wrolpi/app/src/components/MapViewer.js

995 lines
40 KiB
JavaScript
Raw Normal View History

import React, {useCallback, useContext, useEffect, useRef, useState} from "react";
import {useSearchParams} from "react-router";
import maplibregl from "maplibre-gl";
import "maplibre-gl/dist/maplibre-gl.css";
import {Protocol} from "pmtiles";
import layers from "protomaps-themes-base";
import mlcontour from "@acalcutt/maplibre-contour-pmtiles";
import {addMapPin, deleteMapPin, getMapFiles, getMapPins, searchMap, setMapDefaultLocation} from "../api";
import {MAP_VIEWER_URI} from "./Vars";
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
import {Button, Checkbox, Header, Icon, Panel, TextInput, Toggle} from "./ui";
import {SettingsContext, ThemeContext} from "../contexts/contexts";
ui: the monochrome themes draw a black basemap, not a dark one Roland asked why the red overlay sits on the DARK basemap rather than greyscale or black. Measuring it, black is the better answer and the reason is not brightness: flavor distinct colours hued earth luminance dark 19 4 0.014 black 13 0 0.007 Both media filters are the same pure luminance projection -- `0.2126 0.7152 0.0722` -- so they keep brightness and throw hue away. `black` is already achromatic, so filtering it is a hue rotation and nothing else: all thirteen greys were authored as a monochrome hierarchy and every one arrives intact. `dark` has more colours, but four of them are hued (water, parks, up to 25 chroma) and those distinctions do not survive at all -- the features land at whatever brightness their hue happened to carry rather than one anybody chose. Black also emits half the light on the earth fill, which is the point of a night-vision theme. `grayscale` is not an option despite the name: it is a LIGHT theme (earth L0.604), and a luminance filter turns a light map into a bright red one. What made black look impossible in the first commit was the sprite, and that was my mistake: `sprite` and the layer flavor are separate style properties that merely shared one variable. We ship `light` and `dark` sprite sets, and a `black` basemap with `dark` icons is an ordinary style. They are now separate arguments, with `mapSprite()` beside `mapFlavor()`. Fixed in passing: the contour labels branched on `flavor === "dark"` and would have taken the LIGHT branch for `black` -- dark brown text on a black map. Now branches on `light`, so every dark flavor gets the pale label. The new test is the one worth having: the monochrome themes' flavor must paint with greys only, paired with an assertion that `dark` is NOT achromatic. Without that pair the first would pass on any flavor at all and prove nothing about the choice. 1096 jest / 61 suites, 4 e2e themes, clean build. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-02 13:33:51 -06:00
import {mapFlavor, mapSprite} from "../themes/names";
// Terrain DEM file prefix for hillshade and contours.
const TERRAIN_PREFIX = "terrain-";
// Detect full planet files: "planet-20260329.pmtiles" or bare date like "20260329.pmtiles".
function isPlanetFile(name) {
if (name.startsWith("planet-")) return true;
return /^\d{8}\.pmtiles$/.test(name);
}
// Layer groups matching the standalone viewer.
const LAYER_GROUPS = {
"Hillshade": ["hillshade"],
"Contours": ["contour-lines", "contour-labels"],
"Water": ["water"],
"Land Cover": [
"landcover", "landuse_park", "landuse_urban_green", "landuse_beach",
"landuse_aerodrome", "landuse_hospital", "landuse_industrial",
"landuse_pedestrian", "landuse_pier", "landuse_school", "landuse_zoo",
],
"Buildings": ["buildings"],
"Roads - Highway": ["roads_highway"],
"Roads - Major": ["roads_major", "roads_link"],
"Roads - Minor": ["roads_minor", "roads_other"],
"Roads - Tunnels": ["roads_tunnels"],
"Roads - Bridges": ["roads_bridges"],
"Rail & Runways": ["roads_rail", "roads_runway", "roads_taxiway", "landuse_runway"],
"Boundaries": ["boundaries"],
"Labels": ["roads_labels", "places_", "address_label", "pois", "water_label", "water_waterway_label"],
};
// Register PMTiles protocol once.
const protocol = new Protocol();
maplibregl.addProtocol("pmtiles", protocol.tile);
// Planet overview blob — provides global basemap at zoom 0-6.
const MAP_OVERVIEW_URL = "/blobs/map-overview.pmtiles";
// Layers to skip from non-primary sources (they would paint over the primary source's content).
const SKIP_SECONDARY = new Set(["background", "earth"]);
// Set up contour DEM source (must be called once before buildStyle).
let demSource = null;
let terrainUrl = null;
function initContourSource(filename) {
terrainUrl = `/media/map/${filename}`;
if (demSource) return;
demSource = new mlcontour.DemSource({
url: `pmtiles://${window.location.origin}${terrainUrl}`,
encoding: "terrarium",
maxzoom: 8,
worker: true,
});
demSource.setupMaplibre(maplibregl);
}
// sources: [{name, url}] where name is used as the MapLibre source ID and url is the pmtiles:// URL.
// hasTerrain: whether to include hillshade and contour layers.
// scaleUnit: "metric" or "imperial" — affects contour intervals and labels.
ui: the monochrome themes draw a black basemap, not a dark one Roland asked why the red overlay sits on the DARK basemap rather than greyscale or black. Measuring it, black is the better answer and the reason is not brightness: flavor distinct colours hued earth luminance dark 19 4 0.014 black 13 0 0.007 Both media filters are the same pure luminance projection -- `0.2126 0.7152 0.0722` -- so they keep brightness and throw hue away. `black` is already achromatic, so filtering it is a hue rotation and nothing else: all thirteen greys were authored as a monochrome hierarchy and every one arrives intact. `dark` has more colours, but four of them are hued (water, parks, up to 25 chroma) and those distinctions do not survive at all -- the features land at whatever brightness their hue happened to carry rather than one anybody chose. Black also emits half the light on the earth fill, which is the point of a night-vision theme. `grayscale` is not an option despite the name: it is a LIGHT theme (earth L0.604), and a luminance filter turns a light map into a bright red one. What made black look impossible in the first commit was the sprite, and that was my mistake: `sprite` and the layer flavor are separate style properties that merely shared one variable. We ship `light` and `dark` sprite sets, and a `black` basemap with `dark` icons is an ordinary style. They are now separate arguments, with `mapSprite()` beside `mapFlavor()`. Fixed in passing: the contour labels branched on `flavor === "dark"` and would have taken the LIGHT branch for `black` -- dark brown text on a black map. Now branches on `light`, so every dark flavor gets the pale label. The new test is the one worth having: the monochrome themes' flavor must paint with greys only, paired with an assertion that `dark` is NOT achromatic. Without that pair the first would pass on any flavor at all and prove nothing about the choice. 1096 jest / 61 suites, 4 e2e themes, clean build. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-02 13:33:51 -06:00
// `sprite` is deliberately separate from `flavor`: we ship two sprite sets (light, dark)
// and the monochrome themes draw a `black` basemap, which has no sprites of its own.
function buildStyle(sources, flavor = "light", hasTerrain = false, scaleUnit = "metric",
sprite = "light") {
const styleSources = {};
const allLayers = [];
for (let i = 0; i < sources.length; i++) {
const {name, url} = sources[i];
styleSources[name] = {type: "vector", url};
const fileLayers = layers(name, flavor);
if (i === 0) {
// Primary source: use layer IDs as-is.
allLayers.push(...fileLayers);
} else {
// Additional sources: prefix layer IDs and skip background/earth.
for (const layer of fileLayers) {
if (SKIP_SECONDARY.has(layer.id)) continue;
allLayers.push({...layer, id: `${name}:${layer.id}`});
}
}
}
// Add terrain sources and layers if available.
if (hasTerrain && demSource) {
// Read saved visibility to set initial layer state correctly.
const savedVis = getInitialVisibility();
const hillshadeVis = savedVis["Hillshade"] ? "visible" : "none";
const contourVis = savedVis["Contours"] ? "visible" : "none";
// Hillshade from raster-dem source.
styleSources["terrain-dem"] = {
type: "raster-dem",
url: `pmtiles://${terrainUrl}`,
encoding: "terrarium",
tileSize: 512,
};
// Insert hillshade early (after earth/landcover, before roads).
const hillshadeSearch = allLayers.findIndex(l => l.id === "buildings");
const hillshadeIdx = hillshadeSearch === -1 ? allLayers.length : hillshadeSearch;
allLayers.splice(hillshadeIdx, 0, {
id: "hillshade",
type: "hillshade",
source: "terrain-dem",
layout: {visibility: hillshadeVis},
paint: {
"hillshade-shadow-color": "#000000",
"hillshade-illumination-anchor": "viewport",
"hillshade-exaggeration": 0.3,
},
});
// Contour lines from on-the-fly generation.
const isImperial = scaleUnit === "imperial";
const contourUrl = demSource.contourProtocolUrl({
overzoom: 1,
contourLayer: "contours",
elevationKey: "ele",
levelKey: "level",
multiplier: isImperial ? 3.28084 : 1, // meters to feet conversion
thresholds: isImperial ? {
// [minor, major] contour intervals in feet per zoom level.
4: [1000, 5000],
5: [500, 2500],
6: [200, 1000],
7: [200, 1000],
8: [100, 500],
} : {
// [minor, major] contour intervals in meters per zoom level.
4: [500, 2500],
5: [200, 1000],
6: [100, 500],
7: [100, 500],
8: [50, 200],
},
});
const elevUnit = isImperial ? "ft" : "m";
styleSources["contours"] = {
type: "vector",
tiles: [contourUrl],
maxzoom: 8,
};
allLayers.push(
{
id: "contour-lines",
type: "line",
source: "contours",
"source-layer": "contours",
filter: ["!=", ["get", "level"], 1],
layout: {visibility: contourVis},
paint: {"line-color": "rgba(150,100,50,0.3)", "line-width": 0.5},
},
{
id: "contour-lines-major",
type: "line",
source: "contours",
"source-layer": "contours",
filter: ["==", ["get", "level"], 1],
layout: {visibility: contourVis},
paint: {"line-color": "rgba(150,100,50,0.5)", "line-width": 1},
},
{
id: "contour-labels",
type: "symbol",
source: "contours",
"source-layer": "contours",
filter: ["==", ["get", "level"], 1],
layout: {
visibility: contourVis,
"symbol-placement": "line",
"text-field": ["concat", ["get", "ele"], elevUnit],
"text-size": 10,
},
paint: {
ui: the monochrome themes draw a black basemap, not a dark one Roland asked why the red overlay sits on the DARK basemap rather than greyscale or black. Measuring it, black is the better answer and the reason is not brightness: flavor distinct colours hued earth luminance dark 19 4 0.014 black 13 0 0.007 Both media filters are the same pure luminance projection -- `0.2126 0.7152 0.0722` -- so they keep brightness and throw hue away. `black` is already achromatic, so filtering it is a hue rotation and nothing else: all thirteen greys were authored as a monochrome hierarchy and every one arrives intact. `dark` has more colours, but four of them are hued (water, parks, up to 25 chroma) and those distinctions do not survive at all -- the features land at whatever brightness their hue happened to carry rather than one anybody chose. Black also emits half the light on the earth fill, which is the point of a night-vision theme. `grayscale` is not an option despite the name: it is a LIGHT theme (earth L0.604), and a luminance filter turns a light map into a bright red one. What made black look impossible in the first commit was the sprite, and that was my mistake: `sprite` and the layer flavor are separate style properties that merely shared one variable. We ship `light` and `dark` sprite sets, and a `black` basemap with `dark` icons is an ordinary style. They are now separate arguments, with `mapSprite()` beside `mapFlavor()`. Fixed in passing: the contour labels branched on `flavor === "dark"` and would have taken the LIGHT branch for `black` -- dark brown text on a black map. Now branches on `light`, so every dark flavor gets the pale label. The new test is the one worth having: the monochrome themes' flavor must paint with greys only, paired with an assertion that `dark` is NOT achromatic. Without that pair the first would pass on any flavor at all and prove nothing about the choice. 1096 jest / 61 suites, 4 e2e themes, clean build. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-02 13:33:51 -06:00
"text-color": flavor === "light" ? "rgba(150,100,50,0.8)" : "rgba(210,180,140,0.9)",
"text-halo-color": flavor === "light" ? "rgba(255,255,255,0.8)" : "rgba(0,0,0,0.8)",
"text-halo-width": 1,
},
},
);
}
return {
version: 8,
glyphs: "/map-assets/fonts/{fontstack}/{range}.pbf",
ui: the monochrome themes draw a black basemap, not a dark one Roland asked why the red overlay sits on the DARK basemap rather than greyscale or black. Measuring it, black is the better answer and the reason is not brightness: flavor distinct colours hued earth luminance dark 19 4 0.014 black 13 0 0.007 Both media filters are the same pure luminance projection -- `0.2126 0.7152 0.0722` -- so they keep brightness and throw hue away. `black` is already achromatic, so filtering it is a hue rotation and nothing else: all thirteen greys were authored as a monochrome hierarchy and every one arrives intact. `dark` has more colours, but four of them are hued (water, parks, up to 25 chroma) and those distinctions do not survive at all -- the features land at whatever brightness their hue happened to carry rather than one anybody chose. Black also emits half the light on the earth fill, which is the point of a night-vision theme. `grayscale` is not an option despite the name: it is a LIGHT theme (earth L0.604), and a luminance filter turns a light map into a bright red one. What made black look impossible in the first commit was the sprite, and that was my mistake: `sprite` and the layer flavor are separate style properties that merely shared one variable. We ship `light` and `dark` sprite sets, and a `black` basemap with `dark` icons is an ordinary style. They are now separate arguments, with `mapSprite()` beside `mapFlavor()`. Fixed in passing: the contour labels branched on `flavor === "dark"` and would have taken the LIGHT branch for `black` -- dark brown text on a black map. Now branches on `light`, so every dark flavor gets the pale label. The new test is the one worth having: the monochrome themes' flavor must paint with greys only, paired with an assertion that `dark` is NOT achromatic. Without that pair the first would pass on any flavor at all and prove nothing about the choice. 1096 jest / 61 suites, 4 e2e themes, clean build. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-02 13:33:51 -06:00
sprite: `${window.location.origin}/map-assets/sprites/${sprite}`,
sources: styleSources,
layers: allLayers,
};
}
// Groups that start hidden by default (used when no saved state exists).
const DEFAULT_HIDDEN = new Set(["Hillshade", "Contours"]);
const STORAGE_KEY_LAYERS = "wrolpi-map-layers";
const STORAGE_KEY_SCALE = "wrolpi-map-scale";
function getInitialVisibility() {
try {
const saved = localStorage.getItem(STORAGE_KEY_LAYERS);
if (saved) return JSON.parse(saved);
} catch (e) { /* ignore */ }
const v = {};
for (const g of Object.keys(LAYER_GROUPS)) v[g] = !DEFAULT_HIDDEN.has(g);
return v;
}
function LayerControl({map, scaleUnit, onScaleUnitChange, visibilityRef}) {
const [visibility, setVisibility] = useState(() => {
const v = getInitialVisibility();
if (visibilityRef) visibilityRef.current = v;
return v;
});
const [collapsed, setCollapsed] = useState(true);
const [isMobile, setIsMobile] = useState(
() => typeof window !== "undefined" && window.matchMedia("(max-width: 768px)").matches
);
useEffect(() => {
if (typeof window === "undefined") return;
const mq = window.matchMedia("(max-width: 768px)");
const handler = (e) => setIsMobile(e.matches);
mq.addEventListener ? mq.addEventListener("change", handler) : mq.addListener(handler);
return () => {
mq.removeEventListener ? mq.removeEventListener("change", handler) : mq.removeListener(handler);
};
}, []);
const iconOnly = isMobile && collapsed;
const toggleGroup = useCallback((groupName, checked) => {
if (!map) return;
const prefixes = LAYER_GROUPS[groupName];
const allMapLayers = map.getStyle()?.layers || [];
for (const layer of allMapLayers) {
// Match both plain IDs and prefixed IDs (e.g., "water" and "file.pmtiles:water").
const baseId = layer.id.includes(":") ? layer.id.split(":").slice(1).join(":") : layer.id;
if (prefixes.some(p => baseId.includes(p))) {
map.setLayoutProperty(layer.id, "visibility", checked ? "visible" : "none");
}
}
setVisibility(prev => {
const next = {...prev, [groupName]: checked};
if (visibilityRef) visibilityRef.current = next;
try { localStorage.setItem(STORAGE_KEY_LAYERS, JSON.stringify(next)); } catch (e) { /* ignore */ }
return next;
});
}, [map, visibilityRef]);
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
// A floating control drawn over the map canvas: it is part of the app's interface,
// not map data, so it is never filtered — it stays plain, token-colored chrome in
// every theme, same as the nav bar.
const panelStyle = {
position: "absolute",
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
top: "0.625rem",
left: "0.625rem",
zIndex: 1000,
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
background: "var(--panel)",
border: "1px solid var(--border)",
color: "var(--text)",
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
maxHeight: "calc(100vh - 7.5rem)",
overflowY: "auto",
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
minWidth: iconOnly ? 0 : "12.5rem",
ui: scale the whole interface by 10%, off the root font-size The interface reads about 10% smaller than the pre-migration Semantic build, and the drift is not one number. Measured on the QA Pi at 1280px, Semantic's headings were h1 28 / h2 24 / h3 18 / h4 15 / h5 14 against our 26 / 21 / 17 / 14 / 13, and its video card a fixed 290px with an 18px title against our 238px with a 13px one -- while our BODY text and chrome went the other way, a 16px base against Semantic's 14 and 42px buttons against 36. No per-component nudge reproduces what a user comparing the two sees, which is the whole page at once. Ten percent is not a fudge factor: our heading scale times 1.1 is 28.6 / 23.1 / 18.7 / 15.4 / 14.3, which is Semantic's scale to within a pixel at every step. Verified in the browser -- those are the numbers it now renders. `--ui-scale: 1.1` and `html { font-size: calc(100% * var(--ui-scale)) }`, with everything that should grow written in rem: - 102 lengths in ui.css, App.css and tokens.css converted, restricted by property: font-size, padding, margin, gap, width/height, border-radius, offsets. - Borders, outlines and shadows deliberately NOT converted. A hairline is a hairline at any text size. - 46 inline `fontSize` px values across 8 files. `em` values were left alone: they already resolve against the parent. - Ten inline-styled blocks in the library (Card's text block, Loading, Placeholder, ColoredInput, Confirm's action row, CardGroup's gap) moved into ui.css, because an inline px size cannot be reached by a scale or seen by a test. - CardGroup converts its `minWidth` prop to rem, which is what returns a phone to one card per row: as px the track ignored the scale, so a 430px screen fitted two 200px cards where the old build fitted one 290px card. Mantine needs nothing: its 644 size declarations are `calc(Xrem * var(--mantine-scale))` and scale with the root. `theme.scale` must stay at 1 -- it exists to CANCEL a customised root font-size, so setting it to 1.1 as well would leave Mantine's half of the interface at the old size beside our larger text. A test pins that. `zoom: 1.1` on the root was tried first and rejected. It works and it is one line, but it scales at paint: every 1px hairline becomes 1.1px and softens on a 1x display, and viewport units do not shrink to compensate, so `100vh` paints 110% of the screen -- a 95vw modal overflowed sideways and the CBZ viewer's fullscreen overflowed down. Rem scales real lengths, keeps borders crisp, leaves viewport units alone, and preserves a user's own browser font-size preference instead of overriding it. A percentage rather than a px root size, for that last reason. Five guards in themes/ui-scale.test.js, each planted against: the scale itself, that every scalable declaration is rem, that borders are NOT, that no inline px font-size survives anywhere including the gallery, and that Mantine's scale stays 1. Also removes `.ui.modal.themed-confirm`, a dead Semantic-era rule and the source of the one `95vw` that would have overflowed: `.ui.modal` no longer exists and nothing sets that class. Verified: 1141 tests / 65 suites; clean build; browser-checked at 1280px and at a 430px phone; navbar single-row from 780 to 1800px; hairlines still 1px and no horizontal overflow in all four themes. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-03 14:45:08 -06:00
fontSize: '0.8125rem',
};
const headerStyle = iconOnly
? {
padding: 0, cursor: "pointer", userSelect: "none",
display: "flex", alignItems: "center", justifyContent: "center",
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
width: "2.5rem", height: "2.5rem",
}
: {
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
padding: "0.5rem 0.75rem", cursor: "pointer", fontWeight: 600,
display: "flex", justifyContent: "space-between", alignItems: "center",
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
borderBottom: collapsed ? "none" : "1px solid var(--border)", userSelect: "none",
};
return <div style={panelStyle}>
<div
style={headerStyle}
onClick={() => setCollapsed(c => !c)}
title="Layers"
aria-label="Layers"
>
{iconOnly
? <Icon name="map outline" size="large" style={{margin: 0}}/>
: <>
<span>Layers</span>
<Icon name={collapsed ? "chevron right" : "chevron down"} size="small"/>
</>}
</div>
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
{!collapsed && <div style={{padding: "0.25rem 0.75rem 0.5rem"}}>
{Object.keys(LAYER_GROUPS).map(name =>
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
<div key={name} style={{padding: "0.125rem 0"}}>
<Checkbox
label={name}
checked={visibility[name]}
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
onChange={(e) => toggleGroup(name, e.currentTarget.checked)}
/>
</div>
)}
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
<div style={{borderTop: "1px solid var(--border)", marginTop: "0.375rem", paddingTop: "0.375rem"}}>
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
<Toggle
label={scaleUnit === "imperial" ? "Imperial" : "Metric"}
checked={scaleUnit === "imperial"}
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
onChange={(e) => onScaleUnitChange(e.currentTarget.checked ? "imperial" : "metric")}
/>
</div>
</div>}
</div>;
}
const PIN_COLORS = ["red", "blue", "green", "yellow", "orange", "purple"];
function AddPinDialog({lat, lon, onSubmit, onCancel}) {
const [label, setLabel] = useState("");
const [color, setColor] = useState("red");
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
// A floating dialog over the canvas — page chrome, not map data, so it is never
// filtered and keeps flat, bordered surfaces like the rest of the app (no shadow).
return <div style={{
position: "absolute", top: "50%", left: "50%", transform: "translate(-50%, -50%)",
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
zIndex: 1002, background: "var(--panel)", border: "1px solid var(--border)", color: "var(--text)",
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
padding: "1rem", minWidth: "15.625rem",
}}>
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
<div style={{fontWeight: 600, marginBottom: "0.5rem"}}>Add Pin</div>
<div style={{fontSize: '0.75rem', color: "var(--muted)", marginBottom: "0.5rem"}}>{lat.toFixed(4)}, {lon.toFixed(4)}</div>
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
<TextInput
placeholder="Label"
value={label}
onChange={e => setLabel(e.target.value)}
autoFocus
onKeyDown={e => e.key === "Enter" && label.trim() && onSubmit(label.trim(), color)}
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
style={{marginBottom: "0.5rem"}}
/>
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
<div style={{display: "flex", gap: "0.375rem", marginBottom: "0.75rem"}}>
{PIN_COLORS.map(c =>
<div
key={c}
onClick={() => setColor(c)}
style={{
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
width: "1.5rem", height: "1.5rem", borderRadius: "50%", background: c, cursor: "pointer",
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
border: color === c ? "3px solid var(--text)" : "2px solid var(--border)",
}}
/>
)}
</div>
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
<div style={{display: "flex", gap: "0.5rem", justifyContent: "flex-end"}}>
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
<Button role='cancel' size='sm' onClick={onCancel}>Cancel</Button>
<Button
role='primary'
size='sm'
onClick={() => label.trim() && onSubmit(label.trim(), color)}
disabled={!label.trim()}
>
Add
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
</Button>
</div>
</div>;
}
// Coordinate regex: "45.5, -122.6" or "45.5 -122.6"
const COORD_REGEX = /^\s*(-?\d+\.?\d*)\s*[,\s]\s*(-?\d+\.?\d*)\s*$/;
function MapSearch({map}) {
const [query, setQuery] = useState("");
const [results, setResults] = useState([]);
const [showResults, setShowResults] = useState(false);
const [searchMarker, setSearchMarker] = useState(null);
const debounceRef = useRef(null);
const containerRef = useRef(null);
// Close results when clicking outside.
useEffect(() => {
const handler = (e) => {
if (containerRef.current && !containerRef.current.contains(e.target)) {
setShowResults(false);
}
};
document.addEventListener("click", handler);
return () => document.removeEventListener("click", handler);
}, []);
const doSearch = useCallback((q) => {
if (!q || q.length < 2) {
setResults([]);
setShowResults(false);
return;
}
// Check for coordinate input first.
const coordMatch = q.match(COORD_REGEX);
if (coordMatch) {
const lat = parseFloat(coordMatch[1]);
const lon = parseFloat(coordMatch[2]);
if (lat >= -90 && lat <= 90 && lon >= -180 && lon <= 180) {
setResults([{name: `${lat}, ${lon}`, kind: "coordinates", lat, lon, min_zoom: 0, source: "input"}]);
setShowResults(true);
return;
}
}
// Get current map center for proximity bias.
let lat = null, lon = null;
if (map) {
const center = map.getCenter();
lat = center.lat;
lon = center.lng;
}
app: api helpers report failure instead of returning an empty value (#736) Six helpers turned a non-OK response into an empty value, so their consumers rendered a server error as "nothing here": * getConfigBackups -> {dates: []}: "No backups available for this config." A thrown request also left "Loading backups..." forever. * getInventoryBackups -> []: "No backups yet." * getDownloaders -> {downloaders: []} on a throw, the error body on a non-OK response (which then crashed the .map): the RSS form's Downloader select sat empty. * searchMap -> {results: []}: "No map locations found". * getMapSearchStatus -> {indexed: [], missing: []}: "No maps are installed". * flasherSearch -> [null, null], then `|| []`: "No .bin firmware found". Invariants now enforced and tested: * On a non-OK response each helper returns undefined; an empty value only ever comes from the server. * Each consumer shows an error for a failed or thrown fetch and keeps its "none" message for a confirmed empty result. The Downloader select is disabled and says it is loading or could not fetch. A failed map status is unknown, never "no maps are installed". * The map search tab loads from the first render, and a query shorter than the two characters the request needs now says so instead of "No map locations found". Archive browsers and recent tags are left as they are: a failed fetch there hides an optional picker or suggestion list without claiming anything. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-26 17:47:32 -06:00
// A failed search (undefined) shows no dropdown, same as finding nothing: this is an
// as-you-type suggestion list, not a results page.
searchMap(q, 8, 0, lat, lon).then(data => {
app: api helpers report failure instead of returning an empty value (#736) Six helpers turned a non-OK response into an empty value, so their consumers rendered a server error as "nothing here": * getConfigBackups -> {dates: []}: "No backups available for this config." A thrown request also left "Loading backups..." forever. * getInventoryBackups -> []: "No backups yet." * getDownloaders -> {downloaders: []} on a throw, the error body on a non-OK response (which then crashed the .map): the RSS form's Downloader select sat empty. * searchMap -> {results: []}: "No map locations found". * getMapSearchStatus -> {indexed: [], missing: []}: "No maps are installed". * flasherSearch -> [null, null], then `|| []`: "No .bin firmware found". Invariants now enforced and tested: * On a non-OK response each helper returns undefined; an empty value only ever comes from the server. * Each consumer shows an error for a failed or thrown fetch and keeps its "none" message for a confirmed empty result. The Downloader select is disabled and says it is loading or could not fetch. A failed map status is unknown, never "no maps are installed". * The map search tab loads from the first render, and a query shorter than the two characters the request needs now says so instead of "No map locations found". Archive browsers and recent tags are left as they are: a failed fetch there hides an optional picker or suggestion list without claiming anything. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-26 17:47:32 -06:00
const found = data?.results || [];
setResults(found);
setShowResults(found.length > 0);
}).catch(e => console.error(e));
}, [map]);
const handleInput = useCallback((e) => {
const q = e.target.value;
setQuery(q);
clearTimeout(debounceRef.current);
debounceRef.current = setTimeout(() => doSearch(q), 300);
}, [doSearch]);
const handleSelect = useCallback((result) => {
if (!map) return;
// Remove old search marker.
if (searchMarker) {
searchMarker.remove();
}
const zoom = result.kind === "coordinates" ? 14 : Math.max(map.getZoom(), 12);
map.flyTo({center: [result.lon, result.lat], zoom});
const marker = new maplibregl.Marker({color: "#6435c9"})
.setLngLat([result.lon, result.lat])
.setPopup(new maplibregl.Popup({offset: 25}).setText(result.name))
.addTo(map);
marker.togglePopup();
setSearchMarker(marker);
setShowResults(false);
setQuery(result.name);
}, [map, searchMarker]);
const handleKeyDown = useCallback((e) => {
if (e.key === "Escape") {
setShowResults(false);
e.target.blur();
}
}, []);
const formatPopulation = (pop) => {
if (!pop) return "";
if (pop >= 1_000_000) return `${(pop / 1_000_000).toFixed(1)}M`;
if (pop >= 1_000) return `${(pop / 1_000).toFixed(0)}K`;
return `${pop}`;
};
const resultSubtext = (r) => {
if (r.kind === "coordinates") return "";
const parts = [];
if (r.region) parts.push(r.region);
const detail = r.kind_detail || r.kind || "";
if (detail && !r.region) parts.push(detail.replace(/_/g, " "));
const pop = formatPopulation(r.population);
if (pop) parts.push(`pop. ${pop}`);
return parts.join(" \u00b7 ");
};
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
// Floating search chrome over the canvas — page interface, not map data, so it is
// never filtered. Flat surfaces: bordered, no box-shadow, token colors only.
return <div ref={containerRef} style={{
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
position: "absolute", top: "0.625rem", left: "50%", transform: "translateX(-50%)",
zIndex: 5, width: "20rem", maxWidth: "calc(100% - 100px)",
}}>
<input
type="text"
placeholder="Search places..."
value={query}
onChange={handleInput}
onFocus={() => results.length > 0 && setShowResults(true)}
onKeyDown={handleKeyDown}
style={{
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
width: "100%", padding: "0.5rem 0.75rem", border: "1px solid var(--border)",
ui: scale the whole interface by 10%, off the root font-size The interface reads about 10% smaller than the pre-migration Semantic build, and the drift is not one number. Measured on the QA Pi at 1280px, Semantic's headings were h1 28 / h2 24 / h3 18 / h4 15 / h5 14 against our 26 / 21 / 17 / 14 / 13, and its video card a fixed 290px with an 18px title against our 238px with a 13px one -- while our BODY text and chrome went the other way, a 16px base against Semantic's 14 and 42px buttons against 36. No per-component nudge reproduces what a user comparing the two sees, which is the whole page at once. Ten percent is not a fudge factor: our heading scale times 1.1 is 28.6 / 23.1 / 18.7 / 15.4 / 14.3, which is Semantic's scale to within a pixel at every step. Verified in the browser -- those are the numbers it now renders. `--ui-scale: 1.1` and `html { font-size: calc(100% * var(--ui-scale)) }`, with everything that should grow written in rem: - 102 lengths in ui.css, App.css and tokens.css converted, restricted by property: font-size, padding, margin, gap, width/height, border-radius, offsets. - Borders, outlines and shadows deliberately NOT converted. A hairline is a hairline at any text size. - 46 inline `fontSize` px values across 8 files. `em` values were left alone: they already resolve against the parent. - Ten inline-styled blocks in the library (Card's text block, Loading, Placeholder, ColoredInput, Confirm's action row, CardGroup's gap) moved into ui.css, because an inline px size cannot be reached by a scale or seen by a test. - CardGroup converts its `minWidth` prop to rem, which is what returns a phone to one card per row: as px the track ignored the scale, so a 430px screen fitted two 200px cards where the old build fitted one 290px card. Mantine needs nothing: its 644 size declarations are `calc(Xrem * var(--mantine-scale))` and scale with the root. `theme.scale` must stay at 1 -- it exists to CANCEL a customised root font-size, so setting it to 1.1 as well would leave Mantine's half of the interface at the old size beside our larger text. A test pins that. `zoom: 1.1` on the root was tried first and rejected. It works and it is one line, but it scales at paint: every 1px hairline becomes 1.1px and softens on a 1x display, and viewport units do not shrink to compensate, so `100vh` paints 110% of the screen -- a 95vw modal overflowed sideways and the CBZ viewer's fullscreen overflowed down. Rem scales real lengths, keeps borders crisp, leaves viewport units alone, and preserves a user's own browser font-size preference instead of overriding it. A percentage rather than a px root size, for that last reason. Five guards in themes/ui-scale.test.js, each planted against: the scale itself, that every scalable declaration is rem, that borders are NOT, that no inline px font-size survives anywhere including the gallery, and that Mantine's scale stays 1. Also removes `.ui.modal.themed-confirm`, a dead Semantic-era rule and the source of the one `95vw` that would have overflowed: `.ui.modal` no longer exists and nothing sets that class. Verified: 1141 tests / 65 suites; clean build; browser-checked at 1280px and at a 430px phone; navbar single-row from 780 to 1800px; hairlines still 1px and no horizontal overflow in all four themes. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-03 14:45:08 -06:00
fontSize: '0.875rem', background: "var(--panel)", color: "var(--text)",
}}
/>
{showResults && results.length > 0 && <div style={{
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
background: "var(--panel)", border: "1px solid var(--border)", borderTop: "none",
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
maxHeight: "18.75rem", overflowY: "auto",
}}>
{results.map((r, i) => <div
key={i}
onClick={() => handleSelect(r)}
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
onMouseEnter={e => e.currentTarget.style.background = "var(--head)"}
onMouseLeave={e => e.currentTarget.style.background = "transparent"}
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
style={{padding: "0.5rem 0.75rem", cursor: "pointer", borderBottom: "1px solid var(--border)"}}
>
<div style={{fontWeight: 500}}>{r.name}</div>
ui: scale the whole interface by 10%, off the root font-size The interface reads about 10% smaller than the pre-migration Semantic build, and the drift is not one number. Measured on the QA Pi at 1280px, Semantic's headings were h1 28 / h2 24 / h3 18 / h4 15 / h5 14 against our 26 / 21 / 17 / 14 / 13, and its video card a fixed 290px with an 18px title against our 238px with a 13px one -- while our BODY text and chrome went the other way, a 16px base against Semantic's 14 and 42px buttons against 36. No per-component nudge reproduces what a user comparing the two sees, which is the whole page at once. Ten percent is not a fudge factor: our heading scale times 1.1 is 28.6 / 23.1 / 18.7 / 15.4 / 14.3, which is Semantic's scale to within a pixel at every step. Verified in the browser -- those are the numbers it now renders. `--ui-scale: 1.1` and `html { font-size: calc(100% * var(--ui-scale)) }`, with everything that should grow written in rem: - 102 lengths in ui.css, App.css and tokens.css converted, restricted by property: font-size, padding, margin, gap, width/height, border-radius, offsets. - Borders, outlines and shadows deliberately NOT converted. A hairline is a hairline at any text size. - 46 inline `fontSize` px values across 8 files. `em` values were left alone: they already resolve against the parent. - Ten inline-styled blocks in the library (Card's text block, Loading, Placeholder, ColoredInput, Confirm's action row, CardGroup's gap) moved into ui.css, because an inline px size cannot be reached by a scale or seen by a test. - CardGroup converts its `minWidth` prop to rem, which is what returns a phone to one card per row: as px the track ignored the scale, so a 430px screen fitted two 200px cards where the old build fitted one 290px card. Mantine needs nothing: its 644 size declarations are `calc(Xrem * var(--mantine-scale))` and scale with the root. `theme.scale` must stay at 1 -- it exists to CANCEL a customised root font-size, so setting it to 1.1 as well would leave Mantine's half of the interface at the old size beside our larger text. A test pins that. `zoom: 1.1` on the root was tried first and rejected. It works and it is one line, but it scales at paint: every 1px hairline becomes 1.1px and softens on a 1x display, and viewport units do not shrink to compensate, so `100vh` paints 110% of the screen -- a 95vw modal overflowed sideways and the CBZ viewer's fullscreen overflowed down. Rem scales real lengths, keeps borders crisp, leaves viewport units alone, and preserves a user's own browser font-size preference instead of overriding it. A percentage rather than a px root size, for that last reason. Five guards in themes/ui-scale.test.js, each planted against: the scale itself, that every scalable declaration is rem, that borders are NOT, that no inline px font-size survives anywhere including the gallery, and that Mantine's scale stays 1. Also removes `.ui.modal.themed-confirm`, a dead Semantic-era rule and the source of the one `95vw` that would have overflowed: `.ui.modal` no longer exists and nothing sets that class. Verified: 1141 tests / 65 suites; clean build; browser-checked at 1280px and at a 430px phone; navbar single-row from 780 to 1800px; hairlines still 1px and no horizontal overflow in all four themes. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-03 14:45:08 -06:00
{resultSubtext(r) && <div style={{fontSize: '0.6875rem', color: "var(--muted)"}}>{resultSubtext(r)}</div>}
</div>)}
</div>}
</div>;
}
export default function MapViewer() {
const mapContainer = useRef(null);
const mapRef = useRef(null);
const scaleControlRef = useRef(null);
const activeFilesRef = useRef([]);
const hasTerrainRef = useRef(false);
const layerVisibilityRef = useRef(null);
const [mapReady, setMapReady] = useState(false);
const [error, setError] = useState(null);
const [contextMenu, setContextMenu] = useState(null);
const [addingPin, setAddingPin] = useState(null); // {lat, lon} when pin dialog is open
const markersRef = useRef([]);
const pinsVisibleRef = useRef(true);
const [scaleUnit, setScaleUnit] = useState(() => {
try { return localStorage.getItem(STORAGE_KEY_SCALE) || "metric"; } catch (e) { return "metric"; }
});
const [searchParams, setSearchParams] = useSearchParams();
const {theme} = useContext(ThemeContext);
const {settings, fetchSettings} = useContext(SettingsContext);
// Use URL params if provided, otherwise fall back to saved default location, then 0,0,z2.
const hasUrlParams = searchParams.has("lat") || searchParams.has("lon") || searchParams.has("z");
const defaultLoc = settings?.map_default_location;
const initialLat = hasUrlParams
? (parseFloat(searchParams.get("lat")) || 0)
: (parseFloat(defaultLoc?.lat) || 0);
const initialLon = hasUrlParams
? (parseFloat(searchParams.get("lon")) || 0)
: (parseFloat(defaultLoc?.lon) || 0);
const initialZoom = hasUrlParams
? (parseFloat(searchParams.get("z")) || 2)
: (parseFloat(defaultLoc?.zoom) || 2);
useEffect(() => {
if (!mapContainer.current) return;
let map;
let destroyed = false;
const init = async () => {
const mapSources = [];
// Fetch user's PMTiles files first to determine what base layer is needed.
let hasTerrain = false;
let terrainFileName = null;
let planetFile = null;
try {
const data = await getMapFiles();
if (data?.files) {
const regionFiles = [];
for (const f of data.files) {
if (f.name.startsWith(TERRAIN_PREFIX) && f.name.endsWith('.pmtiles')) {
hasTerrain = true;
terrainFileName = f.name;
} else if (isPlanetFile(f.name)) {
if (!planetFile || f.size > planetFile.size) {
planetFile = f;
}
} else {
regionFiles.push(f);
}
}
if (planetFile) {
// Planet covers everything — sole vector source, no overview needed.
mapSources.push({name: planetFile.name, url: `pmtiles:///media/map/${planetFile.name}`});
} else {
// Use regional extracts.
for (const f of regionFiles) {
mapSources.push({name: f.name, url: `pmtiles:///media/map/${f.name}`});
}
// Add planet overview blob as low-zoom base layer (if available).
try {
const resp = await fetch(MAP_OVERVIEW_URL, {method: "HEAD"});
if (resp.ok) {
mapSources.unshift({name: "map-overview", url: `pmtiles://${MAP_OVERVIEW_URL}`});
}
} catch (e) {
console.debug("Planet overview not available:", e);
}
}
}
} catch (e) {
console.error("Failed to fetch map files:", e);
}
if (destroyed) return;
activeFilesRef.current = mapSources;
hasTerrainRef.current = hasTerrain;
// Initialize contour source if terrain data is available.
if (hasTerrain) {
initContourSource(terrainFileName);
}
try {
map = new maplibregl.Map({
container: mapContainer.current,
ui: the monochrome themes draw a black basemap, not a dark one Roland asked why the red overlay sits on the DARK basemap rather than greyscale or black. Measuring it, black is the better answer and the reason is not brightness: flavor distinct colours hued earth luminance dark 19 4 0.014 black 13 0 0.007 Both media filters are the same pure luminance projection -- `0.2126 0.7152 0.0722` -- so they keep brightness and throw hue away. `black` is already achromatic, so filtering it is a hue rotation and nothing else: all thirteen greys were authored as a monochrome hierarchy and every one arrives intact. `dark` has more colours, but four of them are hued (water, parks, up to 25 chroma) and those distinctions do not survive at all -- the features land at whatever brightness their hue happened to carry rather than one anybody chose. Black also emits half the light on the earth fill, which is the point of a night-vision theme. `grayscale` is not an option despite the name: it is a LIGHT theme (earth L0.604), and a luminance filter turns a light map into a bright red one. What made black look impossible in the first commit was the sprite, and that was my mistake: `sprite` and the layer flavor are separate style properties that merely shared one variable. We ship `light` and `dark` sprite sets, and a `black` basemap with `dark` icons is an ordinary style. They are now separate arguments, with `mapSprite()` beside `mapFlavor()`. Fixed in passing: the contour labels branched on `flavor === "dark"` and would have taken the LIGHT branch for `black` -- dark brown text on a black map. Now branches on `light`, so every dark flavor gets the pale label. The new test is the one worth having: the monochrome themes' flavor must paint with greys only, paired with an assertion that `dark` is NOT achromatic. Without that pair the first would pass on any flavor at all and prove nothing about the choice. 1096 jest / 61 suites, 4 e2e themes, clean build. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-02 13:33:51 -06:00
style: buildStyle(mapSources, mapFlavor(theme), hasTerrain, scaleUnit, mapSprite(theme)),
center: [initialLon, initialLat],
zoom: initialZoom,
attributionControl: true,
preserveDrawingBuffer: true,
});
map.addControl(new maplibregl.NavigationControl(), "top-right");
map.addControl(new maplibregl.FullscreenControl(), "top-right");
const sc = new maplibregl.ScaleControl({maxWidth: 200, unit: scaleUnit});
map.addControl(sc, "bottom-left");
scaleControlRef.current = sc;
map.addControl(new maplibregl.GeolocateControl({
positionOptions: {enableHighAccuracy: true},
trackUserLocation: true,
showUserHeading: true,
showAccuracyCircle: true,
}), "top-right");
// Pin toggle button.
const pinToggle = document.createElement("button");
pinToggle.type = "button";
pinToggle.title = "Toggle pins";
pinToggle.textContent = "📍";
pinToggle.style.cssText = "font-size:18px;width:29px;height:29px;border:none;cursor:pointer;background:white;border-radius:4px;display:flex;align-items:center;justify-content:center;";
pinToggle.addEventListener("click", () => {
pinsVisibleRef.current = !pinsVisibleRef.current;
const display = pinsVisibleRef.current ? "" : "none";
pinToggle.style.opacity = pinsVisibleRef.current ? "1" : "0.5";
for (const m of markersRef.current) {
m.getElement().style.display = display;
}
});
const pinControl = {
onAdd: () => {
const container = document.createElement("div");
container.className = "maplibregl-ctrl maplibregl-ctrl-group";
container.appendChild(pinToggle);
return container;
},
onRemove: () => {},
};
map.addControl(pinControl, "top-right");
// Right-click context menu (desktop).
map.on("contextmenu", (e) => {
if (destroyed) return;
setContextMenu({
x: e.point.x,
y: e.point.y,
lat: e.lngLat.lat,
lon: e.lngLat.lng,
});
});
map.on("click", () => setContextMenu(null));
map.on("movestart", () => setContextMenu(null));
// Long-press context menu (mobile).
let longPressTimer = null;
let touchStartPos = null;
const canvas = map.getCanvas();
canvas.addEventListener("touchstart", (e) => {
if (e.touches.length !== 1) return;
touchStartPos = {x: e.touches[0].clientX, y: e.touches[0].clientY};
longPressTimer = setTimeout(() => {
const rect = canvas.getBoundingClientRect();
const point = {x: touchStartPos.x - rect.left, y: touchStartPos.y - rect.top};
const lngLat = map.unproject(point);
setContextMenu({x: point.x, y: point.y, lat: lngLat.lat, lon: lngLat.lng});
}, 500);
}, {passive: true});
canvas.addEventListener("touchmove", (e) => {
if (longPressTimer && touchStartPos) {
const dx = e.touches[0].clientX - touchStartPos.x;
const dy = e.touches[0].clientY - touchStartPos.y;
if (Math.abs(dx) > 10 || Math.abs(dy) > 10) {
clearTimeout(longPressTimer);
longPressTimer = null;
}
}
});
canvas.addEventListener("touchend", () => {
clearTimeout(longPressTimer);
longPressTimer = null;
});
// Update URL on pan/zoom (only while mounted).
map.on("moveend", () => {
if (destroyed) return;
const center = map.getCenter();
const zoom = map.getZoom();
window.history.replaceState(null, '',
`${window.location.pathname}?lat=${center.lat.toFixed(4)}&lon=${center.lng.toFixed(4)}&z=${zoom.toFixed(1)}`
);
});
map.on("load", () => {
if (!destroyed) {
mapRef.current = map;
// Apply saved layer visibility on initial load.
const savedVis = getInitialVisibility();
const allMapLayers = map.getStyle()?.layers || [];
for (const [groupName, visible] of Object.entries(savedVis)) {
const prefixes = LAYER_GROUPS[groupName];
if (!prefixes) continue;
for (const layer of allMapLayers) {
const baseId = layer.id.includes(":") ? layer.id.split(":").slice(1).join(":") : layer.id;
if (prefixes.some(p => baseId.includes(p))) {
map.setLayoutProperty(layer.id, "visibility", visible ? "visible" : "none");
}
}
}
setMapReady(true);
}
});
} catch (e) {
console.error("MapLibre failed to initialize:", e);
if (!destroyed) setError(e.message);
}
};
init();
return () => {
destroyed = true;
if (map) {
map.remove();
}
mapRef.current = null;
setMapReady(false);
// Clean up map params from URL so they don't leak to other pages.
const url = new URL(window.location);
url.searchParams.delete("lat");
url.searchParams.delete("lon");
url.searchParams.delete("z");
window.history.replaceState(null, '', url.pathname + url.search);
};
}, []); // eslint-disable-line react-hooks/exhaustive-deps
// Rebuild style when theme changes.
useEffect(() => {
const map = mapRef.current;
if (!map || !mapReady) return;
const center = map.getCenter();
const zoom = map.getZoom();
ui: the monochrome themes draw a black basemap, not a dark one Roland asked why the red overlay sits on the DARK basemap rather than greyscale or black. Measuring it, black is the better answer and the reason is not brightness: flavor distinct colours hued earth luminance dark 19 4 0.014 black 13 0 0.007 Both media filters are the same pure luminance projection -- `0.2126 0.7152 0.0722` -- so they keep brightness and throw hue away. `black` is already achromatic, so filtering it is a hue rotation and nothing else: all thirteen greys were authored as a monochrome hierarchy and every one arrives intact. `dark` has more colours, but four of them are hued (water, parks, up to 25 chroma) and those distinctions do not survive at all -- the features land at whatever brightness their hue happened to carry rather than one anybody chose. Black also emits half the light on the earth fill, which is the point of a night-vision theme. `grayscale` is not an option despite the name: it is a LIGHT theme (earth L0.604), and a luminance filter turns a light map into a bright red one. What made black look impossible in the first commit was the sprite, and that was my mistake: `sprite` and the layer flavor are separate style properties that merely shared one variable. We ship `light` and `dark` sprite sets, and a `black` basemap with `dark` icons is an ordinary style. They are now separate arguments, with `mapSprite()` beside `mapFlavor()`. Fixed in passing: the contour labels branched on `flavor === "dark"` and would have taken the LIGHT branch for `black` -- dark brown text on a black map. Now branches on `light`, so every dark flavor gets the pale label. The new test is the one worth having: the monochrome themes' flavor must paint with greys only, paired with an assertion that `dark` is NOT achromatic. Without that pair the first would pass on any flavor at all and prove nothing about the choice. 1096 jest / 61 suites, 4 e2e themes, clean build. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-02 13:33:51 -06:00
map.setStyle(buildStyle(activeFilesRef.current, mapFlavor(theme), hasTerrainRef.current,
scaleUnit, mapSprite(theme)));
map.once("style.load", () => {
map.setCenter(center);
map.setZoom(zoom);
// Re-apply layer visibility state after style rebuild.
const vis = layerVisibilityRef.current;
if (vis) {
const allMapLayers = map.getStyle()?.layers || [];
for (const [groupName, visible] of Object.entries(vis)) {
const prefixes = LAYER_GROUPS[groupName];
if (!prefixes) continue;
for (const layer of allMapLayers) {
const baseId = layer.id.includes(":") ? layer.id.split(":").slice(1).join(":") : layer.id;
if (prefixes.some(p => baseId.includes(p))) {
map.setLayoutProperty(layer.id, "visibility", visible ? "visible" : "none");
}
}
}
}
});
}, [theme, mapReady, scaleUnit]);
const handleScaleUnitChange = useCallback((unit) => {
const map = mapRef.current;
const sc = scaleControlRef.current;
if (map && sc) {
map.removeControl(sc);
const newSc = new maplibregl.ScaleControl({maxWidth: 200, unit});
map.addControl(newSc, "bottom-left");
scaleControlRef.current = newSc;
}
try { localStorage.setItem(STORAGE_KEY_SCALE, unit); } catch (e) { /* ignore */ }
setScaleUnit(unit);
}, []);
const handleCopyCoords = useCallback(() => {
if (!contextMenu) return;
// Match decimal precision to current zoom level so pasting into search doesn't over-zoom.
const map = mapRef.current;
const zoom = map ? map.getZoom() : 10;
const decimals = zoom < 4 ? 1 : zoom < 8 ? 2 : zoom < 12 ? 3 : 4;
const text = `${contextMenu.lat.toFixed(decimals)}, ${contextMenu.lon.toFixed(decimals)}`;
navigator.clipboard.writeText(text).catch(() => {
window.prompt("Copy coordinates:", text);
});
setContextMenu(null);
}, [contextMenu]);
const handleExportPng = useCallback(() => {
const map = mapRef.current;
if (!map) return;
setContextMenu(null);
// Force a fresh render, then capture.
map.once("render", function() {
const c = map.getCenter();
const z = map.getZoom();
const filename = `wrolpi-map_${c.lat.toFixed(4)}_${c.lng.toFixed(4)}_z${z.toFixed(0)}.png`;
map.getCanvas().toBlob(function(blob) {
if (!blob) return;
const url = URL.createObjectURL(blob);
const a = document.createElement("a");
a.href = url;
a.download = filename;
a.click();
URL.revokeObjectURL(url);
});
});
map.triggerRepaint();
}, []);
// --- Pins ---
const loadPins = useCallback(async () => {
const map = mapRef.current;
if (!map) return;
// Remove existing markers.
for (const m of markersRef.current) m.remove();
markersRef.current = [];
const data = await getMapPins();
if (!data?.pins) return;
data.pins.forEach((pin, index) => {
const el = document.createElement("div");
el.style.width = "24px";
el.style.height = "24px";
el.style.borderRadius = "50% 50% 50% 0";
el.style.background = pin.color || "red";
el.style.border = "2px solid white";
el.style.boxShadow = "0 1px 4px rgba(0,0,0,0.4)";
el.style.transform = "rotate(-45deg)";
el.style.cursor = "pointer";
const popup = new maplibregl.Popup({offset: 15}).setHTML(
`<div style="font-size:13px">` +
`<strong>${pin.label || "Pin"}</strong><br/>` +
`<small>${pin.lat.toFixed(4)}, ${pin.lon.toFixed(4)}</small><br/>` +
`<button onclick="window.__deleteMapPin(${pin.id})" ` +
`style="margin-top:4px;padding:2px 8px;cursor:pointer;color:red;border:1px solid red;background:none;border-radius:3px;">` +
`Delete</button></div>`
);
const marker = new maplibregl.Marker({element: el})
.setLngLat([pin.lon, pin.lat])
.setPopup(popup)
.addTo(map);
if (!pinsVisibleRef.current) {
marker.getElement().style.display = "none";
}
markersRef.current.push(marker);
});
}, []);
// Global callback for popup delete buttons.
useEffect(() => {
window.__deleteMapPin = async (pinId) => {
await deleteMapPin(pinId);
await loadPins();
};
return () => { delete window.__deleteMapPin; };
}, [loadPins]);
// Load pins when map is ready.
useEffect(() => {
if (mapReady) loadPins();
}, [mapReady, loadPins]);
const handleAddPin = useCallback(() => {
if (!contextMenu) return;
setAddingPin({lat: contextMenu.lat, lon: contextMenu.lon});
setContextMenu(null);
}, [contextMenu]);
const handleSetDefaultLocation = useCallback(async () => {
const map = mapRef.current;
if (!map) return;
const center = map.getCenter();
const zoom = map.getZoom();
await setMapDefaultLocation(center.lat, center.lng, zoom);
await fetchSettings();
setContextMenu(null);
}, [fetchSettings]);
const handleSubmitPin = useCallback(async (label, color) => {
if (!addingPin) return;
const {lat, lon} = addingPin;
setAddingPin(null); // Prevent double-submit.
await addMapPin(lat, lon, label, color);
await loadPins();
}, [addingPin, loadPins]);
if (error) {
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
return <Panel>
<Header as="h3">Map failed to load</Header>
<p>Your browser may not support WebGL. You can access the standalone map viewer
at <a href={MAP_VIEWER_URI}>{MAP_VIEWER_URI}</a></p>
<p>Error: {error}</p>
ui: thirty-five files off Semantic, leaving fifteen Phase 2, part 4. Baseline 50 -> 15. Migrated: all eight calculators, the rest of inventory and collections, the Map/Zim/Docs cluster, Videos, Files, Search, Archive, DashboardPage, DatesSelector, HelpModal, Events, Tags, Flasher's neighbours, admin Settings and Status, Download, and App. Also the three shared modules everything imports: api.js and customHooks.js swap the Semantic toast library for the library's shim (same call shape), and useForm.js renders the library's Message. useForm accepts both the old `{content, header, positive|negative}` shape and the new `{kind, title}`, because the pages building those objects migrate separately. Two library gaps closed, both found by call sites needing them: ActionInput is a forwardRef now. Its ref reaches the input rather than the wrapper, because callers use it to focus or select the field -- a ref to the surrounding div would do neither, which is why one call site had resorted to hand-copying the component's markup. Header takes a `color` token name. Three separate call sites had reached for an inline style to colour a heading that carries meaning ("3 items to add" against "2 to remove"), which is how a hex ends up in markup. Two files needed repair after being interrupted mid-edit: MapViewer had a Segment left in its error branch, and VideoPlayer had lost the useContext import its own code still used. Neither jest nor tsc noticed; only `npm run build` did, which is the whole reason that check exists. Fifteen files remain: Archive, BulkTagModal, Channels, Docs, FileBrowser, FilePreview, Flasher, Nav, Playlists, Theme.tsx, Upload, VideoPlayer, Zim, ControllerPage, admin/Downloads. Theme.tsx is deliberately last -- the unmigrated files still import its Semantic-backed wrappers, so it can only go once they have.
2026-07-30 12:21:30 -06:00
</Panel>;
}
return <div ref={mapContainer} style={{position: "relative", width: "100%", height: "85vh"}}>
{mapReady && <MapSearch map={mapRef.current}/>}
{mapReady && <LayerControl map={mapRef.current} scaleUnit={scaleUnit} onScaleUnitChange={handleScaleUnitChange} visibilityRef={layerVisibilityRef}/>}
{contextMenu && <div
style={{
position: "absolute",
left: contextMenu.x,
top: contextMenu.y,
zIndex: 1001,
background: "white",
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
borderRadius: "0.375rem",
boxShadow: "0 2px 8px rgba(0,0,0,0.3)",
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
padding: "0.25rem 0",
ui: scale the whole interface by 10%, off the root font-size The interface reads about 10% smaller than the pre-migration Semantic build, and the drift is not one number. Measured on the QA Pi at 1280px, Semantic's headings were h1 28 / h2 24 / h3 18 / h4 15 / h5 14 against our 26 / 21 / 17 / 14 / 13, and its video card a fixed 290px with an 18px title against our 238px with a 13px one -- while our BODY text and chrome went the other way, a 16px base against Semantic's 14 and 42px buttons against 36. No per-component nudge reproduces what a user comparing the two sees, which is the whole page at once. Ten percent is not a fudge factor: our heading scale times 1.1 is 28.6 / 23.1 / 18.7 / 15.4 / 14.3, which is Semantic's scale to within a pixel at every step. Verified in the browser -- those are the numbers it now renders. `--ui-scale: 1.1` and `html { font-size: calc(100% * var(--ui-scale)) }`, with everything that should grow written in rem: - 102 lengths in ui.css, App.css and tokens.css converted, restricted by property: font-size, padding, margin, gap, width/height, border-radius, offsets. - Borders, outlines and shadows deliberately NOT converted. A hairline is a hairline at any text size. - 46 inline `fontSize` px values across 8 files. `em` values were left alone: they already resolve against the parent. - Ten inline-styled blocks in the library (Card's text block, Loading, Placeholder, ColoredInput, Confirm's action row, CardGroup's gap) moved into ui.css, because an inline px size cannot be reached by a scale or seen by a test. - CardGroup converts its `minWidth` prop to rem, which is what returns a phone to one card per row: as px the track ignored the scale, so a 430px screen fitted two 200px cards where the old build fitted one 290px card. Mantine needs nothing: its 644 size declarations are `calc(Xrem * var(--mantine-scale))` and scale with the root. `theme.scale` must stay at 1 -- it exists to CANCEL a customised root font-size, so setting it to 1.1 as well would leave Mantine's half of the interface at the old size beside our larger text. A test pins that. `zoom: 1.1` on the root was tried first and rejected. It works and it is one line, but it scales at paint: every 1px hairline becomes 1.1px and softens on a 1x display, and viewport units do not shrink to compensate, so `100vh` paints 110% of the screen -- a 95vw modal overflowed sideways and the CBZ viewer's fullscreen overflowed down. Rem scales real lengths, keeps borders crisp, leaves viewport units alone, and preserves a user's own browser font-size preference instead of overriding it. A percentage rather than a px root size, for that last reason. Five guards in themes/ui-scale.test.js, each planted against: the scale itself, that every scalable declaration is rem, that borders are NOT, that no inline px font-size survives anywhere including the gallery, and that Mantine's scale stays 1. Also removes `.ui.modal.themed-confirm`, a dead Semantic-era rule and the source of the one `95vw` that would have overflowed: `.ui.modal` no longer exists and nothing sets that class. Verified: 1141 tests / 65 suites; clean build; browser-checked at 1280px and at a 430px phone; navbar single-row from 780 to 1800px; hairlines still 1px and no horizontal overflow in all four themes. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-03 14:45:08 -06:00
fontSize: '0.8125rem',
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
minWidth: "11.25rem",
}}
>
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
<div style={{padding: "0.375rem 0.75rem", color: "#666", fontSize: '0.75rem', borderBottom: "1px solid #eee"}}>
{(() => {
const map = mapRef.current;
const zoom = map ? map.getZoom() : 10;
const d = zoom < 4 ? 1 : zoom < 8 ? 2 : zoom < 12 ? 3 : 4;
return `${contextMenu.lat.toFixed(d)}, ${contextMenu.lon.toFixed(d)}`;
})()}
</div>
<div
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
style={{padding: "0.5rem 0.75rem", cursor: "pointer", userSelect: "none"}}
onMouseEnter={e => e.target.style.background = "#f0f0f0"}
onMouseLeave={e => e.target.style.background = "transparent"}
onClick={handleCopyCoords}
>
Copy GPS coordinates
</div>
<div
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
style={{padding: "0.5rem 0.75rem", cursor: "pointer", userSelect: "none"}}
onMouseEnter={e => e.target.style.background = "#f0f0f0"}
onMouseLeave={e => e.target.style.background = "transparent"}
onClick={handleExportPng}
>
Export as PNG
</div>
<div
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
style={{padding: "0.5rem 0.75rem", cursor: "pointer", userSelect: "none"}}
onMouseEnter={e => e.target.style.background = "#f0f0f0"}
onMouseLeave={e => e.target.style.background = "transparent"}
onClick={handleAddPin}
>
Add Pin Here
</div>
<div
review: close the mixed-unit gaps the scale left behind Seven review comments, all valid. The common thread is the one the guard was supposed to catch and did not: geometry converted to rem beside geometry that stayed px. The guard was narrower than its name. It missed: - `transform`, `flex` and `flex-basis`, so `translateX(26px)` and `flex: 1 1 200px` were invisible on properties nobody thought to list; - every NEGATIVE length, because the lookbehind treated the minus sign as part of the preceding token -- `left: -7px` and `margin-top: -2.5px` position the eyelet on a tag whose diamond HAD been converted, and `right: -3px` / `bottom: -3px` place the corner badge on a converted icon. Each is a piece of geometry that stopped tracking the piece it belongs to. With both fixed it found nine values, including one no comment mentioned (`flex: 1 1 200px`). All nine converted, except the three in a dead `.toggle` / `.slider` block: nothing has referenced those classes since Toggle became a Mantine Switch, and it was ALREADY broken -- track and knob in rem, travel still `translateX(26px)`, so the thumb stopped ~4px short of a 54px track. Dead CSS that is also wrong is a trap for whoever revives it, so it is deleted. The tag eyelet is centred with `translateY(-50%)` rather than a negative margin of half its own height, so the two pieces cannot drift again whatever the scale. Also from the review: - A runtime check, which is what the source guards were missing. Three cases in ui-layout.cy.js measure the real thing: the root font-size is the declared multiple of 16, our text AND a Mantine control both grow by the same factor when the scale changes, and a hairline does not. The second is the one that catches `theme.scale` being set and silently cancelling the root. - `max-height: max(..., 56.25cqw)` now follows a plain rem declaration. An engine without container query units throws out the whole declaration, which would have left the poster with NO cap -- a portrait cover standing as tall as the card is wide. Older engines clamp; newer ones take the second. - `pxToRem` in components/ui/scale.ts, replacing the `/ 16` written out in CardGroup and Placeholder. - The map overlays: 20 values across the layer panel, add-pin dialog, search panel and context menu, whose type had grown 10% while their padding had not. Left in px on purpose: the context menu's `left`/`top` (a real click coordinate), maplibre's own `ScaleControl` maxWidth, and the `calc(100% - 100px)` reserve for maplibre's px-sized controls. The `100vh - 120px` reserve DID become rem, since the chrome it reserves for is ours. - The keyboard-shortcuts spec no longer deletes `.mantine-Notifications-root` from the DOM. The reviewer was right that it is brittle; the endpoints are mocked instead, now that the probe named them: `/api/tag`, `/api/events/feed`, `/api/files/search` and `/api/files/worker_status`. Two of those were already "mocked" as `/api/tags` and `/api/events`, which nothing calls -- the wrong paths were the reason a toast appeared at all. Verified under CI's condition (`forceNetworkError` on everything unmocked): 12 passing with no DOM surgery. What is NOT done, deliberately: about 176 numeric geometry values remain in 23 feature files. The map overlays were the visible ones. Sweeping the rest is a change to 23 files nobody has reviewed, so instead the guard's name now says "in our stylesheets" rather than implying the whole app, and a new ratchet keeps numeric geometry out of the component library so the eleventh inline `gap: 10` cannot be added there. Both new guards planted against. Verified: 1142 jest / 65 suites, 251 ui-layout + 10 modal-stack component tests, all 50 e2e specs, clean build, map overlays measured to track the scale (panel minWidth 200 -> 220, search 320 -> 352) and screenshotted.
2026-08-03 19:37:25 -06:00
style={{padding: "0.5rem 0.75rem", cursor: "pointer", userSelect: "none"}}
onMouseEnter={e => e.target.style.background = "#f0f0f0"}
onMouseLeave={e => e.target.style.background = "transparent"}
onClick={handleSetDefaultLocation}
>
Set as Default View
</div>
</div>}
{addingPin && <AddPinDialog
lat={addingPin.lat}
lon={addingPin.lon}
onSubmit={handleSubmitPin}
onCancel={() => setAddingPin(null)}
/>}
</div>;
}