On the release branch mobile-parity's <=767px head rules land before the
colour merge, so 'identical to ab7a100' cannot hold; run with
COLOR_BASE_COMMIT=<commit before the colour merge> there. Measured: vs ab7a100
the only difference is those two documented rules.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- The self-review follow-up sends intent:'review' + review_image_ids instead of
putting the render in mockup_ids, so the backend builds a critique+patch
instruction instead of 'build toward this mockup'. Render still deleted after.
- types/sitesmith.ts: SitesmithStatus = server statuses (adds IMAGES_BUSY,
TOO_MANY_JOBS, SERVER_BUSY, ...) + client statuses (TIMEOUT, ABORTED, ...);
network status strings are narrowed with a type guard (unknown -> ERROR);
the hook builds SendResultErr without casts.
- statusMessages.ts: one specific message per status (a Record over the union,
so a new status without a message fails tsc), used by the modal.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Conflicts: package.json (kept both test:browser:mobile and the colour scripts);
head-parity test (kept E_icons from mobile-parity/main AND the E_brand_swatches /
F_swatches_no_brand cases from color-system; pinned bytes re-derived later
against the merged WHP script).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
npm run test:browser:mobile; MOBILE_PARITY_RECORD/BASELINE compare desktop boxes against a base-commit snapshot.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Found by the real-editor e2e: picking over the header or its logo box returned the page colour behind it.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
craft/scripts/gen-sitesmith-catalog.mjs emits a compact JSON catalog (props with
TS types/enums/defaults, token-aware colour props, animation + icon support,
child rules, section recipes, templates, palettes) for the WHP panel prompt.
src/sitesmith/catalog.drift.test.ts regenerates and compares to the committed
whp file so drift fails.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Published pages/Preview ship a head reset (*{margin:0;padding:0}, img block,
a inherit, body line-height 1.6 / #1f2937 / 16px); the editor canvas had none,
so h1-h6/p margins, line-height and inherited size/colour differed.
- editor.css: one commented block scoping the reset to .canvas-device-frame,
every selector zero-specificity (:where) so inline styles / class rules win
exactly as when published; editor chrome untouched (snapshot identical).
- tokens.ts: canvas heading/link token rules made zero-specificity too.
- utils/css-reset.ts: single source for the reset declarations (html-export
imports it; test asserts byte-identity with the backend head and that the
editor.css block carries the same declarations).
- scripts/e2e-canvas-parity.mjs (npm run test:browser:canvas): real editor vs
published HTML in Firefox + Chromium, boxes within 1px + computed styles.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
HeroSimple emitted 'background' together with backgroundImage/Size/Position
that were undefined. React assigns '' for those on first render and Chromium
and Firefox then wipe the shorthand's gradient (background-image: none), so
the canvas hero rendered white. Build the style with only the relevant keys
(shared by render and toHtml); exported HTML is byte-identical.
Adds a real-browser (Chromium+Firefox) regression test: npm run test:browser:bg,
and toHtml byte-identity assertions.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
