Add UI-09: monochrome dark map tiles

Today's OSM tiles are the default tan/cream basemap, which looks nothing like the mockups regardless of HUD styling. Recommends switching to CARTO's free Dark Matter raster tiles over a client-side color filter on OSM tiles, since the mockups depend on the map itself being dark, not just tinted.

Flags a real interaction with V3-11's offline tile cache: it's keyed by (z,x,y) with no provider tag, so switching tile sources without a cache-busting step would silently serve stale OSM tiles under CARTO's URL. That's folded into this ticket's implementation, not left as a followup bug.

docs/ui-redesign/README.md

@@ -27,6 +27,7 @@ v3 held to.
 | [UI-06](UI-06-plan-and-route-planning.md) | Plan & Route Planning redesign | M | UI-01, UI-03 | Not started |
 | [UI-07](UI-07-rides-history.md) | Rides History redesign | M | UI-01, UI-03 | Not started |
 | [UI-08](UI-08-theme-modern-professional-dark.md) | Theme migration to Modern Professional Dark | S | — | Not started |
+| [UI-09](UI-09-monochrome-dark-map-tiles.md) | Monochrome dark map tiles | S | — | Not started |
 
 ## Dependencies
 
@@ -38,8 +39,11 @@ UI-01 ──┬──► UI-02
 UI-08 ──► UI-03 ──┬──► UI-04 ──► UI-05
                   ├──► UI-06
                   └──► UI-07
+UI-09 ──┬──► UI-05
+        ├──► UI-06
+        └──► UI-07
 
-no dependencies: UI-01 · UI-08
+no dependencies: UI-01 · UI-08 · UI-09
 ```
 
 ## Suggested order
@@ -49,15 +53,17 @@ no dependencies: UI-01 · UI-08
 2. **UI-08** — theme. Every screen ticket after this needs the palette settled once, not
    patched per-screen. Supersedes V3-16's safety-orange direction — see that ticket's
    note.
-3. **UI-02** — skeleton map. Small, and exercises the background-map plumbing UI-01 just
+3. **UI-09** — dark map tiles. Small, no dependencies, and every screen with a map looks
+   wrong without it regardless of how correct the HUD chrome on top is.
+4. **UI-02** — skeleton map. Small, and exercises the background-map plumbing UI-01 just
    built while it's fresh.
-4. **UI-03** — the component kit (`GlassPanel`, `FloatingPill`, `PulsingLocationMarker`).
+5. **UI-03** — the component kit (`GlassPanel`, `FloatingPill`, `PulsingLocationMarker`).
    Build once, consumed by every screen ticket.
-5. **UI-04** — the customizable HUD widget system (drag/resize/persist layout, a
+6. **UI-04** — the customizable HUD widget system (drag/resize/persist layout, a
    visibility toggle in Settings). This is its own ticket, not folded into UI-05, because
    drag-and-resize-with-persisted-layout is a genuinely separate piece of engineering
    from "redesign the record screen" — a small widget-layout editor in miniature.
-6. **UI-05 / UI-06 / UI-07** — the three screen redesigns, in any order once everything
+7. **UI-05 / UI-06 / UI-07** — the three screen redesigns, in any order once everything
    above is done. Each is independently shippable.
 
 ## Things named explicitly by the person who commissioned this, not inferred from the mockups
@@ -77,6 +83,10 @@ no dependencies: UI-01 · UI-08
   mockup's tap-to-expand interaction is a fixed layout with one temporary focus state;
   this goes further — the rider chooses the layout and which stats exist on screen at
   all, and it persists.
+- **The map itself is monochrome dark** (UI-09) — today's default OSM tan/cream tiles are
+  nothing like the mockups regardless of how correct the HUD chrome on top is. A real
+  tile-provider decision, not a color filter thrown on as an afterthought — see that
+  ticket for the trade-off against just filtering OSM's existing tiles.
 
 ## The Map screen is the design north star
 

docs/ui-redesign/UI-09-monochrome-dark-map-tiles.md

@@ -0,0 +1,98 @@
+# UI-09 — Monochrome dark map tiles
+
+**Depends on** nothing · **Size** S · **Status** Not started
+
+## Goal
+Replace the tan/cream default OpenStreetMap raster style with a monochrome dark map, so
+the map actually looks like it belongs in a dark-mode HUD instead of a light basemap
+punched through a dark overlay.
+
+## Context
+`RideMap`, the Plan/Route Planning canvas, and every Stitch mockup all assume a dark or
+near-monochrome map underneath the HUD. Today's `TileLayer` (`ride_map.dart`) points at
+`tile.openstreetmap.org`, OSM's standard light/tan default style — the actual on-device
+result is nothing like the mockups, regardless of how dark the chrome on top of it is.
+
+This affects every screen that shows a map (Map HUD, Plan, Route Planning, and UI-01's
+dimmed background on every other tab), so it's worth settling once, early, rather than
+each screen ticket discovering the same mismatch independently.
+
+## Design
+Two real options, not one obvious answer:
+
+**Option A — switch tile provider to CARTO's "Dark Matter" basemap.**
+`https://{s}.basemaps.cartocdn.com/dark_all/{z}/{x}/{y}{r}.png` (subdomains a-d, retina
+`{r}` variant available, max native zoom 20). Purpose-built dark, desaturated/monochrome
+cartography — closest to the mockups with the least engineering. Free for reasonable
+usage same as OSM's own tiles, but it's a **different third party with its own usage
+policy and attribution requirement** ("© OpenStreetMap contributors © CARTO", not just
+OSM's own attribution) — a real dependency to add, not a style tweak.
+
+**Option B — keep OSM tiles, apply a color filter client-side.** Wrap `TileLayer` in a
+`ColorFiltered` widget using a `ColorFilter.matrix` that desaturates and inverts/darkens
+the tan basemap into a monochrome dark look. No new third party, no new attribution, and
+V3-11's offline tile cache keeps working against the exact tiles it already knows about
+— but the result is a filtered light map, not tiles actually designed for dark
+presentation, and will read as slightly "off" next to CARTO's purpose-built version
+(road/building contrast tuned for light backgrounds doesn't always invert cleanly).
+
+**Recommendation: Option A.** The mockups' whole aesthetic depends on the map itself
+being dark, not just tinted dark, and CARTO's dark tiles are a well-established, free,
+no-API-key option already widely used in the Flutter/Leaflet ecosystem for exactly this.
+The added attribution line and a second host to trust are a small, one-time cost.
+
+**V3-11 interaction:** the offline tile cache (`FileTileCache`/`CachedTileProvider`) is
+already keyed by `TileKey(z, x, y)` with no provider-specific data baked in — but that
+means a cache built against OSM tiles and a cache built against CARTO tiles are
+**indistinguishable to the cache** despite being visually different. Switching providers
+without a plan here would silently serve stale, wrong-looking cached OSM tiles once
+CARTO's URL is live. This needs a cache-busting or provider-tagging fix as part of this
+ticket, not a separate one — see Implementation.
+
+## Implementation
+1. Add a `provider`-qualified cache key (or a cache-version bump that invalidates
+   existing V3-11 cache contents wholesale) so switching tile sources can't silently
+   mix cached tiles from two visually different sources.
+2. Update `RideMap`'s `TileLayer` (and the Route Planner's separate `TileLayer` instance
+   — see `route_planner_screen.dart`) to CARTO's dark tile URL template, correct
+   `subdomains`, and `maxNativeZoom: 20`.
+3. Update the attribution requirement — check whether `flutter_map`'s
+   `RichAttributionWidget`/`AttributionWidget` is in use anywhere already, or add one;
+   OSM-only attribution is no longer sufficient once CARTO tiles are in the mix.
+4. Update `tileUserAgent`/any OSM-specific comments in `ride_map.dart` that assumed OSM's
+   tile servers specifically (the "respect OSM's usage policy" reasoning still applies in
+   spirit to CARTO's own policy, but the specific server being addressed changes).
+5. Re-verify V3-11's tile math/download flow still works end-to-end against the new URL
+   template (tile enumeration and caching are provider-agnostic already; only the fetch
+   URL construction needs updating).
+
+## Acceptance criteria
+- [ ] The map renders in a dark, desaturated/monochrome style matching the Stitch
+      mockups, not OSM's default tan basemap
+- [ ] Correct attribution for both OpenStreetMap and CARTO is shown wherever the map is
+      shown
+- [ ] Switching tile providers does not silently serve stale, visually-mismatched
+      cached tiles from before the switch (V3-11's cache is invalidated or
+      provider-tagged)
+- [ ] Both `RideMap` and the Route Planner's independent `TileLayer` are updated
+      consistently — one dark style everywhere, not just on the screens that happened to
+      get touched first
+
+## Tests
+- Existing `ride_map`/route-planner widget tests updated for the new tile URL and still
+  passing
+- Unit: cache-key/versioning change is exercised the same way V3-11's eviction tests
+  are — a cache built under the old scheme does not get served for the new provider
+
+## Risks
+- **A second third-party tile host is a real dependency, not a free color swap** — its
+  own usage policy, its own risk of being blocked if misused, and its own attribution
+  obligation. Treat it with the same care V3-11 already applies to OSM (rate-limit
+  respecting, no bulk prefetch, real user agent).
+- If CARTO's free tier ever proves insufficient or is discontinued, Option B (a color
+  filter over OSM tiles) is the fallback with no new third-party dependency — worth
+  keeping in mind as a documented Plan B, not re-deriving from scratch later.
+
+## Out of scope
+A fully custom/self-hosted vector tile style — much larger scope, and not needed to hit
+"looks like the mockups" today.