@ids-group-ltd/ids-design-system 0.7.0 → 0.8.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +609 -0
- package/fesm2022/ids-group-ltd-ids-design-system.mjs +2779 -1019
- package/fesm2022/ids-group-ltd-ids-design-system.mjs.map +1 -1
- package/migrations/0.5.1.json +46 -0
- package/migrations/0.6.3.json +11 -0
- package/migrations/0.7.0.json +11 -0
- package/migrations/0.7.1.json +145 -0
- package/migrations/0.8.0.json +179 -0
- package/migrations/AUTHORING.md +18 -0
- package/package.json +1 -1
- package/schematics/collection.json +1 -1
- package/styles/_dropdown-overlay.scss +11 -2
- package/styles/_field-box.scss +17 -0
- package/styles/_field-label.scss +34 -0
- package/styles/_focus.scss +58 -0
- package/styles/_overlay-motion.scss +80 -9
- package/styles/_surface-motion.scss +50 -0
- package/styles/_tokens.scss +55 -7
- package/styles/_typography.scss +2 -2
- package/styles/ds.scss +2 -0
- package/types/ids-group-ltd-ids-design-system.d.ts +943 -142
- /package/schematics/ng-add/{index.js → index.cjs} +0 -0
package/migrations/0.5.1.json
CHANGED
|
@@ -102,6 +102,52 @@
|
|
|
102
102
|
"--ds-menu-item-danger-label-color"
|
|
103
103
|
]
|
|
104
104
|
},
|
|
105
|
+
{
|
|
106
|
+
"id": "modal-card-danger-class-rename",
|
|
107
|
+
"kind": "class-rename",
|
|
108
|
+
"old": "ds-modal .danger",
|
|
109
|
+
"new": ".error",
|
|
110
|
+
"note": "The status-vocabulary rename also renamed the class the card puts on its root: [danger] used to emit `danger`, [error] now emits `error`. The input-rename item fixes the BINDING; a consumer stylesheet that keys off the class needs this one, and nothing in the binding rename would find it. Scope strictly to selectors under ds-modal / ds-modal-card.",
|
|
111
|
+
"searchHints": [
|
|
112
|
+
"ds-modal",
|
|
113
|
+
"ds-modal-card",
|
|
114
|
+
".modal.danger",
|
|
115
|
+
"&.danger"
|
|
116
|
+
]
|
|
117
|
+
},
|
|
118
|
+
{
|
|
119
|
+
"id": "menu-item-danger-class-rename",
|
|
120
|
+
"kind": "class-rename",
|
|
121
|
+
"old": "ds-menu-item .danger",
|
|
122
|
+
"new": ".error",
|
|
123
|
+
"note": "Same rename on ds-menu-item: the host used to carry `danger`, it now carries `error`. Consumer STYLING only.",
|
|
124
|
+
"searchHints": [
|
|
125
|
+
"ds-menu-item",
|
|
126
|
+
".danger"
|
|
127
|
+
]
|
|
128
|
+
},
|
|
129
|
+
{
|
|
130
|
+
"id": "stat-danger-class-rename",
|
|
131
|
+
"kind": "class-rename",
|
|
132
|
+
"old": "ds-stat .danger",
|
|
133
|
+
"new": ".error",
|
|
134
|
+
"note": "ds-stat's variant value moved danger -> error, and the host class follows it (host binding [class.error]). Consumer STYLING only; the value-remap item covers the binding.",
|
|
135
|
+
"searchHints": [
|
|
136
|
+
"ds-stat",
|
|
137
|
+
".danger"
|
|
138
|
+
]
|
|
139
|
+
},
|
|
140
|
+
{
|
|
141
|
+
"id": "progress-danger-class-rename",
|
|
142
|
+
"kind": "class-rename",
|
|
143
|
+
"old": "ds-progress .danger",
|
|
144
|
+
"new": ".error",
|
|
145
|
+
"note": "Same as ds-stat: variant value danger -> error with a matching host class. Consumer STYLING only.",
|
|
146
|
+
"searchHints": [
|
|
147
|
+
"ds-progress",
|
|
148
|
+
".danger"
|
|
149
|
+
]
|
|
150
|
+
},
|
|
105
151
|
{
|
|
106
152
|
"id": "progress-bar-warn-token-rename",
|
|
107
153
|
"kind": "token-rename",
|
package/migrations/0.6.3.json
CHANGED
|
@@ -22,6 +22,17 @@
|
|
|
22
22
|
"<ds-table",
|
|
23
23
|
"rowInteractive"
|
|
24
24
|
]
|
|
25
|
+
},
|
|
26
|
+
{
|
|
27
|
+
"id": "table-component-subclass-member-collisions",
|
|
28
|
+
"kind": "behavior",
|
|
29
|
+
"old": null,
|
|
30
|
+
"new": null,
|
|
31
|
+
"note": "Only affects a consumer that EXTENDS the DS TableComponent class. Subclassing a DS component is unsupported (design-system.md - Components), but where it happens this release breaks the build: the new column API added ~76 members to TableComponent, so a subclass that already used one of those names no longer compiles — TS4114 requires an 'override' modifier, TS2416 fires where the signature differs. Collisions observed: zebra, columnOrder, orderedColumns, visibleColumns, toggleColumn, onColumnDrop, moveColumn, resetColumns, columnFilters, exportCsv, onResizeMove, onResizeEnd. Resolution per member: where the base member means the same thing, keep the name and add `override` with a base-compatible signature (zebra must adopt the base's booleanAttribute transform); otherwise RENAME the subclass's own member — the private/protected ones are internal to the component and its own template, so renaming is behaviour-preserving, but check the new name against the base's member list so it does not collide again. Build the consumer right after the bump: the compiler enumerates every collision. A consumer that only uses the <ds-table> element needs no action.",
|
|
32
|
+
"searchHints": [
|
|
33
|
+
"extends TableComponent",
|
|
34
|
+
"extends "
|
|
35
|
+
]
|
|
25
36
|
}
|
|
26
37
|
],
|
|
27
38
|
"informational": [
|
package/migrations/0.7.0.json
CHANGED
|
@@ -246,6 +246,17 @@
|
|
|
246
246
|
"<ds-select",
|
|
247
247
|
"reserveMessage"
|
|
248
248
|
]
|
|
249
|
+
},
|
|
250
|
+
{
|
|
251
|
+
"id": "table-component-subclass-member-collisions",
|
|
252
|
+
"kind": "behavior",
|
|
253
|
+
"old": null,
|
|
254
|
+
"new": null,
|
|
255
|
+
"note": "Only affects a consumer that EXTENDS a DS component class. Subclassing is unsupported (design-system.md - Components), but where it happens this release can break the build the same way 0.6.3 did: TableComponent gained total, sortingOrder, animateHeight and scrollToTopOnPageChange, and most other component classes changed too, so a subclass that already used one of those names no longer compiles — TS4114 requires an 'override' modifier, TS2416 fires where the signature differs. Resolution per member: where the base member means the same thing, keep the name and add `override` with a base-compatible signature; otherwise RENAME the subclass's own member, checking the new name against the base's member list first. Build the consumer right after the bump: the compiler enumerates every collision. A consumer that only uses the DS elements needs no action.",
|
|
256
|
+
"searchHints": [
|
|
257
|
+
"extends TableComponent",
|
|
258
|
+
"extends "
|
|
259
|
+
]
|
|
249
260
|
}
|
|
250
261
|
],
|
|
251
262
|
"informational": [
|
|
@@ -0,0 +1,145 @@
|
|
|
1
|
+
{
|
|
2
|
+
"version": "0.7.1",
|
|
3
|
+
"items": [
|
|
4
|
+
{
|
|
5
|
+
"id": "surface-motion-partial-a-la-carte",
|
|
6
|
+
"kind": "scss-api",
|
|
7
|
+
"old": "@use 'styles/overlay-motion';",
|
|
8
|
+
"new": "@use 'styles/surface-motion';\n@use 'styles/overlay-motion';",
|
|
9
|
+
"note": "New GLOBAL partial `styles/_surface-motion.scss` holds the shared entrance keyframes (ds-fade-in, ds-scale-in, ds-slide-in, ds-slide-edge-in) that every DS surface now names through --ds-<surface>-enter-keyframes. Consumers of the bundled entry (`@use 'styles/ds';`) need NO change — ds.scss forwards it. Only à-la-carte consumers that @use DS partials individually must add it, next to `styles/overlay-motion`: without it the keyframes do not exist and modal / drawer / bottom-sheet / toast / banner / alert / loading-overlay / overlay panels render with no entrance (silently — no error). Order does not matter; @use anywhere in the styles entry is enough. If a consumer's styles.scss has no `styles/overlay-motion` line, add `@use 'styles/surface-motion';` beside whatever DS partials it does @use.",
|
|
10
|
+
"searchHints": [
|
|
11
|
+
"styles/overlay-motion",
|
|
12
|
+
"styles/ds",
|
|
13
|
+
"@use 'styles/",
|
|
14
|
+
"ids-design-system/styles"
|
|
15
|
+
]
|
|
16
|
+
},
|
|
17
|
+
{
|
|
18
|
+
"id": "table-sortkey-non-null-when-unsorted",
|
|
19
|
+
"kind": "behavior",
|
|
20
|
+
"old": "ds-table[sortKey]",
|
|
21
|
+
"new": null,
|
|
22
|
+
"note": "`ds-table` no longer nulls `sortKey` when the sort cycle reaches \"unsorted\" — the key keeps naming the column the user last touched and `sortDir === null` alone means unsorted. Fixes a header desync for hosts that map \"off\" back to their own default sort. ACTION: any host that treats a non-null `sortKey` as \"a sort is active\" must gate on `sortDir` instead — e.g. `if (sortKey())` becomes `if (sortDir())`, and a server query built as `orderBy: sortKey()` must skip the clause when `sortDir()` is null. Hosts that already read both, or that pass `sortingOrder: ['asc','desc']` (no unsorted state), need no change. The table's own rendering (icons, aria-sort) and its client-side sorting were already gated on `sortDir` and are unchanged.",
|
|
23
|
+
"searchHints": [
|
|
24
|
+
"sortKey",
|
|
25
|
+
"sortDir",
|
|
26
|
+
"sortChange",
|
|
27
|
+
"(sortChange)",
|
|
28
|
+
"[(sortKey)]"
|
|
29
|
+
]
|
|
30
|
+
},
|
|
31
|
+
{
|
|
32
|
+
"id": "field-reserve-message-default-off",
|
|
33
|
+
"kind": "behavior",
|
|
34
|
+
"old": "ds-input[reserveMessage]",
|
|
35
|
+
"new": null,
|
|
36
|
+
"note": "Every message-bearing field now defaults `reserveMessage` to FALSE (was true since 0.7.0). The resolution chain is unchanged - per-field input, then `DS_CONFIG.field.reserveMessage`, then the default - only the final fallback flipped. VISIBLE CONSEQUENCE: a field with no hint/error no longer holds a message line, so a form gets ~23px shorter per field and the controls below shift down the first time a hint or error appears. Affects all 14 fields: input, textarea, select, autocomplete, multi-select, checkbox, radio, radio-group, switch, slider, date-picker, time-picker, color-picker, otp-input. ACTION: a form that validates and must not reflow should turn it back on. App-wide in one line - `provideDsConfig({ field: { reserveMessage: true } })` - which is also how to restore 0.7.0 behaviour exactly; or per field with `[reserveMessage]=\"true\"`. Consumers that had written `[reserveMessage]=\"false\"` to opt OUT can delete those bindings, and any tighter sub-stack added to compensate for the reserved gap (e.g. a `var(--space-1)` wrapper) should be re-checked - with the row collapsed, a normal 16px stack is correct again.",
|
|
37
|
+
"searchHints": [
|
|
38
|
+
"reserveMessage",
|
|
39
|
+
"[reserveMessage]",
|
|
40
|
+
"provideDsConfig",
|
|
41
|
+
"field: {",
|
|
42
|
+
"DS_CONFIG"
|
|
43
|
+
]
|
|
44
|
+
},
|
|
45
|
+
{
|
|
46
|
+
"id": "ds-config-nested-provide-merges",
|
|
47
|
+
"kind": "behavior",
|
|
48
|
+
"old": "provideDsConfig",
|
|
49
|
+
"new": null,
|
|
50
|
+
"note": "A nested `provideDsConfig` (on a Route, component or directive) now MERGES with the config above it, section by section, instead of replacing it wholesale. Previously the provider used `useValue`, so a subtree that set one flag silently reset every unrelated section to its default. NO ACTION for the normal case - a single root-level `provideDsConfig` behaves exactly as before, and nested provision was neither documented nor safe to use. Only relevant if a consumer deliberately relied on a nested call to RESET a section back to defaults: that now needs the section spelled out explicitly, e.g. `provideDsConfig({ datePicker: { locale: 'en-US', dateFormat: undefined, firstDayOfWeek: 1 } })`. Enables the new `dsReserveMessages` directive to scope one field flag without disturbing the rest.",
|
|
51
|
+
"searchHints": [
|
|
52
|
+
"provideDsConfig",
|
|
53
|
+
"DS_CONFIG",
|
|
54
|
+
"providers: [provideDsConfig"
|
|
55
|
+
]
|
|
56
|
+
},
|
|
57
|
+
{
|
|
58
|
+
"id": "exported-classes-gained-members",
|
|
59
|
+
"kind": "behavior",
|
|
60
|
+
"old": "extends TableComponent",
|
|
61
|
+
"new": null,
|
|
62
|
+
"note": "Several exported component classes CHANGED their member surface this range, which breaks a consumer that EXTENDS one (subclassing a DS component is unsupported - design-system.md section Components - but where it happens the names collide: TS4114 demands an `override` modifier, TS2415 fires on a redeclared private member, TS2416 where a signature differs). New members: all 14 message-bearing fields (ds-autocomplete, ds-checkbox, ds-color-picker, ds-date-picker, ds-input, ds-multi-select, ds-otp-input, ds-radio, ds-radio-group, ds-select, ds-slider, ds-switch, ds-textarea, ds-time-picker) gained `reserveMessageResolved`, and CheckboxComponent, SwitchComponent and RadioComponent also gained a protected `devMode`; InputComponent and TextareaComponent also gained `control` and `focus`; SearchComponent gained `focus`; TableComponent gained `rowId` and `rowKey`; PaginationComponent gained `showPageSize`; TimePickerComponent gained `label`, `required`, `optional`, `labelId`, `fullWidth`; DatePickerComponent gained `fullWidth`; AppLayoutComponent gained `railOpen`, `railLabel`, `compact`, `closeRail` plus the private `navigation`, `destroyRef`, `watchDesktopBreakpoint`. ONE REMOVAL: TooltipDirective (dsTooltip) dropped its protected `onOpen`, split into `onPointerEnter` / `onFocusIn` / `onFocusOut` plus a private `scheduleShow`, `describedElement`, `focusedTarget` - a subclass that overrode `onOpen` fails to compile because its `override` names a member the base no longer declares; move a mouse-path body to `onPointerEnter` and a keyboard-path body to `onFocusIn`. ACTION: rename the subclass's own member, or add `override` where the base's semantics genuinely match. Build after upgrading - the compiler enumerates every collision. Consumers who only use the <ds-...> elements need no change.",
|
|
63
|
+
"searchHints": [
|
|
64
|
+
"extends TableComponent",
|
|
65
|
+
"extends InputComponent",
|
|
66
|
+
"extends TextareaComponent",
|
|
67
|
+
"extends PaginationComponent",
|
|
68
|
+
"extends TimePickerComponent",
|
|
69
|
+
"extends DatePickerComponent",
|
|
70
|
+
"extends SearchComponent",
|
|
71
|
+
"extends TooltipDirective",
|
|
72
|
+
"extends CheckboxComponent",
|
|
73
|
+
"extends SwitchComponent",
|
|
74
|
+
"extends RadioComponent",
|
|
75
|
+
"extends AppLayoutComponent",
|
|
76
|
+
"extends ",
|
|
77
|
+
"override readonly"
|
|
78
|
+
]
|
|
79
|
+
},
|
|
80
|
+
{
|
|
81
|
+
"id": "app-layout-scroll-reset-path-only",
|
|
82
|
+
"kind": "behavior",
|
|
83
|
+
"old": "LayoutScrollService",
|
|
84
|
+
"new": null,
|
|
85
|
+
"note": "`ds-app-layout` no longer resets the content scroll on EVERY router NavigationEnd - only when the PATH changes (the part before `?` or `#`). A sort, page or filter change that rewrites just the query string or fragment now KEEPS the reader's scroll position. ACTION only if a screen relied on the old blanket reset: inject the still-public `LayoutScrollService` and call `scrollToTop()` from the handler that changes the filter, which is per-interaction and more precise than the old behaviour. No action for anyone who found the old reset disruptive.",
|
|
86
|
+
"searchHints": [
|
|
87
|
+
"LayoutScrollService",
|
|
88
|
+
"scrollToTop",
|
|
89
|
+
"ds-app-layout",
|
|
90
|
+
"NavigationEnd"
|
|
91
|
+
]
|
|
92
|
+
},
|
|
93
|
+
{
|
|
94
|
+
"id": "modal-title-flex-moved-to-heading",
|
|
95
|
+
"kind": "class-rename",
|
|
96
|
+
"old": "ds-modal-card .title",
|
|
97
|
+
"new": ".heading",
|
|
98
|
+
"note": "STYLING ONLY, and only for a consumer who styles the modal's internal DOM. The `[eyebrow]` slot wrapped `<h2 class=\"title\">` together with the eyebrow in a new `<div class=\"heading\">`, and the `flex: 1` that used to sit on the `<h2>` moved to that wrapper. A consumer rule that targeted `.modal-card .title` for the flex slot must retarget `.heading`; rules that target `.title` for typography or colour are unaffected. Scope strictly to selectors under ds-modal / ds-modal-card.",
|
|
99
|
+
"searchHints": [
|
|
100
|
+
"ds-modal",
|
|
101
|
+
"ds-modal-card",
|
|
102
|
+
".modal-card .title",
|
|
103
|
+
".title",
|
|
104
|
+
"flex: 1"
|
|
105
|
+
]
|
|
106
|
+
},
|
|
107
|
+
{
|
|
108
|
+
"id": "table-client-sort-blanks-last-and-numeric-collation",
|
|
109
|
+
"kind": "behavior",
|
|
110
|
+
"old": "ds-table[clientSort]",
|
|
111
|
+
"new": null,
|
|
112
|
+
"note": "`clientSort` changes how existing columns order in two ways, both applying with or without the new per-column `sortValue`. (1) BLANKS SINK: a cell reading as no-data - empty string, or one of the em dash, en dash, hyphen, 'n/a', 'never', 'none' matched case-insensitively - now sorts LAST in both ascending and descending order. Previously an empty string was just the smallest value, so it floated to the top on ascending. (2) NATURAL TEXT COLLATION: the text fallback now passes { numeric: true } to localeCompare, so 'item 9' precedes 'item 10' instead of following it. ACTION: none for most tables - both changes are what a reader expects. Only if a screen deliberately relied on blanks sorting first, or on the old pure-lexical order of strings containing digits, supply a per-column `sortValue` that maps the values the way that screen wants.",
|
|
113
|
+
"searchHints": [
|
|
114
|
+
"clientSort",
|
|
115
|
+
"sortChange",
|
|
116
|
+
"sortKey",
|
|
117
|
+
"sortDir",
|
|
118
|
+
"[columns]",
|
|
119
|
+
"TableColumn"
|
|
120
|
+
]
|
|
121
|
+
}
|
|
122
|
+
],
|
|
123
|
+
"informational": [
|
|
124
|
+
"Purely additive, opt-in inputs — no consumer action: `ds-table` [rowId] accessor (row identity for tracking/selection/expansion; still defaults to row.id), `ds-date-picker`/`ds-time-picker` [fullWidth], `ds-pagination` [showPageSize], and public `focus(options?)` on `ds-search` / `ds-input` / `ds-textarea`.",
|
|
125
|
+
"`ds-table`'s row generic relaxes from `{ id: string }` to `object` so keyless data can supply [rowId]; rows that keep an `id` type-check exactly as before, and the resolver is public as `rowKey(row)`.",
|
|
126
|
+
"New override tokens, all defaulting to the current values: --ds-<surface>-enter-keyframes / -enter-duration / -enter-easing (+ -enter-offset / -enter-scale) and -leave-duration / -leave-easing across overlay panels, modal, drawer, bottom sheet, toast, banner, alert and loading overlay; --ds-modal-eyebrow-* (5); --ds-modal-title-font-size; --ds-date-picker-trigger-radius; --ds-date-picker-panel-radius; --ds-checkbox-radius; --ds-popover-min-width; --ds-menu-min-width / -max-width; --ds-toast-min-width / -max-width; --ds-table-cell-pad-{top,right,bottom,left}; --ds-table-head-border-color / -weight / -tracking; --ds-pagination-label-font-size / -weight / -color.",
|
|
127
|
+
"`ds-modal` / `ds-modal-card` gained an `[eyebrow]` projection slot above the title; unfilled it renders nothing, so existing dialogs are unaffected.",
|
|
128
|
+
"`DS_CONFIG` gained a `field.reserveMessage` key — `provideDsConfig({ field: { reserveMessage: false } })` sets the message-row default for all 14 message-bearing fields. The per-field input still wins; its type widens to `boolean | null` (null = not set), so existing bindings and bare attributes behave identically.",
|
|
129
|
+
"Visual-only: overlay panels now enter FROM their trigger edge (a panel above its trigger rises instead of sliding down), a selected dropdown option keeps its tint under hover/keyboard-active, and a read-only `ds-modal` no longer draws an empty footer strip.",
|
|
130
|
+
"ds-time-picker gained label / required / optional inputs (additive) — a date + time form row now aligns without a hand-rolled label; existing usages without label render exactly as before.",
|
|
131
|
+
"Visual-only in ds-time-picker: the clock glyph moved from before the value to the trailing edge and grew 14px -> 16px, the value weight dropped from medium to regular, and the trigger gap narrowed 8px -> 4px, all to match ds-date-picker / ds-select / ds-input. Recover the old weight with --ds-time-picker-value-weight: var(--font-weight-medium).",
|
|
132
|
+
"Value typography of both pickers is now tokenised: --ds-date-picker-value-font-family / -font-size / -weight (mono default, shared with the ghost mask) and --ds-time-picker-value-font-family / -font-size / -weight, plus --ds-time-picker-required-color. Defaults unchanged except the time value weight noted above.",
|
|
133
|
+
"ds-date-picker [fullWidth] now pushes the calendar glyph to the trailing edge instead of leaving it packed beside the value; only affects fields that opt into fullWidth.",
|
|
134
|
+
"`ds-app-layout` gained an opt-in responsive rail: below the `lg` breakpoint the rail becomes an off-canvas overlay driven by the new two-way `[(railOpen)]` model, with `railLabel` (default 'Navigation') naming it for a screen reader. At/above `lg` nothing changes and `railOpen` has no effect, so an existing desktop-only consumer is unaffected. New tokens --ds-app-layout-rail-bg-color / -rail-shadow / --ds-app-layout-scrim-color all default to current values. PREREQUISITE: the breakpoint is read at runtime from the emitted `--bp-lg` custom property, with no literal px fallback - a consumer whose global styles never emit the `--bp-*` scale keeps the desktop layout at every width (dev builds log a console.error saying so). Include the DS global styles, or the bundled `styles/ds` entry, to get the compact behaviour.",
|
|
135
|
+
"`inheritDsConfig(config)` is a new public export beside `provideDsConfig`: it reads the nearest ancestor DS_CONFIG (skipSelf, optional) and shallow-merges per top-level section. Call it from an injection context if you need the merging behaviour in your own factory provider or directive metadata.",
|
|
136
|
+
"`dsReserveMessages` is a new directive: put it on a form and every DS field DECLARED inside reserves its message row, without a per-control binding. Presence-driven (no input), merges with the config above it, and a field's own [reserveMessage] still wins. Projected <ng-content> is not covered - it keeps the injector of the template it was written in.",
|
|
137
|
+
"Visual-only: dialogs opened through `ModalService` no longer carry the `ds-overlay-pane` panel class, so the shared overlay-panel entrance is no longer layered on top of the modal's own card/backdrop animation.",
|
|
138
|
+
"Known gap, not a regression: the ds-time-picker label wiring (aria-labelledby / aria-required) landed on the SINGLE-mode trigger only; range mode's start/end segment triggers are not wired to the label yet.",
|
|
139
|
+
"`TableColumn` gained an optional `sortValue: (row) => string | number | Date | null | undefined` so a column sorts by its domain value instead of the rendered text; formatted money, percents and dates sorted wrong before. Purely additive - a column without it keeps today's behaviour. Only affects sorting, not the header filter's value list.",
|
|
140
|
+
"`ds-checkbox`, `ds-switch` and `ds-radio` log a dev-mode warning when a caption is projected (`<ds-checkbox>text</ds-checkbox>`) instead of passed via [label]; the text never rendered and failed silently. All three share the [label]-only contract. Dev only - the detection node is gated on isDevMode(), so production is unchanged. Projection itself is still unsupported. The directive doing the detection is internal (not exported from public-api).",
|
|
141
|
+
"Known gap, not a regression: `--ds-overlay-enter-scale` is the one entrance token NOT forwarded onto a portaled overlay pane, so setting it per instance (`ds-select { --ds-overlay-enter-scale: 0.9 }`) has no effect on select / autocomplete / multi-select / date / time / colour picker / popover / tooltip / menu submenu panels. Scope it globally until it is forwarded; the other four (-enter-keyframes / -duration / -easing / -offset) do forward.",
|
|
142
|
+
"Bugfix, no action: `dsTooltip` now puts `aria-describedby` on the element that actually took focus (recorded from the bubbling `focusin`) rather than unconditionally on its host, so a tooltip on a host that is not itself focusable — `ds-menu-item` carries `role=\"none\"` and holds the real button inside — finally reaches a screen reader. A pointer-only open still describes the host.",
|
|
143
|
+
"Visual-only fixes: `ds-pagination` pins its internal rows-per-page select to `reserveMessage: false` (it inherited the app default and pushed the pager off the page buttons' baseline), and `ds-stat`'s card fills a stretched host so a KPI row with mixed captions paints equal-height tiles. Consumers can drop the global workarounds for both (`ds-pagination ds-field-message { display: none }` and `ds-stat { height: 100% }`)."
|
|
144
|
+
]
|
|
145
|
+
}
|
|
@@ -0,0 +1,179 @@
|
|
|
1
|
+
{
|
|
2
|
+
"version": "0.8.0",
|
|
3
|
+
"items": [
|
|
4
|
+
{
|
|
5
|
+
"id": "table-client-sort-replaced-by-sort-mode",
|
|
6
|
+
"kind": "removed-api",
|
|
7
|
+
"old": "ds-table[clientSort]",
|
|
8
|
+
"new": "sortMode",
|
|
9
|
+
"note": "`ds-table`'s boolean `clientSort` is REMOVED and replaced by `sortMode: 'client' | 'server' | 'none'`. The boolean could not express \"nothing sorts\", so a column marked `sortable: true` on a table with no sorting drew a header that looked live, showed a caret on hover and did nothing — mm-broker shipped several detail tables that way and only caught it in a browser walkthrough. ACTION, mechanical: `[clientSort]=\"true\"` becomes `sortMode=\"client\"`; a bare `clientSort` attribute becomes `sortMode=\"client\"`; `[clientSort]=\"false\"` or `[clientSort]=\"expr\"` becomes `[sortMode]=\"expr ? 'client' : 'server'\"`. A table that had NO clientSort binding and reorders in its own `(sortChange)` handler must now say `sortMode=\"server\"` — this is the one case that is not a find/replace, because the DEFAULT CHANGED: unset now resolves to `'client'`, so a table that previously only emitted `sortChange` will ALSO sort its own rows unless it declares `'server'`. That default flip is deliberate — it makes `sortable: true` always do something — but for a SERVER-PAGED table (one that binds `total`) client sorting reorders the current page only, so the rows look sorted and are not; dev builds warn about exactly that combination. Tables that render a sortable column but sort nowhere should declare `sortMode=\"none\"`, which now drops the button, the caret and `aria-sort` instead of leaving a dead control. NOTE for TypeScript callers: the union is exported as `TableSortMode`, so a field, computed or method that supplies the mode to `[sortMode]` can be typed with `import type { TableSortMode } from '@ids-group-ltd/ids-design-system';` rather than repeating the literals.",
|
|
10
|
+
"searchHints": [
|
|
11
|
+
"clientSort",
|
|
12
|
+
"[clientSort]",
|
|
13
|
+
"sortChange",
|
|
14
|
+
"(sortChange)",
|
|
15
|
+
"sortable: true",
|
|
16
|
+
"TableColumn",
|
|
17
|
+
"ds-table"
|
|
18
|
+
]
|
|
19
|
+
},
|
|
20
|
+
{
|
|
21
|
+
"id": "exported-class-members-grew",
|
|
22
|
+
"kind": "behavior",
|
|
23
|
+
"old": "extends InputComponent",
|
|
24
|
+
"new": null,
|
|
25
|
+
"note": "This release's single class-member-collision entry: it covers EVERY exported class whose member surface changed, so a consumer who EXTENDS a DS component can resolve them in one pass. Subclassing a DS component is unsupported (design-system.md § Components), but where it happens the names collide: TS4114 demands `override`, TS2415 fires on a redeclared private member, TS2416 on a differing signature. (1) The 14 message-bearing FIELD classes (Input, Textarea, Select, DatePicker, TimePicker, Autocomplete, Checkbox, Switch, RadioGroup, Radio, OtpInput, Slider, ColorPicker, MultiSelect) REMOVED the protected `messageState` computed, and DatePickerComponent also removed `messageText`; all 14 ADDED the `error`, `warning`, `errors` and `touched` inputs plus protected `message` and `invalidResolved`, with DatePickerComponent also gaining private `localError` and `dateInvalid`. (2) Both CVA base classes (`DsControlValueAccessor`, `DsBooleanControlValueAccessor`) gained a protected `touchedSelf`. (3) `AppLayoutComponent` gained public `railOverlay` and `railId`. (4) `TableComponent` REMOVED the `clientSort` input and gained `sortModeResolved`, `sortableCol()`, `exporting`, `runExport()`, `virtualizing`, `virtualRowHeight`, `virtualWindow`, `windowData` and `isStriped()`, plus the `persistLayout`, `storageKey`, `exportButton`, `exportFilename`, `exportFetcher`, `virtualScroll` and `virtualRowSize` inputs. (5) `ButtonComponent` gained `ariaControls`; `IconButtonComponent` gained `ariaExpanded` and `ariaControls`. (6) `CheckboxComponent`, `SwitchComponent` and `RadioComponent` REMOVED the protected `devMode`. (7) `LayoutScrollService` gained `scroller`, `scrollTo()`, `configureRestore()`, `consumePendingOffset()` and `contentReady()`. (8) The `close()` method on SelectComponent, AutocompleteComponent, MultiSelectComponent, DatePickerComponent, TimePickerComponent, ColorPickerComponent and PopoverComponent gained an optional `cause` parameter, as did TimePickerComponent's `onEscape()` — an override declared as `close()` still satisfies the base, so this one only bites a subclass that re-declares the parameter list differently. ACTION: rename the subclass's own member, or add `override` where the base's semantics genuinely match; a subclass that overrode `messageState` should move its logic to an `error` / `warning` binding on the element instead, since the message is now resolved centrally. Consumers who only use the <ds-...> elements in templates need no change.",
|
|
26
|
+
"searchHints": [
|
|
27
|
+
"extends InputComponent",
|
|
28
|
+
"extends SelectComponent",
|
|
29
|
+
"extends TextareaComponent",
|
|
30
|
+
"extends DatePickerComponent",
|
|
31
|
+
"extends CheckboxComponent",
|
|
32
|
+
"extends AppLayoutComponent",
|
|
33
|
+
"extends TableComponent",
|
|
34
|
+
"extends ButtonComponent",
|
|
35
|
+
"extends IconButtonComponent",
|
|
36
|
+
"extends DsControlValueAccessor",
|
|
37
|
+
"extends ",
|
|
38
|
+
"messageState",
|
|
39
|
+
"override readonly"
|
|
40
|
+
]
|
|
41
|
+
},
|
|
42
|
+
{
|
|
43
|
+
"id": "select-label-class-renamed-to-label",
|
|
44
|
+
"kind": "class-rename",
|
|
45
|
+
"old": "ds-select .select-label",
|
|
46
|
+
"new": ".label",
|
|
47
|
+
"note": "STYLING ONLY, and only for a consumer who styles ds-select's internal DOM. The label element is now `.label`, the same class every other field in the family uses, so a rule written as `.select-label` (or the doubled `.select-label.select-label` needed to out-rank the shimmed component rule) silently matches nothing after the upgrade. Retarget to `.label`, scoped to the element: `ds-select .label { … }`. In most cases the rule can be DELETED instead — it existed to make the select's label match the input's, and they now match by default. Scope strictly to selectors under ds-select; `.label` is also the class ds-input, ds-textarea, ds-date-picker, ds-time-picker and ds-radio-group use.",
|
|
48
|
+
"searchHints": [
|
|
49
|
+
"select-label",
|
|
50
|
+
".select-label",
|
|
51
|
+
"ds-select",
|
|
52
|
+
"label",
|
|
53
|
+
"field-label"
|
|
54
|
+
]
|
|
55
|
+
},
|
|
56
|
+
{
|
|
57
|
+
"id": "radio-group-label-class-and-type",
|
|
58
|
+
"kind": "class-rename",
|
|
59
|
+
"old": "ds-radio-group .group-label",
|
|
60
|
+
"new": ".label",
|
|
61
|
+
"note": "TWO changes to the same element. (1) CLASS: `ds-radio-group`'s label is now `.label`, so consumer CSS targeting `.group-label` under ds-radio-group stops matching — retarget or delete. Note `.group-label` is ALSO the class `ds-sidenav-group` uses for its heading, which is unchanged; only rules scoped under ds-radio-group are affected. (2) VISIBLE: the label now draws the family's type - `--font-weight-bold` at `--font-size-xs` - instead of its own `--font-weight-medium` at `--font-size-s` on `--line-height-flat`. It was the only field label drawing its own, and in a form beside an input or select it read as a different kind of element. ACTION: none for most screens, since matching the family is the point. To keep the old look, scope the family tokens: `ds-radio-group { --ds-field-label-weight: var(--font-weight-medium); --ds-field-label-font-size: var(--font-size-s); }`.",
|
|
62
|
+
"searchHints": [
|
|
63
|
+
"group-label",
|
|
64
|
+
".group-label",
|
|
65
|
+
"ds-radio-group",
|
|
66
|
+
"radio-group",
|
|
67
|
+
"label"
|
|
68
|
+
]
|
|
69
|
+
},
|
|
70
|
+
{
|
|
71
|
+
"id": "ng-add-schematic-now-loads",
|
|
72
|
+
"kind": "behavior",
|
|
73
|
+
"old": "ng add @ids-group-ltd/ids-design-system",
|
|
74
|
+
"new": null,
|
|
75
|
+
"note": "`ng add` never actually ran its schematic in any released version, so the DS global styles were NEVER wired automatically - every consumer either did it by hand or shipped without them. The package is `\"type\": \"module\"` while the compiled schematic was CommonJS, so Node refused to load it: \"exports is not defined in ES module scope\". The build wrote a nested schematics/package.json with {\"type\":\"commonjs\"} to fix exactly that, and `npm pack` silently DROPPED it - npm treats a directory holding a package.json as a separate package root - so the marker existed in dist/ and was absent from the tarball. The schematic is now `.cts` compiled to `.cjs`, which survives packing. ACTION: consumers who added `@use '@ids-group-ltd/ids-design-system/styles/ds';` to their global stylesheet by hand need do nothing - ng add prepends the same line and skips when it is already present. Consumers who never wired the styles (and whose components therefore render unthemed) can now run `ng add` and have it work, or add that one line themselves.",
|
|
76
|
+
"searchHints": [
|
|
77
|
+
"ng add",
|
|
78
|
+
"ng-add",
|
|
79
|
+
"styles/ds",
|
|
80
|
+
"@use '@ids-group-ltd",
|
|
81
|
+
"schematics"
|
|
82
|
+
]
|
|
83
|
+
},
|
|
84
|
+
{
|
|
85
|
+
"id": "menu-item-radius-default-now-square",
|
|
86
|
+
"kind": "behavior",
|
|
87
|
+
"old": "ds-menu",
|
|
88
|
+
"new": null,
|
|
89
|
+
"note": "DEFAULT CHANGED, visual: a `ds-menu` row's hover/active fill was `var(--radius-md)` (8px) and is now SQUARE — the new `--ds-menu-item-radius` token defaults to `0`. In the same change `ds-menu` dropped its own `padding: var(--space-1)` to `padding: 0` and gained `overflow: hidden`, so rows now run edge to edge and the panel's own corner clips the first and last row. The reason: a curve at zero inset can never be concentric with the panel's, and inside a `ds-popover` (which drops its padding for a menu) the row's 8px sat against the popover's 12px corner and left visible slivers at the top and bottom of the list. ACTION: none for consumers happy with the new look. To restore the old rounded highlight set `ds-menu { --ds-menu-item-radius: var(--radius-md); }`; to also restore the old 4px inset around the rows add your own `ds-menu { padding: var(--space-1); }`, because the panel padding has NO token. Setting a radius without restoring some padding is the one combination to avoid — the corners have nothing to curve away from.",
|
|
90
|
+
"searchHints": [
|
|
91
|
+
"ds-menu",
|
|
92
|
+
"ds-menu-item",
|
|
93
|
+
"--ds-menu-item-radius",
|
|
94
|
+
"menu-item",
|
|
95
|
+
"ds-popover"
|
|
96
|
+
]
|
|
97
|
+
},
|
|
98
|
+
{
|
|
99
|
+
"id": "menu-divider-spacing-default-now-zero",
|
|
100
|
+
"kind": "behavior",
|
|
101
|
+
"old": "ds-menu-divider",
|
|
102
|
+
"new": null,
|
|
103
|
+
"note": "DEFAULT CHANGED, visual: `ds-menu-divider` carried `padding: var(--space-1) 0` (4px above and below) on top of the `gap: var(--space-0-5)` (2px) that `ds-menu` already puts between rows, so a separated group read as a hole — 6px + line + 6px where adjacent items sat 2px apart. The divider's padding now defaults to `0` through the new `--ds-menu-divider-spacing` token, so the line takes one row's place instead of adding a second gap. Menu items keep their own 8px/12px padding, so the line still separates without help. ACTION: none unless you want the old airier boundary back, in which case set `ds-menu { --ds-menu-divider-spacing: var(--space-1); }`. Affects only menus that actually use `ds-menu-divider`.",
|
|
104
|
+
"searchHints": [
|
|
105
|
+
"ds-menu-divider",
|
|
106
|
+
"--ds-menu-divider-spacing",
|
|
107
|
+
"menu-divider",
|
|
108
|
+
"ds-menu"
|
|
109
|
+
]
|
|
110
|
+
},
|
|
111
|
+
{
|
|
112
|
+
"id": "time-picker-panel-radius-now-panel-role",
|
|
113
|
+
"kind": "behavior",
|
|
114
|
+
"old": "ds-time-picker",
|
|
115
|
+
"new": null,
|
|
116
|
+
"note": "DEFAULT CHANGED, visual: the `ds-time-picker` dropdown panel's corner radius was `var(--radius-md)` (8px) and is now `var(--ds-radius-panel)` (`--radius-lg`, 12px). It was the one floating panel in the library still drawing the CONTROL radius — its own sibling `ds-date-picker` has always drawn the panel role, so the two pickers rendered different corners side by side in the same form. The radius was also the only property on that element with no public hook (fill, border and shadow all had one). ACTION: none unless you want the old 8px corner back, in which case set `ds-time-picker { --ds-time-picker-panel-radius: var(--radius-md); }` — the new hook is forwarded onto the portaled overlay pane, so a per-instance override on the element does reach the panel. Affects only the open dropdown; the trigger's radius is unchanged and still resolves through `--ds-time-picker-trigger-radius`.",
|
|
117
|
+
"searchHints": [
|
|
118
|
+
"ds-time-picker",
|
|
119
|
+
"--ds-time-picker-panel-radius",
|
|
120
|
+
"--ds-time-picker-trigger-radius",
|
|
121
|
+
"time-panel",
|
|
122
|
+
"time-picker"
|
|
123
|
+
]
|
|
124
|
+
},
|
|
125
|
+
{
|
|
126
|
+
"id": "focus-partial-needs-use-line",
|
|
127
|
+
"kind": "code-add",
|
|
128
|
+
"old": "@use '@ids-group-ltd/ids-design-system/styles",
|
|
129
|
+
"new": "@use '@ids-group-ltd/ids-design-system/styles/focus';",
|
|
130
|
+
"note": "ONLY for a consumer who cherry-picks DS style partials individually instead of importing the bundled `styles/ds` entry. The `data-focus` focus-style switch AND the `@media (forced-colors: active)` fallback both live in the new `styles/_focus.scss` partial. A consumer of the bundled entry gets both automatically and must do NOTHING — the partial is `@forward`ed from `styles/ds`. A consumer who @uses partials à la carte gets NEITHER until they add this line to the same global stylesheet: without it, the `data-focus` attribute silently does nothing, and — the part that matters for accessibility — focused elements have NO visible indicator at all in Windows High Contrast / forced-colors mode, because forced-colors drops box-shadow and the shadow-only indicator was all they had. Site: the global stylesheet that already @uses the other DS partials. Do NOT add it alongside a `styles/ds` import; that entry already includes it.",
|
|
131
|
+
"searchHints": [
|
|
132
|
+
"@use '@ids-group-ltd/ids-design-system/styles",
|
|
133
|
+
"styles/ds",
|
|
134
|
+
"styles/focus",
|
|
135
|
+
"ids-design-system/styles",
|
|
136
|
+
"data-focus",
|
|
137
|
+
"forced-colors"
|
|
138
|
+
]
|
|
139
|
+
},
|
|
140
|
+
{
|
|
141
|
+
"id": "combobox-tokens-renamed-to-picker",
|
|
142
|
+
"kind": "removed-api",
|
|
143
|
+
"old": "--combobox-offset / --combobox-maxh",
|
|
144
|
+
"new": "--picker-offset / --picker-maxh",
|
|
145
|
+
"note": "The two global tokens sizing the listbox panel shared by ds-autocomplete and ds-multi-select kept the name of a component that no longer exists - ds-combobox was renamed to ds-autocomplete on 2026-07-07. ACTION, only if you overrode them: rename --combobox-offset to --picker-offset and --combobox-maxh to --picker-maxh. These are global scale tokens, not --ds-* component tokens, so nothing else changes and the values are unchanged.",
|
|
146
|
+
"searchHints": [
|
|
147
|
+
"--combobox-offset",
|
|
148
|
+
"--combobox-maxh",
|
|
149
|
+
"combobox"
|
|
150
|
+
]
|
|
151
|
+
}
|
|
152
|
+
],
|
|
153
|
+
"informational": [
|
|
154
|
+
"Every field gained `error` and `warning` beside `hint`, resolved on a fixed ladder: `error` (an explicit string, shown at once) → validation errors → `warning` → `hint`. Only ONE line renders, so a field's height does not depend on how many sources are populated. Purely additive: a field that binds only `hint` behaves exactly as before, and when something is invalid but no source produced text the `hint` still renders in the error state — so the pre-existing `[invalid]` + `[hint]=\"errorText\"` idiom keeps working untouched.",
|
|
155
|
+
"Fields now read validation state from the bound form control, so `error` is optional. Signal forms binds `errors` and `touched` automatically (those input names are Angular 21's `FormUiControl` contract) and already drove the value through the `ControlValueAccessor` DS fields have always implemented. Reactive and template-driven forms need NO new binding: the field resolves `NgControl` after first render and subscribes to `control.events`. Errors wait for the first blur, or for the parent form to be submitted, so a `required` error never greets an untouched form. `invalid` is now OR-ed with what the control reports; the explicit input still wins.",
|
|
156
|
+
"`DS_CONFIG` gained `field.errorMessages` — one shared map from validation-error kind to text, for errors whose validator carried no message (Angular's built-ins never do). A message on the error itself always wins, so `required(path, { message })` in signal forms or `{ myError: 'text' }` from a plain validator overrides the map per case. An error that resolves to no text anywhere warns in dev instead of rendering an error style with an empty line.",
|
|
157
|
+
"`dsRailTrigger` is a new directive: put it on your own burger anywhere inside `ds-app-layout` and it toggles `[(railOpen)]` and keeps `aria-expanded` / `aria-controls` pointing at the rail panel, so the trigger and the panel are related for a screen reader. Both attributes are dropped while the rail is not an overlay, where `aria-expanded=\"false\"` would contradict what the user sees. `ds-app-layout` also exposes `railId` and the read-only `railOverlay` signal for a consumer that builds its own trigger.",
|
|
158
|
+
"`ds-table` `sortMode=\"none\"` is the way to render a column that is sortable in principle but not on this screen: the header renders as plain text, with no button, caret or `aria-sort`, instead of a control that does nothing.",
|
|
159
|
+
"Bugfix, no action: a CLOSED compact `ds-app-layout` rail is now `inert`, and closing it (Escape, scrim click, navigation) returns focus to the control that claims the rail via `aria-controls`. Before, the off-canvas panel kept `role=\"dialog\"`/`aria-modal=\"true\"`, left every nav link tabbable and stranded the caret inside the hidden panel, because cdkTrapFocusAutoCapture restores focus only on directive destroy. A consumer whose custom burger sets neither `dsRailTrigger` nor `aria-controls` gets the inert fix but no focus return - add `aria-controls` pointing at `ds-app-layout`'s `railId` to get it.",
|
|
160
|
+
"Bugfix, no action: NESTING one `ds-app-layout` inside another now works. Three separate faults are fixed. (1) The rail's open state was written as `:host(.rail-open) &`, which Angular's emulated encapsulation compiles to a DESCENDANT selector, and every instance of a component shares one encapsulation attribute — so an inner layout's rail opened whenever an outer one was open, and the inner scrim swallowed clicks meant for the outer. The state now lives on the elements it styles (`.rail.open` / `.scrim.open`). (2) Below `lg` the rail was `position: fixed`, resolving against the viewport, so an embedded layout threw its rail over the top-left of the WINDOW; rail and scrim are now `absolute` against a `position: relative` host. (3) The host now also sets `isolation: isolate`, so each layout z-indexes its rail (`--layer-drawer`) and scrim in its own stacking context instead of the one root context, where an inner layout's rail could paint over an outer layout's open rail. All three are identical to the old behaviour for a single full-viewport shell, which is the normal case.",
|
|
161
|
+
"Bugfix, no action: `--ds-overlay-enter-scale` is now forwarded onto the portaled overlay pane, so scoping it per instance (`ds-select { --ds-overlay-enter-scale: 0.9; }`) finally reaches select / multi-select / autocomplete / date / time / colour picker / popover / tooltip / menu-submenu panels. 0.7.1 shipped it as a known gap — the token was resolved on the pane but absent from the forwarded set, while the other four enter tokens did forward. A GLOBAL override worked before and is unaffected.",
|
|
162
|
+
"The field family gained one label contract: --ds-field-label-font-size / -weight / -color / -transform / -tracking, resolved by ds-input, ds-textarea, ds-select, ds-date-picker, ds-time-picker and ds-radio-group through a shared mixin. Defaults equal the values those components hard-coded, so nothing moves unless you set a token. There is deliberately no per-component token set - custom-property scoping already narrows a retune to one element (`ds-select { --ds-field-label-transform: uppercase; }`) - and -transform/-tracking make the uppercase tracked label reachable through tokens for the first time. ds-input's --ds-input-label-color/-fs/-fw still take precedence where set and now fall through to the family token rather than the raw scale.",
|
|
163
|
+
"`--ds-field-height` is new, and the --field-h-sm/-md/-lg scale is now actually read by ds-input, ds-select, ds-date-picker and ds-time-picker - previously only -lg was, the other two sizes reaching past the scale to the raw --hit-min/--hit-cozy. Values are identical, so no height moves on upgrade. NO ACTION, but two things become possible that silently did nothing before: retuning --field-h-md (or -sm) now moves every field, and --ds-field-height pins one height regardless of the size class, scopable per component. NOT a fix for unequal picker heights - those were already equal as of 0.7.1's aligned trigger geometry.",
|
|
164
|
+
"ds-button gained --ds-button-font-size / -weight / -tracking / -height / -pad-x / -pad-y, and reflects data-variant / data-color on its host element for external targeting. Defaults equal the hard-coded values, so nothing moves. NO ACTION - but a consumer who wrote .button.button.button { font-size; font-weight; letter-spacing; padding; height } to out-rank the shimmed .button.md can replace it with the tokens, set on the host. Per-variant --ds-button-* values were ALREADY reachable from the host before this change (the colour/fill resolvers read the hook first and never declare it), so an indirection layer added to work around that can go too.",
|
|
165
|
+
"ds-app-layout gained opt-in restoreScroll + restoreScrollTimeout, and LayoutScrollService gained contentReady() and consumePendingOffset(). Off by default, so NO ACTION. Turn it on for a list-and-details app and going back returns the reader to their place; forward links still land at the top. A 'back to list' control that adds a history entry instead of traversing must pass state: { dsRestoreScroll: true } to be restored. Note that an app which scrolls on navigation itself (an anchor jump, a URL that names a section) writes after the layout and wins, so restoreScroll is a no-op on those routes.",
|
|
166
|
+
"New role layer between the foundation scales and components: --ds-radius-{inner,control,panel,pill}, --ds-elevation-{raised,overlay,floating,modal}, and --ds-motion-{enter-panel,enter-surface,enter-large,exit,exit-large}-{duration,easing}. Every role defaults to the value its members already used, so NO ACTION and nothing moves visually. What becomes possible: retune one family without hitting the others - re-round panels while buttons keep their radius, lift every overlay in one line, make panel entrances snappier once. The old --shadow-popover/-modal/-toast/-card aliases still work and now resolve THROUGH the elevation roles. A per-component --ds-<comp>-radius / -shadow / -enter-duration override still wins over its role. Note: two motion resolvers (ds-loading-overlay, the toast leave) deliberately stay on the raw scale because they use --ease-standard, which no role carries.",
|
|
167
|
+
"New global focus-style switch: set data-focus=\"ring\" | \"field\" | \"edge\" | \"outline\" on <html>. Omitting it keeps the current split (halo on fields, double ring on controls), so NO ACTION for a consumer of the bundled `styles/ds` entry — it picks up both the switch and the forced-colors fallback automatically. (A consumer who cherry-picks partials à la carte DOES have one line to add; that is the actionable `focus-partial-needs-use-line` item.) Also new, and not part of the switch: under @media (forced-colors: active) every focused element gets a CanvasText outline, because forced-colors drops box-shadow and a shadow-only focus indicator disappeared there entirely.",
|
|
168
|
+
"ds-table gained [persistLayout] + storageKey (round-tripping columnOrder, columnWidths and hiddenKeys through localStorage) and [exportButton] + exportFilename + exportFetcher. All opt-in and off by default, so NO ACTION. A consumer who wrote their own localStorage effect for the three column models, or their own export button, can delete both - together with the sortValue accessor these were the last three reasons to keep a wrapper around ds-table. exportCsv() is unchanged for existing callers and now accepts an optional row set as its second argument.",
|
|
169
|
+
"New component ds-password-rules: the live requirement checklist under a new-password field. Purely additive, NO ACTION. A consumer with a hand-rolled password checklist can replace it - pass the rule set as { label, test } pairs (the policy stays yours, it has to mirror your server), bind [value] to the field and [showUnmet] to whether the user has tried to submit. Exposes metCount and allMet for gating a submit button, and the standalone passwordRulesMet(value, rules) function — the same predicate that backs allMet — for gating one without rendering the list. Also exported: the DsPasswordRule ({ label, test }) and DsPasswordRuleState ('pending' | 'met' | 'failed') types for typing the rule set, plus a listId input that names the <ul> so a field's aria-describedby can point at the checklist.",
|
|
170
|
+
"ds-input gained ariaDescribedBy, merged with the field's own message id rather than replacing it - for describing a field with a password-rule list or a policy note. Additive, NO ACTION. Note that setting aria-describedby on the <ds-input> element itself does NOT work: it lands on the wrapper, not the inner input.",
|
|
171
|
+
"New component ds-auth-shell: the scaffold for the screens in front of the login (sign-in, forgot password, reset, MFA), counterpart to ds-app-layout. Purely additive, NO ACTION. A consumer with a hand-rolled auth scaffold can replace it: brand lockup + optional tagline + one card holding the form + a support line. The brand row is the ds-sidenav row - same brand / logoLetter / [dsBrandName] contract and the same letter-mark tile - so it is NOT the string-splitting brandAccent the ports used; project [dsBrandName] to tint a slice of the wordmark. Card body is the default slot, the support line goes in [dsAuthSupport] (a real <footer>, so it carries the contentinfo landmark). cardElevation / cardPadding pass through to ds-card. Tokens: --ds-auth-shell-min-height, -card-max-width, -gap, -bg-color, -mark-bg-color, -mark-label-color. WORTH KNOWING when porting: the shell owns the full-height box itself, so do NOT wrap it in your own 100vh/100dvh container - the ports that stacked two scrolled a page whose content fits. It also adds env(safe-area-inset-*) padding, and centres with auto margins rather than centre alignment so a form taller than the box stays reachable when the shell is embedded at min-height 100%.",
|
|
172
|
+
"New component ds-filter-bar: a Priority+ filter row - as many filters as fit inline, the rest behind a \"+N more\" panel. Purely additive, NO ACTION. A consumer with a hand-rolled filter row (or one that wraps to two lines on a laptop) can replace it. Each filter goes in an <ng-template [dsFilterItem]=\"180\"> carrying the width it should occupy; the bar decides the split from those numbers plus a ResizeObserver on the row. Trailing controls (search, Clear) carry the dsFilterBarTrailing attribute - mark each control, do NOT wrap them in a div: the bar owns that group's gap and alignment, and a projected wrapper is a plain block whose children stack. Tokens: --ds-filter-bar-gap, --ds-filter-bar-panel-gap. Inputs moreLabel and panelLabel exist for translation. WORTH KNOWING: the declared width is a contract, not a hint - the slot is sized to exactly that number, so declare what the control actually occupies. And because a filter that collapses is MOVED into the panel, its view is destroyed and re-created: internal state such as an open dropdown or focus does not survive that, while its value does (it lives in your signal or form control). Content that reaches the bar with neither slot marker is dropped, with a dev-mode console warning.",
|
|
173
|
+
"ds-table gained virtual scrolling: [virtualScroll]=\"true\" plus virtualRowSize (px). Opt-in and off by default, so NO ACTION. Only the rows in view stay in the DOM (~25 regardless of row count), padded by two spacer rows so the scrollbar still spans the whole data set. REQUIRES maxHeight - a bounded scroller is the viewport it windows against, and without one nothing is virtualized (dev warning). CONSTRAINTS in virtual mode, all enforced rather than assumed: every row is exactly virtualRowSize tall (the CSS height comes from that same input), table-layout becomes fixed, and cells are kept to one line - so declare column widths, and do not rely on wrapping cell text there. Row EXPANSION is suppressed with a dev warning (an expanded row is taller than virtualRowSize and every offset below it would be wrong); pageSize alongside virtualScroll is also warned about, since virtual scrolling is what pagination was for. Sorting, filtering, selection, sticky header, sticky first column and column resizing all keep working. NOT built on the CDK virtual-scroll viewport on purpose: that wraps the table in a transformed content wrapper, and the transform breaks this table's sticky header and sticky first column. STILL OPEN: lazy loading (fetching slices from a server as they scroll into view) is NOT part of this - see the backlog item.",
|
|
174
|
+
"FIXED in ds-table: with stickyHeader and stickyFirstColumn both on, the first body row painted OVER the header in the pinned columns, so the header read as that row's data. --table-sticky-corner-z was already there to keep the corner above the body, but the sticky-header rule out-specified it and both ended up at the same z-index, where DOM order puts the body on top. No action needed - it is a pure CSS fix. If you worked around it with your own z-index on DS table cells, that override can go.",
|
|
175
|
+
"ds-checkbox / ds-switch / ds-radio now RENDER a projected caption instead of dropping it, and expose the caption's colour and size. Additive, NO ACTION - [label] stays the documented API and still wins when both are supplied. If you added a dev-warning workaround after 0.7.1 (moving markup-bearing captions into a string), you can move them back: <ds-checkbox>I accept the <a href=...>Terms</a></ds-checkbox> now works. ONE DIFFERENCE: the required asterisk comes with [label] only - a projected caption owns its own marker, since only your markup knows where it belongs. New hooks: --ds-checkbox-label-color / -label-font-size, the same pair on ds-switch, and --ds-radio-label-font-size (--ds-radio-label-color already shipped). The font-size hook beats the size class. These are deliberately NOT part of the --ds-field-label-* contract: a label above an input is a heading for the field and reads secondary, a caption beside a checkbox is the content being agreed to and reads primary. If you were overriding the caption colour with a descendant selector against .control-text, replace it with the token.",
|
|
176
|
+
"Overlay panels now fade out when DISMISSED and still close instantly when something is CHOSEN. No action - behaviour only. Escape, a press outside, or the page scrolling away fades the panel over --ds-motion-exit-duration (120ms); picking an option, pressing the trigger again, or Tab still closes instantly, because a fade on a deliberate choice reads as the panel lagging behind it. Applies to select, autocomplete, multi-select, date-picker, time-picker, color-picker and popover. Tune with --ds-overlay-leave-duration / --ds-overlay-leave-easing, or set the duration to 0ms to get the old always-instant close everywhere. Reduced motion already zeroes it. A fading panel is pointer-events: none, so it cannot swallow a click aimed at what is behind it. If you built anything that assumed the overlay pane is removed from the DOM synchronously on Escape, note that it now lives for one exit duration longer (the component state - open(), overlayRef - is still cleared immediately).",
|
|
177
|
+
"design-system.md now states the --ds-* token contract once instead of repeating it per component, and carries a warning that its hand-written API and token lists have drifted from the code (ds-button is documented with its pre-color/variant API and the pre-rename --ds-button-bg / -fg names). For the authoritative surface use the auto-generated Component API tables and token reference in ds-docs, which are extracted from source on every build. No action - documentation only."
|
|
178
|
+
]
|
|
179
|
+
}
|
package/migrations/AUTHORING.md
CHANGED
|
@@ -113,6 +113,14 @@ family with "point the channel back at `--primary-h/s/l` to keep the old look".)
|
|
|
113
113
|
**Catch-all:** if you can't confidently fit a breaking/changed entry to a mechanical kind, make it
|
|
114
114
|
`behavior` (never drop it). Reading the changelog completely matters more than perfect kinds.
|
|
115
115
|
|
|
116
|
+
**One rename can need TWO items.** When an input, a `variant` value or a tone also changes the class
|
|
117
|
+
the component puts in the DOM, the `input-rename` / `value-remap` fixes the consumer's *binding* and
|
|
118
|
+
finds nothing in their stylesheet — a selector like `.modal.danger` survives the upgrade and silently
|
|
119
|
+
matches nothing. Emit a `class-rename` alongside it, element-scoped (`ds-modal .danger`), and say in
|
|
120
|
+
the `note` that only consumer STYLING of the internal class is affected. This is how 0.5.1's
|
|
121
|
+
`danger` → `error` vocabulary rename left a consumer with a dead destructive-modal stripe: every
|
|
122
|
+
binding item was present and correct, the class items were missing (added retroactively 2026-08-03).
|
|
123
|
+
|
|
116
124
|
## actionable vs informational
|
|
117
125
|
|
|
118
126
|
Decide by CONTENT, not by which `###` section it sits in — an actionable change can hide in
|
|
@@ -121,6 +129,16 @@ Actionable = existing consumer code stops working or *behaving as before*. Every
|
|
|
121
129
|
components, new optional inputs, additive tokens whose defaults equal the old role token, pure
|
|
122
130
|
bugfixes, visual polish) → `informational`, so a reader knows silence = safe.
|
|
123
131
|
|
|
132
|
+
**Members added to an exported class are actionable even when the feature is additive.** Subclassing
|
|
133
|
+
a DS component is unsupported (`design-system.md` § Components), but a consumer who does it breaks on
|
|
134
|
+
a name collision — TS4114 demands `override`, TS2416 fires when the signature differs — so a release
|
|
135
|
+
that only ADDS can fail their build. Emit ONE `behavior` item per release naming the classes whose
|
|
136
|
+
member surface grew, with the added member names where few enough to list and the count where not,
|
|
137
|
+
`searchHints` that locate the subclass (`extends TableComponent`, `extends `), and a `note` giving the
|
|
138
|
+
resolution: rename the subclass's own member, or add `override` where the base's semantics genuinely
|
|
139
|
+
match. Keep the feature bullets themselves in `informational` — the item is the collision, not the
|
|
140
|
+
feature.
|
|
141
|
+
|
|
124
142
|
## Per-entry checklist
|
|
125
143
|
|
|
126
144
|
- [ ] `id` is a stable, unique kebab slug.
|
package/package.json
CHANGED
|
@@ -92,8 +92,12 @@
|
|
|
92
92
|
flex: 0 0 auto;
|
|
93
93
|
}
|
|
94
94
|
|
|
95
|
-
|
|
96
|
-
|
|
95
|
+
// `:not(.is-selected)` is load-bearing: this branch outranks `.is-selected`
|
|
96
|
+
// on specificity, so without it a hovered (or keyboard-active) selected row
|
|
97
|
+
// repaints to the neutral surface while KEEPING the selected label colour —
|
|
98
|
+
// unreadable as soon as a brand gives the selected row a dark fill.
|
|
99
|
+
&:hover:not(.is-disabled, .is-selected),
|
|
100
|
+
&.is-active:not(.is-disabled, .is-selected) {
|
|
97
101
|
background: var(--surface-secondary);
|
|
98
102
|
}
|
|
99
103
|
|
|
@@ -107,6 +111,11 @@
|
|
|
107
111
|
color: var(--primary-strong);
|
|
108
112
|
|
|
109
113
|
.option-icon { color: var(--primary-strong); }
|
|
114
|
+
|
|
115
|
+
&:hover:not(.is-disabled),
|
|
116
|
+
&.is-active:not(.is-disabled) {
|
|
117
|
+
background: var(--primary-muted);
|
|
118
|
+
}
|
|
110
119
|
}
|
|
111
120
|
|
|
112
121
|
&.is-disabled {
|
|
@@ -0,0 +1,17 @@
|
|
|
1
|
+
// One control-box height for the field family.
|
|
2
|
+
//
|
|
3
|
+
// `--field-h-sm`/`-md`/`-lg` have shipped as the field height scale since the
|
|
4
|
+
// token file was written, but only `-lg` was ever read: every field spelled the
|
|
5
|
+
// other two as the raw `var(--hit-min)` / `var(--hit-cozy)` instead. A consumer
|
|
6
|
+
// retuning `--field-h-md` — the obvious lever, and the one the token name
|
|
7
|
+
// advertises — therefore changed nothing. Same values, so routing through the
|
|
8
|
+
// scale is visually neutral; it just makes the knob real.
|
|
9
|
+
//
|
|
10
|
+
// `--ds-field-height` is the public override and beats the size class, because a
|
|
11
|
+
// consumer setting it has already decided. Per-instance is CSS-variable scoping:
|
|
12
|
+
// `ds-select { --ds-field-height: 44px; }`.
|
|
13
|
+
@mixin field-box-height($size: md) {
|
|
14
|
+
--_field-box-height: var(--field-h-#{$size});
|
|
15
|
+
|
|
16
|
+
height: var(--ds-field-height, var(--_field-box-height));
|
|
17
|
+
}
|
|
@@ -0,0 +1,34 @@
|
|
|
1
|
+
// One label contract for the field family.
|
|
2
|
+
//
|
|
3
|
+
// Five components hard-coded the identical rule and ds-select named its element
|
|
4
|
+
// `.select-label`, so a form mixing a select with an input rendered labels in two
|
|
5
|
+
// sizes and colours out of the box, and retuning them from an app needed
|
|
6
|
+
// `.field.field .label.label, .select-label.select-label` to out-rank the shimmed
|
|
7
|
+
// component rule. One class (`.label`), one token set, resolved here.
|
|
8
|
+
//
|
|
9
|
+
// A MIXIN, not a global rule: a global `.label` would match consumer markup that
|
|
10
|
+
// happens to use the same class name — the leak that global `.field`/`.check`
|
|
11
|
+
// specimens caused in ds-docs. Each field pays ~6 declarations of its own CSS.
|
|
12
|
+
//
|
|
13
|
+
// Per-instance retune is CSS-variable scoping, so no per-component token set is
|
|
14
|
+
// needed: `ds-select { --ds-field-label-transform: uppercase; }` reaches exactly
|
|
15
|
+
// one component. ds-input keeps its older `--ds-input-label-*` trio winning above
|
|
16
|
+
// this, since those shipped in 0.7.x.
|
|
17
|
+
@mixin field-label {
|
|
18
|
+
--_field-label-font-size: var(--ds-field-label-font-size, var(--font-size-xs));
|
|
19
|
+
--_field-label-weight: var(--ds-field-label-weight, var(--font-weight-bold));
|
|
20
|
+
--_field-label-color: var(--ds-field-label-color, var(--text-secondary));
|
|
21
|
+
--_field-label-transform: var(--ds-field-label-transform, none);
|
|
22
|
+
--_field-label-tracking: var(--ds-field-label-tracking, normal);
|
|
23
|
+
|
|
24
|
+
display: flex;
|
|
25
|
+
align-items: center;
|
|
26
|
+
|
|
27
|
+
// --field-gap-req is the label↔asterisk token; the old rules all spelled its
|
|
28
|
+
// value out as var(--space-1) instead of naming it.
|
|
29
|
+
gap: var(--field-gap-req);
|
|
30
|
+
font: var(--_field-label-weight) var(--_field-label-font-size) / var(--line-height-base) var(--font-sans);
|
|
31
|
+
color: var(--_field-label-color);
|
|
32
|
+
text-transform: var(--_field-label-transform);
|
|
33
|
+
letter-spacing: var(--_field-label-tracking);
|
|
34
|
+
}
|