* Add Batched Mode
React has an unfortunate quirk where updates are sometimes synchronous
-- where React starts rendering immediately within the call stack of
`setState` — and sometimes batched, where updates are flushed at the
end of the current event. Any update that originates within the call
stack of the React event system is batched. This encompasses most
updates, since most updates originate from an event handler like
`onClick` or `onChange`. It also includes updates triggered by lifecycle
methods or effects. But there are also updates that originate outside
React's event system, like timer events, network events, and microtasks
(promise resolution handlers). These are not batched, which results in
both worse performance (multiple render passes instead of single one)
and confusing semantics.
Ideally all updates would be batched by default. Unfortunately, it's
easy for components to accidentally rely on this behavior, so changing
it could break existing apps in subtle ways.
One way to move to a batched-by-default model is to opt into Concurrent
Mode (still experimental). But Concurrent Mode introduces additional
semantic changes that apps may not be ready to adopt.
This commit introduces an additional mode called Batched Mode. Batched
Mode enables a batched-by-default model that defers all updates to the
next React event. Once it begins rendering, React will not yield to
the browser until the entire render is finished.
Batched Mode is superset of Strict Mode. It fires all the same warnings.
It also drops the forked Suspense behavior used by Legacy Mode, in favor
of the proper semantics used by Concurrent Mode.
I have not added any public APIs that expose the new mode yet. I'll do
that in subsequent commits.
* Suspense in Batched Mode
Should have same semantics as Concurrent Mode.
* Use RootTag field to configure type of root
There are three types of roots: Legacy, Batched, and Concurrent.
* flushSync should not flush batched work
Treat Sync and Batched expiration times separately. Only Sync updates
are pushed to our internal queue of synchronous callbacks.
Renamed `flushImmediateQueue` to `flushSyncCallbackQueue` for clarity.
* Rewrite ReactFiberScheduler
Adds a new implementation of ReactFiberScheduler behind a feature flag.
We will maintain both implementations in parallel until the new one
is proven stable enough to replace the old one.
The main difference between the implementations is that the new one is
integrated with the Scheduler package's priority levels.
* Conditionally add fields to FiberRoot
Some fields only used by the old scheduler, and some by the new.
* Add separate build that enables new scheduler
* Re-enable skipped test
If synchronous updates are scheduled by a passive effect, that work
should be flushed synchronously, even if flushPassiveEffects is
called inside batchedUpdates.
* Passive effects have same priority as render
* Revert ability to cancel the current callback
React doesn't need this anyway because it never schedules callbacks if
it's already rendering.
* Revert change to FiberDebugPerf
Turns out this isn't neccessary.
* Fix ReactFiberScheduler dead code elimination
Should initialize to nothing, then assign the exports conditionally,
instead of initializing to the old exports and then reassigning to the
new ones.
* Don't yield before commit during sync error retry
* Call Scheduler.flushAll unconditionally in tests
Instead of wrapping in enableNewScheduler flag.
This took a while, but I'm happy I went through it. Some key moments - recursively flushing effects, flushing microtasks on each async turn, and my team's uncompromising philosophy on code reuse. Really happy with this. I still want to expand test coverage, and I have some more small related todos, but this is good to land. On to the next one.
Soundtrack to landing this - https://open.spotify.com/track/0MF8I8OUo8kytiOo8aSHYq?si=gSWqUheKQbiQDXzptCXHTg
* hacked up act(async () => {...})
* move stuff around
* merge changes
* abstract .act warnings and stuff. all renderers. pass all tests.
* move testutils.act back into testutils
* move into scheduler, rename some bits
* smaller bundle
* a comment for why we don't do typeof === 'function'
* fix test
* pass tests - fire, prod
* lose actContainerElement
* tighter
* write a test for TestRenderer
it's an odd one, because not only does sync act not flush effects correctly, but the async one does (wut). verified it's fine with the dom version.
* lint
* rewrote to move flushing logic closer to the renderer
the scheduler's `flushPassiveEffects` didn't work as expected for the test renderer, so I decided to go back to the hack (rendering a dumb container) This also makes reactdom not as heavy (by a few bytes, but still).
* move it around so the delta isn't too bad
* cleanups
fix promise chaining
propagate errors correctly
test for thenable the 'right' way
more tests!
tidier!
ponies!
* Stray comment
* recursively flush effects
* fixed tests
* lint, move noop.act into react-reconciler
* microtasks when checking if called, s/called/calledLog, cleanup
* pass fb lint
we could have globally changed our eslint config to assume Promise is available, but that means we expect a promise polyfill on the page, and we don't yet. this code is triggered only in jest anyway, and we're fairly certain Promise will be available there. hence, the once-off disable for the check
* shorter timers, fix a test, test for Promise
* use global.Promise for existence check
* flush microtasks
* a version that works in browsers (that support postMessage)
I also added a sanity fixture inside fixtures/dom/ mostly for me.
* hoist flushEffectsAndMicroTasks
* pull out tick logic from ReactFiberScheduler
* fix await act (...sync) hanging
- fix a hang when awaiting sync logic
- a better async/await test for test renderer
* feedback changes
- use node's setImmediate if available
- a warning if MessageChannel isn't available
- rename some functions
* pass lint/flow checks (without requiring a Promise polyfill/exclusion)
* prettier
the prettiest, even.
* use globalPromise for the missed await warning
* __DEV__ check for didWarnAboutMessageChannel
* thenables and callbacks instead of promises, pass flow/lint
* tinier. better.
- pulled most bits out of FiberScheduler
- actedUpdates uses callbacks now
* pass build validation
* augh prettier
* golfing 7 more chars
* Test that effects are not flushed without also flushing microtasks
* export doesHavePendingPassiveEffects, nits
* createAct()
* dead code
* missed in merge?
* lose the preflushing bits
* ugh prettier
* removed `actedUpdates()`, created shared/actingUpdatesScopeDepth
* rearrange imports so builds work, remove the hack versions of flushPassiveEffects
* represent actingUpdatesScopeDepth as a tuple [number]
* use a shared flag on React.__SECRET...
* remove createAct, setup act for all relevant renderers
* review feedback
shared/enqueueTask
import ReactSharedInternals from 'shared/ReactSharedInternals';
simpler act() internals
ReactSharedInternals.ReactShouldWarnActingUpdates
* move act() implementation into createReactNoop
* warnIfNotCurrentlyActingUpdatesInDev condition check order
* Add more info to invalid hook call error message
* Update other renderers + change call to action
* Update related tests for new hooks error message
* Fix lint errors
* Throw away old shallow renderer state on type change
This worked in function components but was broken for classes. It incorrectly retained the old instance even if the type was different.
* Remove _previousComponentIdentity
We only needed this because we didn't correctly reset based on type. Now we do so this can go away.
* Use _reset when unmounting
* Use arbitrary componentIdentity
There was no particular reason it was set to element.type. We just wanted to check if something is a render phase update.
* Support Hook state updates in shallow renderer
* Support React.memo in ReactShallowRenderer
ReactShallowRenderer uses element.type frequently, but with React.memo
elements the actual type is element.type.type. This updates
ReactShallowRenderer so it uses the correct element type for Memo
components and also validates the inner props for the wrapped
components.
* Allow Rect.memo to prevent re-renders
* Support memo(forwardRef())
* Dont call memo comparison function on initial render
* Fix test
* Small tweaks
* Import Scheduler directly, not via host config
We currently schedule asynchronous tasks via the host config. (The host
config is a static/build-time dependency injection system that varies
across different renderers — DOM, native, test, and so on.) Instead of
calling platform APIs like `requestIdleCallback` directly, each renderer
implements a method called `scheduleDeferredCallback`.
We've since discovered that when scheduling tasks, it's crucial that
React work is placed in the same queue as other, non-React work on the
main thread. Otherwise, you easily end up in a starvation scenario where
rendering is constantly interrupted by less important tasks. You need a
centralized coordinator that is used both by React and by other
frameworks and application code. This coordinator must also have a
consistent API across all the different host environments, for
convention's sake and so product code is portable — e.g. so the same
component can work in both React Native and React Native Web.
This turned into the Scheduler package. We will have different builds of
Scheduler for each of our target platforms. With this approach, we treat
Scheduler like a built-in platform primitive that exists wherever React
is supported.
Now that we have this consistent interface, the indirection of the host
config no longer makes sense for the purpose of scheduling tasks. In
fact, we explicitly do not want renderers to scheduled task via any
system except the Scheduler package.
So, this PR removes `scheduleDeferredCallback` and its associated
methods from the host config in favor of directly importing Scheduler.
* Missed an extraneous export
* Swap expect(ReactNoop) for expect(Scheduler)
In the previous commits, I upgraded our custom Jest matchers for the
noop and test renderers to use Scheduler under the hood.
Now that all these matchers are using Scheduler, we can drop
support for passing ReactNoop and test roots and always pass
Scheduler directly.
* Externalize Scheduler in noop and test bundles
I also noticed we don't need to regenerator runtime in noop anymore.
* Replace test renderer's fake Scheduler implementation with mock build
The test renderer has its own mock implementation of the Scheduler
interface, with the ability to partially render work in tests. Now that
this functionality has been lifted into a proper mock Scheduler build,
we can use that instead.
* Fix Profiler tests in prod
* Throw in tests if work is done before emptying log
Test renderer already does this. Makes it harder to miss unexpected
behavior by forcing you to assert on every logged value.
* Convert ReactNoop tests to use jest matchers
The matchers warn if work is flushed while the log is empty. This is
the pattern we already follow for test renderer. I've used the same APIs
as test renderer, so it should be easy to switch between the two.
* expose unstable_interact for batching actions in tests
* move to TestUtils
* move it all into testutils
* s/interact/act
* warn when calling hook-like setState outside batching mode
* pass tests
* merge-temp
* move jsdom test to callsite
* mark failing tests
* pass most tests (except one)
* augh IE
* pass fuzz tests
* better warning, expose the right batchedUpdates on TestRenderer for www
* move it into hooks, test for dom
* expose a flag on the host config, move stuff around
* rename, pass flow
* pass flow... again
* tweak .act() type
* enable for all jest environments/renderers; pass (most) tests.
* pass all tests
* expose just the warning from the scheduler
* don't return values
* a bunch of changes.
can't return values from .act
don't try to await .act calls
pass tests
* fixes and nits
* "fire events that udpates state"
* nit
* 🙄
* my bad
* hi andrew
(prettier fix)
* Avoid importing Scheduler directly
The reconciler should not depend directly on Scheduler. This adds it to
the host config for the renderer instead.
(Except for `scheduler/tracing` imports, which are used only by the
profiling build. I've left those imports as-is, though I'm open to
directing those through the host config, too.)
* Make throwaway root id longer to appease Brian
* Improve Reducer Hook's lazy init API
* Use generic type for initilizer input
Still requires an `any` cast in the case where `init` function is
not provided.
* Move DEV-only function right above where it's used
I don't like looking at this top-level function #petty
* Use different dispatchers for functions & classes
Classes support readContext, but not any of the other dispatcher
methods. Function support all methods.
This is a more robust version of our previous strategy of checking
whether `currentlyRenderingFiber` is null.
As a next step, we can use a separate dispatcher for each phase of the
render cycle (mount versus update).
* Use separate dispatchers for mount and update
* Remove mount code from update path
Deletes mount-specific code from the update path, since it should be
unreachable. To continue supporting progressive enhancement (mounting
new hooks at the end of the list), we detect when there are no more
current hooks and switch back to the mount dispatcher. Progressive
enhancement isn't officially supported yet, so it will continue to warn.
* Factoring nits
* Fix Flow
Had to cheat more than I would like
* More Flow nits
* Switch back to using a special dispatcher for nested hooks in DEV
In order for this strategy to work, I had to revert progressive
enhancement support (appending hooks to the end). It was previously a
warning but now it results in an error. We'll reconsider later.
* Always pass args to updateState and updateReducer
Even though the extra args are only used on mount, to ensure
type consistency.
* Turned enableHooks feature flag on everywhere
* Removed useHooks feature flag from tests (now that it's on by default)
* Remove useHooks feature flag entirely
Whenever we do this, Rollup needs to materialize this as an object.
This causes it to also add the Babel compatibility property which is
unnecessary bloat. However, since when we use these, we leak the object
this often also deopts any compiler optimizations.
If we really need an object we should export default an object.
Currently there is an exception for DOMTopLevelEventTypes since
listing out the imports is a PITA and it doesn't escape so it should
get properly inlined. We should probably move to a different pattern
to avoid this for consistency though.
* Add debug tools package
* Add basic implementation
* Implement inspection of the current state of hooks using the fiber tree
* Support useContext hooks inspection by backtracking from the Fiber
I'm not sure this is safe because the return fibers may not be current
but close enough and it's fast.
We use this to set up the current values of the providers.
* rm copypasta
* Use lastIndexOf
Just in case. I don't know of any scenario where this can happen.
* Support ForwardRef
* Add test for memo and custom hooks
* Support defaultProps resolution
* [scheduler] Deadline object -> shouldYield
Instead of using a requestIdleCallback-style deadline object, expose a
method Scheduler.shouldYield that returns true if there's a higher
priority event in the queue.
* Nits
* [Synchronous Suspense] Suspending a class outside concurrent mode
When a class component suspends during mount outside concurrent mode,
change the tag so it's not mistaken for a completed component. For
example, we should not call componentWillUnmount if it is deleted.
* PR nits
* Allow arbitrary types to be wrapped in pure
This creates an outer fiber that container the pure check and an inner
fiber that represents which ever type of component.
* Add optimized fast path for simple pure function components
Special cased when there are no defaultProps and it's a simple function
component instead of class. This doesn't require an extra fiber.
We could make it so that this also works with custom comparer but that
means we have to go through one extra indirection to get to it.
Maybe it's worth it, donno.
* Introduce elementType field
This will be used to store the wrapped type of an element. E.g. pure and
lazy.
The existing type field will be used for the unwrapped type within them.
* Store the unwrapped type on the type field of lazy components
* Use the raw tags for lazy components
Instead, we check if the elementType and type are equal to test if
we need to resolve props. This is slightly slower in the normal case
but will yield less code and branching.
* Clean up lazy branches
* Collapse work tag numbering
* Split IndeterminateComponent out from Lazy
This way we don't have to check the type in a hacky way in the
indeterminate path. Also, lets us deal with lazy that resolves to
indeterminate and such.
* Missing clean up in rebase
* Store the start time on `updateQueue` instead of `stateNode`
Originally I did this to free the `stateNode` field to store a second
set of children. I don't we'll need this anymore, since we use fragment
fibers instead. But I still think using `updateQueue` makes more sense
so I'll leave this in.
* Use fragment fibers to keep the primary and fallback children separate
If the children timeout, we switch to showing the fallback children in
place of the "primary" children. However, we don't want to delete the
primary children because then their state will be lost (both the React
state and the host state, e.g. uncontrolled form inputs). Instead we
keep them mounted and hide them. Both the fallback children AND the
primary children are rendered at the same time. Once the primary
children are un-suspended, we can delete the fallback children — don't
need to preserve their state.
The two sets of children are siblings in the host environment, but
semantically, for purposes of reconciliation, they are two separate
sets. So we store them using two fragment fibers.
However, we want to avoid allocating extra fibers for every placeholder.
They're only necessary when the children time out, because that's the
only time when both sets are mounted.
So, the extra fragment fibers are only used if the children time out.
Otherwise, we render the primary children directly. This requires some
custom reconciliation logic to preserve the state of the primary
children. It's essentially a very basic form of re-parenting.
* Use `memoizedState` to store various pieces of SuspenseComponent's state
SuspenseComponent has three pieces of state:
- alreadyCaptured: Whether a component in the child subtree already
suspended. If true, subsequent suspends should bubble up to the
next boundary.
- didTimeout: Whether the boundary renders the primary or fallback
children. This is separate from `alreadyCaptured` because outside of
strict mode, when a boundary times out, the first commit renders the
primary children in an incomplete state, then performs a second commit
to switch the fallback. In that first commit, `alreadyCaptured` is
false and `didTimeout` is true.
- timedOutAt: The time at which the boundary timed out. This is separate
from `didTimeout` because it's not set unless the boundary
actually commits.
These were previously spread across several fields.
This happens to make the non-strict case a bit less hacky; the logic for
that special case is now mostly localized to the UnwindWork module.
* Hide timed-out Suspense children
When a subtree takes too long to load, we swap its contents out for
a fallback to unblock the rest of the tree. Because we don't want
to lose the state of the timed out view, we shouldn't actually delete
the nodes from the tree. Instead, we'll keep them mounted and hide
them visually. When the subtree is unblocked, we un-hide it, having
preserved the existing state.
Adds additional host config methods. For mutation mode:
- hideInstance
- hideTextInstance
- unhideInstance
- unhideTextInstance
For persistent mode:
- cloneHiddenInstance
- cloneUnhiddenInstance
- createHiddenTextInstance
I've only implemented the new methods in the noop and test renderers.
I'll implement them in the other renderers in subsequent commits.
* Include `hidden` prop in noop renderer's output
This will be used in subsequent commits to test that timed-out children
are properly hidden.
Also adds getChildrenAsJSX() method as an alternative to using
getChildren(). (Ideally all our tests would use test renderer #oneday.)
* Implement hide/unhide host config methods for DOM renderer
For DOM nodes, we hide using `el.style.display = 'none'`.
Text nodes don't have style, so we hide using `text.textContent = ''`.
* Implement hide/unhide host config methods for Art renderer
* Create DOM fixture that tests state preservation of timed out content
* Account for class components that suspend outside concurrent mode
Need to distinguish mount from update. An unfortunate edge case :(
* Fork appendAllChildren between persistent and mutation mode
* Remove redundant check for existence of el.style
* Schedule placement effect on indeterminate components
In non-concurrent mode, indeterminate fibers may commit in an
inconsistent state. But when they update, we should throw out the
old fiber and start fresh. Which means the new fiber needs a
placement effect.
* Pass null instead of current everywhere in mountIndeterminateComponent
* Jest + test renderer helpers for concurrent mode
Most of our concurrent React tests use the noop renderer. But most
of those tests don't test the renderer API, and could instead be
written with the test renderer. We should switch to using the test
renderer whenever possible, because that's what we expect product devs
and library authors to do. If test renderer is sufficient for writing
most React core tests, it should be sufficient for others, too. (The
converse isn't true but we should aim to dogfood test renderer as much
as possible.)
This PR adds a new package, jest-react (thanks @cpojer). I've moved
our existing Jest matchers into that package and added some new ones.
I'm not expecting to figure out the final API in this PR. My goal is
to land something good enough that we can start dogfooding in www.
TODO: Continue migrating Suspense tests, decide on better API names
* Add additional invariants to prevent common errors
- Errors if user attempts to flush when log of yields is not empty
- Throws if argument passed to toClearYields is not ReactTestRenderer
* Better method names
- toFlushAll -> toFlushAndYield
- toFlushAndYieldThrough ->
- toClearYields -> toHaveYielded
Also added toFlushWithoutYielding
* Fix jest-react exports
* Tweak README
* pure
A higher-order component version of the `React.PureComponent` class.
During an update, the previous props are compared to the new props. If
they are the same, React will skip rendering the component and
its children.
Unlike userspace implementations, `pure` will not add an additional
fiber to the tree.
The first argument must be a functional component; it does not work
with classes.
`pure` uses shallow comparison by default, like `React.PureComponent`.
A custom comparison can be passed as the second argument.
Co-authored-by: Andrew Clark <acdlite@fb.com>
Co-authored-by: Sophie Alpert <sophiebits@fb.com>
* Warn if first argument is not a functional component
