fix: preserve alpha when interpolating in oklch

[mahirhir]

The shared hue interpolator in src/interpolator/_hsx.js reverses the polar color so the hue ends up at index 0 for the hue-blending math. For oklch it does that with Array.prototype.reverse():

xyz0 = col1.oklch().reverse();
xyz1 = col2.oklch().reverse();

rgb2oklch appends the alpha channel when it is below 1, so oklch() returns [L, c, h, a] for a semi-transparent color. reverse() flips the whole array to [a, h, c, L], and the next line [hue0, sat0, lbv0] = xyz0 then reads alpha as the hue, hue as the chroma and chroma as the lightness. So any oklch interpolation/mix where an input has alpha < 1 comes out wrong.

The hcl branch right above it does not have this problem: Color.prototype.hcl was switched to reverse3 (in 27a2bb3, resolving #359), which only reverses the first three components and leaves alpha in place. The oklch branch was left on the old reverse(). This change uses the same reverse3 helper there.

Mixing a half-transparent red with blue, the rgb part should match the opaque mix (only the alpha channel should differ):

// before
chroma.mix('#ff000080', '#0000ff', 0.5, 'oklch').hex('rgb') // '#ff00ff'  (wrong)
chroma.mix('#ff0000',   '#0000ff', 0.5, 'oklch').hex('rgb') // '#ba00c2'  (opaque, correct)

// after
chroma.mix('#ff000080', '#0000ff', 0.5, 'oklch').hex('rgb') // '#ba00c2'
chroma.mix('#ff000080', '#0000ff', 0.5, 'oklch').hex()      // '#ba00c2bf'

Added a regression test in test/mix.test.js and a changeset. The full suite passes.

[changeset-bot]

🦋 Changeset detected

Latest commit: 57b5ef5

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
chroma-js	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR