Module Support
* Fix Server-Side Render of CSS Modules
* Fix Jest Snapshots
https://github.com/facebook/jest/pull/8492
* Fix snapshots
* Add test for CSS module edit without remounting
* Add tests for dev and production style being applied
* Add missing TODOs
* Include/exclude should only be applied to issuer, not the CSS file itself
* Add CSS modules + node_modules tests
* Test that content is correct
* Create Multi Module Suite
* Add client-side navigation support for CSS
* Add tests for client-side nav
* Add some delays
* Try another fix
* Increase timeout to 3 minutes
* Fix test
* Give all unique directories
* WIP
* Move data experiment
* Do not throw away rules
* Remove test code
* Correct next data behavior
* Add support for async composing
* Remove unnecessary workaround
* Rename Field
* WIP: rename get static paths
* Adjust logic
* Add warning
* rename constant
* Check for error when using old method
* Add friendly error message
* Test for message of invalid keys
* replace index with dots
* Polyfilling fetch and object-assign
* Polyfilling corejs object-assign
* Adding object-assign in polyfills.js. IE11 does not support Object.assign
* Fixing failing test
* Updating object.assign polyfill to fix aliasing
* Updating test case value to match new build stats
* Increasing the size of default build to 225kb
* Fixing defer-script test case to not include polyfill.js
* Revert README.md
* Re-design the polyfill approach based on PR feedback
* Adding comment and fixing test case
* Rename polyfills chunk
* Extract aliases into helper
* Remove extra new line
* Fix TypeScript typings
* Adding _internal_fetch alias
* Adjust build manifest plugin
* Build manifest plugin changes - adding a separate entry for polyfills
* Rename polyfills entry in build-manifest.json
* Remove old comment
* Fix TS
* Set key
* Polyfills already added
* Filtring polyfill.module.js
* Fix test
* Add __internal_fetch to alias rule
* Adjust name
* bump size
* ignore polyfills
* sigh
* Add initial bit for plugins
* Add checks for needed metadata values
* Add test
* Initial plugins changes
* Add handling for _app middleware
* Add loading of _document middleware and
handling of multiple default export syntaxes
* Fix insert order for middleware member expression
* Remove early return from middleware plugin from testing
* Add tests for current plugin middlewares
* Update test plugin package.json
* Update handling for class default export
* Update to use webpack loader instead of babel
plugin, remove redundant middleware naming, and add field for required env for plugins
* Add middleware to support material-ui use case
and example material-ui plugin
* Update tests and remove tests stuff from google analytics plugin
* Remove old plugin suite
* Add init-server middleware
* Exit hard without stack trace when error in collecting plugins
* Add on-error-client and on-error-server and update
to run init-server with next-start in serverless mode
* Update init-client for google analytics plugin
* Add example Sentry plugin and update with-sentry-simple
* Remove middleware field/folder and use src dir for plugins
* Add post-hydration middleware and update
material-ui plugin
* Put plugins code behind flag
* Update chromedriver
* Revert "Update chromedriver"
This reverts commit 1461e978e677f7da05e29e0415ec614a04bf65f9.
* Update lock file
* Remove un-needed _app for sentry example
* Add auto loading of scoped packages, add plugins
config for manually listing plugins, and update
to only collect plugins once
* Update example plugins
* Expose plugins' config
* Rename plugin lifecycles and add babel-preset-build
* Rename other methods with unstable
* Update log when plugin config overrides auto-detecting
* Enable granular chunks config for all chunk types
Adding comment
Bug fix
* Update index.test.js
* Update index.test.js
* Fix test cases. Adding next/link to not trigger a different bug
* Removing obsolete comment
* Pass config.experimental.cpus to export during build
Currently, there is no way of specifying the number of worker threads of
`next export` when run as part of `next build`.
I suggest a sane default should be to just use the same amount of
workers that were used during the build process which currently seems to
be configured through `config.experimental.cpus`.
This setting is already respected in the two other places where
jest-workers are in use: The TerserPlugin and the staticCheckWorkers in
`next build`.
* Only enable worker threads if there is more than 1 worker
Multiple worker threads can cause problems when certain dependencies are
being used, see e.g. https://github.com/zeit/next.js/issues/7894
This patch allows disabling of worker threads by setting
`config.experimental.cpus = 1`.
The benefit of spawning 1 worker thread, if there is any at all, should
very limited anyways, so the workload can just as well be processed in
the main thread.
* Disable parallel build for firebase authentication example
* Add integration test to cover #7894
* Rename test suite and add worker_threads config
* Disable worker_threads by default
* Update index.test.js
* Use workerThreads config for TerserPlugin
* Update to use workerThreads config in
TerserPlugin for consistency
* Disable node 12 specific test
This pull request is a temporary addition that uses the `x-now-route-params` header in serverless.
This header returns the regex groups with indexes, not their named variants.
As a result, we must use the getRouteMatcher utility to reverse this into Next.js' expected names.
Since this got complex, I've added a test for it. We should probably remove this behavior sooner than later.
* Add buildId to SPR data routes
* Update buildId replace in serverless loader
* Use new RegExp and add comment
* Apply suggestions from code review
Co-Authored-By: Joe Haddad <joe.haddad@zeit.co>
* Test to ensure client transition and handle / data route
* Informative Error for Invalid Global CSS
This adds a helpful error message with a (basic) err.sh link for invalid Global CSS usage.
We'll want to expand on this topic more and offer alternatives when CSS Modules support lands.
* Update expected error message
* initial commit for SPRv2
* Add initial SPR cache handling
* update SPR handling
* Implement SPR handling in render
* Update tests, handle caching with serverless next
start, add TODOs, and update manifest generating
* Handle no prerender-manifest from not being used
* Fix url.parse error
* Apply suggestions from code review
Co-Authored-By: Joe Haddad <joe.haddad@zeit.co>
* Replace set with constants in next-page-config
* simplify sprStatus.used
* Add error if getStaticProps is used with getInitialProps
* Remove stale TODO
* Update revalidate values in SPR cache for non-seeded routes
* Apply suggestions from code review
* Remove concurrency type
* Rename variable for clarity
* Add copying prerender files during export
* Add comment for clarity
* Fix exporting
* Update comment
* Add additional note
* Rename variable
* Update to not re-export SPR pages from build
* Hard navigate when fetching data fails
* Remove default extension
* Add brackets
* Add checking output files to prerender tests
* Adjust export move logic
* Clarify behavior of export aggregation
* Update variable names for clarity
* Update tests
* Add comment
* s/an oxymoron/contradictory/
* rename
* Extract error case
* Add tests for exporting SPR pages and update
/_next/data endpoint to end with .json
* Relocate variable
* Adjust route building
* Rename to unstable
* Rename unstable_getStaticParams
* Fix linting
* Only add this when a data request
* Update prerender data tests
* s/isServerless/isLikeServerless/
* Don't rely on query for `next start` in serverless mode
* Rename var
* Update renderedDuringBuild check
* Add test for dynamic param with bracket
* Fix serverless next start handling
* remove todo
* Adjust comment
* Update calculateRevalidate
* Remove cache logic from render.tsx
* Remove extra imports
* Move SPR cache logic to next-server
* Remove old isDynamic prop
* Add calling App getInitialProps for SPR pages
* Update revalidate logic
* Add isStale to SprCacheValue
* Update headers for SPR
* add awaiting pendingRevalidation
* Dont return null for revalidation render
* Adjust logic
* Be sure to remove coalesced render
* Fix data for serverless
* Create a method coalescing utility
* Remove TODO
* Extract send payload helper
* Wrap in-line
* Move around some code
* Add tests for de-duping and revalidating
* Update prerender manifest test
* Replace worker-farm with jest-worker
* Apply suggestions from code review
Co-Authored-By: Joe Haddad <joe.haddad@zeit.co>
* Remove semaphores on top of jest-worker, unwind
terser worker, and remove extra error log
* Fix dynamic APIs with query params
If you define a dynamic API such as
```
pages/
api/
[id].js
```
This api becomes available at `/api/[id]`. If you send a request with
a query parameter the value of `req.query.id` will include the query
string as well as the path parameter.
E.g. the request to `/api/2?test=123` will result in `req.query`
being
```javascript
{
id: "2?test=123",
test: "123",
}
```
instead of
```javascript
{
id: "2",
test: "123",
}
```
* Fix url parse in serverless loader
* Add serverless test
* Add custom profiler / tracer
This captures both client/production builds and everything before and after is cpu profiled, which makes sure that the whole process instead of just webpack is included in the final report.
* Update tests
* Update profiling-plugin.js
* Rename profiling-plugin.js to profiling-plugin.ts
* Update profiling-plugin.ts
* Update and rename profiling-plugin.ts to profiling-plugin.js
* Update webpack-config.ts
* Update profiling-plugin.js
* Add types
* Add missing type
* Add back TypeScript linting
* Remove tracing-js
* Remove accidental support for invalid JavaScript
Next.js inadvertently supported users mixing ES Modules and CommonJS due to a Babel bug that was fixed in v7.
This removes the hack in favor of the new Babel `sourceType` feature.
**This will result in broken code for users who relied on this non-standard conforming behavior.**
* Fix comment
* Toggle feature behind modern flag
* Add checking for hybrid AMP during static check
* Update to always hit server for prerender data
* Add removing of SPR code when not used
* Add checking for hybrid AMP during static check
* Update to always hit server for prerender data
* Add removing of SPR code when not used
* Update dead code elimination
* Refactor SplitChunksPlugin configs and add experimental chunking strategy
* Use typeDefs for SplitChunksConfig
* Modify build manifest plugin to create runtime build manifest
* Add support for granular chunks to page-loader
* Ensure normal behavior if experimental granularChunks flag is false
* Update client build manifest to remove iife & implicit global
* Factor out '/_next/' prepending into getDependencies
* Update packages/next/build/webpack-config.ts filepath regex
Co-Authored-By: Jason Miller <developit@users.noreply.github.com>
* Simplify dependency load ordering in page-loader.js
* Use SHA1 hash to shorten filenames for dependency modules
* Add scheduler to framework cacheGroup in webpack-config
* Update page loader to not duplicate script tags with query parameters
* Ensure no slashes end up in the file hashes
* Add prop-types to framework chunk
* Fix issue with mis-attributed events
* Increase modern build size budget--possibly decrement after consulting with @janicklasralph
* Use module.rawRequest for lib chunks
Co-Authored-By: Daniel Stockman <daniel.stockman@gmail.com>
* Dasherize lib chunk names
Co-Authored-By: Daniel Stockman <daniel.stockman@gmail.com>
* Fix typescript errors, reorganize lib name logic
* Dasherize rawRequest, short circuit name logic when rawRequest found
* Add `scheduler` package to test regex
* Fix a nit
* Adjust build manifest plugin
* Shorten key name
* Extract createPreloadLink helper
* Extract getDependencies helper
* Move method
* Minimize diff
* Minimize diff x2
* Fix Array.from polyfill
* Simplify page loader code
* Remove async=false for script tags
* Code golf `getDependencies` implementation
* Require lib chunks be in node_modules
* Update packages/next/build/webpack-config.ts
Co-Authored-By: Joe Haddad <timer150@gmail.com>
* Replace remaining missed windows compat regex
* Trim client manifest
* Prevent duplicate link preload tags
* Revert size test changes
* Squash manifest size even further
* Add comment for clarity
* Code golfing 🏌️♂️
* Correctly select modern dependencies
* Ship separate modern client manifest when module/module enabled
* Update packages/next/build/webpack/plugins/build-manifest-plugin.ts
Co-Authored-By: Joe Haddad <timer150@gmail.com>
* Remove unneccessary filter from page-loader
* Add lookbehind to file extension regex in page-loader
* v9.0.3
* Update examples for Apollo with AppTree (#8180)
* Update examples for Apollo with AppTree
* Fix apolloClient being overwritten when rendering AppTree
* Golf page-loader (#8190)
* Remove lookbehind for module replacement
* Wait for build manifest promise before page load or prefetch
* Updating modern-only chunks inside the right entry point
* Fixing ts errors
* Rename variable
* Revert "Wait for build manifest promise before page load or prefetch"
This reverts commit c370528c6888ba7fa71162a0854534ed280224ef.
* Use proper typedef for webpack chunk
* Re-enable promisified client build manifest
* Fix bug in getDependencies map
* Insert check for granularChunks in page-loader
* Increase size limit temporarily for granular chunks
* Add 50ms delay to flaky test
* Set env.__NEXT_GRANULAR_CHUNKS in webpack config
* Reset size limit to 187
* Set process.env.__NEXT_GRANULAR_CHUNKS to false if selectivePageBuilding
* Update test/integration/production/test/index.test.js
Co-Authored-By: Joe Haddad <timer150@gmail.com>
* Do not create promise if not using chunking PR
* Experimental: Serverless Trace target
The Serverless Trace target produces Serverless-handler wrapped entrypoints, but does not bundle all of `node_modules`.
This behavior increases bundling performance to be more akin to `target: 'server'`.
This mode is expected to be used with smart platforms (like [ZEIT Now](https://zeit.co/now) that can trace a program to its minimum dependencies.
* Use more generic variables
* Add asset relocator for production mode of serverless trace
* Verify Firebase compatiblity
* Revert "Add asset relocator for production mode of serverless trace"
This reverts commit 8404f1dcf28b60edab41a56c94b38dcd3fddec20.
* Add serverless trace tests
* Add _isLikeServerless helper
* Make constants
* Fix export
* Update packages/next-server/server/config.ts
Co-Authored-By: JJ Kasper <jj@jjsweb.site>
* Use a global helper for is like serverless
* Update import for isTargetLikeServerless
* Update packages/next/build/index.ts
Co-Authored-By: JJ Kasper <jj@jjsweb.site>
* Module/nomodule implementation based on RFC 7563
* Remove comment
* Fixing issue with building amp pages
* Fixing test cases for serverless mode
* Adding safari 10 nomodule fix. Preloading modern js by default
* Fixing size-limit integration test
* Bug fix
* Adding testcase for modern build
* Trigger rebuild
* Setting default crossOrigin value
* Moving modern config option inside experimental flag
* Adding nomodule attribute to safari-fix script
* Changing safari10NomoduleFix default value to true
* Removing safari-fix flag
* Changing .es6 to .module
* Disable modern default
* Removing default crossOrigin value. Setting modern flag to false by default. Fixed test cases
* Remove confusing defaults and mark required instead
* Adjust blacklist
* Move behavior of page marking
* Fixing childCompiler errors not being captured
* Tweak names
* Revert
* whoops
* Fixing bug with page-loader.js
* Changing modern babel cache name
* Rename helper
* Iterate over both bundles
* Correctly clamp bundle sizes
* Revert test
* Add modern mode tests
* Fix test
* test
* test2
* Compile for the current version of Node.js by default when producing server builds.
* Update packages/next/build/babel/preset.ts
Co-Authored-By: Joe Haddad <timer150@gmail.com>
* Include server variation in Babel cacheKey
* Add prerender PageConfig option
* Update PageConfig type
* Add inlining of data when pre-render is set and add tests
* Update types import
* Add check for props
* Rename prerender to experimentalPrerender for now
* Re-require all module information for client-side initializers list
We really should re-work how our loadable code words, because this passes an excessive number of modules to the client.
In the ideal world, we only pass the entry module id of each chunk, but there's no point in investing the time until we switch to webpack 5 which completely revamps the chunk graph API.
* Get a unique set of ids
* Add export const config support and make withAmp a no-op
* Use babel plugin for PageConfig
* Fix serverless-loader exports
* Add backwards compatibility for withAmp
* Run prettier over packages/**/*.js
* Run prettier over packages/**/*.ts
* Run prettier over examples
* Remove tslint
* Run prettier over examples
* Run prettier over all markdown files
* Run prettier over json files
* Add canonicalBase config to allow setting
absolute path for canonical link
* Make sure canonicalBase is set for
export and serverless
* Move canonicalBase to amp.canonicalBase
* Update tests with canonicalBase config
* Update tests
* run lint-fix
* Fix canonicalBase config parsing
* Fix canonicalBase during export
* Update amphtml tests
* Update escape string regexp operators
* temp
* Extract getRouteRegex func
* First iteration of dynamic routing for production only
* Correctly order prod
* Add serverless support
* Single line it
* noop routes
* Format doc
* Fix dynamic routing for dev
* Add flag for dynamic routing
* Update packages/next-server/lib/router/router.ts
Co-Authored-By: JJ Kasper <jj@jjsweb.site>
* remove example
* Add router tests
* Format code
* Sort routes
* Update to not use posix path methods
* Add automatic exporting of pages with no getInitialProps
* Add support for exporting serverless to static
and serving the html files during next start
* Fix missing runtimeEnv when requiring page, re-add warning
when trying to export with serverless, and update tests
* Update flying-shuttle test
* revert un-used pagesManifest change
* remove query.amp RegExp test
* Fix windows backslashes not being replaced
* Re-enable serverless support for next start
* bump
* Fix getInitialProps check
* Fix incorrect error check
* Re-add check for reserved pages
* Fix static check
* Update to ignore /api pages and clean up some tests
* Re-add needed next.config for test and correct behavior
* Update RegExp for ignored pages for auto-static
* Add checking for custom getInitialProps in pages/_app
* Update isPageStatic logic to only use default export
* Re-add retrying to CircleCi
* Update query during dev to only have values
available during export for static pages
* Fix test
* Add warning when page without default export is
found and make sure to update pages-manifest
correctly in flying-shuttle mode
* Fix backslashes not being replaced
* Integrate auto-static with flying-shuttle
and make sure AMP is handled in flying-shuttle
* Add autoExport for opting in
* Disable babel-loader cache in flying shuttle mode
* Move cleaning up old babel-loader caches
* Change option name for disabling cache on babel-loader
* Add retrying to CircleCi and run prepublish separately to get better logs in Azure
* Apply workaround for Azure node 10 and ncc
* Update incorrect webpack imports
* Use ncced autodll-webpack-plugin
* Move autodll-webpack-plugin to devDependencies
* Move babel-loader into next all the way converting
it to TypeScript and added tracking of babel-loader
cache so we can clear previous files
* Add license from babel-loader
* Apply suggestions from code review
Co-Authored-By: ijjk <jj@jjsweb.site>
* Make sure to catch errors in next-babel-loader
* Add Flying Shuttle plugin to server compilation
* Ignore build artifacts in flying shuttle manifest
* Add comments explaining what's going on
* Emit Shuttle manifest after both compilations
* Add WithAmp to enable AMP support for
pages instead of .amp.js
* Update handling for exporting AMP
* Fix ampPath in export for / path and
revert isAmp logic to handle right
* Update amphtml test suite
* Add handling for noDirtyAmp during
export and update amp-export test suite
* Update serverless and export-default-map
test suites
* Update require-page tests
* Make async-to-promises babel plugin experimental
* Move excludes for asyncToPromises behind flag too
* Move other configs behind flag
* Re-add original exclude item
* Set custom cacheIdentifier for babel-
loader that we can increment
* Update to consider external babel config
for cacheIdentifier
* Update to follow babel loader's cacheIdentifier behavior
* Disable Terser plugin when using flying shuttle
* Add a new minify loader
* Downgrade terser
* Turn on minify loader
* Only compress source / mangle output
* Add amp request header and check for amp
query in serverless mode
* Add serverless test for an AMP page
* Update to handle querystring weirdness
* Update serverless test config
* Get all modules included in build
* Add tests
* Get all modules contained per entry chunk
* Sort files
* Add specialized page entry to manifest
* Split manifest into pages and chunks key
* Update test
* Use relative paths to build directory
* Update test
* [WIP] Use a shared module cache
* ID modules in development
* Revert "ID modules in development"
This reverts commit 0613d92fa2c8c7fa11a5ff5b7770d784af1cec63.
* Remove context replacement
* Only enable shared runtime in prod
* Sort settings
* Add shared runtime experimental setting
* only enable shared runtime in serverless
* Add a new field to webpack types
* Revert "Add a new field to webpack types"
This reverts commit d35fa02207fbfd0085da0fc56aac42c4ff7c34c9.
* Add HashedChunkIdsPlugin to make consistent chunk ids
* Revert "Revert "Add a new field to webpack types""
This reverts commit 338219049e1432038f90c91928b010bbb1267999.
* Make it optional
* Remove record ids
* Revert "Remove record ids"
This reverts commit 15c22dbcda72466c382397c91d02295620f62326.
Fixes https://github.com/zeit/now-builders/issues/168
For some reason with a certain mix of deps `...` is not supported in webpack's parsing.
By default it is supported as all our tests passed before and we have deployed Next.js apps on v2 already.
It looks like :
```
Pages sizes after gzip:
┌ / (196 B)
├ /_app (11.5 kB)
├ /_error (4.44 kB)
├ /blog (196 B)
└ /blog/page (195 B)
```
(style inspired from now-cli : https://github.com/zeit/now-cli/blob/canary/src/util/output/builds.js)
I'll add dynamic chunks in a separate PR.
@timneutkens Do you want to keep `_app` and `_error` or filter them out ? I think it's a good idea to keep them, because `_app` can get pretty large and it would encourage code splitting in that case.
Extends on #5927, instead of `.default` we'll expose `.render` which is semantically more correct / mirrors the naming of the custom server API.
I've updated the spec in #5927 to reflect this change.
(copied from #5927):
```js
const http = require('http')
const page = require('./.next/serverless/about.js')
const server = new http.Server((req, res) => page.render(req, res))
server.listen(3000, () => console.log('Listening on http://localhost:3000'))
```
**This does not change existing behavior.**
building to serverless is completely opt-in.
- Implements `target: 'serverless'` in `next.config.js`
- Removes `next build --lambdas` (was only available on next@canary so far)
This implements the concept of build targets. Currently there will be 2 build targets:
- server (This is the target that already existed / the default, no changes here)
- serverless (New target aimed at compiling pages to serverless handlers)
The serverless target will output a single file per `page` in the `pages` directory:
- `pages/index.js` => `.next/serverless/index.js`
- `pages/about.js` => `.next/serverless/about.js`
So what is inside `.next/serverless/about.js`? All the code needed to render that specific page. It has the Node.js `http.Server` request handler function signature:
```ts
(req: http.IncomingMessage, res: http.ServerResponse) => void
```
So how do you use it? Generally you **don't** want to use the below example, but for illustration purposes it's shown how the handler is called using a plain `http.Server`:
```js
const http = require('http')
// Note that `.default` is needed because the exported module is an esmodule
const handler = require('./.next/serverless/about.js').default
const server = new http.Server((req, res) => handler(req, res))
server.listen(3000, () => console.log('Listening on http://localhost:3000'))
```
Generally you'll upload this handler function to an external service like [Now v2](https://zeit.co/now-2), the `@now/next` builder will be updated to reflect these changes. This means that it'll be no longer neccesary for `@now/next` to do some of the guesswork in creating smaller handler functions. As Next.js will output the smallest possible serverless handler function automatically.
The function has 0 dependencies so no node_modules are required to run it, and is generally very small. 45Kb zipped is the baseline, but I'm sure we can make it even smaller in the future.
One important thing to note is that the function won't try to load `next.config.js`, so `publicRuntimeConfig` / `serverRuntimeConfig` are not supported. Reasons are outlined here: #5846
So to summarize:
- every page becomes a serverless function
- the serverless function has 0 dependencies (they're all inlined)
- "just" uses the `req` and `res` coming from Node.js
- opt-in using `target: 'serverless'` in `next.config.js`
- Does not load next.config.js when executing the function
TODO:
- [x] Compile next/dynamic / `import()` into the function file, so that no extra files have to be uploaded.
- [x] Setting `assetPrefix` at build time for serverless target
- [x] Support custom /_app
- [x] Support custom /_document
- [x] Support custom /_error
- [x] Add `next.config.js` property for `target`
Need discussion:
- [ ] Since the serverless target won't support `publicRuntimeConfig` / `serverRuntimeConfig` as they're runtime values. I think we should support build-time env var replacement with webpack.DefinePlugin or similar.
- [ ] Serving static files with the correct cache-control, as there is no static file serving in the serverless target
This PR Fixes#4920
So the problem is that when a next.js application is built on windows, the `pages-manifest.json` file is created with backslashes. If this built application is deployed to a linux hosting enviroment, the server will fail when trying to load the modules.
```
Error: Cannot find module '/user_code/next/server/bundles\pages\index.js
```
My simple solution is to modify the `pages-manifest.json` to always use linux separator (`/`), then also
modify `server/require.js` to, when requiring page, replace any separator (`\` or `/`) with current platform-specific file separator (`require('path').sep`).
The fix in `server/require.js` would be sufficient, but my opinion is that having some cross-platform consistency is nice.
This change was tested by bulding an application in windows and running it in linux and windows, aswell as building an application in linux and running it in linux and windows. The related tests was also run.
# Conflicts:
# test/integration/production/test/index.test.js
- Replaces taskr-babel with taskr-typescript for the `next` package
- Makes sure Node 8+ is used, no unneeded transpilation
- Compile Next.js client side files through babel the same way pages are
- Compile Next.js client side files to esmodules, not commonjs, so that tree shaking works.
- Move error-debug.js out of next-server as it's only used/require in development
- Drop ansi-html as dependency from next-server
- Make next/link esmodule (for tree-shaking)
- Make next/router esmodule (for tree-shaking)
- add typescript compilation to next-server
- Remove last remains of Flow
- Move hoist-non-react-statics to next, out of next-server
- Move htmlescape to next, out of next-server
- Remove runtime-corejs2 from next-server
* Remove flow-typed
* Remove flow types
* Remove the last types
* Bring back taskr dependency
* Revert "Bring back taskr dependency"
This reverts commit 38cb95d7274d63fe63c6ac3c95ca358a28c17895.
* Bring back preset-flow as it’s used for tests
* Revert "Revert "Bring back taskr dependency""
This reverts commit b4c933ef133f4039f544fb10bf31d5c95d3b27a2.
Takes advantage of caching between builds for Terser, also makes writing caches for babel-loader faster by disabling compression.
Results for zeit.co (350 pages):
Without cache:
[4:16:22 PM] Compiled server in 1m
[4:16:57 PM] Compiled client in 2m
✨ Done in 125.83s.
With cache:
[4:19:38 PM] Compiled client in 17s
[4:19:50 PM] Compiled server in 29s
✨ Done in 31.79s.
Note: these results are from my multi-core Macbook Pro 2017, exact specs:
MacBook Pro (13-inch, 2017, Four Thunderbolt 3 Ports)
- 3,3 GHz Intel Core i5
- 16 GB 2133 MHz LPDDR3
- Intel Iris Plus Graphics 650 1536 MB
The `without cache` build runs uglify in parallel, so without cache is likely to take longer on environments where you have only 1 core available.
The `with cache` build however runs in a single thread, so the results should be similar.
Clears the way a bit for #4943, also makes _document.js less complex, and will allow us to move `__NEXT_DATA__` to a `application/json` script tag.
Also this causes a slightly smaller bundle size 😌
