Implement option B card metadata and solid bottom edges

data/tv-spec/spec.md

@@ -1,6 +1,6 @@
 # Wolfskin TV specification — revision 6
 
-2026-11-06. Revision 6 adds the approved direction-A motion and loading states from [the motion reference](https://pages.justinvisser.org/wolfskin/tv-motion-speed.html), frozen with SHA-256 checksums at `data/tv-v2/motion-2026-11-06/`. Revision-5 feedback below still applies. Unchanged static geometry, navigation, typography and colours remain governed by the approved frames and the [written specification](https://pages.justinvisser.org/wolfskin/tv-spec). The revision-5 prototype delta remains `data/tv-v2/feedback-2026-11-06/`.
+2026-11-07. Revision 6 adds the approved direction-A motion and loading states from [the motion reference](https://pages.justinvisser.org/wolfskin/tv-motion-speed.html), frozen with SHA-256 checksums at `data/tv-v2/motion-2026-11-06/`, and the subsequently approved **option B** [card metadata frames](https://pages.justinvisser.org/wolfskin/tv-card-meta-frames.html). Revision-5 feedback below still applies except where the card amendment explicitly supersedes it. Unchanged navigation, typography and colours remain governed by the [written specification](https://pages.justinvisser.org/wolfskin/tv-spec). The revision-5 prototype delta remains `data/tv-v2/feedback-2026-11-06/`.
 
 ## Who’s watching
 
@@ -83,6 +83,27 @@ Discover key traversal skips pending lanes in both directions. If every remainin
 Native **1920×1080** before/after frames and the frame-budget inventory are published at [the motion build proof](https://pages.justinvisser.org/wolfskin/tv-motion-build/). Before frames use the pre-change surface sources in this same isolated worktree; after frames use the shipped Compose surfaces. No device access, installation or optimized APK is part of this lane. On-box acceptance remains: no series layout change after first content, Home scrolling no worse than **39.6% janky / p90 65 ms**. Native JVM rendering does not measure that GPU/frame-timing target.
 
 
+## Card metadata and bottom state line — revision 6, option B
+
+Film cards show year, duration and a **24 px watched check**. Episodes show their own duration and check, without a year; series show year and check, never an aggregate runtime. Resuming and unwatched cards have no check. Compact metadata omits media-type words, season counts, character roles, ownership and remaining-time prose. Duration rounds total minutes before formatting hours/minutes; narrow layouts tighten metadata gaps from **12 → 8 → 4 px**, never shrink text or ellipsise duration.
+
+Movies, Shows, Anime, Home See all and person filmography always show their card-width title/subtitle. Titles are one-line ellipses: **28 px / 32 px line**, medium muted at rest, bold foreground when focused. Metadata is **24 px / 26 px line**. The **96 px** poster caption band starts its text **38 px** below unscaled art. Home poster captions remain focus-only; Home landscape titles stay on the art with focus-only metadata below it. Landscape shelves reserve **64 px** below art. Loading placeholders reserve the same title/subtitle slots and bands; each shelf still shares the motion lane's single pulse layer.
+
+| Geometry at 1920×1080 | Pixels |
+|---|---:|
+| Library row pitch | 524 |
+| Home poster shelf pitch | 632 |
+| Home wide / tile shelf pitch | 522 / 441 |
+| Initial library art row tops | 372 / 896 |
+| Pinned library art row tops / second-row bottom | 168 / 692 / 1076 |
+| Focus ring to caption / caption to next focused row ring | 20.48 / 26.48 |
+
+One bottom state line replaces the independent resume bar. Unwatched and Discover have no state line; watched has a full-width **3 px accent** line; resume has proportional **6 px accent** fill on a full-width `color.progressTrack` track. Left/right are zero, with **2 px bottom backing** and **2 px opaque separator** above the fill, clipped by the art's **16 px** radius. Focus scales the art/line by **1.08**, with its **4 px** straddle ring drawn independently above them.
+
+Justin's additional correction is binding even where a reference frame leaks: **the lowest card row must contain card/edge/fill colour, never art**. Clip artwork out of the backing band before painting it; merely overdrawing after a rounded clip can retain poster colours in antialiased corners. The backing is opaque `color.bg` (**#07070a**). Precompose and quantise the track over that backing once to preserve the approved **#3d3d3f** browser pixels instead of native alpha-rounding differences; fill is **#cc9fff**. The approved Annie Hall and Locke treatment frames themselves expose artwork below their inset state line; the native build deliberately makes that row solid.
+
+Native same-fixture comparisons and the exhaustive real-art bottom-row proof are published in [the card metadata build gallery](https://pages.justinvisser.org/wolfskin/tv-card-meta-build/). `CardBottomPixelTest` checks white Annie Hall, dark Locke, colourful Apocalypse Now and Avatar Aang at rest and individually focused; every pixel across each full bottom row must equal an art-free rendering of the same clipped backing/ring, while straight-edge and fill/track pixels additionally match exact state colours. `CardGeometryTest` measures narrow fitting, the 24 px check, grid/shelf pitches, pinned rows and ring clearances from native surfaces. Visible reference differences are listed in the gallery and remain merge blockers unless explicitly approved.
+
 ## Sources and verification
 
 Authoritative changed values live in `wolfskin-design/tokens.json`, `title-tokens.json` and `discover-tokens.json`; regenerate Kotlin, CSS and the drift manifest with `python3 wolfskin-design/generate_tokens.py`.