| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
| |
consistency-check tasks
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
defaults, boxes
Acting on the per-package face audit (every package's faces checked against the live daemon and source):
Removed 11 nonexistent mu4e faces — mu4e-cited-1-face through -7-face, mu4e-attach-number-face, mu4e-compose-header-face, mu4e-moved-face. Newer mu4e dropped these (citations now inherit message-cited-text-N), so they themed nothing and showed phantom samples in the preview. Confirmed absent in the installed mu4e source before removing.
Boxed the faces that actually use :box, now that the attribute exists: magit-branch-current and magit-branch-remote-head (flat line box, matching :box t), and the released-button buttons — two flycheck list faces, the telega box-button family, and the slack button/dialog/reaction faces.
Restored the structural defaults the seeds had dropped relative to each package's own defface: underline on org-link/-cite/-cite-key, lsp-face-highlight-read/-rename, flycheck-error/-warning/-info, mu4e-header-highlight-face, calibredb-search-header-highlight-face, telega-describe-section-title, slack-message-output-header/-channel-button; bold on org-special-keyword, lsp-face-highlight-write, all four git-gutter faces, flycheck-error-list-highlight, the erc nick cluster (nick-default, button-nick-default, nick-prefix, my-nick-prefix, nick-msg, notice, pal, button), org-noter's two faces, and three slack label faces; italic on dirvish-media-info-property-key; height on lsp-details-face (0.8), shr-sup (0.8), calibredb-current-page-button (1.1).
Left the taste-call contradictions for review (shr-h3/h4-6, telega-msg-user-title, erc-action-face) — these are theme-design choices, not clear defects. dashboard, dired, org-drill, elfeed, signel, and pearl were clean.
|
| |
|
|
|
|
|
|
| |
The mode-line box in the preview was hardcoded — it showed a box the generated theme couldn't actually produce, since build-theme.el never emitted :box. Made :box a real face attribute instead: a per-face box object (style line/raised/pressed, width, color) stored on UI and package faces, set from a "box" dropdown in both tables, rendered from the attribute everywhere (the mode-line bars, the package previews via ofs, the UI table preview cells), and exported through build-theme.el's --attrs as a proper :box plist (released/pressed → :style *-button; line → :line-width + optional :color). The hardcoded box is gone; mode-line and mode-line-inactive now default to the released-button box that is the Emacs default, so the preview and the export agree.
This also gives the package faces that genuinely use :box a way to represent it — the face audit found several (magit-branch-current/-remote-head, two flycheck list faces, the telega button family, ~15 slack button/dialog faces).
Tests: build-theme gains box-conversion + ui-box-emit ERT tests (24/24); the app-core deep-equal tests account for the new box slot; all 9 browser gates, 20 python, and 55 node tests stay green.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Emacs does
The mock buffer rendered several faces wrong, verified against the running daemon and emacs -Q:
- highlight was flat monochrome text. Like region, the highlight face is background-only, so Emacs lets the syntax foreground show through (face merge: highlight over a keyword keeps the keyword color, adds the highlight background). Both now go through one overlay helper that keeps token colors and only overrides the foreground when the face sets one.
- cursor was an empty block at end-of-line. Emacs draws a box cursor on the character at point, in the frame background over the cursor color — now it sits on a real glyph.
- the overlay faces, the line-number gutter, and the paren faces ignored weight/underline/etc. They now honor bold/italic/underline/strike, matching the table preview.
- the fringe showed only its background. It now draws a continuation indicator in the fringe foreground.
- the mode line had no box. It now carries the 3D released-button box that is the Emacs default.
Defaults seeded to the theme-independent Emacs core styles (from emacs -Q): error/warning/success are bold; lazy-highlight and show-paren-match are underlined (link already was).
Added a #mocktest gate pinning these — overlay faces keep token colors, the cursor is on a glyph, styles render, the fringe indicator is present, the mode line has its box. Verified it goes red when the rendering regresses.
|
| |
|
|
|
|
|
|
|
|
| |
cover them
The pure helpers that were still stranded in app.js — normHex, ratingColor, textOn, and the filename-slug logic — had no unit tests because app.js can't be imported (it runs its bootstrap and references the data placeholders at load). Moved them into importable modules so they can be tested directly: a new app-util.js holds the color/UI-boundary trio, and slugify joins app-core.js. app.js keeps thin wrappers, so no call site changed and the built DOM is byte-identical.
textOn needs rl from colormath, so generate.py's inline strip now drops import lines as well as export lines — app-util.js imports rl for its tests, and the import is stripped on inline where rl is already in the page. _faces in generate.py also gets direct tests for its prefix-strip and label derivation.
New: 12 node tests (normHex, ratingColor, textOn, slugify) and 7 python tests (_faces, app-util integrity, the import strip). Coverage: app-util.js 100/100/100, app-core.js 100/94.9/100, colormath.js 100/96/100 (line/branch/func); generate.py 89% lines (the rest is the __main__ writer and the optional seed-env branch). No bugs surfaced — the logic was correct, just untested.
|
| | |
|
| |
|
|
|
|
|
|
| |
The refactor's goal was to make the app logic testable; this realizes it. Pulled the pure package-face model and the dropdown option list into app-core.js — nameToHex, buildPkgmap, packagesForExport, mergePackagesInto, effResolve (the inherit-chain resolver behind pkgEffFg/pkgEffBg), and optList — with every dependency passed as a parameter so there is no DOM and no module-global reliance. generate.py inlines it into the page the same way it inlines colormath.js (strip exports, placeholder, integrity check), so the browser runs the same code the tests import.
app.js keeps thin wrappers (pname, seedPkgmap, ddList, pkgEffFg, pkgEffBg) that pass the live PALETTE / APPS / PKGMAP into the core, so no call site changed and the built DOM is byte-identical to before.
test-app-core.mjs adds 18 Normal/Boundary/Error tests over the extracted logic — name resolution, the seed/export/merge round trip, the inherit chain including a cycle that must terminate at null, and the "(gone)" dropdown entry — plus an inline-integrity check that the page carries the core verbatim. The node suite goes 25 to 43 tests; python templating gains the app-core integrity assertion.
|
| |
|
|
|
|
|
|
| |
The syntax table had its own sort (srt + a D{} direction map) that read state directly — MAP[kind] for the color column, cell text for elements. The UI and package tables used a separate, more general system (srtTable / cellVal / applyTableSort) that reads the rendered cells. Pointed the syntax headers at srtTable('legbody', col) and deleted srt, so all three tables share one sort.
The mapping is exact: the legtable color cell is a swatch dropdown whose data-val is the hex, which cellVal reads — same key srt sorted on — and the elements cell is text. First-click direction stays ascending. The syntax table sorts on click only; it doesn't opt into the cross-rebuild persistence the UI and package tables get from applyTableSort, which preserves its prior behavior.
Added a #sorttest gate: sort was previously untested, and this collapses two systems into one. It checks the syntax table sorts by color ascending, reverses on a second click, sorts by element name, and that the UI and package tables still sort. The asc/desc pair is self-validating — a no-op sort can't pass both.
|
| |
|
|
|
|
|
|
| |
generate.py was 1378 lines, ~1300 of them a single triple-quoted string holding the whole app. Moved the <style> block to styles.css and the <script> body to app.js, and generate.py now inlines both through placeholders the same way it already inlines colormath.js, then fills the data placeholders. It drops to ~500 lines (the remaining bulk is the package face-data dicts, a later stage).
The generated page is byte-identical to before — every hash gate, the node suite, the spliced-script parse, and the new #locktest stay green. Two integrity tests guard the splice: styles.css inlines verbatim, and app.js reaches the page exactly as fill_data renders it. Both go red if the splice wiring is dropped.
Living in real files instead of a Python string kills the backslash-doubling bug class (str.replace is literal, so escapes survive), gives the CSS and JS real editor tooling, and opens the app logic to unit testing — the point of the whole refactor.
|
| |
|
|
| |
Adds a browser hash gate covering the two lock behaviors no existing gate touched: locking a row disables its control (syntax swatch div via data-locked, UI select via .disabled, both through the shared mkLockCell), and clear-unlocked wipes unlocked rows to default while leaving locked rows untouched across all three tiers. This is the characterization net for the generate.py extraction refactor — it proves the upcoming CSS/JS move preserves lock behavior. Verified it goes red when a lock guard is removed.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
The design session added a batch of tool features. They landed together in the page generator, so this commits them as one.
A custom swatch color dropdown replaces the native select in the syntax table, showing a color square, name, and hex per row, because native option backgrounds render unreliably on Linux Chrome.
THEME_STUDIO_SEED=<file.json> seeds the tool's starting palette, assignments, bold, italic, and UI from a theme.json at generate time, so opening the tool on a given palette is one regenerate with no import step.
Lock toggles sit on all three tables. A closed lock marks a decided element/color pairing and disables that row's editing controls. Clear-unlocked resets the rest to default and leaves the locked rows alone. Locks persist in the saved theme, with keys namespaced (bare syntax, ui:, pkg:) so the tiers don't collide.
The UI faces table gained a contrast column, and the three tables now share one column order: name, lock, color, style, contrast. The bg and fg rows are renamed to "bg (ground)" and "fg".
|
| |
|
|
|
|
| |
The page generator's risky logic, the export-strip that inlines colormath.js and the placeholder substitution that fills in the sample and palette data, had no tests. A bug there ships a broken theme-studio.html the JS tests can't see, since they import colormath.js directly and never look at the assembled page.
I pulled the strip into a strip_exports function and guarded the file write behind __main__, so importing generate.py builds the page in memory without writing it. test_generate.py asserts the strip removes export lines, preserves the body, and rstrips. It asserts the assembled page has every placeholder filled and carries the colormath body verbatim. And it guards that colormath.js keeps its export on a single line, since the line-based strip would leave a multi-line export's continuation behind. That was the exact failure that bit during the first integration.
|
| |
|
|
|
|
|
|
| |
The picker's two heaviest pieces of pure logic lived as strings inside generate.py, reachable only through the single-scenario browser hash tests. I moved them into colormath.js, where they get the same direct Node testing the color math has: planeCell(L,C,H) returns a C×L plane cell's color or flags it out of gamut, and paletteWarnings(palette, threshold, cap) does the pairwise ΔE analysis and returns the too-close pairs, the overflow count, and each color's nearest neighbor. The page now calls both. The inline copies are gone.
The new Node tests cover what the hash tests never could: empty, single, and identical-color palettes; the strict threshold boundary; the cap and overflow count; closest-first ordering; the C=0 achromatic case; and a plane cell pinned to oklch2hex's clamped flag so the plane and the commit path agree on the gamut edge.
The refactor preserves behavior: the page renders identically, guarded by the existing #deltatest and #planetest characterization gates.
|
| |
|
|
|
|
|
|
| |
Perceptual-metrics Phase 4b, the last piece of the OKLCH editor. In OKLCH mode the picker's square becomes a Chroma (x) by Lightness (y) plane at the current hue. The crosshair maps to (C, L) and the hue strip selects H. The unreachable region is greyed out so the sRGB gamut boundary at that hue is visible, and the AA/AAA contrast mask overlays on top of the reachable colors.
The per-cell in-gamut test is forward-only: oklch to oklab to linear-rgb plus a channel-range check, never the binary search, which stays in oklch2hex for committing a chosen color. colormath.js now exports oklab2lrgb, inGamut, and lrgb2hex (with direct Node tests, including one that pins inGamut to oklch2hex's clamped flag so the plane and the commit path agree on the boundary). The rendered bitmap caches on hue, dimensions, mask, and background, so dragging C and L at a fixed hue reuses it.
HSV stays untouched: the square keeps its saturation/value gradient and the existing contrast mask. A #planetest headless guard asserts the crosshair lands at the color's (C, L), an out-of-gamut cell renders as the grey fill, and an in-gamut cell renders as a real color.
|
| |
|
|
|
|
|
|
| |
Perceptual-metrics Phase 4a. The picker gains an edit-model toggle (HSV / OKLCH) held in its own pkModel state, orthogonal to the existing AA/AAA contrast mask (pkMode). The two never share state: one is how you edit the color, the other is what constraint you mask.
In OKLCH mode the picker shows L, C, and H as paired range and number inputs that drive the color through oklch2hex, updating the hex field, swatch, readouts, and the HSV cursor. When the requested chroma is out of sRGB gamut the dials snap to the reachable color and a "chroma clamped to sRGB" line appears. HSV stays the default, and the SV square keeps editing in HSV (the C×L plane is Phase 4b). An SV drag in OKLCH mode refreshes the dials so the two surfaces stay consistent.
A #oklchtest headless guard asserts switching to OKLCH preserves the color, toggling the mask leaves pkModel alone, switching the model leaves pkMode alone, the dials drive the color to a known OKLCH target, and an out-of-gamut chroma raises the clamp status.
|
| |
|
|
|
|
| |
Perceptual-metrics Phase 3. renderPalette now runs a pairwise OKLab ΔE over the palette and warns on any pair below the named DELTAE_MIN threshold (0.02). The warning lists the closest pairs first, caps at five, and appends "and N more" so a noisy palette never hides the count. Each chip's tooltip gains its nearest-neighbor ΔE.
paletteDeltas computes the pairs and the per-color nearest distance in one pass, feeding both the chip titles and the warning list. Palette names go through esc before they reach the warning markup. A #deltatest headless guard asserts a near-identical pair fires and names itself, a spread palette stays quiet, and a tight cluster caps at five in ascending order with the overflow suffix.
|
| |
|
|
| |
The .pinfo2 row had the font and margin but not the flex space-between the .pinfo row above it uses, so the OKLCH and APCA spans ran together as "90°APCA Lc -39". With space-between the OKLCH coordinates sit under the hex and the APCA Lc sits under the WCAG ratio, matching the row above.
|
| |
|
|
|
|
|
|
| |
Perceptual-metrics Phase 2. The picker now shows a second readout row under the WCAG ratio: the OKLCH coordinates (L, C, hue°) and the signed APCA Lc against the ground color. WCAG and all three contrast tables are untouched, so APCA stays picker-only for v1.
APCA Lc carries its sign convention in a tooltip and in the README: positive is dark text on a light background, negative is light text on a dark background, so a light color on dupre's dark ground reads negative.
pkReadout drives the new spans from the inlined colormath functions, and a #readouttest headless guard loads dupre-blue and asserts the spans match both the live computation and the known OKLCH reference, with WCAG unchanged.
|
| |
|
|
|
|
|
|
| |
Perceptual-metrics Phase 1. generate.py inlines the colormath.js body into the page script, stripping the ES-module export so one source feeds both the browser and the Node tests. The page's own lin, rl, contrast, rating, hsv2rgb, rgb2hsv, hex2rgb, and rgb2hex copies move into colormath.js. normHex, textOn, and ratingColor stay in the page as UI-boundary helpers.
rl now reuses colormath's canonical lin (0.04045 cutoff) instead of the old 0.03928 form. The two are byte-identical on every #rrggbb: no 8-bit channel falls between the cutoffs (10/255 = 0.0392, 11/255 = 0.0431), confirmed over 200k random pairs with zero contrast change and no AA/AAA flips.
test-colormath.mjs adds Normal/Boundary/Error cases for the migrated helpers, a seeded hsv-rgb round-trip property test, and an inline-integrity check that the generated page carries the colormath.js body verbatim, so the inlined copy and the tested module can't drift.
|
| |
|
|
| |
openPicker positioned the crosshair while the picker was still display:none, so its client dimensions read 0 and the crosshair pinned to the top-left corner instead of the shown color. I moved display:block ahead of paintPicker so the element has layout when its size is read, then added a headless #cursortest that opens the picker on a saturated color and checks the crosshair lands off-corner.
|
| |
|
|
| |
Clicking a UI or package face already called the flash, but two things hid it: the flashtok animation was scoped to #codepre and .ex cells, so the class landed on the mock-frame and package-preview spans without ever animating; and the flash never scrolled its target into view, so an element below the fold of a preview's scroll box flashed unseen. The animation is now global, and the flash scrolls its element into view (the first of several, via a new flashEls helper). Clicking a face in either tier now lands the viewport on the matching preview element and lights it up.
|
|
|
The tool authors themes from scratch -- palette, faces across every tier, live preview, export to a loadable deftheme. It never selects among existing themes, so "selector" mis-described it. Renamed the directory, the generated HTML and its title, the design spec, and every reference in the code, README, tests, and todo. No behavior change.
|
