FiberNode stateNode could be null
So I get TypeError:
```
at performWorkOnRoot (/tmp/my-project/node_modules/react-dom/cjs/react-dom.development.js:11014:24) TypeError: Cannot read property '_warnedAboutRefsInRender' of null
at findDOMNode (/tmp/my-project/node_modules/react-dom/cjs/react-dom.development.js:15264:55)
```
* Updates inside controlled events (onChange) are sync even in async mode
This guarantees the DOM is in a consistent state before we yield back
to the browser.
We'll need to figure out a separate strategy for other
interactive events.
* Don't rely on flushing behavior of public batchedUpdates implementation
Flush work as an explicit step at the end of the event, right before
restoring controlled state.
* Interactive updates
At the beginning of an interactive browser event (events that fire as
the result of a user interaction, like a click), check for pending
updates that were scheduled in a previous interactive event. Flush the
pending updates synchronously so that the event handlers are up-to-date
before responding to the current event.
We now have three classes of events:
- Controlled events. Updates are always flushed synchronously.
- Interactive events. Updates are async, unless another a subsequent
event is fired before it can complete, as described above. They are
also slightly higher priority than a normal async update.
- Non-interactive events. These are treated as normal, low-priority
async updates.
* Flush lowest pending interactive update time
Accounts for case when multiple interactive updates are scheduled at
different priorities. This can happen when an interactive event is
dispatched inside an async subtree, and there's an event handler on
an ancestor that is outside the subtree.
* Update comment about restoring controlled components
* ReactDOM.flushControlled
New API for wrapping event handlers that need to fire before React
yields to the browser. Previously we thought that flushSync was
sufficient for this use case, but it turns out that flushSync is only
safe if you're guaranteed to be at the top of the stack; that is, if
you know for sure that your event handler is not nested inside another
React event handler or lifecycle. This isn't true for cases like
el.focus, el.click, or dispatchEvent, where an event handler can be
invoked synchronously from inside an existing stack.
flushControlled has similar semantics to batchedUpdates, where if you
nest multiple batches, the work is not flushed until the end of the
outermost batch. The work is not guaranteed to synchronously flush, as
with flushSync, but it is guaranteed to flush before React yields to
the browser.
flushSync is still the preferred API in most cases, such as inside
a requestAnimationFrame callback.
* Test that flushControlled does not flush inside batchedUpdates
* Make flushControlled a void function
In the future, we may want to return a thenable work object. For now,
we'll return nothing.
* flushControlled -> unstable_flushControlled
Removes the `useSyncScheduling` option from the HostConfig, since it's
no longer needed. Instead of globally flipping between sync and async,
our strategy will be to opt-in specific trees and subtrees.
* Bump deps to Jest 22
* Prevent jsdom from logging intentionally thrown errors
This relies on our existing special field that we use to mute errors.
Perhaps, it would be better to instead rely on preventDefault() directly.
I outlined a possible strategy here: https://github.com/facebook/react/issues/11098#issuecomment-355032539
* Update snapshots
* Mock out a method called by ReactART that now throws
* Calling .click() no longer works, dispatch event instead
* Fix incorrect SVG element creation in test
* Render SVG elements inside <svg> to avoid extra warnings
* Fix range input test to use numeric value
* Fix creating SVG element in test
* Replace brittle test that relied on jsdom behavior
The test passed in jsdom due to its implementation details.
The original intention was to test the mutation method, but it was removed a while ago.
Following @nhunzaker's suggestion, I moved the tests to ReactDOMInput and adjusted them to not rely on implementation details.
* Add a workaround for the expected extra client-side warning
This is a bit ugly but it's just two places. I think we can live with this.
* Only warn once for mismatches caused by bad attribute casing
We used to warn both about bad casing and about a mismatch.
The mismatch warning was a bit confusing. We didn't know we warned twice because jsdom didn't faithfully emulate SVG.
This changes the behavior to only leave the warning about bad casing if that's what caused the mismatch.
It also adjusts the test to have an expectation that matches the real world behavior.
* Add an expected warning per comment in the same test
* Harden tests around init/addition/update/removal of aliased attributes
I noticed some patterns weren't being tested.
* Call setValueForProperty() for null and undefined
The branching before the call is unnecessary because setValueForProperty() already
has an internal branch that delegates to deleteValueForProperty() for null and
undefined through the shouldIgnoreValue() check.
The goal is to start unifying these methods because their separation doesn't
reflect the current behavior (e.g. for unknown properties) anymore, and obscures
what actually happens with different inputs.
* Inline deleteValueForProperty() into setValueForProperty()
Now we don't read propertyInfo twice in this case.
I also dropped a few early returns. I added them a while ago when we had
Stack-only tracking of DOM operations, and some operations were being
counted twice because of how this code is structured. This isn't a problem
anymore (both because we don't track operations, and because I've just
inlined this method call).
* Inline deleteValueForAttribute() into setValueForAttribute()
The special cases for null and undefined already exist in setValueForAttribute().
* Delete some dead code
* Make setValueForAttribute() a branch of setValueForProperty()
Their naming is pretty confusing by now. For example setValueForProperty()
calls setValueForAttribute() when shouldSetAttribute() is false (!). I want
to refactor (as in, inline and then maybe factor it out differently) the relation
between them. For now, I'm consolidating the callers to use setValueForProperty().
* Make it more obvious where we skip and when we reset attributes
The naming of these methods is still very vague and conflicting in some cases.
Will need further work.
* Rewrite setValueForProperty() with early exits
This makes the flow clearer in my opinion.
* Move shouldIgnoreValue() into DOMProperty
It was previously duplicated.
It's also suspiciously similar in purpose to shouldTreatAttributeValueAsNull()
so I want to see if there is a way to unify them.
* Use more specific methods for testing validity
* Unify shouldTreatAttributeValueAsNull() and shouldIgnoreValue()
* Remove shouldSetAttribute()
Its naming was confusing and it was used all over the place instead of more specific checks.
Now that we only have one call site, we might as well inline and get rid of it.
* Remove unnecessary condition
* Remove another unnecessary condition
* Add Flow coverage
* Oops
* Fix lint (ESLint complains about Flow suppression)
* Fix treatment of Symbol/Function values on boolean attributes
They weren't being properly skipped because of the early return.
I added tests for this case.
* Avoid getPropertyInfo() calls
I think this PR looks worse on benchmarks because we have to read propertyInfo in different places.
Originally I tried to get rid of propertyInfo, but looks like it's important for performance after all.
So now I'm going into the opposite direction, and precompute propertyInfo as early as possible, and then just pass it around.
This way we can avoid extra lookups but keep functions nice and modular.
* Pass propertyInfo as argument to getValueForProperty()
It always exists because this function is only called for known properties.
* Make it clearer this branch is boolean-specific
I wrote this and then got confused myself.
* Memoize whether propertyInfo accepts boolean value
Since we run these checks for all booleans, might as well remember it.
* Fix a crash when numeric property is given a Symbol
* Record attribute table
The changes reflect that SSR doesn't crash with symbols anymore (and just warns, consistently with the client).
* Refactor attribute initialization
Instead of using flags, explicitly group similar attributes/properties.
* Optimization: we know built-in attributes are never invalid
* Use strict comparison
* Rename methods for clarity
* Lint nit
* Minor tweaks
* Document all the different attribute types
* Deduplication of warn selected on option
- Wrote a failing test
- Deduplication when selected is set on option
* Ran yarn preitter
* Fixed PR request
- Moved dedupe test to above
- Moved && case to seperate if to seperate static and dynamic things
- Render'd component twice
* Actually check for deduplication
* Minor nits
* Harden tests around init/addition/update/removal of aliased attributes
I noticed some patterns weren't being tested.
* Call setValueForProperty() for null and undefined
The branching before the call is unnecessary because setValueForProperty() already
has an internal branch that delegates to deleteValueForProperty() for null and
undefined through the shouldIgnoreValue() check.
The goal is to start unifying these methods because their separation doesn't
reflect the current behavior (e.g. for unknown properties) anymore, and obscures
what actually happens with different inputs.
* Inline deleteValueForProperty() into setValueForProperty()
Now we don't read propertyInfo twice in this case.
I also dropped a few early returns. I added them a while ago when we had
Stack-only tracking of DOM operations, and some operations were being
counted twice because of how this code is structured. This isn't a problem
anymore (both because we don't track operations, and because I've just
inlined this method call).
* Inline deleteValueForAttribute() into setValueForAttribute()
The special cases for null and undefined already exist in setValueForAttribute().
* Delete some dead code
* Make setValueForAttribute() a branch of setValueForProperty()
Their naming is pretty confusing by now. For example setValueForProperty()
calls setValueForAttribute() when shouldSetAttribute() is false (!). I want
to refactor (as in, inline and then maybe factor it out differently) the relation
between them. For now, I'm consolidating the callers to use setValueForProperty().
* Make it more obvious where we skip and when we reset attributes
The naming of these methods is still very vague and conflicting in some cases.
Will need further work.
* Rewrite setValueForProperty() with early exits
This makes the flow clearer in my opinion.
* Move shouldIgnoreValue() into DOMProperty
It was previously duplicated.
It's also suspiciously similar in purpose to shouldTreatAttributeValueAsNull()
so I want to see if there is a way to unify them.
* Use more specific methods for testing validity
* Unify shouldTreatAttributeValueAsNull() and shouldIgnoreValue()
* Remove shouldSetAttribute()
Its naming was confusing and it was used all over the place instead of more specific checks.
Now that we only have one call site, we might as well inline and get rid of it.
* Remove unnecessary condition
* Remove another unnecessary condition
* Add Flow coverage
* Oops
* Fix lint (ESLint complains about Flow suppression)
* ValidateDOMNesting tests(#11299)
* Rewrite tests using only public API.
* Modified the tests to prevent duplication of code.
* Code review changes implemented.
* Removed the .internal from the test file name as
its now written using public APIs.
* Remove mutation
* Remove unnecessary argument
Now that we pass warnings, we don't need to pass a boolean.
* Move things around a bit, and add component stack assertions
* Fix autoFocus for hydration content when it is mismatched
* Add a test for mismatched content
* Fix a test for production
* Fix a spec description and verify console.error output
* Run prettier
* finalizeInitialChildren always returns `true`
* Revert "finalizeInitialChildren always returns `true`"
This reverts commit 58edd228046bcafcbcd04a70cb5e78520b50a07e.
* Add a TODO comment
* Update ReactServerRendering-test.js
* Update ReactServerRendering-test.js
* Rewrite the comment
* Ensure value and defaultValue do not assign functions and symbols
* Eliminate assignProperty method from ReactDOMInput
* Restore original placement of defaultValue reservedProp
* Reduce branching. Make assignment more consistent
* Control for warnings in symbol/function tests
* Add boolean to readOnly assignments
* Tweak the tests
* Invalid value attributes should convert to an empty string
* Revert ChangeEventPlugin update. See #11746
* Format
* Replace shouldSetAttribute call with value specific type check
DOMProperty.shouldSetAttribute runs a few other checks that aren't
appropriate for determining if a value or defaultValue should be
assigned on an input. This commit replaces that call with an input
specific check.
* Remove unused import
* Eliminate unnecessary numeric equality checks (#11751)
* Eliminate unnecessary numeric equality checks
This commit changes the way numeric equality for number inputs works
such that it compares against `input.valueAsNumber`. This eliminates
quite a bit of branching around numeric equality.
* There is no need to compare valueAsNumber
* Add test cases for empty string to 0.
* Avoid implicit boolean JSX props
* Split up numeric equality test to isolate eslint disable command
* Fix typo in ReactDOMInput test
* Add todos
* Update the attribute table
I updated ReactDOMInput.synchronizeDefaultValue such that it assignes
the defaultValue property instead of the value attribute. I never
followed up on the ChangeEventPlugin's on blur behavior.
* Inline HTML and SVG configs into DOMProperty
* Replace invariants with warnings
These invariants can only happen if *we* mess up, and happen during init time.
So it's safe to make these warnings, as they would fail the tests anyway.
* Clearer variable naming
* Use defaultValue instead of setAttribute('value')
This commit replaces the method of synchronizing an input's value
attribute from using setAttribute to assigning defaultValue. This has
several benefits:
- Fixes issue where IE10+ and Edge password icon disappears (#7328)
- Fixes issue where toggling input types hides display value on dates
in Safari (unreported)
- Removes mutationMethod behaviors from DOMPropertyOperations
* initialValue in Input wrapperState is always a string
* The value property is assigned before the value attribute. Fix related tests.
* Remove initial value tests in ReactDOMInput
I added these tests after removing the `value` mutation
method. However they do not add any additional value over existing
tests.
* Improve clarity of value checks in ReactDOMInput.postMountWrapper
* Remove value and defaultValue from InputWithWrapperState type
They are already included in the type definition for HTMLInputElement
* Inline stringification of value in ReactDOMInput
Avoids eagier stringification and makes usage more consistent.
* Use consistent value/defaultValue presence in postMountHook
Other methods in ReactDOMInput check for null instead of
hasOwnProperty.
* Add missing semicolon
* Remove unused value argument in ReactDOMInput test
* Address cases where a value switches to undefined
When a controlled input value switches to undefined, it reverts back
to the initial state of the controlled input.
We didn't have test coverage for this case, so I've added two describe
blocks to cover both null and undefined.
* Use `this` inside invokeGuardedCallback
It's slightly odd but that's exactly how our www fork works.
Might as well do it in the open source version to make it clear we rely on context here.
* Move invokeGuardedCallback into a separate file
This lets us introduce forks for it.
* Add a www fork for invokeGuardedCallback
* Fix Flow
* WIP:use public API
* ReactPortal shifted to shared:all passed
* wrote createPortal method for ReactNoop.(#11299)
* imported ReactNodeList type into ReactNoop.(#11299)
* createPortal method implemented.(#11299)
* exec yarn prettier-all.(#11299)
API for batching top-level updates and deferring the commit.
- `root.createBatch` creates a batch with an async expiration time
associated with it.
- `batch.render` updates the children that the batch renders.
- `batch.then` resolves when the root has completed.
- `batch.commit` synchronously flushes any remaining work and commits.
No two batches can have the same expiration time. The only way to
commit a batch is by calling its `commit` method. E.g. flushing one
batch will not cause a different batch to also flush.
* Move ReactFiberTreeReflection to react-reconciler/reflection #11659
* Use * for react-reconciler
We don't know the latest local version, and release script currently doesn't bump deps automatically.
* Remove unused field
* Use CommonJS in entry point for consistency
* Undo the CommonJS change
I didn't realize it would break the build.
* Record sizes
* Remove reconciler fixtures
They're unnecessary now that we run real tests on reconciler bundles.
* Rename escapeText util. Test quoteAttributeValueForBrowser through ReactDOMServer API
* Fix lint errors
* Prettier reformatting
* Change syntax to prevent prettier escape doble quote
* Name and description gardening. Add tests for escapeTextForBrowser. Add missing tests
* Improve script tag as text content test
* Update escapeTextForBrowser-test.js
* Update quoteAttributeValueForBrowser-test.js
* Simplify tests
* Move utilities to server folder
* Extract Jest config into a separate file
* Refactor Jest scripts directory structure
Introduces a more consistent naming scheme.
* Add yarn test-bundles and yarn test-prod-bundles
Only files ending with -test.public.js are opted in (so far we don't have any).
* Fix error decoding for production bundles
GCC seems to remove `new` from `new Error()` which broke our proxy.
* Build production version of react-noop-renderer
This lets us test more bundles.
* Switch to blacklist (exclude .private.js tests)
* Rename tests that are currently broken against bundles to *-test.internal.js
Some of these are using private APIs. Some have other issues.
* Add bundle tests to CI
* Split private and public ReactJSXElementValidator tests
* Remove internal deps from ReactServerRendering-test and make it public
* Only run tests directly in __tests__
This lets us share code between test files by placing them in __tests__/utils.
* Remove ExecutionEnvironment dependency from DOMServerIntegrationTest
It's not necessary since Stack.
* Split up ReactDOMServerIntegration into test suite and utilities
This enables us to further split it down. Good both for parallelization and extracting public parts.
* Split Fragment tests from other DOMServerIntegration tests
This enables them to opt other DOMServerIntegration tests into bundle testing.
* Split ReactDOMServerIntegration into different test files
It was way too slow to run all these in sequence.
* Don't reset the cache twice in DOMServerIntegration tests
We used to do this to simulate testing separate bundles.
But now we actually *do* test bundles. So there is no need for this, as it makes tests slower.
* Rename test-bundles* commands to test-build*
Also add test-prod-build as alias for test-build-prod because I keep messing them up.
* Use regenerator polyfill for react-noop
This fixes other issues and finally lets us run ReactNoop tests against a prod bundle.
* Run most Incremental tests against bundles
Now that GCC generator issue is fixed, we can do this.
I split ErrorLogging test separately because it does mocking. Other error handling tests don't need it.
* Update sizes
* Fix ReactMount test
* Enable ReactDOMComponent test
* Fix a warning issue uncovered by flat bundle testing
With flat bundles, we couldn't produce a good warning for <div onclick={}> on SSR
because it doesn't use the event system. However the issue was not visible in normal
Jest runs because the event plugins have been injected by the time the test ran.
To solve this, I am explicitly passing whether event system is available as an argument
to the hook. This makes the behavior consistent between source and bundle tests. Then
I change the tests to document the actual logic and _attempt_ to show a nice message
(e.g. we know for sure `onclick` is a bad event but we don't know the right name for it
on the server so we just say a generic message about camelCase naming convention).
Fixes a case where changing the name and checked value of a radio button in the same update would lead to checking the wrong radio input. Also adds a DOM test fixture for related issue.
Related issues:
https://github.com/facebook/react/issues/7630
* Use const/let in more places (#11467)
* Convert ReactDOMFiberTextarea to const/let
* Convert ReactDOMSelection to const/let
* Convert setTextContent to const/let
* Convert validateDOMNesting to const/let
* Replace Object.assign by Object Spread
* Convert ReactDOMFiberOption to Object Spread
* Convert ReactDOMFiberTextarea to Object Spread
* Convert validateDOMNesting to Object Spread
* Don't call idle callback unless there's time remaining
* Expiration fixture
Fixture that demonstrates how async work expires after a certain interval.
The fixture clogs the main thread with animation work, so it only works if the
`timeout` option is provided to `requestIdleCallback`.
* Pass timeout option to requestIdleCallback
Forces `requestIdleCallback` to fire if too much time has elapsed, even if the
main thread is busy. Required to make expiration times work properly. Otherwise,
async work can expire, but React never has a chance to flush it because the
browser never calls into React.
* Update Flow
* Fix createElement() issue
The * type was too ambiguous. It's always a string so what's the point?
Suppression for missing Flow support for {is: ''} web component argument to createElement() didn't work for some reason.
I don't understand what the regex is testing for anyway (a task number?) so I just removed that, and suppression got fixed.
* Remove deleted $Abstract<> feature
* Expand the unsound isAsync check
Flow now errors earlier because it can't find .type on a portal.
* Add an unsafe cast for the null State in UpdateQueue
* Introduce "hydratable instance" type
The Flow error here highlighted a quirk in our typing of hydration.
React only really knows about a subset of all possible nodes that can
exist in a hydrated tree. Currently we assume that the host renderer
filters them out to be either Instance or TextInstance. We also assume
that those are different things which they might not be. E.g. it could
be fine for a renderer to render "text" as the same type as one of the
instances, with some default props.
We don't really know what it will be narrowed down to until we call
canHydrateInstance or canHydrateTextInstance. That's when the type is
truly refined.
So to solve this I use a different type for hydratable instance that is
used in that temporary stage between us reading it from the DOM and until
it gets refined by canHydrate(Text)Instance.
* Have the renderer refine Hydratable Instance to Instance or Text Instance
Currently we assume that if canHydrateInstance or canHydrateTextInstance
returns true, then the types also match up. But we don't tell that to Flow.
It just happens to work because `fiber.stateNode` is still `any`.
We could potentially use some kind of predicate typing but instead
of that I can just return null or instance from the "can" tests.
This ensures that the renderer has to do the refinement properly.
* Enable User Timing API integration with a feature flag
* Expose a way to toggle user timing flag in www
* Update ReactNativeCSFeatureFlags.js
* Update ReactFeatureFlags.js
* Fix dead code elimination for feature flags
Turning flags into named exports fixes dead code elimination.
This required some restructuring of how we verify that flag types match up. I used the Check<> trick combined with import typeof, as suggested by @calebmer.
For www, we can no longer re-export `require('ReactFeatureFlags')` directly, and instead destructure it. This means flags have to be known at init time. This is already the case so it's not a problem. In fact it may be better since it removes extra property access in tight paths.
For things that we *want* to be dynamic on www (currently, only performance flag) we can export a function to toggle it, and then put it on the secret exports. In fact this is better than just letting everyone mutate the flag at arbitrary times since we can provide, e.g., a ref counting interface to it.
* Record sizes
* Convert ReactDOM to const/let
* Convert ReactDOMComponentTree to const/let
* Convert ReactDOMComponentTree to const/let
* Convert getNodeForCharacterOffset to const/let
* Convert getTextContentAccessor to const/let
* Convert inputValueTracking to const/let
* Convert setInnerHTML to const/let
* Convert setTextContent to const/let
* Convert validateDOMNesting to const/let
