max / makeover

2.5.2: give tokyonight a raised surface Its page and raised were the identical hex, #1a1b26, so the theme had no ramp at all rather than a shallow one and nothing in it could lift off anything. It was the one entry in CANNOT_LIFT_OFF_THE_PAGE that was an omission rather than a judgment about someone else's palette. Raised becomes #292e42, which is upstream Tokyo Night's own bg_highlight rather than a colour picked here. That puts 0.079 oklab L between page and raised, above nord's 0.055 and well clear of the 0.05 line. The derived well and bevel pair still pass their own degeneracy tests, so the deeper face costs nothing downstream. The other nineteen entries stay. They are shallow ramps in published palettes, which is a different claim from a missing one, and re-ramping them waits on every app being migrated and eyeballed.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

Author: Max Johnson <me@maxj.phd> · 2026-08-15 21:10 UTC

Signed with PGP, not checked

Commit: 2e03a85ad835eeaf45e35e909a7797e9084fdd3a

Parent: b4775a7

3 files changed, +10 insertions, -3 deletions

M Cargo.toml +1 -1

			@@ -1,6 +1,6 @@
1	1		[package]
2	2		name = "makeover"
3		-	version = "2.5.1"
	3	+	version = "2.5.2"
4	4		edition = "2024"
5	5		description = "Shared theme loading for the make-family apps: TOML theme files parsed into intent-based color tokens, with perceptual derivations and WCAG contrast."
6	6		license = "MIT"

M src/lib.rs +8 -1

			@@ -1974,6 +1974,14 @@
1974	1974		// highlighting and owe our depth model nothing. Failing here says a theme
1975	1975		// cannot hold a bevel, not that it is wrong. Shrinking the list is the fix;
1976	1976		// growing it is a regression in the theme, not in this derivation.
	1977	+	//
	1978	+	// tokyonight left the list on 2026-08-15, and it is the only entry that could
	1979	+	// leave without a judgment call about someone else's palette. Its page and
	1980	+	// raised were the identical hex, so it had no ramp at all rather than a
	1981	+	// shallow one, and the fix is upstream's own `bg_highlight` (#292e42, 0.079
	1982	+	// above the page) rather than a color we picked. The other nineteen are
	1983	+	// shallow ramps in published palettes, which is a different claim, and they
	1984	+	// stay deferred until every app is migrated and eyeballed.
1977	1985		#[test]
1978	1986		fn raised_is_distinct_from_page() {
1979	1987		// Below this, a raised surface and the page under it are one surface to
			@@ -1999,7 +2007,6 @@
1999	2007		"rosepine",
2000	2008		"rosepine-dawn",
2001	2009		"solarized-dark",
2002		-	"tokyonight",
2003	2010		];
2004	2011
2005	2012		let mut flat: Vec<String> = Vec::new();

M themes/tokyonight.toml +1 -1

			@@ -11,7 +11,7 @@
11	11
12	12		[surface]
13	13		page = "#1a1b26"
14		-	raised = "#1a1b26"
	14	+	raised = "#292e42"
15	15		sunken = "#283457"
16	16		overlay = "#16161e"
17	17