test(trip-cover): cover-strip + retina-cap regression; bump user pin

- trip-header/trips-list specs now assert the cropZoom wide-strip aspect
  and the all-or-nothing retina rule (1x-only for sub-2w sources), and a
  new regression on us-canada-mex-2024 covers the reported portrait-blur
- remove the orphaned AE4 test (no-photos-demo fixture was deleted)
- backlog: full-res pixelfed reimport (luxury quality item)
- plan: record the cover-quality follow-up + AE4 fixture removal
- bump user pin -> 536ca20 (trip-description-hero cover fix)

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RDS6t8wcpbwKvvrxykVQ5K

docs/working/backlog.md

@@ -15,6 +15,12 @@ Ideas and improvements not yet planned or scheduled.
 
 ---
 
+## Content quality — luxury improvements (much later)
+
+- [ ] **Re-import pixelfed photos at full resolution** — the current import pulled pixelfed's optimised web renditions, so imported images cap at ~1440px on the long edge (portraits are 700–1200px wide). This is fine for the feed and 1x banners, but the retina cover 2x only kicks in for genuinely wide (≥1440px) sources, so auto-picked trip banners are currently 1x-only. Find the original high-quality versions in the local filesystem and re-import them (or point the pipeline at the originals rather than the pixelfed web renditions). Purely a quality upgrade — no functional gap; future content shot/stored at full res won't have this ceiling.
+
+---
+
 ## GPX Manager (`/gpx-manager`)
 
 - [ ] **Polish the UI** — the current design is functional but bare; align with the Field Notes aesthetic, add better empty states, drag-and-drop upload area

docs/working/plans/2026-07-05-trip-description-and-hero.md

@@ -274,10 +274,26 @@ intentionally-deferred findings.
   non-image selection (e.g. a `.gpx` from the trip page media) can no longer
   route a non-image Medium into `cropResize`. Also hardens R11.
 - Test quality: replaced a vacuous `toContainText` in the description-clamp
-  spec with real clamp/un-clamp assertions; added the missing **AE4** coverage
-  (trip with no cover + no entry images → text-only header, no banner) via a
-  `no-photos-demo` fixture; corrected an R11 over-claim in the trips-list spec
-  header comment.
+  spec with real clamp/un-clamp assertions; corrected an R11 over-claim in the
+  trips-list spec header comment.
+
+**Follow-up (2026-07-07)**
+
+- **Banner/card cover quality fix.** The macro used `cropResize`, which
+  *fits-inside* preserving aspect ratio — so a portrait fallback source was
+  handed back as a ~165px sliver that the `object-fit:cover` box then upscaled
+  into a blur (reported on `us-canada-mex-2024`). Switched to **`cropZoom`**
+  (crop-to-fill → a real w×h cover strip). Retina is now **all-or-nothing**: the
+  2x `srcset` descriptor is emitted only when the source is genuinely ≥2×w
+  (`cover.width >= 2w`), else 1x-only — no upscaling, no intermediate widths.
+  Note: imported pixelfed photos cap at ~1440px wide, so auto-picked covers are
+  usually 1x-only; see `docs/working/backlog.md` (full-res re-import, luxury).
+- **AE4 fixture removed.** The `no-photos-demo` fixture (and its browser test)
+  was deleted at the user's request — it surfaced as stray demo content in the
+  trip list. AE4 (no cover + no images → no banner) is a trivial else-branch of
+  the shared macro's `{% if cover %}` guard, covered by construction alongside
+  the R7/AE3 fallback tests. A regression test for the reported portrait-blur
+  bug now lives in `trip-header.spec.js` against `us-canada-mex-2024`.
 
 **Intentionally deferred — explicit plan override (do not re-flag)**