We also close the connection when the window is in the background and re-connect when it is brought to the foreground. This prevents us from using up too many connections.
This PR adds the `_error`-page as an `404.html`-export, even when it is not explicitly defined in a custom `exportPathMap`.
It also fixes two false negative tests related to this. Previously the tests were matching the fallback 404-page from the test-server, rather than the `404.html`-page from next, which was actually not being generated. The test server is also not set up to serve `/404.html` as `/404` which the tests now reflect.
**Caveat**
In its current state, this PR removes `/404.html` from the `defaultPathMap` passed to the custom `exportPathMap`-functions, since it instead adds it after that function is run. While it is possible that someone is relying on this to exist, it is to my knowledge undocumented and also unlikely to be used for anything but merging it into the custom pathMap.
Since this would now merge `undefined` which would result in it being added later on anyway, I deemed it safe, but would be happy to undo that part of the PR if necessary as it was only cleanup.
**Examples**
As a way to demonstrate what this PR does, this is how examples changed:
* `basic-export` - Behaviour is unchanged, still has a `404.html`
* `with-static-export` - Now has a `404.html`
if `&=1` is used to point to the `amphtml` version of a page it get's escaped causing it to become `&amp=1` this fixes it to stay as `&=1`
Closes: #7038
* Remove client bundles for AMP pages
after build since they are not used
* Remove trailing white space
* Use async-sema to limit removing AMP client bundles
* Bring AMP client bundle removing
semaphore concurrency down to 20
* Don't check blocked pages when
deleting AMP client bundles
* Update client bundle removing for AMP pages
* Add error handling for removing client AMP pages
* rethrow error unless ENOENT during
deleting AMP client pages
* Handle error during removing AMP client
pages the same during dev
* Fix throwing instead of rejecting
* Make sure next/config is set before requiring page
* Update error check
* return on reject
* Fix next/config
* Update amphtml and canonical rels, put amp behind
experimental flag again, and update checking for amp query
* Fix typescript error
* Re-add flag to next.config.js
Not sure when it was introduced, but on the latest Next v8.0.4, there's no descriptive name attached to the component returns from `withRouter`.
This makes it harder to debug and write test cases for components wrapped by `withRouter`.
* Only refresh the page when the active
page is updated in AMP mode
* Update handling of page reload to make sure it
still refreshes after a change to another page
* Update checking to be more accurate
* Fix amp-dev not being loaded without
experimental.amp and remove next.config from amp example
* Remove old with-amp example and
rename experimental-amp to with-amp
* update example name
Co-Authored-By: ijjk <jj@jjsweb.site>
* Update comment wording
Co-Authored-By: ijjk <jj@jjsweb.site>
* Use document for reload to keep scroll position
Co-Authored-By: ijjk <jj@jjsweb.site>
* fallback to reloading on error
Co-Authored-By: ijjk <jj@jjsweb.site>
* Update with-amp example readme
* Make sure AmpContext is available during _document
render and update filtering of script tags in AMP mode
* Update amphtml test to make sure
AmpContext is set for _document render
* Fix stray comma in render.tsx
* 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
* Remove CLI and negative page building
* Remove select all tests
* Update test of getSpecifiedPages
* Prune old tests from snapshot
* Remove old flying shuttle tests
* Fix re-rendering on client after navigating
back from external site
* Clean up tagging before unload
* Add check for history.state 1/2
Co-Authored-By: ijjk <22380829+ijjk@users.noreply.github.com>
* Add check for history.state 2/2
Co-Authored-By: ijjk <22380829+ijjk@users.noreply.github.com>
* Add check for options
Co-Authored-By: ijjk <22380829+ijjk@users.noreply.github.com>
* Add test for navigating to external site and back
also added testing in safari and firefox
* Add test for query in url
* Disable Terser plugin when using flying shuttle
* Add a new minify loader
* Downgrade terser
* Turn on minify loader
* Only compress source / mangle output
* Do not clear the console
Its rude to clear the console, you may be sharing output with other processes even in tty mode.
* Remove unused dependency
* Dedupe and cleanup dev output without clearing
* use logError
* Remove exit handler
* Add next helper
* Add log helpers
* Switch store to log helpers and a shallow object compare
* Update other files to use new logging utility
* request => build
* Update ready on messages
* Use case insensitive matching
* Resolve Next.js loaders first
Next.js loaders should be resolved first, then node_modules. There is no need for NEXT_PROJECT_ROOT_NODE_MODULES anymore since all our loaders are internal
* Move wasm to end of array
Since wasm isn’t the main target for Next.js it makes no sense to try and resolve it as the first value, as it would incur a file not found for nearly all files in the compilation
Checking the stats first. But at least until https://github.com/zeit/next.js/issues/6950 is figured out we should disable it for serverless builds as it's also disabled for zeit.co serverless builds.
