Sharpen the sun's halo rule: white in-app, amber only outside#260
Merged
Conversation
The sun wore an amber halo in two places: the resting dashboard companion and the departing hand-off. Only one of those had a reason. Sharpen it to one rule - the sun's body may be warm, but the light it casts is white. Amber is not a day/night signal and not a role badge; it belongs to the sun as it lives *outside* our own surfaces, where it has to announce itself over content we don't control (the Little Sun overlay, the home-screen widgets). The departing hand-off still warms, because it is becoming that Little Sun - the warming is the hand-off, not a state. So the companion settle drops `warmth: 1`. It still reads sunlit through its warm disc rim (--minded-sun-face-edge), and it keeps the snug reach, which is about the bottom-edge clip (#106), not colour. The upshot is that the everyday lift off the bottom bar into an intervention is now a pure size and position morph rather than doubling as a colour change - one sun, one light. The cool half of the glow axis is untouched (the up-drag let-go read), and the moon never warmed anyway. Rule written down in CLAUDE.md and as a block comment atop sunSettle.ts, with tests guarding every in-app settle against amber creeping back. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KvG5EVDfN7tVKWwatiMAUj
Dropping the companion's amber left four comments asserting the opposite - that the canonical amber is "shared with the resting companion", that the sun "carries a warm halo at all times". Corrected at all four sites (sunAnimationUtils.ts, _variables.scss, LittleSun.kt, Sun.tsx), each now pointing at the rule rather than restating it. Add docs/sun-halo.md: the rule plus a full map of what the sun actually looks like on every surface - the four knobs that vary, a per-state table for day and for night, and the four out-of-app surfaces where the amber legitimately lives. That map existed nowhere before, so "which sun is which colour" could only be answered by reading five files. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KvG5EVDfN7tVKWwatiMAUj
The rule said "the widgets sit on the user's wallpaper, so they wear amber". Only half of each widget does. Both platforms have two faces: a small transparent one that floats on the wallpaper, and a prompt card that paints the app's own sky behind the sun. The card is our surface, so the amber bloom had no business being there. Sharpen the test to what the sun *stands on* rather than which process draws it, and give the card a white bloom: - Android: ic_sun_widget_day_on_sky.xml, a twin of ic_sun_widget_day differing only in the bloom's four gradient stops. drawableFor() takes onOwnSky; PromptCard passes true. widget_preview_card.xml (the picker still that mirrors the card) uses it too. previewImage stays amber - the pre-31 gallery renders it on the launcher's surface, not ours. - iOS: CompanionSun gains onOwnSky, switching the same stops at the same 0.28 alpha to white; the systemMedium PromptCard passes true. The disc keeps its faint gold rim on both - that is the body, not the halo, and the in-app companion carries the same rim. The moon needs no twin: it never warms on any surface. Rule text updated in CLAUDE.md, sunSettle.ts, docs/sun-halo.md, docs/sun-companion-widget.md and LittleSun.kt, and docs/sun-halo.md now splits the out-of-app table into the two cases. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KvG5EVDfN7tVKWwatiMAUj
… twins Review of the branch found the reference doc I added states a falsehood and that five comments still assert the old behaviour - two of them inside blocks these commits edited. Also that nothing enforced the "keep the two files in step" instruction written into three of them. - docs/sun-halo.md said the Android Little Sun has an amber gradient face. It doesn't: LittleSun.kt's SUN_COLOR is Color.White. I had conflated the *native* overlay (30dp, white disc) with the *CSS* one the WebView draws (40px, amber gradient face) - there are two Little Suns on Android, with different sizes and faces. Split into rows and noted why sunSettle.ts keeps a departing target for each. - Corrected the five surviving stale comments: Sun.tsx x2 (the glow-axis note still naming companion rest as warm; "warm, symmetric rest halo"), CompanionSun.swift's header, ic_sun_widget_day.xml's header (now says which twin it is and why), and the iOS widget README. - Fixed a pre-existing error in the block these commits document: DEPART_GLOW_INTENSITY cited COMPANION_REST_GLOW as 1.8; it is 1.25 (1.8 is the hover glow), and the claim that the companion wears a broad bloom contradicts SNUG_GLOW_REACH ten lines below. - FingerRest's comments called its amber "the companion's own light" - the thing this branch just made white. It's the *little sun's* amber; reworded, and the pad itself deliberately left as-is. New sunHaloMirror.test.ts guards the native half across the language boundary, following widgetClockMirror.test.ts: the two drawables may differ only in the bloom's four stops, the on-sky bloom is white and the wallpaper one the canonical amber, the disc is identical on both, iOS matches Android stop for stop (including 0x47 == 0.28), and each call site passes onOwnSky by what the sun stands on. Verified by mutation - amber on the on-sky bloom, a dropped onOwnSky, and an iOS colour drift each fail it, and all three go green again when reverted. sunSettle.test.ts's phase list is now a Record keyed by Exclude<SunPhase,"departing">, so an eighth phase fails to compile rather than silently escaping the guard (verified: TS2741). 574 tests pass, lint clean. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KvG5EVDfN7tVKWwatiMAUj
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
2 participants
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.
The sun wore an amber halo in two places: the resting dashboard companion
and the departing hand-off. Only one of those had a reason.
Sharpen it to one rule - the sun's body may be warm, but the light it
casts is white. Amber is not a day/night signal and not a role badge; it
belongs to the sun as it lives outside our own surfaces, where it has
to announce itself over content we don't control (the Little Sun overlay,
the home-screen widgets). The departing hand-off still warms, because it
is becoming that Little Sun - the warming is the hand-off, not a state.
So the companion settle drops
warmth: 1. It still reads sunlit throughits warm disc rim (--minded-sun-face-edge), and it keeps the snug reach,
which is about the bottom-edge clip (#106), not colour. The upshot is
that the everyday lift off the bottom bar into an intervention is now a
pure size and position morph rather than doubling as a colour change -
one sun, one light.
The cool half of the glow axis is untouched (the up-drag let-go read),
and the moon never warmed anyway.
Rule written down in CLAUDE.md and as a block comment atop sunSettle.ts,
with tests guarding every in-app settle against amber creeping back.
Co-Authored-By: Claude Opus 5 noreply@anthropic.com
Claude-Session: https://claude.ai/code/session_01KvG5EVDfN7tVKWwatiMAUj