Add enum-specific condition negation

[nmn]

Adds the enum-specific condition logic that #1870 uses for conditional assignments. This PR adds the planner and its tests; the public conditional-call API is connected in the next PR.

For an enum with comfortable, compact, and spacious states, the next PR enables this call:

density({
  default: 'comfortable',
  ':hover': 'compact',
  ':active': 'spacious',
});

Hover and active can both match. Ordinary styles can let two rules compete over color; an enum writes a different custom property for each state. Simply activating compact and spacious in separate rules would leave both states active.

The planner therefore excludes active from the compact branch. This is the planned CSS structure, using readable names for the classes and state variables:

:where(.densityReset) {
  --density-comfortable: ;
  --density-compact: ;
  --density-spacious: ;
}

.selection {
  --density-comfortable: initial;
}

.selection:where(:hover):not(:where(:active)) {
  --density-comfortable: ;
  --density-compact: initial;
}

.selection:where(:active) {
  --density-comfortable: ;
  --density-spacious: initial;
}

With hover alone, compact is selected. While pressed, spacious is selected even if hover also matches. With neither condition, comfortable remains selected. initial activates a state; the empty value deactivates it.

The default is set unconditionally, and matching branches explicitly clear it. Container branches also clear earlier states when they match, so an element with no eligible container keeps its default.

Nested conditions are excluded as a whole. If spacious requires both active and a wide viewport, compact remains available while hovering on a narrow viewport. Branches that select the same state do not need exclusions from one another.

Ordinary styles keep the simpler output from #1867. This stronger negation is isolated to enum assignments, where multiple active custom properties would otherwise break stylex.match.

Validated with selector snapshots, all combinations of hover/active/focus, container fallback cases, and the existing ordinary-style snapshots across 1,025 Babel tests.

PR 5 of 6. Builds on #1868. Next: #1870.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated
stylex	Ignored		Sep 10, 2026 9:57am UTC

[github-actions]

workflow: benchmarks/perf

Comparison of performance test results, measured in operations per second. Larger is better.
yarn workspace v1.22.22
yarn run v1.22.22
$ node ./compare.js /tmp/tmp.lrqLWHocoM /tmp/tmp.YGKPfCqMOx

Results	Base	Patch	Ratio
babel-plugin: stylex.create
· basic create	576	575	1.00	-
· complex create	62	63	1.02	+
babel-plugin: stylex.createTheme
· basic themes	454	462	1.02	+
· complex themes	32	32	1.00
Done in 0.08s.
Done in 0.34s.

[github-actions]

workflow: benchmarks/size

Comparison of minified (terser) and compressed (brotli) size results, measured in bytes. Smaller is better.
yarn workspace v1.22.22
yarn run v1.22.22
$ node ./compare.js /tmp/tmp.507cnz4YLE /tmp/tmp.CE76K3hgAz

Results	Base	Patch	Ratio
@stylexjs/stylex/lib/cjs/stylex.js
· compressed	1,571	1,571	1.00
· minified	5,346	5,346	1.00
@stylexjs/stylex/lib/cjs/inject.js
· compressed	1,793	1,793	1.00
· minified	4,915	4,915	1.00
benchmarks/size/.build/bundle.js
· compressed	496,732	496,732	1.00
· minified	4,847,825	4,847,825	1.00
benchmarks/size/.build/stylex.css
· compressed	99,735	99,735	1.00
· minified	740,755	740,755	1.00
Done in 0.07s.
Done in 0.30s.