/* ===================================================================== ROYALBUILDER.ai · lockup.css THE FINAL LOCKUP, and the small amount of page furniture the four layer pages need that neither royal-system.css nor front-door.css already ships. WHY THIS FILE EXISTS AT ALL. The Royal Ai lockup is the header logo on every page of every site in the family (ruled 7/28/2026). It is INLINED as an SVG and d, never ed, so the page's own Montserrat and CSS can reach it. An inlined , then referenced with . It has to be defined once and only once: the file carries three gradients - gNavyR, gNavyO and gGoldO - and a second copy of those IDs on the same page makes the browser resolve whichever it met first, silently. hidden alone is not enough on every engine, so the block is also pulled out of flow and given no box at all. --------------------------------------------------------------------- */ .lockupdefs{position:absolute;width:0;height:0;overflow:hidden; pointer-events:none} /* --------------------------------------------------------------------- 2. THE MARK, ONE SIZE SUPERSEDED 2026-08-02 - this section used to be headed "THE MARK, TWO SIZES" and it carried the arithmetic for both: a front-door size reproducing --fd-mark's min(620px, 94vw) width as a height off the old 694-unit wordmark, and a smaller deep-page size chosen to match the optical weight of the castle-plus-words sitemark the brandbar used to carry. Both sizes and both derivations are preserved verbatim at _superseded\2026-08-02-header-law\css\lockup.css. There is one size now, and it is the same one every site in the family uses. A front door and a deep page are the same brand at the same height; the wrappers still differ, the mark does not. --------------------------------------------------------------------- */ /* --------------------------------------------------------------------- THE HEADER LAW, 2026-08-02 (RAI.v7/ROVER v3 PART TWO C-2 / C-1b). THE MARK IS A CUT PNG NOW. assets\header-lockup-mitre-v11.png, 577 x 192 true pixels, carried by an - not by the inlined are the asset's true pixel size and are there only to reserve the box against layout shift - they are not a size instruction, and CSS must never restate them. FOUR RULES WERE RETIRED HERE and every one of them is preserved verbatim at _superseded\2026-08-02-header-law\css\lockup.css - the bare rule that set only width and display for the inlined mark, the front-door size it took inside .markwrap (derived from the old 694-unit wordmark width), the deep-page size it took inside .brandbar, and that rule's narrow-screen override inside a max-width:767.98px query. Their literal text is deliberately NOT restated in this comment: a commented-out size reads as a live one to every grep that audits this family, and the audit has to be able to see one size and only one. WHY THE TWO WRAPPER-SCOPED SIZES HAD TO GO, caught by MEASUREMENT and not by reading: the law's rule is `.brandmark`, specificity (0,1,0). Those two were `.markwrap .brandmark` and `.brandbar .brandmark`, specificity (0,2,0). A more specific rule wins NO MATTER WHERE IT SITS in the cascade, so this tree kept rendering its header at 114px while every other tree rendered 64px, even though all eight carried a byte-identical size string. The static check passed. The family was still broken. That is exactly the failure C-1b/4 predicts: "Uniform means those ten numbers match — not that the CSS matches. Eyeballing one page at a time will not catch drift and never has." It was found by _royal-system\_build\contact-sheet.html, which loads the real shipping pages and measures the rendered box in each. THE SAME TRAP, ONE STEP FURTHER, and this tree fell into it tonight: a first pass moved the size into each page's own