Bugzilla

Hmm this test case is really depending on implementation.  From [the test case](https://searchfox.org/mozilla-central/rev/b72e9d7d63bf499d1d8168291b93d4ec7fde236e/testing/web-platform/tests/css/css-scroll-snap/input/keyboard.html#51-52);

```
  const scrollCounter = new ScrollCounter(t, scroller);
  await keyPress(scroller, "ArrowDown");
  await waitForScrollEnd(scroller, scrollTop, 400);
  // Make sure we don't jump directly to the new snap position.
  assert_greater_than(scrollCounter.count, 2);
```

So, this test case supposes that browsers do snap once after the scroll trigger by ArrowDown key finished?  In our implementation the scroll by key press directly goes toward the snap point, the reason why it often passes, I mean there are 3 scroll events, is the scroll is async. And getting 2 events is just by chance.

Hmm this test case is really depending on implementation.  From [the test case](https://searchfox.org/mozilla-central/rev/b72e9d7d63bf499d1d8168291b93d4ec7fde236e/testing/web-platform/tests/css/css-scroll-snap/input/keyboard.html#51-52);

```
  const scrollCounter = new ScrollCounter(t, scroller);
  await keyPress(scroller, "ArrowDown");
  await waitForScrollEnd(scroller, scrollTop, 400);
  // Make sure we don't jump directly to the new snap position.
  assert_greater_than(scrollCounter.count, 2);
```

So, this test case supposes that browsers do snap once after the scroll triggered by ArrowDown key finished?  In our implementation the scroll by key press directly goes toward the snap point, the reason why it often passes, I mean there are 3 scroll events, is the scroll is async. And getting 2 events is just by chance.

Hmm this test case is really depending on implementation.  From [the test case](https://searchfox.org/mozilla-central/rev/b72e9d7d63bf499d1d8168291b93d4ec7fde236e/testing/web-platform/tests/css/css-scroll-snap/input/keyboard.html#51-52);

```
  const scrollCounter = new ScrollCounter(t, scroller);
  await keyPress(scroller, "ArrowDown");
  await waitForScrollEnd(scroller, scrollTop, 400);
  // Make sure we don't jump directly to the new snap position.
  assert_greater_than(scrollCounter.count, 2);
```

So, this test case supposes that browsers do snap once after the scroll triggered by ArrowDown key finished?  In our implementation the scroll by key press directly goes toward the snap point, the reason why it often passes, I mean there are 2 scroll events, is the scroll is async. And getting 2 events is just by chance.