Stacked on #32862 and #32842.
This means that Activity boundaries now act as boundaries which can have
their effects mounted independently. Just like Suspense boundaries, we
hydrate the outer content first and then start hydrating the content in
an Offscreen lane. Flowing props or interacting with the content
increases the priority just like Suspense boundaries.
This skips emitting even the comments for `<Activity mode="hidden">` so
we don't hydrate those. Instead those are deferred to a later client
render.
The implementation are just forked copies of the SuspenseComponent
branches and then carefully going through each line and tweaking it.
The main interesting bit is that, unlike Suspense, Activity boundaries
don't have fallbacks so all those branches where you might commit a
suspended tree disappears. Instead, if something suspends while
hydration, we can just leave the dehydrated content in place. However,
if something does suspend during client rendering then it should bubble
up to the parent. Therefore, we have to be careful to only
pushSuspenseHandler when hydrating. That's really the main difference.
This just uses the existing basic Activity tests but I've started work
on port all of the applicable Suspense tests in SelectiveHydration-test
and PartialHydration-test to Activity versions.
Stacked on #32851 and #32900.
This implements the equivalent Configs for ActivityInstance as we have
for SuspenseInstance. These can be implemented as comments but they
don't have to be and can be implemented differently in the renderer.
This seems like a lot duplication but it's actually ends mostly just
calling the same methods underneath and the wrappers compiles out.
This doesn't leave the Activity dehydrated yet. It just hydrates into it
immediately.
I think this was probably just copy-paste from the Suspense path.
It shouldn't matter what the previous state of an Offscreen boundary
was. What matters is that it's now hidden and therefore if it suspends,
we can just leave it as is without the tree becoming inconsistent.
Found this bug while working on Activity. There's a weird edge case when
a dehydrated Suspense boundary is a direct child of another Suspense
boundary which is hydrated but then it resuspends without forcing the
inner one to hydrate/delete.
It used to just leave that in place because hiding/unhiding didn't deal
with dehydrated fragments.
Not sure this is really worth fixing.
I found a bug even before the Activity hydration stuff.
If we're hydrating an Offscreen boundary in its "hidden" state it won't
have any content to hydrate so will trigger hydration errors (which are
then eaten by the Offscreen boundary itself). Leaving it not prewarmed.
This doesn't happen in the simple case because we'd be hydrating at a
higher priority than Offscreen at the root, and those are deferred to
Offscreen by not having higher priority. However, we've hydrating at the
Offscreen priority, which we do inside Suspense boundaries, then it
tries to hydrate against an empty set.
I ended up moving this to the Activity boundary in a future PR since
it's the SSR side that decided where to not render something and it only
has a concept of Activity, no Offscreen.
https://github.com/facebook/react/pull/32863/commits/1dc05a5e2222e18fc3a2062ee1bd957109e21344#diff-d5166797ebbc5b646a49e6a06a049330ca617985d7a6edf3ad1641b43fde1ddfR1111
Since `hidden` is a prop on arbitrary DOM elements it's a common mistake
to think that it would also work that way on `<Activity>` but it
doesn't. In fact, we even had this mistakes in our own tests.
Maybe there's an argument that we should actually just support it but we
also have more modes planned.
So this adds a warning. It should also already be covered by TypeScript.
It used to be that in `__DEV__` we wrapped this `renderWithHooks`,
`checkDidRenderIdHook` pair in calls to `setIsRendering()`. However,
that dev-only bookkeeping was removed in
https://github.com/facebook/react/pull/29206 leaving this redundant
check which runs identical code in dev and in prod.
## Test Plan
* Manually confirm both cases are the same
* GitHub CI tests
We use the Update flag to track if a View Transition had any mutations
or relayout. Unlike the other usage of it, this is just temporary state
during the commit phase.
Normally the flags gets used in the render phase and we reset it when we
rerender but in the case of "nested" updates, those trees didn't update.
We're only looking for relayouts. So we need to manually reset it before
we start using it.
We probably shouldn't abuse the Update flag for this and instead use
something like temporary state on ViewTransitionState.
Stacked on #32838.
We don't always type the Props of built-ins. This adds typing for most
of the built-ins.
When we did type them, we used to put it in the `ReactFiber...Component`
files but any public API like this can be implemented in other renderers
too such as Fizz. So I moved them to `shared/ReactTypes` which is where
we put other public API types (that are not already built-in to Flow).
That way Fizz can import them and assert properly when it accesses the
props.
Even if the `enableSuspenseyImages` flag is off.
Started View Transitions already wait for Suspensey Fonts and this is
another Suspensey feature that is even more important for View
Transitions - even though we eventually want it all the time. So this
uses `<ViewTransition>` as an early opt-in for that tree into Suspensey
Images, which we can ship in a minor.
If you're doing an update inside a ViewTransition then we're eligible to
start a ViewTransition in any Transition that might suspend. Even if
that doesn't end up animating after all, we still consider it Suspensey.
We could try to suspend inside the startViewTransition but that's not
how it would work with `enableSuspenseyImages` on and we can't do that
for startGestureTransition.
Even so we still need some opt-in to trigger the Suspense fallback even
before we know whether we'll animate or not. So the simple solution is
just that `<ViewTransition>` opts in the whole subtree into Suspensey
Images in general.
In this PR I disable `enableSuspenseyImages` in experimental so that we
can instead test the path that only enables it inside `<ViewTransition>`
tree since that's the path that would next graduate to a minor.
Behind the `enableSrcObject` flag. This is revisiting a variant of what
was discussed in #11163.
Instead of supporting the [`srcObject`
property](https://developer.mozilla.org/en-US/docs/Web/API/HTMLMediaElement/srcObject)
as a separate name, this adds an overload of `src` to allow objects to
be passed. The DOM needs to add separate properties for the object forms
since you read back but it doesn't make sense for React's write-only API
to do that. Similar to how we'll like add an overload for
`popoverTarget` instead of calling it `popoverTargetElement` and how
`style` accepts an object and it's not `styleObject={{...}}`.
There are a number of reason to revisit this.
- It's just way more convenient to have this built-in and it makes
conceptual sense. We typically support declarative APIs and polyfill
them when necessary.
- RSC supports Blobs and by having it built-in you don't need a Client
Component wrapper to render it where as doing it with effects would
require more complex wrappers. By picking Blobs over base64,
client-navigations can use the more optimized binary encoding in the RSC
protocol.
- The timing aspect of coordinating it with Suspensey images and image
decoding is a bit tricky to get right because if you set it in an effect
it's too late because you've already rendered it.
- SSR gets complicated when done in user space because you have to
handle both branches. Likely with `useSyncExternalStore`.
- By having it built-in we could optimize the payloads shared between
RSC payloads embedded in the HTML and data URLs.
This does not support objects for `<source src>` nor `<img srcset>`.
Those don't really have equivalents in the DOM neither. They're mainly
for picking an option when you don't know programmatically. However, for
this use case you're really better off picking a variant before
generating the blobs.
We may support Response objects in the future too as per
https://github.com/whatwg/fetch/issues/49
When `startTime` still has its initial value of `-1.1` we must not call
`logComponentMount`. This can occur when rendering a `'next/dynamic'`
component with `{ssr: false}` in a client component, for example.
Unfortunately, I didn't manage to reproduce this scenario in a unit
test.
Safari has a bug where if you put a block element inside an inline
element and the inline element has a `view-transition-name` assigned it
finds it as duplicate names.
https://bugs.webkit.org/show_bug.cgi?id=290923
This adds a warning if we detect this scenario in dev mode.
For the case where it renders into a single block, we can model this by
making the parent either `block` or `inline-block` automatically to fix
the issue. So we do that to automatically cover simple cases like
`<a><div>...</div></a>`. This unfortunately causes layout/styling thrash
so we might want to delete it once the bug has been fixed in enough
Safari versions.
Found a bug that occurs during a specific combination of very subtle
implementation details.
It occurs sometimes (not always) when 1) a transition is scheduled
during a popstate event, and 2) as a result, a new value is passed to an
already-mounted useDeferredValue hook.
The fix is relatively straightforward, and I found it almost
immediately; it took a bit longer to figure out exactly how the scenario
occurred in production and create a test case to simulate it.
Rather than couple the test to the implementation details, I've chosen
to keep it as high-level as possible so that it doesn't break if the
details change. In the future, it might not be trigger the exact set of
internal circumstances anymore, but it could be useful for catching
similar bugs because it represents a realistic real world situation —
namely, switching tabs repeatedly in an app that uses useDeferredValue.
We've known we've wanted this for many years and most of the
implementation was already done for Suspensey CSS. This waits to commit
until images have decoded by default or up to 500ms timeout (same as
suspensey fonts).
It only applies to Transitions, Retries (Suspense), Gesture Transitions
(flag) and Idle (doesn't exist). Sync updates just commit immediately.
`<img loading="lazy" src="..." />` opts out since you explicitly want it
to load lazily in that case.
`<img onLoad={...} src="..." />` also opts out since that implies you're
ok with managing your own reveal.
In the future, we may add an opt in e.g. `<img blocking="render"
src="..." />` that opts into longer timeouts and re-suspends even sync
updates. Perhaps also triggering error boundaries on errors.
The rollout for this would have to go in a major and we may have to
relax the default timeout to not delay too much by default. However, we
can also make this part of `enableViewTransition` so that if you opt-in
by using View Transitions then those animations will suspend on images.
That we could ship in a minor.
Stacked on #32815.
To be able to differentiate mounted subtrees from updated subtrees. This
adds a yellow entry above the component subtree that mounted. This is
added both to the render phase, mutation effect phase, layout effect
phase and passive effect phase.
<img width="962" alt="Screenshot 2025-04-03 at 10 41 02 PM"
src="https://github.com/user-attachments/assets/13777347-07e8-458c-9127-8675ef08b54f"
/>
Ideally we could probably give an annotation to the component instead of
adding a whole other line which is also a color that's kind of
distracting. However, not all components are included and keeping track
of which one is the first one below is kind of annoying. Adding a marker
to all components is kind of noisy. So this is a compromise. It's only
one per depth so it won't make it too deep even on larger trees.
If this is an unmount, those are added to the mutation effect phase for
the layout unmounts and passive unmount effect phase. Since these never
have a render, they're not in the render phase.
<img width="1010" alt="Screenshot 2025-04-03 at 11 05 57 PM"
src="https://github.com/user-attachments/assets/ab39f27e-13be-4281-94fa-9391bb293fd2"
/>
For showing / hiding `<Activity>` the terminology "Reconnect" and
"Disconnect" is used instead.
This fixes two bugs with commit phase effect tracking.
I missed, or messed up the rebase for, deletion effects when a subtree
was deleted and for passive disconnects when a subtree was hidden.
The other bug is that when I started using self time
(componentEffectDuration) for color and for determining whether to
bother logging an entry, I didn't consider that the component with
effects can have children which end up resetting this duration before we
log. Which lead to most effects not having their components logged since
they almost always have children.
We don't necessarily have to push/pop but we have to store at least one
thing on the stack unfortunately. That's because we have to do the
actual log after the children to get the right end time. So might as
well use the push/pop strategy like the rest of them.
Rename "Suspended" commit to "Suspended on CSS" since that's the only
reason for this particular branch. This will not hold true because with
suspended images and with view transitions those can also be the reason.
So in the future we need to add those.
Only log "Blocked" in the components track if we yield for 3ms or
longer. It's common to have like 1-2ms yield times for various reasons
going on which is not worth the noise to consider "blocking".
Rename "Blocked" to "Update" in the Blocking/Transition tracks. This is
when a setState happens and with stack traces it's where you should look
for the stack trace of the setState. So we want to indicate that this is
the "Update".
I only added the "Blocked" part if we're blocked for more than 5ms
before we can start rendering - indicating that some other track was
working at the same time and preventing us from rendering.
Stacked on #32793.
This is meant to model the intended semantics of `addTransitionType`
better. The previous hack just consumed all transition types when any
root committed so it could steal them from other roots. Really each root
should get its own set. Really each transition lane should get its own
set.
We can't implement the full ideal semantics yet because 1) we currently
entangle transition lanes 2) we lack `AsyncContext` on the client so for
async actions we can't associate a `addTransitionType` call to a
specific `startTransition`.
This starts by modeling Transition Types to be stored on the Transition
instance. Conceptually they belong to the Transition instance of that
`startTransition` they belong to. That instance is otherwise mostly just
used for Transition Tracing but it makes sense that those would be able
to be passed the Transition Types for that specific instance.
Nested `startTransition` need to get entangled. So that this
`addTransitionType` can be associated with the `setState`:
```js
startTransition(() => {
startTransition(() => {
addTransitionType(...)
});
setState(...);
});
```
Ideally we'd probably just use the same Transition instance itself since
these are conceptually all part of one entangled one. But transition
tracing uses multiple names and start times. Unclear what we want to do
with that. So I kept separate instances but shared `types` set.
Next I collect the types added during a `startTransition` to any root
scheduled with a Transition. This should really be collected one set per
Transition lane in a `LaneMap`. In fact, the information would already
be there if Transition Tracing was always enabled because it tracks all
Transition instances per lane. For now I just keep track of one set for
all Transition lanes. Maybe we should only add it if a `setState` was
done on this root in this particular `startTransition` call rather
having already scheduled any Transition earlier.
While async transitions are entangled, we don't know if there will be a
startTransition+setState on a new root in the future. Therefore, we
collect all transition types while this is happening and if a new root
gets startTransition+setState they get added to that root.
```js
startTransition(async () => {
addTransitionType(...)
await ...;
setState(...);
});
```
Stacked on #32792.
It's tricky to associate a specific `addTransitionType` call to a
specific `startTransition` call because we don't have `AsyncContext` in
browsers yet. However, we can keep track if there are any async
transitions running at all, and if not, warn. This should cover most
cases.
This also errors when inside a React render which might be a legit way
to associate a Transition Type to a specific render (e.g. based on props
changing) but we want to be a more conservative about allowing that yet.
If we wanted to support calling it in render, we might want to set which
Transition object is currently rendering but it's still tricky if the
render has `async function` components. So it might at least be
restricted to sync components (like Hooks).
Stacked on #32788.
Normally we track `addTransitionType` globally because of the async gap
that can happen in Actions where we lack AsyncContext to associate it
with a particular Transition. This unfortunately also means it's
possible to call outside of `startTransition` which is something we want
to warn for.
We need to be able to distinguish whether `addTransitionType` is for a
regular Transition or a Gesture Transition though.
Since `startGestureTransition` is only synchronous we can track it
within that execution scope and move it to a separate set. Since we know
for sure which call owns it we can properly associate it with that
specific provider's `ScheduledGesture`.
This does not yet handle calling `addTransitionType` inside the render
phase of a gesture. That would currently still be associated with the
next Transition instead.
Stacked on #32785.
This is now replaced by `startGestureTransition` added in #32785.
I also renamed the flag from `enableSwipeTransition` to
`enableGestureTransition` to correspond to the new name.
Stacked on #32783. This will replace [the `useSwipeTransition`
API](https://github.com/facebook/react/pull/32373).
Instead, of a special Hook, you can make updates to `useOptimistic`
Hooks within the `startGestureTransition` scope.
```
import {unstable_startGestureTransition as startGestureTransition} from 'react';
const cancel = startGestureTransition(timeline, () => {
setOptimistic(...);
}, options);
```
There are some downsides to this like you can't define two directions as
once and there's no "standard" direction protocol. It's instead up to
libraries to come up with their own conventions (although we can suggest
some).
The convention is still that a gesture recognizer has two props `action`
and `gesture`. The `gesture` prop is a Gesture concept which now behaves
more like an Action but 1) it can't be async 2) it shouldn't have
side-effects. For example you can't call `setState()` in it except on
`useOptimistic` since those can be reverted if needed. The `action` is
invoked with whatever side-effects you want after the gesture fulfills.
This is isomorphic and not associated with a specific renderer nor root
so it's a bit more complicated.
To implement this I unify with the `ReactSharedInternal.T` property to
contain a regular Transition or a Gesture Transition (the `gesture`
field). The benefit of this unification means that every time we
override this based on some scope like entering `flushSync` we also
override the `startGestureTransition` scope. We just have to be careful
when we read it to check the `gesture` field to know which one it is.
(E.g. I error for setState / requestFormReset.)
The other thing that's unique is the `cancel` return value to know when
to stop the gesture. That cancellation is no longer associated with any
particular Hook. It's more associated with the scope of the
`startGestureTransition`. Since the schedule of whether a particular
gesture has rendered or committed is associated with a root, we need to
somehow associate any scheduled gestures with a root.
We could track which roots we update inside the scope but instead, I
went with a model where I check all the roots and see if there's a
scheduled gesture matching the timeline. This means that you could
"retain" a gesture across roots. Meaning this wouldn't cancel until both
are cancelled:
```
const cancelA = startGestureTransition(timeline, () => {
setOptimisticOnRootA(...);
}, options);
const cancelB = startGestureTransition(timeline, () => {
setOptimisticOnRootB(...);
}, options);
```
It's more like it's a global transition than associated with the roots
that were updated.
Optimistic updates mostly just work but I now associate them with a
specific "ScheduledGesture" instance since we can only render one at a
time and so if it's not the current one, we leave it for later.
Clean up of optimistic updates is now lazy rather than when we cancel.
Allowing the cancel closure not to have to be associated with each
particular update.
This is some overdue refactoring. The two types never made sense. It
also should be defined by isomorphic since it defines how it should be
used by renderers rather than isomorphic depending on Fiber.
Clean up hidden classes to be consistent.
Fix missing name due to wrong types. I choose not to invoke the
transition tracing callbacks if there's no name since the name is
required there.
Portals and `<ViewTransition>` are tricky because they leave the React
tree. You might think of a Portal's container conceptually as also being
part of a React tree but that's not quite how they're modeled today.
They're more like their own roots. So instead, of trying to find a
conceptual place in the React tree we treat Portals as their own root.
We have two ways of tracking whether an update to a ViewTransition
boundary has occurred. Either a DOM mutation has happened within it, or
a resize of a child has caused it to potentially relayout its parent.
Normally that just follows the tree structure of React, but not when
it's a Portal.
When it's a Portal we don't know which DOM parent it might have
affected. For all we know it's at the root (and in fact, in most cases
that's where Portals go).
With this PR we mark the root as having been affected by a mutation or
resize. This means that the whole document will animate and we can't
optimize away from it. This ensures that a mutation to the root of a
Portal doesn't go unanimated with other things are animating such as its
parent.
You can regain this optimization by adding a `<ViewTransition>` boundary
directly inside the Portal itself so it owns its own animation. If that
DOM node is also absolutely positioned it doesn't leak.
Conversely this also means that a mutation inside a Portal doesn't
affect its React parent so it won't trigger its parent's animation if
this was the only thing animating. That could be unfortunate if this
container is actually inside the same React parent. However, because
this would have been an update we would've marked it for "maybe
animating" and updates can't only get their animations cancelled if the
root is cancelled, in practice this will actually animate anyway.
Typically we mark the name of things that might animate in the snapshot
phase. At the same time we track that should call startViewTransition
too. However, we don't do this for "enter" since they're only marked
later. Leading to having just an "enter" not to animate unless there's
at least another update too.
This tracks if there's a ViewTransitionComponent in the tree that
enters. Luckily we know that from the static flag so we don't have to
traverse it.
Stacked on https://github.com/facebook/react/pull/32734
In React a ViewTransition class of `"none"` doesn't just mean that it
has no class but also that it has no ViewTransition name. The default
(`null | undefined`) means that it has no specific class but should run
with the default built-in animation. This adds this as an explicit
string called `"auto"` as well.
That way you can do `<ViewTransition default="foo" enter="auto">` to
override the "foo" just for the "enter" trigger to be the default
built-in animation. Where as if you just specified `null` it would be
like not specifying enter at all which would trigger "foo".
It was always confusing that this is not a CSS class but a
view-transition-class.
The `className` sticks out a bit among its siblings `enter`, `exit`,
`update` and `share`. The idea is that the most specific definition
override is the class name that gets applied and this prop is really
just the fallback, catch-all or "any" that is applied if you didn't
specify a more specific one.
It has also since evolved not just to take a string but also a map of
Transition Type to strings.
The "class" is really the type of the value. We could add a suffix to
all of them like `defaultClass`, `enterClass`, `exitClass`,
`updateClass` and `shareClass`. However, this doesn't necessarily make
sense with the mapping of Transition Type to string. It also makes it a
bit too DOM centric. In React Native this might still be called a
"class" but it might be represented by an object definition. We might
even allow some kind of inline style form for the DOM too. Really this
is about picking which "animation" that runs which can be a string or
instance. "Animation" is too broad because there's also a concept of a
CSS Animation and these are really sets of CSS animations (group,
image-pair, old, new). It could maybe be `defaultTransition`,
`enterTransition`, etc but that seems unnecessarily repetitive and still
doesn't say anything about it being a class.
We also already have the name "default" in the map of Transition Types.
In fact you can now specify a default for default:
```
<ViewTransition default={{"navigation-back": "slide-out", "default": "fade-in"}}>
```
One thing I don't like about the name `"default"` is that it might be
common to just apply a named class that does it matching to
enter/exit/update in the CSS selectors (such as the `:only-child` rule)
instead of doing that mapping to each one using React. In that can you
end up specifying only `default={...}` a lot and then what is it the
"default" for? It's more like "all". I think it's likely that you end up
with either "default" or the specific forms instead of both at once.
Starting a View Transition is an async sequence. Since React can get a
sync update in the middle of sequence we sometimes interrupt that
sequence.
Currently, we don't actually cancel the View Transition so it can just
run as a partial. This ensures that we fully skip it when that happens,
as well as warn.
However, it's very easy to trigger this with just a setState in
useLayoutEffect right now. Therefore if we're inside the preparing
sequence of a startViewTransition, this delays work that would've
normally flushed in a microtask. ~Maybe we want to do the same for
Default work already scheduled through a scheduler Task.~ Edit: This was
already done.
`flushSync` currently will still lead to an interrupted View Transition
(with a warning). There's a tradeoff here whether we want to try our
best to preserve the guarantees of `flushSync` or favor the animation.
It's already possible to suspend at the root with `flushSync` which
means it's not always 100% guaranteed to commit anyway. We could treat
it as suspended. But let's see how much this is a problem in practice.
This adds early logging when two ViewTransitions with the same name are
mounted at the same time. Whether they're part of a View Transition or
not.
This lets us include the owner stack of each one. I do two logs so that
you can get the stack trace of each one of the duplicates.
It currently only logs once for each name which also avoids the scenario
when you have many hits for the same name in one commit. However, we
could also possibly log a stack for each of them but seems noisy.
Currently we don't log if a SwipeTransition is the first time the pair
gets mounted which could lead to a View Transition error before we've
warned. That could be a separate improvement.
This got moved into the functional component and class component case
statements here:
https://github.com/facebook/react/commit/0de1233fd180969f7ffdfc98151922f2466ceb1f.
So that we could separate the error case for class components.
However, due to a faulty rebase this got restored at the top as well.
Leading to double component renders being logged.
In the other offscreen reconnect passes we don't do this in each case
statement but still once at the top. The reason this doesn't matter is
because use the PerformedWork flag and that is only set for function and
class components. Although maybe it should be set for expensive DOM
components too and then we have to remember this.
We currently have the ability to have a separate animation for a
ViewTransition that relayouts but doesn't actually have any internal
mutations. This can be useful if you want to separate just a move from
for example flashing an update.
However, we're concerned that this might be more confusion than its
worth because subtle differences in mutations can cause it to trigger
the other case. The existence of the property name might also make you
start looking for it to solve something that it's not meant for.
We already fallback to using the "update" property if it exists but
layout doesn't. So if we ever decide to add this back it would backwards
compatible. We've also shown in implementation that it can work.
This implements `getRootNode(options)` on fragment instances as the
equivalent of calling `getRootNode` on the fragment's parent host node.
The parent host instance will also be used to proxy dispatchEvent in an
upcoming PR.
Followup from https://github.com/facebook/react/pull/32499
Manual mode is unused and has some bugs such as revealing hidden
boundaries when manually toggling. We also want to change how manual
mode works, and do some refactors to Activity to make it easier to
support. For now we'll remove it, then add it back after the other
changes we have planned.
`focus()` was added in https://github.com/facebook/react/pull/32465.
Here we add `focusLast()` and `blur()`. I also extended `focus` to take
options.
`focus` will focus the first focusable element. `focusLast` will focus
the last focusable element. We could consider a `focusFirst` naming or
even the `focusWithin` used by test selector APIs as well.
`blur` will only have an effect if the current `document.activeElement`
is one of the fragment children.
This does the same thing for `measureUpdateViewTransition` that we did
for `measureNestedViewTransitions` in
https://github.com/facebook/react/pull/32612/commits/e3cbaffef05c7b476c07f7495e06788a9503e636.
If a boundary hasn't mutated and didn't change in size, we mark it for
cancellation. Otherwise we add names to it. The different from the
CommitViewTransition path is that the "old" names are added to the
clones so this is the first time the "new" names.
Now we also cancel any boundaries that were unchanged. So now the root
no longer animates. We still have to clone them. There are other
optimizations that can avoid cloning but once we've done all the layouts
we can still cancel the running animation and let them just be the
regular content if they didn't change. Just like the regular
fire-and-forget path.
This also fixes the measurement so that we measure clones by adjusting
their position back into the viewport.
This actually surfaces a bug in Safari that was already in #32612. It
turns out that the old names aren't picked up for some reason and so in
Safari they looked more like a cross-fade than what #32612 was supposed
to fix. However, now that bug is even more apparent because they
actually just disappear in Safari. I'm not sure what that bug is but
it's unrelated to this PR so will fix that separately.
ViewTransition uses the `useId` algorithm to auto-assign names. This
ensures that we could animate between SSR content and client content by
ensuring that the names line up.
However, I missed that we need to bump the id (materialize it) when we
do that. This is what function components do if they use one or more
`useId()`. This caused duplicate names when two ViewTransitions were
nested without any siblings since they would share name.
This PR separates Activity to it's own element type separate from
Offscreen. The goal is to allow us to add Activity element boundary
semantics during hydration similar to Suspense semantics, without
impacting the Offscreen behavior in suspended children.
There's two ways to find updated View Transitions.
One is the "commit/measureNestedViewTransitions" pass which is used to
find things in unchanged subtrees. This can only lead to the relayout
case since there's can't possibly be any mutations in the subtree. This
is only triggered when none of the direct siblings have any mutations at
all.
The other case is "commit/measureUpdateViewTransition" which is for a
ViewTransition that itself has mutations scheduled inside of it which
leads to the "update" case.
However, there's a case between these two cases. When a direct sibling
has a mutation but there's also a ViewTransition exactly at the same
level. In that case we can't bail out on the whole set of children so we
won't trigger the "nested" case. Previously we also didn't trigger the
"commit/measureUpdateViewTransition" case because we first checked if
that had any mutations inside of it at all. This leads to neither case
picking up this boundary.
We could check if the ViewTransition itself has any mutations inside and
if not trigger the nested path.
There's a simpler way though. Because
`commit/measureUpdateViewTransition` is actually just optimistic. The
flags are pessimistic and we don't know for sure if there will actually
be a mutation until we've traversed the tree. It can sometimes lead to
the "relayout" case. So we can just use that same path, knowing that
it'll just lead to the layout pass. Therefore it's safe to just remove
this check.
This is a nit but a Config should not have to know anything about the
internals of Fibers. Ideally it shouldn't even access them but we have
some cases where we need pointers back in like for this fragment.
The way we've typically abstracted this is using the
`ReactFiberTreeReflection` helper that's in the `react-reconciler`. Such
as in the event system.
https://github.com/facebook/react/blob/f3c956006a90dc68210bd3e19497d10fb9b028d3/packages/react-dom-bindings/src/events/ReactDOMEventListener.js#L22-L26
We sometimes cheat but we really should clean this up such that a
`Fiber` is actually an opaque type to the Configs and it can never dot
into it without using a helper.
So this just moves `traverseFragmentInstanceChildren` to
ReactFiberTreeReflection so that the ConfigDOM doesn't ever dot into its
fields itself. It just passes the Fiber through back into the
react-reconciler. I had to add a wrapper to read the `.child` to avoid
that being assumed too. I also noticed that FragmentInstanceType is not
actually passed through so that argument is unnecessary.
Sebastian Markbåge
Jordan Eldredge
Hendrik Liebau
Andrew Clark
lauren
Ricky
Jack Pope
Sebastian "Sebbie" SilbermannandJack Pope