Top and Cheapest were two words for one state: both meant "the lowest price
among the stations on screen", differing only in the candidate set (radius vs
Favourites). Which word a surface used had drifted by device — the phone's
rows said Cheapest while the watch said Top for the identical state. Top is
gone from the enum, the Key legend, the badge and the watch's hand-rolled
label; Cheapest survives because it names the measure and is already the word
the favourites copy uses.
Closest is now a real station label rather than a verdict. StationRecommendationBadge
gains .closest with a total precedence — Best value, then Cheapest, then Closest
— because a row has room for one label and proximity is the weakest claim: a
short drive says nothing about the price, and unlike the other two it reports no
saving at all. The badge sits above the price, where the other two sit. The
closest row's verdict capsule is suppressed: its only word for that state was
"Closest", so printing it below as well would state the same fact twice, three
lines apart. Other rows keep their Saves / Break even / Costs capsules unchanged.
"Closest station" becomes "Closest", read from the single
FuelSavingsResult.baselineHeadline constant so the badge, the verdict and the Key
cannot word it differently. The Key legend renders all three station labels with
the shared StationLabelBadge; the old KeyStationLabelIndicator, which
special-cased the closest station to a wider verdict pill so it read as a
different class of thing from the two above it, is deleted.
One closest icon everywhere. The listing and detail verdict symbols used
location.circle and location.circle.fill for the state the Key taught as
location.north.fill; both now read StationLabelKind.closest.symbol, so the three
surfaces resolve one glyph from one place.
The detail sheet gains the Closest capsule below the map, ordered to match the
Key. Favourites gains the nearest favourite as its Closest, measured on the same
one-way distance its rows print and restricted to favourites with a price for the
active fuel — matching the radius tab, where the closest station is by definition
the savings baseline.
closestStationID flows from the one scaffold baseline the savings lookup already
compares against, so a row's badge and the reference station cannot disagree.
No savings calculation, threshold, ranking, eligibility or station-selection
logic changed.
The Key sheet's detent is now bound rather than fixed, so dragging it up to read
the Station labels section sticks instead of snapping back; a QA launch argument
opens it at full height, since the 0.6 detent puts that section below the fold on
every device. That argument is read unconditionally, unlike the other harness
hooks which are #if DEBUG-gated.
Tests: 278 pass, 9 new — the label precedence ladder across every combination,
capsule order in the sheet, Closest alone, and the retired label pinned absent.
The More info sheet now lists every ranking a station qualifies for, directly under the map and above the price and distance: the same StationLabelBadge the listing rows render, in a flow layout so two or three capsules wrap to a new line rather than truncating at large Dynamic Type.
The rules are not restated. StationRanking defers to StationRecommendationBadge.badge for the Best Value gate, and takes Cheapest and Top as the flags the host list already computed for its own rows, so the sheet cannot award a capsule the row behind it withheld. A host passes one of the two, never both: Stations sends isCheapestInRange, Favourites sends isTopFavourite. No calculation, ranking, eligibility or savings logic changed.
Rename the card heading Best Value to Value breakdown, since the working behind it can end in Costs rather than a saving, and drop the green section tint so the card matches the neutral background used elsewhere in the sheet. The Saves, Break even and Costs capsules and every breakdown row are untouched.
Rename the explanation link to How does FuelBoard calculate value? It names the calculation rather than the feature, of which the Best Value gate is only the last step. The sheet's own title stays short: the full question truncated to 'How does FuelBoard...' as a navigation title at both inline and large display modes, so the row carries the question and the sheet names the topic.
Tests: 269 pass, 5 new covering capsule selection, order and sheet-vs-row agreement on Best Value.
One adaptive tone palette and one shape rule now drive every capsule on
Stations, Favourites, the detail sheet and the Key.
- CapsuleTone: the light/dark palette pairs in one place, resolved through a
UIColor dynamic provider so a List row that re-renders without rebuilding
its body still follows an appearance switch. Every pair clears 4.5:1
foreground-on-background in both appearances (4.96-6.72:1).
- AdaptiveCapsule: capsule at one drawn line, rounded rect at two or more,
radius 8 -> 10, measuring content before padding so insets cannot inflate
the line count.
- Listing verdict capsules move from white-on-muted-solid to the tinted
palette, so Saves, Break even, Costs and Closest station read the same in
the rows and in the detail sheet.
- Best Value is purple, not green: green stays reserved for a financial
saving, which is what the Saves capsule beside it means.
- Price dots standardise to 12pt in the app (widget keeps its own 6pt dot;
PriceRatingDot now takes a diameter so the difference is deliberate).
- Station labels are sentence case with their recommended symbols (medal, tag,
location.north.fill, trophy) and StationLabelKind replaces the Key's
string-literal dispatch, so renaming a label can no longer drop a row.
- Key says "Costs", matching the capsule's own headline (was "Costs more").
FuelBoard Contributor
FuelBoard Contributor