/* Theme Name: Elsteiner Summer Theme URI: https://elsteinergames.com Author: Elsteiner Games Description: Successor to elsteiner-2026. v1 block theme; design system authored Tier-1 onward per docs/areas/website/ia/. Version: 0.7.41 Requires at least: 6.4 Requires PHP: 8.0 Text Domain: elsteiner-summer */ /* The Version above is the cache-busting `?ver=` on EVERY asset functions.php enqueues — tokens.css, web-blocks.css, web-blocks-local.css and eg-web.js all read `wp_get_theme()->get( 'Version' )`. Bump it with any change to those files or returning visitors keep the cached copy and the deploy is invisible. 0.6.5 — horizontal-scroll edge cue: web-blocks-local.css gains the mask and eg-web.js the state that drives it, so a stale pair would leave the CSS waiting on a data attribute nothing sets. 0.6.7 — el/score-calculator at 360px: web-blocks-local.css stacks the seat row and lifts the radio and caption sizes. CSS-only, but a returning phone is exactly the visitor who would otherwise keep the broken cached copy. 0.6.10 — F12 table bottom-border override (below) + F24 heading scale (theme.json): raised h1/h2/h3 to fluid clamp() sizes, and fixed the real reason they read too small in the first place — theme.json's styles.elements.{h1,h2,h3} hand-typed var(--wp--preset--font-size--h1) etc, but WP kebab-cases numeric-suffixed slugs when generating the preset variable, so the actual custom property is --wp--preset--font-size--h-1 (confirmed on staging's global-styles-inline-css). The old references pointed at a variable that never existed, so h1/h2/h3 never picked up a theme.json font-size at all — same bug for the paired line-height custom properties. This file also carries its own ?ver=, so it needs the same bump discipline as the files above. 0.6.11 — Printables unit (feedback 2026-08-19, F8): :has() opt-out so the score-sheets .eg-dlgrid escapes the shared game_extra template's reading measure, plus the intro/card-row gap. CSS-only, same cache-busting reasoning as every entry above. 0.6.12 — batch fix after first staging deploy (feedback 2026-08-19): rules TOC on the live game_extra rules route (reading-measure reach- through; the F12 table-border rule already covered it via the pre-existing inc/eg-prose-wrap.php wrap, verified live, not changed), archhero-to-grid/facets gap on /games/, /puzzles/, /decks/, dlgrid/ actions-secondary margin-block + 3-up columns on score-sheets, footer newsletter heading scoped back down off the theme.json h2 default, home "All seven games" heading padding-breakpoint parity with the grid/hero below it. CSS + templates only, same cache-busting reasoning as every entry above. 0.7.0 — MINOR, not a patch: the 2026-08-19 coder handoff (LOCKED web library) lands as a design-system step, not a fix. tokens.css and web-blocks.css are replaced wholesale from the drop; the web type scale becomes FLUID (T2 ruling, scale C: h1 40px at 360 to 54px at 1280, display 40-68, h2 24-34, h3 20-25, lead 19-24) and theme.json carries the same clamps with the designer's cqw written as vw, replacing the interim 4vw/3vw/2vw clamps 0.6.10 shipped. N6-page-head arrives as the ONE page top (.eg-pagehead, supersedes the .eg-archhero band); F2's game strip gains a linked red value and a two-up grid below 640; the mark law's struck name is retired (.eg-facet--none .eg-facet-name is now `color: inherit`, so a no-results option is quiet ink and a dead pointer and nothing else). Every one of those is a returning-visitor cache risk on its own, and the header logo silently grows from 20px to 25px on desktop because .eg-logo-text rides --eg-w-fs-h3 (deliberately NOT pinned, per the T2 logo-rider-B ruling). 0.7.4 — Slovenian presentation merge (Greg's ruling 2026-08-20, ticket S): the three-player game stops being a page and becomes a ?players=3 mode of /games/slovenian-tarok/, so the published count goes 7 to 6 and the templates here lose a card each. NO CSS in this file changed — the switch reuses the A2 facet chip wholesale and its placement rules live in el-blocks/blocks/single-game/style.css, cache-busted by that block's own block.json version (0.1.0 to 0.2.0). The bump is here because front-page.html and games-archive.html changed their copy ("Seven games, one deck" to "Six", "All seven games" to "All six", the "7 games" eyebrow to "6 games") and a returning visitor holding a cached page should not be the last one still counting seven. 0.7.5 — the C family lands (drop R4b, 2026-08-22, ticket CT): C1 contract table, C2 scope filter, C3 contract disclosures. assets/design/web-blocks.css was re-cut from the drop's blocks.css with scripts/split_web_bundle.py and grew by exactly 50 rules, all of them .eg-ctw* / .eg-scopes* / .eg-cdetail* (0 removed, 0 changed — the rest of the layer is byte-for-byte what 0.7.0 installed, and tokens.css did not move at all). NO CSS in this file changed: the block ships its own two-rule style.css for the one thing the designer layer cannot carry (a filtered hidden with [hidden] needs to beat the below-640 `display: block` stacking rule). The bump is here because web-blocks.css itself is enqueued from this file's ?ver= and a returning reader holding the cached copy would get the new markup with none of its styling. 0.7.7 — the Z family lands as A8/A9 (drop R4d, 2026-08-22, ticket PB): A8 category-band, A9 category-intro. The /puzzles/ archive and every per-game puzzles page now build their five type bands from these units instead of the interim F22/F23/F15 markup: an A9 head (live COUNT, the ratified one-sentence LINE from eg_puzzle_type_meta(), an ALL link) atop an A8 row that SAMPLES the type at most 8 cards deep with a tail plate, never the whole type. assets/design/web-blocks.css was re-cut from the drop's blocks.css with scripts/split_web_bundle.py: 26 selectors added (.eg-qband*, .eg-qintro*, .eg-qmore*), 0 removed, 0 changed — nothing outside the Z family moved. tokens.css gains exactly one property inside its existing .eg-web block, --eg-w-qcard (21rem, measured not rounded); the drop's own tokens.css carries a much wider unrelated diff (print- layer type scale / heading-spacing tokens already ratified 2026-08-19 but never installed here) that is NOT taken by this bump — flagged in DROP-INTEGRATION-STATUS.md instead, since it rides on a different ratification than this ticket's write set covers. The bump is here because web-blocks.css and tokens.css are both enqueued from this file's ?ver=. Interim .puzzle-grid__band* and .game-puzzle-list__type-group* block CSS retired (blocks/puzzle-grid/style.css, blocks/game-puzzle-list/style.css) along with the markup it styled; the interim .eg-cmp-scroll trailing-edge fade is gone from these rows with it (MANIFEST.md law: a faded live link reads as a disabled one). 0.7.10 — PM coder update of 2026-08-30 applied (ticket CU). Two halves. (a) The book masters are re-run through scripts/build_book_content.py, so derived/book-content/ and el-blocks/assets/book-content/ pick up the three Calling the King contract renames (Farbendreier "No-Trump", Farbensolo "No-Trump Solo", Sechserdreier "Blind Solo" — "Blind Six" and "Grand No-Trump Solo" are gone), King Ultimo announced by the called King's holder, the Forehand's opening pass, Strawman losing its 35:35 draw and Calling the Twenty losing "forced Solo" from Expansions. EG_BOOK_IMPORT_VERSION goes 5 -> 6 in the same commit because that content is baked into post_content at import and cannot self-heal. (b) Difficulty leaves every reader-facing puzzle surface (ruling §2: the reader sees Example or Puzzle, never a pip or the words easy/tricky). blocks/game-puzzle-list/render.php drops the eg-badge--diff dot badge from the A7 card foot; blocks/puzzle-board/render.php drops the eg-wpuzzle-diff dots from the read-reveal bar and the five-star el-stars mark from the legacy header, and stops reading _puzzle_difficulty entirely. The meta and its admin column/metabox in includes/cpt-puzzle.php are untouched — it is internal editorial data. The GAME facts strip keeps its own difficulty dots (§7 leaves that ruling alone); this is a different mark on a different surface. No CSS changed: .eg-badge--diff / .eg-wpuzzle-diff / .el-stars rules stay in place, unreferenced by these two blocks, because .el-stars is still used by the _partials preview. The bump is here because this is a reader-facing markup change on the puzzle pages and a cached copy would keep showing the retired pips. 0.7.11 — Example/Puzzle tier lands (ticket TR, PM ruling 2026-08-30 §2, item 2 unblocked). scripts/build_book_content.py derives a fixed 'example'|'puzzle' tier per puzzle straight from its source difficulty string (beginner-friendly/easy -> example, real-play/tricky -> puzzle; numeric fallback 1 -> example, 2/3 -> puzzle) into puzzles.json; inc/book-content-import.php carries it to _puzzle_tier meta (EG_BOOK_IMPORT_VERSION 6 -> 7, since existing puzzle rows otherwise pass the unchanged-content check and never pick up the new field). includes/cpt-puzzle.php gets the one shared reader, eg_puzzle_tier() (falls back to _puzzle_difficulty when _puzzle_tier isn't written yet), and eg_puzzle_tier_label(). blocks/puzzle-board/render.php's read-reveal shell splits on it: Example renders the answer open beside the question with no
/reveal control at all; Puzzle keeps exactly today's closed
reveal (so the analytics reveal event, untouched, only ever fires on Puzzles now — correct per ruling, Examples have nothing to reveal). Both blocks/puzzle-board/render.php's eg-wpuzzle-bar and blocks/game-puzzle-list/render.php's eg-qcard-foot gain a second eg-badge showing the tier word beside the existing id/type badge. Existing .eg-badge / .eg-web styling only, no new CSS. Regenerated bundle: 120 example / 216 puzzle across 336 puzzles. 0.7.12 — DROP-full-library-2026-08-30 taken (ticket FL). Standing web/design-delivery/web-library/ refreshed from the drop (fixes the stale, pre-R5b tokens.css MANIFEST flags as stale-list item 1; the standing blocks.css was also a leftover combined book+web copy, replaced by the drop's clean pre-cut web bundle). assets/design/web-blocks.css re-cut via scripts/split_web_bundle.py: 681 rules, 1 selector added (.eg-web table td strong + em, the native-names-on-tables rule), 0 removed, K1 hand centring (.eg-web .eg-hand justify-content: center) confirmed present, matching MANIFEST exactly. The drop's tokens.css carries a much wider diff (R5b book-scale restore + four sticky-colour tokens + C10 vign-below), all confirmed .eg-book-scoped and unreferenced by web-blocks.css (MANIFEST says as much); not taken into this theme's tokens.css, print-layer work out of this ticket's scope. Eleven corrected card-preview faces re-pulled and normalised (card-back, clubs-cavalier, diamonds-1, hearts-1/2, hearts-king, spades-10/queen, tarok-9/13/18); same paths, no template change. Known discrepancy (not fixed, per MANIFEST + ticket): the drop's own web/puzzle-card.html specimen still shows a DIFFICULTY dots slot; the site stays ahead of it (0.7.11's tier badges, zero pips), confirmed unchanged by this drop via PHP harness on blocks/game-puzzle-list/render.php. 0.7.35 — the F1 secondary tiles burst out of their own grid tracks at tablet width. The library fixes the column count at four past its 640 container query, but the primary tile takes about half that row, so the four tracks compute to 94px while a label needs about 110px; the tiles then overflowed their borders and the container. 0.7.34's fifth tile made it plainly visible. The column count now follows the width, and because auto-fit collapses unused tracks, 1280 still shows all five in one row. Scoped to the same 640 container tier the library's own four-column rule uses: unscoped it also reached the narrow tier and collapsed the phone layout from two tiles per row to one, which measurement caught. 0.7.34 — the ruled name lands (coder addendum 2026-09-17, section D6, standing since 2026-08-30). Reader copy says scorekeeper, and the route is /games/{id}/scorekeeper/ with a permanent redirect from the old /calculator/ address. The printed book was the risk: six pointer targets in assets/book-pointers.json resolved to /calculator, and a printed QR goes through that map, so the map moved in the same commit and nothing printed breaks. Existing game_extra posts migrate their slug on scaffold sync (version 23). Also section D4: every game hub now links to its variants page, labelled "More rules and expansions", which is the label the designer is moving to in round R10, and house law C5 applied to two of our own strings that claimed the site holds "the full" catalogue. Section D2: a review notice baked into a page is now hidden from readers and still shown to editors, so /traditional-scoring/ stops opening with a DRAFT box. 0.7.33 — page furniture was trapped in the reading measure (ticket W-E). One root cause, five symptoms: the extras hub's four action tiles rendered 87px wide with their labels painting outside the tile borders; the score sheets' download grid dropped to two tracks and wrapped the third card; the About pillars sat in 186px columns; and page titles hung 206px left of the body under them. .eg-actions and .eg-pillars now join .eg-dlgrid's existing :has() escape from the 44rem prose cap, and the two units that declare their own --eg-w-page-max are bled past main's padding so that cap can actually bind. Running prose keeps its measure — only furniture moved. The N6 band also reaches the four pages that were each inventing their own title treatment (extras, about, dictionary, glossary), which closes the title/body edge break as a side effect. Last arrow dingbat gone from 404. 0.7.32 — W-F. The glossary shipped literal markdown asterisks in six terms (the R3 parser ran definitions through the inline-markdown converter but not the terms, so a native word italicised in the master arrived as "(*Koenigrufen*)"), and all 64 terms were red underlined links pointing at their own row — clicking one did nothing. Terms now render their emphasis and are plain ink per the library; the row keeps its id so existing deep links still resolve. Score sheets: the "Also for this game" row was a bare secondary strip borrowing the E3 chooser's label class, with two tiles both typed reference; it is a complete F1 unit now (primary "Read full rules", then calculator and puzzles) and the pending-cheat-sheet note no longer touches the row below it. EG_BOOK_SCAFFOLD_VERSION 22. 0.7.31 — the rules pages were unreadable and their navigation unreachable (ticket W-C). Paragraph rhythm: the design layer's rule is a CHILD combinator (.eg-prose > * + *) but WordPress inserts an .entry-content wrapper, so every paragraph on an 11,661px page had 0px of separation and headings butted their first paragraph; the rhythm now reaches through the wrapper using the library's own --eg-w-sp-5. Sticky TOC: body carried overflow-x:hidden, which per spec forces overflow-y:auto and made BODY the nearest sticky ancestor instead of the viewport, so the TOC scrolled away and left a 288px column empty for 11,000px; html already guards the horizontal axis, so the body rule is gone and the TOC clamps at 32px. Mobile gutter: the page-head's 360 tier used a different spacing token from every body unit (36px vs 52px); one token now. Tap targets brought to the library's 44px minimum: TOC summary and links, scope chips, a band's "all" link, deck game names — padding only, type sizes untouched. 0.7.30 — second Teams harvest pass: each chapter's Cheat Sheet ladder is read as well as its core ladder, which gives zwanzigerrufen a Teams column it did not have (3 of 8 contracts). Contracts with no structured data anywhere in the masters keep an honestly empty cell — koenigrufen's eight advanced contracts have no Teams field in any chapter section, so no value was invented for them. EG_BOOK_IMPORT_VERSION 13. 0.7.29 — regression sweep of 23 pages x 2 widths after the fix wave. Two residues closed: the /puzzles/ landing still emitted its page-head band as
(the block half of the 0.7.26/0.7.27 landmark fix — the single-game and single-deck bands stay
because theirs sit inside an
, where that is correct); and the About page's one em dash turned out to live in the SSOT at ssot/company/brand.json (a positioning pillar's summary), not in the page, so it is fixed at source and the bundle re-synced. Book-side, reported not patched: prose arrows survive in generated content ("Normal Game > Farberl > Solo" on /start20/, "(void > play a Tarok)" on rules pages) — the masters own that wording. 0.7.28 — the contract table gains a TEAMS column (2 vs 2 / 1 vs 3 / all vs all). W-A's duplicate-table fix had removed the chapter ladder that carried it, and no surface reproduced it — while the PM ruling of 2026-08-30 makes it load-bearing ("the teams column on every contract table carries the difference" between a partner No-Trump and a solo one). The data is now harvested from the chapter's own ladder into a new contracts-meta.json bundle asset and joined to rows by SSOT contract id, never by row order; a contract with no harvested value renders an empty cell rather than a guess, and a game whose ladder carries no team data (zwanzigerrufen) renders no column at all. EG_BOOK_IMPORT_VERSION 12. 0.7.27 — the el/page-head BLOCK also emitted its band as
, giving every game_extra page a second banner landmark (the archive templates were fixed in 0.7.26; this is the dynamic-block half of the same defect). A
whose nearest sectioning root is is a banner, and this band sits directly in
. It is a page title, so it is a
now. 0.7.26 — two residues from 0.7.25's frame work, caught on live re-measure. (a) The /games/ and /decks/ archive templates rendered the page-head band with tagName "header", so every archive carried a SECOND banner landmark beside the site header; the band is a page title, not a site banner, so it is a plain group now. (b) The band still floated 24px below the header: 0.7.25 cancelled main's 48px padding but not the 24px margin-block-start WordPress's constrained layout puts on main itself. 0.7.25 — the design-review fix wave (tickets H2, W-A, W-B). H2, theme frame: header, main and footer now share ONE frame (main was on 72rem while every other .eg-web frame used --eg-w-page-max 1180px, which is why the logo hung left of the H1); .eg-feature bleeds edge to edge like the page head, so its red rule reads as a section edge instead of a floating stripe; game cards are white again (a stale AA hotfix used a substring selector, [class*="game-card"], that clobbered .eg-gcard's own --eg-card-white); duplicate banner/contentinfo landmarks removed (every template-part carried tagName, so WP wrapped the part's own
in a second one) and the duplicate id="newsletter"; the newsletter privacy link now passes AA at 4.97:1 (it lost to .eg-footer a on source order and rendered paper-on-paper at 1.11:1); the page-head band meets the header instead of floating under 48px of white. W-A, content: the rules pages printed the contract table TWICE (the chapter's own ladder table beside the C1 block) — suppressed at build time; native parentheticals can no longer break across three lines; puzzle titles drop the em dash and stop calling an Example a Puzzle ("Calling the King: Example C1"). W-B, game page: contract group headings carry the house band rule (2px ink, upright caps) instead of an invented italic label, and their names now match the rules page (English-leading, book spelling). 0.7.24 — arrow dingbats removed from reader-facing copy. The design library uses NO arrows in any web template (its pager unit is explicit: "Plain links; 44px minimum target"), yet the puzzle prev/next nav shipped left/right arrows, the game_extra breadcrumb shipped a left arrow, and the community hub shipped "Browse " + arrow. Words carry it: the breadcrumb now reads "Back to {Game}". Also gives that breadcrumb link the library's 44px minimum target (it measured 17px tall live). 0.7.23 — the footer newsletter form is now the library's G8 email-CTA unit (ticket H1) instead of a hand-rolled lookalike: bordered paper-2 panel, 48px controls (was 21px browser default, below the 44px touch minimum), Inter uppercase label on --eg-track-caps, and the house privacy line, which the lookalike omitted entirely. view.js follows G8's own structure (STATUS is a sibling of the form, not a descendant). 0.7.22 — page-reference stripping left visible wreckage in reader prose. A page ref never stands alone in the masters: it is parenthesised, fenced by commas, or introduced by a preposition. The build stripped only the "p. XX" token, so every rules page shipped "are at the back,, for tables" and koenigrufen shipped "the contract table on tells you what each contract takes"; the glossary shipped orphaned trailing commas after "See also:". scripts/build_book_content.py now removes each wrapper WITH its reference (longest form first) and tidies doubled/orphaned commas; unit-checked against all three shapes plus two no-op sentences. Content regenerated, EG_BOOK_IMPORT_VERSION 10 forces the re-import. 0.7.21 — the newsletter block's stylesheet was never enqueued: its block.json declared render + viewScript but no `style` key, so blocks/newsletter-signup/style.css (which exists and is correct) never loaded. Two live consequences on every page's footer form, both found in the 2026-08-30 critical homepage review: the form rendered as raw browser chrome (21px inputs against the 48px --eg-control-h token), and the ANTI-BOT HONEYPOT — hidden by that stylesheet with position/left, never display:none — rendered as a visible 153px text field directly above the real email input, so a human filling it had their signup silently rejected as bot traffic. Both block.json copies now declare the style. 0.7.20 — HOTFIX for 0.7.18: the SEOPress social filters (og/twitter title + description) echo their value VERBATIM, so the bare strings 0.7.18 returned printed as visible text above the site header on every route seo-meta.php owns (home, games, decks, glossary, puzzles) — caught by Greg on the home page, "lines without context in the top". Social filters now return complete tags; the titles_* filters keep taking bare strings, which is their own separate contract. 0.7.19 — crawl rev 3 wave 2 (tickets R4/R5). R4: slovenian facts strip stamps the inactive mode hidden server-side (the "Players 4 3" no-JS double-print); the engine-dump Summary paragraph (game.json description) is off every game page — the Greg-supplied standfirst carries that job; contract-band derivation unified into inc/contract-bands.php (the 3p/4p bucket split is genuine SSOT variant data and stays). R5: contract reader-prose map extended to all six games (26 new entries, 3 documented book-side gaps); slovenian native-vs-Simplified VP split captioned at page and per-contract level per the book's own ruling; difficulty glyphs become eg-dot markup at build time (meta-description leak closed); EG_BOOK_IMPORT_VERSION 9 forces the re-import. 0.7.18 — external crawl rev 2/3 wave 1 (tickets R1/R2/R3 + orchestrator fix-ups). R1: SEO title/canonical/description fixed on virtual routes (games, decks, glossary, puzzles archive) via SEOPress filters — root cause was template_include routing that never set query flags, so SEOPress saw every virtual route as the blog index; em dashes retired from reader-facing site copy (10 strings). R2: contract-table reader prose replaces SSOT engine notes (koenigrufen map, hard leak filter for all games), C2 scope filter now removes bands server-side (works with JS off, chips are plain links), English-leading contract names. R3: /glossary/ now syncs from the book's 10-glossary.md (59 entries) instead of the engine vocabulary; /dictionary/ drops VERIFY/placeholder/dash rows; per-game puzzle pages drop the vestigial same-page ALL links. Orchestrator: advanced contract badges on game pages now link with ?scope=advanced (the default rules view no longer shows that band), and the merged "Advanced / expansion contracts" heading reads "Advanced contracts" (no expansion-scope contract exists). 0.7.17 — CP: Greg's 14 coder pointers (2026-08-30) applied to el/score-keeper-koenigrufen, CALC-LAB-C2 ported (drop drop-calculator-lab-c2-2026-08-30). Sheet now prints ONE VP number per player per hand (CON/BON are entry-only fields on the Record-a-hand form) with a bold red-rule TOTAL foot row; no pre-printed empty rows; empty-state copy lifted from the lab; exactly two actions (Record a hand / New game), entry form closed by default; "First to 50 victory points" fact removed from the block; live per-row Hand VP while typing; a − sign-toggle button beside every numeric field plus typed minus, always rendering the true minus sign (U+2212); cells sized for a signed three-digit number, right-aligned, tabular figures; buttons are 4px radius (never pill), primary red-deep / secondary ink outline, Cancel never primary; sheet cells are DISPLAY only (paper table, hairline rules, paper-2 head/foot bands, 1.6px red head rule); below 640px the sheet restructures into per-player strips, two up (lab's .c2-strips copied wholesale), zero horizontal scroll at 360px with four players and −215. Values now reference library tokens (var(--eg-…)) instead of the lab's standalone copies. The block lives in el-blocks (own block.json version 0.1.0 → 0.2.0, its own cache-bust path); this theme-changelog entry is the record of intent per ticket CP, not a claim that this file's ?ver= gates the block's CSS. 0.7.16 — the theme stylesheet was DEAD since 0.7.7: the 0.7.7 changelog prose contained selector globs with the character pair star-slash (".eg-qband" + star + slash + ...) which terminates a CSS comment early; the exposed prose mis-parsed and the browser dropped EVERY rule in this file (styleSheets entry: 0 cssRules, verified via CDP on staging). Fixed by rewriting the two prose lines without the pair. Never write bare glob chains like x-star-slash-y-star inside CSS comments. Also: the revived 0.7.7-era .eg-pagehead bleed needed !important physical margins to beat WP constrained layout's margin auto !important on game_extra tops; the redundant 0.7.15 duplicate rule is removed in favour of the original. 0.7.15 — N6 page-head full bleed (ruled 2026-08-19, T1a: "the paper-2 band paints full bleed"). WP constrained layout was boxing .eg-pagehead to the content column (900px on game_extra, 1200px on singles/archives) so the band read as a rectangle around the headline (Greg, 2026-08-30). Theme reach-through only — the design layer is untouched: container-unit bleed (body.eg-web is the inline-size container; cqw excludes the scrollbar where vw would not), max-width freed from WP's clamp. The formula self-cancels when the parent is already full width. 0.7.14 — score-keeper polish after Greg's live review: legend copy simplified (CON "Contract points", BON "Bonus points" — the made/failed and trull/ultimo/valat wording belonged to the deleted rules engine); the "＋ Record a hand" button now toggles to "Cancel" while the entry panel is open, so only one live action path shows at a time. 0.7.13 — score keeper rebuilt to a dumb sheet (Greg's ruling 2026-08-30, ticket SK). el/score-keeper-koenigrufen: setup now asks only the four player names (no contract-table question, no book/house mode split); recording a hand asks CON and BON per player and the sheet sums them to VP, replacing the deleted declarer picker, contract dropdown, card-point readout and bonus checkboxes. render.php's config shrinks to game/name/seats/sessionEnd — no more embedded SSOT contract table. view.js's .eg-kc-table drops rowspan (one row per hand, three cells per player) and gains data-label attributes so style.css can reflow it into stacked hand cards below 640px with zero horizontal scroll, instead of forcing the old fixed 34rem min-width. NO CSS in THIS file changed — the block ships its own style.css, cache-busted independently, and its view.js is cache-busted via its own index.asset.php version (egsk010 -> egsk011). The bump is here per ticket SK's instruction to keep this file's version counter moving with the feedback round (0.7.12 already taken by ticket FL). 0.7.8 — /puzzles/ becomes a landing (Greg's ruling, 2026-08-22, ticket PZ, supersedes 0.7.7's category-destination IA on the archive): the 0.7.7 A8/A9 wiring stays live on the per-game pages, uncapped (no sample, no tail plate, tail hrefs repointed from a /games/{id}/puzzles/ {type}/ route that never existed to a same-page #type-{slug} anchor); el/puzzle-grid (blocks/puzzle-grid/render.php) drops its A2 facet rail, A3 chips, A4 result bar, and all A8 card rows and instead lists one A9 intro entry per game with a published puzzle, in the book's order, linking to that game's own puzzles page — plus a new type legend (eg-deflist, the same markup el/glossary-list already emits) so Decision/Exchange/Play/Counting/Ruling are explained before a reader hits them. No CSS in web-blocks.css or tokens.css changed: every class the new markup uses (eg-qband, eg-qintro*, eg-deflist) already existed for this exact purpose, and blocks/puzzle-grid/style.css and blocks/game-puzzle-list/style.css only lose dead rules for markup render.php no longer emits. The bump is here anyway because this is a reader-facing markup change on the site's most-linked page (a cached /puzzles/ copy would keep showing the retired facet rail) — same cache-busting reasoning as every prior entry. 0.7.6 — the E family lands (drop R4c, 2026-08-22, ticket PR): E1 download-card, E2 downloads-grid, as amended by Greg's generic-icon ruling the same day. assets/design/web-blocks.css was re-cut from the drop's blocks.css with scripts/split_web_bundle.py: 4 selectors added (the eg-dl--form/eg-dl--ref thumb variants), 1 removed (.eg-dl-thumbpending, no longer needed once every card always has a glyph), 4 changed (.eg-dl base padding, .eg-dl-thumb/-thumbimg simplified to full-width object-fit:contain, .eg-dl-body gains its own horizontal padding, .eg-dlgrid's track sizing). The old 640-and-up side-by-side @container block for .eg-dl/.eg-dl-body is gone outright (the card stacks at every width now). Nothing outside the .eg-dl and .eg-dlgrid selectors moved; tokens.css did not change. Also added assets/design/miniatures/ sheet-generic.svg (new asset directory, SVG only, no normalize_web_assets budget needed) and retired two interim theme overrides this drop obsoletes (see "Printables unit" below for the line-by-line justification of what stayed and what left). The bump is here because web-blocks.css itself is enqueued from this file's ?ver=. 0.6.13 — verifier findings: archhero gap rule needed body.eg-web prefix (web-blocks.css's later `.eg-web .eg-grid { margin: 0 auto }` tied and won on order — measured margin-top 0 on /decks/ and /games/); rules TOC desktop open state moved to view.js (closed
content cannot be shown by CSS; Chromium uses content-visibility). */ /* ------------------------------------------------------------------------ Base — mobile-first defaults. ------------------------------------------------------------------------ */ /* Border-box everywhere — without this, blocks that combine width:100% with their own padding (e.g. WP inner alignfull groups) overflow their parent at narrow viewports. Critical for 360px mobile. */ *, *::before, *::after { box-sizing: border-box; } html { font-size: 16px; text-rendering: optimizeLegibility; -webkit-font-smoothing: antialiased; /* Belt-and-braces: prevent any rogue child from forcing a horizontal scrollbar on the document. Deliberately the ONLY overflow-x guard in this pair (see the removed body rule below, W-C finding 2) — the root element's own overflow, when not visible, is what the UA propagates to the real viewport scrollport per spec; that is enough on its own for the belt-and-braces goal. */ overflow-x: hidden; } body { margin: 0; font-family: var(--wp--preset--font-family--serif, Cormorant Garamond, Georgia, serif); font-size: 17px; line-height: 1.6; color: var(--wp--preset--color--ink, #1a1a1a); background: var(--wp--preset--color--paper, #faf8f3); /* W-C finding 2 (blind review 2026-08-30): this body-level overflow-x: hidden was redundant with html's own (above) for the horizontal- scrollbar guard, but NOT harmless: per the CSS overflow spec, when one axis is non-visible the other computes to auto, so body's overflow-y silently became `auto` too — making BODY itself (whose box height auto-fits its content, so it has zero internal scroll range) the nearest "scrolling ancestor" for every position:sticky descendant on the page, ahead of the real viewport. That's why .eg-rules-toc (blocks/rules-toc/style.css) never stuck: verified live on koenigrufen/rules/ — getBoundingClientRect().top tracked window.scrollY 1:1 (391.9 / -2608.1 / -5608.1 at scrollY 0/3000/6000, an exact -3000/-3000 delta = static, not sticky) before this rule was removed; after removal (isolated test via an injected stylesheet, confirmed only then written here), top clamped to the intended 32px at every scrollY. html's own overflow-x:hidden above still guards against a horizontal scrollbar (confirmed no regression: document scrollWidth stayed <= clientWidth after this change), since the UA propagates the root element's own non-visible overflow to the viewport regardless of what body declares. max-width: 100vw (kept) is a plain box constraint, not an overflow declaration, so it does not re-create the scroll-container bug. */ max-width: 100vw; } /* Site title (header logo) — display serif. */ .wp-block-site-title { font-family: var(--wp--preset--font-family--display, Playfair Display, Georgia, serif); font-weight: 700; letter-spacing: -0.01em; } .wp-block-site-title a { color: var(--wp--preset--color--ink, #1a1a1a); text-decoration: none; } /* Navigation — sans-serif. */ .wp-block-navigation { font-family: var(--wp--preset--font-family--sans, Inter, system-ui, sans-serif); font-weight: 500; } .wp-block-navigation a { color: var(--wp--preset--color--ink, #1a1a1a); text-decoration: none; } .wp-block-navigation a:hover { text-decoration: underline; } /* Headings — display serif. */ h1, h2, h3, .wp-block-heading { font-family: var(--wp--preset--font-family--display, Playfair Display, Georgia, serif); font-weight: 700; line-height: 1.15; letter-spacing: -0.01em; margin: 1.5em 0 0.5em; text-wrap: balance; } h1, .wp-block-heading[class*="has-text-align-"] { margin-top: 0; } /* Body links. */ a { color: var(--wp--preset--color--red, #b22222); text-underline-offset: 2px; } /* ------------------------------------------------------------------------ Layout container — keeps content readable at all viewports. ------------------------------------------------------------------------ */ main.wp-block-group.alignfull > .wp-block-group:not(.alignfull), main.wp-block-group.alignfull > * { /* Constrain top-level main content to 1200px and add side padding */ max-width: 1200px; margin-left: auto; margin-right: auto; } /* Wide-screen: more breathing room. */ @media (min-width: 1500px) { main.wp-block-group.alignfull > * { max-width: 1200px; } } /* Mobile padding for any direct main child that doesn't have explicit padding. */ @media (max-width: 768px) { main.wp-block-group.alignfull > *, main.wp-block-group.alignfull { padding-left: 1.25rem; padding-right: 1.25rem; } } /* ------------------------------------------------------------------------ Header / footer (parts/header.html, parts/footer.html) — already use alignfull + their own padding; don't double up. ------------------------------------------------------------------------ */ header.wp-block-group.alignfull, footer.wp-block-group.alignfull { max-width: none; margin-left: 0; margin-right: 0; padding-left: 1.25rem; padding-right: 1.25rem; } @media (min-width: 768px) { header.wp-block-group.alignfull, footer.wp-block-group.alignfull { padding-left: 2rem; padding-right: 2rem; } } header.wp-block-group > .wp-block-group { max-width: 1200px; margin-left: auto; margin-right: auto; } /* ------------------------------------------------------------------------ Mobile nav — hamburger toggle. The Navigation block's mobile overlay needs visible toggle styling. ------------------------------------------------------------------------ */ @media (max-width: 768px) { .wp-block-navigation__responsive-container-open, .wp-block-navigation__responsive-container-close { color: var(--wp--preset--color--ink, #1a1a1a); } .wp-block-navigation__responsive-container.is-menu-open { background: var(--wp--preset--color--paper, #faf8f3); } } /* ------------------------------------------------------------------------ Pillars block — 3-column cards on desktop, stacked on mobile. Markup: section.el-pillars > ul.el-pillars__list > li.el-pillars__card ------------------------------------------------------------------------ */ section.el-pillars { width: 100%; margin: 0 auto; } ul.el-pillars__list { list-style: none; margin: 0; padding: 0; display: grid; gap: 1.5rem; grid-template-columns: 1fr; } @media (min-width: 720px) { ul.el-pillars__list { grid-template-columns: repeat(3, 1fr); gap: 2rem; } } li.el-pillars__card { padding: 1.75rem 1.5rem; background: var(--wp--preset--color--paper, #faf8f3); border: 1px solid var(--wp--preset--color--ink-3, #e0d8c8); border-radius: 8px; list-style: none; } .el-pillars__card-title { font-family: var(--wp--preset--font-family--display, Playfair Display, Georgia, serif); font-size: 1.25rem; font-weight: 700; margin: 0 0 0.5rem; color: var(--wp--preset--color--ink, #1a1a1a); } .el-pillars__card-summary { font-family: var(--wp--preset--font-family--serif, Cormorant Garamond, Georgia, serif); font-style: italic; font-size: 1rem; line-height: 1.5; color: var(--wp--preset--color--ink-2, #444); margin: 0 0 1rem; } ul.el-pillars__card-points { list-style: disc; margin: 0; padding-left: 1.25rem; font-family: var(--wp--preset--font-family--sans, Inter, system-ui, sans-serif); font-size: 14px; line-height: 1.6; color: var(--wp--preset--color--ink-2, #444); } ul.el-pillars__card-points li { margin-bottom: 0.25rem; } /* ------------------------------------------------------------------------ Brand hero — big tagline + centered CTAs. ------------------------------------------------------------------------ */ .el-brand-hero { padding: 3rem 1.25rem 2.5rem; background: var(--wp--preset--color--paper, #faf8f3); text-align: center; box-sizing: border-box; width: 100%; overflow-wrap: break-word; } @media (min-width: 768px) { .el-brand-hero { padding: 6rem 2rem 4rem; } } .el-brand-hero__inner { max-width: 800px; margin: 0 auto; padding: 0 0.5rem; box-sizing: border-box; } .el-brand-hero__tagline { font-family: var(--wp--preset--font-family--display, Playfair Display, Georgia, serif); font-size: clamp(1.6rem, 6vw, 4rem); font-weight: 800; line-height: 1.1; letter-spacing: -0.02em; margin: 0 0 1rem; color: var(--wp--preset--color--ink, #1a1a1a); text-wrap: balance; overflow-wrap: break-word; hyphens: auto; } /* Mobile: ensure long tagline never overflows the viewport. */ @media (max-width: 480px) { .el-brand-hero__tagline { font-size: 1.5rem; line-height: 1.2; letter-spacing: -0.01em; } .el-brand-hero { padding: 2rem 1rem 1.5rem; } .el-brand-hero__inner { padding: 0; } } @media (max-width: 380px) { .el-brand-hero__tagline { font-size: 1.35rem; } } /* Force every direct child of
to respect viewport width. */ main.wp-block-group.alignfull, main.wp-block-group.alignfull > * { box-sizing: border-box; max-width: 100%; } .el-brand-hero__mission { font-family: var(--wp--preset--font-family--serif, Cormorant Garamond, Georgia, serif); font-size: clamp(1.05rem, 2vw, 1.25rem); font-style: italic; color: var(--wp--preset--color--ink-2, #444); margin: 0 auto 2rem; max-width: 600px; line-height: 1.5; } .el-brand-hero__ctas { display: inline-flex; gap: 0.75rem; flex-wrap: wrap; justify-content: center; } .el-brand-hero__cta { display: inline-block; padding: 0.75rem 1.5rem; font-family: var(--wp--preset--font-family--sans, Inter, system-ui, sans-serif); font-size: 15px; font-weight: 500; text-decoration: none; border-radius: 6px; transition: transform 0.1s ease; } .el-brand-hero__cta:hover { transform: translateY(-1px); } .el-brand-hero__cta--primary { background: var(--wp--preset--color--ink, #1a1a1a); color: var(--wp--preset--color--paper, #faf8f3); } .el-brand-hero__cta--secondary { background: transparent; color: var(--wp--preset--color--ink, #1a1a1a); border: 1px solid var(--wp--preset--color--ink-3, #ccc); } /* ------------------------------------------------------------------------ Featured game — card with name + meta + flavor + CTA. ------------------------------------------------------------------------ */ .el-featured-game { text-align: center; padding: 2rem 1.25rem; } .el-featured-game__inner { max-width: 720px; margin: 0 auto; } .el-featured-game__label { font-family: var(--wp--preset--font-family--sans, Inter, system-ui, sans-serif); font-size: 12px; font-weight: 600; text-transform: uppercase; letter-spacing: 0.1em; color: var(--wp--preset--color--ink-3, #888); margin: 0 0 0.5rem; } .el-featured-game__name { font-family: var(--wp--preset--font-family--display, Playfair Display, Georgia, serif); font-size: clamp(1.75rem, 4vw, 2.5rem); font-weight: 700; margin: 0 0 0.5rem; } .el-featured-game__players { font-family: var(--wp--preset--font-family--sans, Inter, system-ui, sans-serif); font-size: 14px; color: var(--wp--preset--color--ink-2, #444); margin: 0 0 1rem; } .el-featured-game__flavour { font-family: var(--wp--preset--font-family--serif, Cormorant Garamond, Georgia, serif); font-style: italic; font-size: 1.1rem; color: var(--wp--preset--color--ink-2, #444); margin: 0 0 1.5rem; line-height: 1.5; } .el-featured-game__cta { display: inline-block; padding: 0.6rem 1.25rem; font-family: var(--wp--preset--font-family--sans, Inter, system-ui, sans-serif); font-size: 14px; font-weight: 500; background: var(--wp--preset--color--ink, #1a1a1a); color: var(--wp--preset--color--paper, #faf8f3); border-radius: 6px; text-decoration: none; } /* ------------------------------------------------------------------------ Game-card grid — games archive shows multiple cards. F4/F25 root cause: .eg-grid is a `wp:group` block rendered by WP core as `is-layout-flow` (its className points at a hand-written CSS-grid rule in web-blocks.css rather than a block-level "layout":{"type":"grid"} attribute), so WP's global styles still inject `:root :where(.is-layout-flow) > * { margin-block: 24px 0 }` on every child except the first (verified live: matched rule on the 2nd/3rd .eg-gcard, absent on the 1st). Under `display:grid` that pushes row 1's 2nd/3rd column siblings down 24px while the true first child (Strawman on /games/) keeps margin-top:0 and sits flush against the archive header band above it — off the baseline the rest of its row sits on. The grid's own `gap` (web-blocks.css) already provides consistent spacing, so drop WP's injected flow margin for grid children entirely, on both /games/ (.eg-gcard) and /decks/ (.eg-dcard) grids. The old `.el-game-card` / `.wp-block-el-game-card` rules that used to live in this section are removed, not kept alongside the fix: el/game- card's own co-located style.css already documents them as dead ("the old .el-game-card* rules matched no markup this block emits any more") — except `.wp-block-el-game-card` itself, which get_block_wrapper_ attributes() still emits on the real , so its `display:block` rule was silently overriding web-blocks.css's `.eg-gcard { display:flex }` the whole time (higher specificity: two classes under `main.wp-block-group >` beats `.eg-web .eg-gcard`). ------------------------------------------------------------------------ */ .eg-grid > .eg-gcard, .eg-grid > .eg-dcard { margin-block: 0; } /* Batch fix item 2 (feedback 2026-08-19): the SAME margin-block:0 fix above removed the only vertical gap that used to sit between .eg-archhero (the header band, .eg-web .eg-archhero in web-blocks.css — no bottom margin/ padding of its own) and the card grid or facet rail directly below it — /games/ and /decks/ use .eg-grid, /puzzles/ (el/puzzle-grid) uses .eg-facets next. Both are now flush against the band. Adjacent-sibling selector, no template change needed. */ /* Specificity note (verified live 2026-08-19): web-blocks.css loads AFTER this file and sets `.eg-web .eg-grid { margin: 0 auto; ... }` (0,2,0), which tied and beat the original `.eg-archhero + .eg-grid` (0,2,0) on order — computed margin-top was 0 on /decks/ and /games/. Prefix with body.eg-web (0,2,1) so the gap actually applies; horizontal auto centering from web-blocks.css is preserved because only margin-top is set here. */ body.eg-web .eg-archhero + .eg-grid, body.eg-web .eg-archhero + .eg-facets { margin-top: var(--eg-w-sp-5, 2rem); } /* Batch fix item 2: .eg-archhero (the band's own box, background-color) has no width rule of its own — it's simply 100% of
's available width — while its child .eg-archhero-inner centers ITS content at max-width:var(--eg-w-page-max) with its own horizontal padding. Because --eg-w-page-max (1180px, tokens.css) never actually binds inside this theme's ~1088px main content column, .eg-archhero-inner's real limiting factor is that same padding, which .eg-grid's cards share too — so the coloured band background reads as wider than the text/cards it frames by exactly that padding, on both sides (most visible on the right, against the grid's last card). Background-color paints the full border-box regardless of padding, so padding alone can't shrink what's visible — shrink the box itself (width + auto margins) by the same inset instead, then pull .eg-archhero-inner back out by that same amount so its own padding lands the TEXT exactly where it already sat: only the visible background box moves, not the copy. Same tokens/breakpoint as .eg-archhero-inner / .eg-grid (web-blocks.css) so the two stay in lock-step at every width. */ /* body.eg-web prefix on both: web-blocks.css's own `.eg-web .eg-archhero- inner` rule (2 classes) already sets `margin: 0 auto`, so a same- specificity override here would lose on source order alone — the extra type selector matches this file's existing override convention (see the game-card / puzzle-teaser sections above) rather than reaching for !important. */ body.eg-web .eg-archhero { width: calc(100% - 2 * var(--eg-w-sp-4, 1rem)); margin-inline: auto; } body.eg-web .eg-archhero-inner { margin-inline: calc(var(--eg-w-sp-4, 1rem) * -1); } @container eg-web (min-width: 640px) { body.eg-web .eg-archhero { width: calc(100% - 2 * var(--eg-w-sp-5, 1.5rem)); } body.eg-web .eg-archhero-inner { margin-inline: calc(var(--eg-w-sp-5, 1.5rem) * -1); } } /* The two .eg-archhero blocks above now serve ONE page: the (currently drafted) blog index, el/blog-archive-hero. /games/, /decks/, /puzzles/ and the game and deck singles moved to N6-page-head below on 2026-08-19. They stay because A1 is still a live block in the library, not because anything in this theme's shipped nav reaches it today. */ /* ------------------------------------------------------------------------ N6 page head — the ONE page top (drop 2026-08-19, T1a, LOCKED tier). ------------------------------------------------------------------------ */ /* The ruling is "paper-2 band paints FULL BLEED across eg-web; text keeps the grid left edge" — the exact opposite of what the A1 band above does, where we SHRANK the band to stop it out-reaching the cards. So the band escapes the content column instead, and the inner puts the text back on the grid. Both halves are box arithmetic against this theme's own content column (`.eg-web .wp-site-blocks > main.wp-block-group`, further up this file: max-width 72rem, margin-inline auto, padding-inline clamp(1rem,4vw,2.5rem)), not a new layout: BAND. main is centred with symmetric padding, so the horizontal centre of any block inside it is the viewport centre; `margin-inline: calc(50% - 50vw)` therefore lands the band's left edge at 0 and `width: 100vw` carries it to the far edge, at every viewport, whether the page top is a direct child of main (the archive templates) or sits inside the block's own
(el/single-game, el/single-deck, el/puzzle-grid). max-width:none is needed because WP's constrained layout caps every child at contentSize. TEXT. The inner then re-creates the column the cards sit in: the same 72rem cap, and padding = main's own inline padding PLUS the sp-4/sp-5 the design system already gives .eg-grid / .eg-pagehead-inner. Because 72rem (1152px) is below both --eg-w-page-max (1180) and WP's contentSize, the grid's own caps never bind and the two left edges agree at every width: below 1152 both are padding + sp-N from the viewport edge, above it both are (100vw - 1152)/2 + padding + sp-N. Only padding-inline is set, so the drop keeps ownership of the band's vertical rhythm (sp-6/sp-5, and sp-7/sp-6 in its own >=640 container rule). body.eg-web prefix for the usual source-order reason: web-blocks.css loads after this file and its own `.eg-web .eg-pagehead-inner` is the same two-class specificity. */ /* When the page head IS the first thing in main, the band must meet the header. 0.7.25 cancelled main's 48px padding-top, but WordPress's own constrained-layout rule (:root :where(.is-layout-constrained) > * { margin-block-start: 24px }) still pushed main itself down, leaving a white strip above a full-bleed band. Measured live 2026-08-30: header bottom 65, band top 89. Dynamic blocks (single-game, single-deck) wrap the band one level inside their own wrapper, so both shapes are matched. */ body.eg-web .wp-site-blocks > main.wp-block-group:has(> .eg-pagehead:first-child), body.eg-web .wp-site-blocks > main.wp-block-group:has(> :first-child > .eg-pagehead:first-child) { margin-top: 0; } /* F1 secondary tiles: column COUNT must follow the available width, not be fixed. The library pins four columns once the container passes 640 (web-blocks.css, the eg-web 640 container query). At 768 the primary tile already takes about half the row, leaving the secondary grid roughly 410px, so four tracks compute to 94px each while a label like "Scorekeeper" has a min-content width near 110px: the tiles then burst out of their own tracks and past the container, which is visible on every game hub. Adding the fifth tile (the variants link, addendum section D4) made it worse, so the floor is set here rather than left to the library. auto-fit collapses the tracks it does not need, so 1280 still renders all five tiles in one row. Scoped to the same 640 container tier the library's four-column rule uses, so the narrow tier keeps its two-up layout. */ @container eg-web (min-width: 640px) { body.eg-web .eg-actions-secondary { grid-template-columns: repeat(auto-fit, minmax(9rem, 1fr)); } } body.eg-web .eg-pagehead { max-width: none; /* cqw, not vw: the eg-web container's width is the real layout width, so the bleed lands exactly on the viewport edges instead of overhanging by half the scrollbar (measured -7.5px with 50vw). */ width: 100cqw; /* physical longhands + !important: WP constrained layout puts margin-left/right: auto !important on direct children of a layout container (game_extra tops), which beats any non-important margin. */ margin-left: calc(50% - 50cqw) !important; margin-right: calc(50% - 50cqw) !important; /* The mobile media rule below (main > * padding at <=768px) would pad the band's own box on archive templates, shifting the inner text 20px right of the grid. The band's text inset belongs to -inner alone. */ padding-inline: 0; } /* H2 finding B (2026-08-30, coordinator addendum): "the paper-2 band paints full bleed" (N6's own RULED note above) reads as a band that starts the page, not one floating 72px below it. Header bottom to .eg-pagehead top measured 65/137 live; a local tokens.css+web-blocks.css+style.css harness reproducing header + main + pagehead (post finding-1/-4 fixes, so a single, unwrapped
) isolates the remaining piece to exactly main's own inline `padding-top: var(--wp--preset--spacing--7)` (48px, theme.json) — .eg-pagehead sat at main's own top padding edge with `margin-top: 0`, confirmed gap 48px before this rule, 0 after. Only templates whose FIRST rendered thing inside
is the page-head band lose their top inset — decks-archive.html / games-archive.html / single-game_extra.html render .eg-pagehead as main's direct child; el/single-deck and el/single-game (render.php) wrap it one level deeper in the block's own get_block_wrapper_attributes() element first (verified by reading render.php, not assumed), so the selector is a DESCENDANT combinator, not `>` — still `:first-child` so it only fires when .eg-pagehead is the first thing in ITS OWN parent (the true page top), not some unrelated nested occurrence. All five share the same 48px main padding-top, so one token-driven rule covers every one of them without a template-by-template edit. Templates with no page-head band (front page, plain pages, prose singles) never match `:first-child` here and keep the top spacing they already have. */ body.eg-web main.wp-block-group.alignfull .eg-pagehead:first-child { margin-top: calc(-1 * var(--wp--preset--spacing--7, 48px)); } /* Measured fix 2026-08-20 (verifier round 2, second pass): the band text column mirrors the CONTENT GRID's left edge, which is main's column (1200px cap + main padding) PLUS .eg-grid's own inline padding. Both components are viewport-dependent: main padding is spacing-6 above 768px and 1.25rem below (this file's own media query); the grid padding is eg-w-sp-4 below the 640px container width and sp-5 above. Rule order matters: the >=769px media rule is written LAST so it wins the 640-768 overlap against the container rule. W-C finding 3 (blind review 2026-08-30): the base tier's target no longer holds live — re-measured on koenigrufen/rules/, koenigrufen/, /games/ and /decks/ at 360px, every one of them the same way: the page title sat at x=36 while every body-content unit beneath it (.eg-gcard, .eg-dcard, .wp-block-paragraph in .eg-prose, .eg-actions, .el-contract-index — a different block on every page) sat at x=52, a consistent 16px step, mobile-only (1280px already matches at x=89 on both, untouched). This is universal across templates, not one page's local wrapper, so the general rule below is the right scope. 52px = main's own real padding-left (32px — WP's per-block "constrained layout" style, not this file's 1.25rem child rule, wins the cascade on
itself) + the theme's standing 1.25rem local inset every body-content wrapper still carries on top of it. Swapping this tier's second addend from eg-w-sp-4 (16px) to the same --wp--preset--spacing--6 (32px) token the >=769px tier below already uses reproduces that 52px exactly (measured after: title and body both x=52) without inventing a new constant. The >=640px container-query tier is untouched — out of this finding's measured scope. */ body.eg-web .eg-pagehead-inner { max-width: 1200px; padding-inline: calc(1.25rem + var(--wp--preset--spacing--6, 2rem)); } @container eg-web (min-width: 640px) { body.eg-web .eg-pagehead-inner { padding-inline: calc(1.25rem + var(--eg-w-sp-5, 1.5rem)); } } @media (min-width: 769px) { body.eg-web .eg-pagehead-inner { padding-inline: calc(var(--wp--preset--spacing--6, 2rem) + var(--eg-w-sp-5, 1.5rem)); } } /* Same gap the A1 band needed, for the same reason (.eg-pagehead declares no bottom margin of its own, and .eg-grid's `margin: 0 auto` in web-blocks.css ties on specificity and wins on order without the body prefix). */ body.eg-web .eg-pagehead + .eg-grid, body.eg-web .eg-pagehead + .eg-facets { margin-top: var(--eg-w-sp-5, 2rem); } /* Carried forward by hand from the designer's own web bundle: the 2026-08-19 drop used to ship one combined blocks.css, and scripts/split_web_bundle.py kept only selectors mentioning .eg-web, so the designer's two `html` rules (`scrollbar-gutter: stable` and the P1-P4 prompt scroll lock) were dropped on the way in. Since 2026-10-18 web-blocks.css is a verbatim copy of the canon web bundle and carries both rules itself; this declaration stays as the theme's own statement of the invariant (same value, so it is idempotent). scrollbar-gutter is load-bearing for the full-bleed band above: without it a page that gains a scrollbar reserves no gutter and 100vw overshoots the viewport by the scrollbar's width (body's overflow-x:hidden clips it, but the band still reads off-centre). The prompt lock (`html.eg-prompt-open { overflow: hidden }`) is inert until something sets that class; it belongs with P1-P4 when they are wired. */ html { scrollbar-gutter: stable; } /* ------------------------------------------------------------------------ H2 finding 2 (2026-08-30, blind design review) — .eg-feature full bleed. ------------------------------------------------------------------------ */ /* web-blocks.css gives `.eg-feature` no width rule of its own (only `.eg-feature-inner` gets `max-width: var(--eg-w-page-max); margin: 0 auto`), so as a plain child of
it inherited main's own box — 1180px + main's own clamp() inset (see the "Global content column" rule above) — and its 3px top rule read as a short stripe with paper visible past both ends instead of running the full width of the viewport, same as N6 .eg-pagehead did before that band got its own bleed treatment. SAME technique, reused rather than reinvented: escape main's column with `width: 100cqw` + physical `margin-left/right: calc(50% - 50cqw) !important` (cqw not vw for the same scrollbar-gutter reason the pagehead comment above documents; !important to beat WP constrained layout's `margin-left/right: auto !important` on direct main children). No padding-inline reset is needed here (unlike .eg-pagehead): .eg-feature itself declares no padding of its own — only .eg-feature-inner does, and that inner column already runs on the same `var(--eg-w-page-max)` frame finding 1 put main back onto, so it keeps the text where it already sat. */ body.eg-web .eg-feature { max-width: none; width: 100cqw; margin-left: calc(50% - 50cqw) !important; margin-right: calc(50% - 50cqw) !important; } /* ------------------------------------------------------------------------ Rule-list (rules page). ------------------------------------------------------------------------ */ .el-rule-entry { padding: 1rem 0; border-bottom: 1px solid var(--wp--preset--color--ink-3, #ddd); } .el-rule-entry:last-child { border-bottom: none; } .el-rule-entry__title { font-family: var(--wp--preset--font-family--display, Playfair Display, Georgia, serif); font-size: 1.25rem; margin: 0 0 0.5rem; } .el-rule-entry__title a { color: var(--wp--preset--color--ink, #1a1a1a); text-decoration: none; } .el-rule-entry__desc { color: var(--wp--preset--color--ink-2, #444); font-size: 0.95rem; } .el-rule-entry__scopes { list-style: none; padding: 0; margin: 0.5rem 0 0; display: inline-flex; gap: 0.5rem; flex-wrap: wrap; } .el-rule-entry__scope { display: inline-block; padding: 0.15rem 0.5rem; font-family: var(--wp--preset--font-family--sans, Inter, system-ui, sans-serif); font-size: 11px; font-weight: 500; text-transform: uppercase; letter-spacing: 0.05em; border-radius: 3px; background: var(--wp--preset--color--paper-2, #f0ebde); color: var(--wp--preset--color--ink-2, #444); } /* ------------------------------------------------------------------------ Footer columns — stack on mobile, 3-up on desktop (already controlled by wp:columns block markup; here we just refine). ------------------------------------------------------------------------ */ footer .wp-block-columns { margin-top: 0; margin-bottom: 0; } footer ul.is-style-no-bullets { list-style: none; padding-left: 0; margin: 0; } footer ul.is-style-no-bullets li { margin-bottom: 0.25rem; } footer a { color: var(--wp--preset--color--ink-2, #444); } /* Batch fix item 4 (feedback 2026-08-19): .el-newsletter-signup__heading (blocks/newsletter-signup/render.php) is a bare

with no CSS of its own anywhere (not in that block's style.css, not in web-blocks.css), so it's one of the few headings on the site that actually falls through to theme.json's global h2 element style — 0.6.10's h-N var fix (see the changelog note at the top of this file) made that style resolve correctly for the first time, at var(--wp--preset--font-size--h2)'s clamp max (2rem/32px). That reads oversized next to the rest of the footer (12px column labels, 14-17px body copy) — restore a footer- appropriate size without reverting the theme.json fix or touching newsletter-signup's own files (outside this ticket's write set). Reuses the h3 preset — WP kebab-cases the numeric-suffixed slug the same way (see the changelog note), so --wp--preset--font-size--h3 does not exist either; go straight to --h-3. */ .el-newsletter-signup__heading { font-size: var(--wp--preset--font-size--h-3, 1.25rem); } /* Batch fix item 5 (feedback 2026-08-19): .eg-sectionhead (web-blocks.css) never gains the >=640px sp-5 horizontal-padding bump that .eg-grid, .eg-pillars and .eg-hero-inner all get at the same container breakpoint — it stays at sp-4 (16px) forever, so on the home page "All seven games" sits 8px left of the hero copy and the card grid beneath it at desktop widths (measured on staging: hero/pillars/grid content all start at the same x, .eg-sectionhead-title starts 8px earlier). Match the same breakpoint the rest of the page-max sections use so the heading lines up with what it introduces. body.eg-web prefix for the same source-order reason as .eg-dlgrid above — web-blocks.css's own `.eg-web .eg-sectionhead` rule is the same 2-class specificity and loads after this file. */ @container eg-web (min-width: 640px) { body.eg-web .eg-sectionhead { padding-inline: var(--eg-w-sp-5); } } /* ------------------------------------------------------------------------ Page templates — apply container constraint to
. ------------------------------------------------------------------------ */ /* Global content column — every page's
sits in a centered, padded column instead of running edge-to-edge (templates render alignfull, which ignored theme.json contentSize). Header/footer are separate template parts with their own inner width, so they stay full-bleed. */ /* Specificity must beat WP's .wp-block-group.alignfull (0,2,0) — main itself carries alignfull, so a plain `.wp-site-blocks > main` (0,1,1) loses. */ /* H2 finding 1 (2026-08-30, blind design review): this box used its own ad hoc 72rem (1152px) cap, a leftover from before the .eg-web design system dropped in. Every OTHER .eg-web frame — .eg-header-inner, .eg-footer-inner, .eg-hero-inner, .eg-feature-inner, .eg-grid, .eg-pillars, .eg-listhdr, .eg-sectionhead, .eg-archhero-inner — uses the ONE page-max token instead: `max-width: var(--eg-w-page-max)` (1180px, tokens.css). 1152 vs 1180 (a 28px gap, doubled to ~56px of total drift once each side re-centers) is exactly the "header/footer sit on a different frame than main" gap measured live (box edges 57/1209 for main vs 43/1223 for header/footer @1280 — verified again with a local tokens.css+web-blocks.css+style.css harness at viewport 1280: .eg-web .wp-site-blocks > main computed left:56.5 / right:1208.5 / width:1152 before this fix). Swapping in the token makes main's own box land on the SAME 1180px frame header and footer already use — one constant, one source of truth, instead of two numbers that were meant to agree and drifted. This was already known to bite: the .eg-archhero bleed/pull-out hack further up this file exists ONLY because "--eg-w-page-max never actually binds inside this theme's ~1088px main content column" (its own comment) — with main now on the same token, that inner mismatch closes at the source instead of needing a per-band workaround. Padding-inline (the clamp below) is left as designed — it is a separate, already-tuned inset budget shared by .eg-pagehead-inner's own measured-fix formula (LOCKED tier), not respecified here. */ .eg-web .wp-site-blocks > main.wp-block-group { max-width: var(--eg-w-page-max, 1180px); margin-inline: auto; padding-inline: clamp(1rem, 4vw, 2.5rem); padding-bottom: 4rem; box-sizing: border-box; } /* Tighter reading measure for long-form single content (blog posts, pages). This wins over the block layout, which is usually what we want and once cost an afternoon: /family-tree/ declared contentSize 800 / wideSize 1100 on both the main group and post-content, the section correctly carried .alignwide, and it still rendered at 704px, because 44rem caps the post-content WRAPPER before any child alignment is considered. A page whose content is a table rather than prose opts out with .is-wide-page on its main (see templates/page-family-tree.html). */ .single .wp-site-blocks > main > :where(.wp-block-post-content, .wp-block-post-title), .page .wp-site-blocks > main > :where(.wp-block-post-content, .wp-block-post-title) { max-width: 44rem; margin-inline: auto; } /* The opt-out. The title keeps the reading measure; only the content escapes, so a wide table does not drag its heading out to 1100px with it. */ .wp-site-blocks > main.is-wide-page > .wp-block-post-content { max-width: none; } /* ------------------------------------------------------------------------ Rules TOC on game_extra (F11 batch fix, feedback 2026-08-19). ------------------------------------------------------------------------ */ /* Reach-through for templates/single-game_extra.html's .eg-rules-layout (F11 batch fix, feedback 2026-08-19): el/rules-toc placed the group between
and wp:post-content, so the direct-child selector above no longer matches post-content on ANY game_extra sub-page — restore the same narrow measure through the extra level of nesting for every page that isn't the rules chapter. Descendant (not direct-child) on the last hop: inc/eg-prose-wrap.php inserts its own `.eg-prose` div between .eg-rules-layout and wp:post-content's own output for every "prose" extra slug (rules, variants, kontra, holding, ...), so the real depth is 1 level for score-sheets/puzzles/calculator/extras but 2 for prose pages — .wp-block-post-content only ever appears once inside .eg-rules-layout either way, so the descendant combinator is exact, not a broad match. */ .single-game_extra .wp-site-blocks > main > .eg-rules-layout .wp-block-post-content { max-width: 44rem; margin-inline: auto; } /* The rules chapter itself opts back out: el/rules-toc only ever renders on the "rules" extra (blocks/rules-toc/render.php returns '' everywhere else), so :has(.eg-rules-toc) is true precisely when this IS the rules page — no second class or template needed. The chapter needs the wider column (~1090px) so the >=1024px sticky sidebar (blocks/rules-toc/ style.css) has room beside it instead of crushing it into 44rem. */ .single-game_extra .wp-site-blocks > main > .eg-rules-layout:has(.eg-rules-toc) .wp-block-post-content { max-width: none; margin-inline: 0; } /* W-C finding 1 (feedback 2026-08-19, blind review 2026-08-30): every rules page reads as one undifferentiated wall of text. Root cause: web-blocks.css (design layer, out of this ticket's write set) sets `.eg-web .eg-prose > * + * { margin-top: var(--eg-w-sp-5) }` — a CHILD combinator — but on the live game_extra "rules" route WordPress inserts a `.entry-content` wrapper between `.eg-prose` and the actual paragraphs/headings (inc/eg-prose-wrap.php wraps wp:post-content's own output), so the real DOM is `.eg-prose > .entry-content > p`, never `.eg-prose > p` directly — the rule's selector never matches and 10 consecutive paragraphs measured 0px apart. Measured live (koenigrufen/rules/, 1280px): before 0px gap / 0px margin-top; after 24px / 24px, matching --eg-w-sp-5's own value (not hardcoded — the var is used directly, no fallback needed since tokens.css defines it unconditionally ahead of this file). Added alongside the original selector (kept, for any .eg-prose usage without the wrapper) rather than replacing it — purely additive, no specificity fight, so no !important needed. This also closes the second half of the finding (headings butting against their first paragraph): h2/h3 are direct .entry-content children too, so a paragraph immediately after a heading is a `* + *` match and now gets the same 24px. */ .eg-web .eg-prose > .entry-content > * + * { margin-top: var(--eg-w-sp-5); } /* Re-skin our existing el-* CTA buttons to the designer's button treatment so they inherit the accessible palette (white on deep red passes AA; the old hero CTAs failed color-contrast under the new tokens). */ body.eg-web a[class*="__cta"], body.eg-web button[class*="__cta"] { display: inline-block; text-decoration: none !important; font-weight: 600; border-radius: var(--eg-btn-radius, 6px); padding: 0.7em 1.4em; line-height: 1.2; /* default (unmodified cta) = accessible primary; --secondary overrides */ background: var(--eg-btn-bg, #8a0f22); color: var(--eg-btn-fg, #fff) !important; border: 1.5px solid transparent; } body.eg-web a[class*="__cta--primary"], body.eg-web button[class*="__cta--primary"] { background: var(--eg-btn-bg, #8a0f22); color: var(--eg-btn-fg, #fff) !important; border: 1.5px solid transparent; } body.eg-web a[class*="__cta--primary"]:hover, body.eg-web button[class*="__cta--primary"]:hover { background: var(--eg-btn-bg-hover, #7A0619); } body.eg-web a[class*="__cta--secondary"], body.eg-web a[class*="__cta--ghost"], body.eg-web button[class*="__cta--secondary"] { background: transparent !important; color: var(--eg-btn2-fg, #2a2320) !important; border: 1.5px solid var(--eg-btn2-border, #2a2320); } /* Every OTHER action button (form submits, WP element buttons, block submit/reveal buttons) gets the accessible primary treatment — deep red + white — so no button anywhere renders the buggy red-text-on-dark. */ body.eg-web .wp-element-button, body.eg-web button[type="submit"], body.eg-web input[type="submit"], body.eg-web [class*="__submit"], body.eg-web [class*="__reveal-btn"], body.eg-web [class*="__answer-submit"] { background: var(--eg-btn-bg, #8a0f22) !important; color: var(--eg-btn-fg, #fff) !important; border: 1.5px solid transparent !important; text-decoration: none !important; border-radius: var(--eg-btn-radius, 6px); font-weight: 600; } body.eg-web .wp-element-button:hover, body.eg-web button[type="submit"]:hover, body.eg-web [class*="__submit"]:hover { background: var(--eg-btn-bg-hover, #7A0619) !important; } /* Never let the site link colour bleed into any button. */ body.eg-web button, body.eg-web .wp-element-button { text-decoration: none; } /* H2 finding 3 (2026-08-30, blind design review): this AA-contrast hotfix pre-dates el/game-card's H4-game-card rewrite (see the block's own style.css: "now emits H4-game-card markup ... eg-gcard, eg-gcard-media, eg-gcard-body, eg-gcard-title, eg-gcard-desc, eg-gcard-meta"). Its intent was the meta row's own fill ("4 players" text); its selector `[class*="game-card"]` is a substring match that ALSO catches the block's own core-generated wrapper class `wp-block-el-game-card` — the same class that carries `eg-gcard` — so it force-painted the WHOLE card paper (#FBF7F1) over web-blocks.css's `.eg-web .eg-gcard { background: var(--eg-card-white) }` (measured live: .eg-gcard computed rgb(251,247,241), the page ground, not rgb(255,255,255), the specimen's card white). The companion `__players`/`__meta` BEM selectors below never matched the rewritten markup either (dead code, same class the block's own style.css already flags) — dropped along with the background line rather than left as inert legacy selectors pointing at markup that no longer exists. If a meta-row AA fix is still needed for the current .eg-meta-item / .eg-meta-label markup, that is a fresh, separately-scoped fix, not a revival of this one. */ /* Dark cards (puzzle teaser has a near-black tabletop bg) rendered red text + red links = the red-on-black bug. Force light text + a light CTA inside any dark teaser/card so contrast passes. */ body.eg-web [class*="puzzle-teaser"] :where(h1, h2, h3, h4, p, span, time, a) { color: #F4EFE6 !important; } body.eg-web [class*="puzzle-teaser"] a[class*="cta"], body.eg-web [class*="puzzle-teaser"] [class*="__cta"] { background: #F4EFE6 !important; color: #241f28 !important; border-color: #F4EFE6 !important; } body.eg-web [class*="puzzle-teaser"] a:not([class*="cta"]) { text-decoration: underline; text-underline-offset: 2px; } /* H2 finding A (2026-08-30, coordinator addendum, CDP matched-rules dump on .eg-ecta-privacy a): the newsletter's G8 eg-ecta unit is a light --eg-paper-2 panel that H1 ported into the dark --eg-ink footer. Both colliding rules live in the generated design layer (assets/design/ web-blocks.css, never hand-edited): `.eg-web .eg-ecta-privacy a { color: var(--eg-ink-3) }` (the panel's own correct intent) loses to `.eg-web .eg-footer a { color: var(--eg-paper) }` (paper-on-paper-2, 1.11:1) purely because the footer rule sits later in that generated file at the SAME specificity (0,2,1). Same fix pattern as every other .eg-web-vs- .eg-footer collision in this file (see the body.eg-web prefix notes above, e.g. .eg-archhero + .eg-grid): add the type selector to win on specificity instead of source order, scoped to the panel's own privacy line only so ordinary footer links keep their paper colour. --eg-ink-3 on --eg-paper-2 is the SAME pairing the coordinator already measured for the privacy line's surrounding sentence — 4.97:1, clears the 4.5:1 AA minimum. text-decoration-color set to match (the footer's own translucent paper underline would otherwise stay all but invisible on the light panel too). */ body.eg-web .eg-ecta-privacy a { color: var(--eg-ink-3); text-decoration-color: var(--eg-ink-4); } /* ------------------------------------------------------------------------ Tables — rules-chapter tables (F12). ------------------------------------------------------------------------ */ /* assets/design/web-blocks-local.css styles every core/table inside .eg-prose (the rules chapter + its scoring tables — .eg-prose is used ONLY by el/rules-content, see blocks/rules-content/render.php) with a border-bottom on every row's cells, but then explicitly zeroes it again on the LAST row (`tbody tr:last-child td { border-bottom: none; }`), so every one of those tables is missing its closing bottom edge — reads as "top borders only". That rule lives outside this ticket's write set (assets/design/), so the fix is an override here instead of an edit there. This stylesheet is enqueued BEFORE web-blocks-local.css (functions.php: 'elsteiner-summer' has no deps, 'eg-web-local' depends on 'eg-web-blocks' which depends on 'eg-web-tokens'), so a same-specificity rule here would still lose on source order — !important forces the win regardless of load order, matching this file's own existing convention for the same kind of cross-file override (see the button and game-card sections above). */ body.eg-web .eg-prose .wp-block-table tbody tr:last-child td { border-bottom: 1px solid var(--wp--preset--color--ink-4, #ddd) !important; } /* ------------------------------------------------------------------------ Printables unit (feedback 2026-08-19; re-audited for the R4c drop, 2026-08-22, ticket PR — see the retirement note at the end of this section for what left and why) ------------------------------------------------------------------------ */ /* F8: the score-sheets page's cards (.eg-dlgrid, assets/design/ web-blocks.css) are capped and centred (R4c: `repeat(auto-fit, minmax(19rem,22rem))`, was `repeat(auto-fit, minmax(21rem,1fr))`), but templates/single-game_extra.html sets contentSize:900px on
, and that page's .wp-block-post-content is ALSO capped to 44rem by the "Tighter reading measure" rule above ("The opt-out" section) — either width is well under what three 22rem tracks plus two 20px gaps need (1096px, the R4c drop's own MANIFEST.md measurement against the site's 1180px page-max). single-game_extra.html is the SHARED template for every game_extra sub-page (rules, variants, puzzles, calculator, score-sheets, ...), so widening it (or bumping contentSize) would widen every one of those, not just the printables page. Rather than add a per-page template just for this, :has() scopes the escape to exactly the post-content that carries .eg-dlgrid — the same "title keeps the measure, content escapes" opt-out already established above, applied without an .is-wide-page class on
. Specificity (0,4,1) beats the existing (0,2,1) reading-measure rule outright, so !important is not needed here. STILL NEEDED after R4c: the drop changed the grid's track sizing, not the page template or this ancestor chain, so the escape this rule buys is exactly as necessary as it was for the 21rem floor. F11 batch fix (feedback 2026-08-19) added el/rules-toc's .eg-rules-layout group between
and wp:post-content on every game_extra sub-page (see "Rules TOC on game_extra" above), so score-sheets' post-content is no longer a direct child of
either — both the plain selector and this :has() escape need the same extra level added. W-E (feedback round 2026-08-19, blind review 2026-08-30): the SAME squeeze hits two more furniture units for the same reason — F1's .eg-actions (koenigrufen/extras/, 87px tiles with labels painting outside the tile border) and H2's .eg-pillars (/about/, three columns forced to 186px). Not a new root cause, so not a new rule: added to this :has() list rather than duplicated. This escape alone is enough for .eg-actions (it declares no max-width of its own — same as the working single-game/render.php instance of it, which sits directly in
with no post-content ancestor to escape from at all, landing at ~1116px there); .eg-dlgrid and .eg-pillars need the further bleed rule below as well, since their own `max-width: var(--eg-w-page-max)` (web-blocks.css) still can't outgrow this now-uncapped but still main-padding-bound post-content box on its own. */ .single .wp-site-blocks > main > .wp-block-post-content:has(.eg-dlgrid, .eg-actions, .eg-pillars, .eg-sk, .game-puzzle-list), .page .wp-site-blocks > main > .wp-block-post-content:has(.eg-dlgrid, .eg-actions, .eg-pillars, .eg-sk, .game-puzzle-list), .single-game_extra .wp-site-blocks > main > .eg-rules-layout .wp-block-post-content:has(.eg-dlgrid, .eg-actions, .eg-pillars, .eg-sk, .game-puzzle-list) { max-width: none; } /* W-E: .eg-dlgrid and .eg-pillars both declare their OWN `max-width: var(--eg-w-page-max); margin: 0 auto` (web-blocks.css) — authored to bind directly against the true page frame, same token main's own box (further up this file) uses. The :has() escape above only removes the 44rem reading-measure cap; it leaves these two still bound by main's own padding-inline (32px, always, from templates/*.html's own inline style — the LOCKED clamp() a few rules up in this file does not actually win against that inline longhand, so the real, measured inset is the flat 32px/48px below, not the clamp formula), plus another 1.25rem this file's own "mobile padding for any direct main child" rule (above) adds at <=768px on whichever ancestor sits directly under
(.eg-rules-layout for game_extra, .wp-block-post-content itself for page.html — same net inset either way). Undo exactly that, symmetrically (both sides get the identical value, so width + margin-inline sums back to 100% of the *un-widened* box — never over-constrained, no !important needed, and no 100cqw viewport trick either: unlike .eg-pagehead's edge-to-edge band, these need to land CENTERED on main's own frame, which a cqw-relative margin does not do once max-width caps the result below 100cqw — verified against a local tokens.css+web-blocks.css+style.css harness before landing this rule, cqw gave x:0 instead of main's own x:42.5 at 1280). Measured after (1280): .eg-dlgrid and .eg-pillars both x:42.5 w:1180, exactly main's own box — three 352px .eg-dlgrid tracks fit (content 1132px, R4c's own "3×352+40=1096" floor), .eg-pillars reaches 1180/345px matching the specimen. .eg-actions is deliberately NOT in this rule — see the comment above. */ /* SK1 (2026-10-30): the scorekeeper section is drawn to sit straight in the page frame with its own 16px gutter (.eg-sk padding) and its own page-max cap; inside the extra template's two paddings it got 256px at 360 and truncated eight names. Same bleed as the grids. The puzzle bands (eg/game-puzzle-list) had the same squeeze: 302px cards in a 256px column, cut at the right edge (2026-10-30). The bands carry their own page-max and gutter, like the scorekeeper. */ body.eg-web .wp-block-post-content .eg-dlgrid, body.eg-web .wp-block-post-content .eg-pillars, body.eg-web .wp-block-post-content .eg-sk, body.eg-web .wp-block-post-content .game-puzzle-list { margin-inline: calc(-1.25rem - var(--wp--preset--spacing--6, 2rem)); width: calc(100% + 2.5rem + 2 * var(--wp--preset--spacing--6, 2rem)); } @media (min-width: 769px) { body.eg-web .wp-block-post-content .eg-dlgrid, body.eg-web .wp-block-post-content .eg-pillars, body.eg-web .wp-block-post-content .eg-sk, body.eg-web .wp-block-post-content .game-puzzle-list { margin-inline: calc(-1 * var(--wp--preset--spacing--6, 2rem)); width: calc(100% + 2 * var(--wp--preset--spacing--6, 2rem)); } } /* F8: intro paragraph and the card row read as touching with no rule between them. web-blocks.css (designer-owned) sets no margin-top on .eg-dlgrid, so this doesn't fight anything — added here rather than there per the write-set split for this ticket. STILL NEEDED after R4c: the re-cut layer still emits no margin-top for .eg-dlgrid (confirmed by diffing the re-cut file against the installed one), so nothing upstream closed this gap. */ body.eg-web .eg-dlgrid { margin-top: var(--eg-w-sp-5, 2rem); } /* Batch fix item 3 (feedback 2026-08-19): .eg-dlgrid and .eg-actions- secondary (web-blocks.css) are both `display:grid`/`display:flex`, but WP still classes their wp:group wrapper `is-layout-flow` (the className attribute points at hand-written CSS rather than a block "layout": {"type":"grid"|"flex"} attribute — same F4/F25 root cause already fixed for .eg-grid above), so WP's global styles inject `:root :where(.is-layout-flow) > * { margin-block: 24px 0 }` on every child but the first. Under grid/flex that margin lands INSIDE each card's own cell/row instead of between rows, so cards start at different vertical offsets within an otherwise aligned row ("staggered tops", .eg-dlgrid; "Also for this game" row, .eg-actions-secondary). Drop it, same as .eg-grid > .eg-gcard above. STILL NEEDED after R4c: the bug is in WP's own block-attribute vs. hand-written-CSS mismatch, not in the design layer, and R4c changed neither .eg-dlgrid's display:grid nor its className. */ .eg-dlgrid > *, .eg-actions-secondary > * { margin-block: 0; } /* RETIRED 2026-08-22 (ticket PR, R4c drop). Two overrides removed here, both obsoleted by the same drop: 1. The 3-up grid-forcing override (`@container eg-web (min-width:1024px) { body.eg-web .eg-dlgrid { grid-template-columns: repeat(auto-fit, minmax(16rem,1fr)); ... } }`). It existed because the OLD `minmax(21rem,1fr)` track let 1fr stretch two cards to half the row at large widths and still miscounted at mid widths, so this file hand-lowered the floor to 16rem as a fallback. R4c's own `minmax(19rem,22rem)` plus `justify-content:center` is a MEASURED cap-and-floor pair (MANIFEST.md: 22rem is the largest track that lets three fit in the page-max content box; 23rem missed by 12px and silently dropped to two columns) — it is the fix this file's own comment was reaching for by hand, now done properly upstream and without a fixed-1fr-stretch failure mode. Re-adding a floor override on top of an already-capped track would only refight a problem the drop already solved, and risks under/overshooting the drop's own measured number. 2. The stale note that eg_download_card() emits no thumb and needs no override to close the gap. v21 of book-scaffold.php (ticket PR) brings the thumb slot BACK (see EG_BOOK_SCAFFOLD_VERSION's changelog) as one shared generic glyph. No override is needed for it EITHER: the re-cut web-blocks.css carries full `.eg-dl-thumb` / `.eg-dl-thumbimg` styling, including the eg-dl--form / eg-dl--ref variant treatments (verified by diffing the re-cut file against the previously installed one — only .eg-dl* / .eg-dlgrid selectors moved, nothing else in the layer did), so the old comment is simply superseded rather than replaced with a new rule. Not verified live (no local WordPress in this worktree — see 00-shared.md "Verification you CAN do"): the exact pixel width single-game_extra.html's post-content column reaches after the :has() escape, which the retired override's own comment estimated at "~1024-1056px" against a floor math that no longer exists. Flagged in the PR report as a staging-verify item rather than asserted here. */ /* ------------------------------------------------------------------------ Tap targets (W-C finding 4, blind review 2026-08-30) — 44px minimum, per the library's own pager law ("Plain links; 44px minimum target"). .eg-rules-toc__summary / .eg-rules-toc__item a (blocks/rules-toc/) and .eg-qintro-all (blocks/puzzle-grid/) and .eg-dgame-name (blocks/single-deck/) are fixed in each block's own style.css — this ticket owns all three. .eg-facet is NOT: it's emitted by both blocks/contract-table/render.php (a concurrent sibling ticket's write set) and blocks/single-game/render.php (out of this ticket's write set), and its sizing rule lives in the generated assets/design/web-blocks.css, which this ticket must never hand-edit — so this is the one tap-target fix that lands here instead, as a cascade override, the same pattern this file already uses for the rules-chapter table border fix above. !important note: verified via the actual live order (koenigrufen/rules/) that this stylesheet (a ) prints AFTER every block's own inline style tag (register_block_type_from_metadata prints block CSS inline, earlier in ) but BEFORE web-blocks.css (also a , later). So a block's own style.css (e.g. puzzle-grid's) beats web-blocks.css on a same-specificity tie by source order alone — no !important needed there — but THIS file, sitting between the two, needs !important to beat web-blocks.css's later .eg-facet rule; matches this file's own documented convention (see "Tables" section above). */ body.eg-web .eg-facet { min-height: 44px !important; } /* Coordinator addendum (ticket W-E, 2026-08-30): the 404 page's single link (templates/404.html) is plain paragraph markup with no block of its own, so it never got the pager law's "44px minimum target" the breadcrumb link already carries (blocks/extra-breadcrumb/style.css — same inline-flex/align-items/min-height shape, reused here rather than reinvented). No !important: this selector (0,3,1) is not competing with anything from web-blocks.css for this class. */ body.eg-web .eg-404-back a { display: inline-flex; align-items: center; min-height: 44px; }