Fix flex shrink explosion when the shrink total is a positive float residual

Summary:
Fixes https://github.com/react/yoga/issues/2038
Fixes https://github.com/react/yoga/issues/1651
Closes https://github.com/react/yoga/issues/2039
Closes https://github.com/react/yoga/issues/1855

Thanks to sebholl for the analysis and the reproduction in https://github.com/react/yoga/issues/2038.

## Issue

When a flex line overflows, the first pass freezes the items that hit their minimum size and removes them from the shrink total. The second pass then divides the remaining free space by that total.

The total was computed by subtracting the frozen items flex shrink factors from the float total. When every item freezes, the total should become 0, but float subtraction leaves a small residual.

### Example

All items have `flexShrink: 1`:

```
row (width: 350, flexDirection: row)
├── A  width: 100.1  minWidth: 90
├── B  width: 200.3  minWidth: 190
└── C  width: 300.7  minWidth: 290
```

The items need 601.1 but the row is only 350, so they must shrink by 251.1. Each item would go below its `minWidth`, so the first pass freezes all three (at 90 / 190 / 290) and removes them from the shrink total. With no items left, the total should be 0. But in float, `-601.1 + 100.1 + 200.3 + 300.7 = 0.000061`.

That is above the `1e-6` guard added in https://github.com/react/yoga/issues/1974 (for https://github.com/react/yoga/issues/1665), so the second pass divides the remaining free space by it: `-220 / 0.000061 ≈ -3.6 million`. Instead of staying at their min, the items grow to about **361M / 722M / 1084M** (expected **90 / 190 / 290**).

### Solution

With non-negative `flexShrink`, every scaled shrink factor is ≤ 0, so a real total is never positive. A positive total can only be this float residual, and dividing by it is what makes the items explode. The second pass now treats a positive total like the near-zero case of the `1e-6` guard from https://github.com/react/yoga/issues/1974 and uses the existing fallback. Negative residuals (which the min clamp already handles) and all other layouts are unchanged. Fixing the fallback for `flexShrink < 1` (https://github.com/react/yoga/issues/931) is left for a separate PR.

#### Why not `inexactEquals` (|total| < 0.0001)?
The leftover grows with the size of the total. For example, 60 chips of about 100pt each in a 1000pt row give a total of about 6000, and the leftover is then often larger than 0.0001. In a test with rows like this, about half of them still exploded with `inexactEquals`. With the sign check, none did, because a real shrink total is never positive.

X-link: https://github.com/react/yoga/pull/2044

Reviewed By: javache

Differential Revision: D126038629

Pulled By: cortinico

fbshipit-source-id: dc1ec3370783e227ce4477bcd632ae5907034193
This commit is contained in:
nishan (o^▽^o) authored and meta-codesync[bot] committed 2026-10-20 08:42:12 -07:00
1 parent d63648bac6
commit 24a3be5f99
1 file changed
+6 -2
@@ -1138,12 +1138,16 @@ static float distributeFreeSpaceSecondPass(
// Use a relative epsilon guard instead of exact equality: after the
// first pass removes all constrained items,
// totalFlexShrinkScaledFactors may be near-zero rather than exactly 0
// due to floating-point cancellation. Dividing by a near-zero value
// due to floating-point cancellation. Dividing by a near-zero value
// produces a gigantic childSize that overwhelms the min/max clamp.
// A positive total is always such a residual: with non-negative
// flexShrink a real total is never positive
// (https://github.com/react/yoga/issues/2038).
const float shrinkFactorMagnitude =
std::abs(flexLine.layout.totalFlexShrinkScaledFactors);
if (yoga::isDefined(flexLine.layout.totalFlexShrinkScaledFactors) &&
shrinkFactorMagnitude < 1e-6f) {
(shrinkFactorMagnitude < 1e-6f ||
flexLine.layout.totalFlexShrinkScaledFactors > 0)) {
childSize = childFlexBasis + flexShrinkScaledFactor;
} else {
childSize = childFlexBasis +