Skip to content

[pull] main from react:main - #669

Merged
pull[bot] merged 2 commits into
code:mainfrom
react:main
Sep 13, 2026
Merged

pull[bot] merged 2 commits into
code:mainfrom
react:main

Conversation

@pull

@pull pull Bot commented Sep 13, 2026 •

Copy link
Copy Markdown

See Commits and Changes for more details.


Created by pull[bot] (v2.0.0-alpha.4)

Can you help keep this open source service alive? 💖 Please sponsor : )

connorshea and others added 2 commits September 13, 2026 10:51
## Summary

AI Disclosure: This was assisted using Claude Code, Fable 5.1. This PR
description was fully written by me, manually, and the code was reviewed
and tested by me manually.

This improves the accuracy of the animation-related event props
supported by React.

This adds `onAnimationCancel` alongside the existing `onAnimationStart`,
`onAnimationIteration`, and `onAnimationEnd` props.

Adds `onAnimationCancel` to React DOM, completing the animation event
family alongside `onAnimationStart`, `onAnimationIteration`, and
`onAnimationEnd`.
[`animationcancel`](https://developer.mozilla.org/en-US/docs/Web/API/Element/animationcancel_event)
is a CSS Animations Level 1 event fired when an animation stops before
it has completed.

This changeset is based on the similar changes in #27345 for the
transition events.

Note that MDN's browser compat data [currently
claims](https://developer.mozilla.org/en-US/docs/Web/API/Element/animationcancel_event)
that this is not supported in Chrome or Edge (at least not as
`onanimationcancel`), but this change will work in Chrome 83+ due to
React passing things through `addEventListener` anyway, and also the MDN
docs are [outdated about
this](mdn/browser-compat-data#29376), Chrome
has started supporting `onanimationcancel` directly [since earlier this
year](https://issues.chromium.org/issues/41404325).

Also note that the animation and transition events on DOMEventName can
probably be made to no longer do this vendor prefixing dance? I didn't
do that in this PR and it should be done separately, but they've not
been prefixed in a long time now and the comment rationalizing the
current code mentions Android 4.x, which was last released over a decade
ago now. So it seems like something that has just never been cleaned up.

## How did you test this change?

- Added tests to match the sibling properties' tests.
- Ran `yarn test`, `yarn lint`, and `yarn prettier` and all passed.

Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…me names (#37608)

## Summary

V8 prefixes async call sites with `async ` when it prints a stack, like
this:

```
Error: boom
    at inner (/tmp/asy.js:1:44)
    at async outerName (/tmp/asy.js:2:30)
```

`parseStackTraceFromChromeStack` captures `async outerName` as the frame
name and strips the prefix here:

```js
} else if (name.startsWith('async ')) {
  name = name.slice(5);
  isAsync = true;
}
```

`'async '` is six characters, so `slice(5)` leaves the space behind and
the frame name comes back as `' outerName'` rather than `'outerName'`.

I noticed it while reading the parser, and it is not purely cosmetic:
that name is the first element of the `ReactFunctionLocation` returned
by `extractLocationFromComponentStack` and
`extractLocationFromOwnerStack`, which `backend/fiber/renderer.js`
stores as `instance.source`. Any async component whose frame reaches
that path is recorded under a name with a stray leading space.

The same line exists in
`packages/react-server/src/ReactFlightStackConfigV8.js`, which the
DevTools file is a copy of. After review I fixed it in this PR as well,
in a second commit. There it only matters on the fallback path that
parses an already formatted stack string, when the error's `stack` was
read or assigned before React reaches it. #37130 is open on that file
too, but it does not touch these lines.

## How did you test this change?

I first confirmed the format V8 actually emits, rather than assuming it:

```
$ node -e 'async function inner(){await null;throw new Error("boom")}
async function outerName(){await inner()}
outerName().catch(e=>console.log(e.stack))'
Error: boom
    at inner ([eval]:1:52)
    at async outerName ([eval]:2:33)
```

Then I added a case to the existing `extractLocationFromComponentStack`
block in `utils-test.js`. Against `main` it fails with exactly the
leading space:

```
  ● utils › extractLocationFromComponentStack › should strip the async prefix from a frame name

    - Expected  - 1
    + Received  + 1

      Array [
    -   "Comments",
    +   " Comments",
        "https://react.dev/_next/static/chunks/848-122f91e9565d9ffa.js",
        5,
        9236,
      ]
```

With the one-character fix applied:

```
$ yarn test --build --project=devtools -r=experimental utils-test
PASS packages/react-devtools-shared/src/__tests__/utils-test.js
Tests:       63 passed, 63 total
```

I also ran the whole DevTools project before and after to check I was
not moving anything else. Both runs end at `9 failed, 4 failed suites`,
the same test names each time (`componentStacks`, `console`,
`inspectedElement`, `legacy/inspectElement`), so those failures are
pre-existing on `main` in my environment and unrelated to this change.
The only difference between the two runs is my new test: 586 passed
before, 587 after.

For the Flight side I added `ReactFlightStackConfigV8-test.js`, which
assigns a formatted stack to an error and checks what `parseStackTrace`
returns. Against `main` it fails with the same leading space (`"
outerName"`); with the fix it passes on stable and experimental in
development. It is gated to `__DEV__` because that fallback goes through
the DEV-only stack cache. `ReactFlightServer-test` and
`ReactFlightAsyncDebugInfo-test` still pass next to it (23 tests).

`prettier` and `eslint` are clean on all changed files. `yarn flow
dom-node` reported no errors for the DevTools commit; I did not rerun
Flow after the one-character Flight change.

AI tools used
@pull pull Bot locked and limited conversation to collaborators Sep 13, 2026
@pull pull Bot added the ⤵️ pull label Sep 13, 2026
@pull
pull Bot merged commit ccea5fd into code:main Sep 13, 2026
15 checks passed
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants