* Add Visibility flag for hiding/unhiding trees
There's `beforeblur` logic in the snapshot phase that needs to visit
every Suspense boundary whose visibility is toggled. Right now it does
that by visiting Placement and Deletion effects. That includes many
unrelated nodes.
By adding a new flag specifically for toggling Visibility, we will only
visit the relevant Suspense (and Offscreen) boundaries, instead of all
nodes that have a Placement.
Potential follow-ups (not urgent):
- The `beforeblur` logic also has a check to see whether the visibility
was toggled on or off. It only cares about things being hidden. As a
follow up, I can split the Visibility flag into separate Hide/Show
flags, and only visit Hide.
- Now that this is separate from Update, we can move the rest of the
Suspense's layout effects (like attaching retry listeners) to the
passive phase.
* Gate behind createEventHandle feature flag
Only need to visit deleted and hidden trees during the snapshot phase
if the experimental `createEventHandle` flag is enabled. Currently,
it's only used internally at Facebook, not open source.
* Remove dead code branch
This function is only called when initializing roots/containers (where we skip non-delegated events) and in the createEventHandle path for non-DOM nodes (where we never hit this path because targetElement is null).
* Move related functions close to each other
* Fork listenToNativeEvent for createEventHandle
It doesn't need all of the logic that's needed for normal event path.
And the normal codepath doesn't use the last two arguments.
* Expand test coverage for non-delegated events
This changes a test to fail if we removed the event handler Sets. Previously, we didn't cover that.
* Add DEV-level check that top-level events and non-delegated events do not overlap
This makes us confident that they're mutually exclusive and there is no duplication between them.
* Add a test verifying selectionchange deduplication
This is why we still need the Set bookkeeping. Adding a test for it.
* Remove Set bookkeeping for root events
Root events don't intersect with non-delegated bubbled events (so no need to deduplicate there). They also don't intersect with createEventHandle non-managed events (because those don't go on the DOM elements). So we can remove the bookeeping because we already have code ensuring the eager subscriptions only run once per element.
I've moved the selectionchange special case outside, and added document-level deduplication for it alone.
Technically this might change the behavior of createEventHandle with selectionchange on the document, but we're not using that, and I'm not sure that behavior makes sense anyway.
* Flow
* bump package to latest
* update files to respect lint
* disable object-type-delimiter rule to work with prettier
* disable rule to let flow check pass
We don't need to visit passive effect nodes during before mutation.
The only reason we were previously was to schedule the root-level
passive effect callback as early as possible, but now that
`subtreeFlags` exists, we can check that instead.
This should reduce the amount of traversal during the commit phase,
particularly when mounting or updating large trees that contain many
passive effects.
Large legacy applications are likely to be difficult to update to handle this feature, and it wouldn't add any value– since newer APIs that require this resilience are not legacy compatible.
This is done so that any effects scheduled by the shallow render are thrown away.
Unlike the code this was forked from (in ReactComponentStackFrame) DevTools should override the dispatcher even when DevTools is compiled in production mode, because the app itself may be in development mode and log errors/warnings.
* Deprecate old test script commands
* Update PR template test script
* Add test-stable and test-www-classic
* Update circle test names
* Rename test-www-classic to test-classic
* Missed some job renames
* Missed some more job renames
* update all facebook.github.io links
* facebookincubator links : update some outdated links and fix two other broken links where they are actually the latest updated ones
* Improve error message by expanding the object in question
* Don't warn for key/ref getters
* Error if refs are passed in server components or to client components
<time> tag has been supported by Chrome since Chrome 62.0.
Remove workarounds which were in place to avoid friction with
versions before Chrome 62.
Signed-off-by: Shivam Sandbhor <shivam.sandbhor@gmail.com>
Technically this change is unnecessary, since the feature is controlled by a flag, but since we decided not to ship this in v17– I'm going to remove it for now entirely.
When a link opens a URL in a new tab with target="_blank", it is very simple for the opened page to change the location of the original page because the JavaScript variable window.opener is not null and thus "window.opener.location can be set by the opened page. This exposes the user to very simple phishing attacks.
Adds a new prop to the Suspense component type,
`unstable_expectedLoadTime`. The presence of this prop indicates that
the content is computationally expensive to render.
During the initial mount, React will skip over expensive trees by
rendering a placeholder — just like we do with trees that are waiting
for data to resolve. That will help unblock the initial skeleton for the
new screen. Then we will continue rendering in the next commit.
For now, while we experiment with the API internally, any number passed
to `unstable_expectedLoadTime` will be treated as "computationally
expensive", no matter how large or small. So it's basically a boolean.
The reason it's a number is that, in the future, we may try to be clever
with this additional information. For example, SuspenseList could use
it as part of its heuristic to determine whether to keep rendering
additional rows.
Background
----------
Much of our early messaging and research into Suspense focused on its
ability to throttle the appearance of placeholder UIs. Our theory was
that, on a fast network, if everything loads quickly, excessive
placeholders will contribute to a janky user experience. This was backed
up by user research and has held up in practice.
However, our original demos made an even stronger assertion: not only is
it preferable to throttle successive loading states, but up to a certain
threshold, it’s also preferable to remain on the previous screen; or in
other words, to delay the transition.
This strategy has produced mixed results. We’ve found it works well for
certain transitions, but not for all them. When performing a full page
transition, showing an initial skeleton as soon as possible is crucial
to making the transition feel snappy. You still want throttle the nested
loading states as they pop in, but you need to show something on the new
route. Remaining on the previous screen can make the app feel
unresponsive.
That’s not to say that delaying the previous screen always leads to a
bad user experience. Especially if you can guarantee that the delay is
small enough that the user won’t notice it. This threshold is a called a
Just Noticeable Difference (JND). If we can stay under the JND, then
it’s worth skipping the first placeholder to reduce overall thrash.
Delays that are larger than the JND have some use cases, too. The main
one we’ve found is to refresh existing data, where it’s often preferable
to keep stale content on screen while the new data loads in the
background. It’s also useful as a fallback strategy if something
suspends unexpectedly, to avoid hiding parts of the UI that are already
visible.
We’re still in the process of optimizing our heuristics for the most
common patterns. In general, though, we are trending toward being more
aggressive about prioritizing the initial skeleton.
For example, Suspense is usually thought of as a feature for displaying
placeholders when the UI is missing data — that is, when rendering is
bound by pending IO.
But it turns out that the same principles apply to CPU-bound
transitions, too. It’s worth deferring a tree that’s slow to render if
doing so unblocks the rest of the transition — regardless of whether
it’s slow because of missing data or because of expensive CPU work.
We already take advantage of this idea in a few places, such as
hydration. Instead of hydrating server-rendered UI in a single pass,
React splits it into chunks. It can do this because the initial HTML
acts as its own placeholder. React can defer hydrating a chunk of UI as
long as it wants until the user interacts it. The boundary we use to
split the UI into chunks is the same one we use for IO-bound subtrees:
the <Suspense /> component.
SuspenseList does something similar. When streaming in a list of items,
it will occasionally stop to commit whatever items have already
finished, before continuing where it left off. It does this by showing a
placeholder for the remaining items, again using the same <Suspense />
component API, even if the item is CPU-bound.
Unresolved questions
--------------------
There is a concern that showing a placeholder without also loading new
data could be disorienting. Users are trained to believe that a
placeholder signals fresh content. So there are still some questions
we’ll need to resolve.
Commit phase durations (layout and passive) are stored on the nearest (ancestor) Profiler and bubble up during the commit phase. This bubbling used to be implemented by traversing the return path each time we finished working on a Profiler to find the next nearest Profiler.
This commit removes that traversal. Instead, we maintain a stack of nearest Profiler ancestor while recursing the tree. This stack is maintained in the work loop (since that's where the recursive functions are) and so bubbling of durations has also been moved from commit-work to the work loop.
This PR also refactors the methods used to recurse and apply effects in preparation for the new Offscreen component type.
This PR double invokes effects in __DEV__ mode.
We are thinking about unmounting layout and/or passive effects for a hidden tree. To understand potential issues with this, we want to double invoke effects. This PR changes the behavior in DEV when an effect runs from create() to create() -> destroy() -> create(). The effect cleanup function will still be called before the effect runs in both dev and prod. (Note: This change is purely for research for now as it is likely to break real code.)
**Note: The change is fully behind a flag and does not affect any of the code on npm.**
* Don't call onCommit et al if there are no effects
Checks `subtreeFlags` before scheduling an effect on the Profiler.
* Fix failing Profiler tests
The change to conditionally call Profiler commit hooks only if updates were scheduled broke a few of the Profiler tests. I've fixed the tests by either:
* Adding a no-op passive effect into the subtree or
* Converting onPostCommit to onCommit
When possible, I opted to add the no-op passive effect to the tests since that that hook is called later (during passive phase) so the test is a little broader. In a few cases, this required adding awkward act() wrappers so I opted to go with onCommit instead.
Co-authored-by: Brian Vaughn <bvaughn@fb.com>
There were a few pairs of commit phase functions that were almost
identical except for one detail. I've refactored them a bit to
consolidate their implementations:
- Lifted error handling logic when mounting a fiber's passive hook
effects to surround the entire list, instead of surrounding each effect.
- Lifted profiler duration tracking to surround the entire list.
In both cases, this matches the corresponding code for the layout phase.
The naming is still a bit of a mess but I'm not too concerned because
my next step is to refactor each commit sub-phase (layout, mutation)
so that we can store values on the JS stack. So the existing function
boundaries are about to change, anyway.
Andrew Clark
Dan Abramov
Sebastian Silbermann
oltrep
IDrissAitHafid
adasq
Brian Vaughn
Dominic Gannaway
Simen Bekkhus
Saikat Guha
Ricky
Nick Reiley
Sebastian Markbåge
Shivam Sandbhor
Paul Doyle
Eugene Maslovich
John Wilson
Dmitriy Kovalenko
Luna Ruan
6h057
Todor Totev