chore(deps): update @tenphi/tasty to 3.1.0 - #1341
Merged
Merged
Conversation
Points at 0.0.0-snapshot.5776c4c to run CI and Chromatic against the change before it releases. Colour-token opacity is now applied with `color-mix()` rather than the token's channel components, so every `#token.N` value emits `color-mix(in oklab, var(--x-color) N%, transparent)`. The colours are identical — mixing premultiplied against `transparent` leaves the channels alone — and the `--*-color-rgb` companions are untouched, so the raw CSS in GlobalStyles that reads them directly still works. One test pinned the old alpha syntax as a proxy for "the compound rule survived the merge". Rewritten to match the token it derives from instead, which is what it was actually guarding. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
🦋 Changeset detectedLatest commit: d58ca86 The changes in this PR will be included in the next version bump. This PR includes changesets to release 1 package
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 |
Contributor
📦 NPM canary releaseDeployed canary version 0.0.0-canary-9dc6bba. |
Contributor
🧪 Storybook is successfully deployed!
|
Contributor
🏋️ Size limit report
Compared against main at 947bd47 — run 32488249408, 2026-08-21T13:42:30Z.To see which modules changed, download the size-limit-statoscope-report artifact from this run and open report.html. |
Re-points at 0.0.0-snapshot.5a79c6d. Review of the upstream PR replaced `color-mix()` with relative colour syntax: mixing against `transparent` composed alpha instead of replacing it, and broke a percentage-valued opacity variable. `#token.N` now emits `oklch(from var(--x-color) l c h / .N)`, which also restores the previous computed-value shape. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
tenphi
changed the title
Re-points at 0.0.0-snapshot.73ab317. Chromatic on the previous snapshot caught disabled+selected buttons on the `current` theme rendering 2.5x more opaque: the upstream change had moved `#current.N` from composing its inherited alpha to replacing it, and this kit's `current` ramp is authored against the composition (`CURRENT_OUTLINE_STYLES` notes ".18 renders as ~.07"). `#current` is back to `color-mix()`, unchanged from 3.0.2. The generated CSS now differs from baseline only in the form of the token fades — same colour, same alpha. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
124.03 kB in CI, 26 bytes over the 124 kB entry, and all of it the dependency: both sides rebuilt here read 123.63 kB on tasty 3.0.2 and 124.03 kB on the snapshot, +0.40 kB, with `All` moving by the same 0.40 kB — the signature of a change inside Tasty's always-included core, as the 3.0.1 -> 3.0.2 note in the same entry describes. Raised to 125 kB rather than shaved, which is what that entry's own note asks for once a change needs more than its 330 B of headroom. `All` is left at 510 kB; it still passes at 509.37 kB. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Replaces the snapshot pin now that the release is out, and merges main. 3.1.0 carries both colour changes: `light-dark()`/`contrast-color()` treated as colours, and colour-token opacity applied with relative colour syntax (`oklch(from var(--x-color) l c h / .N)`) rather than the channel components. `#current.N` still composes with the alpha it inherits, so the `current` ramp is untouched, and the `--*-color-rgb` companions still hold numeric components for the raw CSS in GlobalStyles that reads them. Both size budgets move, and neither is this branch's own growth. Measured with main merged and the kit's source held constant, tasty is +0.39 kB on `All` (509.55 -> 509.94 kB) and +0.40 kB on the Button entry (123.63 -> 124.03 kB) — matching deltas, i.e. a change inside Tasty's always-included core. `All` had only 450 B of headroom on main before this, so 510 -> 515 kB; the Button entry goes 124 -> 125 kB, which its own note asks for rather than shaving something real. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
tenphi
changed the title
Merged
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Updates
@tenphi/tastyto the released 3.1.0 (was pinned to a snapshot while tenphi/tasty#268 was in review) and mergesmain.What changes upstream
3.1.0 carries two colour changes, #266 and #268:
Colour-token opacity is applied with CSS relative colour syntax instead of the token's channel components:
The channels are copied and the alpha slot is written, so the colour is identical and the alpha is replaced rather than composed. Because the outer function is
oklch(), the computed value keeps the shape it had before. It no longer needs the token decomposed into channels first, which is what makes acolor-mix()/light-dark()token — or a--x-colordeclared in plain CSS — fade correctly instead of falling back to an@propertyinitial value.Modern colour functions —
light-dark()andcontrast-color()joincolor-mix(),color-contrast()andcolor()as recognised colours, so a call lands in the colour slot of whichever style property uses it.What this touches here
Every
#token.Nvalue in the kit re-serialises:#surface-text.04,#purple.10, the#row-overlaystate map inTableBase, and so on. Chromatic reported1057 tests unchangedon the snapshot this release was cut from — the colours really are identical.Two things deliberately did not change, both of which this kit depends on:
#current.Nstill composes with the alpha it inherits. Thecurrentramp is authored against that —CURRENT_OUTLINE_STYLESnotes that.18"renders as ~.07" because the disabled label already dimmedcurrentcolorto.4. An earlier snapshot had unified#currentonto replacement, and Chromatic on this PR caught it: disabled+selected buttons rendered 2.5× more opaque. Fixed upstream before release.--*-color-rgbcompanions are still generated with numeric components, so the raw CSS inGlobalStyles.tsxthat reads one directly —rgb(var(--primary-color-rgb) / .2)— keeps working. Verified resolving to109 113 176viapnpm probe globals.One test changed
row-state-matrix.test.tsxasserted/--row-overlay-color:\s*rgb\(/on theselected & hoveredcompound rule. That pinned the old alpha syntax as a proxy for what the test actually guards — per its own comment, that the compound rule doesn't vanish into the state-map merge. Rewritten to match the token it derives from:Nothing else in 2094 unit tests or 123 browser tests needed touching.
Both size budgets move
Neither is this branch's own growth. Measured with
mainmerged and the kit's source held constant:Matching +0.4 kB deltas on both entries, which per
.size-limit.cjs's own diagnostic is what a change inside Tasty's always-included core looks like.Allhad only 450 B of headroom onmainbefore this, so any dependency bump was going to trip it; it is rounded to the next 5 kB step per the convention in that file. The Button entry goes to 125 kB, which its own note asks for once a change needs more than its 330 B. Both raises are recorded in the file with the measurements.The cost is real and worth stating plainly: ~0.4 kB gzipped reaches every consumer who imports anything, because it is in the parser and style handlers. There is nothing to tree-shake.
Verified locally
pnpm test— 2093 pass, 1 skipped (97 files)pnpm test:browser— 123 pass (10 files)pnpm lint— clean (only the pre-existingconsistent-token-usage/maskSizewarnings)pnpm size— 509.94/515 kB and 124.03/125 kB, both measured after a freshpnpm buildnpx tsc --noEmit— pre-existingNodeListOfiteration errors insrc/test/probe/only, unchanged frommainand not part of CI🤖 Generated with Claude Code